It was almost 20 years after the pioneering studies of Earle and associates that Berwald and Sachs (1963) demonstrated the “transformation” of hamster embryo cells growing in cell cul- tures. These investigators utilized cells dissociated from hamster embryos by proteolytic diges- tion and dispersion of the cells, essentially as single cells. These cells were plated directly onto feeder layers of cells (normal cells having had their growth arrested by radiation or chemical means) in appropriate containers, usually petri dishes, at relatively low density (number of cells per plate) and were allowed to attach and grow. After seeding, appropriate carcinogenic agents were administered either in the medium or as ionizing radiation. The cultures were allowed to grow for 8 to 10 days, and the resultant colonies were fixed and stained. A diagram of the proce- dure is shown in Figure 14.4. Transformed cell colonies were identified by specific morphologi- cal characteristics distinguishing transformed from control or solvent-treated cells. The morphological characteristics demonstrated were those of small fusiform cells that grew in an irregular, random criss-cross pattern and did not appear to exhibit contact inhibition of move- ment or growth, as was found in untreated control cells in culture (Figure 14.5). The lack of “spontaneous” transformation in Syrian hamster embryo (SHE) cells as compared with the rela- tive frequency of spontaneous transformation in mouse (Sanford et al., 1970) and rat cells in culture (Morel-Chany et al., 1985) was apparently critical to the success of this system. On the other hand, human fibroblastic cells rarely exhibited the morphologic transformation character- istics seen in Figure 14.4 (Hayflick, 1977). The mechanism of such spontaneous morphological transformation is not clear, although Rubin and associates (Rubin et al., 1990; Ellison and Rubin, 1992) have proposed several mechanisms based on studies of aneuploid cell lines. “Transformation” of such cell lines is discussed later in this chapter. That “morphological trans- formation” itself may not be a reflection of neoplastic transformation was seen by the studies of Evans et al. (1972), demonstrating that sera from different species and even different sexes
Figure 14.4 Protocol for the assay of morphological transformation of hamster embryo cells by chemi- cals or radiation. Midterm hamster embryos were removed from pregnant uteri, minced, enzymatically dis- sociated, and seeded as single cells on feeder layers (normal cells having had their growth arrested by radiation or chemical means). The newly seeded cells are treated with either radiation or chemicals, and the resultant colonies developing from the hamster embryo cells (normal and transformed) are scored after 8 to
10 days of incubation. (Adapted from Hall and Hei, 1986, with permission of the authors and publisher.)
within the same species resulted in different rates and yields of morphological transformation in culture as well as neoplastic development on inoculation of 106 or more cells into an appropriate mouse host.
Some of the characteristics of the experiments by Berwald and Sachs were quite interest- ing, especially the fact that the incidence of transformation was extremely high, exceeding 80% in some experiments. This argued rather strongly that the mechanism of the transformation by chemicals was not a simple mutation, because of the high incidence of conversion. These au- thors also demonstrated the transformation in vitro of hamster embryo cells (Borek and Sachs,1968) and of human cells (Borek, 1980) by x-irradiation and by infection with oncogenic pa- povaviruses (cf. Sachs, 1974).
Studies by Huberman et al. (1976) indicated that cell transformation in culture is due to a mutagenic event in view of the constant ratio of transformation to mutation to ouabain resistance induced by carcinogenic hydrocarbons in vitro. While these studies did indicate an approximate 20-fold greater rate of transformation than of mutation, studies by Barrett and Ts’o (1978) indi- cated that morphological transformation in primary SHE cell cultures occurred with a frequency from 20 to 500 times higher than mutation at two different loci. This finding strongly suggests that the two processes are not identical. A review by Parodi and Brambilla (1977) of much of the literature on the relation between mutation and transformation in cell culture concluded that there was “an absolute difference between structural mutations and transformation.” Further- more, Elmore et al. (1983) demonstrated that increased mutation rates were not a necessary fac- tor in carcinogen-induced transformation of human fibroblasts. Similarly, Kaden et al. (1989)
Figure 14.5 An artist’s conception of the microscopic appearance of normal and transformed cells in culture. The single-cell layer of spindle cells is characteristic of normal cells, which ultimately exhibit con- tact inhibition of replication, whereas transformed cells exhibit cytologic anaplasia and “piling up” of one cell on another, with lack of contact inhibition of replication.
