WO2009154790A2 - Novel metastasis suppresor genes and uses thereof - Google Patents

Abstract

The invention relates to methods for detecting a metastatic cancer, methods for inhibiting tumor metastatic properties, methods for treating tumor metastasis, methods for identifying agents for the treatment of tumor metastasis, and methods for screening for modulators of tumor metastatic properties.

Description

NOVEL METASTASIS SUPPRESSOR GENES AND USES THEREOF

RELATED APPLICATIONS This application claims priority under 35 U.S.C. §119 from U.S. provisional application serial number 61/074,366, filed June 20, 2008, and U.S. provisional application serial number 61/110,169, filed October 31, 2008, the contents of which are incorporated herein in their entirety.

FIELD OF THE INVENTION The invention relates to methods for detecting a metastatic cancer, methods for inhibiting tumor metastatic properties, methods for treating tumor metastasis, methods for identifying agents for the treatment of tumor metastasis, and methods for screening for modulators of tumor metastatic properties.

BACKGROUND OF THE INVENTION

The progression to metastasis involves a series of steps, commonly known as the metastatic cascade, which is under the control of multiple signals generated both internally and externally to the cell. It is well appreciated that many oncogenic factors affect a spectrum of processes from nascent primary tumor formation to an established metastatic state, and many of these factors are targets for cancer therapy. There is also a growing understanding that some proteins have the capacity to inhibit one or more steps of the metastatic cascade without necessarily affecting tumorgenicity (Berger et al., Cancer Biol. Ther. 4, 805-812, (2005)). While genes whose products have this metastatic inhibitory potential may reveal new therapeutic strategies, their discovery has been difficult to achieve in a systematic and efficient way. Thus, few metastasis suppressor genes have been identified and the full range of cellular functions that influence tumor metastatic properties remains elusive.

SUMMARY OF INVENTION The invention, in some aspects, provides methods for inhibiting tumor metastatic properties, methods for treating tumor metastasis, methods for identifying agents for the treatment of tumor metastasis, and methods for screening for modulators of tumor metastatic properties. We developed and carried out a genome-wide RNA interference (RNAi) screen to identify metastasis suppressor genes by using a novel three- dimensional cell culture assay system that facilitates assessment of a broad range of metastatic properties (i.e., invasion of the tumor border, expansion into the extracellular matrix and colonization of the secondary site) on a genome-wide scale. We identified 80 genes that influence one or more metastatic properties by affecting various cellular functions, including cell cycle regulation, cell communication, signal transduction, cytoskeletal structure, intracellular transport, metabolic and energy pathways, and protein metabolism. Thus, our efforts have shed new light on the array of cellular functions that influence tumor metastasis. The combination of our customized culture system with in vivo tumor metastasis and primary tumor growth assays has enabled us to gain broad insight into the importance of these metastasis suppressor genes (MSGs) in tumor metastasis. As disclosed herein, these MSGs provide a novel basis for identifying compounds and compositions that modulate tumor metastasis. Moreover, these MSGs provide new opportunities for treating metastatic disease.

Metastasis suppressor genes encode proteins that inhibit one or more steps required for metastasis without affecting primary tumor formation (Steeg, P. S., Nat. Rev. Cancer 3, 55-63, (2003)). Here we describe a genome-wide RNA interference screening strategy to identify new metastasis suppressor genes. In some aspects, the screen involves expressing a plurality shRNAs in poorly metastatic B 16-FO mouse melanoma cells (Fidler, IJ., Cancer Res. 35, 218-334, (1975)) and identifying those shRNA(s) that enhance one or more metastatic properties of the B 16-FO cells. Candidate shRNAs were selected based upon enhanced formation of satellite colonies in a three-dimensional cell culture system. Individual B 16-FO knockdown cell lines were then tested in a mouse tail vein injection assay for their ability to promote lung metastasis. Using this approach we discovered 22 genes that suppress metastasis without affecting primary tumor growth. Mining of cancer gene-profiling databases reveals that five of these genes are significantly down-regulated in metastatic melanoma, and 15 are down-regulated in metastases of multiple tumor types. Knockdown of one of these genes, Gasl, also increases metastasis in a murine spontaneous metastasis assay. We find that Gasl is substantially down-regulated in B16-F10 cells, which contributes to the high metastatic potential of this cell line; accordingly, we show that knockdown of Gasl protects B 16-FO cells from apoptosis. We also find that the human GASl gene is down-regulated in metastatic melanoma cell lines and tissue samples. We conclude that GASl is a melanoma metastasis suppressor gene. Cancer profiling database mining reveals that in addition to GASl, four of the genes we identified in the screen are significantly down- regulated in metastatic melanoma, and 14 are down-regulated in metastases of multiple tumor types. Thus, the genome-wide shRNA screen we have developed reveals genes that, on the basis of both experimental and clinical evidence, are new metastasis suppressors. According to some aspects of the invention, methods of detecting a metastatic cancer in a subject are provided. In some embodiments, the methods comprise obtaining a clinical sample from a subject having, or suspected of having, cancer, and determining a level of expression of at least one metastasis suppressor gene (MSG) in the clinical sample from the subject. In some embodiments, if expression of the MSG is reduced compared with a control value (e.g., a historical reference level, a threshold level below which a metastatic cancer is indicated, a level of expression of the MSG in a control sample, e.g., a sample from a subject having a non-metastatic cancer, etc.) the subject has a metastatic cancer. In some embodiments, the cancer is a melanoma, breast, prostate, ovarian, liver, sarcoma, colon, lung, bladder, gastric, head, neck, seminoma, Ewing's sarcoma, cervical or renal cancer. In some embodiments, the MSG is GASl. In other embodiments, the MSG is selected from: ACTA2, ADAMTS16, AGL, ALG6, ATGl, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orβ4, PRAMEF8, 4931422A03RIK, and KIAAl 276.

In some embodiments, the step of determining if expression of a MSG is reduced comprises comparing the level of expression of the MSG in the clinical sample with the level of expression of the MSG in a control sample, wherein a decrease in expression of the MSG in the clinical sample compared with the control sample indicates that the MSG is reduced. Typically a decrease in expression of the MSG in a clinical sample compared with a control sample indicates that the MSG is reduced when it is a statistically significant decrease. The decrease may be up to about 5%, about 10%, about 20%, about 30%, about 40%, about 50%, about 60%, about 70%, about 80%, about 90%, about 95%, about 99%, or more.

A variety of control values may be used. In some embodiments, the control value is a historical reference level below which a metastatic cancer is indicated. In some embodiments, the control value is a historical reference level at or below which a metastatic cancer is indicated. In other embodiments, the control value is a threshold level (e.g., a predetermined threshold level) at or below which a metastatic cancer is indicated. In other embodiments, the control value is a threshold level below which a metastatic cancer is indicated. In other embodiments, the control value is a level of expression of the MSG in a control sample. A variety of control samples may be used. In some embodiments, the control sample is from a normal tissue, e.g. , a normal tissue from the same subject from which the clinical sample was obtained. In other embodiments, the control sample is a non-metastatic cancer tissue. The methods are not limited to control samples from non-metastatic tissues or from subjects not having a metastatic cancer. In some embodiments, the control sample is from a metastatic tissue sample and an expression level of the MSG which is at or below the control sample indicates that the subject has a metastatic cancer.

In some embodiments, the step of determining the level of expression of a MSG comprises measuring the level of an mRNA of the MSG. In other embodiments, the step of determining the level of expression of a MSG comprises measuring the level of a protein encoded by the MSG. In other embodiments, the step of determining the level of expression of a MSG comprises measuring the level of a genomic locus comprising a MSG, e.g., to detect a mutation, e.g., a deletion, translocation, inversion, etc., that reduces or eliminates expression of the MSG. In some embodiments, the methods involve determining the level of expression of a plurality of different MSGs, e.g., in a multiplex reaction, e.g., using a nucleic acid array or protein detection array. Accordingly, in some embodiments, if expression of a plurality of MSGs are reduced, the cancer is identified as a metastatic cancer (See, e.g., Table 2). The plurality of MSGs which are indicative of a metastatic cancer may be at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 20, at least 30, at least 40, at least 50, at least 60, at least 70, at least 80, or more MSGs. In certain embodiments, the plurality of MSGs comprises one or more MSGs selected from: ACTA2, ALG6, CCDC39, CCNA2, CTSO, CPA2, DPPA3, GASl, PDYN, PHKAl, SETD2, SLC9A3R2, THSD7B, TOMM70A, and ZNF294. In certain embodiments, the plurality of MSGs are selected from: ACTA2, ALG6, CCDC39, CCNA2, CTSO, CPA2, DPPA3, GASl, PDYN, PHKAl, SETD2, SLC9A3R2, THSD7B, TOMM70A, and ZNF294.

According to some aspects, kits for detecting a metastatic cancer in a subject are provided. In some embodiments, the kits comprise at least one container having disposed therein a reagent for detecting expression of a metastasis suppressor gene, and a label and/or instructions for use of the kit in detecting a metastatic cancer based on expression of a metastasis suppressor gene. In some embodiments, the kits comprises a device for measuring expression of a plurality of MSGs in parallel, e.g., a nucleic acid array, a protein detection array, a bead-based nucleic acid assay system, etc.

According to one aspect of the invention, methods for inhibiting one or more metastatic properties in a cell are provided. The methods include increasing the activity of one or more metastasis suppressor genes (MSGs). In certain embodiments the MSGs are one or more genes of: ACTA2, ADAMTS16, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orβ4, PRAMEF8, 4931422A03RIK, and KIAAl 276. In one embodiment, the MSG is GASl. In some embodiments, increasing the activity of one or more MSGs may involve contacting a cell with a compound or composition efficacious at increasing the activity of the one or more MSGs. In some embodiments, the cell has reduced activity of the one or more MSGs induced by the compound or composition. In some embodiments the cell is in vitro or ex vivo. In some embodiments the cell is in vivo. In other embodiments the cell is a tumor cell. In certain embodiments the tumor is non-metastatic. In certain other embodiments the tumor is metastatic. In some embodiments the composition efficacious at increasing the activity of the one or more MSGs is a gene therapy. In certain embodiments the gene therapy comprises delivery of a therapeutically effective amount of an expression construct encoding one or more of: ACTA2, ADAMTSl 6, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CDS,

CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPP A3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NMEl, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orf24, PRAMEF8, 4931422 A03 RIK, and KIAA1276. In one embodiment, the gene therapy comprises delivery of a therapeutically effective amount of an expression vector encoding GASl.

According to one aspect of the invention, methods for treating a subject having, or at risk of having, a tumor metastasis are provided. The methods comprise administering to the subject an effective amount of a compound or composition that increases the activity of one or more MSGs. In certain embodiments, the one or more MSGs are one or more of: ACTA2, ADAMTSl 6, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTC 17, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orβ4, PRAMEF8, 4931422A03RIK, and KJAAl 276. In one embodiment, the MSG is GASl. In some embodiments the composition that increases the activity of one or more MSGs is a gene therapy. In certain embodiments the gene therapy comprises delivery of a therapeutically effective amount of an expression vector encoding one or more of:

ACTA2, ADAMTS16, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTC 17, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orβ4, PRAMEF8, 4931422 A03RIK, and KIAAl 276. In one embodiment, the gene therapy comprises delivery of a therapeutically effective amount of an expression vector encoding GASl.

According to one aspect of the invention, methods for identifying compounds or compositions useful as pharmacological agents for the modulation of one or more metastatic properties are provided. The methods comprise contacting a cell with a compound or composition and assaying for the increased expression of one or more MSGs. In certain embodiments the MSGs are one or more of: ACTA2, ADAMTS16, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPP A3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP,

MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDMl 3, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orβ4, PRAMEF8, 4931422 A03 RIK, and KIAAl 276. In one embodiment, the MSG is GASl. In some embodiments the compound or composition contacts the cell for a predetermined period of time. In certain embodiments, the predetermined period of time is about 14 days. In some embodiments the cell is grown in an environment wherein one or more metastatic properties is measured. In certain embodiments the environment is in vitro. In certain other embodiments the environment is in vivo. In some embodiments the cell is a tumor cell. In certain embodiments the tumor is non-metastatic. In certain other embodiments the tumor is metastatic. In some embodiments the cell contacted by the compound or composition has reduced expression of one or more MSGs. In some embodiments the composition is a gene therapy. In certain embodiments the gene therapy comprises delivery of a therapeutically effective amount of an expression vector encoding one or more of: ACTA2, ADAMTS16, AGL, ALG6, ATGl, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NMEl, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SSl 8Ll, TDPOZ2, THSDlB, TOPBPl, TOMMlOA, TTCIl, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orβ4, PRAMEF8, 4931422 A03 RIK, and KIAAl 216. In one embodiment the gene therapy comprises delivery of a therapeutically effective amount of an expression vector encoding GASl. In certain embodiments the in vitro environment further comprises collagen. In certain embodiments the in vitro environment comprising collagen further comprises a basement membrane. In certain embodiments the in vitro environment comprising collagen and further comprising basement membrane further comprises a fibrin gel. In certain embodiments the metastatic properties comprise movement of the cell through the collagen gel, the basement membrane, and/or the fibrin gel.

According to one aspect of the invention, a method for screening for modulators of one or more metastatic properties is provided. The method comprises transducing test cells and control cells with pools of a plurality of retroviruses, wherein individual retroviruses in the plurality comprise a nucleic acid encoding a product capable of affecting expression of at least one gene encoded in the genome of the transduced cells; isolating test cells with one or more altered metastatic properties compared with control cells; and identifying the transduced nucleic acid in the isolated test cells. In one embodiment the product capable of affecting expression is an shRNA or shRNA-mir. In certain embodiments the shRNA or shRNA-mir is directed against the at least one gene encoded in the genome of the transduced cells. In some embodiments, the cell is grown in an environment wherein one or more metastatic properties are measured. In certain embodiments the environment is in vitro or ex vivo. In certain other embodiments the environment is in vivo. In some embodiments the cell is a tumor cell. In certain embodiments the tumor is non-metastatic. In certain other embodiments the tumor is metastatic. In certain other embodiments the tumor is implanted. In certain other embodiments an in vivo assay for alterations in one or more tumor metastatic properties is provided. In certain other embodiments the isolating comprises resecting metastatic tumor tissue. In certain other embodiments the isolating comprises resecting primary tumor tissue. In certain embodiments the in vitro or ex vivo environment further comprises collagen. In certain embodiments the in vitro or ex vivo environment comprising collagen further comprises a basement membrane. In certain embodiments the in vitro or ex vivo environment comprising collagen and further comprising basement membrane further comprises a fibrin gel. In certain embodiments the metastatic properties comprise movement of the cell through the collagen gel, the basement membrane, and/or the fibrin gel. In some embodiments the identifying comprises cloning the nucleic acid. In certain embodiments the identifying further comprises sequencing the nucleic acid. In some embodiments a cancer gene database is mined to determine expression of the nucleic acid in metastatic and non-metastatic tumors, hi some embodiments where the environment is in vitro or ex vivo the method further comprises subjecting the isolated cells to an in vivo assay for alterations in one or more metastatic properties. In some embodiments the plurality of retroviruses comprise sequence complementary to a portion of the mRNA sequence of each of substantially all known protein coding genes of the transduced cell's genome.

In some aspects, the invention provides methods for identifying a modulator of at least one metastatic property. In some embodiments, the methods involve contacting a plurality of cells, in a three-dimensional culture system, with a plurality of expression vectors comprising inserts and identifying one or more inserts that alter at least one metastatic property of a cell. In certain embodiments, at least one insert is a coding sequence for a functional RNA, optionally wherein the functional RNA is a miRNA, a shRNA, or an shRNA-mir. In some embodiments, the methods involve contacting a plurality of cells with a plurality of expression vectors, which comprise an shRNA gene operably-joined to a regulatory sequence, and identifying one or more of the expression vectors in the plurality that alter at least one metastatic property of a cell.

These and other aspects of the invention, as well as various advantages and utilities, will be more apparent with reference to the detailed description of the invention. Each aspect of the invention can encompass various embodiments as will be understood by the following description.

BRIEF DESCRIPTION OF DRAWINGS

Figure 1 depicts a genome- wide shRNA screen for the identification of candidate metastasis suppressor genes. (A) Schematic summary of the screen. Cells were embedded in collagen (gray), coated in Matrigel™ (orange) and sandwiched in fibrin (blue). (B) 3-D cell culture assay. Collagen/Matrigel™-embedded cells are shown in the center, surrounded by migrating satellite colonies in the fibrin matrix. (C) 3-D cell culture assay of B 16-FO cells stably transduced with a NS shRNA or a representative shRNA pool.

Figure 2 depicts identification of 22 genes the knockdown of which increase metastasis without affecting primary tumor growth. (A) Mouse tail vein metastasis assays showing lungs from mice injected with B 16-FO or B16-F10 cells, or B 16-FO cells stably expressing a NS shRNA or an shRNA directed against one of 22 genes. (B) Quantitation of the results presented in (A). Error bars represent SEM (n=3 mice). (C) Subcutaneous flank injections. Primary tumor growth at day 16 following subcutaneous injection of B 16-FO KD cell lines. Error bars represent SEM (n=3 mice).

Figure 3 depicts that down-regulation of Gas 1 in B16-F10 cells contributes to their high metastatic potential. (A) qRT-PCR analysis of candidate gene expression in B 16-FO and B16-F10 cells. Gene expression was normalized to Gapdh, and presented as the ratio of expression in B 16-FO relative to B16-F10 cells. A ratio >1 (above the red line) indicates down-regulation in B 16-F 10 cells. Error bars represent s.d. (n=3). (B) Three dimensional cell culture assay showing B 16-F 10 cells stably expressing Gasl or empty vector. (C) Left, mouse tail vein metastasis assay showing lungs from mice injected with B16-F10 cells stably expressing Gasl or empty vector. Right, quantitation. Error bars represent s.e.m. (n=3 mice). (D) Primary tumor growth at day 16 following subcutaneous injection of B16-F10 cells stably expressing Gasl or empty vector. Figure 4 depicts that down-regulation of Gasl in B16-F10 cells contributes to their high metastatic potential. (A) Left, spontaneous metastasis assay showing lungs from mice injected with B16-F0 cells stably expressing a Gasl or non-silencing shRNA. Right, quantitation. (B) Top, in vivo single-cell fluorescence imaging of the lung following injection of fluorescently labeled B 16-FO cells stably expressing a Gasl or NS shRNA into the tail vein. Bottom, quantitation. (C) qRT-PCR analysis of Gasl expression in immortalized human melanocyte MeI-STV (vector) cells relative to MeI-STR (Ras- expressing) cells. Error bars represent s.d. (n=3). Fig 4D depicts that knockdown of Gasl in the mouse breast cancer cell line 67NR increases lung metastasis. Mouse tail vein metastasis assay were performed; lungs are shown from mice injected with 67NR cells stably expressing a Gasl or non-silencing (NS) shRNA. Figure 5 depicts that Gasl suppresses metastasis by promoting apoptosis of cells following dissemination to secondary sites. (A) Left, in vivo single-cell fluorescence imaging of the lung following tail vein injection of fluorescently labeled B 16-FO cells stably expressing a Gasl or NS shRNA. Right, quantitation; error bars represent SEM (n=3 mice); at 1 hr p=0.7364 and at 3 hr p=0.0148. (B) Left, representative images of B16-F0/NS or B16-F0/Gasl KD cells colocalizing with a fluorescent pan-caspase probe, monitored at 3 hr after injection. Right, quantitation of the percentage of apoptotic B 16- FO/NS or B16-F0/Gasl KD cells. Error bars represent SEM (n=3 mice).

