In Chapter 2 the distinctions between benign and malignant neoplasia were discussed, with the caveat that the distinction was largely artificial but convenient. In the context of the stages of neoplastic development, the imperfections of the behavioral classification of neoplasia become more apparent. While virtually all preneoplastic lesions may be considered as benign in the be- havioristic classification, a significant number of benign lesions also occur during progression, such as adenomas of the thyroid, stomach, colon, ovary, and liver, as well as a few mesenchymal neoplasms (Henson and Albores-Saavedra, 1986). Still, the key distinction between benign and malignant neoplasms resides in their ability to extend beyond the primary growth by the process of invasion into adjacent tissues and to successfully metastasize, with resultant independent sec- ondary growths. In recent years the characteristics and to some extent the mechanisms involved in the development of these anatomical extensions of neoplasia, invasion, and metastasis have been clarified.
Initial considerations of the characteristics and mechanisms of invasive neoplastic growth ar- gued that the increased mobility and rapidity of growth of the neoplastic cell, together with a capacity for proteolysis and a decrease in pH (possibly owing to the high glycolysis of the neo- plasm; Chapter 15), were responsible for its ability to invade normal tissue (cf. Sylvén, 1968). Other mechanisms also were invoked to explain tumor cell invasion, including differences in intracellular osmotic pressure between normal and neoplastic cells, the degree of host reactions such as inflammation, the anatomical structure of the tissue being invaded, and the loss of “con- tact inhibition,” a process whereby normal cells are inhibited in growth and motion by adjacent cells. All of these were thought to be factors in neoplastic cell invasion both in vivo and in vitro.
During the last decade, these descriptive characteristics and mechanisms of the process of neoplastic invasion of normal tissues have been greatly advanced by a better understanding of the molecular nature of cell-cell interaction in vivo as well as the interaction of cells with their immediate environment as defined by their surrounding extracellular matrix.
The Extracellular Matrix
In general terms, the extracellular matrix (ECM) consists of a mixture of macromolecules, ions, and substrates that are specific for any given tissue. The ECM thus provides a distinct environ- ment for different cells of the organism (Scott-Burden, 1994). The ECM is largely made up of a complex blend of macromolecules, some of which, once formed, may exist for the life span of the organism (Hay, 1991). The macromolecules of the ECM largely but not entirely consist of fiber-forming elements (collagen and related molecules) and glycoprotein and proteoglycan “packing” components. A diagram of the ECM associated with vascular elements is given in Figure 10.2, in the upper part of which are depicted endothelial cells resting on a basement mem- brane. Basement membranes are heterogeneous, highly specialized, electron-dense structures constructed from components of a number of the extracellular matrix proteins. They vary from 20 to 200 nm in thickness and function to separate epithelia and endothelia from their underlying connective tissue. Basement membranes provide anchorage for adjacent cells as well as stimuli for cell differentiation, cell migration, and cell phenotype (Stanley et al., 1982; Weber, 1992).
Of the ECM molecules depicted in Figure 10.2, collagen is the predominant species, mak- ing up approximately one-third of the protein in the human body (cf. Scott-Burden, 1994). There are more than 14 different species of collagen molecules (van der Rest and Garrone, 1991), with
Figure 10.2 Diagrammatic representation of the ECM in association with vascular elements. Endothe- lial cells (EC) rest on a basement membrane (BM) of the internal vascular wall, whereas smooth muscle cells (SMC) are surrounded by molecules of the ECM. The internal elastic lamina (EL) acts as a fenestrated layer between the subendothelial intima and smooth muscle cells of the vascular medial layer. Molecules of the ECM shown schematically are tenascin (TN), proteoglycans (PG), laminin (LN), fibronectin (FN), thrombospondin (TSP), and collagen fibers which are depicted as large dashed lines. (Adapted from Scott- Burden, 1994, with permission of the author and publishers.)
