Cancer arises from the deregulation of intracellular signaling pathways leading to uncontrolled cell proliferation and tumour formation. In many cases, cancer cells in the primary tumour disseminate and colonise distant tissues and organs to form secondary tumours by the process of metastasis. Metastasis is a complex multistep process, involving migration of cancer cells from the primary tumour, their systemic spread by the circulatory system, followed by the colonisation and growth of these cells into tumours at secondary sites. Metastatic tumours are responsible for the majority of cancer deaths. Understanding the mechanisms of metastasis is therefore crucial to understanding carcinogenesis, predicting the likelihood of primary cancer spread and devising new strategies for the treatment of metastatic cancer. Numerous pathways can affect metastasis and it is now clear that inactivation of members of the metastasis suppressor gene family plays a central role in this process in many human cancers. These genes suppress metastasis but not primary tumour growth. To date, over 20 metastasis suppressors have been discovered, which can act at various stages along the metastatic pathway. In this review we discuss the different mechanisms of action of selected metastasis suppressor genes to illustrate their diversity of action.
The development of cancer stems from cellular transformation and the ability of cancer cells to evade normal regulated processes. Cancer cells accumulate a series of defects in several regulatory processes, leading to tumourigenesis and malignancy. A number of key hallmarks are believed to be important during the development of cancer and malignancy, including self-sufficiency in growth signals, insensitivity to growth-inhibitory signals, ability to evade apoptosis, unlimited replicative potential and sustained angiogenesis. The final stages of cancer include the ability of cells to invade tissues and metastasise to distant tissues and organs.1,2
The majority of deaths among cancer patients are due to metastatic tumours rather than the primary tumour, due to impedance of the function of vital organs.2 There has been a growing interest in the study of metastasis, to gain a deeper understanding of this process with the aim of improving cancer prognosis and treatment. Metastasis is a complex, multi-step process where cancer cells disseminate from the primary tumour, colonise distant tissues and organs and grow into secondary metastatic tumours. It is proposed that cells within a primary tumour can undergo a process termed the epithelial to mesenchymal transition (EMT), to become less adherent and more motile.3,4 This aids in their ability to break through the basement membrane of an organ or tissue, allowing the cells to enter the vasculature by the process of intravasation and travel via the lymphatic or circulatory systems. The circulating tumour cells can get arrested on the lymphatic or blood vessel walls, before leaving the system and invading their new environment by the process of extravasation. For the cancer cells to proliferate and colonise the new secondary site, they are thought to undergo a mesenchymal to epithelial transition, losing their motility but leading to increased adhesion and proliferation (figure 1).5 Metastasis is a very inefficient process, with only a small proportion of cancer cells acquiring the capacity to survive each step along the metastatic cascade. Metastatic cancer cells generally colonise specific tissues and organs that are permissive for their survival and growth. In 1889, Paget observed that breast cancer metastases preferentially colonise the liver rather than other organs such as the spleen, which is subject to a similar amount of circulation. He hypothesised that the tumour cells (seeds) are distributed equally across the body and only invade organs which provide a favourable environment (soil), facilitating their colonisation. This led to the ‘seed and soil’ theory to describe organ specificity of metastatic cancer cells.6
Host-tumour interactions involving the reciprocal interaction between tumour cells and their surrounding micro-environment, inflammatory and other stromal cells, play an important role in determining which tissues and organs are colonised by cancer cells.7-14 At a molecular level, numerous enzymes such as matrix metalloproteinases, cytokines, chemokines and growth factors are important for promoting remodelling of the extracellular matrix (ECM) and for facilitating cancer cell survival, proliferation and colonisation at the secondary site.7-18
The discovery of tumour suppressor genes, such as the Retinoblastoma gene (Rb), which are mutated and inactivated in many human cancers leading to their transformation and uncontrolled proliferation,19 prompted the search for genes that may be involved in the regulation of metastasis. Early studies identified potential metastasis suppressor genes by loss of heterozygosity, comparative genomic hybridisation and karyotype analysis of chromosome abnormalities in human tumours. Chromosomes containing potential metastasis suppressor genes were then individually introduced into cells by microcell-mediated transfer. This method has been instrumental in identifying numerous metastasis suppressors.20 Subsequently, positional cloning or differential gene expression studies were used to narrow down to a region within the chromosome and eventual identification of a specific gene.21-24 To confirm its function as a metastasis suppressor, the gene is transfected into a competent cell line with low expression or activity of this gene. Various in vitro assays that measure phenotypes associated with metastasis, such as motility, invasion and colonisation abilities, are then evaluated. However, to validate a gene’s metastasis suppressor function, studies must be completed to show that its expression reduces metastasis without affecting tumourigenicity in vivo.25 To date, over 20 metastasis suppressor genes that act at various stages of the metastatic process have been identified (table 1).26 We will discuss the roles of a selection of metastasis suppressors to highlight their diverse mechanisms of action.
