Shravan K. Chintala, Ph.D. and Jasti S. Rao, Ph.D.

Department of Neurosurgery, Box 064, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA.

Received 9/28/96; Accepted 10/17/96; On-line 11/01/96

TABLE OF CONTENTS

1. Abstract

2. Introduction

3. Extracellular matrix components

3.1. Proteoglycans

3.2. Myelin

3.3. Fibronectin

3.4. Vitronectin

3.5. Tenascins

3.6. Laminins

3.7. Type IV collagen

4. Gliomas

4.1. Invasion

5. ECM receptors and cell adhesion molecules

5.1. Gangliosides

5.2. Cluster differentiation 44

5.3. Neural cell adhesion molecule

5.4. Integrins

6. ECM turnover and degradation

6.1. Matrix metalloproteases

6.2. Plasminogen activators

7. Perspectives

8. Acknowledgments

9. References

1. ABSTRACT

The invasion of glioma into normal brain tissue is a major challenge to clinical intervention because these tumors often highly infiltrate the surrounding brain tissue. Total surgical resection of gliomas is impossible, and recurrence of tumor growth is a common phenomenon; patients have a mean survival time of 8 - 12 months. Although in recent years substantial progress has been made toward understanding the invasive behavior of gliomas in vitro and in vivo, the factors responsible for the extensive infiltration are still poorly documented. This review focuses on recent research concerning the invasion of gliomas, as well as the extracellular matrix components, and the proteolytic enzymes involved.

A better understanding of cell-matrix interactions will help in developing therapeutic strategies to decrease the invasion of gliomas.

2. INTRODUCTION

The extracellular matrix (ECM) is a dynamic milieu that plays a pivotal role in regulating cellular functions during normal and pathological remodeling processes such as embryonic development, tissue repair, inflammation, tumor invasion, and metastasis. Although the ECM contains mainly collagens and noncollagenous glycoproteins such as glycosaminoglycans and proteoglycans, its composition is probably unique for each cell type within an organ. In the nervous system, in contrast to other organ systems, the structure and function of the ECM have not yet been completely determined. The brain is largely free of a well-defined ECM, except where mesodermally derived endothelial cells invade the CNS to establish vasculature. The parenchyma of the CNS, however, appear to be filled with a relatively amorphous matrix that contains mainly hyaluronic acid (1) and little collagen and other fibrous proteins.

A well-defined ECM exists in the form of a true basement membrane, cerebral vasculature, and the glial limitans externa. The latter is a basement membrane that covers the brain's entire cortical surface and also separates astrocyte foot processes from pial cells and the subarachnoid space (2). The cerebral vascular basement membrane, which surrounds the blood vessels of the brain, contains type I, III, and IV collagens, fibronectin, laminin, and heparan sulfate proteoglycans (Table 1). Recent in vitro studies have shown that both immature astrocytes and normal leptomeningial cells can synthesize basement membrane molecules. For this review we limit our focus to the ECM components of the human brain that have been proposed as potential candidates for mediating glioma cell invasion.

3. EXTRACELLULAR MATRIX COMPONENTS

3.1. Proteoglycans

Proteoglycans and glycosaminoglycans (GAG) are abundantly present in the brain parenchyma. Proteoglycans contain a core protein with one or more covalently bound glycosaminoglycan side chains. The major proteoglycans in the CNS are chondroitin sulfate (CS) and heparan sulfate. CS is a polymer consisting of alternating units of N-acetylgalactosamine and glucuronic acid. In the mature brain, CS, is located in the cytoplasm of some neurons (5-7) and astrocytes, and in myelinated and unmyelinated axon fibers, but not in oligodendrocytes and myelin (8). In contrast, CS proteoglycans are the major proteoglycans of the white and gray matters of the brain, located predominantly in the extracellular spaces of granular and molecular layers in the immature cerebellum, and they are believed to be associated with cell differentiation and migration in the central nervous system (9). Both heparan sulfate and chondroitin sulfate are present in the basement membrane (10). Heparan sulfate is found as membrane protein in synaptic vesicles and in the ECM of the neuromuscular junction (11, 12), and it is also present in the basement membranes of the Schwann's cells (13). Heparan sulfate proteoglycans have been found to induce cell motility (14).

Hyaluronic acid (HA or hyaluronan), a high molecular-weight proteoglycan found in the extracellular matrix, is the only proteoglycan that does not contain a core protein. A polysaccharide comprised of repeating disaccharide units of D-glucuronic acid and N-acetyl glucosamine, HA is found in most extracellular matrices and at the cell surface. Brain parenchyma contain several other proteoglycans, such as tenascin and cytotactin, in addition to glial hyaluronate binding protein (GHAP or hyaluronectin).

Glycosaminoglycans are found mainly within tumor tissue (15); their hyaluronic acid content has been shown to increase transiently during tumor cell migration and is usually found at the interface between tumor mass and host tissues (16). Apart from its major role in tumor cell invasion, HA has been implicated in many cell functions including neural crest migration (17). Elevated levels of HA have been correlated with tumor cell invasiveness (18). HA interacts with other extracellular matrix proteins via hyaluronan binding proteins and receptors such as cluster differentiation 44 (CD44). Recent studies have demonstrated that suppression of CD44 expression decreases migration and invasion of human glioma cells (19). Further, intracranial injection of CD44-suppressed cells led to relatively localized tumor growth and with a reduced malignant behavior, whereas control cells showed extensive invasion typical of highly malignant gliomas (17). Koochekpour et al. (20) demonstrated that HA-mediated cell detachment involved its high-affinity receptors, CD44. They also showed that HA induces cell detachment, stimulates migration, and promotes invasion through interaction with CD44. In 1994, Merzak et al. (21) showed for the first time that CD44 is involved in glioma cell invasion. Thiery et al. (22) had reported that antisense CD44 oligonucleotides inhibited the invasion of glioma cells in vitro, and that several components of basement membrane molecules, such as fibronectin, laminin, vitronectin, and collagen I, might be involved in CD44-mediated invasion of human glioma cells. The proposal is that once synthesized, HA becomes hydrated, opening up the extracellular space to invading tumor cells. This process is suggested to be facilitated through CD44-HA interaction. The role of HA in glioma cell invasion is complicated however, by the fact that astrocytes synthesize HA themselves (23), rather than tumor cells stimulating host fibroblasts to make HA (24). Our own results showed that invasion of glioblastoma cells was significantly increased in the presence of HA (25).

3.2. Myelin

Gliomas also disseminate along the myelinated fiber tracts of white matter, which results in distant spread of tumor cells through the corpus callosum into the contralateral hemisphere (26-28). Caroni and Schwab (29) showed that C6 rat glioma cells attach to and spread on crude extracts of CNS myelin, and that the spreading of C6 glioma cells on myelin extracts depends on the expression of a membrane-bound metalloendoprotease (30). More recently, Giese et al. (31) found that established glioma cell lines as well as primary cells isolated from glioblastoma biopsy specimens, attach and migrate on crude myelin extracts. Although these findings indicated that glioma cells migrate and spread along existing anatomical structures such as myelin, the responsible component/molecule of the crude myelin extract has not been identified. As Giese et al. (31) suggested, the three-dimensional conformation of the associated proteins could be changed during the migrating process, and this in turn could influence the behavior of the glioma cells. It seems, however, that neither neural cell adhesion molecules nor integrins mediate the adhesion of glioma cells to crude myelin extracts.

3.3. Fibronectin

Fibronectin is an M_r 500,000 glycoprotein found in most extracellular matrices as aggregates or fibrils. Fibronectin consists of two polypeptide chains of approximately M_r 250,000 linked by interchain disulfide bonds. Fibronectin has three homologous repeats type I (about 45 amino acids long with two disulfide bonds), type II (about 60 residues with two disulfide bonds), and a type III (about 90 residues with no disulfide bonds). The first major adhesive protein identified, fibronectin has many biological functions involving cell adhesion, migration, and invasion. Fibronectin mediates a variety of adhesive events by binding to fibrinogen/fibrin, collagen, heparan sulfate, and hyaluronic acid. Fibronectin is found at the gliomesenchymal junction of tumors and in tumor-associated blood vessels, and it is expressed by glioblastoma cell lines in vitro (32, 33). An arg-gly-asp (RGD) sequence in the third fibronectin repeat functions as an integrin receptor for most cells (34). Kochi et al. (35) showed that astrocytomas and glioblastomas do not express fibronectin and that fibronectin was confined to proliferating vessel walls and leptomeninges.