Figure 6 depicts that down-regulation of GASl in human metastatic melanoma cell lines and tissue samples. (A) GASl expression in benign nevi, primary melanoma and metastatic melanoma. Data were obtained through the Oncomine database; a Student's t-test, performed directly though the Oncomine 3.0 software, showed the difference in gene expression between the samples was significant. (B) Immunoblot analysis of GASl expression in human melanoma cell lines derived from primary tumors (UACC-257) or metastatic sites (LOX IMVI, MALME-3M, SK-MEL-2 and SK-MEL- 5). Tubulin was monitored as a loading control. (C) Left, quantitation of detectable GASl expression in primary (n=47) and lymph node metastatic (n=17) melanoma samples. Right, representative examples of a primary melanoma and lymph node metastasis analyzed for expression of GASl (red) and the melanoma marker HMB45 (green); nuclei were stained with DAPI (blue). Merged images are shown; scale bar represents 20 μm.

Figure 7 depicts three-dimensional cell culture assays of 78 B 16-FO knockdown cell lines. Collagen/Matrigel™-embedded cells are shown in the center, surrounded by migrating satellite colonies in the fibrin matrix. B 16-FO cells expressing a non-silencing (NS) shRNA and B16-F10 cells are shown as controls.

Figure 8 depicts an analysis of target gene expression in the B 16-FO knockdown cell lines. (A) Quantitative real-time RT-PCR (qRT-PCR) confirmed in all cases that expression of the target gene was decreased in each B 16-FO KD cell line. Error bars indicate SEM (n=3). (B) Immunoblot analysis for Gasl, Acta2 and Ccna2. Actin is shown as a loading control.

Figure 9 depicts three-dimensional cell culture assays and target gene analysis for each of the candidate 22 B 16-FO knockdown cell lines using a second, unrelated shRNA. For all 22 genes, a second, unrelated shRNA directed against the same target also resulted in increased formation of satellite colonies in the three-dimensional assay (A) and knockdown of target gene expression (B). Error bars indicate SD (n=3).

Figure 10 depicts that Gasl does not affect proliferation or apoptosis in cultured B16-F0 or B16-F10 cells. (A) Proliferation of B16-F0 cells expressing a Gasl or non- silencing (NS) shRNA, and B16-F10 cells ectopically expressing Gasl or empty vector was assessed at 24 and 48 hours. Error bars indicate SEM (n=3). (B) Levels of apoptosis in the indicated cultured cells lines were determined by immunoblotting for Parp-1.

Cleaved Parp-1, a marker of apoptosis, is indicated by an arrow.

Figure 1 IA depicts that Gasl suppresses metastasis by promoting apoptosis of cells following dissemination to secondary sites (24 hour time point). Quantitation of in vivo single-cell fluorescence imaging of the lung 24 hours following tail vein injection of fluorescently labeled B 16-FO cells stably expressing a Gasl or NS shRNA. Error bars represent SEM (n=3 mice); p=0.0027. Figure 1 IB depicts that ectopic expression of the anti-apoptotic protein Bcl-2 increases satellite colony formation in the 3-D cell culture assay. B 16-FO cells stably expressing Bcl-2 or vector were grown for 8 days in the 3-D cell culture system. Satellite colonies were counted under IOOX magnification. Error bars indicate SEM (n=3). Figure 11 C depicts that knockdown of GASl in UACC-257 cells increases their ability to form satellite colonies in the 3-D cell culture assay. Human melanoma UACC-257 cells stably expressing a GASl or NS shRNA were grown for 8 days in the 3-D cell culture system. Error bars indicate SEM (n=3). Figure 1 ID depicts that GASl expression is reduced in MeI-STR cells relative to MeI-STV cells. qRT-PCR analysis of Gas 1 expression in immortalized human melanocyte MeI-STV (vector) cells relative to MeI-STR (Ras-expressing) cells. Error bars represent SD (n=3). The results show that GASl expression is reduced ~6-fold in MeI-STR cells relative to MeI-STV cells. Figure 1 IE depicts that knockdown of Gasl does not increase the invasion of Bl 6-FO cells. Transwell/Boyden chamber assays comparing the invasion of B 16- F0/Gasl KD, B16-F0/NS, and B16-F10 cells. Cells were placed in the upper compartment of the Transwell/Boyden chamber and invasion and migration into the lower chamber was measured 48 hours later. Invasion percentage was determined by dividing the number of cells that invaded by the number of cells that migrated. Error bars indicate SEM (n=3). Figure 1 IF depicts that Gasl does not affect expression of GIi 1, a marker of Sonic hedgehog signaling, in Bl 6 mouse melanoma cells. Immunoblot analysis monitoring GIi 1 expression in B 16-FO cells expressing a Gasl or non-silencing (NS) shRNA, and in B16-F10 cells ectopically expressing Gasl or empty vector. Actin was monitored as a loading control. The results show that Gasl knockdown in B 16-FO cells or Gasl over-expression in B16-F10 cells does not affect Shh signaling.

Figure 12 depicts down-regulation of ACTA2, CTSO, SLC9A3R2 and DPPAS in human metastatic melanoma cell lines and tissue samples. (A) ACTA2, CTSO and SLC9A3R2 expression in benign nevi, primary melanoma and metastatic melanoma. (B) DPP A3 expression in advanced versus early stage melanoma. Data were obtained through the Oncomine database; a Student's t-test, performed directly though the

Oncomine 3.0 software, showed the difference in gene expression between the samples was significant.

Figure 13 depicts that knockdown of Gasl in the mouse breast cancer cell line 67NR increases lung metastasis. Left, mouse tail vein metastasis assay showing lungs from mice injected with 67NR cells stably expressing a Gasl or NS shRNA. Right, quantitation of lung weight. In contrast to the experiments performed with Bl 6 cells, in which the number of metastases could be directly counted, here metastasis was quantified by weighing the lungs; direct counting was not possible due to the large number of metastases, the large size of the metastatic nodules, and the lack of pigmentation. Error bars represent SEM (67NR/NS, n=9 mice; 67NR/Gasl KD, n=12 mice).

DETAILED DESCRIPTION OF THE INVENTION

Tumor metastasis utilizes, at least in part, the ability of metastatic cells to adhere to the proteins of the extracellular matrix (ECM), to migrate, and to survive at a distant location. Cell culture systems that have been developed where cancer cell lines are grown in a three-dimensional environment in vitro or ex vivo and have provided a controlled environment to study various tumor metastatic properties. These three- dimensional systems have shown that a variety of aggressive human tumor cell lines, including those derived from colorectal, prostate and non-small lung cancers, develop satellite colonies separate from the primary tumors in vitro or ex vivo (Doillon, C.J., et al., Anticancer Res. , 24, 2169-2177, (2004)). The present invention, in some aspects, presents a new strategy for the genome- wide discovery of genes that affect tumor metastasis. In some aspects, the present invention provides a novel three-dimensional cell culture assay system implemented in a genome-wide RNA interference (RNAi) screen to identify metastasis suppressor genes. Combination of the three-dimensional culture system with in vivo tumor metastasis and in vivo primary tumor growth assays has provided broad insight into the importance of these metastasis suppressor genes (MSGs) in tumor metastasis. As disclosed herein, these MSGs also provide a novel basis for identifying compounds and compositions that modulate tumor metastasis and provide new opportunities for treating metastatic disease. Cancer is disease characterized by uncontrolled cell proliferation and other malignant cellular properties. Cancer cells can arise from a number of genetic and epigenetic perturbations that cause defects in mechanisms controlling cell migration, proliferation, differentiation, and growth that lead to tumor formation and/or metastasis. As used herein, the term cancer includes, but is not limited to, the following types of cancer: breast cancer; biliary tract cancer; bladder cancer; brain cancer including glioblastomas and medulloblastomas; cervical cancer; choriocarcinoma; colon cancer; endometrial cancer; esophageal cancer; gastric cancer; hematological neoplasms including acute lymphocytic and myelogenous leukemia; T-cell acute lymphoblastic leukemia/lymphoma; hairy cell leukemia; chronic myelogenous leukemia, multiple myeloma; AIDS-associated leukemias and adult T-cell leukemia/lymphoma; intraepithelial neoplasms including Bowen's disease and Paget' s disease; liver cancer; lung cancer; lymphomas including Hodgkin's disease and lymphocytic lymphomas; neuroblastomas; oral cancer including squamous cell carcinoma; ovarian cancer including those arising from epithelial cells, stromal cells, germ cells and mesenchymal cells; pancreatic cancer; prostate cancer; rectal cancer; sarcomas including leiomyosarcoma, rhabdomyosarcoma, liposarcoma, fibrosarcoma, and osteosarcoma; skin cancer including melanoma, Merkel cell carcinoma, Kaposi's sarcoma, basal cell carcinoma, and squamous cell cancer; testicular cancer including germinal tumors such as seminoma, non-seminoma (teratomas, choriocarcinomas), stromal tumors, and germ cell tumors; thyroid cancer including thyroid adenocarcinoma and medullar carcinoma; and renal cancer including adenocarcinoma and Wilms tumor. Other cancers will be known to one of ordinary skill in the art. In one embodiment the cancer is a melanoma.

Tumors resulting from uncontrolled cell proliferation can be either benign or malignant. Whereas benign tumors remain localized in a primary tumor that remains localized at the site of origin and that is often self limiting in terms of tumor growth, malignant tumors have a tendency for sustained growth and an ability to spread or metastasize to distant locations. "Metastasis" as used herein refers to this spreading of malignant tumor cells and involves a diverse repertoire of malignant properties. These metastatic properties, as used herein, include cell invasion into tissues adjacent to primary tumors, migration through adjacent tissue, entry into the bloodstream or lymphatic system, dissemination through the bloodstream or lymphatic system, exit from the bloodstream or lymphatic system, and implantation at distant sites where new tumors can form. Other metastatic properties include aberrant cell proliferation, growth, survival. Still other metastatic properties will be apparent to one of ordinary skill in the art.

As used herein, "modulators of tumor metastasis" are endogenous or exogenous gene products, compounds, or compositions that affect one or more tumor metastatic properties. For example, tumor metastatic properties that can be affected include cell invasion, migration, entry and exit into blood vessels and the lymphatic systems, dissemination, implantation, proliferation, growth, and survival. Moreover, modulators of tumor metastasis can either enhance or suppress any one or more of the above tumor metastatic properties. As used herein, "enhance" means to increase or intensify tumor metastasis or a tumor metastatic property. For example, an increase in proliferation of a cell caused by modulator is an enhancement in a tumor metastatic property. As used herein, "suppress" means to reduce or inhibit, e.g., to reduce or inhibit tumor metastasis or a tumor metastatic property.. For example, a decrease in cell proliferation caused by modulator could be considered a suppression of a metastatic property. Suppression may, or may not, be complete. For example, cell proliferation may, or may not, be decreased to a state of complete arrest for a modulator's effect to be considered one of suppression. Moreover, suppression of a metastatic property may include the prevention of the acquisition or expression of a metastatic propertu, and the reduction of already existing metastatic property.

As used herein, "metastatic cancer cells" are cancer cells that metastasize. Correspondingly, non-metastatic cancer cells are cells that do not metastasize. Non- metastatic cancer cells may acquire mutations and/or epigenetic alterations that result in a conversion to metastatic cancer cells. Moreover, non-metastatic cancer cells may be perturbed genetically, such as with RNA interference, and/or chemically, such as with small-molecule compounds, to enhance certain metastatic properties. Similarly, as described herein, metastatic cancer cells may be perturbed genetically, such as with RNA interference and/or chemically such as with small-molecule compounds, to suppress one or more metastatic properties. The term "substantially non-metastatic" may be used in an experimental context with reference to a cell line or cells that are substantially less metastatic than a control metastatic cell line, or cells. "Substantially non-metastatic" may refer to one or more quantitative or qualitative assessments of one or more tumor metastatic properties, such as number of metastatic tumors or size of metastatic tumors. In these examples, a substantially non-metastatic cell line may produced substantially fewer (e.g., a statistically significant "fewer") metastatic tumors or substantially smaller metastatic tumors (e.g., a statistically significant "smaller") than a metastatic cell line. The term "substantially non-metastatic" includes cells and cell lines that are entirely non- metastatic. In one embodiment, a three-dimensional cell culture system is used to screen for novel regulator(s) of tumor metastasis. In some embodiments, a three-dimensional cell culture system comprises one or more cells embedded in an extracellular matrix. Typically, the cells embedded in an extracellular matrix are in a culture chamber such as a culture dish or plate well. A plate well may be in a multi-well plate having a number of wells selected from: 6, 12, 24, 96, 384, and 1536, but it is not so limited. The extracellular matrix may comprise one or more components such as collagen, fibrin, basement membrane, fibronectin, laminin, fibrillin, elastin, glycosaminoglycans, chitosan, alginate, proteoglycans, hyaluronan or other glycosaminoglycans. In some embodiments, the extracellular matrix in which the cells are embedded may comprise collagen selected from the group consisting of collagen I, II, III, IV, V, VI, VII, VIII, IX, X, XI and XII. In some embodiments, the three-dimensional cell culture system provides an experimental framework for systematically and efficiently identifying genes whose products enhance or suppress tumor metastasis. Moreover the three-dimensional culture system, as described herein, can be used to systematically and efficiently identify modulators of tumors metastasis. In some embodiments, the system comprises test cells and control cells that are embedded in collagen, coated with basement membrane matrix, such as the commercially available Matrigel product (BD Biosciences), and sandwiched into fibrin gel. These collagen, basement membrane, and fibrin layers provide an extracellular environment wherein metastatic cells have the appropriate extracellular cues to stimulate various aspects of their repertoire of metastatic properties, including migration and the formation of colonies removed from the original colony ("satellite colonies"). In contrast, non-metastatic cells remain viable and grow locally. As described herein, test or control cells in this three-dimensional in vitro or ex vivo system could be primary cells, non-immortalized cell lines, immortalized cell lines, transformed immortalized cell lines, benign tumor derived cell lines, malignant tumor derived cell lines, or transgenic cell lines. More than one set of control cells may be provided, such as non-metastatic and metastatic tumor derived cell lines. Cells in this system may be subjected to one or more genetic or chemical perturbations and then incubated for a predetermined time. The predetermined time is a time sufficient to produce a change in one or more tumor metastatic properties (e.g., as reflected in the number of satellite colonies) in a control cell. In one embodiment, an RNAi-based screen identifies genes that modulate tumor metastatic properties when knocked down. The methods of this screen are applicable to the use of libraries comprising RNAi based modalities consisting of from a single gene to all, or substantially all, known genes in a organism under investigation. In one embodiment the screen uses a mouse shRNA-mir library comprising about 62,400 shRNA-mirs directed against about 28,000 genes that are divided into pools, which are packaged into retrovirus particles and used to stably transduce substantially non- metastatic cancer cells (see Examples). Methods for viral packaging and transduction of cells, including those described herein, are well known to one of ordinary skill in the art. The library described herein, utilizes a mir-30-based shRNA (shRNAmir) expression vector in which shRNA is encoded in carrier that it is flanked by approximately 125 bases 5' and 3' of the pre-miR-30 sequence. Expression vectors can employ either polymerase II or polymerase III promoters to drive expression of these shRNAs and result in functional siRNAs in cells. The former polymerase permits the use of classic protein expression strategies, including inducible and tissue-specific expression systems. Other library compilations, such Lentiviral-based systems and libraries directed against human sequences, are readily available and well known to one of ordinary skill in the art. Also, the method is readily conducive to screening cDNA-based expression libraries to identify genes that modulate tumor metastatic properties when exogenously expressed. An expression vector is one into which a desired sequence may be inserted, e.g. , by restriction and ligation, such that it is operably joined to regulatory sequences and may be expressed as an RNA transcript. An expression vector typically contains an insert that is a coding sequence for a protein or for a functional RNA such as an shRNA, a miRNA, or an shRNA-mir. Vectors may further contain one or more marker sequences suitable for use in the identification of cells that have or have not been transformed or transfected with the vector. Markers include, for example, genes encoding proteins that increase or decrease either resistance or sensitivity to antibiotics or other compounds, genes that encode enzymes whose activities are detectable by standard assays known in the art (e.g., β-galactosidase or alkaline phosphatase), and genes that visibly affect the phenotype of transformed or transfected cells, hosts, colonies or plaques (e.g., green fluorescent protein). As used herein, a coding sequence (e.g., protein coding sequence, miRNA sequence, shRNA sequence) and regulatory sequences are said to be "operably" joined when they are covalently linked in such a way as to place the expression or transcription of the coding sequence under the influence or control of the regulatory sequences. If it is desired that the coding sequences be translated into a functional protein, two DNA sequences are said to be operably joined if induction of a promoter in the 5' regulatory sequences results in the transcription of the coding sequence and if the nature of the linkage between the two DNA sequences does not (1) result in the introduction of a frame-shift mutation, (2) interfere with the ability of the promoter region to direct the transcription of the coding sequences, or (3) interfere with the ability of the corresponding RNA transcript to be translated into a protein. Thus, a promoter region would be operably joined to a coding sequence if the promoter region were capable of effecting transcription of that DNA sequence such that the resulting transcript might be translated into the desired protein or polypeptide. It will be appreciated that a coding sequence need not encode a protein but may instead, for example, encode a functional RNA such as an miRNA, shRNA or shRNA-mir.

