Molecular mechanisms of transformation of C3H/10T1/2 C1 8 mouse embryo cells and diploid human fibroblasts by carcinogenic metal compounds.

Carcinogenic arsenic, nickel, and chromium compounds induced morphological and neoplastic transformation but no mutation to ouabain resistance in 10T1/2 mouse embryo cells; lead chromate also did not induce mutation to ouabain or 6-thioguanine resistance in Chinese hamster ovary cells. The mechanism of metal-induced morphological transformation was likely not due to the specific base substitution mutations measured in ouabain resistance mutation assays, and for lead chromate, likely not due to this type of base substitution mutation or to frameshift mutations. Preliminary data indicate increases in steady-state levels of c-myc RNA in arsenic-, nickel-, and chromium-transformed cell lines. We also showed that carcinogenic nickel, chromium, and arsenic compounds and N-methyl-N-nitro-N-nitrosoguanidine (MNNG) induced stable anchorage independence (Al) in diploid human fibroblasts (DHF) but no focus formation or immortality. Nickel subsulfide and lead chromate induced Al but not mutation to 6-thioguanine resistance. The mechanism of induction of Al by metal salts in DHF was likely not by the type of base substitution or frameshift mutations measured in these assays. MNNG induced Al, mutation to 6-thioguanine resistance, and mutation to ouabain resistance, and might induce Al by base substitution or frameshift mutations. Dexamethasone, aspirin, and salicylic acid inhibited nickel subsulfide, MNNG, and 12-O-tetrade-canoylphorbol-13-acetate (TPA)-induced Al in DHF, suggesting that arachidonic acid metabolism and oxygen radical generation play a role in induction of Al. We propose that nickel compounds stimulate arachidonic acid metabolism, consequent oxygen radical generation, and oxygen radical attack upon DNA.(ABSTRACT TRUNCATED AT 250 WORDS) ImagesFigure 1.

Significant effort has gone into studying the molecular mechanisms of metal carcinogenesis, and the molecular basis of metal carcinogenesis is beginning to be understood. Metal compounds induce chromosomal aberrations in cultured mam-malian cells (28,29); however, except for hexavalent chromium compounds and platinum coordination complexes, carcinogenic metal compounds are inactive or at best weakly active in bacterial and mammalian cell mutagenesis assays (7,8,30). The in vitro mammalian and bacterial mutagenesis assays currently available do not effectively detect the DNA alterations caused by carcinogenic metal salts (28,29). In this manuscript, I review recent studies from my laboratory on induction of morphological transformation in C3H/1OTl/2 Cl 8 (1OTl/2) mouse embryo cells and anchorage independence in diploid human fibroblasts by carcinogenic metal salts and hypothesize the possible mechanisms of metal salt-induced cell transformation.

Materials and Methods
C3H/1OTl/2 Cl 8 mouse embryo fibroblasts were cultured according to the meth-ods of Reznikoff et al. (31). Assays were conducted to quantitate chemically induced cytotoxicity and morphological transformation (23,24,32,33,34). Assays to detect chemically induced mutation to ouabain resistance in 1OTl/2 cells were performed as previously reported (23,24,33). Diploid human fibroblasts were derived from circumcised human foreskins and cultured according to the methods of Biedermann and Landolph (26,24. Assays to detect metal-induced cytotoxicity, anchorage independence, and mutation to 6-thioguanine resistance and ouabain resistance were also as described by Biedermann and Landolph (26,27).

Resufts
Induction ofMorphological Transformation of 1OT1/2 Cells by CartinogenicArsenic, Nickel, and Chromium Compounds 10T1/2 cells are an aneuploid, immortal mouse cell line derived from embryos of C3H mice that are contact inhibited, nontumorigenic, and have a low frequency of spontaneous transformation (31). These cells exhibit a high frequency of morphological transformation when treated with chemical carcinogens or radiation (7,8,23,24,(32)(33)(34)(35). They are also a useful model cell culture system used to study the mechanisms of chemically induced neoplastic transformation and to detect carcinogens by their ability to induce morphological transformation of these cells (7,8,34,35).
We have been studying the ability of carcinogenic metal salts to induce morphological transformation of 10T1/2 cells to gain insight into the mechanisms of metal carcinogenesis. Our cumulative studies indicate that carcinogenic arsenic, nickel, and chromium compounds induce morphological transformation of 1OT1/2 cells (7,8).