3.4. Vitronectin

Vitronectin, originally known as S-protein and primarily found in the liver, is a multicellular serum protein that promotes cell adhesion, spreading, and migration of a variety of cell types (36). Vitronectin is found mainly in serum, skin, and wound tissue (37, 38). Vitronectin is absent in normal brain and early-stage glioblastoma but late-stage glioblastomas have recently been shown to express vitronectin (39). Vitronectin supports cell adhesion through integrins. Apart from modulating plasminogen (40) and plasminogen activator inhibitor (PAI-I) (40-43) vitronectin promotes cell adhesion mediated by integrins alphavß3, alphavß5, alphaIIß5 and alphavß1. To date, alphavß3 integrin appears to be specific for vitronectin, by the finding that low to high metastatic potential of human melanoma was correlated with an increased expression of vitronectin receptor alphavß3 (44). Gladson and Cheresh (39) recently showed that vitronectin may play a role during local invasion of glioblastoma. A recent study showed that, compared to fibronectin, laminin, and collagen IV, vitronectin was a poor adhesive and migratory protein for glioblastoma cells U-251 and SF-767 (45).

3.5. Tenascins

Tenascins are large disulfide-linked heterodimeric extracellular glycoproteins. Tenascin-C, the original member of a family of heterodimeric extracellular matrix proteins, is implicated in adhesion and migration of human glioma cells (46). However, the recent discovery that tenascin knock-out mice develop normally (47) and the tenascin's contradictory effects on adhesion and migration (48-50) raised a controversy concerning their basic role. To date, three members of the tenascin family have been identified, tenascin-C, tenascin-R, and tenascin-X. All known forms of tenascins consist of heptad repeats, epidermal growth factor (EGF)-like repeats, fibronectin type III repeats and a fibrinogen domain. Tenascin-C (also known as myotendinous antigen, glioma mesenchymal extracellular matrix, hexabrachion, J1-200/220, and cytotactin) is mainly found during normal embyrogenesis; it is prominent in the developing central nervous system and in developing connective tissues, and it is overexpressed in tumors (51, 52). Tenascin-C is expressed in malignant breast carcinomas (53), in wound healing (54, 55) and during newt limb regeneration (56). Tenascin-C is upregulated in most types of carcinomas including melanomas and gliomas (57-59). In the central nervous system, tenascin-C is found to be synthesized by glial and neural crest cells (60-62) and by satellite cells of the peripheral nervous system (63). Tenascin-R (also known as J1-160/180, januscin, and restrictin), so far identified in rat and chicken, seems to be specific to the central and peripheral nervous system (64-65). Tenascin-X mainly expressed in skeletal and heart muscles, was originally reported as a partial sequence encoded by gene X (66), but little is known about its function. Although tenascin was categorized as an antiadhesive protein (67-69), recent data show that tenascin supports cell adhesion involving the RGD sequence in the third fibronectin type III repeat that it seems to be mediated by V3 integrin (70, 71). A recent study showed that tenascin enhances U251.3 glioma cell migration by an RGD independent mechanism (46).

3.6. Laminins

Laminins are a large group of adhesion glycoproteins found in all basement membranes and in hyperplastic blood vessels in gliomas, gliosarcomas, and menigiomas as an integral part of the glial limitans externa (72,74). Laminin is a structural glycoprotein found predominantly in basement membranes (75-77). It plays a role in migration, neurite outgrowth, proliferation, and differentiation (78, 79). The first identified laminin was purified from Engelbreth-Holm-Swarm sarcoma, and recent literature identifies at least six other forms. All the seven known laminins are composed of three subunits designated as alpha, ß, and gamma and association of disulfide-linked subunits give rise to different laminins (80). Laminin interacts with a variety of basement membrane components such as entactin/nidogen, type IV collagen, and heparan sulfate (80). The other laminin isoforms known are laminin-2 (merosin), laminin-3 (S-laminin), laminin-4 (S-merosin), laminin-5 (kalinin/nicein/epiligrin), laminin-6 (K-laminin), and laminin-7 (Ks-laminin) (78, 80).

3.7. Type IV collagen

Interstitial collagens are located in the leptomeninges and the fibromuscular layer of large blood vessels in the brain. Type IV collagen, mainly present in capillaries and large blood vessels, is the principal collagenous constituent of most basement membranes. The most common molecular form of type IV collagen is a heteropolymeric molecule, [a1(IV)]2 a2 (IV); other types of homopolymeric forms, [a1(IV)]3 and [a2(IV)]3, occur as well, and three additional type IV collagen chains are known to exist in kidney (81). Type IV collagen is secreted and assembled as a procollagen molecule, in which, each chain has an apparent molecular weight of 160,000-180,000. In rotary shadowed electron microscope preparations, type IV collagen appears as a rod with a knob like non-helical domain at the COOH- terminal end. Two to four monomers can be associated to form network like appearance in most basement membranes. Recent studies have shown that the glioblastoma cells are also capable of synthesizing type IV collagen in vitro (33). In an immunohistochemical study, Bjerkvig et al., (82) showed that type IV collagen was strongly expressed in tumor spheroids from rat glioma cell line BT4C but was negative in monolayers, and fibronectin was strongly expressed in BT4C and BT4Cn cell lines. In an immunofluorescence study, Bellon et al., (83) demonstrated that type IV collagen was localized to the subendothelial basement membrane of blood vessels in gliomas. Similar results were reported earlier by Rutka et al., (84).

4. GLIOMAS

Gliomas, the most common primary brain tumors, account for more than 40% of all CNS neoplasms. Human astrocytic brain tumors have been divided into pilocytic (juvenile) astrocytoma, low-grade astrocytoma, anaplastic astrocytoma, and glioblastoma. Astrocytomas are defined as tumors comprised predominantly neoplastic astrocytes. Pilocytic astrocytoma (World Health Organization [WHO] grade I) is the most common brain tumor in children. These tumors are typically located in midline structures, e.g., the optic nerves, third ventricle, thalamus, medial temporal lobe, brain stem, and cerebellum. Pilocytic astrocytoma very rarely progresses to anaplasia. Low-grade astrocytoma (WHO grade II) shows a consistent tendency to diffusely infiltrate the surrounding brain parenchyma. These neoplasms typically occur in young adults; they may be localized in any CNS region, including the spinal cord, but they tend to localize in cerebral hemispheres. Anaplastic astrocytoma (WHO grade III) is characterized by neoplastic fibrillary or gemistocytic astrocytes. This astrocytoma has an inherent and often rapid tendency to progress to glioblastoma. Glioblastoma multiforme (WHO grade IV) is the most frequently occurring and most malignant brain tumor that typically affects adults. The most common location of glioblastoma is the frontotemporal region, but the parietal lobes are also often affected. Glioblastomas usually infiltrate through the corpus callosum, with extensions into the contralateral hemisphere.

4.1. Invasion

Human gliomas are characterized by highly diffuse infiltrative growth into the surrounding normal brain tissue, which makes surgical resection difficult (28, 85). In contrast to other tumor types, gliomas rarely metastasize outside the central nervous system (28). In an experimental model system using C6 glioma cells, laminin, was shown to be the principal constituent of the basement membrane of the blood vessels (86, 87), and it appeared to be responsible, in part, for the exclusion of glioblastoma cells from blood vessels (88). The role of laminin is, however, still controversial. During invasion, the tumor cells break up their cell-cell and cell-matrix interactions, become motile and subsequently forge a path through the digested ECM. This is accomplished by changes in the expression of cytoskeletal proteins, cell adhesion molecules and matrix-degrading proteases. Whether the tumor cells invade in response to existing ECM components or themselves synthesize autologous ECM and then migrate to the newly laid matrix is still a debated issue. Tumor cell migration is also influenced by migrating signals, such as ECM components, either by chemo-and/or haptotactic mechanisms (89). Liotta (90) proposed a three-step hypothesis for tumor cell migration, suggesting that cellular invasion is the result of highly coordinated and complex mechanism that include a series of cell-matrix interactions: (1) modification of cell-cell and cell-matrix attachments, (2) proteolytic modification of the ECM, and (3) invasion through proteolytically modified ECM.

These events must be coordinated and integrated so that the leading edge of the invasive cells makes new matrix contacts while the trailing edge breaks previously formed ones. ECM proteolysis and cell migration are not mutually independent events. Cell attachment may influence protease production, and protease activity can alter cell attachment and spreading. Understanding how these events are coordinated will bring a better appreciation of the underlying mechanisms of cell invasion and help to identify new targets for therapeutic interventions. In most cases, degradation of the basement membrane results in the solubilization of ECM fragments that may have chemotactic effects.