The precise nature of the regulatory sequences needed for gene expression may vary between species or cell types, but shall in general include, as necessary, 5' non- transcribed and 5' non-translated sequences involved with the initiation of transcription and translation respectively, such as a TATA box, capping sequence, CAAT sequence, and the like. Such 5' non-transcribed regulatory sequences will include a promoter region that includes a promoter sequence for transcriptional control of the operably joined gene. Regulatory sequences may also include enhancer sequences or upstream activator sequences as desired. The vectors of the invention may optionally include 5' leader or signal sequences. The choice and design of an appropriate vector is within the ability and discretion of one of ordinary skill in the art. Exemplary regulatory sequences for expression of interfering RNA (e.g., shRNA, miRNA) are disclosed herein. One of skill in the art will be aware of these and other appropriate regulatory sequences for expression of interfering RNA, e.g., shRNA, miRNA, etc. In some embodiments, a virus vector for delivering a nucleic acid molecule is selected from the group consisting of adenoviruses, adeno-associated viruses, poxviruses including vaccinia viruses and attenuated poxviruses, Semliki Forest virus, Venezuelan equine encephalitis virus, retroviruses, Sindbis virus, and Ty virus-like particle. Examples of viruses and virus-like particles which have been used to deliver exogenous nucleic acids include: replication-defective adenoviruses (e.g., Xiang et al., Virology 219:220-227, 1996; Eloit et al., J Virol. 7:5375-5381, 1997; Chengalvala et al., Vaccine 15:335-339, 1997), a modified retrovirus (Townsend et al., J. Virol. 71 :3365-3374, 1997), a nonreplicating retrovirus (Irwin et al., J. Virol. 68:5036-5044, 1994), a replication defective Semliki Forest virus (Zhao et al., Proc. Natl. Acad. ScL USA 92:3009-3013, 1995), canarypox virus and highly attenuated vaccinia virus derivative (Paoletti, Proc. Natl. Acad. ScL USA 93:11349-11353, 1996), non-replicative vaccinia virus (Moss, Proc. Natl. Acad. ScL USA 93:11341-11348, 1996), replicative vaccinia virus (Moss, Dev. Biol. Stand. 82:55-63, 1994), Venzuelan equine encephalitis virus (Davis et al., J. Virol. 70:3781-3787, 1996), Sindbis virus (Pugachev et al., Virology 212:587-594, 1995), lentiviral vectors (Naldini L, et al., Proc Natl Acad Sci USA. 1996 Oct 15;93(21):11382-8) and Ty virus-like particle (Allsopp et al., Eur. J. Immunol 26:1951-1959, 1996).

Another virus useful for certain applications is the adeno-associated virus, a double-stranded DNA virus. The adeno-associated virus is capable of infecting a wide range of cell types and species and can be engineered to be replication-deficient. It further has advantages, such as heat and lipid solvent stability, high transduction frequencies in cells of diverse lineages, including hematopoietic cells, and lack of superinfection inhibition thus allowing multiple series of transductions. The adeno- associated virus can integrate into human cellular DNA in a site-specific manner, thereby minimizing the possibility of insertional mutagenesis and variability of inserted gene expression. In addition, wild-type adeno-associated virus infections have been followed in tissue culture for greater than 100 passages in the absence of selective pressure, implying that the adeno-associated virus genomic integration is a relatively stable event. The adeno-associated virus can also function in an extrachromosomal fashion.

In general, other useful viral vectors are based on non-cytopathic eukaryotic viruses in which non-essential genes have been replaced with the gene of interest. Non- cytopathic viruses include certain retroviruses, the life cycle of which involves reverse transcription of genomic viral RNA into DNA with subsequent proviral integration into host cellular DNA. In general, the retroviruses are replication-deficient (i.e., capable of directing synthesis of the desired transcripts, but incapable of manufacturing an infectious particle). Such genetically altered retroviral expression vectors have general utility for the high-efficiency transduction of genes in vivo. Standard protocols for producing replication-deficient retroviruses (including the steps of incorporation of exogenous genetic material into a plasmid, transfection of a packaging cell lined with plasmid, production of recombinant retroviruses by the packaging cell line, collection of viral particles from tissue culture media, and infection of the target cells with viral particles) are provided in Kriegler, M., "Gene Transfer and Expression, A Laboratory Manual," W.H. Freeman Co., New York (1990) and Murry, EJ. Ed. "Methods in Molecular Biology," vol. 7, Humana Press, Inc., Clifton, New Jersey (1991).

Various techniques may be employed for introducing nucleic acid molecules of the invention into cells, depending on whether the nucleic acid molecules are introduced in vitro or in vivo in a host. Such techniques include transfection of nucleic acid molecule-calcium phosphate precipitates, transfection of nucleic acid molecules associated with DEAE, transfection or infection with the foregoing viruses including the nucleic acid molecule of interest, liposome-mediated transfection, and the like. Other examples include: N-TER™ Nanoparticle Transfection System by Sigma-Aldrich, FectoFly™ transfection reagents for insect cells by Polyplus Transfection, Polyethylenimine "Max" by Polysciences, Inc., Unique, Non- Viral Transfection Tool by Cosmo Bio Co., Ltd., Lipofectamine™ LTX Transfection Reagent by Invitrogen, SatisFection™ Transfection Reagent by Stratagene, Lipofectamine™ Transfection Reagent by Invitrogen, FuGENE® HD Transfection Reagent by Roche Applied Science, GMP compliant in vivo-jetPEI™ transfection reagent by Polyplus Transfection, and Insect GeneJuice® Transfection Reagent by Novagen. Other RNAi-based modalities may be employed in the methods disclosed herein, such as siRNA-based oligonucleotides and/or siRNA-based oligonucleotides modified to alter potency, target affinity, the safety profile and/or the stability, for example, to render them resistant or partially resistant to intracellular degradation. Modifications, such as phosphorothioates, for example, can be made to oligonucleotides to increase resistance to nuclease degradation, binding affinity and/or uptake. In addition, hydrophobization and bioconjugation enhances siRNA delivery and targeting (De Paula et al., RNA 13(4):431-56, (2007)) and siRNAs with ribo-difiuorotoluyl nucleotides maintain gene silencing activity (Xia et al., ASC Chem. Biol. , 1 (3): 176-83,(2006)). siRNAs with amide-linked oligoribonucleosides have been generated that are more resistant to Sl nuclease degradation (Iwase, R et al., 2006 Nucleic Acids Symp, Ser 50: 175-176). In addition, modification of siRNA at the 2'-sugar position and phosphodiester linkage confers improved serum stability without loss of efficacy (Choung et al., Biochem.

Biophys. Res. Commun. 342(3):919-26, (2006)). In one study, 2'-deoxy-2'-fluoro-beta- D-arabinonucleic acid (FANA)-containing antisense oligonucleotides compared favorably to phosphorothioate oligonucleotides, 2'-O-methyl-RNA/DNA chimeric oligonucleotides and siRNAs in terms of suppression potency and resistance to degradation (Ferrari, N et al., Ann N Y Acad Sc, 1082: 91-102, (2006)).

Other molecules that can be screened using the methods of the invention and/or used in the methods of the invention include sense and antisense nucleic acids (single or double stranded), ribozymes, peptides, DNAzymes, peptide nucleic acids (PNAs), triple helix forming oligonucleotides, antibodies, and aptamers and modified form(s) thereof directed to sequences in gene(s), RNA transcripts, or proteins. Antisense and ribozyme suppression strategies have led to the reversal of a tumor phenotype by reducing expression of a gene product or by cleaving a mutant transcript at the site of the mutation (Carter et al., Br. J. Cancer, 67(5):869-76, (1993); Lange et al., Leukemia, 6(l l):1786-94 (1993); Valera et al., J. Biol. Chem., 269(46):28543-6 (1994); Dosaka-Akita et al., Am. J. Clin. Pathol, 102(5):660-4 (1994); Feng et al., Cancer Res., 55(10):2024-8 (1995); Quattrone et al., Cancer Res., 55(l):90-5 (1995); Lewin et al., Nat Med, 4(8):967-71 (1998)). For example, neoplastic reversion was obtained using a ribozyme targeted to an H-Ras mutation in bladder carcinoma cells (Feng et al., Cancer Res., 55(10):2024-8 (1995)). Ribozymes have also been proposed as a means of both inhibiting gene expression of a mutant gene and of correcting the mutant by targeted trans-splicing (Sullenger et al., Nature, 371(6498):619-22 (1994); Jones et al., Nat. Med, 2(6):643-8 (1996)). Ribozyme activity may be augmented by the use of, for example, non-specific nucleic acid binding proteins or facilitator oligonucleotides (Herschlag et al., Embo J., 13(12):2913-24 (1994); Jankowsky et al., Nucleic Acids Res., 24(3):423-9 (1996)). Multitarget ribozymes (connected or shotgun) have been suggested as a means of improving efficiency of ribozymes for gene suppression (Ohkawa et al., Nucleic Acids Symp Ser., (29):121-2, (1993)). Triple helix approaches have also been investigated for sequence-specific gene suppression. Triplex forming oligonucleotides have been found in some cases to bind in a sequence-specific manner (Postel et al., Proc. Natl. Acad. Sci. U.S.A., 88(18):8227-31 (1991); Duval-Valentin et al., Proc. Natl. Acad. Sci. U.S.A., 89(2):504-8 (1992); Hardenbol et al., Proc. Natl. Acad. Sci. U.S.A. , 93(7):2811 -6 (1996); Porumb et al.,

Cancer Res., 56(3):515-22 (1996)). Similarly, peptide nucleic acids have been shown to inhibit gene expression (Hanvey et al., Antisense Res. Dev., l(4):307-17 (1991); Knudsen et al., Nucleic Acids Res., 24(3):494-500 (1996); Taylor et al., Arch. Surg., 132(11):1177-83 (1997)). Minor-groove binding polyamides can bind in a sequence- specific manner to DNA targets and hence may represent useful small molecules for future suppression at the DNA level (Trauger et al., Chem. Biol, 3(5):369-77 (1996). In addition, suppression has been obtained by interference at the protein level using dominant negative mutant peptides and antibodies (Herskowitz, Nature, 329(6136):219- 22 (1987); Rimsky et al., Nature, 341(6241):453-6 (1989); Wright et al., Proc. Natl. Acad. Sci. USA 86(9):3199-203 (1989)). In some cases suppression strategies have lead to a reduction in RNA levels without a concomitant reduction in proteins, whereas in others, reductions in RNA have been mirrored by reductions in protein.

The diverse array of suppression strategies that can be employed includes the use of DNA and/or RNA aptamers that can be selected to target, for example, a protein of interest such as a MSG. For example, in the case of age related macular degeneration (AMD), anti-VEGF aptamers have been generated and have been shown to provide clinical benefit in some AMD patients (Ulrich, H, et al. Comb. Chem. High Throughput Screen, 9: 619-632 (2006)). Suppression and replacement using aptamers for suppression in conjunction with a modified replacement gene and encoded protein that is refractory or partially refractory to aptamer-based suppression could be used in the invention.

In one embodiment, experimental systems are contemplated in which a large set of samples, such as the genome-wide shRNA-mir library disclosed herein, is screened without pooling. Such systems could make use of high-throughput biological techniques and equipment, such laboratory automation and sample tracking processes well known to one of ordinary skill in the art. In one embodiment cells are embedded in collagen, coated with a basement membrane matrix (Matrigel), sandwiched into a fibrin gel, and then incubated for about predetermined period of time {e.g., about two weeks). Other three dimensional cell culture systems are known to one of ordinary skill in the art and could be employed to assess certain tumor metastatic properties. In certain embodiments, substantially non- metastatic control B 16-FO cells produced between about five and about ten satellite colonies whereas B16-F10 melanoma-derived cells, a metastatic positive control cell, produced between about 1000 and about 2000 satellite colonies. Other non-metastatic, or substantially non-metastatic control cells, and metastatic control cells isolated from melanoma tissue or tissues of other cancers described herein, are contemplated. Also, other tumor metastatic properties can be assessed, either quantitatively or qualitatively, such as the distance of the colonies from the primary tumor site, and/or the size of the colonies. Moreover, differences of orders of magnitude other then those shown here may be considered significant, or substantial, in other contexts. In one embodiment, a mouse tail vein/lung metastasis assay is described in which cells were injected intravenously and metastasized to lungs (see Examples). Other like assays are contemplated as useful in accordance with the invention, and are known to one of ordinary skill in the art. For example, cells may be injected intraperitoneally or subcutaneously, followed by an assessment of the formation of metastatic tumors at a site different from the injection site. Moreover, following administration of a test cell, or a plurality thereof, to a subject, the cell or cells can metastasize to one or more distant organ sites other than lung, including one or more of the lymph nodes, skeletal muscle, heart, brain, liver, spleen, kidney, or spinal cord.

In other embodiments in silico sources can be investigated to obtain information on molecular functions, as well as genetic, biological, and clinical features of gene products that are discovered to modulate one or more metastatic properties. Cancer gene databases, and the like, may be mined to, for example, determine expression of the gene products in metastatic and non-metastatic tumors. In other embodiments, in silico sources can serve as a starting point for identifying candidate modulators of one or more metastatic properties. Subsequent studies involving the in vivo or ex vivo and/or in vivo systems described herein may be performed to further investigate the ability of candidate modulators to affect one or more metastatic properties. In one embodiment certain metastasis suppressor genes identified herein provide a basis for identifying compounds useful as pharmacological agents for the suppression of various tumor metastatic properties and/or treatment of metastatic tumors. In this aspect compounds are contacted with test cells (and preferably control cells) at a predetermined dose. In one embodiment the dose may be about up to InM. In another embodiment the dose may be between about InM and about 10OnM. In another embodiment the dose may be between about 10OnM and about lOuM. In another embodiment the dose may be at or above lOuM. Following incubation for an appropriate predetermined time, the effect of compounds on the expression of the one or more metastasis suppressors genes (MSGs), for example as outlined in Table 1, is determined by an appropriate method known to one of ordinary skill in the art, such as quantitative RT-PCR. Compounds that substantially alter the expression of one or more metastasis suppressors genes can be used for treatment and/or can be examined further.

In one embodiment, quantitative RT-PCR is employed to examine the expression of tumor suppressor genes. Other methods known to one of ordinary skill in the art could be employed to analyze mRNA levels, for example microarray analysis, cDNA analysis, Northern analysis, and RNase Protection Assays. In some embodiments, analysis of tumor suppressor protein levels is performed to examine tumor suppressor gene expression. Protein levels can be determined using any appropriate method known in the art. Exemplary methods include immunoassays, immunoblotting, immunoprecipitation, mass spectroscopy, spectrophotometry, enzymatic assays, and ELISA. Other appropriate methods will be apparent to the skilled artisan.

In some embodiments, oligonucleotide arrays (e.g., microarrays, mRNA detection arrays, genomic DNA detection arrays) or bead-based nucleic acid assay systems are provided for evaluating expression of a plurality of MSGs in parallel. In some embodiments, the oligonucleotide arrays consist essentially of immobilized nucleic acid probes (e.g., oligonucleotide probes) that hybridize with a plurality of MSGs, and optionally one or more control genes. In some embodiments, the oligonucleotide arrays consist essentially of immobilized nucleic acid probes that hybridize with at least 2, at least 5, at least 10, at least 20, at least 30, at least 40, at least 50, at least 60, at least 70, at least 80, or more MSGs. In some embodiments, the oligonucleotide arrays consist essentially of immobilized nucleic acid probes (e.g., oligonucleotide probes) that hybridize with a plurality of MSGs selected from: ACTA2, ADAMTS 16, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTC17, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orf24, PRAMEF8, 4931422A03RIK, and KIAA1276.

In other embodiments, protein detection systems (e.g., protein detection arrays, e.g., antibody arrays) are provided for evaluating expression of a plurality of MSGs in parallel. In some embodiments, protein detection arrays are provided that consist essentially of antibodies or fragments of antibodies, which are immobilized on a solid support, that specifically bind to a plurality of proteins encoded by MSGs, and optionally one or more control proteins. In some embodiments, the protein detection arrays consist essentially of at least 2, at least 5, at least 10, at least 20, at least 30, at least 40, at least 50, at least 60, at least 70, at least 80, or more antibodies or fragments of antibodies, which are immobilized on a solid support, that specifically bind to a plurality of proteins encoded by MSGs, e.g., MSGs selected from: ACTA2, ADAMTS 16, AGL, ALG6,

ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CP A2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC 19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTC17, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orf24, PRAMEF8, 4931422A03RIK, and KIAA1276. In one embodiment, further evaluation of compounds that substantially alter the expression of MSGs employs the three-dimensional cell culture system. Putative tumor metastasis modulator compounds can be examined for their ability to affect tumor metastatic properties in vitro or ex vivo. Test cells in this system may be subjected to a one or more chemical perturbations (e.g., contacted with a compound) and incubated for a time sufficient to produce a change in one or more metastatic properties (e.g., number of satellite colonies) in a control cell. Compounds that substantially alter the metastatic properties of the test cells can be used for treatment and/or can be examined further.

In one embodiment further evaluation of compounds that alter one or more tumor metastatic properties of the test cells in the in vitro or ex vivo three-dimensional cell culture assays and/or that substantially alter the expression of metastatic suppressor genes are examined to determine if they affect metastasis in vivo, for example by using the mouse tail vein/lung metastasis assays described herein. Incubation of test compounds at various dose and time schedules are contemplated wherein after a predetermined period of time, for example about two weeks, mice are examined for evidence of tumor metastasis. Typically this predetermined time is a time sufficient to produce a significant change in tumor metastasis in a control experiment. Moreover, the effects of compounds can be determined in the context of different dose and time schedules.

In some cases, evaluation of modulators of one or more metastatic properties, such as compounds or compositions, may involve implantation of tumors comprising test and control cells. Tumors can be implanted by various routes known to one of ordinary skill in the art and described herein. In vivo evaluation of modulators of one or more metastatic properties may comprise one or more assays for alterations in one or more tumor metastatic properties. In certain cases, assaying for alterations in one or more tumor metastatic properties may involve resecting metastatic or primary tumor tissue and performing an assay on the resected tissue, such as by various histological staining, hybridization, and/or immunolabeling techniques known in the art. Cells may also be isolated from resected tumor tissues, and subjected to various cytological assays known in the art, so as to assess cellular and molecular mechanisms influencing one or more metastatic properties. Cell lines may also be established from cells in resection tissue. One aspect of the invention contemplates the treatment of a subject having or at risk of having tumor metasasis. In other aspects the invention provides methods for detecting a metastatic cancer in a subject. As used herein a subject is a mammalian species, including but not limited to a dog, cat, horse, cow, pig, sheep, goat, chicken, rodent, or primate. Subjects can be house pets (e.g., dogs, cats), agricultural stock animals (e.g., cows, horses, pigs, chickens, etc.), laboratory animals (e.g., mice, rats, rabbits, etc.), zoo animals (e.g., lions, giraffes, etc.), but are not so limited. Preferred subjects are human subjects. The human subject may be a pediatric, adult or a geriatric subject.