found no simple correlation between spontaneous mutation rate and the malignant phenotype in Chinese hamster embryo fibroblast lines. The same conclusion had been drawn for the chemical and radiation induction of transformation in the mouse C3H/10T1/2 fibroblast cell line (Landolph and Heidelberger, 1979; Chan and Little, 1982). However, the fact that some struc- tural alteration in DNA is associated with the neoplastic transformation in vitro was evidenced by the finding of Barrett et al. (1978) that ultraviolet irradiation of cells that had incorporated the base analog bromodeoxyuridine into their DNA were transformed. Neither of these two pertur- bations alone induced this transformation. Since it is well known that bromodeoxyuridine incor- poration into mammalian cell DNA induces mutations by itself and that mutagenesis is markedly enhanced by the ultraviolet irradiation of such treated cells, these authors concluded that trans- formation resulted, at least in part, from the ultraviolet-induced structural alteration in DNA con- taining bromodeoxyuridine. Double-stranded DNA breaks induced by electroporated restriction enzymes induced morphological transformation in a mouse cell line, C3H10T1/2 (Borek et al.,1991). Transformation of SHE cells is also accompanied in many instances by the induction of aneuploidy. This is especially notable with both the synthetic estrogen diethylstilbestrol (Tsutsui et al., 1983) and asbestos fibers (Oshimura et al., 1984; Dopp et al., 1995). The normal estrogen,17β-estradiol, was also effective in inducing both cell transformation and numerical chromo- some changes in SHE cells (Tsutsui et al., 1987). However, male sex hormones, both synthetic and natural, exhibited only a very weak transforming effect on SHE cells (Lasne et al., 1990). While diethylstilbestrol may induce mutations at some but not all loci studied, no DNA adducts occurred in SHE cells at concentrations up to 10 µg/mL (Hayashi et al., 1996). Changes in the 32P-postlabeling pattern of DNA adducts did occur in SHE cells in the presence of estradiol and some of its metabolites (Hayashi et al., 1996). Thus, while there is evidence that DNA damage can be correlated with morphological transformation, the quantitative relationships between mu- tation and transformation are not identical.
Studies by Heidelberger (1973) with the prostate system confirmed and extended these investigations. In Heidelberger’s studies, the organ cultures treated with hydrocarbons were dis-persed into cell suspensions and transformed cell lines were produced. DiPaolo et al. (1971a) confirmed the findings of Berwald and Sachs and demonstrated that, by inhibiting the cellular toxicity of carcinogenic hydrocarbons in hamster embryo cell cultures with various noncarcino- genic chemicals, one may dissociate the transforming property from the toxic metabolic prop- erty of carcinogenic hydrocarbons (DiPaolo et al., 1971b). The cytotoxicity of carcinogenic polycyclic hydrocarbons for cell cultures appears to depend on the presence and inducibility of the microsomal aryl hydroxylase complex, the regulation of which is mediated through the prod- uct of the Ah locus (Chapter 7). Utilizing human mammary epithelial cells, Stampfer et al. (1981) demonstrated that inhibition of growth of such cells was 50 to 100 times more sensitive to the presence of benzo[a]pyrene than were cultured fibroblasts. Nebert and associates (cf. Benedict et al., 1972) have shown that the metabolism of the hydrocarbon is necessary for cyto- toxicity in that the inhibitors of hydrocarbon metabolism also inhibit the toxicity. Furthermore, Peterson and associates (1979) have demonstrated that the mutagenic lesion(s) produced by car- cinogenic agents in cell culture is not associated with the cytotoxicity of the agent. On the other hand, Poiley et al. (1980) demonstrated that in a large number of normal hamster embryo cell preparations, those cells that were consistently more easily transformed by polycyclic hydrocar- bons had consistently higher levels of aryl hydrocarbon hydroxylase and that the enzyme could be induced to a higher level by treatment with appropriate agents.