Non-metastatic clone 23 (NM23) was the first metastasis suppressor gene identified.27 Analysis of tumours from human hepatocellular carcinoma and gastric cancer demonstrated a negative correlation between expression of NM23 and metastasis.28,29 Transfection of NM23 into metastatically competent breast,30 melanoma,31 gastric,32 and oral squamous carcinoma cell lines,33 resulted in reduced metastasis in vivo. Re-expression of NM23 induced a reduction in cell motility of human breast cancer cells and murine melanoma cells.34 In early stage HD3 subline HT29 colon carcinoma cells, NM23 promotes transforming-growth factor (TGF-β)–induced adherence.35 TGF-β has opposing effects on cells, depending on the stage of tumour progression. During the early stages, TGF-β acts as a tumour suppressor, while in the later stages, it promotes EMT and hence metastasis.36 The product of this NM23 gene is a protein histidine-kinase and site-directed mutagenesis, demonstrating that its enzymatic activity is important for its function.37,38 NM23 regulates the Ras/MAPK signalling pathway. Therefore, overexpression of NM23 in MDA-MB-435 breast cancer cells reduces mitogen-activated protein kinase (MAPK) activity.39 NM23 co-precipitates with and phosphorylates the kinase suppressor of Ras on Serine 392, which is a binding site for the 14-3-3 kinase suppressor of Ras inhibitor.40 Therefore, phosphorylation of kinase suppressor of Ras by NM23 contributes to reduced Ras/MAPK signalling. Numerous studies have demonstrated a correlation between tumour progression and deregulation of the Ras/MAPK signalling pathway. For example, increased expression and activity of MAPK is associated with lymph node metastases in breast cancer.41 Increased activity of the Ras/MAPK signalling pathway can play several roles during tumorigenesis and metastasis, such as regulating apoptosis, cell migration and angiogenesis.42 At a molecular level, this pathway can impinge on various molecules to regulate metastasis, such as increasing the production of matrix metallopeptidase-9,43 and regulating the EMT.44
Microcell mediated transfer of chromosome 11 into the breast cancer cell line, MDA-MB-435, significantly reduced the metastatic potential of these cells in nude mice.45 Further analysis identified the metastasis suppressor function to a novel gene termed, breast cancer metastasis suppressor 1 (BRMS1). Initial metastasis studies with MDA-MB-435 and MDA-MB-231 breast cancer cells expressing BRMS1 showed that although these cells were still locally invasive, there was a significant reduction in lymph node and lung metastases.46 In addition to breast cancer, BRMS1 also reduces the metastasis of melanoma,47 ovarian,48 and non-small cell lung cancer cell lines.49 BRMS1 regulates various aspects of cell behavior. One example is the regulation of homotypic gap junctions, which are involved in intercellular communication to regulate the ability of cells to detach from primary tumours and/or respond to signals during transportation or at the secondary site.50 BRMS1 can also increase the susceptibility of cells to anoikis, which is programmed cell death induced by detachment from the extracellular matrix, thereby decreasing the likelihood of circulating cancer cells reaching and colonising secondary sites.51
BRMS1 is a protein of 246 amino acids and can regulate numerous cellular pathways.46 Yeast two-hybrid screens identified the transcriptional regulators, retinoblastoma binding protein 1 and mammalian Sin3 as BRMS1 interacting proteins. These interactions were confirmed by co-immunoprecipitation studies of lysates from MDA-MB-231 breast cancer cells expressing BRMS1.52 BRMS1 recruits the retinoblastoma binding protein 1/mammalian Sin3/histone deacetylase transcriptional repressor complex to repress transcription of various pro-metastatic genes such as osteopontin and urokinase-type plasminogen activator.53,54 BRMS1 also reduces transcription of the epidermal growth factor receptor (EGFR) to decrease AKT signalling.55 Microarray studies demonstrate that BRMS1 regulates the expression of numerous genes, such as those of the major histocompatibility complex and genes involved in protein localisation and secretion.56 Therefore, BRMS1 metastasis suppressor function is at least in part mediated through regulation of the expression of different genes that play important roles in metastasis.