As used herein "treatment", or "treating", includes amelioration, cure or maintenance (i.e., the prevention of relapse) of a disorder. Treatment after a disorder has started aims to reduce, ameliorate or altogether eliminate the disorder, and/or its associated symptoms, to prevent it from becoming worse, or to prevent the disorder from re-occurring once it has been initially eliminated (i.e., to prevent a relapse).

As used herein, a "clinical sample" refers to an isolated biomolecule, such as DNA, RNA, or protein, an isolated cell, an isolated tissue, saliva, gingival secretions, cerebrospinal fluid (spinal fluid), gastrointestinal fluid, mucus, urogenital secretions, synovial fluid, blood, serum, plasma, urine, cystic fluid, lymph fluid, ascites, pleural effusion, interstitial fluid, intracellular fluid, ocular fluids, seminal fluid, mammary secretions, vitreal fluid, stool, and nasal secretions. In some instances, a clinical sample is a tissue biopsy, such as a tumor or cancer biopsy. However, clinical samples are not so limited and other exemplary clinical samples will be readily apparent to one of ordinary skill in the art. As used herein, "obtaining a clinical sample" refers to any process for directly or indirectly acquiring a clinical sample from a subject (patient). For example, a clinical sample may be obtained (e.g., at a point-of-care facility) by procuring a tissue sample (e.g., a cancer tissue sample) from a subject or procuring a specimen, such as a blood or urine sample, produced by the subject. Alternatively, a clinical sample may be obtained by receiving the clinical sample (e.g., at a laboratory facility) from one or more individuals who procured the sample from the subject

As used herein, "gene therapy" is a therapy focused on treating genetic diseases, such as cancer, by the delivery of one or more expression vectors encoding therapeutic gene products, including polypeptides or RNA molecules, to diseased cells. Methods for construction and delivery of expression vectors are disclosed herein and will be known to one of ordinary skill in the art. One embodiment herein contemplates the use of gene therapy to deliver one or more expression vectors encoding one or more metastasis suppressor genes, particularly one or more of the metastasis suppressor genes in Table 1, to inhibit or prevent tumor metastasis, or to treat a subject in need thereof.

As used herein a "therapeutically effective amount" is an amount of a compound or composition capable of sufficiently and substantially inhibiting tumor metastasis, or one or more tumor metastatic properties. A therapeutically effective amount can refer to any compounds or compositions described herein that have tumor metastasis inhibitory properties. Methods for establishing a therapeutically effective amount for any compounds or compositions described herein are known to one of ordinary skill in the art. As used herein pharmacological agents comprise compounds or compositions and a pharmaceutically acceptable carrier that have therapeutic utility, i.e., that facilitate delivery of compounds or compositions in a therapeutically effective amount.

As described above compounds or compositions that substantially alter the expression of one or more metastasis suppressor genes and/or that are potential modulators of metastatic properties and can be discovered using the disclosed test methods. Examples of types of compounds or compositions that may be tested include, but are not limited to: anti-metastatic agents, cytotoxic agents, cytostatic agents, cytokine agents, antiproliferative agents, immunotoxin agents, gene therapy agents, angiostatic agents, cell targeting agents, etc. As disclosed herein test compounds can be examined by the in vivo and/or in vitro or ex vivo experimental systems described herein. For in vivo studies, a test compound can be administered before cells are administered; at the same or about the same time as cells are administered, or after cells are administered. Cells and test compound(s) can be administered by the same or different routes. The effect on tumor development can be assessed by determining whether the test compound affects the size, location, and/or number of primary and/or metastatic tumors in the subject.

The following provides further examples of test compounds and is not meant to be limiting. Those of ordinary skill in the art will recognize that there are numerous additional types of suitable test compounds that may be tested using the methods, cells, and/or animal models of the invention. Test compounds can be small molecules (e.g., compounds that are members of a small molecule chemical library). The compounds can be small organic or inorganic molecules of molecular weight below about 3,000 Daltons. The small molecules can be, e.g., from at least about 100 Da to about 3,000 Da (e.g., between about 100 to about 3,000 Da, about 100 to about 2,500 Da, about 100 to about 2,000 Da, about 100 to about 1,750 Da, about 100 to about 1,500 Da, about 100 to about 1,250 Da, about 100 to about 1,000 Da, about 100 to about 750 Da, about 100 to about 500 Da, about 200 to about 1500, about 500 to about 1000, about 300 to about 1000 Da, or about 100 to about 250 Da). Test compounds can also be microorganisms, such as bacteria (e.g. , Escherichia coli, Salmonella typhimurium, Mycobacterium avium, or Bordetella pertussis), fungi, and protists (e.g., Leishmania amazonensis), which may or may not be genetically modified. See, e.g., U.S. Patents No. 6,190,657 and 6,685,935 and U.S. Patent Applications No. 2005/0036987 and 2005/0026866.

The small molecules can be natural products, synthetic products, or members of a combinatorial chemistry library. A set of diverse molecules can be used to cover a variety of functions such as charge, aromaticity, hydrogen bonding, flexibility, size, length of side chain, hydrophobicity, and rigidity. Combinatorial techniques suitable for synthesizing small molecules are known in the art (e.g. , as exemplified by Obrecht et al., Solid-Supported Combinatorial and Parallel Synthesis of Small-Molecular-Weight Compound Libraries, Pergamon-Elsevier Science Limited (1998)), and include those such as the "split and pool" or "parallel" synthesis techniques, solid-phase and solution- phase techniques, and encoding techniques (see, for example, Czarnik, A. W., Curr. Opin. Chem. Biol, 1:60 (1997)). In addition, a number of small molecule libraries are publicly or commercially available (e.g., through Sigma- Aldrich, TimTec (Newark, DE), Stanford School of Medicine High-Throughput Bioscience Center (HTBC), and ChemBridge Corporation (San Diego, CA).

Compound libraries screened using the new methods can comprise a variety of types of test compounds. A given library can comprise a set of structurally related or unrelated test compounds, hi some embodiments, the test compounds are peptide or peptidomimetic molecules, hi some embodiments, test compounds include, but are not limited to, peptide analogs including peptides comprising non-naturally occurring amino acids, phosphorous analogs of amino acids, amino acids having non-peptide linkages, or other small organic molecules. In some embodiments, the test compounds are peptidomimetics (e.g., peptoid oligomers, e.g., peptoid amide or ester analogues, D- peptides, L-peptides, oligourea or oligocarbamate); peptides (e.g., tripeptides, tetrapeptides, pentapeptides, hexapeptides, heptapeptides, octapeptides, nonapeptides, decapeptides, or larger, e.g., 20-mers or more); cyclic peptides; other non-natural or unnatural peptide-like structures; and inorganic molecules (e.g., heterocyclic ring molecules). Test compounds can also be nucleic acids. The test compounds and libraries thereof can be obtained by systematically altering the structure of a first "hit" compound that has a chemotherapeutic (e.g., anti- metastatic) effect, and correlating that structure to a resulting biological activity (e.g., a structure-activity relationship study).

Such libraries can be obtained using any of the numerous approaches in combinatorial library methods known in the art, including: biological libraries; peptoid libraries (libraries of molecules having the functionalities of peptides, but with a novel, non-peptide backbone which are resistant to enzymatic degradation but which nevertheless remain bioactive; see, e.g., Zuckermann, et al., J. Med. Chem., 37:2678-85 (1994); spatially addressable parallel solid phase or solution phase libraries; synthetic library methods requiring decon volution; the "one-bead one-compound" library method; and synthetic library methods using affinity chromatography selection (Lam, Anticancer Drug Des. 12:145 (1997)). Examples of methods for the synthesis of molecular libraries can be found in the art, for example in: DeWitt et al. , Proc. Natl. Acad. Sci. USA 90, 6909 (1993); Erb et al., Proc. Natl. Acad. Sci. USA, 91 :11422 (1994); Zuckermann et al., J. Med. Chem., 37:2678 (1994); Cho et al., Science, 261 :1303 (1993); Carrell et al.,

Angew. Chem. Int. Ed. Engl, 33:2059 (1994); Carell et al., Angew. Chem. Int. Ed. Engl, 33:2061 (1994); and in Gallop et al., J Med. Chem., 37:1233 (1994). Libraries of compounds can be presented in solution (e.g., Houghten, Biotechniques, 13:412-421 (1992)), or on beads (Lam, Nature, 354:82-84 (1991)), chips (Fodor, Nature, 364:555- 556 (1993)), bacteria (Ladner, USP 5,223,409), spores (Ladner, U.S. Patent No.

5,223,409), plasmids (Cull et al., Proc. Natl Acad. Sci. USA, 89:1865-1869 (1992)) or on phage (Scott et al., Science, 249:386-390 (1990); Devlin, Science, 249:404-406 (1990); Cwirla et al., Proc. Natl. Acad. Sci. USA, 87:6378-6382 (1990); Felici, J. MoI Biol, 222:301-310 (1991); Ladner, supra.). Certain results may be clinically beneficial, such as if the test compound is a suppressor of tumor metastasis, or activator of tumor metastatic suppressor genes, such as those disclosed herein (See Table 1). Still other clinically beneficial results include: (a) inhibition or arrest of primary tumor growth, (b) inhibition of any tumor metastatic properties and (c) extension of survival of a test subject. Compounds with clinically beneficial results are potential chemotherapeutics, and may be formulated as such.

Compounds identified as having a chemotherapeutic or anti-metastatic effect can be selected and systematically altered, e.g., using rational design, to optimize binding affinity, avidity, specificity, or other parameters. Such optimization can also be screened for using the methods described herein. Thus, one can screen a first library of small molecules using the methods described herein, identify one or more compounds that are "hits," (by virtue of, for example, induction of expression of one or more MSGs and/or their ability to reduce the size and/or number of rumors, e.g. , at the original site of implantation and at metastasis sites), and subject those hits to systematic structural alteration to create a second library of compounds structurally related to the hit. The second library can then be screened using the methods described herein.

A variety of techniques useful for determining the structures of compounds are known and can be used in the methods described herein (e.g. , NMR, mass spectrometry, gas chromatography equipped with electron capture detectors, fluorescence, and absorption spectroscopy).

Assays of chemotherapeutic activity of test compounds may be conducted in vitro or ex vivo and/or in vivo using cells and methods of the invention. For example, a test compound may be administered to a nonhuman subject to which has been administered (e.g., implanted or injected with) a plurality of the cells described herein, e.g., a number of cells sufficient to induce the formation of one or more tumors (e.g., metastatic tumors). The nonhuman subject can be, e.g. , a rodent (e.g. , a mouse). The test compound can be administered to the subject by any regimen known in the art. For example, the test compound can be administered prior to, concomitant with, and/or following the administration of cells of the invention. A test compound can also be administered regularly throughout the course of the method, for example, one, two, three, four, or more times a day, weekly, bi-weekly, or monthly, beginning before or after cells of the invention have been administered. In other embodiments, the test compound is administered continuously to the subject (e.g., intravenously). The dose of the test compound to be administered can depend on multiple factors, including the type of compound, weight of the subject, frequency of administration, etc. Determination of dosages is routine for one of ordinary skill in the art. Typical dosages are 0.01-200 mg/kg (e.g., 0.1-20 or 1-10 mg/kg).

The size and/or number of tumors in the subject can be determined following administration of the tumor cells and the test compound. The size and/or number of tumors can be determined non-invasively by any means known in the art. For example, tumor cells that are fluorescently labeled (e.g., by expressing a fluorescent protein such as GFP) can be monitored by various tumor-imaging techniques or instruments, e.g., non-invasive fluorescence methods such as two-photon microscopy. The size of a tumor implanted subcutaneously can be monitored and measured underneath the skin. To determine whether a compound affects tumor formation or metastasis, the size and/or number of tumors in the subject can be compared to a reference standard (e.g., a control value). A reference standard can be a control subject which has been given the same regimen of administration of tumor cells and test compound, except that the test compound is omitted or administered in an inactive form. Alternately, a compound believed to be inert in the system can be administered. A reference standard can also be a control subject which has been administered non-tumor cells and test compound, non- tumor cells and no test compound, or non-tumor cells and an inactive test compound. The reference standard can also be a numerical figure or figures representing the size and/or number of tumors expected in an untreated subject. This numerical figure(s) can be determined by observation of a representative sample of untreated subjects. A reference standard may also be the test animal before administration of the compound. The assay methods disclosed herein are amenable to high-throughput screening (HTS) implementations. In some embodiments, the screening assays of the invention are high throughput or ultra high throughput (e.g., Fernandes, P. B., Curr Opin Chem Biol. 1998 2:597; Sundberg, S A, Curr Opin Biotechnol. 2000, 11 :47). HTS includes testing of up to, and including, 100,000 compounds per day, whereas ultra high throughput (uHTS) includes screening in excess of 100,000 compounds per day. The assay methods disclosed herein may be carried out in a multi-well format, for example, a 96-well, 384- well format, or 1,536-well format, and are suitable for automation. In the high throughput assays of the invention, it is possible to screen several thousand different compounds or compositions in a single day. In particular, each well of a microtiter plate can be used to run a separate assay against a selected test compound, or, if concentration or incubation time effects are to be observed, a plurality of wells can contain test samples of a single compound. It is possible to assay many plates per day; assay screens for up to about 6,000, 20,000, 50,000, or more than 100,000 different compounds are possible using the assays of the invention. Typically, HTS implementations of the assays disclosed herein involve the use of automation. In some embodiments, an integrated robot system consisting of one or more robots transports assay microplates between multiple assay stations for compound, cell and/or reagent addition, mixing, incubation, and finally readout or detection. In some aspects, an HTS system of the invention may prepare, incubate, and analyze many plates simultaneously, further speeding the data- collection process. High throughput screening implementations are well known in the art. Exemplary methods are also disclosed in High Throughput Screening: Methods and Protocols (Methods in Molecular Biology) by William P. Janzen (2002) and High- Throughput Screening in Drug Discovery (Methods and Principles in Medicinal Chemistry) (2006) by Jόrg Huser, the contents of which are both incorporated herein by reference in their entirety.

The practice of the present invention employs, unless otherwise indicated, conventional techniques of molecular biology (including recombinant techniques), microbiology, cell biology, biochemistry and immunology, which are within the skill of the art. Such techniques are explained fully in the literature, such as, Molecular Cloning: A Laboratory Manual, second edition (Sambrook et al. 1989) Cold Spring Harbor Press; Oligonucleotide Synthesis (M.J. Gait, ed. 1984); Methods in Molecular Biology, Humana Press; Cell Biology: A Laboratory Notebook (J.E. Cellis, ed. 1998) Academic Press; Animal Cell Culture (RJ. Freshney, ed. 1987); Introduction to Cell and Tissue Culture ( J.P. Mather and P.E. Roberts 1998) Plenum Press; Cell and Tissue Culture: Laboratory Procedures (A. Doyle, J.B. Griffiths, and D.G. Newell, eds. 1993-8) J. Wiley and Sons; Methods in Enzymology (Academic Press, Inc.); Handbook of Experimental Immunology (D.M. Weir and CC. Blackwell, eds.); Gene Transfer Vectors for Mammalian Cells (J.M. Miller and M.P. Calos, eds. 1987); Current Protocols in Molecular Biology (F.M. Ausubel et al., eds. 1987); PCR: The Polymerase Chain Reaction, (Mullis et al., eds. 1994); Current Protocols in Immunology (J.E. Coligan et al., eds. 1991); Short Protocols in Molecular Biology (Wiley and Sons 1999); Immunobiology (CA. Janeway and P. Travers, 1997); Antibodies (P. Finch 1997); Antibodies: a practical approach (D. Catty., ed., IRL Press 1988-1989); Monoclonal antibodies : a practical approach (P. Shepherd and C. Dean, eds., Oxford University Press 2000); Using antibodies: a laboratory manual (E. Harlow and D. Lane (Cold Spring Harbor Laboratory Press 1999); The Antibodies (M. Zanetti and J.D. Capra, eds., Harwood Academic Publishers 1995); and Cancer: Principles and Practice of Oncology (V.T. DeVita et al., eds., J.B. Lippincott Company 1993).

EXAMPLES Example 1: METHODS Cell lines and cell culture

Mouse melanoma cell lines B 16-FO (ATCC#CRL-6322) and B16-F10 (ATCC#CRL-6475) were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS) at 37°C and 5% CO₂. 67NR is a BALB/C mouse-derived breast cancer cell line, as described previously (Aslakson, CJ. and Miller, F.R., Cancer Res., 52, 1399-1405 (1992)). 67NR cells were grown in DMEM medium supplemented with 2 mM L-glutamine, 100 U/mL penicillin, 100 ug/mL streptomycin, 2.5 x 10^*5 M 2-mercaptoethanol, and 10% FBS. MeI-STR and MeI-STV cell lines, immortalized primary human melanocytes expressing RasV12 or empty vector, respectively, were described previously (Gupta, P.B. et al., Nat. Genet., 37, 1047- 1054 (2005)) and provided by R. A. Weinberg (Whitehead Institute, MIT). UACC-257, LOX IMVI, SK-MEL-2, SK-MEL-5 and MALME-3M melanoma cell lines were obtained from the ATCC and were grown as recommended by the supplier. To generate stable Gasl -expressing B16-F10 cell lines, Gasl was subcloned from the expression vector pcDNA3-Gasl (a generous gift of CM. Fan, Carnegie Institution of Washington, USA) into the retroviral vector pQCXI-puro (Clontech). shRNA screen

The mouse shRNA^mir library (release 2.16; Open Biosystems) was obtained through the University of Massachusetts Medical School shRNA library core facility. Ten retroviral pools, each comprising -6000 shRNA clones, were generated with titers of ~2xl O⁵ pfu ml^"1. These retroviral stocks were produced following co-transfection into the PhoenixGP packaging cell line (a gift from G. Nolan, Stanford University, USA) with plasmids expressing VSV-G (pCI-VSVG; a gift from G. Nolan) and gag-pol (Ory, D.S. et al., Proc. Natl. Acad. Sci USA, 93, 11400-11406 (1996)). B16-F0 cells (1.2xlO⁶) were transduced at a multiplicity of infection of 0.2 with the retroviral stocks in 100 mm plates, and two days later selected for resistance to puromycin (2 μg ml^"1) for seven days. Satellite colonies were isolated from the fibrin matrix and expanded for genomic DNA isolation. To identify the candidate shRNAs, the shRNA region of the transduced virus was PCR amplified (using primers PSM2-forward, 5'-

GCTCGCTTCGGC AGCACATATAC-S' (SEQ ID NO: 47) and PSM2-reverse, 5'- GAGACGTGCTACT TCCATTTGTC-3' (SEQ ID NO: 48)) and cloned into pGEM-T Easy (Promega). A minimum of 96 clones were sequenced per pool, using primer PSM2-seq 5'-GAGGGCCTATTTCCCATGAT-S' (SEQ ID NO: 49).

