Print This Page
Cancer Genes and Genomics
BAC array CGH revolutionizes high-resolution analysis of genetic changes in cancer cells
The use of high density BAC arrays for array-based comparative genomic hybridization (aCGH) analysis to identify genome-wide losses, gains and amplifications in primary tumors and cell lines by Norma Nowak, PhD and colleagues (Nature Genet 2001; 29:263:459) is a major contribution to the genomic analysis of cancer cells.
The development of BAC arrays was built on the pioneering effort of RPCI scientists who created large insert BAC clones as part of an NCI subcontract to provide the substrate for the sequencing of the human genome. Over 70% of the human genome sequence currently available in the public databases was derived from these human BAC libraries. These resources have since been made generally available to the scientific community which has led to innovative applications of BAC technology for the study of human cancer genetics.
One of these creative innovations has been the development of high density microarrays of BACs to detect the hallmark copy number abnormalities of cancer cells. The current RPCI custom aCGH platform carries ~19K BAC clones giving almost uninterrupted coverage of the human genome and was designed from the RPCI-11 BAC library (Cowell and Nowak, Adv Cancer Res 2003; 90:91). This array represents the highest density human array in current common usage. Significant application of these arrays has been in the identification of constitutional chromosome rearrangements in cancer susceptibility syndromes (Cowell et al., Br J Cancer 2004; 90:860) and other genetic disorders (Drazinic et al., Am J Med Genet 2005; 134:282).
The small amounts of DNA required for this analysis and the fast turnaround time coupled with high resolution of the definition of the breakpoints involved in the CNAs makes this an ideal approach for defining structural chromosome changes which could potentially replace conventional cytogenetics in this diagnostic arena to detect and characterize syndrome-related deletions and duplications.
Dr Nowak, Dr Hicks, Mary Reid, PhD (Cancer Prevention and Population Science, Daniel Gaile, PhD (Biostatistics), Maureen Sullivan, DDS, Thom Loree, MD and Nestor Rigual, MD propose to use a multi-dimensional, discovery-driven approach to perform a comprehensive, genome-wide survey of head and neck squamous cell carcinoma (HNSCC), in our effort to identify and validate genetic alterations in the progression of HNSCC. Studies examine cases of early stage disease oral leukoplakia (OL), primary tumor (no cervical lymph node metastases), recurrent disease and late stage (primary tumor and cervical lymph node metastases with/without distant metastases) and will use array based aCGH to detect copy number changes in OL and tumor samples.
Aberrations in genomic copy number present in early lesions, will likely be present in epithelial tissue that appears histologically normal, and may be responsible for the high recurrence and mortality rate of HNSCC. Studies to validate this approach demonstrate that copy number alterations (CNAs) can be robustly and reproducibly detected by aCGH in DNA isolated from challenging tumor types and sources, including archival materials, low DNA yield and heterogeneous tissues (Nowak et al., 2007; Genet Med 2007; 9:585.
In collaboration with Dr Allan Bradley (Sanger Institute, UK) Yuejin Yu, PhD was involved in establishing a whole-genome mouse BAC microarray mostly using the RPCI RP-23 library with a modest ~1-Mb resolution (Chung et al., Genome Res 2004; 14:188). Custom mouse BAC arrays have also been generated by Genetics Program Members that have improved this resolution to <500 kb and these arrays are being used within the Program in the analysis of genetic changes in the mouse models of cancer theme (Hackett et al., Cancer Res 2003; 63:5266; Li et al., Human Molec Genet 2007; 16:11:1359. Recently, these arrays were also used to define polymorphic variations in copy number of specific regions throughout the mouse genome between different strains of mice commonly used in the identification of tumor modifier loci (Snijders et al., Genome Res 2005; 2:302).
aCGH has also been used extensively to characterize the subtle genetic abnormalities in cancer cells that are not detectable using lower resolution chromosome based approaches. Implicating genes in cancer development through analysis of structural chromosome changes requires high resolution definition of specific breakpoints to accurately delineate the critical regions of the genome that are involved.
Thus, in studies by Dr Cowell of low grade oligodendrogliomas aCGH has defined small, consistent chromosome deletions on chromosomes 11 and 13 (Rossi et al., Genes Chromosomes Cancer 2005; 44:85) as well as novel amplicons in high grade gliomas (Rossi et al., Genes Chromosomes Cancer 2005; 44:392). The characterization of previously unidentified amplicons in tumors, especially where they occur in combination, offers an opportunity to investigate the relationship between these abnormalities and subtle differences in clinical outcomes within tumor types.
