Principles of Metastasis

In spite of the advances made in diagnostic approaches, surgical techniques, adjuvant chemotherapy, and radiation therapy, metastasis to distant sites remains the leading cause of cancer related death [1-3]. Unfortunately these metastatic foci can also be resistant to conventional therapies [4]. A significant amount of research has focused on the biologic and molecular events involved in this process, however, much remains unknown. The purpose of this chapter is to provide an overview of the complex process of metastasis.

Metastatic Cascade

The process of metastasis is a series of sequential events or steps that must be completed for successful dissemination of a tumor to occur [5]. The metastatic cancer cell has often been referred to as the decathlete of cells for its ability to leave the primary tumor, enter the circulation, evade host defenses and physical stress, exit the vasculature and recognize an appropriate distant organ where it must interact in a sometimes hostile environment, divide, and continue to grow [6]. Because of theses multiple complex steps one can imagine that metastasis is a highly inefficient process. In fact, experimental models have shown that less than 0.01% of cells injected intravenously actually form viable tumor foci [5]. Additionally, the process of metastasis is not random and each step can be rate limiting, with failure to succeed at any point resulting in disruption of the process.
Several steps in this process, often referred to as the metastatic cascade, are discussed in the following.

Angiogenesis

As a tumor grows beyond a size that can be sustained by simple diffusion of nutrients from the immediate environment, hypoxic areas arise and the tumor initiates development of its own blood supply. This is usually the result of increased production of pro-angiogenic factors such as basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), and platelet-derived growth factor (PDGF) either by the tumor or the surrounding host cells in response to tumor signals [8,9]. At the same time, suppression occurs of anti-angiogenic factors such as interferon alpha, thrombospondin, and endogenous protease inhibitors [5]. New vessel growth from existing vasculature has been believed to be the end result of this imbalance of factors; however, evidence is increasing to support the hypothesis that circulating endothelial progenitor cells (CEPs) derived from the bone marrow also play a significant role in tumor angiogenesis [10]. Regardless of the source, vascularization of tumor foci greatly increases the potential for metastasis.

Detachment

Cells must detach from the primary mass to start the journey to a distant location. To do this, down regulation occurs in cell-to-cell adhesion molecules such as E-cadherin as well as alterations in how the cell interacts with the extracellular matrix (ECM) [11]. Change in cell mobility is also required and is associated with altered cytoskeletal elements in response to tumor-derived factors acting in an autocrine fashion, as well as host-secreted factors. Tumor cells may also secrete proteases that facilitate detachment [5].

Invasion

Once a cell separates from the primary tumor and enters the ECM it must gain access to the vasculature. Degradative proteolytic enzymes are produced by the tumor or host support cells to facilitate this process. These include cathepsins, matrix metalloproteinases, and plasminogen activators. An increase in these enzymes has been associated with greater metastatic potential in several tumor types [12]. Tumor cells that bind to basement membrane components such as laminin or fibronectin also have a metastatic advantage.

Survival in the Circulation

At this point tumor cells must avoid host immune defenses and are susceptible to destruction by immune effector cells such as lymphocytes, monocytes, and natural killer cells. Other reasons for circulatory death include mechanical stresses associated with turbulent flow and high oxygen tension leading to toxicity. The host may also trigger apoptosis of tumor cells in the circulation through a variety of signals [11].

Arrest and Extravasation

Cells or aggregates of cells (emboli) must stop at a distant vascular bed to begin the extravasation process. This arrest can be facilitated by the mechanical size of the emboli compared with the capillary or, more likely, is mediated by tumor cell attachment to specific cell surface markers such as E-selectin or CD-44. Arrest can also occur when endothelial cells are damaged and the basement membrane is exposed [5]. Cells can then attach to basement membrane components such as laminin, type IV collagen, and other proteoglycans similar to the process that occurs during invasion into the vascular space. These interactions are facilitated by cell surface adhesion molecules such as integrins [11].