Individual knockdown cell lines were generated by retroviral transduction of 2x10⁵ B 16-FO cells (in 6-well plates) with the respective shRNA. shRNAs were either obtained from the Open Biosystems library or the RNAi Consortium (Table 4). Three-dimension cell culture assay Cells were embedded in collagen gel at a density of 5x10⁴ cells per 200 μl gel in

96- well plates, as previously described (Doillon, CJ. et al., Anticancer Res. 24, 2169- 2177 (2004)). After two to three hrs, the gels were removed from the well, soaked in growth factor-reduced Matrigel™ (BD Biosciences) for two min, sandwiched into a fibrin gel laid down in wells of a 24-well (or 10 cm) plate, and incubated for 14 days at 37°C with 5% CO₂ in culture medium. The media was renewed every other day, and the ability of cells to migrate from the gel into the fibrin was also assessed every other day. The antifibrinolytic agent aprotinin (Sigma) was added to the culture media at 100 U ml^"1. Satellite colonies were stained using a solution of 0.2% methylene blue in 50% methanol. Mouse metastasis assays

For the experimental (tail vein injection) assay, B 16-FO cells (2x10⁵) stably expressing a candidate shRNA were suspended in 200 μl PBS and injected in the lateral tail vein of three C57BL/6 mice (Taconic). Lungs were harvested 14 days post injection and fixed in formalin. Metastases were counted and statistical analysis (One-way ANOVA) was performed using GraphPad Prism version 3.00 for Windows (GraphPad Software). For the spontaneous metastasis assay, B16-F0 cells (IxIO⁶) stably transduced with a Gasl or NS shRNA were injected into the foot pad of syngeneic C57BL/6 mice. When the primary tumor reached a size of 100 mm³ it was excised, and the mice were examined for lung metastases four weeks later. All experiments were performed in accordance with the Institutional Animal Care and Use Committee (IACUC) guidelines. Tumor formation assays Tumor formation assays were performed essentially as described previously

(Theodoro, J.G., et al., Science 313, 968-971 (2006)). Briefly, 5xlO⁴ B16-F0 knockdown cell lines were suspended in 60 μl Matrigel™ and injected subcutaneously into the right flank C57BL/6 mice (3 mice per shRNA). Tumor dimensions were measured every two days from the time of appearance of the tumors, and tumor volume was calculated using the formula π/6 x (length) x (width)². Statistical analysis was performed as described above. Quantitative RT-PCR

Total RNA was isolated using TRIZOL (Invitrogen). Reverse transcription was performed in triplicate using Superscript II Reverse Transcriptase (Invitrogen) as per the manufacturer's instructions, followed by quantitative real-time PCR. To monitor target gene knockdown, total RNA was isolated seven days following retroviral transduction and puromycin selection. The sequences of the primers used for quantitative real-time PCR are provided in Table 5. Single-cell fluorescence imaging. Experiments were performed as previously described (Tsai, Y.C. et al. Nat Med

13, 1504-9 (2007)). Briefly, B16-F0 cells (IxIO⁶) stably transduced with a Gasl or NS shRNA were fluorescently labeled with CellTracker Green (Invitrogen) and injected into the tail vein of C57BL/6 mice. At various time points following injection, mice were euthanized and their lungs inflated by slow intra-tracheal injection of PBS followed by imaging by epifluorescence microscopy (Leica). Immunofluorescence.

Immunofluorescence was performed using the MElOOl malignant melanoma, metastatic malignant melanoma and nevus tissue array (Biomax), which contains samples from 56 cases of malignant melanoma, 20 cases of metastatic (lymph node or fatty tissue) malignant melanoma and 24 cases of nevus (normal tissue). The array was hybridized overnight at 4°C with a biotinylated anti-human GASl affinity purified polyclonal antibody (RD System, #BAF2636) followed by incubation with a Cy3- conjugated secondary antibody (Sigma, ExtrAvidin-Cy3 #E4142) for 1 hour at room temperature. As a control, samples were stained with the melanoma marker HMB45 (Kapur, R. P. et al., J Histochem Cytochem. 40, 207-212 (1992)). The microarray was incubated for 1 hour at room temperature with a mouse anti-human HMB45 monoclonal antibody (DAKO # M0634), followed by incubation with a FITC-conjugated secondary antibody (Invitrogen, Alexa Fluor® 488 goat anti-mouse # A21 121) for l hour at room temperature. Cell nuclei were visualized with 4',6-diamidino-2-phenylindole (DAPI). Imaging was done by epifluorescence microscopy (Leica). Only HMB45-positive samples were scored for GASl expression. Oncomine database searches

The Haqq (Haqq, C. et al., Proc. Natl. Acad. ScI USA 102, 6092-6097 (2006)) and Smith (Smith, A.P. et al., Cancer Biol. Ther. 4, 1018-1029 (2005)) melanoma microarray datasets were accessed using the Oncomine Cancer Profiling Database (oncomine.org). The Haqq melanoma dataset includes 10 normal mole, 5 primary melanoma and 17 metastatic melanoma samples; the Smith melanoma dataset includes 6 early stage (normal, benign nevi, melanoma in situ) and 7 advanced stage (vertical growth phase melanoma, metastatic growth phase melanoma, melanoma positive lymph nodes) samples. Histograms depicting gene expression in each sample, as well as a Student's /-test giving a P value for the comparison of candidate gene expression between the groups, were obtained directly through the Oncomine 3.0 software. Analogous approaches were used to analyze expression of candidate genes in other cancers.

Example 2: Genome-wide RNA; Screen for Modulators of Tumor Metastases. Metastatic dissemination of a primary tumor to a secondary site is the major cause of deaths from solid tumors (reviewed in Gupta, G.P et al., Cell 127, 679-695, 2006; and Nguyen, D.X. et al., Nat. Rev. Genet. 8, 341-352 (2007)). The progression to metastasis involves a series of discrete steps, commonly known as the metastatic cascade, which minimally includes: invasion of the tumor border, intravasation into vascular structures, survival during transport to the secondary site, extravasation, and colonization of the secondary site (reviewed in Gupta, G.P et al., Cell 127, 679-695, 2006; and Steeg, P.S., Nat. Med. 12, 895-904 (2006)). The complex process of metastasis is controlled by multiple genes that either increase or decrease metastatic potential (reviewed in Nguyen, D.X. et al., Nat. Rev. Genet. 8, 341-352 (2007), and Berger, J.C. et al., Cancer Biol. Ther. 4, 805-812 (2005)). Although many genes have been identified that promote metastasis, a relatively small number of metastasis suppressor genes have been documented Rinker-Schaefer, CW. et al., Clin. Cancer Res. 12, 3882-3889 (2006). This is due, at least in part, to a lack of experimental approaches for the systematic identification of genes that specifically inhibit metastasis.

Three-dimensional (3-D) cell culture systems comprising cancer cell lines grown in matrices of collagen and fibrin provide an ex vivo model system for studying tumor cell invasion and expansion into the extracellular matrix (Doillon, CJ. et al., Anticancer Res. 24, 2169-2177 (2004)). Using this bi-composite gel technology system, it has been shown that a variety of "aggressive" human tumor cell lines, including colorectal, prostate and non-small lung carcinoma cells, rapidly develop satellite colonies separate from the primary tumors. The formation of these satellite colonies provides a cell culture model that recapitulates several critical steps of the metastatic process.

Using a 3-D cell culture system as an initial selection, we performed a genome- wide small hairpin RNA (shRNA) screen to identify genes that, when knocked down, increase formation of satellite colonies (Figure Ia). A mouse shRNA library comprising -62,400 shRNAs directed against ~28,000 genes was divided into 10 pools, which were packaged into retrovirus particles and used to stably transduce poorly metastatic Bl 6-FO mouse melanoma cells (Fidler, I.J., Cancer Res. 35, 218-224 (1975)). The cells were embedded in collagen, coated with a basement membrane matrix (Matrigel™), sandwiched into a fibrin gel, and then incubated for two weeks. Figure Ib shows, as expected, that B 16-FO cells produced few (5-10) satellite colonies whereas B16-F10 cells, a highly metastatic counterpart of B 16-FO cells (Fidler, I. J., Cancer Res. 35, 218- 224 (1975)), produced numerous (1000-2000) satellite colonies. Transduction of B 16-FO cells with an shRNA pool, but not a non-silencing (NS) shRNA control, also resulted in the appearance of many satellite colonies; the results from a representative pool are shown (Figure Ic).

Table 1 : Metastasis suppressor genes that tested positive in the tail vein/lung metastasis assay. Biological Gene Name fcjπcSoπ symbol

CeI cycle reputation Ccnβ2 cyctn A2

Gasi growth βπest specific 1

CeI communication/ anas olfactoiyreoeptoT 198

9gπsl transduction Pdyn pfodjfiGrøhn

Cytøsfceletoπ Acta2 βcfn, alpha 2, smooti muscle, aorta

Develop me πt Ofφβ3 developmental pluripotency-assoάatad 3

Transport Tαmm70a kaπstocase of outer m-bchondiial membrane 70 homαtog A (yβasl)

MetaboisπV Afgδ asparagjne-inked gljcoεytatiαn 6 homotog(yeasQ

Energy Pathways Hsd3b2 hydrDxy-da-a-5-stβπαid dehydrogenase, 3 beta- and steroid deβa-isαmerasβ 2

Ptikβt phoaphorytase kinase alpha 1

Protein me tabofsm Cpβ2 cartXDcypepf dase A2, panσβatc

Ctso cathθpsin O

Unknown 4931422AOSHk RJKEN cONA 4931422A03 gene

B830t»aK06RΛ RIKEN cONA B630019KD6 gene

Ccdc39 coied-ODi domain containing 39 βϊ 436563 predicted gene, EG436563

Ptdm13 PRdoman contarang 13

Prrςβ prof πe rich G-carbαxygluiamic add prolan 3

SeU2 SET domain containing 2

Sc9a3r2 solute carrier famiy 9 (sodMnVhydrogen exchanger), iscrform 3 regulator 2

ThsdTb ffinombospondl n, type I domain containing 7B

Zfp294 zinc finger protein 294

Following the two-week incubation, satellite colonies from each pool were isolated, genomic DNA was extracted and the shRNAs were identified by sequence analysis. Positive candidates were confirmed in independent single- well 3-D cell culture assays by stably transducing B 16-FO cells with a single shRNA directed against each candidate gene (Figure 7). In total, the screen identified 80 genes (Table 3) that, following shRNA-mediated knockdown, promoted formation of B 16-FO satellite colonies in the fibrin matrix (Table 1).

Example 3: Jn vivo Assessment of Metastases Suppressor Genes

We next asked whether knockdown of any of the candidate genes could promote metastasis. B 16-FO cells (~2xlO⁵) stably transduced with an shRNA against the candidate gene were injected into the tail veins of syngeneic C57BL/6 mice, and 14 days later the mice were examined for lung metastases. Figure 2a shows, as expected, that injection of B 16-FO cells alone or B 16-FO cells stably expressing a NS shRNA resulted in very few lung metastases (typically <10 per lung),whereas injection of B 16-Fl 0 cells resulted in a substantially higher number (-200 per lung). Of the 80 genes identified in the 3-D cell culture assay, knockdown of 22 genes led to a significant increase in the number of lung metastases relative to the NS shRNA (Table 1 , Figure 2a, and quantitated in Figure 2b). Quantitative real-time RT-PCR (qRT-PCR) confirmed in all cases that expression of the target gene was decreased in each B 16-FO knockdown cell line (Figure 8). For all genes, a second, unrelated shRNA directed against the same target also resulted in increased formation of satellite colonies in the 3-D assay (Figure 9). The 22 genes are involved in a variety of processes including cell cycle regulation (Ccna2 and Gasl), cell communication and signal transduction (Olfrl98 and Pdyri), cytoskeletal function (Actal), development (Dppa3), intracellular transport (Tomm70ά), metabolism/energy pathways (Alg6, Hsd3b2 and Phkal), and protein metabolism (Cpa2 and Ctsό).

Example 4: In vivo Assessment of Metastasis Suppressor Genes in Primary Tumors

An essential criterion of a metastasis suppressor gene is that it does not affect growth of the primary tumor (Steeg, P.S., Nat. Rev. Cancer 3, 55-63 (2003)). We therefore tested whether knocking down any of the candidate genes affected the ability of B 16-FO cells to form primary tumors. B 16-FO cells (5x10⁴) stably expressing an shRNA directed against each candidate gene were subcutaneously injected into the flanks of C57BL/6 mice, and tumor volume was measured at day 16. Figure 2c shows that knockdown of any of the 22 genes did not significantly affect primary tumor growth.

Metastasis suppressor genes are often expressed at lower levels in highly metastatic tumor cells relative to poorly- or non-metastatic tumor cells or normal cells (reviewed in Steeg, P.S., Nat. Rev. Cancer 3, 55-63 (2003)). To test whether any of the 22 candidate genes conformed to this pattern, we first compared their expression in highly metastatic B 16-F 10 and poorly metastatic B 16-FO cells. The qRT-PCR results of Figure 3a show that the majority of genes were modestly down-regulated (1.5-3 fold) in Bl 6-F 10 cells compared to B 16-FO cells. Three genes (Dppa3, Olfrl98 and Pdyri) were up-regulated in B16-F10 cells, and two genes (Acta2 and Tdpozl) were unchanged. Most notably, one gene, Gasl (Growth arrest-specific 1), was markedly down-regulated (~11 fold) in B16-F10 cells. Gasl encodes a pleiotropic 45-kDa glycosylphosphatidylinositol (GPI)-anchored cell surface protein (Stebel, M. et al., FEBS Lett. 481, 152-158 (2000)) that has been shown to play a role in both negative (Del Sal, G., et al., Cell 70, 595-607 (1992); Lee, C.S., et al., Proc. Natl. Acad. ScI USA 98, 11347-11352 (2001); Mellstrom, B. et al. MoI Cell Neurosci. 19, 417-29 (2002); Zamorano, A., et al., J. Neurosci. Res. 71, 256-63 (2003); Zamorano, A., et al., Neurobiol. Dis. 15, 483-91 (2004)) and positive regulation of cell growth (Liu, Y., et al., Dev. Biol. 236, 30-45 (2001); Liu, Y., et al., Dev. Biol. 236, 30-45 (2001)), as well as in embryonic development through regulation of Sonic Hedgehog signaling (Lee, C.S., et al., Proc. Natl. Acad. Sci. USA 98, 11347-11352 (2001); Cobourne, M.T., et al., Development 131, 2875-2885 (2004); Allen, B.L., et al., Genes Dev. 21, 1244-1257 (2007); Martinelli, D.C. et al., Genes Dev. 21, 1231-1243 (2007); Seppala, M. et al., J. Clin. Invest. 117, 1575-1584 (2007)). It is likely that Gasl mediates different responses in a cell-type specific manner.

To determine whether down-regulation of Gasl contributed to the high metastatic capacity of B16-F10 cells, we tested whether ectopic expression of Gas 1 in B16-F10 cells could reduce their ability to metastasize. Significantly, ectopic expression of Gasl decreased the number of satellite colonies in the 3-D cell culture assay ( Figure 3b) and the number of lung metastases in the mouse tail vein metastasis assay (Figure 3 c) relative to the empty vector control. Ectopic Gasl expression had no effect on the ability of B16-F10 cells to form a tumor when injected subcutaneously in mice (Figure 3d), indicating that the reduced metastatic efficiency following ectopic expression of Gasl in B16-F10 cells was not the result of a general decrease in cell proliferation or tumorigenicity.

The above results suggested that Gasl is a metastasis suppressor. However, the experimental (tail vein injection) metastasis assay is somewhat limited in testing for metastatic potential, as it bypasses critical steps of the metastatic cascade such as invasion of the tumor border and intravasation into the vasculature. We therefore tested whether Gasl knockdown could promote metastasis in a more rigorous spontaneous metastasis assay. B 16-FO cells (1x10⁶) stably transduced with a Gasl or NS shRNA were injected into the foot pad of C57BL/6 mice. When the primary tumor reached a size of 100 mm³ it was excised, and four weeks later the mice were examined for lung metastases. Figure 4a shows that the percentage of animals developing lung metastases was significantly higher in mice injected with B 16-FO Gasl knockdown (B16-F0/Gasl KD) cells than those injected with B 16-FO cells expressing an NS (B 16-FO/NS) shRNA. We next tested whether knockdown of Gasl could promote metastasis in another cancer cell line. Mouse breast cancer 67NR cells (~2xlO⁵) were stably transduced with a Gasl or NS shRNA were injected into the tail veins of C57BL/6 mice, and six weeks later the mice were examined for lung metastases. Consistent with the results in Bl 6- FlO cells, knockdown of Gasl led to a significant increase in the number of lung metastases relative to the NS shRNA (Figure 4d).

We next sought to determine the mechanism by which Gasl suppresses metastasis. We reasoned that Gasl could decrease the ability of cells to colonize and/or survive in the lung. To test this possibility, we labeled Bl 6-FO/NS and B16-F0/Gasl KD cells with a fluorescent dye (Khanna, C. et al., Nat. Med. 10, 182-6 (2004)) and quantified their persistence early after their arrival in the lung following injection into the tail vein. Figure 4b shows that one hour after injection, the number of B16-F0/Gasl KD cells and Bl 6-FO/NS cells in the lung were roughly equivalent, whereas three hours after injection, the number of B16-F0/Gasl KD cells that remained in the lung was significantly higher than the number of B 16-FO/NS cells. Importantly, B 16-FO/Gas 1 KD cells showed decreased caspase activation compared to Bl 6-FO/NS cells, indicating Gasl KD cells exhibited a reduction in apoptosis. Fluorescence imaging established colocalization of B 16-FO cells expressing a Gasl shRNA and cells in which caspase is activated, as evidenced by the accumulation of a fluorescence pan-caspase probe (Figure 5a). Collectively, these results suggest that down-regulation of Gasl promotes metastasis by protecting cells from apoptosis.

