Although it is clear that genomic instability is the ultimate basis for the development of the stage of progression and its primary consequences, invasion and metastasis, one may ask whether spe- cific genes and their products are involved in these last two processes. The critical nature of the expression of cell adhesion components, proteinases, and coagulation mechanisms of impor- tance in invasion and metastasis has already been discussed. However, it is also apparent that specific genes and their products not directly involved in invasive and metastatic processes also influence, both directly and indirectly, these operations. Some of these genes are listed in Table 10.4, along with their functions and the correlation, either positive or negative, with metastatic potential. The Wnt-1 gene was found disrupted in several but not all primary mammary carcino- mas in mice induced by the mouse mammary tumor virus through a mechanism of insertional mutagenesis (Chapters 4 and 8). However, no disruption or alteration of the gene was seen in any metastatic lesions, indicating that alteration of the gene either inhibited or was not relevant to the appearance of metastases (Sarkar, 1990). The mts-1 (p9Ka) gene, which codes for a calcium binding protein, is expressed at high levels in rat metastatic mammary tumor cell lines but not in benign neoplasms (cf. Ponta et al., 1994). In contrast, expression of the wild-type nm23 gene appears to inhibit the development of metastases, while mutation (Wang et al., 1993) or reduced levels of expression of the gene are related to an increased potential for metastases in most (Bev- ilacqua et al., 1989) but not all (Russell et al., 1997) carcinomas. In the case of the L-myc gene, a particular polymorphism was correlated with extensive metastases of lung cancer, especially
Figure 10.12 Incidence of gene amplification involving up to three genes in nonmetastatic and metastatic breast neoplasms. (Adapted from Donovan-Peluso et al., 1991, with permission of the authors and publisher.)
adenocarcinomas in the human (Kawashima et al., 1988). Hendrix and associates (1992,1996) have also demonstrated a correlation of the expression of the intermediate filament proteins vi- mentin and keratins K8 and K18 (Chapter 16) with the metastatic potential of experimental mel- anoma cells. Higher levels of expression of these genes in transfected melanoma cells as well as other naïve neoplastic cells exhibited a higher degree of metastatic and invasive potential. Not listed in the table but of great significance is the suggestion by Glinsky et al. (1997) that meta- static cells exhibit a greater resistance to apoptosis than nonmetastatic neoplastic cells. Concom- itant with this is a diminished level of nuclear calcium-dependent endonucleases and reduced activity of specific caspases.
Table 10.4 shows that most of the examples given are genes whose products function ei- ther in signal transduction pathways or as transcription factors, and all are associated with an interaction of the cell with its environment. Studies have also been undertaken wherein genes, especially those involved in signal transduction, are transfected into cells with a low metastatic potential to determine whether artificial disruption of signal transduction pathways would en- hance metastatic growth. This was found to occur from studies by Liotta and colleagues (cf. Muschel and Liotta, 1988). Their investigations demonstrated that transfection of a mutant ras oncogene into mouse fibroblasts in culture resulted in increased expression of the transfected gene and enhanced expression of the endogenous ras gene. Both of these parameters correlated with a dramatic increase in metastatic growth of the transfected cells in vivo, usually in nude (immunocompromised, Chapter 19) mouse hosts. A striking example of this correlation is seen in Figure 10.13, where transfection of a mutant ras gene (EJ-ras) or wild-type ras (cHras) with the mutant v-myc oncogene into rat embryo fibroblasts showed a dramatic increase in lung me- tastases in nude mice and also a dramatic increase in the activity of collagenase. Transfection of the mutant EJ-ras gene, together with a viral oncogene E1A, resulted in no metastases or en- hanced collagenase activity, although the resultant transfectants were tumorigenic in the immu- nocompromised host (Garbisa et al., 1987), suggesting that collagenase production was necessary for successful metastatic growth but not neoplasia. While these authors presented evidence that some of the transfectants exhibited little or no aneuploidy, other studies indicated that ras transfection of mouse cell lines resulted in enhanced aneuploidy of the metastatic cells (Ichikawa et al., 1990) and that ras transfection accelerates rather than initiates formation of
Figure 10.13 Transfection of the mutant oncogene EJ-ras or the wild-type proto-oncogene cHras with the viral oncogene vmyc into diploid rat embryo fibroblasts (REFs) induces collagenase activity and me- tastasis formation. In contrast, transfection with mutant ras (EJ-ras) together with the adenovirus oncogene E1A eliminates the metastatic capability and increase in collagenase activity, but the cells remain fully neo- plastic. The correlation between metastatic potential and expression of type IV collagenase is apparent. Numbers and letters in parentheses indicate the designation of the specific cell line under study. (Modified from Liotta and Kohn, 1990, with permission of the authors and publisher.)
metastases (Greig et al., 1985). Furthermore, in other rodent cells, transfection of the ras oncogene did not result in enhanced metastatic potential (Baisch et al., 1990; Tuck et al., 1990). As a further complication in the interpretation of such transfection experiments, Kerbel et al. (1987) reported that exposure of SP1, a nonmetastatic mouse mammary adenocarcinoma cell line, to calcium phosphate alone in the absence of DNA resulted in an increase in clones mani- festing metastatic properties. Furthermore, transfection of DNA may induce chromosome re- arrangements and other types of mutations as well as epigenetic changes in recipient cells (Bardwell, 1989).
That enhanced ras expression from either the endogenous gene or a transfected gene was not the only factor involved in enhanced metastatic capability of such cells was demonstrated by Ichikawa et al. (1992), who found that suppression of metastatic ability could occur by genetic means even in the presence of enhanced expression of the transfected ras oncogene. Further- more, transfection of other oncogenes in NIH 3T3, the standard cell line utilized, also resulted in an enhanced potential for metastases (Egan et al., 1987; Greenberg et al., 1989). Transfection of p53 genomic clones into murine carcinoma cells exhibiting a low metastatic capacity led to an enhancement of the expression of the p53 protein together with increased metastatic potential of these cells (Pohl et al., 1988). In analogy to the original mutant ras oncogene studied by Liotta and colleagues, a single point mutation in the met proto-oncogene eliminated metastatic poten- tial but not transformation to neoplasia of rat fibroblasts (Giordano et al., 1997). Although it is somewhat difficult to draw any general conclusions from these transfection experiments, it is clear that, under certain experimental circumstances, transfection of specific genes does enhance the metastatic potential of the cells being studied. In view of the functional characteristics of the small sample of endogenous genes that affect metastatic potential, as noted in Table 10.5, the types of genes used for the transfection experiments have functions overlapping with those indi- cated in the table, e.g., signal transduction. Thus, it is very likely not only that the metastatic phenotype is characterized by alterations in cell adhesion molecules and proteinases but also that alteration of signal transduction pathways plays a major role in the development of the success- ful metastatic phenotype.