As the length of survival continues to improve for patients with systemic cancers, the problem of metastases to the chemotherapeutic sanctuary of the central nervous system (CNS) continues to grow. This review focuses on the pathobiology of brain metastasis, attempting to develop a framework for understanding the steps and molecular underpinnings of the metastatic cascade. In this process, cancer cells escape the primary tumor organ, intravasate into blood vessels, survive the hematogenous dissemination to the CNS, arrest in brain capillaries, extravasate, proliferate, and develop angiogenic abilities to succeed as an established metastasis. Each of the steps of the metastatic cascade is dependent on distinct molecular pathways, the identification of which may be exploited in attempting to halt or slow the development of brain metastases.
The issue of brain metastasis is an increasing problem for oncologists and neurosurgeons. Although improved cancer therapies have increased the survival rates of many cancers, brain metastases continue to have poor overall survival rates. Thus, there continues to be a need for the discovery of new and different targets to prevent a tumor from escaping from its site of origin, migrating to the brain, and continuing to proliferate at distant sites. This process is commonly referred to as the metastatic cascade and is highly dependent on a multitude of cellular processes, each of which must be present and functioning in order for successful metastasis ( Appendix 1 ).
The incidence of brain metastasis is unknown, but estimates range from 140,000 to 170,000 new cases per year. It is well known that certain cancers preferentially metastasize to the brain, specifically lung (50%–60%), breast (15%–20%), melanoma (5%–10%), and renal and colon cancers (4%–6%). Twenty to 30% of all patients who have breast cancer eventually develop central nervous system (CNS) metastasis. The treatment of metastasis to the brain is further complicated by the unique characteristics of the brain. The blood-brain barrier (BBB), with its tight junctions and lack of lymphatic drainage, makes the delivery of chemotherapeutic agents difficult and represents a therapeutic haven from chemotherapy. In addition, the microenvironment of the brain parenchyma is unique. The interstitial fluid of the brain is high in chloride, which may make the parenchyma a hostile environment to many metastatic clones and favor clones of neuroepithelial origin, such as small-cell carcinoma of the lung or melanoma.
This article reviews the known pathobiologic components of the ability of primary tumors to metastasize at each stage of the metastatic cascade. The promoters, contributors, and inhibitors of the cascade are numerous and may become overwhelming without an organizational structure within which to arrange the alphabet soup of chemicals and genes that are involved in this process. By examining the genetic and pathobiologic characteristics of the process in this matter, we hope to create a more structured platform for understanding this complex process that allows the clinician to create a framework of current therapeutics and possible future research.
Escape/intravasation
For a primary tumor to metastasize to a secondary site, it must break free from the primary tumor and enter the vasculature. Although many of the molecular processes discussed later are active at both the escape and extravasation/proliferation stages, they are first encountered within the cascade early on and so are addressed in this section.
The E-cadherin-catenin complex is vital for the maintenance of both normal and tumoral cytoarchitecture as well as a necessary mediator of cell-cell adhesion. In the metastatic escape of a tumor, clone cells have reduced intercellular adhesion and disordered cytoarchitecture, and are thus prone to separation from the primary tumor mass. These clones are then free to invade both locally as well as to continue on to intravasation and further progress in the cascade. Decreased expression of the E-cadherin-catenin complex has been correlated with invasion, metastasis, and unfavorable prognosis. In addition, Shabani and colleagues established a correlation between E-cadherin-catenin complex expression and an increased mindbomb homolog 1 (MIB1) index in metastatic adenocarcinoma.