Previous studies have shown that oncogenic Ras efficiently transforms immortalized primary human melanocytes, and that these RasV12-expressing cells (MEL-STR cells) form metastatic tumors in vivo (Gupta, P.B. et al., Nat. Genet. 37, 1047-1054 (2005)). We therefore asked whether expression of Ras would result in down-regulation of GASl. Figure 4c shows that GASl expression was reduced ~6-fold in MeI-STR cells relative to immortalized primary human melanocytes expressing empty vector (MeI-STV cells). GASl is also down-regulated in Ras-transformed human mammary epithelial cells (BiId, A.H. et al., Nature 439, 353-357 (2006)). These results suggest that promotion of metastasis by Ras occurs, at least in part, through down- regulation of GASl .

Example 5: In silico Assessment of Metastasis Suppressor Genes in Human Cancer

Acquisition of metastatic potential is thought to involve the inactivation or down- regulation of metastasis suppressor genes (Nguyen, D.X. et al., Nat. Rev. Genet. 8, 341- 352, 2007 and Berger, J.C. et al., Cancer Biol. Ther. 4, 805-812 (2005)). The above results raised the possibility that loss of Gasl might be required for development of metastatic melanoma. As a first test of this possibility, we monitored GASl expression in a panel of human melanoma cell lines whose origin (i.e., from either primary or metastatic melanoma) was known. We find that GASl expression is high in UACC-257 malignant melanoma cells, but is dramatically reduced in metastatic melanoma cell lines derived from the lymph node (LOX IMVI), axillary node (SK-MEL-5), lung (MALME- 3M) or skin (SM-MEL-2) (Figure 6b). We also performed immunofluorescence analysis of GASl expression on a panel of human primary and metastatic melanoma samples using a melanoma tissue microarray. Similar to the results in the melanoma cell lines, GASl expression is detectable in the majority of primary melanomas (39/47 samples), but in only 41% of metastatic melanomas (10/17 samples) (Figure 6c).

Finally we asked whether the human homologs of any of the murine putative metastasis suppressor genes we identified were down-regulated in human metastatic tumors. We initially analyzed expression profiles in human metastatic melanoma versus primary melanoma or benign nevi. A search of the publicly-accessible Oncomine cancer profiling database¹ ' revealed in addition to GAS 1 , as expected, four three genes (ACTA2, CTSOnd SLC9A3R2) were significantly down-regulated (p<0.05) in metastatic melanoma compared to primary melanoma and/or benign nevi (Haqq, C. et al., Proc. Natl. Acad. Sci. USA 102, 6092-6097 (2005)) (Figure 4a). In addition, DPP A3 was expressed at significantly lower levels in advanced relative to early stage melanoma (Smith, A.P. et al., Cancer Biol. Ther. 4, 1018-1029 (2005)) (Figure 4b). A search of additional cancers revealed that 15 of the genes we identified, including GASl, were down-regulated in either metastatic versus primary tumor samples, or in late (stages III and IV) relative to early (stage I) disease in multiple cancer types (Table 2).

Table 21 list of genes that are down-regulated in metastatic versus non-metastatic samples, or in late (stages 10 and IV) relative to early (stage I) disease of multiple cancer types

A

ALQβ

CCDC39

CCNA2

CTSO

CPA2

DPPA3

GAS1

PDYN

PHKA1

SETD2

SLC9A3R2

THSD7B

TOMMWA

ZNF294

• indicates down-regutafon in metastatic versus norwnβtastatic samples, or h late relative to earty stages erf various human cancers. ZNF294 is the human hαmolog of mouse Zfp294.

Examples 6: Metastasis Suppressor Genes in Cancer

Metastasis accounts for the majority of cancer deaths arising from solid tumors. It is therefore imperative to understand the basis by which a primary tumor develops the ability to metastasize. In addition, genes that regulate the metastatic process can be used to diagnose and predict disease, and may also provide new therapeutic targets. In this report, we have described an experimental strategy for the systematic identification of metastasis suppressor genes. Using this approach, we have found 80 genes that inhibited formation of satellite colonies in a 3-D tissue-culture assay and have shown that 22 of these genes suppressed metastasis to the lung following mouse tail vein injection. The other 60 genes we identified could function as metastasis suppressors for colonization of organs other than the lung or in alternative steps of the metastatic cascade that were not tested here. Fifteen of the 22 genes we identified are down-regulated in metastases from melanoma or other cancers. Thus, the genome-wide shRNA screen we have developed reveals genes that, on the basis of both experimental and clinical evidence, are new metastasis suppressors.

Among the most interesting of the 22 new metastasis suppressor genes is GASl, which is down-regulated in metastases from melanoma (Figure 6a) as well as breast and prostate cancers (Table 2). Consistent with these findings, we have shown that down- regulation of Gas 1 contributes to the high metastatic potential of B16-F10 mouse melanoma cells. Moreover, we have found that oncogenic Ras down-regulates GASl, explaining at least part of the basis by which Ras promotes metastasis. Interestingly, the chromosomal region harboring GASl is frequently deleted in myeloid malignancies

(Evdokiou, A. et al., Genomics 18, 731-733 (1993)) and bladder cancer (Simoneau, A.R. et al., Cancer Res. 56, 5039-5043 (1996)). Moreover, down-regulation of GASl is associated with the progression of prostate cancer (Bettuzzi, S. et al. Cancer Res. 60, 28- 34 (2000) and Scaltriti, M. et al. Int. J. Cancer 108, 23-30 (2004)). Collectively, these results suggest GASl could be a useful marker of metastatic potential.

Example 7: Genome-wide shRNA Screen Identifies GASl as a Novel Melanoma Metastasis Suppressor Gene

Metastasis suppressor genes inhibit one or more steps required for metastasis without affecting primary tumor formation. Due to the complexity of the metastatic process, the development of experimental approaches for identifying genes involved in metastasis prevention has been challenging. Here we describe a genome-wide RNA interference screening strategy to identify candidate metastasis suppressor genes. Following expression in weakly metastatic B 16-FO mouse melanoma cells, shRNAs were selected based upon enhanced satellite colony formation in a three-dimensional cell culture system and confirmed in a mouse experimental metastasis assay. Using this approach we discovered 22 genes whose knockdown increased metastasis without affecting primary tumor growth. We focused on one of these genes, Gasl, because we found that it was substantially down-regulated in highly metastatic B16-F10 melanoma cells, which contributed to the high metastatic potential of this mouse cell line. We further demonstrated that Gasl has all the expected properties of a melanoma tumor suppressor including: suppression of metastasis in a spontaneous metastasis assay, promotion of apoptosis following dissemination of cells to secondary sites, and frequent down-regulation in human melanoma metastasis-derived cell lines and metastatic tumor samples. Thus, we have developed a genome-wide shRNA screening strategy that enables the discovery of new metastasis suppressor genes. Metastatic dissemination of a primary tumor to a secondary site is the major cause of deaths from solid tumors (reviewed in Gupta and Massague 2006; Nguyen and Massague 2007). The progression to metastasis involves a series of discrete steps, commonly known as the metastatic cascade, which minimally includes: invasion of the tumor border, intravasation into vascular structures, survival during transport to the secondary site, extravasation, and colonization of the secondary site (reviewed in Gupta and Massague 2006; Steeg 2006). The complex process of metastasis is controlled by multiple genes that either increase or decrease metastatic potential (Berger et al. 2005; Nguyen and Massague 2007). Although many genes have been identified that promote metastasis, a relatively small number of metastasis suppressor genes have been documented (Rinker-Schaeffer et al. 2006). This is due, at least in part, to a lack of experimental approaches for the systematic identification of genes that specifically inhibit metastasis.

Three-dimensional (3-D) cell culture systems comprising cancer cell lines grown in matrices of collagen and fibrin provide an ex vivo model system for studying tumor cell invasion and expansion into the extracellular matrix (Doillon et al. 2004). Using this bi-composite gel technology system, it has been shown that a variety of "aggressive" human tumor cell lines, including colorectal, prostate and non-small lung carcinoma cells, rapidly develop satellite colonies separate from the primary tumors. The formation of these satellite colonies provides a cell culture model that recapitulates several critical steps of the metastatic process, including tumor cell motility and invasion, expansion into the collagen matrix, and the ability to survive and form colonies at secondary sites. Here we show how a 3-D cell culture system can be used in conjunction with genome-wide RNA interference screening to identify candidate metastasis suppressor genes. These candidates can be further analyzed and validated, thereby enabling the discovery of new metastasis suppressor genes in specific cancer types.

A genome-wide shRNA screen for the identification of candidate metastasis suppressor genes Using a 3-D cell culture system as an initial selection, we performed a genome- wide small hairpin RNA (shRNA) screen to identify genes that, when knocked down, increase formation of satellite colonies (Fig. IA). A mouse shRNA library comprising -62,400 shRNAs directed against -28,000 genes (Silva et al. 2005) was divided into 10 pools, which were packaged into retrovirus particles and used to stably transduce poorly metastatic B 16-FO mouse melanoma cells (Fidler 1975). The cells were embedded in collagen, coated with a basement membrane matrix (MatrigelTM), sandwiched between two layers of fibrin gel to completely encapsulate the collagen/MatrigelTM plug, and then incubated for two weeks. Figure IB shows, as expected, that B 16-FO cells produced few (5-10) satellite colonies whereas B 16-F 10 cells, a highly metastatic counterpart of B 16-FO cells (Fidler 1975), produced numerous (500-1000) satellite colonies. Transduction of B 16-FO cells with an shRNA pool, but not a non-silencing (NS) shRNA control, also resulted in the appearance of many satellite colonies; the results from a representative pool are shown (Fig. 1C). Following the two-week incubation, satellite colonies from each pool were isolated, genomic DNA was extracted and the shRNAs were identified by sequence analysis. Positive candidates were confirmed in independent single-well 3-D cell culture assays by stably transducing B 16-FO cells with a single shRNA directed against each candidate gene (See, e.g., Fig. 7; quantified in Table 6). In total, the screen identified 80 genes that, following shRNA-mediated knockdown, promoted formation of B 16-FO satellite colonies in the fibrin matrix (Table 3).

Identification of 22 genes whose knockdown increase metastasis without affecting primary tumor growth

We next asked whether knockdown of any of the candidate genes could promote metastasis. B 16-FO cells stably transduced with an shRNA against each of the 80 candidate genes were injected into the tail veins of syngeneic C57BL/6 mice, and 14 days later the mice were examined for lung metastases. Figure 2A shows, as expected, that injection of B16-F0 cells alone or B16-F0 cells stably expressing an NS shRNA resulted in very few lung metastases (typically <10 per lung), whereas injection of Bl 6- FlO cells resulted in a substantially higher number (-200 per lung). Of the 80 genes identified in the 3-D cell culture assay, knockdown of 22 genes led to a significant increase in the number of lung metastases relative to the NS shRNA (Table 1, Fig. 2 A and quantitated in Fig. 2B). Quantitative real-time RT-PCR (qRT-PCR) confirmed in all cases that expression of the target gene was decreased in each B 16-FO knockdown cell line (Fig. 8A). For three of the genes for which suitable antibodies were available (Gasl, Acta2 and Ccna2) target gene knockdown was also confirmed by immunoblot analysis (Fig. 8B). For all genes, a second, unrelated shRNA directed against the same target also resulted in gene knockdown and increased formation of satellite colonies in the 3-D assay (Fig. 9). The 22 genes are involved in a variety of processes including cell cycle regulation (Ccna2 and Gasl), cell communication and signal transduction (Olfrl98 and Pdyn), cytoskeletal function (Acta2), development (Dppa3), intracellular transport (Tomm70a), metabolism/energy pathways (Alg6, Hsd3b2 and Phkal), and protein metabolism (Cpa2 and Ctso).

An essential criterion of a metastasis suppressor gene is that it does not affect growth of the primary tumor (Steeg 2003). We therefore tested whether knocking down any of the candidate genes affected the ability of B 16-FO cells to form primary tumors. B 16-FO cells stably expressing an shRNA directed against each candidate gene were subcutaneously injected into the flanks of C57BL/6 mice, and tumor volume was measured at day 16. Figure 2C shows that knockdown of any of the 22 genes did not significantly affect primary tumor growth.

Down-regulation of Gasl in B16-F10 cells contributes to their high metastatic potential

Metastasis suppressor genes are often expressed at lower levels in highly metastatic tumor cells relative to poorly- or non-metastatic tumor cells or normal cells (reviewed in Steeg 2003). To test whether any of the 22 candidate genes conformed to this pattern, we first compared their expression in highly metastatic B16-F10 and poorly metastatic B 16-FO cells. The qRT-PCR results of Figure 3 A show that all but one of the genes were either expressed equally or at most modestly down-regulated in B16-F10 cells compared to B 16-FO cells. By contrast, one gene, Gasl (Growth arrest-specific 1), was markedly down-regulated (~11 fold) in B16-F10 cells. Gasl encodes a pleiotropic 45-kDa glycosylphosphatidylinositol-anchored membrane protein (Stebel et al. 2000) that has been shown to play a role in both negative (Del Sal et al. 1992; Lee et al. 2001; Mellstrom et al. 2002; Zamorano et al. 2003; Zamorano et al. 2004) and positive regulation of cell growth (Liu et al. 2001; Spagnuolo et al. 2004), as well as in embryonic development through regulation of Sonic Hedgehog signaling (reviewed in Martinelli and Fan 2007b).

To determine whether down-regulation of Gas 1 contributed to the high metastatic capacity of B16-F10 cells, we tested whether ectopic expression of Gas 1 in B16-F10 cells could reduce their ability to metastasize. Significantly, ectopic expression of Gasl decreased the number of satellite colonies in the 3-D cell culture assay (Fig. 3B) and the number of lung metastases in the mouse tail vein metastasis assay (Fig. 3C) relative to the empty vector control. Ectopic Gasl expression had no effect on the ability of Bl 6- FlO cells to form a tumor when injected subcutaneously in mice (Fig. 3D), indicating that the reduced metastatic efficiency following ectopic expression of Gasl in B16-F10 cells was not the result of a general decrease in cell proliferation or tumorigenicity. Consistent with this finding, cultured B 16-FO cells expressing an NS shRNA (B 16- F0/NS) and B 16-FO Gasl knockdown (B16-F0/Gasl KD) cells have comparable proliferation rates and levels of apoptosis (Fig. 10). Gasl knockdown promotes metastasis in a spontaneous metastasis assay

The tail vein injection assay bypasses several critical steps of the metastatic cascade such as invasion of the tumor border and intravasation into the vasculature. We therefore tested whether Gasl knockdown could promote metastasis in a more rigorous "spontaneous metastasis assay" in which B 16-FO cells, stably transduced with a Gasl or NS shRNA, were injected subcutaneously into the foot pad of C57BL/6 mice and approximately six weeks later the animals were examined for lung metastases. Figure 4 shows that the percentage of animals developing lung metastases was substantially higher in mice injected with B16-F0/Gasl KD cells than those injected with B16-F0/NS cells. Gasl suppresses metastasis by promoting apoptosis of cells following dissemination to secondary sites

One way in which a metastasis suppressor can function is by preventing tumor cells from colonizing or surviving at secondary sites. To test whether Gasl functioned in such a manner, we labeled B16-F0/NS and B16-F0/Gasl KD cells with a fluorescent dye and quantified their persistence after arrival in the lung following tail vein injection.

Figure 5 A shows that one hour after injection, the number of B16-F0/Gasl KD cells and B16-F0/NS cells in the lung was roughly equivalent, whereas at later times, three (Fig. 5A) or 24 (Fig. 1 Ia) hours after injection, the number of B16-F0/Gasl KD cells that remained in the lung was significantly higher than the number of B16-F0/NS cells. Notably, apoptosis was reduced in B16-F0/Gasl KD cells compared to B16-F0/NS cells, as evidenced by decreased levels of activated caspase (Fig. 5B). Collectively, these results suggest that Gasl suppresses metastasis by promoting apoptosis of cells following dissemination to secondary sites.

In support of this conclusion, ectopic expression of the anti-apoptotic protein BcI- 2 in B 16-FO cells resulted in an increased number of satellite colonies in the 3-D assay (Fig. 1 Ib), similar to the effect observed upon Gasl knockdown. Notably, previous studies have shown that Bcl-2 expression in B16 melanoma cells blocks apoptosis and increases the number of pulmonary metastases following mouse tail vein injection (Takaoka et al. 1997).

GASl is down-regulated in human metastatic melanoma cell lines and tissue samples Acquisition of metastatic potential is thought to involve the inactivation or down- regulation of metastasis suppressor genes (Berger et al. 2005; Nguyen and Massague 2007). We therefore asked whether progression of human primary to metastatic melanoma might involve loss of GASl. In support of this idea, a search of the publicly- accessible Oncomine cancer profiling database (Rhodes et al. 2007) revealed that GASl was significantly down-regulated (p<0.05) in metastatic melanoma compared to primary melanoma and benign nevi (Haqq et al. 2005)(Fig. 6A).

To confirm and extend this observation, we monitored GASl expression in a panel of human melanoma cell lines that originated from either a primary tumor or a metastasis. The immunoblot of Figure 6B shows that GASl expression was high in UACC-257 cells, derived from a primary melanoma, and markedly reduced in melanoma cell lines derived from metastases isolated from lymph nodes (LOX IMVI and SK-MEL- 5), lung (MALME-3M) or skin (SK-MEL-2). Significantly, knockdown of GASl in UACC-257 cells increased their ability to form satellite colonies in the 3-D cell culture assay (Fig. 1 Ic), analogous to the results with B 16-FO cells. Finally, we also monitored expression of GAS 1 in a pair of well-characterized, experimentally derived human melanoma cell lines: MeI-STV and MeI-STR. MeI-STV cells are immortalized primary human melanocytes that can be transformed by expression of an oncogenic RasV12 allele and the resulting RasV12-transformed cells (MeI-STR cells) can form metastatic tumors in vivo (Gupta et al. 2005). We found that GASl expression was reduced ~6- fold in MeI-STR cells relative to MeI-STV cells (Fig. 1 Id).

Next, we analyzed GASl expression in human primary and metastatic melanoma samples using a melanoma tissue microarray. As a control, samples were stained with the melanoma marker HMB45 (Kapur et al. 1992). Figure 6C shows that GASl expression was detectable in the majority (~83%) of primary melanomas but in only ~41% of metastatic melanomas. Collectively, these results indicate that GASl expression is frequently down-regulated during progression to metastatic melanoma. In this example, we have described an experimental strategy for the systematic identification of candidate metastasis suppressor genes. Using this approach, we found 80 genes that inhibited formation of satellite colonies in a 3-D cell culture assay and have shown that 22 of these genes suppressed metastasis to the lung following mouse tail vein injection. The other 58 genes we identified could function as metastasis suppressors for colonization of organs other than the lung or in alternative steps of the metastatic cascade that were not tested here.

