Sarcoma-producing cell lines derived from clones transformed in vitro by benzo[a]pyrene.

Cells derived from mixed Syrian hamster embryo cultures were treated with pyrene (control) or with benzo[a]pyrene. Transformed clones were obtained only with the carcinogen. Some of the transformed clones were responsible for cell lines that produced tumors when injected into hamsters. These observations provide evidence that chemical-induced oncogenesis can be studied by an in vitro model.

A transforming ras gene in tumorigenic guinea pig cell lines initiated by diverse chemical carcinogens.

Fetal guinea pig cells were transformed by treatment with four different chemical carcinogens including nitroso compounds and polycyclic hydrocarbons. As a consequence of this treatment, oncogenes capable of transforming NIH/3T3 cells became activated in each of five independently established clonal guinea pig cell lines. Molecular characterization of representative NIH/3T3 transformants revealed that the same oncogene was present in each of the cell lines tested. Moreover, detection of this transforming gene paralleled the acquisition of tumorigenic properties by these neoplastic cells.

Differential early viral gene expression in two stages of human papillomavirus type 16 DNA-induced malignant transformation.

Human papillomavirus (HPV) type 16 DNA induces progressive transformation in NIH 3T3 cells. Two types of cell lines, PM3T3G0 and PM3T3Fo, were isolated by G418 or focus selection, respectively, after transfection of cells by a recombinant HPV 16 DNA carrying the neo gene. These cell lines exhibited distinct phenotypes compared with controls. Saturation densities of PM3T3G0 and PM3T3Fo lines were two- to three- and five- to sevenfold greater than that of control NIH 3T3 cells, respectively. Neither cell type required high serum for growth, in contrast to NIH 3T3 cells. PM3T3G0 lines were premalignant, whereas PM3T3Fo lines manifested tumorigenicity within 2 weeks. Subpopulations of three PM3T3G0 lines underwent progressive transformation as reflected by focus formation. Analysis of HPV 16-specific mRNA species demonstrated that high levels of early and late gene expression were detected in premalignant PM3T3G0 lines, whereas relatively low quantities of selected gene messages were expressed in malignant transformants. Thus, high levels of viral gene expression are not crucial for malignant transformation.

Quantitative in vitro transformation of Syrian golden hamster embryo cells with the use of frozen stored cells.

Quantitative transformation by carcinogens of Syrian golden hamster embryo cells from primary, secondary, or tertiary cultures could be obtained from cell pools frozen prior to any culturing with the same efficiency as with fresh cells. Cells retained the ability to activate a wide range of chemical carcinogens when the noncultured hamster cells were cooled at a controlled rate and when the established procedures for culturing cells were followed. The number of experiments that could be done with cells derived from frozen hamster pools could be increased by substitution of a hamster cell line for the feeder layer. The response to chemical carcinogens of different classes including carcinogenic hydrocarbons, aromatic amine derivatives, metal complexes, aflatoxin B1, N-methyl-N'-nitro-N-nitrosoguanidine, and UV irradiation was similar to that reported with fresh cells. Transformation rate increased with increasing carcinogen concentration. The factors important for the breeding of healthy animals and for selection of those appropriate for obtaining reproducible transformation experiments were enumerated.

Equivalency of endothelial cell growth supplement to irradiated feeder cells in carcinogen-induced morphologic transformation of Syrian hamster embryo cells.

Endothelial cell growth supplement (ECGS), an extract of bovine neural tissue with growth-promoting properties for human endothelial and epithelial cells and for mouse BALB/c fibroblast-like cells, can be substituted for feeder cells in a quantitative 7-day Syrian hamster embryo cell colony in vitro model of carcinogenesis. Inclusion of 50 or 100 micrograms ECGS/ml medium throughout the 7-day growth period produced results equal to those obtained with feeder cells. The frequency and morphology of normal fibroblast colonies and carcinogen-induced morphologically transformed cell colony growth in the presence of ECGS were similar to those in the presence of feeder cells. A positive dose-response relationship in transformation by benzo[a]pyrene occurred. The frequency of transformed colonies following UV irradiation and treatment of the cells with the cocarcinogenic tumor promoter 12-O-tetradecanoylphorbol 13-acetate was greatly augmented, and lymphotoxin, a lymphokine with anticarcinogenic activity, reduced transformation. Thus ECGS can substitute for feeder cells in supporting in vitro transformation and eliminates a potential complex source of variability for studies where interaction(s) with feeder cells are a consideration. The mechanics of this model system was simplified, and its versatility for the study of physiologic, carcinogenic, and other pathophysiologic processes was broadened.

Independent expression of chemical carcinogen-induced phenotypic properties frequently associated with the neoplastic state in a cultured guinea pig cell line.

