Title

Markers for invasion in breast and lung carcinoma. Field of the Invention

The present invention relates to markers indicative of invasion in cancers, particularly breast and lung carcinomas. These markers find use in the diagnosis/prognosis of invasion (metastasis) and may thus be useful in identification of suitable treatment regiment for patients. The markers may also be of use in identifying potential therapeutic agents or targets for the prevention or reduction of invasion in cancer patients. Background to the Invention Cancer is a multistep process and occurs as a result of the loss of control of cell division, leading to initial tumour formation, which can then be followed by metastatic spread. A distinguishing feature of malignant cells is their ability to invade surrounding normal tissue, metastasize through the blood and lymphatic systems and re-establish at distant secondary locations. To form metastases, individual tumour cells must break from the primary tumour mass, degrade extracellular matrix, invade the surrounding normal tissue, enter the blood or lymphatic circulation, exit the circulation at a distal tissue and establish satellite colonies within this new tissue environment. This behavour requires the cooperative function of numerous proteins. This metastatic spread of solid tumour is responsible directly or indirectly for most cancer-related deaths. However, none of the functions of metastasizing cells are unique to cancer cells. Angiogenesis, nerve growth cone extension and homing, embryogenesis and trophoblast implantation are all examples of physiological invasion. During embryonic development, motile cells are tightly regulated in order to ensure proper homing and reversion to a non-motile phenotype after migration into a destined location. In contrast, cancer cells have lost the ability to recognise specific targets. Their inappropriate growth signals are accompanied by mechanisms to avoid apoptosis and the potential to elicit angiogenesis for independent nutrient supply. Invasion is not simply due to growth pressure but involves additional genetic deregulation over and above those molecular events that cause uncontrolled proliferation. The difference between the normal process and the pathogenic nature of cancer invasion is therefore one of regulation.

Communication between individual cells in multicellular organisms is essential for their regulation and co-ordination of complex cellular processes such as growth, differentiation, migration and apoptosis. The signal transduction pathways mediating these processes are regulated in part by polypeptide growth factors that generate signals by activating cell surface receptors. The primary mediators of such physiological cell responses are receptor tyrosine kinases (RTKs). That is, in most cells, growth factors mediate cellular activity by means of receptors with intrinsic tyrosine kinase activity. It is widely accepted that cancer cells contain genetic damage that leads to tumourigenesis through deregulation of key signalling pathways. Activation of growth factor receptors and their intrinsic tyrosine kinase activity initiates signalling cascades that involve multiple intracellular signalling pathways, such as the phosphatidylinositol 3-kinase (PI3K) and MAPK pathways. These pathways are responsible for the diverse target actions of these growth factors including increased cell division, cell migration, and

inhibition of apoptosis. Malignant cells aπse as a result of a stepwise progression of genetic events that include the unregulated expression of growth factors or components of their signalling pathways (Fang and Richardson, 2005, Toker and Yoeli-Lerner, 2006). In this way growth factors and their receptors have been shown to play a major role in cancer development One of the RTK families involved in cancer development is the epidermal growth factor receptor (EGFR/ErbB) family, of which erbB2 is a member ErbB receptors are expressed in various tissues of epithelial, mesenchymal and neuronal origin, m which they are involved in the control of diverse biological processes such as proliferation, differentiation, migration and apoptosis. erbB2 overexpression serves as a critical component that couples ErbB receptor tyrosine kinases to the migration/invasion machinery of carcinoma cells. Increased levels of erbB2 have been found to increase tumorigenicity and metastatic potential in vivo, and in particular, deregulation of the erbB2 gene has been detected m approximately 25% of human breast cancers and is associated with poor clinical prognosis. Early reports showed erbB2 was amplified and/or overexpressed in approximately 30% of human breast carcinomas (Slamon et al , 1987, Slamon et al , 1989) A study of 163 tumours from patients with different stages of breast cancer were analysed by Marx et al. (1990) in order to evaluate the distribution of erbB-2. The majority of studies have found the most useful prognostic factor in breast cancer is the number of positive auxiliary lymph nodes, indeed as the number of metastatic nodes increases, survival rates decrease and relapse rates increase Marx et al. (1990) found 50% of patients with three or more positive lymph-nodes were erbB-2 positive Breast cancer with 10 or more positive lymph nodes have a poor prognosis with about 30 per cent of patients alive at 5 years after primary surgery alone (Nemoto et al , 1980). Of 942 invasive ductal carcinomas also found significant association between erbB-2 and tumour grade (Quenel et al., 1995).

Recent reports have also revealed that the kinase domain of erbB-2 is somatically mutated in lung adenocarcinomas, suggesting the mutated erbB2 gene may act as an oncogene in human lung cancers Pfeiffer et al. (1996) examined 186 unselected and systemically untreated patients with non-small cell lung cancer (NSCLC) for erbB-2 status erbB-2 was highly expressed in 26% of tumours The expression of erbB-2 was highest in adenocarcinomas Although this overexpression was found to have no correlation with prognosis, Tsai et al (1996) reported that the intrinsic chemosensitivity of NSCLC cells correlated well with the expression of erbB-2, and transfection of erbB-2 cDNA into low erbB-2 expressing NSCLC significantly enhanced chemoresistance to adriamycin, cisplatm, mitomycin C and VP- 16 (Tsai et al, 1996) Kristiansen et a\ (2001) also found overexpression of erbB2 was also found to correlate with disease-stage and chromosomal in non-small cell lung cancer. A more recent study examined 120 primary lung tumours and identified 4% that had mutations within m erbB2, m the adenocarcinoma subtype of lung cancer, 10% of cases had mutations (Stephens et al., 2004). The initiation and progression of cancers is a complex process, involving an accumulation of genetic aberrations in the cell. Microarrays allow examination of whole genomes simultaneously, allowing a complete investigation of the effects of genetic aberrations in many diseases, including cancer. An in vitro study used human lung cancer cell lines with varying degrees of invasion ability and metastatic potential to determine invasion-related genes (Chen et al., 2001). Hundreds of genes were found significant to the

invasive phenotype, several of which had already been associated with invasion. DNA microarrays have also been used to examine RNA profiles of in vivo models, and have proven powerful tools in the determination of markers of clinical significance (Kim et al, 2006).

In the present study, the genetic profiles of a non-invasive MCF7H3 cell line were compared to those of an invasive, erbB2 overexpressing cell line (MCF7H3erbB2). As a results of this work several genes were found to be associated with an invasive phenotype, of these, RPS6KA3,TFPI1 and THBSl were found to have functional effect.

Ribosomal protein S6 kinase, 9OkDa, polypeptide 3 (RPS6KA3) is one of the four p90 ^rek family genes (RSKl to RSK4), it maps to Xp22 and encodes a 90 kDa ribosomal S6 serine/threonine kinase (Guimiot et al, 2004). The RSK genes are a subfamily of mitogen-activated protein kinase-activated protein kinases (MAPKAPKs) and are downstream effectors of mitogen-activated protein kinase (MAPK) (Zhao et al, 1996) that contain two distinct kinase catalytic domains in a single polypeptide chain. Using a specific inhibitor for the RSK family, it was discovered that they played an important role in proliferation, and gene- silencing of RPS6KA3 using RNAi resulted in a 57% decrease in proliferation of MCF7 cells (Smith et al., 2005). A similar study carried out using a prostate cancer cell line confirmed these results (Clarke et al., 2005). Both reports also observed a 50% increase in RSK family members in breast and prostate cancer tissue compared to normal tissue, which would suggest dysregulation in cancer cells. Although no direct association is documented between RPS6KA3 and invasion, many of its substrates have been identified as key players in tumour progression; these include Estrogen receptor α, cyclic AMP response element-binding protein (CREB), c-Fos and nuclear factor-κB (Smith et al., 2005).

Tissue factor pathway inhibitor (TFPIl) is an endogenous anticoagulant protein of the serine protease family TFPIl comprises of three Kunitz type domains flanked by peptide segments. TFPIl is mainly produced by microvascular endothelial cells and pooled in the endothelium (50-80%), plasma (10-15%) and platelets (<2.5%) (Werling et al, 1993; Novotony et al, 1989; Sandset, 1996). High plasma levels of TFPIl have been reported in cancer patients with solid tumours, whereas those with leukaemia and related blood malignancies have normal levels of TFPIl (Lindahl et al., 1989, 1992; Inversen et al, 1998, 2002). However, the theory that high TFPIl levels in cancer were a consequence of activated coagulation was disproved by Inversen (1998) who demonstrated there was no correlation between the two. It was concluded that TFPIl was related to the biology of the disease rather than the degree of coagulation (Inversen et al, 1998).

Thrombospondin-1 (THBSl) is a large (450 kDa) glycoprotein that is released into the extracellular matrix by several cell types (cultured endothelial cells, fibroblasts and monocytes have all been found to synthesize and secrete THBSl). It is synthesised by most cells in culture (Adams, 1997). THBSl expression is increased in response to growth factors, heat shock and hypoxia, and is downregulated in response to IL-I β and TNFα. Overexpression of THBSl has been associated with migration in many cancer tissues. THBSl has been located at the border between tumour and stroma in primary tumours, and from here can contribute to tumour progression or deterioration depending on the THBSl receptor repertoire of the tumour (Bastian et al, 2005). The events that take place during tumour progression

enable the tumour to interact with its stromal environment in ways that enhance its ability to proliferate in the primary site and, in highly malignant tumours, to metastasize to distant sites in the body (Brown et al., 1999). Because each cell expresses a different repertoire of receptors, the composition of the THBSl- containing complexes, and therefore the cellular responses, may vary among different cell types. It is not surprising then that experimental evidence has indicated THBSl can be both pro- and anti-invasive (Qian, 2001 ; Moon et al., 2005) (it has also been found to be adhesive and anti-adhesive, can foster and retard metastasis, stimulate and inhibit angiogenesis and increase and reduce proteolytic activity and fibrinolysis (Bornstein, 1995)). In a related study comparing the genetic profiles of lung carcinoma cell lines (DLKPs)of different levels of invasion, again several genes were found to be associated with an invasive phenotype, of these, SlOOAl 3 was found to have functional effect.