In many pathological conditions, ECM remodeling is accompanied by a cellular invasion that can compromise matrix organization and disrupt tissue. Under normal conditions, a negative feedback mechanism may limit the behavior of normal cells, such as endothelial cells, and lymphocytes. Although tumor cells use the same mechanism as normal cells, they seem to lack the appropriate feed back mechanism during invasion (91).

5. ECM RECEPTORS AND CELL ADHESION MOLECULES

As they invade the tissue, tumor cells must adhere to a variety of ECM components and cell surface molecules. These interactions have a profound effect on the tumor cells invasive phenotypic behavior. Over the last several years, extensive progress has been made in identifying ECM components and particular sites on the components that mediate cellular adhesion / invasion. Tumor cells interact with the extracellular matrix by adhesion molecules or matrix receptors. Several classes of matrix receptors have been described, including integrins, gangliosides, CD44, and some members of the immunoglobulin superfamily such as the neural cell adhesion molecule (NCAM) (92).

5.1. Gangliosides

Gangliosides are a class of sialic acid-containing glycosphingolipids that are selectively localized in the cell membrane (93, 94). Although their functional role is not well documented, the observation that cells that had high migratory capacity concomitantly express more gangliosides gave rise to the speculation that gangliosides may be involved in migratory and invasive behavior of glial tumor cells (22). In this context, Merzak et al. (95) showed for the first time that cell-membrane gangliosides are involved in glioma cell invasion in vitro. Moreover, the same group showed that exogenous addition of gangliosides enhanced the adhesion of human glioma cell lines to fibronectin, laminin, vitronectin and collagen I (96), thus consolidating the role of gangliosides in adhesion. The identification of a ganglioside binding site on fibronectin supported the role of gangliosides in fibronectin-mediated cell adhesion (97). Moreover, expression of GD3 ganglioside was found elevated in malignant astrocytomas (98) and gangliosides were shown to regulate the function of integrins (99-101).

5.2. Cluster differentiation 44

Originally known as the lymphocyte homing receptor (also known as Hermes antigen, Pgp-1), cluster differentiation 44 (CD44) is a polymorphic family of glycoproteins that carries N- and O-linked sugars and glycosaminoglycan side chains found to be the major receptor for hyaluronic acid (HA) with a wide cellular distribution (102, 103). CD44 has two isoforms with a molecular mass of 80-90 kDa and 150 kDa respectively (104). CD44 is expressed on both normal and neoplastic astrocytes (105), and it has been shown to be upregulated in neoplastic cells (106). Recent studies have shown that CD44 mediates attachment of cells to various ECM proteins including HA, chondroitin sulfate, laminin, fibronectin, vitronectin, types I and IV collagen and modified basement membrane matrigel (107). Koochekpour et al. (19) recently reported that antibodies against CD44 inhibited the invasion of human glioma cells and they have showed that the factor responsible for the invasion is hyaluronic acid.

5.3. Neural cell adhesion molecule (NCAM)

NCAMs are perhaps the most widely studied adhesion molecules that are expressed on developing neurons in normal brain (108-111). NCAMs are implicated in mediating adhesion to neural elements and have been shown to bind to several types of collagens (112). A glycoprotein, NCAM, is composed of three polypeptides, NCAM-A, NCAM-B, and NCAM-C, with a molecular mass of 180 kDa, 140 kDa, and 120 kDa, respectively. The A and B proteins are phosphorylated, integral membrane proteins, while C is a nonphosphorylated peripheral polypeptide linked to the plasma membrane by a phosphatidylinositol anchor (113, 114). During migratory events in embryogenesis, NCAM is down-regulated, and it is re-expressed once the target organ has been reached and differention is initiated (21). Recent studies (115) showed that NCAM expression is correlated with different modes of invasion of rat glioma cell lines, both in vivo (116), and in vitro (117). NCAM expression has also been shown to correlate with downregulation of 92-kDa gelatinase. The finding that MMP-1 was downregulated by NCAM-B transfection but not in NCAM-C clones, suggested that the NCAM molecule's transmembrane domain, but not its extracellular domain, is involved in signal transduction (117).

5.4. Integrins

Integrins are heterodimeric integral plasma membrane cell-surface receptors that modulate cell-cell and cell-ECM interactions (118). They consist of noncovalently linked alpha and ß subunits. To date, 8ß and 16alpha subunits have been identified; their combinations produce 20 distinct integrins (119-121). Integrins are proposed to be involved in signal transduction from the ECM to the nuclei of cells. The cytoplasmic domain interacts with the cytoskeletal proteins in the vicinity of the cell membrane and is also responsible for the cell adhesion, shape and migration. The binding of cells to different ECM proteins, fibronectin, laminin, thrombospondin, von Willebrand's factor, and tenascin depends on the specific combination of alpha and ß subunits (120). The subunits (alpha, 120-180 kDa; ß, 90-110 kDa) are noncovalently associated and expressed on a wide variety of cell types. Integrins recognize an (RGD) sequence in fibronectin and vitronectin, in addition to Lys-Gln-Ala-Gly-Asp-Val (KQAGDV) in fibrinogen, Glu-Ile-Leu-Asp-Val (EILDV) in alternatively spliced segment of fibronectin, and Gly-Pro-Arg-Pro (GPRP) in fibrinogen (122, 118).

The cytoplasmic domain of the integrin molecule interacts with cytoskeletal proteins such as actin and talin, and activates specific kinases and phosphatases. The extracellular domain of the integrin interacts with ECM components and with integral plasma membrane receptors. Divalent cations, such as calcium and magnesium, play an important role in the function of integrins; a distinct calcium binding site has been identified on an integrin subunit (118). Altered expression of integrins has been shown to correlate with the malignant phenotype of tumors (123, 124). Integrin subunits link the cell's extracellular environment with internal cytoskeletal elements via linking proteins, such as talin and vinculin (125).

6. ECM TURNOVER AND DEGRADATION

Matrix degradation and turnover are important processes in tissue remodeling during development, wound healing, tumor necrosis, and inflammation. Key components of ECM turnover and regulation are the matrix metalloproteases (MMP) and their inhibitors TIMPs (tissue inhibitors of metalloproteases). A recently characterized metalloprotease family consists of membrane-type matrix metalloproteases (MT-MMP) that activate progelatinase A and a cascade of matrix proteases that in turn degrade extracellular matrix.

6.1. Matrix Metalloproteases

Matrix metalloproteases (MMPs) are a group of zinc-dependent enzymes that degrade ECM molecules, proteoglycans, glycoproteins, and various types of collagens (126, 127). MMPs are secreted by cells in a prometalloprotease form. The activation of MMPs in vivo is poorly understood. The MMP family has nine members that can be divided into three distinct groups based on their substrate specificities (table 2).

Often, the activation of the enzyme is completed by an autocatalytic activation that results in the loss of an aminoterminal peptide (130). MMP may be activated in vitro by detergents, organomercurial compounds, and proteolytic enzymes such as plasmin, trypsin, kallikrein and stromelysin (127-131). All these agents induce a conformational change such that the cystein-residue in the propeptide and the zinc molecule in the enzyme is disrupted exposing the zinc molecule (131). This, in turn, results in a partially active intermediate enzyme that cleaves the propeptide by autocatalytic activation and results in an active enzyme. Local invasive growth is one of the key features of primary brain tumors. The most common and malignant brain tumors, glioblastoma multiforme, is characterized by aggressive invasion of the surrounding normal brain. The specific mechanisms for this invasive behavior is still obscure; however, tumor cell interaction with the immediate ECM and subsequent degradation of the extracellular matrix components plays a key role in this process. Studies have shown that both MMP-2 (132) and MMP-9 (133) were significantly higher in malignant brain tumors in vivo. Moreover, MMP-9 facilitates the invasion of glioblastoma cells in vitro (25) and MMP-9 over expression correlates with the malignant progression of gliomas in vivo (134). MMPs function is regulated by naturally occurring protease inhibitors, tissue inhibitors of metalloproteases (TIMPs). The TIMP family consists of three members, TIMP-1 (a 28-kDa protein), TIMP-2 (a 21-kDa protein), and TIMP-3 (a 21-kDa protein) all with broad specificity in inhibiting MMPs (Table 2). However, the ability of these inhibitors to form complexes with latent and proenzyme forms of various types of collagenases differs significantly. Decreased levels of TIMP-1 and TIMP-2 expression favors the invasion of glioblastoma cells (135).