Relationships among carcinogen-induced morphlogic transformation, anchorage independence, susceptibility to growth inhibition by the immunologic hormone lymphotoxin and to natural killer cell and natural lymphoid cell cytotoxicity, natural delayed-type (tuberculin) skin reactivity, and tumorigenicity were investigated with the use of a guinea pig cell culture model of carcinogenesis characterized by sequential, extended developmental stages. Twelve cultures of passage 90 nontumorigenic fibroblast-like line 118 cells derived from a 36-day-old strain 2/N guinea pig fetus were treated with 10 micrograms benzo[a]pyrene or 1 microgram N-nitroso-N-methylnitroguanidine/ml medium for 3 days and maintained in exponential growth. One week later, all cultures exhibited morphologic alteration, and by 8 weeks two-thirds exhibited morphologic transformation. After 4 months, morphologically altered or transformed cells expressed progrssive colony growth in semisolid 0.35% agar medium. Susceptibility to lymphotoxin colony inhibitory activity developed subsequent to morphologic transformation but independently of growth in semisolid medium. Susceptibility to natural lymphoid cell destruction was also independent of growth in semisolid medium. All four phenotypic properties were expressed only after carcinogen treatment and 12 months later remained unaccompanied by either natural delayed-type skin reactivity or by the ability of the cells to grow as tumors in syngeneic guinea pigs of nu/nu mude mice. Thus while morphologic transformation, colony formation in semisolid medium, and susceptibility to lymphotoxin and to nautral lymphoid cell cytotoxicity were frequently associated with neoplasia, they could be expressed independently of one another; individually or collectively in a specific developmental sequence, they were insufficient for expression of either natural delayed-type skin reactivity of tumorigenicity.

Neoplastic transformation of Syrian hamster cells by putative epoxide metabolites of commercially utilized chloroalkenes.

The transforming ability of six epoxides of structurally related chloroalkenes was determined with a quantitative Syrian hamster embryo cell model. The epoxides used were cis-1-chloropropene oxide (c-CPO), trans-1-chloropropene oxide (t-CPO), cis-1,3-dichloropropene oxide, trans-1,3-dichloropropene oxide, trichloroethylene oxide (TCEO), and tetrachloroethylene oxide (PCEO). All six epoxides induced morphologic transformation of Syrian hamster embryo cells and caused cell lethality as reflected in the reduced cloning efficiency; in all instances, transformation was observed with less than 25% toxicity. The potency of the various epoxides was influenced by the difference in stability of the compounds. The results with c-CPO, t-CPO, TCEO, and PCEO were consistent with a linear dose response. The transformation results reflect the carcinogenicity of the parental chloroalkenes tested thus far. Furthermore, if the epoxides are metabolic intermediates of the chloropropene parent chloropropenes, the epoxides are probably proximate carcinogens.

Sister chromatid exchange and chromosome aberration analysis with the use of several carcinogens and noncarcinogens.

Several chemical carcinogens and noncarcinogens were tested for their ability to induce sister chromatid exchanges (SCE) and structural chromosome aberrations in cultured V79-4 Chinese hamster cells. All of the direct-acting carcinogens induced a large increase in SCE frequency. Two chemicals, which are mutagenic in microorganisms but whose carcinogenicity is poorly documented, also increased the frequency of SCE. Carcinogenic polycyclic hydrocarbons caused an increased incidence of SCE only when a metabolizing feeder layer was used, whereas no increase was observed with noncarcinogenic polycyclic hydrocarbons. The other noncarcinogens also did not influence the SCE frequency. Although some chemicals increased the frequency of structural chromosome aberrations, no correlation was found between the frequencies of SCE and aberrations.

Lymphotoxin prevention of diethylnitrosamine carcinogenesis in vivo.

Development of intervention measures to control cancer would be facilitated by being able to monitor in vivo carcinogenesis by in vitro quantitation of early indices of neoplastic transformation to assess the in vivo effectiveness of preventive-therapeutic measures. Pregnant Syrian golden hamsters were used in an in vivo-in vitro transplacental model of carcinogenesis to determine the extent that in vivo administration of immunologic hormone preparations along with chemical carcinogen would prevent morphologic transformation assessed in vitro. Pregnant hamsters at 10-11 days of gestation were given injections ip of 3 mg diethylnitrosamine (DENA)/100 g body weight and were killed 2 days later when fetal cells were seeded for colony formation. The frequency of morphologically transformed colonies was assessed after 7 days of growth. Cloning efficiency and mean transformation frequency after DENA exposure were 3.6% and 1 X 10(-4) per cell seeded, respectively. The ip injection of an immunologic hormone preparation reduced the transformation frequency by 46%. The hormone preparation, containing 10,000 U of lymphotoxin but no detectable interferon, was the ultrafiltered lymphokines (greater than 10,000 mol wt) from phytohemagglutinin-stimulated hamster peritoneal leukocytes. The effect of lymphotoxin on cocarcinogenic exposure of fetal cells to DENA in vivo followed by X-irradiation in vitro was also determined. Cells exposed to 250 rad in vitro had a cloning efficiency of 0.5% and a transformation frequency of 0.4 X 10(-4) per cell seeded. After DENA injection and X-irradiation, the transformation frequency increased to 1 X 10(-4) and was inhibited 64% by lymphotoxin in vivo. Thus immunologic hormones (e.g., lymphotoxin) can prevent carcinogenesis in vivo. Furthermore, in vitro quantitation of transformation is a rapid means for evaluating therapeutic and autochthonous effector mechanisms for their ability to prevent or otherwise modulate carcinogenesis in vivo.