SlOO proteins are small EF hand Ca ²⁺ binding proteins of variable length and sequence that are expressed in a tissue-specific manner. To date more than twenty SlOO family members have been reported (1-3). There are a number of functions associated with SlOO proteins including cell division, motility, secretion, protein synthesis and membrane permeability (4-6). In addition, recent studies have reported associations between SlOO family members and tumour development and progression (7-10). In particular, S100A4, S100A6, SlOOAIwA SlOOAlO have been found to correlate with, and indeed, functionally effect a more aggressive cancer phenotype (7,12,13, 14). In 1996 Wicki et al. identified a new SlOO member termed S100A13 (11). The C-terminus of the S100A13 protein ends with the motif RKK, which is also seen in the metastasis-associated protein S100A4 and SlOOAlO. This protein has since been found to play a crucial role in the release of FGF-I in angiogenesis (15). S100A13 is widely expressed in many types of tissues, with particularly high expression in the thyroid gland. An increase in SlOOAl 3 transcript was found to be associated with a more aggressive invasive phenotype and clustered with several other well- documented cancer/metastasis-related genes. Object of the Invention

One object of the invention was to investigate the key genetic changes in the process of invasion and metastasis of human cancers overexpressing erbB2. This was achieved by examining the genetic profiles of non-invasive MCF7H3 compared to invasive MCF7H3erbB2, using microarray analysis. Key genetic changes in the process of invasion and metastasis of human cancers was also investigated by examining the genetic profiles of DLKP lung carcinoma cell lines of different levels of invasion. A further object of the invention was to identify markers of invasion which would find use as diagnostic or prognostic markers of cancer invasion. The identification of such markers may allow the treatment regimen for individual patients to be determined depending on the likelihood of the patient developing metastases. A further object was to identify markers which would be targets for future anti-invasion interventions and could be used in the development of anti-invasion therapeutic agents which could prevent or retard invasion in cancer. Summary of the Invention

According to the present invention there is provided use as a marker of invasion in cancer of at least one of the oligonucleotide DNA molecules encoding the markers selected from the group consisting of TFPIl ;

RPS6KA3; BACE2; FARPl; MRPS33; C10orf58; HRSP12; GTF2IRD2; MORN2; LOC255654;

VPS45A; GCSH; ATP6V1C1 ; VATl ; LPHNl ; MAP1LC3B; Clorβ l ; Pl 17; NDUFA3; C19orf21 ;

ZNRFl ; RABL5; MRPS6; RBM25; ADCYl ; MGC45438; TMC4; LOC402617; HIST1H2AC; VPS37C;

TMEM46; ClorflOό; TPCNl ; 203326_x_at ; 227929_at; ATP9, GLPlR, KCNJ8, Ioc51159, PLXDCl, S100A13, SCHIPl, a corresponding mRNA molecule, a polypeptide or protein encoded by any one of these mRNA, DNA or cDNA molecules, or an antibody raised against such a polypeptide or protein.

The invention also provides methods of diagnosing cancers with an invasive tendency comprising determining the levels of one or more of the said markers in a patient and comparing that with a control value. The control value may be obtained from a subject who is known not to have cancer. The presence of or an increased level compared to a control of some markers may be indicative of invasion whilst the absence of or a decreased level compared to a control of other markers may be indicative of invasion. This information is found under the heading of 'fold change' in Table 1. This refers to the change in the level of expression of a particular mRNA transcript in an invasive compared to non-mvasive cell line.

Therefore a positive fold change indicates up-regulation in relation to invasion and these genes may be pro- invasive, and a negative fold change indicates down-regulation and such genes may be anti-invasive.

The probe Ids and the gene/cDNA sequence Ids referred to herein and in Table 1 are Affymetrix target sequences. They are available at www.affymetrix.com, and can be found by entering the probeset ID.

Target sequences are generated from sequence and annotation data obtained from multiple databases such as GenBank, RefSeq and dbEST. The target sequence is the portion of a transcript reference sequence that is interrogated by a probe set. The sequences 203326_x_at and 227929_at have are no gene symbols available.

The markers ATP9, GLPlR, KCNJ8, Ioc51159, PLXDCl, S100A13, SCHIPl are all markers of invasion in lung carcinoma while the remainder of the markers are markers of invasion in breast carcinomas.

TFPIl and RPS6KA3 are markers of invasion in both carcinomas. The invention also provides a diagnostic assay of invasion in cancer of at least one of the oligonucleotide

DNA molecules encoding the markers selected from the group consisting of TFPIl ; RPS6KA3; BACE2;

FARPl ; MRPS33; C10orf58; HRSP12; GTF2IRD2; M0RN2; LOC255654; VPS45A; GCSH;

ATP6V1C1 ; VATl ; LPHNl ; MAP1LC3B; ClorG l ; Pl 17; NDUFA3; C19orf21 ; ZNRFl ; RABL5;

MRPS6; RBM25; ADCYl ; MGC45438; TMC4; LOC402617; HIST1H2AC; VPS37C; TMEM46; ClorflOό; TPCNl; 203326_x_at ; 227929_at; ATP9, GLPlR, KCNJ8, Ioc51159, PLXDCl, S100A13,

SCHIPl, a corresponding mRNA molecule, a polypeptide or protein encoded by any one of these mRNA,

DNA or cDNA molecules, or an antibody raised against such a polypeptide or protein.

In a still further aspect the invention provides use in a method of identifying therapeutic agents which can prevent or retard invasion in cancer, of at least one of the oligonucleotide DNA molecules encoding the markers selected from the group consisting of TFPIl ; RPS6KA3; BACE2; FARPl ; MRPS33;

C10orf58; HRSP12; GTF2IRD2; M0RN2; LOC255654; VPS45A; GCSH; ATP6V1C1; VATl ;

LPHNl; MAP1LC3B; Clorβl; Pl 17; NDUF A3; C19orf21; ZNRFl; RABL5; MRPS6; RBM25;

ADCYl; MGC45438; TMC4; LOC402617; HIST1H2AC; VPS37C; TMEM46; ClorflOό; TPCNl ;

203326_x_at ; 227929_at; ATP9, GLPlR, KCNJ8, Ioc51159, PLXDCl , S100A13, SCHIPl, a corresponding mRNA molecule, a polypeptide or protein encoded by any one of these mRNA, DNA or cDNA molecules, or an antibody raised against such a polypeptide or protein. The invention also provides a method of identifying therapeutic agents which can prevent or retard invasion in cancer comprising contacting a cancer cell or cell line, or an animal suffering from cancer with a test therapeutic agent and determining the level of expression of one or more of the above- mentioned markers. Alternatively the method may comprise contacting a polypeptide or protein of one or more of the above-mentioned markers with a test therapeutic agent and determining if the agent modulates the activity of the protein or polypeptide, when compared with a control. Preferably in the above uses, methods or assays, more than one marker is used. For example at least five or at least ten and more preferably all of the markers are used.

Particularly preferred for use in the assays, methods or uses of the invention for invasion in lung carcinoma are the markers ATP9, GLPlR, KCNJ8, Ioc51159, PLXDCl, S100A13, SCHIPl, TFPIl and RPS6KA3. Particularly preferred for use in the assays, methods or uses of the invention for invasion in breast carcinoma are the markers TFPIl ; RPS6KA3; BACE2; FARPl ; MRPS33; C10orf58;

HRSP 12; GTF2IRD2; M0RN2; LOC255654; VPS45A; GCSH; ATP6V1C1 ; VATl; LPHNl; MAP1LC3B; Clorβl; P117; NDUFA3; C19orf21; ZNRFl; RABL5; MRPS6; RBM25; ADCYl; MGC45438; TMC4; LOC402617; HIST1H2AC; VPS37C; TMEM46; ClorflOό; TPCNl; 203326_x_at ; and 227929_at. The assay may be a real-time PCR assay, a customised micro-array assay or a histochemical assay. All such assays are well known to those of skill in the art. Where the assay is a histochemical assay, the antibody may be labeled with a suitable label. Suitable labels include coloured labels, fluorescent labels and radioactive labels. The invention also provides a solid support onto which one or more of the DNA sequences, mRNAs, polypeptides, proteins or antibodies as described above, have been fixed. The invention also provides diagnostic kits for invasion in cancer comprising marker sequences, DNAs, mRNAs, polypeptides, proteins or antibodies as described above.

In a still further aspect the invention provides a method of prevention or treatment of invasion in cancer comprising administering to a patient in need of such treatment, an inhibitor of the protein product of a pro-invasive sequence shown in Table 1, or an agent which can silence a sequence shown in Table 1. The agent which silences the gene may by an siRNA molecule directed against any of the mRNA sequences or an antibody directed against the protein product of any of the mRNA sequences. The invention also provides a method of treatment of invasion in cancer comprising administering to a patient in need of such treatment, an activator of a marker sequence or the protein product of an anti- invasive marker sequence shown in Table 1. Detailed Description of the Invention

Cell lines DLKP is a poorly differentiated human lung squamous cell line, previously described (Law et al, 1992; McBride et ah, 1998). DLKP4E is a clonal subpopulation of eIF4E cDNA transfected DLKP.