Membrane-type matrix metalloproteases are, recently discovered, distinct metalloproteases that are expressed on the surface of tumor cells. MT-MMPs activate gelatinase and also function as receptors for gelatinase A (136-138), inducing efficient pericellular proteolysis. So far, MT-MMP-1, MT-MMP-2, and MT-MMP-3 have been identified (139). MT-MMP expression has been found to correlate with the invasive behavior of HT-1080 and NIH3T3 cells (140). Our recent studies showed MT-MMP-1 to be highly upregulated in human glioblastomas (141).

6.2. Plasminogen activators

Plasminogen activators, (PAs) which belong to the serine protease family, are involved in matrix degradation. Two types of PA: urokinase-type plasminogen activator (u-PA) and tissue-type plasminogen activator (t-PA) convert plasminogen to plasmin, a broad-spectrum enzyme that cleaves fibrin, fibronectin, proteoglycans, and laminin (142, 143). uPA is involved in tissue remodeling during wound healing, inflammatory cellular migration, neo-vascularization and tumor cell invasion, while tPA, a key enzyme in thrombosis is involved in the dissolution of clots in blood vessels and the maintenance of hemostasis in the vasculature (144). tPA is expressed during brain development (145) and may be involved in neuronal migration and neurite outgrowth (146, 147). On the other hand, tPA is totally absent in glioblastoma, colon, lung and breast metastasis. However, anaplastic astrocytoma, low grade glioma and meningioma contain normal levels of tPA (148).

uPA is a highly specific serine protease that is secreted as zymogen, as a single chain (sc) polypeptide (scu-PA), and is readily converted into an active two-chain uPA (tcu-PA) by plasmin, kallikrein, cathepsin B, and nerve growth factor-gamma (143, 149-152). Receptor-bound active tcu-PA increases the rate at which plasminogen converts to plasmin (142, 143). The major substrate for uPA is plasmin, which has trypsin-like broad-spectrum protease activity, and which activates MMP (153, 154). uPA is synthesized by various tissues and cells, including those of CNS. One of the plasminogen activator, uPA binds to a specific cell surface receptor, uPAR and provides a localized cell surface proteolysis for the degradation of ECM. Proteolytic activity of uPA is modulated by its cell surface receptor, as well as by plasminogen activator inhibitors (PAIs). Recent studies demonstrated that overexpression of uPA was associated with malignant progression of human gliomas (155) and high affinity binding uPARs significantly contributes to the invasive behavior of gliomas in vitro (156). In situ hybridization studies have shown that uPAR mRNA was significantly higher in glioblastomas than in low grade glioma and normal brain (157).

Plasminogen activators are regulated by plasminogen activator inhibitors (PAIs) and protease nexin-I (PN-1). PAI-1 is a 46-54 kDa glycoprotein with high specificity for both uPA and tPA (158). Recent immunohistochemical and biochemical studies (159) have shown that the amount of PAI-1 was higher in malignant brain tumors. Using northern blot and in situ hybridization techniques, Yamamoto et al. (160) have shown higher levels of PAI-1 mRNA in glioblastoma than normal brain tissue. Recent evidence also indicates a correlation between the presence of tumor necrosis and higher levels of PAI-1 in glioblastomas (161). PAI-2, originally found in placental tissue (162), has a stronger affinity for uPA than tPA; it also exists in plasma. PAI-3, a 50-kDa inhibitor, has stronger affinity for uPA than tPA. The protease nexin-1 is a 43-50 kDa secreted glycoprotein, that inhibits uPA and plasmin. In the ECM, nexin-1 binds to thrombin, urokinase and plasmin (163). PN-1 is present in normal human brain (164), is produced mainly by astrocytes (165) and by glioblastoma cells (166, 167). Higher PN-1 levels have been observed in an experimental rat 9L gliosarcoma tumor than in normal brain (167).

7. PERSPECTIVE

We briefly summarized the molecular repertoire of extracellular matrix components of the brain and the role of ECM receptors and proteases involved in glioma cell invasion. Although different ECM components have been identified as having a role in glioma cell invasion, so far no single molecule has been identified as being responsible for the invasion. Understanding the significance of the ECM molecules will have a profound effect on anti-invasion strategies. One target would be the inhibition of proteolytic enzymes by an antisense method. Our recent results showed that antisense oligonucleotide transfection might be a suitable therapeutic approach to the inhibition of tumor cell invasion (Y. Go et al., unpublished observations). Inhibiting the synthesis of ECM components does not seem to be effective because tumor cells can find ECM molecules in the normal brain. Another approach might be to target integrin signaling in cell-matrix interaction. Whether this strategy can be exploited remains to be seen.

8. ACKNOWLEDGMENTS

We thank Lore Feldman for review and Sylvia Ledesma in preparation of the article. Supported in part, by NCI grant CA56792, and ACS grant EDT-91 (J.S. Rao).

9. REFERENCES

1. A. Bignami & R. Asher: Some observations on the localization of hyaluronic acid in adult, newborn and embryonal rat brain. Int J Dev Neurosci 10, 45-57 (1992).

2. J.T. Rutka, G. Apodaca, R. Stern & M. Rosenblum: The extracellular matrix of the central and peripheral nervous systems: structure and function. J Neurosurg 69, 155-70 (1988).

3. W. Paulus & C. Tonn: Interaction of glioma cells and extracellular matrix. J Neuro-Oncol 24, 87-91 (1995).

4. W. Paulus, W. Roggendorf & D. Schuppan: Immunohistochemical investigation of collagen subtypes in human gliomas. Virchows Arch-A, Pathological antomy & Histochem 413, 325-332 (1988).

5. H.J. Fryer, G.M. Kelly, L. Molinaro & S. Hockfield: The high molecular weight Cat-301 chondroitin sulfate proteoglycan from brain is related to the large aggregating proteoglycan from cartilage, aggrecan. J Biol Chem 267, 9874-83 (1992).

6. T. Kosaka, C.W. Heizmann & S.C. Fujita: Monoclonal antibody 473 selectively stains a population of GABAergic neurons containing the calcium-binding protein parvalbumin in the rat cerebral cortex. Exp Brain Res 89, 109-14 (1992).

7. S. Zaremba, J. R. Naegele, C.J. Barnstable & S. Hockfield: Neuronal subsets express multiple high-molecular-weight cell-surface glycoconjugates defined by monoclonal antibodies Cat-301 and VC1.1. J Neurosci 10, 2985-95 (1990).

8. D. A. Aquino, R.U. Margolis & R.K. Margolis: Immunocytochemical localization of a chondroitin sulfate proteoglycan in nervous tissue. II. Studies in developing brain. J Cell Biol 99, 1130-9 (1984a).

9. D.A. Aquino, R. U. Margolis & R.K. Margolis: Immunocytochemical localization of a chondroitin sulfate proteoglycan in nervous tissue. I. Adult brain, retina, and peripheral nerve. J Cell Biol 99, 1117-29 (1984b).

10. R.K. Margolis & R.U. Margolis: Structure and distribution of glycoproteins and glycosaminoglycans. In: R.K. and R.U. Margolis Eds, Complex carbohydrates of nervous tissue, Plenum Press, NY 45-73 (1979).

11. K. M. Buckley, E.S. Schweitzer, G. P. Miljanich, L. Clift-O'Grady, P.D. Kushner, L.F. Reichardt & R.B. Kelly: A synaptic vesicle antigen is restricted to the junctional region of the presynaptic plasma membrane. Proc Natl Acad Sci USA 80, 7342-6 (1983).

12. M.J. Anderson & D.M. Fambrough: Aggregates of acetylcholine receptors are associated with plaques of a basal lamina heparan sulfate proteoglycan on the surface of skeletal muscle fibers. J Cell Biol 97, 1396-411(1983).

13. R.P. Bunge & M. B. Bunge: Interrelationship between schwann cell function and extracellular matrix production. Trends Neurosci 6, 499-505 (1983).

14. B.J. Rollins, M.K. Cathcart & L.A. Culp. Fibronectin-protcoslycan binding as the molecular basis for fibroblast adhesion to extracellular matrices. In: M.I. Harowitz Ed. The Glycoconjugate, Academic Press, NY 3, 289-329 (1982).

15. B. Glimelius, B. Norling, B. Westermark & A. Wasteson: Composition and distribution of glycosaminoglycans in cultures of human normal and malignant glial cells. Biochem J 172, 443-56 (1978).

16. R.V. Iozzo: Proteoglycans and neoplastic-mesenchymal cell interactions. Human Pathol 15, 2-10 (1984).

17. B.P. Toole, R.L. Goldberg, G. Chi-Rosso, C.B.Underhill & R.W. Orkin: Hyaluronate-cell interactions In: R.L. Trelstad Ed., The role of extracelular matrix in development. Liss, NY 43-66 (1984).