Another family of adhesion and signaling receptor proteins are the integrins. They mediate both cell migration and tumor invasion via the triggering of multiple signal transduction pathways. They are therefore vital in the complex cascade of regulation of such processes as gene expression, growth control, cytoskeletal architecture, and apoptosis. In an animal model of human nonsmall-cell lung cancer (NSCLC), blocking of the α 3 β 1 -integrin significantly decreased brain metastasis. Researchers at Oxford showed that the blockage of the β 1 -integrin subunit prevented tumor cell adhesion to the vascular basement membrane (VBM) and attenuated metastasis establishment and growth in vivo. Furthermore, focal adhesion kinase (FAK) is known to be a key mediator in integrin signaling and therefore is believed to play a role in metastatic migration and proliferation. Dephosphorylation and therefore the inhibition of FAK at the Y397 locus via activated oncogene (rat sarcoma) has been shown to promote tumor migration via the facilitation of focal adhesion turnover at the leading edge of tumor cells.
Another aspect of the ability of a tumor cell to escape the local site is its ability to break down or functionally remodel the extracellular matrix (ECM). Degradation of the ECM via proteolytic enzymes is believed to clear a pathway for invasion. This proteolytic activity has been located on the cell membrane at the advancing edge of invading tumor cells. The ECM proteolysis may also release factors that promote cell proliferation and angiogenesis for contribution to later steps in the cascade. Neurotrophins (NTs) are known to promote brain invasion via enhancing the production of the ECM proteolytic enzyme heparinase. Heparinase is an endo-β- d -glucuronidase that cleaves the heparin sulfate chains of the ECM. It is the dominant mammalian heparin sulfate degradative enzyme and is known to destroy both the ECM and the BBB. NTs have been found at the tumor-brain interface of melanoma, and there are reports of the p75 NT receptor functioning as a molecular determinant for brain metastasis.
Next on the list of molecular degraders of the ECM are the plasminogen activators and their inhibitors. Plasmin is a tumor-associated serine protease activated by urokinase-type plasminogen activator (uPA). The production and release of uPA has been well documented in human cancers. The uPA binds to the receptor uPA-R (CD87), the activity of which is regulated by the action of plasminogen activator inhibitor type 1 and 2 (PAI-1/2) on the cell membrane and causes urokinase to convert plasminogen to plasmin. The proteolytic activity of plasmin then degrades components of the ECM including fibrin, fibronectin, proteoglycans, and laminin. Further, plasmin activates other proteolytic enzymes, with resultant local invasion and migration. As far back as 1994, researchers have found that there are high levels of uPA in metastatic tumors, that uPA correlates with necrosis and edema, and that there is an inverse correlation with levels of uPA of a tumor and survival. In addition, high levels of uPA and absent tissue plasminogen activator correlate with aggressiveness and decreased survival.
The matrix metalloproteases (MMPs) are a family of 20 proteolytic enzymes that have also been well established as functioning to degrade the ECM in metastasis. Their expression is regulated via cytokines, and the ECM metalloprotease inducer is found on the surface of tumor cells. With induction and stimulation, there is an ECM breakdown and a tumor cell migration. MMP activity is known to correlate with invasiveness, metastasis, and poor prognosis. One study found that MMP-2 is present in all metastatic brain tumors tested regardless of the site of origin and that the level of activity inversely correlated with survival. However, although MMP-9 was found by Arnold and colleagues to be upregulated in all brain metastases and primary brain tumors, there was an inability to correlate upregulation with survival. The tissue inhibitor of metalloprotease 1 (TIMP-1) overexpression in a murine model was shown to reduce the incidence of brain metastasis by 75% compared with wild type, showing that inhibitors of MMPs suppress brain metastasis.
The properties of the tumor cell membranes may contribute to the local invasiveness and migratory capability of tumor clones. In a study on brain-specific breast cancer metastasis, Khaitan and colleagues showed that the increased expression of KCNMA1, the gene that encodes for the pore-forming α-subunit of the large-conductance calcium and voltage-activated potassium channel big-conductance type potassium channel that is known to be upregulated in breast cancer, has led to greater invasiveness and transendothelial migration. Furthermore, there has been increased interest in the scientific role of the family of membrane proteins known as aquaporins. Among their many functional roles, aquaporins are known to facilitate tumor migration, as seen in aquaporin-dependent tumor angiogenesis and metastasis via a mechanism of facilitated water transport in the lamellipodia of migrating cells.