Two invasive, non-drug exposed, variants of the DLKP cell line (KP1,KP2) were also used, along with a non-invasive, Vincristine-resistant (VCR) variant, an invasive Taxotere-resistant DLKP variant (TX2) and a non- invasive Taxotere-resistant line (TXl).

MCF7 human breast adenocarcinomas, cultured as described previously (Favre et ah, 1994). MCF7pcDNA is a clonal subpopulation of pcDNA transfected MCF7, MCF74E is a clonal subpopulation of eIF4E cDNA transfected MCF7, and MCF74Emut is a clonal subpopulation of eIF4E mut cDNA transfected MCF7. MCF7H3 is a clonal subpopulation of MCF7, and MCF7H3erbB2 is a clonal subpopulation of MCF7H3 transfected with erbB2. SKB R3 is an erbB2-positive, human breast adenocarcinoma, cultured as described by Miller et ah, 1994. Arrays The microarray gene expression experiments which were performed in this body of work were performed using Affymetrix® Ul 33 Plus 2 GeneChips ®. Arrays were used as per manufactures instruction. DChip DNA-Chip Analyzer (dChip) is a software package implementing model-based expression analysis of oligonucleotide arrays (Li and Wong, 2001) and several high-level analysis procedures. High- level analysis in dChip included comparative analysis and hierarchical clustering. Data normalisation as carried out by dChip was downloaded from (http://dchip.org/) along with other data analysis modules. Gene list Generation for TFPIl, THBSl, RPS6KA3, EGRl, TNFAIP8 study

The first step in analysis of microarray data was to identify differentially expressed genes. Average gene expression values were obtained for all probesets for each cell line, and these values were compared to identify genes whose expression levels were significantly different between two cell lines. Genes that were differentially expressed were uncovered using parameters of p-value of <0.05, fold change of 1.2, and expression level > 100. The initial comparison of MCF7H3 versus MCF7H3erbB2, produced a list of 3,349 differentially expressed genes. In order to reduce the number of genes for further analysis, genes differentially expressed in MCF7H3 compared to MCF7 were removed. This was done under the assumption that differential expression in this case was only due to clonal variation, and therefore it was unlikely that these genes were involved in invasion. Similarly, genes differentially expressed in MCF74E and MCF74Emut compared to MCF7 were also removed. The reason for this was neither MCF74E nor MCF74Emut were invasive, and it was assumed that these genes would not play a significant role in the invasive phenotype of MCF7erbB2. The final list of 120 genes specific to MCF7H3erbB2 and invasion contained some overlap due to different probe sets targeting various gene transcripts. Further examination found there were 108 different genes on this list.

Bioinformatics The 108 genes identified through different gene list comparisons (as described above) to be specific to MCF7H3erbB2 and invasion, were further analysed using Pathway Assist®. PathwayAssist is a product aimed at the visualisation and analysis of biological pathways, gene regulation networks and protein interaction maps. It comes with a comprehensive database that gives a snapshot of all information available in PubMed, with the focus on pathways and cell signalling networks. PathwayAssist worked by identifying relationships among genes, small molecules, cell objects and processes and built pathways based on these relationships. Gene list generaήon for S100A13 study:

Transcriptional profiling of a panel of invasive lung cancer cell lines was performed previously in our laboratory using U133A Human Genome Chips (Affymetrix) to identify differentially expressed genes. The cell lines profiled were variants of the poorly differentiated human lung, squamous cell line, DLKP. These cell lines had differing levels of invasiveness, but were derived from the same parent offering a unique opportunity to study the less-well characterised mechanisms of invasion. Two invasive, non-drug exposed, variants of the DLKP cell line (KP1,KP2) were used in the study along with a non-invasive, Vincristine- resistant (VCR) variant, an invasive Taxotere-resistant DLKP variant (TX2) and a non-invasive Taxotere- resistant line (TXl). These variants displayed different degrees of invasiveness when assayed in an in vitro Matrigel invasion assay ranging from non/weakly to strongly invasive. They were assigned a numerical value denoting their relative invasiveness as follows: VCR(O)<TX1(1)<KP2(3.5)<TX2(5)=KP1(5), where VCR is the least and TX2 and KPl are most invasive. This scale was applied to reflect the intermediate degrees of the phenotype in an attempt to gain insight into the more subtle transcriptional changes associated with this transition. In the analysis, genes that did not cross a two-fold threshold above or below the median in at least one cell line were removed from the analysis, as were genes that were consistently called absent by the Affymetrix software. A Welch ANOVA statistical test was used to find genes that were reliably differentially expressed between the five different cell lines. Subsequently, a further Welch ANOVA test was used to find genes in this list that were useful in discriminating between more invasive and less invasive samples. Normalisation and quality control of microarray experiments In order to compare gene expression results from experiments performed using multiple chips, it was necessary to normalise the data obtained following scanning. The purpose of data normalisation was to minimise the effects of experimental and technical variation between microarray experiments so that meaningful biological comparisons could be drawn from the data sets and that real biological changes could be identified. After normalisation using dChip, all data from microarray chips went through several QC steps. Variability caused by several factors in the fields of experimental design, experimental setup, image analysis and data analysis, disguises actual differences in signal intensities and highlights the necessity for quality control. Once satisfactory the quality of the array data was measured using hierarchical clustering. THBSl, TFPIl, TNFAIP8, EGRl, RPS6KA3 siRNA transfection The siRNAs used were chemically synthesized and purchased from Ambion Inc. These siRNAs were 21-23 bps in length and were introduced to the cells via reverse transfection with the transfection agent siPORT™ iVeoFX™ (Ambion inc., 4511). Cell suspensions were prepared at 7x10 ⁵ cells per ml for MCF7, and 3xl0 ⁵ cells per ml for SKBR3, DLKP and DLKP4E (in a 6-well plate). Volume of siPORT™ NeoFX™ per well was 6μl for MCF7, and 2 μl for DLKP, DLKP4E, and SKBR3 S100λ13 siRNA transfection: The optimised conditions for transfection of three independent siRNA sequences to S100A13 (Ambion ID 42125, 263271, 263272) into the invasive cell lines were 2μl NeoFx (Ambion) to transfect 3OnM siRNA in a cell density of 3x10 ⁵ cells per well of a 6- well plate. Solutions of siRNA at a final concentration of 3OnM were prepared in optiMEM (Gibco™, 31985). NeoFX solution was prepared in optiMEM and incubated at room temperature

for 10 minutes. After incubation, an equal volume of NeoFX/optiMEM solution was added to each siRNA/optiMEM solution. These solutions were mixed and incubated for a further 10 minutes at room temperature. Replicates of lOOμl of the siRNA/NeoFX solutions were added to a 6-well plate. ImI of each cell suspension was added to each well at a concentration of 3xlO ⁵ cells per well. The plates were mixed gently and incubated at 37°C for 24 hours. After 24 hours, the transfection mixture was removed from the cells and the plates were fed with fresh medium. QPCR Total RNA was isolated from cells using RNeasy Mini Prep Kit® (QIAGEN, 74104). The Taqman® Real time PCR analysis was preformed using the Applied BioSystems Assays on Demand PCR Kits, using a primer probe pair specific to THBSl (Applied Biosystems, HS00962914_ml). Experiments were preformed in triplicate, following per manufacturer's instructions. The level of mRNA as detected by qPCR was expressed in terms of Relative Quantification. Relative quantification determines the changes in steady-state mRNA levels of a gene across multiple samples and expresses it relative to the levels of an internal control RNA (in this case GAPDH). Western blot Denaturing polyacrylamide gel electrophoresis (SDS-PAGE) followed by immunoblotting for the detection of the antigens recognised by THBSl, TFPIl, RPS6KA3, EGRl and SlOOAl 3 antibodies were performed by the methods of Laemmli (1970) and Towbin et al. (1979) as outlined previously (Moran et al., 1998). Invasion assays Invasion assays were carried out using BD BioCoat™ Growth Factor Reduced MATRIGEL™ Invasion Chambers (BD Biosciences, 354483). Inserts were rehydrated as specified by manufacturers protocol. Cell suspensions were prepared in culture media containing 5% FCS at a concentration of 1 X 10 ⁶ cells /ml. 500μl of Media containing the same concentration of FCS was added to the well of the BD Falcon™ TC Companion Plate. lOOμl of cell suspension was then added into the insert. The invasion assays were then incubated for 48 hours at 37°C, 5% CO ₂ atmosphere. For SKBR3 cell suspensions were prepared in culture media without serum, and 500ul of Media containing 10% FCS was added to the well of the BD Falcon™ TC Companion Plate. The invasion assays were then incubated for 72 hours at 37°C, 5% CO ₂ atmosphere.