Genome-wide aCGH analysis of other tumor types have revealed genetic changes which serve as potential biomarkers for clinical outcome as well as identifying regions of the genome that contain genes related to tumorigenesis. Thus, the unequivocal identification of 1p/19q loss in anaplastic oligodendroglioma, which has been shown to predict response to chemotherapy, using aCGH provides the platform for rapid and objective classification of these tumors for particular treatment regimens (Cowell et al., J Neuropathol Exp Neurol 2004; 63:151).
Amplication of the NYMC gene in neuroblastoma
The high resolution analysis has also been incorporated into clinical trials as a means of investigating response to therapeutic interventions, for example, EGFR inhibitors in adult brain tumors (Lassman et al., Cancer Res 2005; 11:7841) and in investigations of the genetic damage incurred in nuclear fallout areas in Chernobyl (Varma et al., Br J Cancer 2005; 93:699). In these studies by Joseph Geradts, MD, two cohorts of breast cancers from RPCI and Gomel, Belarus, an area exposed to the nuclear fallout from the 1986 Chernobyl accident have been under investigation.
Belarussian tumors were characterized by a higher number of amplifications and fewer deletions. Unsupervised hierarchical clustering clearly distinguished the different genotypes of the RPCI and Belarussian breast cancers suggesting that these differences may be due to radiation exposure. Novel genetic events in gliomas (Cowell et al., Cancer Genet Cytogenet 2004; 151:36), pancreatic cancer (Nowak et al., Cancer Genet Cytogenet 2005; 161:36), bone tumors (Smith et al., Genes Chromosomes Cancer 2006; 45:957), and myeloid sarcomas (Deeb et al., Genes Chromosomes Cancer 2005; 44:373) have revealed critical genetic events in the etiology of these tumors.
Although, the characterization of CNAs in tumors has proved valuable in its own right, the true value lies in the ability to precisely define the nature of these abnormalities so that the critical genetic consequences of these events can be established as a basis for understanding cancer development. In proof-of-principle experiments, Dr Cowell and colleagues have combined the analysis of aCGH with oligonucleotide expression arrays to define the ‘driver’ genes whose expression is altered by these events Rossi et al., Genes Chromosomes Cancer 2005; 44:85; Lo et al., Genes Chromosomes Cancer 2007; 46:53; Lo et al., Brain Pathol 2007; 17:3:282).
Development of novel software facilitates synchronous analysis of microarray data
The combination for synchronous analysis of gene expression data and CGH data has led to the development of software tools to computationally overlay the large datasets derived from CGH, SNP and gene expression platforms (Lo et al., Cancer Inform 2007; 3:307), which provides rapid identification of cancer related genes. This has been achieved by the research environment within the Genetics Program which brings together molecular biologists (Drs Cowell, Nowak, Ionov, and Dr Hawthorn), pathologists (Dr Hicks and George Deeb, MD), computational scientists and bioinformaticians (Ping Liang, PhD) to define priorities in data management and devise solutions to these problems. This analysis has been applied to the analysis of adult and pediatric brain tumors (Lo et al., Genes Chromosomes Cancer 2007; 46:53; Rossi et al., Genes Chromosomes Cancer 2006; 45:290; Lo et al., Brain Pathol 17:3:282), for example, where specific genes involved in losses and amplifications have been identified.
There have been other significant accomplishments that have resulted from the interaction between Dr Nowak and members of the scientific community in the wider application of the BAC libraries to genome research. The technology to identify gene expression patterns of specific genes for which antibodies are not available has been developed through BAC recombineering where fluorescence reporter genes are placed in the context of the regulatory elements of the gene. By creating mice which are transgenic for these BACs the expression of the reporter gene defines the specific cell types which express it endogenously (Gong et al., Nature 2003; 425:917) which forms the underlying basis for the NIH-sponsored GENSAT Program. This technology has been successfully applied by Dr Cowell to the analysis of the expression of the LGI1 metastasis suppressor gene (Head et al, Mamm Genome 2007;18:5:328).High-resolution chromosome analysis identifies underlying genetic events in cancer
In collaboration with Dr Matsui, work by Dr Karpf (Molecular Targets and Experimental Therapeutics Program) analyzed the chromosomes in a colon cancer cell line where the DNA methyltransferase 3A gene had been deleted and demonstrated that these cells were undergoing chromosomal catastrophie with high levels of structural chromosome rearrangements which are presumably related to an inability to maintain the methylation patterns in these cells (Karpf and Matsui, Cancer Res 2005; 65:8635). These findings have profound implication for strategies using demethylating agents as therapeutics which, in light of the above findings, could potentially promote tumor progression. A similar phenomenon, although to a lesser degree, was seen in cells engineered by Keshav Singh, PhD (Cell Stress and Biophysical Therapies Program) to be devoid of mitochondria (Singh et al., Gene 2005; 354:140).