The mitogen-activated protein kinase, kinase 4/stress-activated protein/Erk kinase 1 (MKK4/SEK1) gene, was identified as a metastasis suppressor following introduction of human chromosome 17 via microcell mediated transfer into the highly metastatic AT6.1 prostate cancer cell line.57 When these cells were injected into mice, there was no difference in the size of the ensuing tumours, but a significant decrease in their metastatic ability to the lungs was observed. The region encoding the metastasis suppressor was later narrowed down to ~70 centiMorgan region of DNA,57 and subsequently identified as MKK4.22 Mice inoculated with ovarian cancer cells expressing MKK4 displayed a reduced number of metastases to the liver, small bowel, near the stomach and spleen, and prolonged their survival rate by 70%. The mean survival of the mice increased from 37 to 63 days.58
Loss of heterozygosity on Chromosome 17p has been observed frequently in human ovarian cancers, implicating MKK4 in its pathology.59 Analysis of human clinical ovarian cancer samples by immunohistochemistry demonstrated a significant loss of MKK4 expression in metastases compared to the primary ovarian tumours, supporting the idea that this gene plays an important metastasis suppressor function in these tumours.58 MKK4 acts upstream of the c-Jun NH2-terminal (JNK) and p38 kinase signalling pathways, which respond to stress stimuli.60 In the presence of cellular stress such as irradiation, DNA damage or in response to proinflammatory cytokines, MKK4 is activated by upstream activators and becomes phosphorylated. MKK4 then phosphorylates and activates JNK and p38 kinases, which mediate downstream events.61 The importance of the role of MKK4 as a kinase in the suppression of metastasis was demonstrated in studies where human ovarian cancer cells, SKOV3ip.1, expressing catalytically-inactive MKK4 mutant resulted in significantly more metastases in mice than cells expressing active MKK4.62 Activation of the JNK and p38 pathways typically leads to apoptosis.63 Therefore, MKK4 at least in part, mediates its metastasis suppressor effects by inducing apoptosis, removing the ability of cancer cells to survive, proliferate, migrate and colonise new sites.