Several known tumor-suppressor genes that function at the level of escape and migration/intravasation are worth exploring. The best known of these is the KISS-1 gene on chromosome 1. KISS-1 encodes metastin, which is a ligand of the orphan G protein couples receptor hOT7T175. Lee and colleagues found that the forced expression of KISS-1 suppressed both melanoma and breast metastasis. Other investigators have found an inverse correlation between KISS-1 expression and melanoma progression.
KAI1 (CD82) is another tumor-suppressor gene on chromosome 11p11.2. KAI1 functions to regulate adhesion, migration, growth, and differentiation of tumor cell lines. It has been found to have an inverse correlation with prostate progression as well as breast and melanoma metastasis. In addition, KAI1 is known to be associated with the epidermal growth factor receptor (EGFR), discussed later in this article, and is believed to affect the Rho GTPase pathway, resulting in suppression of lamellipodia formation and migration.
The tumor suppressor, Drg-1 methylated inhibition, has been found to inhibit both liver metastasis and colorectal carcinoma invasion. Overexpression of this gene has been linked to resistance to ironectan chemotherapy. In a murine model of breast cancer metastasis, the Notch signaling pathway was found to be activated via increased Jag2 mRNA, creating a cell line that was both more migratory and more invasive in collagen assays. In addition, inactivation of the Notch pathway significantly decreased the migratory and invasive activity of the studied cell lines.
Arrest/extravasation
The next series of steps in the metastatic cascade involve a complex set of interactions that allow the tumor clones that have invaded the blood stream to arrest at a secondary site and extravasate from the vasculature to establish a metastasis in a new organ. The clones must then survive and proliferate at the secondary site. Although the exact causes of arrest and proliferation at specific sites have not been completely elucidated, one theory is that there are direct neurotropic interactions between tumor clones and the brain along with yet undiscovered brain-specific homing capacity within the tumor cells that result in brain metastasis. Carbonell and colleagues described a process termed vascular cooption, whereby 95% of micrometastasis are observed to grow along the exterior of preexisting vessels before any overt metastatic tumor is detected. The VBM tumor cell interaction is adhesive in nature. This interaction implied that the VBM is the soil for brain metastasis rather than a previously theorized neurotropism. With the VBM as a substrate, tumor cells are able to infiltrate the brain parenchyma. Saito and colleagues showed that the pia-glial membrane, present along the external surface of blood vessels, serves as a scaffold for metastatic tumor cells spreading in an angiocentric pattern, which furthers the hypothesis of a perivascular soil.
A biologic model for metastatic tumor cells describes how they function like macrophages both within the vasculature and during extravasation in a mouse model of CNS metastasis. In this model, the tumor cells expressed multiple properties of macrophages that included morphologic appearance, surface adhesion, phagocytosis, total lipid composition, and expression of CD11b, Iba1, F4/80, CD68, CD45, and CXCR (all genes specifically expressed by macrophages). It is possible that by expressing these molecules, the metastases can mimic macrophages and escape the immune system while traveling through the vascular system.
The exact mechanisms by which cancer cells pass through the BBB is unknown. However, recently, 3 genes that mediate brain-specific breast metastasis have been described. Cyclooxygenase 2 (COX-2, also known as PTGS-2) as well as the EGFR ligand, heparin-binding epidermal growth factor have been linked to metastasis to the lung as well as to the brain. In addition, these genes function to assist extravasation through nonfenestrated capillaries and enhance colonization. The α 2,6 -sialyltransferase ST6GALNAC5 is normally restricted to the brain and when expressed by breast cancer cells, enhances their adhesion to brain endothelium and their passage through the BBB via cell surface glycosylation.