CDNA Sl OQAl 3 cDNA was obtained from Open Biosystems (MHSlOl 1-59294), sub-cloned into pSPORTβ mammalian expression vector and sequenced. Transient cDNA transfection experiments were performed with 5μl Lipofectamine2000 (Life Technologies) and 2μg plasmid in a cell density of 4x10 ⁵ per well of a 6-well plate. Medium was removed after 24 hours and replaced with fresh growth medium. Results Pathway analysis THBSl. TFPlI, RPS6KA3.EGR1, TNFλIP8 study It seemed plausible that a network of directly interacting genes that existed within a group of genes already chosen based on their association with an invasive phenotype, was most likely an invasion-relevant pathway. This

information could, in turn, lead to the discovery of novel genes and/or pathways associated with invasion/metastasis. Pathway Assist® was used to identify what genes, if any, from the list of 108, had direct biological interaction with each other, or previously annotated pathways. Interactions derived from Pathway Assist ® analysis are based on information available in the literature to date, and can only take into consideration well-annotated genes. The result of this analysis was a 9-gene pathway. This pathway showed MAP3K1 was involved in the regulation of RPS6KA3, ESRl, TNFAIP8 and TANK. It also showed MAP3K1 was capable of binding TANK. RPS6KA3 was shown to be involved in regulation of ESRl, possible through binding. ESRl was also found to effect RPS6KA3 expression. PTEN was found to regulate TNFAIP8. EGRl positively regulated PTEN, and was itself bound by EGR3 and positively regulated by ESRl .

Further examination of the literature found two members of this pathway, PTEN and EGRl, interacted with Thrombospondin 1 (THBSl). This gene was not present on the final list but was on the original list of MCF7H3 versus MCF7H3erbB2, with a fold change of- 2.31. What was most interesting about this gene was it linked the 9-gene pathway to tissue factor pathway inhibitor (TFPIl), the gene with the greatest increase of expression (19.77 fold) on the final list of 108 genes. TFPIl was chosen as a target for siRNA knock-down based on its large fold change, and THBSl was chosen because of its association with TFPIl. The other targets chosen for further analysis were tumour necrosis factor alpha-induced protein 8 (TNF AIP8), early growth response 1 (EGRl) and ribosomal protein S6 kniase, 9OkDa, polypeptide 3 (RPS6KA3). All 3 genes were contained within the 9-gene pathway identified by PathWay Assist. The genes chosen for functional analysis are shown in Table 2.

Table 2: Genes specific to MCF7H3erbB2 and related to invasion chosen for further analysis

Further literature searches found 39 of the 108 genes were related to invasion, or processes relevant to invasion. The novel invasion-associated genes are shown in Table 1.

Table 1: Table 1:

HIST1H2AC histone cluster 1, H2ac 215071 s at NOVEL 2.05 0.006756 histone family, member ASSOCIATED WITH L /DEF=Human DNA COLON CANCER sequence from clone RP1-221C16 on chromosome 6. Contains two genes for novel histone 4 family members, two genes for novel histone 1 family members, three genes for novel histone 2B family members, a gene for a novel histone 2A family me... chromosome 20 open 225505 s at NOVEL (no link to 1.71 0.048837 reading frame 81 invasion/metastasis)

FARPl FERM, RhoGEF 201911 s at NOVEL (Differentially 1.68 0.002722 This gene was originally isolated through subtractive http ://scholar.£oog

(ARHGEF) and Expressed After TGFBl hybridization due to its increased expression in le.com/url?sa=U& pleckstrin domain Treatment) Pancreatic differentiated chondrocytes versus dedifferentiated q=http://doi.wiley. protein 1 (chondrocyte- chondrocytes. The resulting protein contains a com/10.1002/gcc.l derived) predicted ezrin-like domain, a DbI homology domain, 0179 and a pleckstrin homology domain. It is believed to be a member of the band 4.1 superfamily whose members link the cytoskeleton to the cell membrane. Two alternatively spliced transcript variants encoding

distinct isoforms have been found for this gene.

No gene gb:M76729.1 =Human 203326 x at NOVEL 1.67 0.028377 symbol pro-alpha-1 (V) collagen available mRNA, complete cds. /PROD=pro-alpha-l type V collagen /FL=gb:D90279.1 gb:NM 000093.1 gb:M76729.1

MRPS33 mitochondrial 218654 s at NOVEL 1.57 0.000997 Mammalian mitochondrial ribosomal proteins are ribosomal protein S33 encoded by nuclear genes and help in protein synthesis within the mitochondrion. Mitochondrial ribosomes (mitoribosomes) consist of a small 28S subunit and a large 39S subunit. They have an estimated 75% protein to rRNA composition compared to prokaryotic ribosomes, where this ratio is reversed. Another difference between mammalian mitoribosomes and prokaryotic ribosomes is that the latter contain a 5S rRNA. Among different species, the proteins comprising the mitoribosome differ greatly in sequence, and sometimes in biochemical properties, which prevents easy recognition by sequence homology. The 28S subunit of the mammalian mitoribosome may play a crucial and characteristic role in translation initiation. This gene encodes a 28S subunit protein that is one of

covalently bound at the cell membrane. Latrophilin-1 has been shown to recruit the neurotoxin from black widow spider venom, alpha-latrotoxin, to the synapse plasma membrane.

MAP1LC3B microtubule-associated 208786 s at NOVEL The product of 1.35 0.011027 The product of this gene is a subunit of neuronal http ://scholar.goog protein 1 light chain 3 this gene is a subunit of microtubule-associated MAPlA and MAPlB proteins, le.com/url?sa=U& beta neuronal microtubule- which are involved in microtubule assembly and q=http://doi.wileγ. associated MAPlA and important for neurogenesis. Studies on the rat homolog com/10.1002/iic.22 MAPlB proteins, which implicate a role for this gene in autophagy, a process 030 are involved in that involves the bulk degradation of cytoplasmic microtubule assembly component. and important for neurogenesis. Studies on the rat homolog implicate a role for this gene in autophagy, a process that involves the bulk degradation of cytoplasmic component.Ref: up- regulated in breast cancer cells after treatment.

Clorβl Chromosome 1 open 225638 at NOVEL 1.32 0.035701 reading frame 31

K*

O

ZNRFl zinc and ring finger 1 223382 s at NOVEL may be involved 1.21 0.031147 In a study identifying genes in rat that are upregulated Araki T, in ubiquitin-mediated in response to nerve damage, a gene which is highly Nagarajan R, protein modifϊcation(In expressed in ganglia and in the central nervous system Milbrandt J. 2001. a study identifying genes was found. The protein encoded by the rat gene Identification of in rat that are contains both a zinc finger and a RING finger motif and genes induced in upregulated in response is localized in the endosome/lysosome compartment, peripheral nerve to nerve damage, a gene indicating that it may be involved in ubiquitin-mediated after injury. which is highly protein modification. The protein encoded by this Expression expressed in ganglia and human gene is highly similar in sequence to that profiling and in the central nervous encoded by the rat gene. novel gene system was found. The discovery. J Biol protein encoded by the Chem 276:34131- rat gene contains both a 34141. zinc finger and a RING finger motif and is localized in the endosome/lysosome compartment, indicating that it may be involved in ubiquitin-mediated protein modification. The protein encoded by this human gene is highly similar in sequence to that encoded by the rat gene. XZNRFl, also called PNIP or peripheral nerve injury protein, a protein

identified in the course of a search for proteins upregulated, and with potential functions, in response to chemical and excitotoxic nervous damage [Araki et al., 2001].

ZNRFl and ZNRF2 are associated with presynaptic vesicles and the presynaptic plasma membrane, respectively,

K* K* and both have been shown to inhibit Ca2p- dependent exocytosis dependent on their ubiquitin-ligase activities [Araki and Milbrandt, 2003].)

RABL5 RAB, member RAS 222742 s at NOVEL (The RAS 1.2 0.015718 http ://scholar.£OO2 oncogene family-like 5 superfamily of GTP le.com/url?sa=U& (guanosine triphosphate) α=http://stke.scien hydrolysis-coupled cemae.org/cai/cont signal transduction relay ent/abstract/2004/ proteins can be 250/rel3 subclassified into RAS, RHO, RAB, and ARF families, as well as the

closely related G family. The members of each family can, in turn, be arranged into evolutionarily conserved branches. These groupings reflect structural, biochemical, and functional conseryation. Recent findings have provided insights into the signaling characteristics of representative

K* members of most RAS superfamily branches. )

MRPS6 mitochondrial 224919 at NOVEL Mammalian 1.2 0.041366 Mammalian mitochondrial ribosomal proteins are http ://scholar.goog ribosomal protein S6 mitochondrial ribosomal encoded by nuclear genes and help in protein synthesis le.com/url?sa=U& proteins are encoded by within the mitochondrion. Mitochondrial ribosomes α=http://doi.wilev. nuclear genes and help (mitoribosomes) consist of a small 28S subunit and a eom/10.1002/pros. in protein synthesis large 39S subunit. They have an estimated 75% protein 20308 within the to rRNA composition compared to prokaryotic http://cancerres.aa mitochondrion (Ref: ribosomes, where this ratio is reversed. Another crjournals.org/cgi/ differentially expressed difference between mammalian mitoribosomes and reprint/67/1/139 in p53 gain-of function prokaryotic ribosomes is that the latter contain a 5S LNCaP androgen- rRNA. Among different species, the proteins comprising dependent prostate the mitoribosome differ greatly in sequence, and carcinoma cells) sometimes in biochemical properties, which prevents prostate, pancreatic easy recognition by sequence homology. This gene encodes a 28S subunit protein that belongs to the

K*

4-

TMC4 transmembrane 226403 at Novel -1.56 0.003781 Mutations of TMCl cause deafness in humans and channel-like 4 mice. TMCl and a related gene, TMC2, are the founding members of a novel gene family. Here we describe six additional TMC paralogs (TMC3 to TMC8) in humans and mice, as well as homologs in other species. cDNAs spanning the full length of the predicted open reading frames of the mammalian genes were cloned and sequenced. All are strongly predicted to encode proteins with 6 to 10 transmembrane domains and a novel conserved 120-amino-acid sequence that we termed the TMC domain. TMCl, TMC2, and TMC3 comprise a distinct subfamily expressed at low levels, whereas TMC4 to TMC8 are expressed at higher levels in multiple tissues. TMC6 and TMC8 are identical to

K* Ul the EVERl and EVER2 genes implicated in epidermodysplasia verruciformis, a recessive disorder comprising susceptibility to cutaneous human papilloma virus infections and associated nonmelanoma skin cancers, providing additional genetic and tissue systems in which to study the TMC gene family.