We first examined the ability of 1OT1/2 cells to phagocytose carcinogenic insoluble nickel compounds. We found that 10T1/2 cells readily phagocytose nickel subsulfide particles, and the phagocytosed particles can be readily seen in phagocytic vesicles (Figure 1). The phagocytosis of nickel subsulfide and crystalline nickel monosulfide by 1OT1/2 cells occurs when low concentrations of these compounds (0-50 pM) are added to the medium ( Figure 2A) (24). Over the concentration range that phagocytic uptake of nickel compounds occurs, we also observed a concentration dependent cytotoxicity in 1OT1/2 cells treated with the insoluble nickel compounds nickel subsulfide and nickel monosulfide ( Figure 2B), With nickel oxide. we did not observe discrete phagocytic vesicles, but nevertheless, we observed cytotoxicity in 1OT1/2 cells treated with from 0 to 1 mM nickel oxide ( Figure 28) (24).
We next studied the ability of carcinogenic nickel compounds to induce morphological transformation in 10T1/2 cells (24). The insoluble carcinogenic nickel compounds nickel subsulfide, nickel monosulfide, and nickel oxide (greenish preparation) all induced dose-dependent transformation in 10T1/2 cells, over the same concentration ranges that were phagocytosed and induced cytotoxicity ( Table 1). This indicated that phagocytosis of particulate nickel compounds is an initial step in the processes of cytotoxicity and cell transformation in 10T1/2 cells (24). Similar observations have been made by Costa and co-workers in Syrian hamster cells (17,18). Nickel subsulfide induced primarily type II foci, and cell lines derived from these foci did not grow in soft agarose ( Table 2). Nickel monosulfide-induced largely type II and occasionally type III foci, and three nickel monosulfide induced type II and type III foci gave rise to cell lines that grew in soft agarose (24). Nickel oxide induced type II and occasionally type III foci, and one of these type II and type III foci gave rise to a cell line that grew in soft agarose and formed fibrosarcomas in nude mice (24).
Conversely, the soluble nickel compounds nickel sulfate and nickel chloride, which are noncarcinogenic or substantially less carcinogenic than the insoluble nickel compounds, did not induce morphological transformation of 1OT1/2 cells (7,8,24).
The inability of soluble nickel compounds to induce cell transformation is consistent with whole animal carcinogenicity studies (1,2,4,5,(7)(8)(9)(10)(13)(14)(15)(16). This indicates that results for induction of morphological transformation in 1OT1/2 cells with nickel compounds correlates with the results of whole animal carcinogenesis assays using nickel compounds. Our cell transformation responses with nickel compounds are therefore specific responses.
We previously developed an assay detecting chemically induced (33) stable (36), specific (37) base substitution mutations (38) to ouabain resistance that results from a mutation in a gene encoding (Na,K)-ATPase activity on murine chromosome 3 (39) and confers a ouabain-   resistant (Na,K)-ATPase activity (35,40). We next studied the ability of nickel compounds to induce mutations in this assay. Concentrations of nickel subsulfide and nickel oxide that induced cytotoxicity and morphological transformation in 10TI/2 cells did not induce base substitution mutations to ouabain resistance in 10T1/2 cells (Table 1) (24). This indicates nickel compounds likely do not induce morphological transformation by inducing the specific, restricted type of base substitution mutations that are detected in an assay for mutation to ouabain resistance (24). Recent preliminary studies from our laboratory indicate that nickel monosul- fide-and nickel oxide-transformed 1OT1/2 cell lines frequently express higher steadystate levels of c-myc RNA ( Table 2). We hypothesize that transformation of 1OT1/2 cells by nickel compounds results in small deletions or small mutations in the 3' or 5' regulatory regions of the c-myc and c-H-ras protooncogenes, leading to activation of these protooncogenes via this mechanism and hence a stabilization of the protooncogene RNAs and the increased steady-state levels and increased half-lives of these mRNAs that we observe. Studies to test this hypothesis critically in our laboratory are in progress (T Sakuramoto, D Evans, T Miura, S Boone, and JR Landolph, unpublished data).