Analysis of these cells by Dr Matsui and Ivan Still PhD, demonstrated an increased occurrence of structural chromosome abnormalities compared with the wild-type cells reinforcing the idea of a crosstalk between the mitochondrial and nuclear genomes. FISH has been used as a valuable means of verifying aCGH based observations (Rossi et al., Genes Chromosomes Cancer 2005; 44:85; Rossi et al., Genes Chromosomes Cancer 2005; 44:392). FISH was used in a collaboration between Terry Beerman, PhD, (Molecular Targets and Experimental Therapeutics Program) and Dr Matsui, to identify DNA breakage through activation of H2AX in a PIKK- independent manner following treatment with hedamycin, a DNA alkylator; in addition, chromosomal alterations and telomere dysfunction were detected following exposure to the enediyne, C-1027 (McHugh et al., Cancer Res 2005; 65:5344; Tu et al., Mol Cancer Ther 2005; 4:1175). FISH has been an important means of analyzing genetic abnormalities in cultured embryonic stem cell lines. In collaboration with Dr Aravinda Chakravarti (Johns Hopkins Cancer Center), Dr Matsui was able to demonstrate that genetic drift had occurred in these cells that had deletions of critical genes such as the FYN oncogene in commonly used stem cell lines. These results have important implications for their widespread use (Maitra et al., Nat Genet 2005; 10:1099).
In collaboration with Dr Baumann (Tumor Immonology and Immunotherapy Program) and Greg Loewen, MD (Cancer Prevention and Population Sciences Program), Dr Matsui also monitored chromosome changes associated with the transformation of human bronchial epithelial cells (Loewen et al., BMC Cancer 2005; 5:145). FISH analysis was also an essential approach to the characterization of the megabase duplications achieved by Dr Yu as part of the chromosome engineering aspect of the Mouse Genetics Theme (Li et al., Hum Molec Genet 2007; 16:11:1359).
A novel genome wide approach for the discovery of tumor suppressor genes
Recent pioneering work by Drs Ionov and Cowell has developed a genome-wide solution to the search for tumor suppressor genes by manipulating the nonsense mediated decay (NMD) pathway (Ionov et al., Oncogene 2004; 23: 639). This approach, which reveals mRNA species carrying chain-terminating codons, is facilitated by the high resolution capabilities of the Affymetrix oligonucleotide arrays provided by Dr Hawthorn. In early studies, genome scans for novel mutated genes successfully identified p300 as a critical tumor suppressor gene in the development of colon cancer (Ionov et al., Proc Natl Acad Sci USA 2004; 101:1273). These studies were extended to the identification of novel tumor suppressor genes in prostate cancer (Rossi et. al., Cancer Genet Cytogenet 2005; 161:97) such as the Janus Kinase, JAK1, which is central to cell proliferation signaling in cancer cells and the activin receptor ACVR2, which is responsible for growth factor signaling (Rossi et al., Cancer Genet Cyotgenet 2005; 163:123).
Recently, Dr Ionov in collaboration with Drs Cowell and Hawthorn described a significant improvement of the NMD technology which has further enhanced the ability to define tumor suppressor genes (Ivanov et al., Oncogene 2007; 26:20:2873). As a result of this development, and as an essential proof-of-concept, many novel tumor suppressor genes have been identified in colon cancers showing microsatellite instability.