18. W. Knudson, C. Biswas, X.Q. Li, R.E. Nemec & B.P. Toole: The role and regulation of tumor-associated hyaluronan. The Biology of Hyaluronan, CIBA Foundation Symposium 143 Chichester, 150-169 (1989).

19. H. Okada, J. Yoshida, M. Sokabe, T. Wakabayashi & M. Hagiwara: Suppression of CD44 expression decreases migration and invasion of human glioma cells. Int J Cancer 66, 255-60 (1996).

20. S. Koochekpour, G.J. Pilkington & A. Merzak: Hyaluronic acid/CD44H interaction induces cell detachment and stimulates migration and invasion of human glioma cells in vitro. Intl J Cancer 63, 450-4 (1995).

21. A. Merzak, S. Koocheckpour & G.J. Pilkington: CD44 mediates human glioma cell adhesion and invasion in vitro. Cancer Res 54, 3988-92 (1994a).

22. J.P. Thiery, J.L. Duband, U. Rutishauser & G.M. Edelman: Cell adhesion molecules in early chicken embryogenesis. Proc Natl Acad Sci USA 79, 6737-41 (1982).

23. G.J. Pilkington: Glioma heterogeneity in vitro: the significance of growth factors and gangliosides. Neuropath App Neurobiol 18, 434-442 (1992).

24. W. Knudson, C. Biswas & B.P.Toole: Interactions between human tumor cells and fibroblasts stimulate hyaluronate synthesis. Proc Natl Acad Sci USA 81, 6767-71 (1984).

25. J. S. Rao, P.A. Steck, P. Tofilon, D. Boyd, F. Ali-Osman, W.G. Stetler-Stevenson, Liotta LA, & R. Sawaya: Role of plasminogen activator and of 92-KDa type IV collagenase in glioblastoma invasion using an in vitro matrigel model. J Neuro-Oncol 18, 129-38 (1993).

26. P.C. Burger, E.R. Heinz, T. Shibata & P. Kleihues: Topographic anatomy and CT correlations in the untreated glioblastoma multiforme. J Neurosurg 68, 698-704 (1988).

27. H.P. Maxwell: The incidence of interhemispheric extension of glioblastoma multiforme through the corpus callosum. J Neurosurg 3, 54-57 (1946).

28. D.S. Russell & L. Rubinstein: Tumors of the central neuroepithelial origin In: Pathology of the Nervous System, 5th Ed. Hodder & Stoughton, London, 532 (1989).

29. P. Caroni & M.E. Schwab: Oligodendrocytes CNS myelin are nonpermissive substrates for neurite out growth and fibroblast spreading in vitro. J Neurosci 8, 2381-93 1988).

30. V.R. Amberger, P.A. Paganetti, H. Seyulberger, J.A. Eldering & M.E. Schwab: Characterization of a membrane-bound metalloendoprotease of rat C6 glioblastoma cells. Cancer Res 54, 4017-25 (1994).

31. A. Giese, L. Kluwe, B. Laube, H. Meissner, M.E. Berens & M. Westphal: Migration of human glioma cells on myelin. Neurosurg 38, 755-64 (1996).

32. M. Schachner, G. Schoonmaker & R.O. Hynes: Cellular and subcellular localization of LETS protein in the nervous system. Brain Res 158, 149-58 (1978).

33. S.K. Chintala, R. Sawaya, Z.L. Gokaslan, G. Fuller & J.S. Rao: Immunohistochemical localization of extracellular matrix proteins in human glioma, both in vivo and in vitro. Cancer Lett 101, 107-14 (1996).

34. E. Ruoslahti E & M.D. Pierschbacher: New perspectives in cell adhesion: RGD and integrins. Science 238, 491-7 (1987).

35. N. Kochi, E. Tani, T. Morimura & T. Itagaki: Immunohistochemical study of fibronectin in human glioma and meningioma. Acta Neuropath 59, 119-26 (1983).

36. J. Silnutzer & D.W.Barnes: Human serum factor. In: D.W. Barnes, D.A. Sirbasku and D.H. Sato Eds., Methods for preparation of media, supplements, and animal cell culture, Liss, NY 245 (1984).

37. K. Dahlback, H. Lofberg, J. Alumets & B. Dahlback: Immunohistochemical demonstration of age-related deposition of vitronectin (S-protein of complement) and terminal complement complex on dermal elastic fibers. J Invest Dermatol 92, 727-33 (1989).

38. F. Grinnell, C.H. Ho, & A. Wysocki: Degradation of fibronectin and vitronectin in chronic would fluid: analysis by cell blotting, immunoblotting, and cell adhesion assays. J Invest Dermatol 98, 410-6 (1992).

39. C.L. Gladson & D.A. Cheresh: Glioblastoma expression of vitronectin and the alpha v beta 3 integrin. Adhesion mechanism for transformed glial cells. J Clin Invest 88, 1924-32 (1991).

40. C. Kost, W. Stuber, H.J. Ehrlich, H. Pannekoek & K.T. Preissner: Mapping of binding sites for heparin, plasminogen activator inhibitor-1, and plasminogen to vitronectin's heparin-binding region reveals a novel vitronectin-dependent feedback mechanism for the control of plasmin formation. J Biol Chem 267, 12098-105 (1992).

41. D. Seiffert & D.J. Loskutoff: Evidence that type 1 plasminogen activator inhibitor binds to the somatomedin B domain of vitronectin. J Biol Chem 266, 2824-30 (1991).

42. Z. Gechtman, R. Sharma, T. Kreizman, M. Fridkin & S. Shaltiel: Synthetic peptides derived from the sequence around the plasmin cleavage site in vitronectin. Use in mapping the PAI-1 binding site. FEBS Lett 315, 293-7 (1993).

43. J. Mimuro, S. Muramatsu, Y. Kurano, Y. Uchida, H. Ikadai, S. Watanabe & Y. Sakata: Identification of the plasminogen activator inhibitor-1 binding heptapeptide in vitronectin. Biochem 32, 2314-20 (1993).

44. S.M Albelda, S.A. Mette, D.E. Elder, R. Stewart, L. Damjanovich, M. Herlyn & C.A. Buck: Integrin distribution in malignant melanoma: association of the beta 3 subunit with tumor progression. Cancer Res 50, 6757-64 (1990).

45. M. E. Berens, M.D. Rief, M. A. Loo & A. Giese: The role of extracellular matrix in human astrocytoma migration and proliferation studied in a microliter scale assay. Clin Exp Metast 12, 405-15 (1994).

46. E.I. Deryugina & M.A. Bourdon: Tenascin mediates human glioma cell migration and modulates cell migration on fibronectin. J Cell Sci 109, 643-52 (1996).

47. Y. Saga, T. Yagi, Y. Ikawa, T. Sakakura & S. Aizawa: Mice develop normally without tenascin. Genes Dev 6, 1821-31 (1992).

48. E.J. Mackie, W. Halfter & D. Liverani: Induction of tenascin in healing wounds. J Cell Biol 107, 2757-67 (1988).

49. W. Halfter, R. Chiquet-Ehrismann & R.P. Tucker RP: The effect of tenascin and embryonic basal lamina on the behavior and morphology of neural crest cells in vitro. Develop Biol 132, 14-25 (1989).

50. A. Lochter, L. Vaughan, A. Kaplony, A. Prochiantz, M. Schachner & A. Faissner: J1/tenascin in substrate-bound and soluble form displays contrary effects on neurite outgrowth. J Cell Biol 113, 1159-7 (1991).

51. H.P. Erickson & M.A. Bourdon: Tenascin: an extracellular matrix protein in specialized embryonic tissues and tumors. Ann Rev Cell Biol 5, 71-92 (1989).

52. E. Aufderheide & P. Ekblom: Tenascin during gut development: appearance in the mesenchyme, shift in molecular forms, and dependence on epithelial-mesenchymal interactions. J Cell Biol 107, 2341-49 (1988).

53. G.K. Koukoulis, V.E. Gould, A. Bhattacharya, J.E. Gould, A.A. Howeedy & I. Virtanen: Tenascin in normal, reactive, hyperplastic, and neoplastic tissues: biologic and pathologic implications. Human Pathol 22, 636-43 (1991) .

54. D.J. Whitby & M.W. Ferguson: The extracellular matrix of lip wounds in fetal, neonatal and adult mice. Development 112, 651-68 (1991).

55. D.J. Whitby, M.T. Longaker, M.R.Harrison, N.S. Adzick & M.W. Ferguson: Rapid epithelialisation of fetal wounds is associated with the early deposition of tenascin. J Cell Sci 99, 583-6 (1991).