We were also able to show that lead chromate induced a low but dose-dependent and reproducible frequency of type III morphological transformation (23), the strongest type of transformation in 1OT1/2 cells (32,34). The transformed cells stably maintained a focus-forming phenotype, grew in soft agarose, and formed fibrosarcomas when injected into nude mice (23). Lead chromate-treated cells had a large number of vacuoles and extruded cytoplasm over the particles of lead chromate, apparently in an attempt to phagocytose lead chromate particles (23). Calcium chromate, potassium dichromate, and strontium chromate did not induce morphological transformation in lOT1/2 cells (23). Strontium chromate was slightly soluble in culture and eventually dissolved and calcium chromate did not induce transformation even when added to the cells as a particulate in acetone suspension (23). This indicated that unique physicochemical properties of insoluble lead chromate particles were responsible for its uptake, likely by phagocytosis, and its ability to induce cytotoxicity and cell transformation (23), and correlated with earlier observations that the slightly soluble hexavalent chromlum compounds are also carcinogenic in animal bioassays (4,5,7,8).
We showed that lead chromate induced morphological transformation in 10T1/2 cells but not mutation to ouabain resistance in 10TI/2 or CHO cells nor mutation to 6-thioguanine resistance in CHO cells (23). In addition, calcium chromate did not induce morphological transformation in 1OT1/2 cells or mutation to ouabain resistance in 10T1/2 or CHO cells but did induce mutation to 6-thioguanine resistance in CHO cells (23). Hence, we speculated that lead chromate induced morphological transformation by a mechanism not involving the specific type of base substitution mutations or frameshift mutations detectable in assays for mutation to ouabain resistance or 6-thioguanine resistance. Most recently, we obtained preliminary evidence that in two lead chromate-transformed 1 OT 1/2 cell lines, the steady-state levels of c-myc RNA are 6to 8-fold higher, and the half-life of c-myc RNA is increased in these cell lines (M Dews and JR Landolph, unpublished data). Our current working hypothesis is that lead chromate induces mutations or small deletions in the 5' or 3' regulatory regions of the c-myc gene, leading to a c-myc RNA with an increased half-life. This may be due either to deletions or mutations generated by the reduction of Cr(VI), leading to either Cr(V) or oxygen radicals generated by intracellular reduction of Cr(VI).
Third, we observed that sodium arsenite induced a low but reproducible yield of morphological transformation in 1OTI/2 cellsUR Landolph and C Troesch, unpublished data). Cloning of arsenitetransformed cell lines yielded cell lines that formed type II and type III foci, grew in soft agarose, and formed fibrosarcomas in nude mice. We also found that sodium arsenite promoted cell transformation initiated by 3-methylcholanthrene. Pentavalent sodium arsenate and potassium arsenate did not induce morphological transformation, indicating that this transformation was specific for the trivalent state of arsenic UR Landolph and CT Troesch, unpublished data). These data are consistent with epidemiological studies indicating that arsenic compounds are carcinogenic to humans (1)(2)(3)6,8,12). A resolution to the apparent paradox of the noncarcinogenicity of arsenic in animal bioassays could be that sodium arsenite is a weak cell-transforming agent and a promoting or co-carcinogenic agent that is not easily detected in the relatively insensitive animal bioassays. A sec-Volume 102, Supplement 3, September 1994 ond possibility is that humans are more sensitive to arsenic-induced carcinogenesis than are rats and mice. Preliminary work from our laboratory has first ruled out the hypothesis that carcinogenic metal compounds induce morphological transformation in lOTl/2 cells by inducing amplification or gross rearrangement of known protooncogenes such as c-myc, converting them into oncogenes ( Table 2). We are now testing a second hypothesis that carcinogenic metal compounds cause mutations or small deletions in the 5' or 3' regulatory regions of specific protooncogenes, converting them into activated oncogenes. We are also testing a third hypothesis that carcinogenic metal salts cause mutational inactivation or deletion of tumor suppressor genes, thereby inactivating them. To date, we have found preliminary evidence for increased steadystate levels of c-myc RNA in lead chromate, nickel oxide, and sodium arsenite transformed lOTl/2 cell lines. We have not found amplification or rearrangements of the c-myc genes in these cell lines (T Sakuramoto, T Miura, M Dews, J Lillehaug, and JR Landolph, unpublished data). We are currently studying whether there are small mutations or deletions in the 3' or 5' regulatory regions of this gene in these transformed cell lines that might account for their activation and higher steady-state levels of the c-myc transcripts ( Table 2).

