CHARACTERISTICS OF THE SURFACE MEMBRANE OF NORMAL AND TRANSFORMED CELLS IN VITRO
In Chapters 10 and 14, the importance of the structure of the external plasma membrane in rela- tion to tumor invasion, tumor metastases, and the phenomenon of contact inhibition of both cell movement and cell replication were discussed. Thus it was appropriate that one of the earliest investigations of the biochemistry and cell biology of transformed cells was related to the sur- face membrane. In 1969 Burger (cf. Burger, 1973) demonstrated that the addition of certain plant proteins, loosely termed agglutinins, would cause the agglutination of virally transformed cells in culture. In contrast, the parent cells from which the transformed cells were derived did not agglutinate when specific plant materials were added to the medium. The purified material responsible for the agglutination was found to be a glycoprotein with a molecular weight of ap- proximately 18,000. This material reacted in a similar manner with a number of neoplastic cell types obtained from neoplasms growing in vivo as well as cells transformed in culture by chem- icals, ionizing radiation, viruses, or “spontaneously.” In many instances, cells that had reverted or lost their transformed phenotype also lost their capacity for agglutination in the presence of the plant agglutinin.
Several other plant agglutinins or lectins, as they are known, have been found to affect neoplastic cells in a similar manner. In addition, some nontransformed cells also exhibit agglu- tinability, thus making the original generalization invalid. Furthermore, Burger and others (cf. Sharon, 1977) demonstrated that treatment of normal cells with trypsin for very short periods of time rendered them agglutinable. This last experiment indicated that normal cells contained re- ceptor sites for the plant agglutinins but that these sites were normally “protected” by some pep- tide components of the surface membrane. Furthermore, it was shown by Sachs (1974) that some variants of polyoma-transformed cells showed varying degrees of agglutination by concanavalin A, another plant lectin. Sachs’ laboratory also demonstrated that normal fibroblasts in mitosis are agglutinated by concanavalin A as well as by the wheat germ lectin, whereas transformed fibroblasts in mitosis are not agglutinated by these lectins.
As can be seen from Table 16.2, lectins from both plant and animal sources interact rather specifically with certain sugars and their derivatives. Those from plant sources react primarily with specific sugar moieties, while lectins or selectins from mammalian sources (Chapter 10) react with more complex oligosaccharide structures. As noted from the table, lectins are proteins with a selective affinity for simple or complex sugars. Their specificity is dependent not only on the presence of the sugar in the terminal position but also on its anomeric confirmation, the at- tachment site to and nature of the subterminal sugar, the number of receptor sites, and the degree of steric hindrance caused by surrounding structures. In the mammal, as exemplified in the hu- man, lectins are involved in the interaction of specific cells with other specific cell types (Chap- ter 10). Their functions in plants is not entirely clear.
The binding of lectins to the carbohydrate moieties listed in Table 16.2 is not by means of covalent linkages but rather through weak molecular interactions such as van der Waals forces. Other studies (cf. Nicolson, 1976) have demonstrated that whether cells were agglutinated by a
Table 16.2 Some Examples of Specific Lectins in Plants and Animals
lectin or not, the total number of lectin molecules bound to the surface was usually the same for both normal and neoplastic cells. The answer to this puzzle became apparent when lectin mole- cules labeled with fluorescent dyes were used to interact with normal and neoplastic cells. Such experiments demonstrated a difference in the surface distribution of lectins bound to normal cells compared with lectins bound to tumor cells. On normal cells, the lectin molecules were distributed randomly, whereas they appeared to be aggregated into clusters on the surface of tu- mor cells. This is shown in Figure 16.2. Furthermore, lectins in their polyvalent form are capable of inducing a redistribution of lectin-binding sites on the plasma membrane of the cell, which then becomes agglutinated. This latter process may extend to the clustering of such sites to form a large mass on one portion of the cell surface, which has been termed a cap. Capping occurs more rapidly on transformed fibroblast cell lines than on untransformed cells under identical circumstances (cf. Nicolson and Poste, 1976) and occurs commonly on normal lymphocytes (see below). Capping involves an interaction of the surface receptor molecules (such as lectin recep- tors as well as a variety of other receptors including those for hormones, growth factors, and so on) with elements of the cytoskeleton, discussed later in this chapter (Bourguignon and Bourgui- gnon, 1984). The effectiveness and rapidity of capping, as well as agglutination, which results from cluster formation of lectin-binding sites, is dependent on these and several other factors including the cell density of the culture (Inbar et al., 1977) and the ATP content of the cells (Vlodavsky et al., 1973) but not the ploidy of the cell (Sivak and Wolman, 1972). In addition, the fatty acid and cholesterol composition of the cell membrane also affects the effectiveness and
Figure 16.2 Pathways of ligand (concanavalin A)-induced receptor redistribution on cells. After ligand binding, initially dispersed receptors may remain dispersed or undergo clustering. The clustered receptor- ligand complexes may coalesce or form patches and eventually caps. (After Nicolson and Poste, 1976, with permission of the authors and publisher.)