Human cervical cancer is associated with human papilloma virus infection; however, this infection is not sufficient to induce transformation and progression of these epithelial cells. Loss of heterozygosity analyses suggests the presence of a tumor suppressor gene (TSG) on chromosome 6p21.3-p25. Dr Coignet, has identified NOL7, mapped this gene to 6p23 and localized the protein to the nucleolus. FISH analysis demonstrated an allelic loss of an NOL7 in cultured tumor cells and human tumor samples. Transfection of NOL7 into cervical carcinoma cells inhibited their growth in mouse xenografts, confirming its in vivo tumor suppressor activity. The induction of tumor dormancy correlated with an angiogenic switch caused by a decreased production of vascular endothelial growth factor and an increase in the production of the angiogenesis inhibitor, thrombospondin-1. These data suggest that NOL7 may function as a TSG in part by modulating the expression of the angiogenic phenotype (Hasina et al., Oncogene 2006; 25:588).
Characterization of novel oncogene-related mechanisms of tumorigenesis
Analysis of the ZNF198/FGFR1 fusion kinase from myeloproliferative disease by Drs Cowell and Dr Baumann, PhD (Tumor Immunology and Immunotherapy Program) has demonstrated that the chimeric protein is constitutively activated and regulates the activity of members of the STAT family of transcription factors. Gene expression array analysis in collaboration with Dr Hawthorn, demonstrated this fusion kinase gene could up-regulate the PAI-2 (SERPINB2) plasminogen inactivator (Kasyapa et al., Blood 2006; 107:3693). As a result of this genetic event it was demonstrated that expression of the ZNF198/FGFR1 fusion gene is associated with specific PAI-2 -mediated resistance to apoptosis which may contribute to the highly malignant nature of leukemic cells carrying this fusion kinase gene.
In addition, Dr Cowell’s group has used mass spectroscopy to demonstrate that the ZNF198/FGFR1 fusion gene also interacts with the cellular splicing machinery which may be another factor contributing to the transforming capabilities of this powerful oncogene (Kasyapa et al., Exp Cell Res 2005; 309:78). Recent work has demonstrated that this fusion kinase affects the formation of PML bodies and prevents the sumoylation of the PML protein which is essential for its function (Kunapuli et al., Exp Cell Res 2006; 312:3739). Disruption of PML bodies is a critical event in the development of promyelocytic leukemia and suggests a common etiology for these two leukemias and hence possibly a common treatment. ZNF198-FGFR1 induces high SERPINB2 expression levels and HSPA1A proteins which have been shown to be important for the stability of the chimeric protein (Kasyapa et al., Blood 2006; 107:3693; Kaysapa et al., J Chem Biol 2007; In press) suggesting that targeting these proteins might be a viable approach for the treatment of myeloprolifierative diseases carrying the ZNF198-FGFR1 rearrangement.
The Src oncogene has also been shown to be significantly upregulated in metastatic cancers and recent work by Dr Gelman shows that v-Src downregulates the STE-20-like kinase, SLK, which normally controls cell shape and motility by regulating actin stress fiber formation, through a casein kinase II-dependent mechanism (Chaar et al., J Biol Chem 2006; 281:28193). Dr Gelman's laboratory has also investigated the role of the focal adhesion kinase, FAK, in Src-mediated oncogenic transformation.
A recent study (Moissoglu et al., J Biol Chem 2003; 278:47946) demonstrated that v-Src could rescue cellular polarity, motility and oncogenic transformation parameters in the absence of FAK. Interestingly, the anchorage-independent growth of FAK-/-[v-Src] cells was roughly 7- to 10-fold higher than FAK+/+[v-Src] controls, correlating with a concomitant super-activation of PI3K, but not ERK or STAT-3, pathways (Moissoglu et al., Biochem Biophys Res Commun 2005; 330:673). Moreover, FAK, Src and PI3K normally form a tripartite complex in which FAK occupies a "favored" SH2 binding site on v-Src facilitating its phosphorylation and activation, whereas PI3K is typically in a lower activation SH3 site on v-Src. Thus, FAK can act either as a positive or negative inducer of Src-mediated oncogenic progression depending on the parameter studied.
Previous work in Dr Coignet’s lab demonstrated that the co-repressor of transcription SMRT is down-regulated, not only in transformed NHL (Song et al., Cancer Res 2005; 65:4554), but also in metastatic tumors from numerous cancers. To explain why the SMRT gene is targeted by genomic rearrangements, giving rise to this down-regulation, Dr Coignet showed that a fragile site was localized within the gene itself which is the basis of a prognostics assay for tumor metastasis (see below).