Induction ofAnchorage Independence in Diploid Human Fibrobiasts
Recently, we began to study the molecular mechanisms by which carcinogenic metal salts induce transformation of cultured diploid human fibroblasts. As a source of human fibroblasts, we used circumcised human foreskins. Previous studies have shown that organic carcinogens could induce anchorage independence in cultured human fibroblasts (41)(42)(43). We showed that nickel compounds [nickel subsulfide, nickel acetate, and nickel subsulfide ( Figure 3) (26), hexavalent chromium compounds (lead chromate, potassium dichromate, calcium chromate, and chromium trioxide) (26,27), and arsenic compounds (sodium arsenite and sodium arsenate, (26)] induced dose-dependent anchorage independence (AI) in diploid human fibroblasts. Metal-induced AI was a stable phenotype (26,2X). To date, we have not observed other transformation phenotypes, such as morphological transformation. All the metal-induced anchorage-independent cell strains had saturation densities comparable to those of normal human fibroblasts, and they all eventually senesced (26,27). The specificity of carcinogenic metal salt-induced Al is indicated by the fact that manganese chloride, mercuric acetate, and calcium chloride, which are not carcinogenic, did not induce AI (27).
Nickel subsulfide did not induce mutation to ouabain resistance or to 6-thioguanine resistance (26) and lead chromate did not induce mutation to 6-thioguanine resistance (27) at concentrations that were cytotoxic, and induced AI in diploid human fibroblasts. Hence, nickel subsulfide and lead chromate likely induced anchorage independence by mutations of the type not easily measured in assays for 6thioguanine resistance. We speculate that metal-induced oxygen radical generation and consequent radical-induced mutations might be part of the mechanism of nickel subsulfide and lead chromate-induced Al in diploid human fibroblasts. Calcium chromate, potassium dichromate, and MNNG did induce mutation to 6-thioguanine resistance and AI over the same concentration ranges, indicating that mutation might be a mechanism by which these compounds induced AI (27).
Recently, we have begun to test the hypothesis that stimulation of arachidonic acid metabolism and consequent generation of oxygen radicals are part of the molecular mechanism by which metal compounds induce AI in diploid human fibroblasts. We have found that nickel subsulfide-induced AI in human fibroblasts was inhibited by dexamethasone, aspirin, and nordihydroguaieretic acid. Similarly, MNNG-induced AI was also inhibited by these three inhibitors of arachidonic acid release (dexamethasone) and metabolism (aspirin, which inhibits cyclooxyenase activity) and oxygen radical persistence [salicylic acid, which scavenges oxygen radicals (40)]. In addition, the tumor promoter TPA induced Al only when applied to the soft agar, and the TPA-induced anchorage-independent phenotype was reversible and only manifested in the presence of TPA. However, induction of AI by TPA was also inhibited by dexamethasone and aspirin (RA Biedermann, JO Nwankwo, J Weng, and JR Landolph, unpublished data; summarized in Table 3).

Conclusions
It is very clear that the mechanisms of metal carcinogenesis are complex and only beginning to become understood, and that the mechanisms of carcinogenesis for each metal are specific to that metal (44). With Environmental Health Perspectives these considerations in mind, however, a number of conclusions and interesting speculations may be drawn from the work in our laboratory on mechanisms of metalinduced cell transformation. For the carcinogenic insoluble nickel compounds such as nickel subsulfide, we (24) and others (17,18) have evidence that these compounds are phagocytosed and that large amounts of these compounds are therefore taken up into individual murine (24), hamster (17,18), and human (26) fibroblasts. In addition, our recent preliminary results indicate that inhibitors of arachidonic acid release and its metabolism by cyclooxygenase and an oxygen radical scavenger (salicylic acid) inhibit nickel subsulfide-induced AI in diploid human fibroblasts. We therefore speculate that nickel subsulfide induces membrane perturbations that activate the prostaglandin synthesis cascade, resulting in generation of oxygen radicals. This could explain our inability to measure mutation to ouabain resistance in 1OT1/2 cells (24) or mutation to ouabain resistance or to 6-thioguanine resistance in diploid human fibroblasts (26) following treatment of these cells with nickel subsulfide, since oxygen radicals do not induce the type of mutation that is easily measured in these assays. At this time, we cannot assess the fraction of nickel-generated oxygen radicals that derive from stimulation of arachidonic acid metabolism pathway versus the fraction that derive from protein-bound nickel ions that bind to DNA and generate oxygen radicals, since it is known that nickel ions bind to proteins that can then bind to DNA (9,45,46). Second, we also have preliminary evidence that there are increased steady-state levels of c-myc RNAs in nickel oxide and nickel monosulfide transformed 1OT1/2 mouse embryo cells (T Sakuramoto, T Miura, D Evans, and JR Landolph, unpublished data). We speculate that this may result from nickel compound-generated oxygen radicals, which may induce mutations in the 3' or 5' regulatory regions of these protooncogenes, leading to enhanced stability of these RNAs and contributing to the induction and maintenance of the transformed phenotype. We know that multiple activated oncogenes may cooperate in cell transformation (47)(48)(49), and we are studying how many oncogenes are activated in nickel transformed cell lines.