Of these 22 new metastasis suppressor gene candidates we focused on Gasl because we found that it was down-regulated in highly metastatic B16-F10 melanoma cells and that this down-regulation contributed to the high metastatic potential of this mouse cell line. We further demonstrated that Gasl has the expected properties of a melanoma tumor suppressor including suppression of metastasis in both experimental (Fig. 2) and spontaneous assays (Fig. 4), promotion of apoptosis following dissemination of cells to secondary sites (Fig. 5B), and frequent down-regulation in human metastasis- derived cell lines and metastatic tumor samples (Fig. 6A-C). The basis by which GASl is down-regulated in metastatic melanoma remains to be determined. Preliminary analysis of the GASl promoter reveals a CpG island (S. G. and M.R.G., unpublished observation), raising the possibility that down-regulation of GASl expression is due to epigenetic silencing.

Our results indicate that Gasl suppresses metastasis, at least in part, by promoting apoptosis following dissemination of cells to secondary sites. To-date, we have found no evidence for a role of Gasl in other steps in the metastatic cascade, such as invasion (Fig. 1 Ie). Consistent with our results, previous studies have reported that Gasl can have a pro-apoptotic activity in specific cell types (Mellstrom et al. 2002; Zamorano et al. 2004). In addition, we note that other metastasis suppressors such as Caspase 8, BRMSl and TIMPs have, like Gasl, been shown to prevent apoptosis following dissemination to secondary sites (reviewed in Stafford et al. 2008). The detailed mechanism by which Gasl promotes apoptosis at the secondary site remains to be determined and there are several reasonable possibilities. For example, loss of Gasl could facilitate immune evasion, increase survival from biophysical forces encountered during metastatic spread, or regulate the response to cell death or survival signals emanating from the microenvironment at the secondary site. The signaling pathways by which Gasl induces apoptosis remain unknown.

Previous studies have shown that during development Gasl binds the secreted morphogen Sonic Hedgehog (Shh) with high affinity (Lee et al. 2001) and regulates Shh signaling (Lee et al. 2001; Cobourne et al. 2004; Allen et al. 2007; Martinelli and Fan 2007a; Seppala et al. 2007). Aberrant activation of Shh signaling has been implicated in the development of several cancers (reviewed in Wetmore 2003) including melanoma (Stecca et al. 2007). However, we found that Shh signaling was not affected by either Gasl knockdown in B 16-FO cells or ectopic expression of Gasl in B16-F10 cells as evidenced by levels of the well-characterized Shh marker GUI (Fig. Hf). Interestingly, Gasl has been reported to bind the receptor tyrosine kinase Ret and to inhibit Ret- mediated survival signaling pathways in neuroblastoma cells (Cabrera et al. 2006;

Lopez-Ramirez et al. 2008). Whether Gasl induces melanoma cell apoptosis through a similar mechanism remains to be determined.

A search of the publicly-accessible Oncomine cancer profiling database revealed that in addition to GASl, three genes (ACTA2, CTSO and SLC9A3R2) were also significantly down-regulated (p<0.05) in metastatic melanoma compared to primary melanoma and/or benign nevi (Haqq et al. 2005) (Fig. 12A). Moreover, DPP A3 is expressed at significantly lower levels in advanced relative to early stage melanoma (Smith et al. 2005) (Fig. 12B). A search of other cancer types revealed that 15 of the genes we identified, including GASl, were down-regulated in either metastatic versus primary tumor samples, or in late (stages III and IV) relative to early (stage I) disease in multiple cancer types (Table 2). Thus, it seems highly likely that some of these other 21 genes are also metastasis suppressors in melanomas or other cancers. Our search of other cancer types showed that, in addition to melanoma, GASl is also down-regulated in breast and prostate cancer metastases (Table 2). Consistent with this observation, we found that knockdown of Gas 1 increased the ability of weakly metastatic mouse breast cancer 67NR cells (Aslakson and Miller 1992) to colonize the lung following mouse tail vein injection (Fig. 13), analogous to the results with B 16-FO melanoma cells. In addition, down-regulation of GASl is associated with the progression of prostate cancer (Bettuzzi et al. 2000; Scaltriti et al. 2004). These considerations strongly suggest that GASl is also a metastasis suppressor gene in cancers other than melanoma.

Materials and methods Cell lines and cell culture

Mouse melanoma cell lines B 16-FO (ATCC#CRL-6322) and B16-F10 (ATCC#CRL-6475) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum at 37°C and 5% CO2. To generate stable Gasl- expressing B16-F10 cell lines, Gasl was subcloned from the expression vector pcDNA3- Gasl (a gift from CM. Fan, Carnegie Institution of Washington, USA) into the retroviral vector pQCXI-puro (Clontech). Melanoma cell lines UACC-257, LOX IMVI, SK-MEL- 2, SK-MEL-5 and MALME-3M were obtained from ATCC and grown as recommended by the supplier.

Three-dimension cell culture assay

Cells were embedded in collagen gel at a density of 5x10⁴ cells per 200 μl gel in 96-well plates, as previously described (Doillon et al. 2004). After two to three hours, the gels were removed from the well, soaked in growth factor-reduced MatrigelTM (BD Biosciences) for two min, sandwiched into a fibrin gel laid down in wells of a 24-well (or 10 cm) plate, and incubated for 14 days at 37⁰C with 5% CO2 in culture medium. The media was renewed every other day, and the ability of cells to migrate into the fibrin was assessed every other day. The antifibrinolytic agent aprotinin (Sigma) was added to the culture media at 100 U/ml. Satellite colonies were stained using a solution of 0.2% methylene blue in 50% methanol.

To perform the shRNA screen, retroviral pools were prepared as previously described (Gazin et al. 2007). Satellite colonies were isolated from the fibrin matrix and expanded for genomic DNA isolation. A minimum of 96 clones were sequenced per pool, as previously described (Gazin et al. 2007). Using this approach we initially identified 113 candidate genes, which were retested by generating individual knockdown cell lines by retroviral transduction of 2x10⁵ B 16-FO cells (in 6- well plates) with the respective shRNA. shRNAs were either obtained from the Open Biosystems library or the RNAi Consortium (e.g., Table 4). A total of 78 candidates were reconfirmed in the independent, single-well 3-D cell culture assays (Figure 7). Metastasis assays

For the experimental assay, 2x10⁵ cells were suspended in 200 μl PBS and injected in the lateral tail vein of three C57BL/6 mice (Taconic). Lungs were harvested 14 days post injection and fixed in formalin. Metastases were counted and statistical analysis (One-way ANOVA) was performed using GraphPad Prism version 3.00 for Windows (GraphPad Software). Experiments were performed in accordance with the Institutional Animal Care and Use Committee (IACUC) guidelines. For the spontaneous metastasis assay, 1x106 cells were injected into the footpad of syngeneic C57BL/6 mice. When the primary tumor reached a size of 100 mm3 it was excised, and the mice were examined for lung metastases four weeks later. Experiments were performed in accordance with the guidelines of the Canadian Council for Animal Care and approved by Laval University institutional Animal Care Committee. Tumor formation assays

Tumor formation assays were performed essentially as described previously (Gazin et al. 2007). Briefly, 5x10⁴ cells were suspended in 60 μl MatrigelTM and injected subcutaneously into the right flank of C57BL/6 mice (three mice per shRNA). Tumor volume was calculated as described (Gazin et al. 2007). Statistical analysis was performed as described above. Quantitative RT-PCR

To monitor target gene knockdown, total RNA was isolated seven days following retroviral transduction and puromycin selection, and qRT-PCR was performed as previously described (Gazin et al. 2007). Primer sequences used for qRT-PCR are provided in Table 5.

Single-cell fluorescence imaging Experiments were performed essentially as previously described (Tsai et al. 2007). B16-F0/NS or B16-F0/Gasl KD cells (1x10⁶) were fluorescently labeled with CellTracker Green (Invitrogen) and injected into the tail vein of C57BL/6 mice. At various time points after injection, mice were euthanized and their lungs fixed in formalin followed by imaging using a Zeiss Axiophot 2 fluorescence microscope. Using Image J software, the number of events larger than 10 pixels were counted in five random fields per lung (n=3 mice), and statistical analysis (Student's t-test) was performed using GraphPad Prism version 3.00 for Windows (GraphPad Software). To monitor caspase activation, at 3 hours after injection of B16-F0/NS or B16-F0/Gasl KD cells mice were injected with a sulforhodamine-conjugated fluoromethylketone derivative of VaI- Ala- Asp (Immunochemistry) and sacrificed 30 min later. The lungs were excised, washed in PBS and snap frozen in OCT embedding medium. Cryosections (10 μm) were analyzed by fluorescence microscopy. Only cells that retained CellTracker Green were counted (~100 per lung; n=3 mice). Cells showing both green and red fluorescence were scored as apoptotic. Tissue microarray analysis

Immunofluorescence was performed using the MElOOl malignant melanoma, metastatic malignant melanoma and nevus tissue array (Biomax), which contains samples from 56 cases of malignant melanoma, 20 cases of metastatic (lymph node or fatty tissue) malignant melanoma and 24 cases of nevus (normal tissue). The array was hybridized overnight at 4°C with a biotinylated anti-human GASl affinity purified polyclonal antibody (RD System, #BAF2636) followed by incubation with a Cy3- coηjugated secondary antibody (Sigma, ExtrAvidin-Cy3 #E4142) for 1 hour at room temperature. As a control, the microarray was incubated for 1 hour at room temperature with a mouse anti-human HMB45 monoclonal antibody (Dako #M0634), followed by incubation with a FITC-conjugated secondary antibody (Invitrogen, Alexa Fluor® 488 goat anti-mouse #A21121) for 1 hour at room temperature. Cell nuclei were visualized with 4',6-diamidino-2-phenylindole (DAPI). Imaging was done by fluorescence microscopy as described above. Only HMB45-positive samples were scored for GASl expression.

Oncomine database searches The Haqq (Haqq et al. 2005) melanoma microarray dataset was accessed using the Oncomine Cancer Profiling Database (www.oncomine.org), and includes 10 normal mole, 5 primary melanoma and 17 metastatic melanoma samples. Histograms depicting GASl gene expression in each sample, as well as a Student's t-test giving a p value for the comparison of candidate gene expression between the groups, were obtained directly through the Oncomine 3.0 software.

References for Example 7

Allen, B.L., Tenzen, T., and McMahon, A.P. 2007. The Hedgehog-binding proteins Gasl and Cdo cooperate to positively regulate Shh signaling during mouse development. Genes Dev. 21 : 1244-1257.

Aslakson, CJ. and Miller, F.R. 1992. Selective events in the metastatic process defined by analysis of the sequential dissemination of subpopulations of a mouse mammary tumor. Cancer Res. 52: 1399-1405.

Berger, J.C., Vander Griend, D.J., Robinson, V.L., Hickson, J.A., and Rinker- Schaeffer, CW. 2005. Metastasis suppressor genes: from gene identification to protein function and regulation. Cancer Biol. Ther. 4: 805-812.

Bettuzzi, S., Davalli, P., Astancolle, S., Carani, C, Madeo, B., Tampieri, A., and Corti, A. 2000. Tumor progression is accompanied by significant changes in the levels of expression of polyamine metabolism regulatory genes and clusterin (sulfated glycoprotein 2) in human prostate cancer specimens. Cancer Res. 60: 28-34.

Cabrera, J.R., Sanchez-Pulido, L., Rojas, A.M., Valencia, A., Manes, S., Naranjo, J.R., and Mellstrom, B. 2006. Gasl is related to the glial cell-derived neurotrophic factor family receptors alpha and regulates Ret signaling. J Biol Chem 281 : 14330-9.

Cobourne, M.T., Miletich, I., and Sharpe, P.T. 2004. Restriction of sonic hedgehog signalling during early tooth development. Development 131 : 2875-2885.

Del Sal, G., Ruaro, M.E., Philipson, L., and Schneider, C. 1992. The growth arrest-specific gene, gasl, is involved in growth suppression. Cell 70: 595-607.

Doillon, C.J., Gagnon, E., Paradis, R., and Koutsilieris, M. 2004. Three- dimensional culture system as a model for studying cancer cell invasion capacity and anticancer drug sensitivity. Anticancer Res. 24: 2169-2177.

Fidler, I.J. 1975. Biological behavior of malignant melanoma cells correlated to their survival in vivo. Cancer Res. 35: 218-224. Gazin, C, Wajapeyee, N., Gobeil, S., Virbasius, CM., and Green, M.R. 2007. An elaborate pathway required for Ras-mediated epigenetic silencing. Nature 449: 1073- 1077.

Gupta, G.P. and Massague, J. 2006. Cancer metastasis: building a framework. Cell 127: 679-695.

Gupta, P.B., Kuperwasser, C, Brunet, J.P., Ramaswamy, S., Kuo, W.L., Gray, J.W., Naber, S.P., and Weinberg, R.A. 2005. The melanocyte differentiation program predisposes to metastasis after neoplastic transformation. Nat. Genet. 37: 1047-1054.

Haqq, C, Nosrati, M., Sudilovsky, D., Crothers, J., Khodabakhsh, D., Pulliam, B.L., Federman, S., Miller, J.R., 3rd, Allen, R.E., Singer, M.I., Leong, S.P., Ljung, B.M., Sagebiel, R.W., and Kashani-Sabet, M. 2005. The gene expression signatures of melanoma progression. Proc. Natl. Acad. Sci. USA 102: 6092-6097.

Kapur, R.P., Bigler, S.A., Skelly, M., and Gown, A.M. 1992. Anti-melanoma monoclonal antibody HMB45 identifies an oncofetal glycoconjugate associated with immature melanosomes. J. Histochem. Cytochem. 40: 207-212.

Lee, C.S., Buttitta, L., and Fan, CM. 2001. Evidence that the WNT-inducible growth arrest-specific gene 1 encodes an antagonist of sonic hedgehog signaling in the somite. Proc. Natl. Acad. Sci. USA 98: 11347-11352.

Liu, Y., May, N.R., and Fan, CM. 2001. Growth arrest specific gene 1 is a positive growth regulator for the cerebellum. Dev. Biol. 236: 30-45.

Lopez-Ramirez, M.A., Dominguez-Monzon, G., Vergara, P., and Segovia, J. 2008. Gasl reduces Ret tyrosine 1062 phosphorylation and alters GDNF-mediated intracellular signaling. Int J Dev Neurosci 26: 497-503.

Martinelli, D. C and Fan, CM. 2007a. Gasl extends the range of Hedgehog action by facilitating its signaling. Genes Dev. 21: 1231-1243.

Martinelli, D.C and Fan, CM.. 2007b. The role of Gasl in embryonic development and its implications for human disease. Cell Cycle 6: 2650-2655.

Mellstrom, B., Cena, V., Lamas, M., Perales, C, Gonzalez, C, and Naraηjo, J.R. 2002. Gasl is induced during and participates in excitotoxic neuronal death. MoI. Cell Neurosci. 19: 417-429.

Nguyen, D.X. and Massague, J. 2007. Genetic determinants of cancer metastasis. Nat. Rev. Genet. 8: 341-352. Rhodes, D.R., Kalyana-Sundaram, S., Mahavisno, V., Varambally, R., Yu, J., Briggs, B.B., Barrette, T.R., Anstet, M.J., Kincead-Beal, C, Kulkarni, P., Varambally, S., Ghosh, D., and Chinnaiyan, A.M. 2007. Oncomine 3.0: genes, pathways, and networks in a collection of 18,000 cancer gene expression profiles. Neoplasia 9: 166- 180.

Rinker-Schaeffer, C.W., O'Keefe, J.P., Welch, D.R., and Theodorescu, D. 2006. Metastasis suppressor proteins: discovery, molecular mechanisms, and clinical application. Clin. Cancer Res. 12: 3882-3889.

Scaltriti, M., Brausi, M., Amorosi, A., Caporali, A., D'Arca, D., Astancolle, S., Corti, A., and Bettuzzi, S. 2004. Clusterin (SGP-2, ApoJ) expression is downregulated in low- and high-grade human prostate cancer. Int. J. Cancer 108: 23-30.

Seppala, M., Depew, M.J., Martinelli, D.C., Fan, CM., Sharpe, P.T., and Cobourne, M.T. 2007. Gasl is a modifier for holoprosencephaly and genetically interacts with sonic hedgehog. J. Clin. Invest. 117: 1575-1584. Silva, J.M., Li, M.Z., Chang, K., Ge, W., Golding, M.C., Rickles, R. J., Siolas, D.,

Hu, G., Paddison, P.J., Schlabach, M.R., Sheth, N., Bradshaw, J., Burchard, J., Kulkarni, A., Cavet, G., Sachidanandam, R., McCombie, W.R., Cleary, M.A., Elledge, S.J., and Hannon, G.J. 2005. Second-generation shRNA libraries covering the mouse and human genomes. Nat. Genet. 37: 1281-1288. Smith, A.P., Hoek, K., and Becker, D. 2005. Whole-genome expression profiling of the melanoma progression pathway reveals marked molecular differences between nevi/melanoma in situ and advanced-stage melanomas. Cancer Biol. Ther. 4: 1018-1029.

Spagnuolo, R., Corada, M., Orsenigo, F., Zanetta, L., Deuschle, U., Sandy, P., Schneider, C, Drake, C.J., Breviario, F., and Dejana, E. 2004. Gasl is induced by VE- cadherin and vascular endothelial growth factor and inhibits endothelial cell apoptosis. Blood 103: 3005-3012.

Stafford, L.J., Vaidya, K.S., and Welch, D.R. 2008. Metastasis suppressors genes in cancer. Int J Biochem Cell Biol 40: 874-891.

Stebel, M., Vatta, P., Ruaro, M.E., Del Sal, G., Parton, R.G., and Schneider, C. 2000. The growth suppressing gasl product is a GPI-linked protein. FEBS Lett. 481 : 152-158. Stecca, B., Mas, C, Clement, V., Zbinden, M., Correa, R., Piguet, V., Beermann, F., and Ruiz, I.A.A. 2007. Melanomas require HEDGEHOG-GLI signaling regulated by interactions between GLIl and the RAS-MEK/AKT pathways. Proc. Natl. Acad. Sci. USA 104: 5895-5900. Steeg, P.S. 2003. Metastasis suppressors alter the signal transduction of cancer cells. Nat. Rev. Cancer 3: 55-63.

Steeg, P.S. 2006. Tumor metastasis: mechanistic insights and clinical challenges. Nat. Med. 12: 895-904.

Takaoka, A., Adachi, M., Okuda, H., Sato, S., Yawata, A., Hinoda, Y., Takayama, S., Reed, J. C, and Imai, K. 1997. Anti-cell death activity promotes pulmonary metastasis of melanoma cells. Oncogene 14: 2971-2977.