speed of agglutination of cells (Hill and Borysenko, 1979). This latter characteristic may be re- lated in part to the microviscosity (fluidity-rigidity behavior) of the cell membrane. However, a number of studies using fibroblast-like cells in culture reported various results from increased fluidity of the plasma membrane of transformed fibroblasts to a lowered fluidity of the mem- brane of these cells compared with that of normal or untransformed fibroblasts (cf. Nicolson,1976; Shinitzky and Inbar, 1976). Normal lymphocytes exhibited a greater microviscosity (lesser fluidity) than malignant lymphocytes grown in suspension (Shinitzky and Inbar, 1976). This phenomenon may be related to the cholesterol content of these cells in that normal lympho- cytes contain about twice the amount of this molecular species as lymphoma cells (Shinitzky and Inbar, 1974). Lectin and antibody-induced capping of receptor sites is readily seen in normal lymphocytes, but may not occur under the same circumstances with neoplastic lymphocytes (In- bar et al., 1973). Furthermore, redistribution of surface receptors may occur on lymphocytes in hypertonic medium even in the absence of any ligands (Yahara and Kakimoto-Sameshima,
1977). Thus, the mobility of receptors and related molecules on the surface of various cell types is dependent on a variety of factors, including the valency of the ligand, microviscosity of the membrane, and interaction of the plasma membrane molecule, receptor or otherwise, with intra- cellular components, including elements of the cytoskeleton and other members of the signal transduction pathway (e.g., Graziadei et al., 1990).
After these extensive studies on differential lectin agglutination of normal and transformed cells in culture, some investigators extended these studies into the in vivo situation. Becker (1974) demonstrated a differential lectin agglutination of fetal and malignant hepatocytes com- pared with adult hepatocytes. The latter cells, even after treatment with protease, cannot be ag- glutinated by concanavalin A, whereas fetal liver cells and hepatoma cells are agglutinated by this lectin. Weiser (1972) also demonstrated that intestinal epithelial cells of the human fetus, but not of the adult, can be agglutinated by concanavalin A. These studies suggest that in at least some cell populations the property of agglutination by lectins is another example of the expres- sion of fetal characteristics by neoplastic cells. In both normal and neoplastic keratinocytes, both in vivo and in vitro, the expression of lectin binding by using a variety of different lectins exhib- ited significant differentiation-dependent expression. Normal keratinocytes in culture bound those lectins that neoplastic keratinocytes also bound with the exception of the Ulex europaeus agglutinin I (peanut agglutinin), which was bound to neoplastic keratinocytes but not to normal cells (Suter et al., 1991).
In accord with the changes in lectin-binding sites described above, biochemical studies of the surface membrane of normal and transformed cells in vitro have shown differences. Surface glycoproteins and gangliosides of cells transformed by viruses, chemicals, and x-rays in vitro show significant but not necessarily common differences when compared with nontransformed cells cultured in vitro (Baker et al., 1980; Srinivas and Colburn, 1984; Glick, 1979). Smets et al. (1978) demonstrated that alterations in membrane glycopeptides of transformed cells did not always correlate with anchorage-independent growth in vitro, although tumors derived from cell lines showed a correlation of the two characteristics. Furthermore, the addition of retinoic acid to transformed mouse fibroblasts greatly increases their adhesive properties, which appear to be related to the effect of this vitamin A derivative on the biosynthesis of cell-surface glycoproteins (Sasak et al., 1980). In agreement with these findings is the demonstration by several in- vestigators of lowered levels in transformed cells of a glycosyltransferase involved in the synthe- sis of glycoproteins and gangliosides on the surface membrane (Roth et al., 1974; Patt and Grimes, 1974).
A frequent finding in transformed cells in culture was the absence or marked decrease of a cell surface glycoprotein of molecular weight of about 250,000 (Chen et al., 1976). This protein had been designated as LETS (large external transformation-sensitive protein). The protein was shown to be absent from the surface of cells transformed by some oncogenic DNA viruses. Inter- estingly, when cells were transformed by a temperature-sensitive mutant of the SV40 virus, the LETS protein did not disappear from the surface of infected cells cultured at the nonpermissive temperature (Shopsis and Sheinin, 1976). Another cell surface protein that is greatly reduced in many transformed cells in culture is fibronectin, a large, adhesive protein found in many cells, in connective tissue, and in plasma (Chapter 10). This protein is involved in the maintenance of the cellular cytoskeleton and the interaction of cells with collagen and with other intercellular mo- lecular matrices. It is now apparent that the LETS protein and fibronectin are quite likely one and the same molecular species (cf. McDonagh, 1981). Several secretory glycoprotein species, one having a molecular weight of 35,000 (Gottesman and Cabral, 1981) and another protein family of molecular weight 60,000 to 62,000 (Senger et al., 1980, 1983) were identified and correlated with transformation of both fibroblastic and epithelial cells in culture. While these early findings posed many more questions than they answered, knowledge of characteristics of the surface membrane of the cells in culture in part laid the foundation for our understanding of cell-cell interaction, which has become so important in our understanding of the processes of invasion and metastases during the stage of progression in vivo.