Nicoletta Sacchi, PhD has investigated aberrant AML1-MTG transcription factors generated in leukemia and secondary myeloid malignancy consequent to chemotherapy and found that the MTG family of proteins act as novel chromatin adapters that network corepressor proteins, histone modifying enzymes and DNA-binding transcription factors (Hoogeveen et al., Oncogene 2002; 21:6703; Rossetti et al., Genomics 2004; 84:1). Fusion of either MTG8 or MTG16 with the transcription factor AML1, which is crucial for definitive hematopoiesis, converts AML1 from a transcriptional activator into a transcriptional repressor .
Dr Coignet found that over expression of the NOTCH1 ligand, JAG2, was also seen in the malignant plasma cells in multiple myeloma patients. Specifically JAG2 induces the secretion of IL6, VEGF, and IGF-1 from the stromal cells in the bone marrow micro-environment. This antibody blocks secretion and decreases the chemoresistance from myeloma cells, chemoresistance that is conferred to myeloma cells through their interaction with the bone marrow micro-environment (Houde et al., Blood 2004; 104:3697).
Dr Coignet’s lab developed an anti-jag2 mab directed against the jag2 binding peptide with its receptor notch. In a SCID/Hu mouse model for human multiple myeloma treatment of seven groups of mice with the anti jag2 antibody as well as isotype controls showed no difference in survival in 2 control groups but important survival increase in the other 5 groups when treated with the jag2 mab (up to 50 percent increase in one group).
Novel cancer genes that affect chromosome segregation
Genes affecting chromosome segregation have frequently been shown to have abnormal function in the development of many cancers. Studies by Dr Cowell’s group cloned the EVI5 gene from a reciprocal translocation breakpoint and showed that it is an essential component of the mitotic mechanism (Faitar et al., Genomics 2005; 86:594), where it is located in the centrosome during interphase but moves to the mitotic spindle during mitosis and then occupies critical positions in the developing cleavage furrow during cytokinesis (Faitar et al., Exp Cell Res 2006; 312:2325). RNAi knockdown of this protein results in failure of cytokinesis.
LTG Mass spectroscopy analysis identifies the RAB11 protein from GST-pulldown
Proteomics analysis of the function of EVI5 shows that it binds other key proteins related to mitosis such as survivin, INCENP and aurora kinase B. EVI5 carries a TBC domain in the N-terminal end of the protein which is considered to be a GTPase activating domains for small Rab-type GTPases. Using linear ion trap mass spectroscopy, RAB11 was shown to specifically interact with EVI5 and be activated by it (Dabbeekeh et al., Oncogene 2007; 26:2804). These observations suggest that EVI5 is marshalling essential membrane components to the cleavage furrow during cytokinesis using RAB11 directed vesicle transport.
Vesicle trafficking has proved to be an important pathway in a variety of human diseases including cancer. Richard Swank, PhD’s group, in collaboration with Rosemary Elliott, PhD and using bioinformatics tools developed by Dr Liang (Liang et al., BMC Genomics 2005; 6:126:1) have investigated fundamental mechanisms of vesicle trafficking in model systems and identified genes that facilitate this process through common protein complexes (Suzuki et al., Proc Natl Acad Sci 2003; 100:1146; Zhang et al., Nat Genet 2003; 33:145; Li et al., Nat Genet 2003; 35:84).
For example, the HPS1 and HPS4 proteins occur together within the BLOC-3 protein complex (Chiang et al., J Biol Chem 2003; 278:20332), and the HPS3, HPS5 and HPS6 proteins are components of the BLOC-2 complex (Gautam et al ., J Biol Chem 2004; 279:12935). Other genetic mechanisms involve Slc357a11 (Chintala et al., Blood 2007; 109:1533) and SLC7A11 (Chintala et al., Proc Natl Acad Sci USA 2005; 102:10964). These findings are important, not only because they identify basic functional mechanisms of these proteins, but also because they predict that therapies which are successful for mutations in one member of a complex may likewise be applicable to mutations within other members of that complex.
The discovery of novel cancer related genes has opened up the possibility of understanding their function with a view to developing better ways of stratifying tumors and developing novel therapeutic strategies. Dr Still has been investigating the role of the novel TACC family of proteins in oncogenesis, particularly in breast cancer where they show altered expression. Previous work showed that the TACCs interact with the mitotic spindle during cell division, and recently he has demonstrated that each of the three TACC proteins interact with different members of the aurora kinase family of mitotic regulators (Tien et al., Mol Cell Proteomics 2004; 3:93).