No gene Homo sapiens, clone 227929 at NOVEL 3.83 0.003001 UNKNOWN symbol IMAGE:5277945, available mRNA

K*

K*

siRNA analysis of genes found differentially expressed in invasive MCF7H3erbB2 compared to noninvasive MCF7H3.

Microarray analysis found that in MCF7H3erbB2, EGRl, TNFAIP8, TFPIl, RPS6KA3 were up- regulated compared to the non-invasive MCF7H3 parent (suggesting a potential pro-invasive role for these genes.), and THBSl was down-regulated (suggesting a potential anti-invasive role). In order to examine the functional effect of these genes in relation to invasion, RNA interference was used to silence the genes in the invasive cell lines DLKP4E and SKBR3, and the resulting effect on invasion observed. Western Blot and qPCR were carried out post - transfection of two to three non-homologous siRNAs to determine the success of siRNA silencing of the target genes. EGRl EGRl siRNA A, B and C were all effective in silencing EGRl mRNA in SKBR3, but in DLKP4E, only siRNA C was effective.

Western blots were carried out using an EGRl specific antibody to detect if EGRl siRNA transfection had had an effect at protein level .Results showed that there was a decrease in EGRl at a protein level in both cell lines. DLKP4E transfected with all 3 EGRl siRNAs showed considerable protein knock-down compared to the non-transfected and scrambled controls. In SKBR3 a reduction in EGRl protein was seen as a result of two non-homologous siRNAs, also indicating efficient silencing. Results from invasion assays showed a significant reduction in the number of invading cells after transfection with EGRl siRNAs. Both photographic evidence and cell counts show that the number of invading cells was halved after EGRl siRNA transfection in DLKP4E. Similar results are also true of SKBR3.

TNFAIP8 TNFAIP8 could be examined only using invasion assays, as an antibody was not commercially available due to the novelty of the target. However, both the photographs of the invasion inserts and the cell counts show a considerable decline in the number of invading cells in DLKP4E. DLKP4E cells transfected with TNFAIP8 siRNA A and B were 60%, and C 50% less invasive than DLKP4E transfected with a scrambled control. Results for SKBR3 siRNA A and B were less impressive, with cell counts showing a 20% to 30% reduction in invasive cells. TNFAIP8 C, with a 65% drop in the number of invading cells, was the most considerable change in SKBR3 cells. Statistical analysis showed changes observed in SKBR3 were significant RPS6KA3 Analysis of microarray data found that RPS6KA3 was increased in both the invasive cell lines, MCF7H3erbB2 and DLKP4E, compared to their non-invasive parent cell lines. A fold change increase of +2.36 (p value 0.006) was seen in MCF7H3erbB2, and +1.2 (p value 0.02) in DLKP4E. To further assess the role of RPS6KA3 in invasion, siRNA was used to target the gene in invasive erbB2 overexpressing (SKBR3) and eIF4E overexpressing (DLKP4E) cell lines. Two non-homologous siRNAs were used for all experiments, both of which had been validated by Ambion

The success of siRNA silencing was measured using both real-time PCR and western blot analysis. RPS6KA3 siRNA succeeded in reducing RPS6KA3 at a protein level, which is evident from Western blot analysis of DLKP4E and SKB R3 transfected with two non-homologous RPS6KA3 siRNAs.

The largest observed decrease in RPS6KA3 mRNA occurred in SKBR3. 48hr post siRNA B transfection, real-time PCR results showed a 50% decrease in mRNA. A 25% decrease in RPS6KA3 at 24hrs was also observed in DLKP4E.

Results from invasion assays showed a considerable reduction in the number of invading cells after transfection with both RPS6KA3 siRNAs A and B. Photographic evidence, along with cell counts show the number of invading cells was reduced by at least 50% after RPS6KA3 siRNA transfection in DLKP4E. SKBR3 also had a dramatic reduction in invading cells after RPS6KA3 siRNA transfection, again with greater than 50% fewer invading cells. The results here show clearly that knock-down of RPS6KA3 caused a decrease in invasion in both cell lines. Proof of knock-down of RPS6KA3 at a protein level, combined with a decrease in invasion after siRNA transfection, validates array analysis which implicated a role for RPS6KA3 in the invasion process. It also implicates a direct role for RPS6KA3 protein in the invasion process.

THBSl THBSl, was down-regulated by -2.3 fold in MCF7H3erbB2 compared to the non-invasive MCF7H3 parent, suggesting a possible anti-invasive role for THBSl. Four cell lines were subsequently chosen to examine the effect of THBSl specific siRNA on in vitro invasion. These were: MCF7 (non-invasive), DLKP (mildly invasive), SKBR3 (invasive) and DLKP4E (invasive). Western Blotting and qPCR were carried out post-transfection of three non-homologous siRNAs to determine the efficacy of siRNA silencing of THBSl. DLKP4E, the most highly invasive cell line, showed 20% to 50% reduction in THBSl mRNA 48hrs after transfection with siRNA A, B and C. In SKBR3 both siRNA A and B caused an 80% to 85% decline in THBSl mRNA after 24hrs, with siRNA C also showing a 40% decrease . THBSl mRNA levels were reduced by 75% to 80% in DLKP at both 24 and 48hr after transfection of all three siRNA's . MCF7, the only non-invasive cell line used in the study, showed a 25% to 40% reduction in mRNA after transfection with all three siRNAs . THBSl protein levels were found to closely correlate with the level of invasion in the untransfected cell lines in that it was difficult to detect in DLKP, DLKP4E and SKBR3, but could be detected in MCF7, where a considerable decrease was observed post-siRNA transfection. THBSl siRNA A and C had the most marked effect in MCF7. After transfection of THBSl siRNA B a protein band was still visible, but clearly reduced compared to the non-transfected and scrambled controls. Protein results mirrored those found at mRNA level in MCF7. Invasion assays were carried out to establish whether THBSl downregulation led to a change in the invasion status of the cells. The most dramatic results were seen in DLKP and MCF7. DLKP, a mildly invasive cell line, showed a 3.5 to 4-fold increase in the number of invading cells when transfected with all three siRNAs. MCF7, a completely non-invasive cell line, became invasive after transfection with THBSl siRNA. SKBR3 showed an increase in invading cells, with cell counts revealing a 1.3 to 1.7 fold increase in THBSl siRNA transfected cells. DLKP4E, based on counts of invading cells, showed a 1.08 fold increase in invasion, a change which was found statistically significant (p-value <0.05). These results demonstrate the dramatic increases in invasion across SKB R3, DLKP and MCF7 caused by transfection of THBSl siRNA, and validate microarray analysis which identified THBSl as a suppressor of invasion.

TFPIl TFPIl was chosen as a target for siRNA based on MCF7H3erbB2 array data analysis that showed a +19.77 fold change in MCF7H3erbB2 (invasive) compared to MCF7H3 (non-invasive). The functional effects of TFPIl were examined in both an erbB2 (SKB R3) and an eIF4E (DLKP4E) overexpressing cell line. Real-time PCR carried out on TFPIl siRNA A, B and C in both DLKP4E and SKBR3 showed significant knock-down of TFPIl niRNA compared to scrambled siRNA trans fections. Therefore, transfection of all 3 TFPIl siRNAs into DLKP4E and SKBR3 resulted in silencing of TFPIl at mRNA level, with overall much greater effect in SKBR3 cells.

72hrs after transfection with TFPIl siRNA, cells were assayed for invasion. DLKP4E results showed a reduction in the number of invading cells when transfected with all 3 TFPIl siRNAs. SKBR3 transfected with TFPIl siRNA showed a dramatic decrease in invasion. Up to an 80% reduction in invading cells was observed after TFPIl siRNA A and B transfection. This result combined with those from real-time PCR implies that siRNA silencing of TFPIl in DLKP4E and SKBR3 decreased invasion. siRNA and cDNA analysis of genes found differentially expressed in invasive DLKP compared to noninvasive DLKP. Array data indicated that SlOOAl 3 expression was increased in the more invasive cell lines in a panel of lung cancer cell lines. Analysis of SlOOAl 3 mRNA by RT-PCR confirmed increased expression in the invasive cell lines P2 and TX2 compared with the weakly invasive cell lines TXl and VCR. S100A13 Having established that SlOOAl 3 transcript levels correlate with a more invasive phenotype, SlOOAl 3 mRNA was knocked down in an attempt to identify a functional role for the protein. Three independent siRNA sequences were transiently transfected into the more invasive cell line, TX2, and target mRNA depletion was monitored over a 72hr period by quantitative RT-PCR. It was important to establish extended suppression of the gene in order to minimise non-specific effects of the transfection procedure in the subsequent invasion assay. Not unexpectedly, measuring the invasive ability of cells that have been recently exposed to transfection reagent will impact greatly on this phenotype. We found 72hrs to be the minimum recovery time required by transfected cells before being transferred to invasion chambers. Transient transfection with 3 independent siRNAs resulted in SlOOAl 3 -specific knockdown of between 60-80% in TX2 cells over this timeframe. Two of the siRNAs maintained significant levels of knockdown up to 72hrs with the third less effective at that stage. These oligos were found to be equally effective in all three invasive cell variants over this period. Cell lines Pl, P2 and TX2 were transiently transfected with S100A13-specific siRNA and scrambled control oligos and allowed to recover for 72 hours. The cells were then assayed for their ability to migrate through Matrigel™ in a chamber-based in vitro invasion assay. After 24 hrs the inserts were stained and invading cells counted. A decrease of up to 80% was seen in cell lines that were depleted in SlOOAl 3 . The pattern of reduction correlated closely with the level of transcript knockdown obtained with each siRNA. It is worth noting the impact of the treatment itself on the phenotype (No treatment v scrambled) hence the long recovery time before performing the functional assay. Delaying the assay further (96 hrs) might reduce this non-specific effect further but there is a risk that the knockdown would 'wear off and Sl 0OA 13 expression would have recovered by then.