The chemokine/receptor system, CXCL12/CXCR4, and the recently discovered alternate receptor, CXCR7, function in the homing of neoplastic cells from the primary site to the target site in metastatic disease. In a study of 56 patients with metastatic lesions to the brain from differing primary sites, Salmaggi and colleagues found that CXCL12 was expressed in tumor cells and tumor vessels, and this expression correlated with shorter survival. In addition, the CXCR7 was expressed by tumor cells as well as by the adjacent brain, and the CXCR4 was present in all samples with a nuclear pattern. However, the expression of these receptors did not correlate with survival. Thus, the expression of CXCL12 may indicate aggressiveness of brain-specific metastasis. Another recently described mediator of organ-specific breast cancer metastasis is the expression of heat-shock protein (HSP27). HSP27 is a chaperone of the small heat-shock protein (sHSP) family. Researchers have been able to associate the expression of HSP27 in brain-specific breast cancer metastatic cell lines with the 36/67-laminin receptor. HSP27 created clusters of chaperone and cochaperone proteins that facilitate brain-specific metastasis. In addition, HSP27 associated these chaperone clusters through kinases to a group of filament proteins that may assist in organ-specific homing.
The wingless integration gene (WNT) and T-cell factor (TCF) pathway, WNT/TCF pathway, and its target genes, homeobox B9 and lymphoid-enhancing factor 1, are mediators of brain-specific chemotactic invasion and colony outgrowth in lung adenocarcinoma. Hyperactivity of this pathway is present in metastatic subpopulations of adenocarcinoma cells. Decreases in the activity of TCF attenuates the ability of the cells to form brain and bone metastasis, which indicates their contribution to brain-specific metastatic of lung adenocarcinoma lesions.
Adding to the list of brain-specific contributors to metastasis, Zhang and colleagues described another brain-specific molecular determinant for metastasis of melanoma in a murine model. The investigators found that the transforming growth factor β2 (TGFβ2) was highly expressed in brain-specific murine melanoma cell lines. After transfection of the TGFβ2 gene into another cell line, an increase was noted in the production of microscopic metastatic lesions to brain parenchyma.
Adhesion of neoplastic cells to the endothelium of metastatic sites via a hyaluronate matrix ligand is mediated by CD44 on chromosome 11p11.2. CD44 encodes a membrane glycoprotein that acts as a receptor for hyaluronic acid and osteopontin. CD44 can be downregulated via DNA methylation, and such downregulation has been correlated with increased tumor grade. In addition, upregulation occurs in 48% of brain metastases studied, especially thyroid, melanoma, and breast. Primary brain tumors express the standard form of CD44, whereas metastatic lesions almost exclusively express the only splicing variant of the gene product, providing clinicians with a possible tumor marker for metastatic potential.
Invasion of brain parenchyma is mediated by the tumor-suppressor gene phosphate and tensin homolog deleted on chromosome 10 (PTEN) or mutated in multiple advance cancers (MMAC1). The PTEN/MMAC gene product and the cytoskeletal protein tensin are similar and interact with actin filaments at focal cell adhesions that inhibit cell migration in the functioning gene, whereas, in an antisense mutation, migration was enhanced. In lung cancer metastasis, 25% of the genes had an inactivating mutation, suggesting that migration and metastatic progression are inhibited by the normally functioning gene.
Angiogenesis and Metastatic Clone Proliferation
Tumor angiogenesis is an important aspect of the ability of a neoplastic population to survive and grow at a secondary site. Failure of vascular growth ultimately restricts the tumor mass to 0.2 mm or to the limits of tissue diffusion distance. There seems to be a balanced interplay of proangiogenic and antiangiogenic factors. Much research has been devoted in recent years to the elucidation of these angiogenic factors as a target for tumor treatment.