The TACC proteins have now been shown to bind to elements of the SWI/SNF chromatin remodeling complex (Lauffart et al., Genomics 2003; 81:192) and also to histone acetyltransferases (Gangisetty et al., Oncogene 2004; 23:2559). This suggests that one of the functions of the TACCs could be to perform an assembly or coordination function, bringing elements of the chromatin remodeling, transcriptional and post transcriptional machinery together in the nucleus.
New insights into the genetic control of metastasis
One particular theme within the Genetics Program relates to a new field, that of cancer metastasis genes. These genes, while not significantly affecting the growth potential of primary tumor cells in vitro or in vivo, have a profound effect on their ability to migrate, invade, metastasize to, and survive and induce neovascularization in distant organs. Given that most cancer-related deaths are associated with metastatic disease, the identification and analysis of this class of genes will be critical to the next generation of diagnostics, prognostics and therapeutics.
Andrei Bakin, PhD has demonstrated that tropomyosin-1 critically contributes toTGF-ß control of cell morphology and motility. TGF-ß, up-regulates tropomyosin-1, which normally functions by binding and stabilizing actin fibers, and inhibits cell motility and invasion. Metastatic tumor cells express low levels of tropomyosin-1 and do not form stable actin fibers in response to TGF-ß. In these cells TGF-ß stimulates migration and invasion. Thus, tropomyosin-mediated actin fiber formation contributes to the control of cell motility and invasion by TGF-ß. The loss of this TGF-ß response marks the metastatic switch in TGF-ß function in cancer progression (Bakin et al., Mol Biol Cell 2004; 15:4682).TGF-ß signaling in breast cancer cells contributes to tumor angiogenesis and invasiveness by stimulating matrix metalloproteinase 9 (MMP-9).
Expression of kinase-inactive ALK5 (TGF-ß type I receptor) reduces tumor invasion and tumor vascularization in orthotopic xenografts in SCID mice. In contrast, constitutively active ALK5-T204D enhances tumor invasion and angiogenesis by stimulating expression of MMP-9. The knock down of MMP-9 by RNA interference (RNAi) reduces tumor invasion and neovasculature formation, and increases tumor cell death. This study demonstrated that MEK-ERK signaling is required for up-regulation of MMP-9 by the TGF-β-ALK5 axis, whereas JNK, p38 MAPK, and Smad4 are not. Thus, the TGF-β-ALK5-MEK-MMP-9 pathway in tumor cells promotes tumor angiogenesis. Given that TGF-β-Smad4 signaling functions as a tumor suppressing process, these data open the possibility for selective suppression of TGF-β-mediated tumor angiogenesis and metastasis by targeting MAPK and MMP-9 (Safina et al., Oncogene, 2007;26:240).
A novel function of TGF-β-Smad signaling in control of glutathione (GSH) metabolic pathways was recently identified. TGF-β suppresses major enzymes of GSH biosynthesis and the GSH-mediated detoxification system, leading to a reduction of GSH and an increase of ROS levels. Smad3 and Smad4 regulate transcriptional repressors that prevent expression of the metabolic genes, suggesting an important role in normal homeostasis and cancer progression (Bakin et al., Free Radic Biol Med 2005; 38:375).
Work in Dr Cowell’s laboratory identified the WAVE3 gene, which is involved in actin cytoskeleton remodeling, as a potential metastasis-inducer (Sossey-Alaoui et al., Mamm Genome 2003; 14:314). WAVE3 promotes tumor metastasis in breast, prostate and colon cancer cells through the regulation of MMP expression (Sossey-Alaoui et al., Exp Cell Res 2005; 308:135). siRNA experiments demonstrated that WAVE3 controls cell migration and invasion through lamellipodia formation and that disruption of its function prevents invasion downstream of a PI3 kinase-dependent pathway (Sossey-Alaoui et al., J Biol Chem 2005; 280:21748).