Discussion

Of the 108 genes found to be related to invasion and specific to MCF7H3erbB2 in one microarray analysis, 39 were found to have been previous linked to invasion in the literature. RPS6KA3, TFPIl, EGRl, THBSl and TNFAIP8, were chosen for functional analysis. These genes not only appeared on the final list of 108, but they also appeared in a network of genes found to have direct interaction with each other, as generated by Pathway assist®. Of the 108 genes specific to MCF7H3erbB2 and related to invasion, 9 were found to have direct biological interaction with each other. This network was generated based on the idea that a pathway existing within a group of genes chosen based on their association with an invasive phenotype, was likely an invasion-relevant pathway. Recent studies have shown that up- regulation of erbB2 in breast cancer causes activation of the AKT/mTOR/4EBPl pathway, which results in the release of eIF4E from inhibitory binding proteins. Therefore it is possible cells over-expressing erbB2 or eIF4E may contain similar mechanisms and/or novel markers for invasion, and so target genes were examined in both eIF4E and erbB2-overexpressing cell lines. A Welch ANOVA statistical test was used to find genes that were reliably differentially expressed between the five different DLKP cell lines. Subsequently, a further Welch ANOVA test was used to find genes in this list that were useful in discriminating between more invasive and less invasive samples. This generated a gene list containing 45 genes. In this dendrogram the cell line clusters are shown along the horizontal with decreasing invasiveness from left to right. It is noteworthy that there is a visible distinction between the 2 most invasive and the 2 least invasive cell line variants but that the P2 variant seems to straddle both cohorts, in this cluster at least. Within this cluster we identified several genes that have been previously associated with cancer and, in particular, invasion or metastasis. Of the 12 genes up regulated in more invasive cells, several have been reporter previously to display disregulated expression in cancer and/or metastases. S100A13 was chosen for functional analysis RPS6KA3 Independent data analysis of two microarray experiments, one comparing invasive MCF7H3erbB2 to non-invasive MCF7H3, and the other comparing highly invasive DLKP4E to poorly invasive DLKP, found increased levels of RPS6KA3 at mRNA level corresponded to an invasive phenotype. A fold change of +2.36 was seen in MCF7H3erbB2, and +1.2 in DLKP4E, suggesting an increase in RPS6KA3 contributed to an increase in invasion. Although a direct link between RPS6Ka3 and invasion has not been shown previously, recent work suggests involvement of RPS6KA3 in breast cancer, with mean levels of RPS6KA3 statistically higher than in normal tissue (being overexpressed in -50% of human breast cancer tissue samples) (Smith et al., 2005). The same study identified the first small-molecule RSK-specific inhibitor, SLOlOl, caused inhibition of proliferation in MCF7 cells, producing a cell-cycle block in Gl phase. Work carried out with prostate cancer tissues produced similar results, with —50% of samples overexpressing RPS6KA3, and inhibition of proliferation after inhibition of RPS6KA3 (with both siRNA and SLOlOl )(Clarke et al, 2005).

A considerable decrease in RPS6KA3 protein was seen in DLKP4E and SKBR3 post-siRNA transfection. This result shows that despite lack of evidence at mRNA level, RPS6KA3 siRNA did function in 'knocking

-down' RPS6KA3 in both cell lines. Real-time PCR detected GAPDH knock-down in these cells under the same conditions and therefore it is unlikely that this result was due to an unsuccessful transfection. Results from the present study have demonstrated that siRNA silencing of RPS6KA3 alone had a considerable effect on invasion in SKBR3 and DLKP4E. RPS6KA3 therefore holds a powerful position in determining cellular response.

RPS6KA3 may also contribute to increased invasion through inhibition of apopotosis. Defective apoptosis can facilitate metastasis by allowing cells to ignore restraining signals from neighboring cells, survive detachment from the extracellular matrix, and persist in hostile environments. The development and maintenance of healthy tissues is dependent on a balance between cell survival and cell death. Disruption of this balance and prevention of apoptotsis contributes to uncontrolled growth and clonal expansion of cancer cells.

The Bcl-2 family member Bad is a pro-apoptotic protein, and phosphorylation of Bad by cytokines and growth factors promotes cell survival in many cell types (Reed, 1998). Phosphorylation of Bad results in its release from Bcl-xl, increasing levels of Bcl-xl in the cell and causing a decrease in apoptosis. She et al., 2002 illustrated that UVB-induced phosphorylation of Bad at serine 112 was mediated through MAP kinase signaling pathways in which RPS6KA3 served as direct mediator. More recent reports have confirmed this and show that RPS6KA3 -mediated phosphorylation of Bad is activated by the Ras signaling pathway (Gu et al., 2004). Results here demonstrate a functional effect of RPS6KA3 in invasion of both DLKP4E and SKBR3. RPS6KA3 overexpression in breast and prostate cancer tissue, along with its obvious association with so many other invasion markers strongly implicates RPS6KA3 in the invasion process. This knowledge combined with the functional effects observed after RPS6KA3 siRNA silencing in DLKP4E and SKBR3, makes RPS6KA3 a probable marker of invasion, and a promising target for future anti-invasion interventions. TFPIl

TFPIl was chosen as a target for siRNA based on MCF7H3erbB2 array data analysis that showed a +19.77 fold change in MCF7H3erbB2 (invasive) compared to MCF7H3 (non-invasive), and also a direct relationship with genes regulated by the 9-gene network. eIF4E translation of complex mRNAs, in particular oncogenes, has previously been associated with cancer progression and is also involved in the translation of erbB2 protein (Yoon et al., 2006). erbB2 can also control the amount of eIF4E available for translation by activating the Akt/mTOR signalling cascade (Zhou et al., 2004). In this way, eIF4E and erbB2 are involved in a positive feed-back loop. Real-time PCR carried out on TFPIl siRNA A, B and C in both DLKP4E and SKBR3 showed significant knock-down of TFPIl mRNA compared to scrambled siRNA transfections. Therefore, transfection of all 3 TFPIl siRNAs into DLKP4E and SKBR3 resulted in silencing of TFPIl at mRNA level, with overall much greater effect in SKBR3 cells.

Western blots carried out showed no significant change in TFPIl protein after transfection with any of the TFPIl siRNAs in DLKP4E TFPIl was not detected in SKBR3, and this may have been due to low levels of TFPIl expression in this cell line.

72hrs after transfection with TFPIl siRNA, cells were assayed for invasion. DLKP4E results showed a reduction in the number of invading cells when transfected with all 3 TFPIl siRNAs. SKB R3 transfected with TFPIl siRNA showed a dramatic decrease in invasion. Up to an 80% reduction in invading cells was observed after TFPIl siRNA A and B transfection. This considerable drop in invasion was obvious from photographs of invasion inserts and was confirmed by counting the number of invading cells per μm . This result combined with those from real-time PCR implies that siRNA silencing of TFPIl in DLKP4E and SKBR3 decreased invasion. THBSl

Microarray analysis found that in MCF7H3erbB2, an invasive breast cancer cell line, THBSl was down- regulated by -2.3 fold compared to the non-mvasive MCF7H3 parent, indicating an anti-mvasive role for THBSl Four cell lines of increasing invasiveness were subsequently chosen to examine the effect of THBSl specific siRNA on in vitro invasion These were: MCF7 (non-mvasive), DLKP (mildly invasive), SKBR3 (highly invasive) and DLKP4E (highly invasive).

Western Blot and qPCR were carried out post-transfection of three non-homologous siRNAs to determine the success of siRNA silencing of THBSl DLKP4E, the most highly invasive cell line, showed a 20% to 50% reduction m THBSl mRNA 48hrs after transfection with siRNA A, B and C. In SKBR3 both siRNA A and B caused an 80% to 85% decline in THBSl mRNA after 24hrs, with siRNA C also showing a 40% decrease. THBSl mRNA levels were reduced by 75% to 80% in DLKP at both 24 and 48hr after transfection of all three siRNA' s. MCF7, the only non-mvasive cell line used in the study, showed a 25% to 40% reduction in mRNA after transfection with all three siRNAs Though insufficient levels of THBSl protein were present in DLKP4E, SKBR3 and DLKP for detection by western blot, THBSl protein was detected in MCF7, and a considerable decrease observed post-siRNA transfection THBSl siRNA A and C had the most marked effect in MCF7, with no detectable band After transfection of THBS 1 siRNA B a protein band was still visible, but clearly reduced compared to the non-transfected and scrambled controls Interestingly, protein results mirrored those found at mRNA level in MCF7 Invasion assays were carried out to establish whether THBSl silencing seen in mRNA and protein led to a change in the invasion status of the cells DLKP4E, already a highly invasive cell line, showed an insignificant change in the number of invading cells. SKBR3 showed an increase in invading cells, with cell counts revealing a 1 3 to 1.7 fold increase in THBS 1 siRNA transfected cells. The most dramatic results were seen in DLKP and MCF7. DLKP, a mildly invasive cell line, showed a 3.5 to 4-fold increase m the number of invading cells when transfected with all three siRNAs. MCF7, a completely non- invasive cell line, became invasive after transfection with THBSl siRNA. These results demonstrate the dramatic increases in invasion across SKBR3, DLKP and MCF7 caused by transfection of THBSl siRNA, and validate microarray analysis which identified THBSl as a potential anti-invasive gene.