Kang-Ai 1 (KAI1) was identified following initial microcell mediated transfer studies of chromosome 11 into the rat AT3.1 prostate cancer cell line. When these cells were injected into mice, it was found that the region p11.2-13 significantly suppressed the number of lung metastases.64 The KAI1 gene was later identified when DNA fragments from chromosome 11p11.2-13 were used as probes to screen cDNA libraries obtained from both metastasis-suppressed and non-suppressed microcell hybrid AT6.1 cells.23 The expression of KAI1 is deregulated in prostate,65 pancreatic,66 non-small cell lung,67 colon,68 colorectal,69 and breast cancers.70 KAI1 affects several cellular functions, such as migration and adhesion, which are often altered in cancer cells during metastasis. Therefore, studies using the stable colon cancer cell lines, BM314 with KAI1 knocked-down and DLD-1 cells overexpressing KAI1, were completed to assess these aspects of KAI1 function.68 DLD-1 cells overexpressing KAI1 displayed reduced phagokinetic motility and migration through a filter coated with reconstituted basement membrane, a measure of invasiveness. The opposite effect was seen with cells expressing reduced KAI1. In addition, cells overexpressing KAI1 displayed a significant increase in binding to ECM components, such as fibronectin. Wound healing assays with fibronectin coated plates showed that knock-down of KAI1 in BM314 cells induced quicker migration on to the fibronectin-coated surface compared to control cells. Therefore, a major mechanism of KAI1 metastasis suppressor function is likely through its ability to reduce cancer cell migration and increased adhesion to the ECM.
KAI1 is a glycosylated protein of 46-60kDa containing peptide motifs, thereby placing it in the tetraspanin family that function as adaptors for large cell surface molecules.71 Although the exact molecular mechanism behind the role of KAI1 as a metastasis suppressor remains to be fully defined, studies to date indicate that it can attenuate signalling of the EGFR pathway. Co-immuniprecipatetion studies showed that KAI1 associates with EGFR.72 In wound healing assays, HB2 human mammary epithelial cells overexpressing KAI1 displayed reduced epidermal growth factor (EGF)-induced migration. Morphological differences were also observed following EGF-induced migration of cells overexpressing KAI1, with cells displaying fewer lamellipodial protrusions. Functional studies indicated that KAI1 promotes EGFR endocytosis, suggesting that this is the mechanism of KAI1 attenuation of EGFR signalling. EGFR signalling is a major pathway involved in promoting the proliferation of many cells and this pathway is deregulated in many cancer types.73 In terms of metastasis, EGFR signalling is known to increase the production of matrix metalloproteinases-9 in breast cancer cells and enhance the invasiveness in prostate cancer cells.74,75
The number of metastasis suppressor genes continues to increase. As already discussed, metastasis suppressors may regulate numerous aspects of cellular behaviour, such as apoptosis, anoikis, maintaining inter-cellular or cellular interactions with the surrounding ECM to regulate EMT. Tumour cells that undergo EMT and intravasion must be able to survive transport through the vasculature, extravasation and evade apoptosis at the new secondary site before establishing new colonies. Currently, over 20 metastasis suppressors that impinge on different aspects of the metastatic cascade have been identified (table 1). It is likely that as new genes in this family are discovered, novel mechanisms of metastasis suppression will be unveiled, providing new insights into this complex process.
Although our understanding of the biological mechanism of metastasis and the action of metastasis suppressors is increasing, significant challenges remain to translate this knowledge into a clinical setting for improved patient outcome. A major goal of clinicians is early detection of the cancer before metastasis occurs. Early detection is associated with better prognosis and treatment is less challenging when the cancer is localised to the primary site and has not metastasised. Metastasis suppressor genes may eventually be useful as prognostic markers to define the likelihood of primary tumour spread and response to therapy. For example, various cancers have shown high expression of metastasis suppressor genes such as NM23 and KAI1 in primary tumours, with a reduction in matched metastases.65,76 Further clinical studies will be needed to determine if expression of these genes can predict outcome and thus provide utility for prognosis or therapeutic responses.
Apart from their prognostic potential, metastasis suppressors may provide new targets for cancer therapy. At this stage there are significant challenges in targeting metastasis suppressors as therapeutic targets, since it is envisaged that compounds would need to activate or restore their activity, as opposed to many anti-cancer compounds that bind and inhibit key molecules, oncogenes and pathways required for cancer cell survival. Nevertheless, anti-cancer drugs such as the histone-deacetylase inhibitor Vorinostat that has broad effects on the expression of many genes, demonstrates that compounds may be developed that activate tumour suppressor genes and pathways.77-79 The development of compounds that increase the expression or activity of metastasis suppressors, could open new possibilities for treatment of cancer.