The most commonly recognized of these neoplastic angiogenic factors is vascular endothelial growth factor (VEGF). Kim and colleagues found that VEGF expression plays a role in the ability of breast cancer cells to metastasize, and that the inhibition of VEGF, via a receptor tyrosine kinase inhibitor, reduces tumor-induced angiogenesis and restricted tumor growth. SSecks (Src-suppressed C kinase substrate) is known to decrease the expression of VEGF via reduction of angiogenesis. It also stimulates the expression of the proangiogenic molecule angiopoietin 1 and may regulate the brain angiogenesis and tight junction formation, therefore regulating BBB differentiation and contributing to angiogenesis.
Yet another angiogenic regulator is a member of the previously mentioned MMP family, the MMP-9/gelatinaseB complex, that may contribute to the switch from vascular quiescence to angiogenesis. PAI-1, the uPA cell-surface receptor mentioned earlier, is often localized to the proliferating vessels in brain metastasis and therefore may also play a role in angiogenesis. Plexin D1 expression in tumor versus nonneoplastic vasculature was explored to determine if plexin D1 is unique to tumor cells and vasculature, thus participating in tumor angiogenesis. Plexin D1 was found to be expressed in neoplastic cells as well as tumor vasculature, whereas its expression in nonneoplastic tissue was restricted to a small subset of activated macrophages, suggesting that plexin D1 may play a significant role in tumor angiogenesis.
A significant contributor to the secondary-site tumor for growth potential in breast cancer is overexpression of hexokinase 2 (HK2), which plays a key role in glucose metabolism and apoptosis. Researchers at the National Cancer Institute found that both mRNA and protein levels of HK2 were increased in brain metastatic derivative cell lines compared with the parental cell line in vitro. In addition, these investigators found that the knockdown of expression reduced cell proliferation, implying that the gene contributes to the proliferation and growth of breast cancer metastasis. They showed that increased expression was associated with poor survival after craniotomy.
At least 2 tumor-suppressor genes that function at the proliferation level of the metastatic cascade have been described. The first gene, NM23, regulates cell growth by encoding for a nucleotide diphosphate protein kinase that interacts with menin, which is a putative tumor suppressor encoded by the gene MEN1. NM23 is believed to reduce signal transduction and thereby decrease anchorage-independent colonization, invasion, and motility. In melanoma, decreased expression is correlated with increased brain metastasis. The second described tumor-suppressor gene, BrMS1, localizes to chromosome 11 in melanoma and breast cancers. BrMS1 prevents disseminated tumor cell growth by restoring the normal gap junction phenotype and maintaining cell-to-cell communication in the primary tumor. Seraj and colleagues found an inverse correlation between the expression of BrMS1 and the metastatic potential in melanoma.
Cascade Nonspecific Metastatic Contributors
Certain molecular contributions cannot be attributed to a specific step in the cascade either because they are active at every level, or, as in most cases, their true function is yet to be discovered. These molecular entities are on the forefront of cancer science and are worth addressing. Zeb-1, the zinc finger E-box homeobox transcription factor, is overexpressed in metastatic cancers. This overexpression leads to epithelial-mesenchymal transition and increased metastasis. The mutation of this gene has been shown to decrease proliferation of progenitor cells in mutant mice. This mutation may be a target for metastatic prevention at the progenitor level.
Several other genetic markers have been located that pertain to metastasis in particular. A deletion of the 4q arm in lung (both small- and nonsmall-cell) metastasis to the brain and bone has been documented. In addition, in NSCLC, the overexpression of 3 genes, CDH2 (N-cadherin),KIFC1, and FALZ, was highly predictive of metastasis to the brain in early and advanced lung cancer. Therefore, these genes may be used to predict a high risk of metastasis early in the diagnosis. In prostate cancer, increased expression of KLF6-SV1, the Kruppel-like factor tumor-suppressor gene, predicted poorer survival and correlated with increased metastasis to lymph, brain, and bone. Overexpression of homeoprotein Six-1, a transcriptional regulator, increased TGFβ signaling, and metastasis in breast cancer, with significantly shortened relapse times. These genes are important in the understanding of the metastatic cascade as well as for further research.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