56. H. Onada, M.I. Poulin, R. A. Tassava & I-M Chiu: Characterization of newt tenascin cDNA and localization of tenascin mRNA during newt limb regeneration by in situ hybridization. Dev Biol 148, 219-32 (1991).

57. R. Chiquet-Ehrismann: Tenascin and other adhesion-modulating proteins in cancer. Sem Cancer Biol 4, 301-10 (1993).

58. M. Higuchi, T. Ohnishi, N. Arita, S. Hiraga & T. Hayakawa: Expression of tenascin in human gliomas: its relation to histological malignancy, tumor dedifferentiation and angiogenesis. Acta Neuropath 85, 481-7 (1993).

59. T. Sakai, H. Kawakatsu, N. Hirota, T. Yokoyama, T. Sakakura & M. Saito: Specific expression of tenascin in human colonic neoplasms. Br J Cancer 67, 1058-64 (1993).

60. R.P. Tucker: The distribution of J1/tenascin and its transcript during the development of the avian cornea. Differentiation 48, 59-66 (1991).

61. R.P. Tucker: The in situ localization of tenascin splice variants and thrombospondin 2 mRNA in the avian embryo. Development 117, 347-58 (1993).

62. R.P. Tucker & S.E. McKay: The expression of tenascin by neural crest cells and glia. Development 112, 1031-9 (1991).

63. B. Wehrle-Haller, M. Koch, S. Baumgartner, J. Spring & M. Chiquet: Nerve-dependent and -independent tenascin expression in the developing chick limb bud. Development 112, 627-637 (1991).

64. R. Chiquet-Ehrismann: Tenascins, a growing family of extracellular matrix proteins. Experientia 51, 853-62 (1995).

65. H.P. Erickson: Tenascin-C, tenascin-R & tenascin-X: a family of talented proteins in search of functions. Curr Opin Cell Biol 5, 869-76 (1993).

66. Y. Morel, J. Bristow, S.E. Gitelman & W.L. Miller: Transcript encoded on the opposite strand of the human steroid 21-hydroxylase/complement component C4 gene locus. Proc Natl Acad Sci USA 86, 6582-6 (1989).

67. E.H. Sage & P. Bornstein: Extracellular proteins that modulate cell-matrix interactions. SPARC, tenascin, and thrombospondin. J Biol Chem 266, 14831-4 (1991).

68. R. Chiquet-Ehrismann, P. Kalla, C.A. Pearson, K. Beck & M. Chiquet: Tenascin interferes with fibronectin action. Cell 53, 383-90 (1988) .

69. M. M. Lotz, C.A. Burdsal, H.P. Erickson & D.R. McClay: Cell adhesion to fibronectin and tenascin: quantitative measurements of initial binding and subsequent strengthening response. J Cell Biol 109, 1795-805 (1989).

70. P. Joshi, C.Y. Chung, I. Aukhil & H.P. Erickson: Endothelial cells adhere to the RGD domain and the fibrinogen-like terminal knob of tenascin. J Cell Sci 106, 389-400 (1993).

71. P. Sriramarao, M. Mendler & M.A. Bourdon: Endothelial cell attachment and spreading on human tenascin is mediated by alpha 2 beta 1 and alpha v beta 3 integrins. J Cell Sci 105:1001-12 (1993).

72. A.L. Prieto, G.M. Edelman & K.L. Crossin: Multiple integrins mediate cell attachment to cytotactin/tenascin. Proc Natl Acad Sci USA 90, 10154-8 (1993).

73. M.T. Giordana, I. Germano, G. Giaccone, A. Mauro, A. Migheli, & D. Schiffer: The distribution of laminin in human brain tumors: an immunohistochemical study. Acta Neuropathol 67, 51-7 (1985).

74. D. Schiffer, M.T. Giordana, A. Mauro & A. Migheli: GFAP, F VIII/RAg, laminin, and fibronectin in gliosarcomas: an immunohistochemical study. Acta Neuropathol 63, 108-16 (1984).

75. R. Timpl: Structure and biological activity of basement membrane proteins. Eur J Biochem 180, 487-02 (1989).

76. K. Beck, I. Hunter & J. Engel: Structure and function of laminin: anatomy of a multidomain glycoprotein. FASEB J 4, 148-60 (1990).

77. P.D. Yurchenco & J.C. Schittny: Molecular architecture of basement membranes. FASEB J 4, 1577-90 (1990).

78. R. Timpl & J.C. Brown: The laminins. Matrix Biol 14, 275-81 (1994).

79. P.D. Yurchenco & J.J. O'Rear: Basal lamina assembly. Curr Opin Cell Biol 6, 674-81 (1994).

80. R.E. Burgeson, M. Chiquet, R. Deutzmann, P. Ekblom, J. Engel, H. Kleinman, G.R. Martin G. Meneguzzi, M. Paulsson & J. Sanes: A new nomenclature for the laminins. Matrix Biol 14, 209-11 (1994).

81. S. Gunwar, J. Saus, M.E. Noelken & B.G. Hudson: Glomerular basement membrane. Identification of a fourth chain, alpha 4, of type IV collagen. J Biol Chem 265, 5466-9 (1990).

82. R. Bjerkvig, O.D. Laerum & G. J. Rucklidge: Immunocytochemical characterization of extracellular matrix proteins expressed by cultured glioma cells. Cancer Res 49, 5424-8 (1989).

83. G. Bellon, T. Caulet, Y Cam, M. Pluot, G. Poulin, M. Pytlinska & M.H. Bernard: Immunohistochemical localisation of macromolecules of the basement membrane and extracellular matrix of human gliomas and meningiomas. Acta Neuropathol 66, 245-52 (1985).

84. J.T. Rutka, C.A. Myatt, J.R. Giblin, R.L. Davis & M.L. Rosenblum: Distribution of extracellular matrix proteins in primary human brain tumours: an immuno-histochemical analysis. Canad Neurol Sci 14, 25-30 (1987).

85. H.J. Scherer: The forms of growth in gliomas and their practical significance. Brain 63, 1-35 (1940).

86. M. Jucker, H.K. Kleinman & D.K. Ingram: Fetal rat septal cells adhere to and extend processes on basement membrane, laminin, and a synthetic peptide from the laminin A chain sequence. J Neurosci Res 28, 507-17 (1991).

87. H.K. Kleinman, G.C. Sephel, K. Tashiro, B.S. Weeks, B.A. Burrous, S.H. Adler, Y. Yamada & G.R. Martin: Laminin in neuronal development. Ann NY Acad Sci 580, 302-10 (1990).

88. J.J. Bernstein & C.A. Woodard: Glioblastoma cells do not intravasate into blood vessels. Neurosurg 36, 124-32 (1995).

89. G. Taraboletti, D.D. Roberts & L.A. Liotta: Thrombospondin-induced tumor cell migration: haptotaxis and chemotaxis are mediated by different molecular domains. J Cell Biol 105, 2409-15 (1987).

90. L.A. Liotta:Tumor invasion and metastases--role of the extracellular matrix: Rhoads Memorial Award lecture. Cancer Res 46, 1-7 (1986).

91. C. H. Graham, I. Connelly, J. R. MacDougall, R.S. Kerbel, W.G. Stetler-Stevenson & P.K. Lala: Resistance of malignant trophoblast cells to both the anti-proliferative and anti-invasive effects of transforming growth factor-beta. Exp Cell Res 214, 93-9 (1994).

92. L.F. Reichardt: Extracellular matrix molecules and their receptors. In: T. Kreis and R Vale Eds., Guidebook to the extracellular matrix and adhesion proteins, Oxford University Press, Oxford, New York, Toronto 3-11 (1993).

93. S.I. Hakomori: Glycosphingolipids in cellular interaction, differentation and oncogenesis. Annu Rev Biochem 50, 733-64 (1981).

94. L. Svennerholm: Chromatographic separation of human brain gangliosides. J Neurochem 10, 613-23 (1963).

95. A. Merzak, S. Koochekpour & G.J. Pilkington: Cell surface gangliosides are involved in the control of human glioma cell invasion in vitro. Neurosci Lett 177, 44-6 (1994).

96. A. Merzak, S. Koochekpour, & G.J. Pilkington: Adhesion of human glioma cell lines to fibronectin, laminin, vitronectin and collagen I is modulated by gangliosides in vitro. Cell Adhesion Commun 3, 27-43 (1995).

97. L.K. Thompson, P.M. Horowitz, K.L. Bentley, D.D. Thomas, J.F. Alderete & R.J. Klebe : Localization of the ganglioside-binding site of fibronectin. J Biol Chem 261, 5209-14 (1986).