EGRl and TNFAIP8, both of which have previously been identified as markers of invasion, were also included in the present study. Results of invasion assays post-TNFAIP8 siRNA transfection showed a considerable decrease in invasive cells. The invasion assay results following EGRl silencing were also impressive, with all three siRNAs causing a considerable drop in the number of invading cells in both DLKP4E and SKBR3. These results serve to validate the process by which the novel markers/targets in the present study were identified. SlOOAl 3

Though SlOOAl 3 had been linked to angiogenesis, it (or other genes mentioned above) has not been linked previously with cancer metastasis. In order to confirm the differential expression of the SlOOAl 3 transcript in the cell line panel, semi-quantitative RT-PCR was performed. The gene was indeed increased by up to 300% in the more invasive compared to the less invasive cell lines in the panel. Interestingly the level of transcript in the P2 variant seemed to be under-estimated in the microarray analysis compared to RT-PCR. Transient transfection with target-specific siRNA reduces cellular levels of SlOOAl 3 transcript Having established that S100A13 transcript levels correlate with a more invasive phenotype, S100A13 mRNA was knocked down in an attempt to identify a functional role for the protein. Three independent siRNA sequences were transiently transfected into the more invasive cell line, TX2, and target mRNA depletion was monitored over a 72hr period by quantitative RT-PCR. Transient transfection with 3 independent siRNAs resulted in S100A13-specific knockdown of between 60-80% in TX2 cells over this timeframe. Two of the siRNAs maintained significant levels of knockdown up to 72hrs with the third less effective at that stage. These oligos were found to be equally effective in all three invasive cell variants over this period.

Knockdown of SlOOAl 3 in invasive lung cancer cell lines reduces their in vitro invasiveness Cell lines Pl, P2 and TX2 were transiently transfected with S100A13-specific siRNA and scrambled control oligos and allowed to recover for 72 hours. The cells were then assayed for their ability to migrate through Matrigel™ in a chamber-based in vitro invasion assay. After 24 hrs the inserts were stained and invading cells counted. A decrease of up to 80% was seen in cell lines that were depleted in SlOOAl 3. The pattern of reduction correlated closely with the level of transcript knockdown obtained with each siRNA. It is worth noting the impact of the treatment itself on the phenotype (No treatment v scrambled) hence the long recovery time before performing the functional assay.

Reduced invasiveness is not a result of decreased cellular profileration

It was unclear at this point whether the observed reduction in invasion was merely a by-product of decreased cellular proliferation or a direct impact on cellular invasiveness. Prior to the invasion assay, cell numbers were counted in the mock and S100A13-specific siRNA transfected cells at the end of the 72 hour recovery period. The treatment was found to have no impact on cell number compared to the control cultures. Transient overexpression of exogenous Sl 0OA 13 is insufficient to induce an invasive phenotype Having established that specific down-regulation of the SlOOAl 3 gene transcript in invasive lung cancer cell variants reduces this phenotype by as much as 80%, the importance of this factor in its development

was investigated. Clearly many genes are differentially expressed in cells with different degrees of invasiveness but it would be useful to identify the control molecules that set the process in motion. The weakly invasive variants, TXl and VCR were transfected with an expression vector containing the S100A13 coding sequence driven by a strong viral promoter. High levels of expression of the exogenous protein were achieved up to 72 hrs post-transfection. It is also apparent from this experiment that the levels of endogenous protein are below the level of detection with the antibody used here. When cultured overnight on Matrigel™ the cells were not found to be any more invasive than vector-only transfected controls. Upregulation of SlOOAl 3 protein, transiently at least, does not transform weakly invasive DLKP cells to a more aggressive phenotype. The array data indicated that SlOOAl 3 expression was increased in the more invasive cell lines in a panel of lung cancer cell lines. Analysis of SlOOAl 3 mRNA by RT-PCR confirmed increased expression in the invasive cell lines P2 and TX2 compared with the weakly invasive cell lines TXl and VCR. This data concurs with a study by Smirnov et al., (2005) who identified SlOOAl 3 as a novel predictor of metastasis following its detection in circulating tumour cells in blood in breast, colorectal and prostate cancers, but not in Lung cancer patients.

To help define the potential role of S100A13 in invasive lung cancer, three independent S100A13- specific siRNA sequences were transfected in the invasive cell lines Pl, P2 and TX2 to examine the effect of this decrease in S100A13 expression on both proliferation and invasion. Additionally, S100A13 cDNA was transiently transfected into the less invasive cell lines TXl and VCR. An SlOOAl 3 siRNA-specific decrease in mRNA levels was confirmed by PCR analysis. Endogenous SlOOAl 3 protein levels were below western blotting detection limits.

Decreasing SlOOAl 3 expression had no obvious effect on DLKP variant proliferation thereby indicating no obvious role for SlOOAl 3 in cell growth or apoptosis in these cell lines, although more specific analysis such as H-thymididine incorporation assays and TUNEL assays would clarify this further. The identification and characterisation of molecular markers that can identify patients at high risk for metastatic spread and improve clinical management, therapeutic outcome and patient survival is of paramount importance. SlOO family members have been identified as molecular markers for metastatic potential. S100A2 is of prognostic significance in lung cancer in that it allows for the discrimination of high- and low- risk patients in the lymph node negative sub group. Clinically this translates as aggressive initial treatment of Sl 00A2 -negative patients thereby avoiding undertreatment. It also allows for the selection of S 100A2 -positive patients who would benefit from less aggressive treatment (Diederichs et al., 2004). Komatsu et al., (2002) associated S100A6 expression with human colorectal adenocarcinoma tumorigenesis and metastasis. SlOOAl 3 has been recently implicated as a marker of angiogenesis in endometriosis (Hayrabedyan et al., 2005) and was identified as a novel predictor of metastasis following global gene expression profiling of circulating tumour cells by Smirnov et al., (2005). Our results show that decreasing S100A13 expression in invasive cell lines can reduce their invasive capacity by more than 80% in siRNA treated cells compared with their scrambled siRNA transfected counterparts. This is the first study to give functional

evidence for the role of S 100Al 3 in invasion, since in vitro invasion can be inhibited by reduced

S100A13 expression.

The words "comprises/comprising" and the words "having/including" when used herein with reference to the present invention are used to specify the presence of stated features, integers, steps or components but does not preclude the presence or addition of one or more other features, integers, steps, components or groups thereof.

It is appreciated that certain features of the invention, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the invention which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable sub-combination.

References

AdamSjJC. Thrombospondιn-1. The international journal of biochemistry & cell biology 1997 29:861-865

Bastian,M, Stemer,M and Schuff-Werner,P. Expression of thrombospondιn-1 in prostate-derived cell lines. International journal of molecular medicine 2005 15:49-56 Bornstein,P. Diversity of function is inherent in matricellular proteins: an appraisal ofthrombospondin 1.

The Journal of cell biology 1995 130:503-506

Bronckart Y, Decaestecker C, Nagy N, Harper L, Schafer BW, Salmon I, Pochet R, Kiss R, Heizman

CW. Development and progression of malignancy in human colon tissues are correlated with expression of specific Ca(2+)-binding SlOO proteins. Histol Histopathol. 2001 Jul;16(3):707-12. Brown,LF, Guidi,AJ, Schnitt,SJ, Van De Water,L, Iruela-Arispe,ML, Yeo,TK, Tognazzi,K and

DvorakjHF. Vascular stroma formation in carcinoma in situ, invasive carcinoma, and metastatic carcinoma of the breast. Clinical cancer research: an official journal of the American Association for

Cancer Research 1999 5:1041-1056

Chen,JJ, Peck,K, Hong,TM, Yang,SC, Sher,YP, Shih,JY, Wu ₅ R, Cheng,JL, Roffler,SR, Wu,CW and Yang,PC. Global analysis of gene expression in invasion by a lung cancer model. Cancer research 2001

61:5223-5230

Donato R. Intracellular and extracellular roles of SlOO proteins. Microsc Res Tech 2003;6:540-551

Donato R. SlOO: A multigenic family of calcium-modulated proteins of the EF-hand type with intracellular and extracellular functional roles. Int J Biochem Cell Biol 2001;33:637-668. El-Rifai W, Moskaluk CA, Abdrabbo MK, Harper J, Yoshida C, Riggins GJ, Frierson HF Jr, Powell SM

Gastric cancers overexpress SlOOA calcium-binding proteins. Cancer Res. 2002 Dec l ,62(23):6823-6.

Fang,JY and Richardson,BC. The MAPK signalling pathways and colorectal cancer. The Lancet

Oncolcgy 2005 6:322-327

Garrett SC, Varney KM, Weber DJ, Bresnick AR J Biol Chem. 2006 Jan 13;281(2).677-80. Epub 2005 Oct 21. Sl 00A4, a mediator of metastasis.

Guimiot,F, Delezoide,AL, Hanauer,A and Simonneau,M. Expression of the RSK2 gene during early human development. Gene expression patterns : GEP 2004 4:111-114

Heizmann CW, Fritz G, Schafer BW. SlOO proteins: Structure, functions and pathology. Front Biosci 2002;7:1356-1368.