Tsai, Y.C., Mendoza, A., Mariano, J.M., Zhou, M., Kostova, Z., Chen, B., Veenstra, T., Hewitt, S.M., Helman, L.J., Khanna, C, and Weissman, A.M. 2007. The ubiquitin ligase gp78 promotes sarcoma metastasis by targeting KAIl for degradation. Nat. Med. 13: 1504-1509.

Wetmore, C. 2003. Sonic hedgehog in normal and neoplastic proliferation: insight gained from human tumors and animal models. Curr. Opin. Genet. Dev. 13: 34- 42.

Zamorano, A., Lamas, M., Vergara, P., Naranjo, J.R., and Segovia, J. 2003. Transcriptionally mediated gene targeting of gasl to glioma cells elicits growth arrest and apoptosis. J. Neurosci. Res. 71: 256-263.

Zamorano, A., Mellstrom, B., Vergara, P., Naranjo, J.R., and Segovia, J. 2004. Glial-specific retrovirally mediated gasl gene expression induces glioma cell apoptosis and inhibits tumor growth in vivo. Neurobiol. Dis. 15: 483-491.

Example 8 Supplemental Materials and methods

Cell lines and culture

MeI-STR and MeI-STV cell lines, immortalized primary human melanocytes expressing RasV12 or empty vector, respectively, were described previously (Gupta et al. 2005) and provided by Robert Weinberg. 67NR cells, a BALB/C mouse-derived breast cancer cell line (Aslakson and Miller 1992), were provided by Fred Miller and were grown in DMEM medium supplemented with 5% fetal bovine serum, 5% newborn calf serum, IX non-essential amino acids (Invitrogen), L-glutamine (2 mM) and IX penicillin/streptomycin (Invitrogen). To generate B 16-FO cell lines expressing Bcl-2, B 16-FO cells were stably transduced with a Bcl-2 retroviral vector (pME2PUROSRαBCL2; Innes et al. 1999), provided by Suzanne Cory, and puromycin selected for 4 days.

Proliferation and apoptosis assays

The proliferation rate of the B 16-FO cells expressing a Gasl shRNA or non- silencing control, or B16-F10 cells ectopically expressing Gasl or vector was determined using CyQU ANT® Cell Proliferation Assay Kit according to the manufacturer (Invitrogen). To evaluate apoptosis in the above cell lines, protein extracts were prepared and the levels of cleaved and non-cleaved Parp-1 were analyzed by immunoblotting using an antibody against Parp-1 (BioMol).

Transwell/Boyden invasion assays

Invasion assays were performed using the CytoSelect 24-Well Cell Migration and Invasion Assay as described by the manufacturer (Cell Biolabs). In brief, B 16-FO cells expressing a Gasl shRNA or non-silencing control, or B16-F10 cells were placed in the upper compartment of the Transwell/Boyden chamber and invasion as well as migration into the lower chamber was measured 48hrs later. Invasion percentage was determined by dividing the number of cells that invaded over the number of cells that migrated. Shh signaling assays

B 16-FO cells were transduced with a Gasl or non-silencing shRNA and 7 days later, cell extracts were prepared and analyzed by immunoblotting using an antibody against GUI (Santa Cruz) or, as a loading control, Actin (Sigma). For B16-F10 cells, extracts were prepared 7 days following ectopic expression of Gasl Oncomine database searches

The Haqq (Haqq et al. 2005) and Smith (Smith et al. 2005) melanoma microarray datasets were accessed using the Oncomine Cancer Profiling Database (www.oncomine.org). The Haqq melanoma dataset includes 10 normal mole, 5 primary melanoma and 17 metastatic melanoma samples; the Smith melanoma dataset includes 6 early stage (normal, benign nevi, melanoma in situ) and 7 advanced stage (vertical growth phase melanoma, metastatic growth phase melanoma, melanoma positive lymph nodes) samples. Histograms depicting gene expression in each sample, as well as a Student's t-test giving a p value for the comparison of candidate gene expression between the groups, were obtained directly through the Oncomine 3.0 software. Analogous approaches were used to analyze expression of candidate genes in other cancers.

References for Example 8

Aslakson, CJ. and Miller, F.R. 1992. Selective events in the metastatic process defined by analysis of the sequential dissemination of subpopulations of a mouse mammary tumor. Cancer Res. 52: 1399-1405. Gupta, P.B., Kuperwasser, C, Brunet, J.P., Ramaswamy, S., Kuo, W.L., Gray,

J.W., Naber, S.P., and Weinberg, R.A. 2005. The melanocyte differentiation program predisposes to metastasis after neoplastic transformation. Nat. Genet. 37: 1047-1054.

Haqq, C, Nosrati, M., Sudilovsky, D., Crothers, J., Khodabakhsh, D., Pulliam, B.L., Federman, S., Miller, J.R., 3rd, Allen, R.E., Singer, M.I., Leong, S.P., Ljung, B.M., Sagebiel, R. W., and Kashani-Sabet, M. 2005. The gene expression signatures of melanoma progression. Proc. Natl. Acad. Sci. USA 102: 6092-6097.

Innes, K.M., Szilvassy, S.J., Davidson, H.E., Gibson, L., Adams, J.M., and Cory, S. 1999. Retroviral transduction of enriched hematopoietic stem cells allows lifelong Bcl-2 expression in multiple lineages but does not perturb hematopoiesis. Exp Hematol 27: 75-87.

Smith, A.P., Hoek, K., and Becker, D. 2005. Whole-genome expression profiling of the melanoma progression pathway reveals marked molecular differences between nevi/melanoma in situ and advanced-stage melanomas. Cancer Biol. Ther. 4: 1018-1029.

Table 3 | List of 80 metastasis suppressor genes that tested positive in the three-dimensional cell culture assay.

Table 5 | Primer sequences for quantitative real-time RT-PCR analyses

Table 6. Quantitation of satellite colony formation in the 3-D assay of Figure 7

Satellite colonies were counted as scored as follows: - (0-10 colonies), + (11-25 colonies), ++ (26-50 colonies) or +++ (>50 colonies).

This invention is not limited in its application to the details of construction and the arrangement of components set forth in the following description or illustrated in the drawings. The invention is capable of other embodiments and of being practiced or of being carried out in various ways. Also, the phraseology and terminology used herein is for the purpose of description and should not be regarded as limiting. The use of "including," "comprising," or "having," "containing," "involving," and variations thereof herein, is meant to encompass the items listed thereafter and equivalents thereof as well as additional items.

Having thus described several aspects of at least one embodiment of this invention, it is to be appreciated various alterations, modifications, and improvements will readily occur to those skilled in the art. Such alterations, modifications, and improvements are intended to be part of this disclosure, and are intended to be within the spirit and scope of the invention. Accordingly, the foregoing description and drawings are by way of example only.

We claim:

Claims

1. A method of detecting a metastatic cancer in a subject, the method comprising: obtaining a clinical sample from a subject suspected of having cancer, and determining a level expression of at least one metastasis suppressor gene (MSG) in the clinical sample, wherein if the level of expression of the MSG is reduced in the clinical sample relative to a control value, the subject has a metastatic cancer.

2. The method of claim 1 , wherein the MSG is GASl .

3. The method of claim 1 , wherein the MSG is selected from: ACTA2, ADAMTSl 6, AGL, ALG6, ATGl, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CDS, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPAS, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orf24, PRAMEF8, 4931422 A03RIK, and KIAA1276.

4. The method of any one of claims 1 to 3, wherein the step of determining the expression level of the MSG comprises comparing the level of expression of the MSG in the clinical sample with the level of expression of the MSG in a control sample, wherein a decrease in expression of the MSG in the clinical sample compared with the control sample indicates that the MSG is reduced, and wherein the level of expression of the MSG in the control sample is the control value.

5. The method of claim 4, wherein the control sample is a non-metastatic cancer tissue.

6. The method of any one of claims 1 to 5, wherein the step of determining the expression level of the MSG comprises measuring the level of an mRNA of the MSG.

7. The method of any one of claims 1 to 5, wherein the step of determining the expression level of the MSG comprises measuring the level of a protein encoded by the MSG.

8. The method of any one of claims 1 to 7, further comprising determining the level of expression of a plurality of different MSGs.

9. The method of claim 8, wherein if expression of a plurality of MSGs are reduced, the subject has a metastatic cancer.

10. The method of claim 8 or 9, wherein the plurality is at least fifteen MSGs.

11. The method of any of claims 8 to 10, wherein the plurality comprises one or more MSGs selected from: ACTA2, ALG6, CCDC39, CCNA2, CTSO, CPA2, DPP A3, GASl, PDYN, PHKAl, SETD2, SLC9A3R2, THSD7B, TOMM70A, and ZNF294.

12. The method of any one of claims 1 to 11, wherein the cancer is a melanoma, breast, prostate, ovarian, liver, sarcoma, colon, lung, bladder, gastric, head, neck, seminoma, Ewing's sarcoma, cervical or renal cancer.

13. A kit for detecting a metastatic cancer in a subject, the kit comprising: at least one container having disposed therein a reagent for detecting expression of a metastasis suppressor gene, and a label and/or instructions for use of the kit in detecting a metastatic cancer based on expression of a metastasis suppressor gene.

14. A method for suppressing one or more metastatic properties in a cell comprising increasing the activity of one or more metastasis suppressor genes (MSGs) in the cell.

15. The method of claim 14, wherein the MSGs are one or more of: ACTA2,

ADAMTSl 6, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12or/24, PRAMEF8, 4931422 A03RIK, and KIAA1276.

16. The method of claim 14, wherein the MSG is GASl.

17. The method of any one of claims 14 to 16, wherein the increasing the activity of one or more MSGs involves contacting the cell with a compound or composition efficacious at increasing the activity of the one or more MSGs.

18. The method of any one of claims 14 to 17, wherein the cell has reduced activity of the one or more MSGs induced by the compound or composition.

19. The method of any one of claims 14 to 18, wherein the cell is in vitro.

20. The method of any one of claims 14 to 18, wherein the cell is in vivo.

21. The method of any one of claims 14 to 20, wherein the cell is a tumor cell.

22. The method of any one of claims 14 to 21 , wherein the tumor cell is non- metastatic.

23. The method of any one of claims 14 to 21 , wherein the tumor cell is metastatic.

24. The method of any one of claims 14 to 23, wherein the composition is a gene therapy.

25. The method of claim 24, wherein the gene therapy comprises delivery of a therapeutically effective amount of an expression construct encoding one or more of: ACTA2, ADAMTS16, AGL, ALG6, ATGl, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP 6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orJ24, PRAMEF8, 4931422 A03RIK, and KIAAl 276.

26. The method of claim 24 or 25, wherein the gene therapy comprises delivery of a therapeutically effective amount of an expression construct encoding GASl.

27. A method for treating a subject having, or at risk of having, a tumor metastasis comprising administering to the subject an effective amount of a compound or composition that increases the activity of one or more MSGs.

28. The method of claim 27, wherein the MSGs are one or more of: ACTA2,

ADAMTSl 6, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPP A3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, MAMTl, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orf24, PRAMEF8, 4931422 A03RIK, andKIAA1276.

29. The method of claim 28, wherein the MSG is GASl.

30. The method of any one of claims 27 to 29, wherein the composition is a gene therapy.

31. The method of claim 30, wherein the gene therapy comprises delivery of a therapeutically effective amount of an expression vector encoding one or more of: ACTA2, ADAMTS16, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPP A3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDM13, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orf24, PRAMEF8, 4931422 A03RIK, andKIAA1276.

32. The method of claim 30 or 31, wherein the gene therapy comprises delivery of a therapeutically effective amount of an expression construct encoding GASl.

33. A method for identifying compounds or compositions useful as pharmacological agents for modulating one or more metastatic properties, comprising contacting a cell with a compound or composition and assaying for the increased expression of one or more MSGs, wherein a compound or composition that increases the expression of one or more MSGs is a compound or composition useful as pharmacological agents for modulating one or more metastatic properties.

34. The method of 33, wherein the MSGs are one or more of: ACTA2, ADAMTSl 6, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPA3, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NME7, NR1D2, N6AMT1, OLFR835, OPTC, OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMEL3, PRDMl 3, PRRG3, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SS18L1, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZC3H13, ZNF294, C12orf24, PRAMEF8, 4931422 A03RIK, andKIAA1276.

35. The method of claim 33 or 34, wherein the MSG is GASl.

36. The method of any one of claims 33 to 35, wherein the compound or composition contacts the cell for a predetermined period of time.

37. The method of claim 36, wherein the predetermined period of time is about 14 days.

38. The method of any of one of claims 33 to 37, wherein the cell is grown in an environment wherein one or more metastatic properties measured.

39. The method of any one of claims 33 to 38, wherein the environment is in vitro.

40. The method of any one of claims 33 to 38, wherein the environment is in vivo.

41. The method of any one of claims 33 to 40, wherein the cell is a tumor cell.

42. The method of claim 41 , wherein the tumor cell is non-metastatic.

43. The method of claim 41, wherein the tumor cell is metastatic.

44. The method of any one of claims 33 to 43, wherein the cell contacted by the compound or composition has reduced expression of one or more MSGs.

45. The method of any one of claims 33 to 44, wherein the composition is a gene therapy.

46. The method of claim 45, wherein the gene therapy comprises delivery of a therapeutically effective amount of one or more expression vectors encoding one or more of: ACTA2, ADAMTS16, AGL, ALG6, ATG7, BAIAP2, BNIPl, B630019K06RIK, CABYR, CAV2, CCDC39, CCNA2, CD5, CDC26, CENTDl, CLDND2, COL24A1, CPA2, CRKL, CTSO, DAZ2, DDHD2, DEGS2, DGKB, DKCl, DMRTAl, DNAJB2, DPPAS, FCHSDl, FIGNLl, GAL, GASl, GIMAP6, GPRASPl, HSD3B2, MAP2K2, MAZ, METTL5, MIPEP, MUC19, NBEAL2, NMEl, NR1D2, N6AMT1, OLFR835, OPTC₁ OR5AC2, OTUD4, PCLO, PCSK6, PDYN, PHKAl, PPPlRlO, PRAMELS, PRDMlS, PRRGS, RAN, RFX2, ROBO2, SCCPDH, SETD2, SLC9A3R2, SLC25A1, SPC25, SSl 8Ll, TDPOZ2, THSD7B, TOPBPl, TOMM70A, TTCl 7, UBNl, UGT2B34, UQCRC2, ZADHl, ZCCHC9, ZCSHlS, ZNF294, C12orf24, PRAMEF8, 49Sl 422 AOSRlK, and KIAAl 276.

47. The method of claim 45 or 46, wherein the gene therapy comprises delivery of a therapeutically effective amount of an expression construct encoding GASl .

48. The method of claim 39, wherein the environment comprises a collagen gel.

49. The method of claim 48, wherein the environment further comprises basement membrane.

50. The method of claim 48, wherein the environment further comprises a fibrin gel.

51. The method of claim 49, wherein the environment further comprises a fibrin gel.

52. The method of claim 49, wherein a metastatic property comprises movement of the cell through the collagen gel and/or the basement membrane.

53. The method of claim 50, wherein a metastatic property comprises movement of the cell through the collagen gel and/or the fibrin gel.

54. The method of claim 51 , wherein a metastatic property comprises movement of the cell through the collagen gel, the basement membrane, and/or the fibrin gel.

55. A method for screening for modulators of one or more metastatic properties comprising the following steps: (i) transducing test cells and control cells with pools of a plurality of retroviruses, wherein individual retroviruses in the plurality comprise a nucleic acid encoding a product capable of affecting expression of at least one gene encoded in the genome of the transduced cells

(ii) isolating test cells with one or more altered metastatic properties compared with control cells, and

(iii) identifying the transduced nucleic acid in the isolated test cells.

56. The method of claim 55, wherein the product capable of affecting expression is an shRNA or shRNA-mir.

57. The method of claim 56, wherein the shRNA or shRNA-mir is directed against the at least one gene encoded in the genome of the transduced cells.

58. The method of any one of claims 55 to 57, wherein the test cells and control cells are grown in an environment wherein one or more metastatic properties are measured.

59. The method of claim 58, wherein the environment is in vitro.

60. The method of claim 58, wherein the environment is in vivo.

61. The method of any one of claims 55 to 60, wherein the test cells are tumor cells.

62. The method of claim 61 , wherein the tumor cells are non-metastatic.

63. The method of claim 61 , wherein the tumor cells are metastatic.

64. The method of any of claims 55 to 63, further comprising performing an assay for alterations in one or more tumor metastatic properties of the test cells.

65. The method of claim 60, wherein the isolating comprises resecting metastatic tumor tissue.

66. The method of claim 60, wherein the isolating comprises resecting primary tumor tissue.

67. The method of claim 59, wherein the environment comprises a collagen gel.

68. The method of claim 67, wherein the environment further comprises basement membrane.

69. The method of claim 67, wherein the environment further comprises a fibrin gel.

70. The method of claim 68, wherein the environment further comprises a fibrin gel.

71. The method of claim 68, wherein a metastatic property comprises movement of the cell through the collagen gel and/or the basement membrane.

72. The method of claim 69, wherein a metastatic property comprises movement of the cell through the collagen gel and/or the fibrin gel.

73. The method of claim 70, wherein a metastatic property comprises movement of the cell through the collagen gel, the basement membrane, and/or the fibrin gel.

74. The method of any one of claims 55 to 57, wherein the identifying comprises cloning the nucleic acid.

75. The method of any one of claims 55 to 74, further comprising sequencing the nucleic acid.

76. The method of any one of claims 55 to 75, further comprising mining a cancer gene database to determine expression of the nucleic acid in metastatic and non- metastatic tumors.

77. The method of any one of claims 55 to 76, further comprising subjecting the isolated test cells to an in vivo assay for alterations in one or more metastatic properties.

78. The method of any one of claims 55 to 77, wherein the plurality of retroviruses comprise sequence complementary to a portion of the mRNA sequence of each of substantially all known protein coding genes of the transduced cell's genome.

79. A method for identifying a modulator of at least one metastatic property comprising: contacting a plurality of cells, in a three-dimensional culture system, with a plurality of expression vectors comprising inserts and identifying one or more inserts that alter at least one metastatic property of a cell.

80. The method of claim 79, wherein at least one insert is a coding sequence for a functional RNA, optionally wherein the functional RNA is a miRNA, a shRNA, or an shRNA-mir.

81. A method for identifying a modulator of at least one metastatic property comprising: contacting a plurality of cells with a plurality of expression vectors, which comprise an shRNA gene operably-joined to a regulatory sequence, and identifying one or more of the expression vectors in the plurality that alter at least one metastatic property of a cell.