98. T.D. Traylor & E.L. Hogan: Gangliosides of human cerebral astrocytomas. J Neurochem 34, 126-31 (1980).

99. D.A. Cheresh, R. Pytela, M.D. Pierschbacher, F.G. Klier, E. Ruoslahti & R.A. Reisfeld: An Arg-Gly-Asp-directed receptor on the surface of human melanoma cells exists in an divalent cation-dependent functional complex with the disialoganglioside GD2. J Cell Biol 105, 1163-73 (1987).

100. G. Mugnai, K. Lewandowska, H.U. Choi, L.C. Rosenberg & L.A. Culp: Ganglioside-dependent adhesion events of human neuroblastoma cells regulated by the RGDS-dependent fibronectin receptor and proteoglycans. Exp Cell Res 175, 229-47 (1988).

101. W.B. Stallcup: Involvement of gangliosides and glycoprotein fibronectin receptors in cellular adhesion to fibronectin. Exp Cell Res 177, 90-102 (1988).

102. L.J. Picker, M. Nakache & E.C. Butcher: Monoclonal antibodies to human lymphocyte homing receptors define a novel class of adhesion molecules on diverse cell types. J Cell Biol 109, 927-37 (1989).

103. B. F. Flanagan, R. Dalchau, A.K. Allen, A.S. Daar & J.W. Fabre: Chemical composition and tissue distribution of the human CDw44 glycoprotein. Immunol 67,167-75 (1989).

104. M.S. Sy, Y.J. Guo & I. Stamenkovic: Distinct effects of two CD44 isoforms on tumor growth in vivo. J Exp Med 174, 859-66 (1991).

105. N. Girgrah, M. Letarte, L.E. Becker, T.F. Cruz, E. Theriault, & M.A. Moscarello: Localization of the CD44 glycoprotein to fibrous astrocytes in normal white matter and to reactive astrocytes in active lesions in multiple sclerosis. J Neuropath Exp Neurol 50, 779-92 (1991).

106. E.J. Quackenbush, S. Vera, A. Greaves, & M. Letarte: Confirmation by peptide sequence co-expression on various cell types of the identity of CD44 and P85 glycoprotein. Mol Immunol 27, 947-55 (1990).

107. B. Radotra, D. McCormick & A. Crockard: CD44 plays a role in adhesive interactions between glioma cells and extracellular matrix components. Neuropath Appl Neurobiol 20, 399-405 (1994).

108. B.A. Cunningham, J.J. Hemperly, B.A. Murray, E.A. Prediger, R. Brackenbury & G.M. Edelman: Neural cell adhesion molecule: structure, immunoglobulin-like domains, cell surface modulation, and alternative RNA splicing. Science 236, 799-806 (1987) .

109. G.M. Edelman: Cell adhesion molecules in the regulation of animal form and tissue pattern. Ann Rev Cell Biol 2, 81-116 (1986) .

110. D. Goldowitz, D. Barthels, N. Lorenzon, A. Jungblut, & W. Wille: NCAM gene expression during the development of cerebellum and dentate gyrus in the mouse. Brain Res Dev Brain Res 52, 151-60 (1990).

111. U. Rutishauser, A. Acheson, A.K. Hall, D.M. Mann & J. Sunshine: The neural cell adhesion molecule (NCAM) as a regulator of ell-cell interactions. Science 240, 53-7 (1988).

112. R. Probstmeier, K. Kuhn & M. Schachner: Binding properties of the neural cell adhesion molecule to different components of the extracellular matrix. J Neurochem 53, 1794-801 (1993).

113. J.M. Lyles, D. Linnemann & E. Bock: Biosynthesis of the D2-cell adhesion molecule: post-translational modifications, intracellular transport, and developmental changes. J Cell Biol 99, 2082-91 (1984).

114. O. Nybroe, M. Albrechtsen, J. Dahlin, D. Linnermann, J.M. Lyles, C. J. Moller & E. Bock: Biosynthesis of the neural cell adhesion molecule: characterization of polypeptide C. J Cell Biol 101, 2310-15 (1985).

115. G.J. Rucklidge, K. Edvardsen & E. Bock: Cell-adhesion molecules and metallo-proteinases: a linked role in tumour cell invasiveness. Biochem Soc Transact 22, 63-8 (1994).

116. A.M. Andersson, N. Moran, H. Gaardsvoll, D. Linnemann, R. Bjerkvig, O.D. Laerum & E. Bock: Characterization of NCAM expression and function in BT4C and BT4Cn glioma cells. Int J Cancer 47, 124-9 (1991).

117. R. Bjerkvig, O.D. Laerum & O. Mella: Glioma cell interactions with fetal rat brain aggregates in vitro and with brain tissue in vivo. Cancer Res 46, 4071-9 (1986).

118. R.O. Hynes: Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69, 11-25 (1992).

119. R.O. Hynes: Integrins: a family of cell surface receptors. Cell 48, 549-54 (1987).

120. E. Ruoslahti: Integrins. J Clin Invest 87, 1-5 (1991).

121. S. Dedhar: Integrin mediated signal transduction in oncogenesis: an overview. Cancer Metast Rev 14, 165-72 (1995).

122. J. D. Loike, B. Sodeik, L. Cao, S. Leucona, J.I. Weitz, P.A. Detmers, S.D. Wright & S.C. Silverstein:CD11c/CD18 on neutrophils recognizes a domain at the N terminus of the alpha chain of fibrinogen. Proc Nat Acad Sci USA 88, 1044-8 (1991).

123. F.G. Giancotti & F. Mainiero: Integrin-mediated adhesion and signaling in tumorigenesis. Biochim Biophy Acta 1198, 47-64 (1994).

124. S. Dedhar: Integrins and tumor invasion. Bioessays 12, 583-90 (1990).

125. K. Burridge, K. Fath, T. Kelly, G. Nuckolls & C. Turner: Focal adhesions: transmembrane junctions between the extracellular matrix and the cytoskeleton. Ann Rev Cell Biol 4, 487-525 (1988).

126. L.M. Matrisian: Metalloproteinases and their inhibitors in matrix remodeling. Trends Genet 6, 121-5 (1990).

127. J.F. Woessner, Jr: Matrix metalloproteinases and their inhibitors in connective tissue remodeling. FASEB J 5, 2145-54 (1991).

128. L.M. Matrisian & B.W. Ennis: Matrix degrading metalloproteinases. J Neuro-Oncol 18, 105-109 (1994).

129. L. M. Matrisian: The matrix-degrading metalloprotienases. BioEssays 14, 455-463 (1992).

130. W.G. Stetler-Stevenson, H.C. Krutzsch, M.P. Wacher, I.M. Margulies & L.A. Liotta: The activation of human type IV collagenase proenzyme. Sequence identification of the major conversion product following organomercurial activation. J Biol Chem 264, 1353-56 (1989).

131. E.B. Springman, E.L. Angleton, H. Birkedal-Hansen & H.E. Van Wart: Multiple modes of activation of latent human fibroblast collagenase: evidence for the role of a Cys73 active-site zinc complex in latency and a "cysteine switch" mechanism for activation. Proc Natl Acad Sci USA 87, 364-8 (1990).

132. R.E. Sawaya, M. Yamamoto, Z.L. Gokaslan, S.W. Wang, S. Mohanam, G.N. Fuller, I.E. McCutcheon, W.G. Stetler-Stevenson, G.L Nicolson & J. S. Rao. Expression and localization of 72 kDa type IV collagenase (MMP-2) in human malignant gliomas in vivo. Clin Exp Metastasis 14, 35-42 (1996).

133. J.S. Rao, P.A. Steck, S. Mohanam, W.G. Stetler-Stevenson, L.A. Liotta & R. Sawaya: Elevated levels of Mr 92,000 type IV collagenase in human brain tumors. Cancer Res 53, 2208-2211 (1993).

134. J.S. Rao, M. Yamamoto, S. Mohanam, G.N. Fuller, Z.L. Gokaslan, W.G. Stetler-Stevenson, M. Nakajima, G.L. Nicholson & R. Sawaya: Expression and localization of 92-kDa type IV collagenase/gelatinase B (MMP-9) in human gliomas. Clin Exp Metastasis 14, 12-8 (1996).

135. S. Mohanam, S.W. Wang, A. Rayford, M. Yamamoto, R. Sawaya, M. Nakajima, L.A. Liotta, G.L. Nicolson, W.G. Stetler-Stevenson & J.S. Rao: Expression of tissue inhibitors of metalloproteinases: negative regulators of human glioblastomas invasion in vivo. Clin Exp Metastasis 13, 57-62 (1995).