In an intra-programmatic collaboration between the laboratories of Dr Cowell and Drs Bakin and Hicks further insights were provided into the role of p38 and MMPs in WAVE3-associated metastasis. This study shows that WAVE3 is elevated in malignant breast cancers and contributes to the metastatic potential of tumor cells by regulating p38 MAPK. Knockdown of WAVE3 by siRNA in the human breast cancer MDA-MB-231 cell line reduced cell motility, migration, and invasion, which correlated with a reduction in active p38MAPK. In xenograft models, tumor cells expressing siRNA-WAVE3, or dominant-negative p38MAPK, showed a significant reduction in the formation of lung metastasis. Thus, these studies provide direct evidence that the WAVE3-p38 axis promotes cancer progression and metastasis (Sossey-Alaoui et al., Am J Pathol 2007; 170:6:2112) and as such may be a target for future therapeutic intervention.
Dr Gelman recently cloned the SSeCKS gene from 6q24, which was shown to be inactivated in prostate cancer progression. In recent studies SSeCKS was shown to inhibit the production of macroscopic lung metastases induced by MatLyLy prostate cancer cells without significantly affecting the growth at the primary-site tumor. Re-expression of SSeCKS using a tetracycline-regulated system allowed the cells to metastasize to the lungs but they could only produce relatively avascular micrometastases (Su et al., Cancer Res 2006; 66:5599).
Suppression of lung metastasis by SSeCKS in an experimental metastasis mouse model shows high numbers of tumors in cells not expressing SSeCKS (top) compared with the same cells forced to reexpress SSeCKS
Expression microarray data indicated that SSeCKS downregulates several genes important for proliferative signaling, such as PDGFR, CDK-2, and cyclin D, and genes involved in angiogenesis, such as HIF-1a and VEGF (Liu et al., BMC Cancer 2006; 6:105). SSeCKS has been shown to function as an angiogenesis antagonist, in as much as conditioned media from SSeCKS over-expressing cells inhibits tube formation of vascular endothelial cells. Indeed, in one study (Su et al., Cancer Res 2006; 6:5599), the tetracycline-regulated re-expression of SSeCKS in MatLyLu cells significantly decreased even primary tumor vascularization when injected subcutaneously in nude mice, yet decreased the production of macroscopic lung metastases by >95%. In this system, the forced re-expression of either VEGF165 or 121 isoforms rescued the ability of SSeCKS-reexpressing MatLyLu cells to form macroscopic lung metastases suggesting that SSeCKS inhibits metastasis by suppressing neovascularization at the metastatic site.
In a further analysis of the functions of SSeCKS, Dr Gelman has shown that its re-expression in v-Src-transformed 3T3 cells inhibits the formation of podosomes, correlating with a decrease in the ability of the cells to invade through Matrigel. Podosomes, also known as invadopodia, are membrane protrusions enriched with MMPs and are considered critical structures that facilitate metastatic invasiveness. SSeCKS facilitates podosome inhibition without affecting the ability of v-Src to phosphorylate cellular substrates, including the podosome protein, Tks5. SSeCKS inhibited podosome formation by inducing RhoA- and Cdc42-dependent cytoskeletal remodeling, specifically by decreasing the activation levels of all the Rho-family GTPase members (Gelman and Gao, Mol Cancer Res 2006; 3:151).The implication of this work is that SSeCKS, a scaffolding protein that controls both proliferative and cytoskeletal signaling pathways, prevents oncogenic progression by maintaining a normalized cytoskeleton. In exciting new studies Dr Gelman’s group have now shown that SSeCKS-deficient mice develop precursor hyperplastic lesions suggestive of prostatic intraepithelial neoplasias. These data indicate that loss of SSeCKS expression, as occurs in many cancers, renders them more tumor and/or metastasis-prone. These data also suggest that SSeCKS is a potential therapeutic target for metastasis and that SSeCKS-/- mice may be used to identify novel anti-cancer compounds.
The LGI1 gene, which was discovered by Dr Cowell, was also identified as a metastasis suppressor gene in high grade glioma cells which have inactivated this gene through promoter methylation. Re-expression of exogenous LGI1 in glioma cells that are null for its activity results in loss of invasion and growth in soft agar (Kunapuli et al., Oncogene 2003; 22:3985). It was subsequently shown that invasion in these cells is controlled by inactivation of the MEK/ERK pathway which results in suppression of MMP production (Kunapuli et al., J Biol Chem 2004; 279:23151). LGI1 is a secreted protein and application of this secreted protein to cancer cells results in the inactivation of ERK signaling suggesting that LGI1 as a peptide may have potent anti-metastasis capabilities; work is underway to design peptidomimics that have the same effect.