We also hypothesize that nickel ions generate oxygen radicals, and that these oxygen radicals are responsible for inacti-vating tumor suppressor genes (50), both known tumor suppressor genes such as the retinoblastoma gene, Rb (51) and the p53 suppressor gene (54, suppressor genes that are just becoming identified and understood such as the suppressor gene on human chromosome 11 (53), and novel suppressor genes just being discovered, such as the senescence-mediating suppressor gene that is a target for nickel-induced cell transformation discovered by Costa's group (54). We also speculate that nickel compounds will both inactivate genes such as the p53 gene and also activate it to a dominantly acting negative oncogene. Further work is in progress in our laboratory and in other laboratories to test these hypotheses.
It still is unclear as to what the active ionic species are that induce chromium carcinogenesis. Chromate induces DNA crosslinks in rat liver and kidney (55) and binds to chromatin and DNA (56). There is much current speculation as to whether Cr(VI) is reduced to Cr(V), which is a proximate carcinogen, or whether it is reduced to lower oxidation states, which might generate oxygen radicals (44). Our work with the strong carcinogen lead chromate (57) has given preliminary results that lead chromate transformed lOT1/2 cell lines have higher steadystate levels of c-myc RNA (M Dews and JR Landolph, unpublished data). We are testing the hypothesis that these higher steadystate levels of c-myc RNA are due to mutations in the 3' or 5' regulatory regions of the c-myc gene, leading to a longer halflife of c-myc RNA and contributing to maintenance of the transformed state. Our results with lead chromate indicate that it does not induce mutation to ouabain resistance in 10T1/2 or CHO cells or mutation to 6-thioguanine resistance in CHO cells (23). Hence, we speculate that this compound may generate Cr(V) and/or generate oxygen radicals that induce mutations or small deletions in the 3' or 5' regulatory regions of the c-myc gene activating it to an oncogene, by leading to an increased halflife of c-myc RNA. Work is in progress in our laboratory to test this hypothesis. Finally, we have also found preliminary evidence for increased steady-state levels of c-myc expression in sodium arsenite transformed cell lines. There was no amplification of the c-myc gene in arsenite-transformed 10T1/2 cell lines UR Lillehaug, D Evans, and JR Landolph, unpublished data). This is interesting, because Lee et al. have found that arsenite induced amplification of the dihydrofolate reductase gene in mouse 3T6 cells (58).
Since we have found that sodium arsenite does not induce mutation to ouabain resistance in 1OT1/2 cells (R Landolph and C Troesch, unpublished data), we again postulate that arsenite induces mutations in 3' or 5' regulatory regions of the c-myc gene in these transformed cell lines. The precise mechanism by which arsenite induces such mutations is not clear. Work in our laboratory is currendy in progress to determine the type of mutations arsenite induces in the cmyc gene that contributes an increased halflife of c-myc RNA and to maintenance of the transformed phenotype (M Dews and JR Landolph, unpublished data).
In the case of arsenic-, nickel-, and chromium-induced cell transformation, we postulate that these compounds also cause mutation and/or deletions in tumor suppressor genes that inactivate these genes. We postulate that a combination of mutation or deletion induces activation of single or multiple protooncogenes into oncogenes, plus a mutational or deletional inactivation of tumor suppressor genes, is part of the mechanism of metal carcinogenesis.