In addition to the systems mentioned above, which are concerned primarily with mesen- chymal cell transformation, a number of differentiated epithelial cell types derived either from endoderm or ectoderm (Chapter 2) have been reported. These include cells derived from rat liver (e.g., Borenfreund et al., 1975; Montesano et al., 1980; Williams et al., 1973), mouse mammary epithelial cells (Miyamoto et al., 1988), rat tracheal epithelial cells (Thomassen et al., 1983), rat bladder (Hashimoto and Kitagawa, 1974), rabbit bladder (Summerhayes et al., 1981), and epi- dermal keratinocytes (Slaga et al., 1978; Yuspa et al., 1983b; Fusenig and Boukamp, 1998). In primary cultures of hepatocytes, the usual chemical hepatocarcinogens may not be effective for their conversion to proximate and ultimate forms because of the loss of phase I activity (Guil- louzo, 1986). On the other hand, polycyclic hydrocarbons and alkylating agents as well as afla- toxin (Schaeffer and Heintz, 1978) and ethionine (Brown et al., 1983) did reportedly induce “transformation” of rat liver epithelial cells in culture. Perhaps the best-characterized of the epi- thelial transformations in vitro is that of keratinocytes, both rodent and human. Fusenig and Boukamp (1998), on the basis of studies of a spontaneously immortalized cell line obtained from normal human keratinocytes maintained in culture for an extended period, have proposed a multistage model for the development of normal keratinocytes into malignant neoplastic cells because of changes seen both in vivo and in vitro (Figure 14.6). Such a series of alterations are analogous to those seen in the stage of progression in both human and rodent neoplastic devel- opment (Table 9.12, Figure 9.8). In this model, however, the conversion of the immortal cell (HaCaT) occurring spontaneously in normal keratinocyte cultures to a “benign cell” requires transfection with the H-ras cellular oncogene. A major finding allowing for such experiments was the demonstration that high levels of calcium ions induced terminal differentiation of mouse epidermal keratinocytes, but continued maintenance may occur at decreased calcium concentra- tions. Treatment with several different types of chemical carcinogens induced the formation of cells that were resistant to the high calcium-induced differentiation of the keratinocytes (Kilkenny et al., 1985).
In general, transformation of mouse keratinocytes by a variety of chemical agents may be readily accomplished, but the transformation of human cells (either mesenchymal or epithelial) to cells capable of neoplastic growth in immunosuppressed animal hosts has met with much less success (McCormick and Maher, 1988; Kuroki and Huh, 1993; Holliday, 1996). Previous results demonstrating the growth of human cells in soft agar (anchorage independence; see below) may
Figure 14.6 Schematic for “stages” of the transformation of normal human keratinocytes to metastatic cells in vivo. The conversion of the immortal cell to the benign cell requires transfection of the H-ras cellu- lar oncogene as well as several other cultural characteristics noted in the figure. The steps thereafter can be demonstrated in vivo as well as in a special model in vitro system of these authors. (Adapted from Fusenig and Boukamp, 1998, with permission of the authors and publisher.)
not be a good method to characterize chemically induced transformation of human cells in vitro (Peehl and Stanbridge, 1981). On the other hand, several investigators (Milo et al., 1996; Gruen- ert et al., 1995; Stampfer and Bartley, 1985) have reported transformation and the formation of immortalized cell lines from both mesenchymal and epithelial human cells. Grafström (1990) reviewed such studies in human epithelial tissues in vitro. The mechanism for this peculiar “re- sistance” of human cells to transformation in vitro, despite the ability of such cells to metabolize polycyclic hydrocarbons and even form DNA adducts with their ultimate forms, is still not com- pletely understood.