Organ Specificity

Clinical as well as experimental observations have shown that certain tumors tend to metastasize to certain locations preferentially. This is exemplified in Paget's "seed and soil" hypothesis, which was proposed to explain the apparent non-random pattern of tumor spread observed in visceral organs and bones [4]. This theory states that specific interactions between the metastasizing tumor cells and the target organ environment must occur and be favorable in order for tumor foci to survive and grow. Another theory of organ preference is based on hemodynamic factors such as the number of metastases that occur in an organ is related to the number of tumor cells delivered there by the flow of blood. In reality the theories are not mutually exclusive [11]. Rodent models demonstrating organ preference of metastatic cells are well documented in the literature. An example of this also exists in a clinical observation of ovarian cancer patients who have peritoneovenous shunts placed to palliate ascites build up. Ovarian cancer cells will readily grow in ascetic fluid and on peritoneal and organ surfaces. They rarely, however, spread beyond the abdominal cavity. When a shunt is placed, millions of tumor cells are dumped into the venous circulation, theoretically increasing the chance of hematogenous metastasis. Clinical observations at the time of patient death by Tarin and colleagues, however, did not show any increase in metastasis outside the abdominal cavity, supporting the need for the appropriate "soil" to ensure metastatic cell growth [4,13]. Our understanding of the tumor-host microenvironment has increased. Each organ likely has a unique set of growth factors, differentially expressed endothelial receptors and cell surface receptors that allow for unique cross talk with available metastasizing cells, encouraging or discouraging tumor growth [4]. One example of this crosstalk is the finding that chemokine receptors may play a role in organ specificity. It has been shown in vitro that breast cancer cells have a high expression of CXCR4 and CCR7 and that the ligands for these receptors are highly expressed in organs where breast cancer preferentially spreads, including the lung and liver. Additionally, blocking the specific receptor ligand interaction can decrease metastasis experimentally [14].

New Considerations

Disseminated Tumor Cells and Tumor Dormancy

The use of sensitive immunocytochemical stains and PCR strategies has allowed for the detection of disseminated tumor cells in sites such as the bone marrow where clinically evident macrometastases rarely occur. One example of this has been reported in 550 breast tumor patients where 30 to 40% of cells taken from the bone marrow had markers specific for epithelial tumors compared with 1% of the 200 non-cancer bearing individuals sampled. The presence of these cells also predicted a worse prognosis [15]. Additional studies have shown that 20 to 40% of patients with carcinomas from various primary sites have cells detectable in the bone marrow even with no evidence of lymph node or distant organ metastasis [3] The implications of the discovery of these cells may indicate that the bone marrow may be a reservoir for malignant cells either to remain dormant or to adapt prior to establishing metastatic foci in other organs.

Genetic and phenotypic profiling of these cells compared with the primary tumor has shown significant differences, supporting the theory that these cells change and develop unique characteristics after leaving the primary tumor, perhaps even early in tumor development [3]. This theory proposes that cells neither proliferate nor die, instead remain quiescent. These cells are probably resistant to traditional chemotherapeutics that typically target dividing cells [1]. The dormancy of cells can be reflected in the lag time that exists between the removal of a primary tumor and the development of detectable metastatic foci. Detection of disseminated cancer cells early in the course of disease might help guide therapy by predicting which patients actually need adjuvant therapy post-surgical excision of the primary tumor. For example 90% of breast cancer patients without lymph node metastasis will have adjuvant therapy recommended, whereas only 20 to 25% of those individuals will likely develop metastasis within the first 10 years, indicating that some patients may be overtreated [16].

Therapeutic considerations also need to be given to these disseminated cells especially when targeted therapies become routinely incorporated into cancer treatment. If the primary tumor expresses one target and the disseminated cells another, one subpopulation may go untreated.

Gene Expression Profiling and the Metastatic Phenotype

In longstanding models of metastasis it has been believed that, within any given tumor, a small subpopulation of cells develop and are genetically altered to acquire the advantages needed to become metastatic [2]. This implies that the ability to metastasize happens late in tumor progression. Several studies have challenged this, including ones in breast cancer using microarray platforms to profile gene expression. These studies have shown that the differences in gene expression of the primary tumor can predict whether the tumor will spread or remain localized. In one study by Van De Vijver et al., tumors from 295 early-stage breast cancer patients were screened using a 70-gene prognosis profile dividing the population into poor and good gene signature groups. Patients with a poor signature (n = 180) had a 10-year survival of 55% compared with 94% for those in the good signature group (n = 115). Additionally the probability of remaining metastasis free was 51% in the poor prognosis group compared with 85% in the good prognosis group. Even in multivariate analysis, the gene profile status remained a strong independent factor at predicting disease outcome [17]. The authors of this study concluded that gene profiling was potentially a more powerful predictive tool than standard clinical and histopathologic criteria. The genetic background of the host may also influence the efficiency of the metastatic process, with certain individuals having the genetic profile that favorably supports tumor spread. If these individuals can be reliably identified, investigation of cancer preventative measures might be warranted [18,19].

Conclusion

As more is learned about the process of metastasis, it is clear that many unanswered questions exist and that new technologies will help elucidate the intricacies of this complex process. From a therapeutic standpoint, it is also evident that not only must the metastatic tumor cell be targeted, but so must the host-produced factors that promote tumor cell growth, survival, invasion, and metastasis.