type 1 collagen the predominant species in the ECM. The collagen structure is that of an ex- tended triple helix bound together by numerous hydrogen bonds, some of which are dependent for their formation on posttranslational hydroxylation of proline and lysine catalyzed by a vita- min C (ascorbic acid)–dependent reaction. In addition, the collagen structure is further cemented by glycosylation of these and other hydroxyl groups, as well as oxidation of the ε-amino groups of lysine, some of which are oxidized by a copper-dependent lysyl oxidase and form covalent crosslinks with amino groups of adjacent collagen molecules, binding the multimolecular struc- ture into strong, stress-resistant collagen fibers. Elastin—which is a key component of the ECM in pliable tissues such as vessels, alveolae of the lung, dermis, and intestine—has a structure somewhat similar to that of collagen, but with more hydrophobic domains that act as “coiled springs” and are related to the protein’s elasticity.
A variety of glycoproteins and proteoglycans are associated with the ECM in normal tis- sues. The diagrammatic structures of several of these molecules are seen in Figure 10.3. The three glycoproteins shown—fibronectin, laminin, and SPARC—are generally representative of the glycoproteins existing in the ECM. All of the glycoproteins have various domains within the protein that interact selectively with cellular and matrix components, as indicated in the dia- gram. Fibronectin is composed of two nearly identical polypeptides linked by disulfide bridges and having a combined molecular weight near 450,000 (cf. Ruoslahti et al., 1982). The molecule is made up of a series of polypeptide repeats, as indicated in the figure (I, II, and III). Fibronectin monomers have varying numbers of these repeats, the number being dependent on differential splicing of fibronectin mRNA in various tissues. As a result, fibronectin monomers may exist in a wide diversity of molecular sizes. SPARC consists of a single polypeptide chain with four do- mains, as shown. In contrast, the laminin structure is more cruciform, consisting of three peptide chains linked together by disulfide bonding in the carboxyl half of the three peptides. Each of the three peptides is encoded by a different gene (cf. Scott-Burden, 1994), but no differential mRNA
Figure 10.3 Diagrammatic structures of glycoproteins and proteoglycans of the extracellular matrix. The three different types of polypeptide repeats in fibro- nectin are noted as types I, II, and III in each of the two chains. (A) The domains are indicated in boxes. (B) The SPARC protein domains are indicated beneath the diagrammatic structure with an indication of the exons above the diagrammatic structure. (C) The association of the three peptide components of laminin are noted in addition to an indication of the regions binding collagen and the cell. (D) The generalized structure of this typical proteoglycan monomer shows the protein core and positions of the N-and O-linked oligosaccharides as well as the hyaluronic acid (HA) binding region. CS, chondroitin sulfate; DS, dermatan sulfate; HS, heparan sulfate; KS, keratan sulfate. Further details are given in the text. (Structures were adapted from the following: A, C: Scott-Burden, 1994; B: Lane and Sage, 1994; D: Wight, 1989 with permission of the authors and publishers.)
splicing has been reported, although at least seven different forms resulting from the association of different gene products have been reported (Timpl and Brown, 1994). Laminin and at least two other ECM glycoproteins, thrombospondin and tenascin, possess a large number of epidermal growth factor–like repeats in their structure. It has been theorized that cleavage of the molecule may produce active epidermal growth factor molecules affecting the growth of tissue (Chapter 14).
The proteoglycans consist of a protein backbone (core protein) that has many threonine and serine residues. To these hydroxyl groups are attached a series of long, linear carbohydrate chains termed glycosaminoglycans, which usually consist of a dimeric repeat structure of acid and amine structures of sugars (cf. Scott-Burden, 1994). Individual proteoglycans may bind to hyaluronic acid chains or in some cases may actually span membranes. A “link” protein may assist in this association, which leads to the formation of large aggregated molecules that largely fill the interstices of the ECM. It must be emphasized that the proteoglycans are an extremely diverse series of molecules, exhibiting the general structure shown in Figure 10.3 but having different oligosaccharides and core proteins, depending on the tissue and the species (Kjellén and Lindahl, 1991).