136. W.L. Monsky, T. Kelly, C.Y. Lin, Y. Yeh, W.G. Stetler-Stevenson, S.C. Mueller & W.T. Chen: Binding and localization of M(r) 72,000 matrix metalloproteinase at cell surface invadopodia. Cancer Res 53, 3159-64 (1993).

137. S. Zucker, U.M. Moll, R.M. Lysik, E.I. DiMassimo, W.G. Stetler-Stevenson, L.A. Liotta & J.W. Schwedes: Extraction of type-IV collagenase/gelatinase from plasma membranes of human cancer cells. Intl J Cancer 45, 1137-42 (1990).

138. H. Sato & M. Seiki: Membrane-type matrix metalloproteinases (MT-MMPs) in tumor metastasis. J Biochem 119, 209-15 (1996).

139. H. Will & B. Hinzmann: cDNA sequence and mRNA tissue distribution of a novel human matrix metalloproteinase with a potential transmembrane segment. Eur J Biochem 231, 602-8 (1995).

140. H. Sato, T. Takino, Y. Okada, J. Cao, A. Shinagawa, E. Yamamoto & M. Seiki: A matrix metalloproteinase expressed on the surface of invasive tumour cells. Nature 370, 61-5 (1994).

141. M. Yamamoto, S. Mohanam, R. Sawaya, G.N. Fuller, M. Seiki, H. Sato, Z.L. Gokaslan, L.A. Liotta, G.L. Nicolson & J.S. Rao: Differential expression of membrane-type matrix metalloproteinase and its correlation with gelatinase A activation in human malignant brain tumors in vivo and in vitro. Cancer Res 56, 384-92 (1996).

142. M. Laiho & J. Keski-Oja: Growth factors in the regulation of pericellular proteolysis: a review. Cancer Res 49, 2533-53 (1989).

143. B.B. Wolf, J.Vasudevan, J. Henkin & S.L. Gonias: Nerve growth factor-gamma activates soluble and receptor-bound single chain urokinase-type plasminogen activator. J Biol Chem 268, 16327-31 (1993).

144. D.A. Hart & A. Rehemtulla: Plasminogen activators and their inhibitors: regulators of extracellular proteolysis and cell function. Comp Biochem Physiol - B: Comp Biochem 90, 691-708 (1988).

145. G. Moonen, M.P. Grau-Wagemans, I. Selak, P.P. Lefebvre, B. Rogister, J.D. Vassalli & D. Belin: Plasminogen activator is a mitogen for astrocytes in developing cerebellum. Brain Res 352, 41-8 (1985).

146. A. Krystosek & N.W. Seeds: Plasminogen activator release at the neuronal growth cone. Science 213, 1532-4 (1981).

147. G. Moonen, M.P. Grau-Wagemans & I. Selak: Plasminogen activator-plasmin system and neuronal migration. Nature 298, 753-5 (1982).

148. A.K. Bindal, M. Hammoud, W.M. Shi, S.Z. Wu, R. Sawaya & J.S. Rao: Prognostic significance of proteolytic enzymes in human brain tumors. J Neuro-Oncol 22, 101-110 (1994).

149. L. Ossowski, D. Biegel & E. Reich: Mammary plasminogen activator: correlation with involution, hormonal modulation and comparison between normal and neoplastic tissue. Cell 16, 929-40 (1979).

150. R.H. Goldfarb, G. Murano, R. Brundage, G.P. Siegal, V. Terranova, S. Garbisa & L.A. Liotta: Degradation of glycoprotein and collagenous components of the basement membrane: studies with urokinase-type plasminogen activator, alpha-thrombin, and plasmin. Sem Thromb Haemost 12, 335-6 (1986).

151. A. Ichinose, K. Fujikawa & T. Suyama: The activation of pro-urokinase by plasma kallikrein and its inactivation by thrombin. J Biol Chem 261, 3486-9 (1986).

152. H. Kobayashi, M. Schmitt, L. Goretzki, N. Chucholowski, J. Calvete, M. Kramer, W.A. Gunzler, F. Janicke & H. Graeff : Cathepsin B efficiently activates the soluble and the tumor cell receptor-bound form of the proenzyme urokinase-type plasminogen activator (Pro-uPA). J Biol Chem 266, 5147-52 (1991).

153. W.T. Chen: Membrane proteases: roles in tissue remodeling and tumour invasion. Curr Opin Cell Biol 4, 802-9 (1992).

154. C. He, S.M. Wilhelm, A.P. Pentland, B.L. Marmer, G.A. Grant, A.Z. Eisen & G.I. Goldberg: Tissue cooperation in a proteolytic cascade activating human interstitial collagenase. Proc Natl Sci Acad USA 86, 2632-36 (1989).

155. M. Yamamoto, R. Sawaya, S. Mohanam, A.K. Bindal., J.M. Bruner, K. Oka, V.H. Rao, M. Tomonaga, G.L. Nicolson & J.S. Rao: Expression and localization of urokinase-type plasminogen activator in human astrocytoma in vivo. Cancer Res. 54, 3656-3661 (1994).

156. S. Mohanam, R. Sawaya, D. Boyd, F. Ali-Osman & J.S. Rao: Modulation of in vitro invasion of human glioblastoma cells by urokinase type plasminogen activator receptor antibody. Cancer Res 53, 4143-4147 (1993).

157. M. Yamamoto, R. Sawaya, S. Mohanam, V.H. Rao, J.M. Bruner, G.L. Nicolson & J.S. Rao: Expression and localization of urokinase-type plasminogen activator receptor in human gliomas. Cancer Res 54, 5016-5020 (1994).

158. P.A. Andreasen, L.S. Nielsen, P. Kristensen, J. Grondahl-Hansen, L. Skriver, & K. Dan : Plasminogen activator inhibitor from human fibrosarcoma cells binds urokinase-type plasminogen activator, but not its proenzyme. J Biol Chem 261, 7644-51 (1986).

159. J.S. Rao, A. Rayford, S. Kono, R.A. Morantz, B.W. Festoff & R. Sawaya: Increased levels of plasminogen activator inhibitor-1 (PAI-1) in human brain tumors. J Neuro-Oncol 17, 215-221 (1993).

160. M. Yamamoto, R. Sawaya, S. Mohanam, D.J. Loskutoff, J.M. Bruner, V.H. Rao, K. Oka, M. Tomonaga, G.L. Nicolson & J.S. Rao: Expression and cellular localization of messenger RNA for plasminogen activator inhibitor type 1 in human astrocytomas in vivo. Cancer Res 54, 3329-3332 (1994).

161. R. Sawaya, M. Yamamoto, O.J. Ramo, M.L. Shi, A. Rayford & J.S. Rao: Plasminogen activator inhibitor-1 in brain tumors: Relation to malignancy and necrosis. Neurosurg 36, 375-381 (1995).

162. B. Astedt, I. Lecander, T. Brodin, A. Lundblad & K. Low: Purification of a specific placental plasminogen activator inhibitor by monoclonal antibody and its complex formation with plasminogen activator. Thromb & Haemost 53, 122-5 (1985).

163. J.B. Baker, D.A. Low, R.L. Simmer & D.D. Cunningham: Protease-nexin: a cellular component that links thrombin and plasminogen activator and mediates their binding to cells. Cell 21, 37-45 (1980).

164. S.L. Wagner, J.W. Geddes, C.W. Cotman, A.L. Lau, D. Gurwitz, P.J. Isackson & D.D. Cunningham: Protease nexin-1, an antithrombin with neurite outgrowth activity, is reduced in Alzheimer disease. Proc Natl Acad Sci USA 86, 8284-8 (1989).

165. D.E. Rosenblatt, C.W. Cotman, M. Nieto-Sampedro, J.W. Rowe & D.J. Knauer: Identification of a protease inhibitor produced by astrocytes that is structurally and functionally homologous to human protease nexin-I. Brain Res 415, 40-8 (1987).

166. S.L. Wagner, A.L. Lau, A. Nguyen, J. Mimuro, D.J. Loskutoff, P.J. Isackson & D.D. Cunningham: Inhibitors of urokinase and thrombin in cultured neural cells. J Neurochem 56, 234-42 (1991).

167. J.S. Rao, J.B. Baker, R.A. Morantz, B. Kimler, R. Evans & B.W. Festoff: Serpin inhibitors of urokinase and thrombin in normal rat brain and the 9L brain tumor: evidence for elevated expression of protease nexin I-like inhibitor and a novel sodium dodecyl sulfate activated tumor antithrombin. Cancer Res 50, 5039-44 (1990).