Iversen,N, Lindahl,AK and Abildgaard,U. Elevated plasma levels of the factor Xa-TFPI complex in cancer patients. Thrombosis research 2002 105:33-36 Iversen,N, Lindahl,AK and Abildgaard,U. Elevated TFPI in malignant disease: relation to cancer type and hypercoagulation. British journal of haematology 1998 102:889-895

Kennedy RD, Gorski JJ, Quinn JE, Stewart GE, James CR, Moore S, Mulligan K, Emberley ED, Lioe TF, Morrison PJ, Mullan PB, Reid G, Johnston PG, Watson PH, Harkin DP. BRCAl and c-Myc associate to transcriptionally repress psoriasin, a DNA damage-inducible gene. Cancer Res. 2005 Nov 15;65(22):10265-72.

Kim,J, Mori,T, Chen,SL, Amersi,FF, Martinez,SR, Kuo,C, Turner,RR, Ye,X, Bilchik,AJ, Morton,DL and Hoon,DS. Chemokine receptor CXCR4 expression in patients with melanoma and colorectal cancer liver metastases and the association with disease outcome. Annals of Surgery 2006 244 : 113 - 120 Kristiansen,G, Yu, Y, Petersen,S, Kaufmann,O, Schluns,K, Dietel,M and Petersen,! Over expression of c- erbBl protein correlates with disease-stage and chromosomal gain at the c-erbB2 locus in non-small cell lung cancer. European journal of cancer 2001 37: 1089-1095

Krop I, Marz A, Carlsson H, Li X, Bloushtain-Qimron N, Hu M, Gelman R, Sabel MS, Schnitt S, Ramaswamy S, Kleer CG, Enerback C, Polyak K. Cancer Res. 2005 Dec 15;65(24): 11326-34. A putative role for psoriasin in breast tumor progression. Law,E, Gilvarry,U, Lynch, V, Gregory ,B, Grant,G and Clynes,M. Cytogenetic comparison of two poorly differentiated human lung squamous cell carcinoma lines. Cancer genetics and cytogenetics 1992 59:111- 118

Lefranc F, Golzarian J, Chevalier C, DeWitte O, Pochet R, Heizman C, Decaestecker C, Brotchi J, Salmon I, Kiss R. Expression of members of the calcium-binding S-100 protein family in a rat model of cerebral basilar artery vasospasm. J Neurosurg. 2002 Aug;97(2):408-15.

Lindahl,AK, Odegaard,OR, Sandset,PM and Harbitz,TB. Coagulation inhibition and activation in pancreatic cancer. Changes during progress of disease. Cancer 1992 70:2067-2072 Lindahl,AK, Odegaard,OR, Sandset,PM and Harbitz,TB. Coagulation inhibition and activation in pancreatic cancer. Changes during progress of disease. Cancer 1992 70:2067-2072 Lindahl,AK, Sandset,PM and Abildgaard,U. The present status of tissue factor pathway inhibitor. Blood coagulation & fibrinolysis : an international journal in haemostasis and thrombosis 1992 3:439-449 Lindahl,AK, Sandset,PM, Abildgaard,U, Andersson,TR and Harbitz,TB. High plasma levels of extrinsic pathway inhibitor and low levels of other coagulation inhibitors in advanced cancer. Acta Chirurgica Scandinavica 1989 155:389-393 Maelandsmo GM, Florenes VA, Mellingsaeter T, Hovig E, Kerbel RS, Fodstad O. Differential expression patterns of S100A2, S100A4 and S100A6 during progression of human malignant melanoma. Int J Cancer. 1997 Aug 22;74(4):464-9.

Marenholz I, Heizmann CW, Fritz G. SlOO proteins in mouse and man: from evolution to function and pathology (including an update of the nomenclature). Biochem Biophys Res Commun. 2004 Oct 1;322(4):1111-22. Review.

Marx.D, Schauer,A, Reiche,C, May,A, Ummenhofer,L, Reles,A, Rauschecker,H, Sauer,R and Schumacher,M. c-erbB2 expression in correlation to other biological parameters of breast cancer. Journal of cancer research and clinical oncology 1990 116:15-20

McBride,S, Meleady,P, Baird,A, Dinsdale,D and Clynes,M. Human lung carcinoma cell line DLKP contains 3 distinct subpopulations with different growth and attachment properties. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 1998 19:88-103 Moon, Y, Bottone,FG,Jr, McEntee,MF and Eling,TE. Suppression of tumor cell invasion by cyclooxygenase inhibitors is mediated by thrombospondιn-l via the early growth response gene Egr-1. Molecular cancer therapeutics 2005 4: 1551-1558

Moran,E, Larkm,A, Cleary,I, Barnes,C, Kennedy,SM, Kelehan,P and Clynes,M. Monoclonal antibodies raised to paraffin wax embedded archival tissue; feasibility study of their potential to detect novel antigenic markers. Journal of immunological methods 1998 219:151-159

Mouta Carreira C, LaVallee TM, Tarantini F, Jackson A, Lathrop JT, Hampton B, Burgess WH, Maciag T. SlOOAl 3 is involved in the regulation of fibroblast growth factor- 1 and p40 synaptotagmin-1 release in vitro. J Biol Chem. 1998 Aug 28;273(35):22224-31. NemotojT, Vana,J, Bedwani,RN, Baker,HW, McGregor,FH and Murphy,GP. Management and survival of female breast cancer: results of a national survey by the American College of Surgeons. Cancer 1980 45:2917-2924

Novotny,WF, Girard,TJ, Miletich,JP and Broze,GJ,Jr. Purification and characterization of the hpoprotein-associated coagulation inhibitor from human plasma Journal of Biological Chemistry 1989 264:18832-18837 Pfeiffer,P, Clausen,PP, Andersen,K and Rose,C. Lack of prognostic significance of epidermal growth factor receptor and the oncoprotein pl85HER-2 in patients with systemically untreated non-small-cell lung cancer' an immunohistochemical study on cryosections. British journal of cancer 1996 74:86-91 Qian,X, Rothman,VL, Nicosia,RF and Tuszynski,GP. Expression of thrombospondιn-1 in human pancreatic adenocarcinomas, role in matrix metalloproteιnase-9 production. Pathology oncology research : POR 2001 7 251-259

Quenel,N, Wafflart,J, Bonichon,F, de Mascarel,I, Trojani,M, Durand,M, AvriLA and Coindre,JM. The prognostic value ofc-erbB2 in primary breast carcinomas, a study

Rehman I, Cross SS, Catto JW, Leiblich A, Mukherjee A, Azzouzi AR, Leung HY, Hamdy FC. Promoter hyper-methylation of calcium binding proteins S100A6 and S100A2 in human prostate cancer. Prostate. 2005 Dec l ;65(4):322-30.

Sandset,PM. Tissue factor pathway inhibitor (TFPI) -an update. Haemostasis 1996 26 Suppl 4:154-165

She,QB, Ma, WY, Zhong,S and Dong,Z. Activation of JNKl, RSK2, and MSKl is involved in serine 112 phosphorylation of Bad by ultraviolet B radiation. The Journal of biological chemistry 2002 277:24039-

24048

Slamon,DJ, Clark,GM, Wong,SG, Levin, WJ, Ullrich,A and McGuire,WL. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 1987

235:177-182

Slamon,DJ, Godolphin,W, Jones,LA, HoIt ₅ JA, Wong,SG, Keith,DE, Levin, WJ, Stuart,SG, Udove,J and

Ullrich,A. Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science 1989

244:707-712 Smith,JA, Poteet-Smith,CE, Xu, Y, Errington,TM, Hecht,SM and Lannigan,DA. Identification of the first specific inhibitor of p90 nbosomal S6 kinase (RSK) reveals an unexpected role for RSK in cancer cell proliferation. Cancer research 2005 65:1027-1034

Stephens P, Hunter C, et al. Lung cancer: intragenic ERBB2 kinase mutations in tumours. Nature. 2004

Sep 30;431(7008):525-6 Toker,A and Yoeh-Lerner,M. Akt signaling and cancer: surviving but not moving on. Cancer research

2006 66:3963-3966

Tsai,CM, Chang,KT, Wu,LH, Chen,JY, Gazdar,AF, Mitsudomi,T, Chen,MH and Perng,RP. Correlations between intrinsic chemoresistance and HER-2/neu gene expression, p53 gene mutations, and cell proliferation characteristics in non-small cell lung cancer cell lines. Cancer research 1996 56:206-209 Werling,RW, Zacharski,LR, Kisiel,W, Bajaj,SP, Memoli,VA and Rousseau,SM. Distribution of tissue factor pathway inhibitor in normal and malignant human tissues. Thrombosis and haemostasis 1993

69:366-369

Wicki R, Schafer BW, Erne P, Heizmann CW Characterization of the human and mouse cDNAs coding for S100A13, a new member of the SlOO protein family. Biochem Biophys Res Commun. 1996 Oct 14;227(2):594-9

Zhang L, Fogg DK, Waisman DM J Biol Chem 2004 Jan 16;279(3):2053-62. Epub 2003 Oct 21. RNA interference-mediated silencing of the SlOOAlO gene attenuates plasmin generation and invasiveness of

Colo 222 colorectal cancer cells

Zhao,Y, Bjorbaek,C and Moller,DE. Regulation and interaction of pp90(rsk) isoforms with mitogen- activated protein kinases The Journal of biological chemistry 1996 271 :29773-29779