Immunomodulation of the lymphoresponsive phases of carcinogenesis: mechanisms of natural immunity.

Biphasic modulation of the frequency of 3-methylcholanthrene-induced tumors in relation to increasing immunocompetence of tumor-bearing mice is indicative of lymphoresponsive stimulatory and inhibitory phases of carcinogenesis; components and mechanisms contributing to the paradoxical ability of the native or natural immune system to stimulate as well as to inhibit the development of cancer were identified with the use of an in vitro model of carcinogenesis. Normal NIH/N Syrian golden hamster spleen leukocytes caused a biphasic modulation of UV-irradiated hamster cell morphologic transformation that was dependent on the leukocyte-to-target cell (L:TC) ratio. Whereas low (less than 50:1) and high (greater than 200:1) L:TC ratios inhibited the transformation frequency, intermediate ratios were less inhibitory and seemingly immunostimulatory. Inhibition alone occurred with antigen- or mitogen-activated leukocytes or with macrophages. An anticarcinogenic lymphocyte hormone, lymphotoxin, eliminated the leukocyte-mediated stimulatory phase; galactose, a lymphotoxin inhibitor, blocked the leukocyte-mediated inhibitory phases. Thus immunoinhibition was mediated through lymphotoxin, and the apparent immunostimulation possibly resulted from immune cell regulation of lymphotoxin activity during the early stages of carcinogenesis.

Loss of leukoregulin up-regulation of natural killer but not lymphokine-activated killer lymphocytotoxicity in human papillomavirus 16 DNA-immortalized cervical epithelial cells.

The sensitivity of human cervical epithelial cells immortalized by transfection with human papillomavirus type 16 (HPV16) DNA, to lysis by natural killer (NK) and lymphokine-activated killer (LAK) lymphocytes was evaluated at progressive stages of transformation. Both early- (10-20 wk) and late- (greater than 30 wk) passage HPV16-immortalized cells were resistant to NK lymphocyte cytotoxicity but sensitive to LAK lymphocyte cytotoxicity at lymphocyte-to-cervical cell ratios ranging from 1:1 to 50:1 in a 4-hour 51Cr release assay. Treatment of early-passage HPV16 DNA-immortalized cells with 2.5 U/mL of the NK lymphocytotoxicity-sensitizing lymphokine, leukoregulin, for 1 hour induced modest sensitivity to NK cells (P less than .05) but markedly up-regulated LAK sensitivity twofold to threefold. At the later passages, leukoregulin up-regulation of sensitivity to NK was lost but remained to LAK lymphocytotoxicity. Similarly, an HPV16-positive human cervical carcinoma cell line, QGU, was also resistant to NK lymphocytotoxicity and sensitive to LAK lymphocytotoxicity; leukoregulin failed to confer sensitivity to the NK-resistant QGU tumor cells and increased their sensitivity to LAK lymphocytotoxicity 1.5-fold to twofold. Although the HPV-immortalized cervical cells containing integrated HPV16 DNA were not tumorigenic, they mimicked the response of established HPV16-positive cervical carcinoma cells. HPV16-immortalized cervical epithelial cells provide a useful model for the study of cytokine modulation of dysplastic and neoplastic cervical epithelial cell sensitivity to natural lymphocytotoxicity.

Mutagenesis of Chinese hamster cells in vitro by combination treatments with methyl methanesulfonate and N-acetoxy-2-acetylaminofluorene.

Mutational synergism was examined in Chinese hamster V79 cells exposed to methyl methanesulfonate followed by N-acetoxy-2-acetylaminofluorene (AcAAF) at different time intervals. Treatment with 500 micron methyl methanesulfonate resulted in 95% survival of cloning ability and induced approximately 4 azaguanine-resistant mutants/10(5) survivors. Seven micron AcAAF produced 10 times as many mutants, and the survival was 7%. Lethal synergism was observed for methyl methanesulfonate treatments followed by 7 micron AcAAF, and the resulting lethality was unaffected by increasing the time interval between treatments from 1 to 48 hr. However, no significant changes in the mutant frequency from that induced by AcAAF alone were found for treatment intervals of 1 to 63 hr. This result contrasts with the 6-fold enhancement of the AcAAF-induced transformation of Syrian hamster embryo cells exposed to the same combination with a 48-hr interval between treatments, as previously reported (Chem.-Biol. Interactions, 9: 351-364, 1974). The difference in the response of these two cell types demonstrates the difficulties in attempting to extrapolate the known correlation between individual mutagen and carcinogen treatments to combination treatments, with different cell types for the two cellular responses.

Excision and postreplication DNA repair capacities, enhanced transformation, and survival of Syrian hamster embryo cells irradiated by ultraviolet light.

The frequency of ultraviolet light (UV)-induced neoplastic transformation of Syrian hamster embryo cells (HEC) is enhanced 3- to 10-fold when the cells are first treated with either X-irradiation or with methyl methanesulfonate. Maximum enhancement occurs when the interval between the two treatments is 48 hr. The relevance of UV-induced transformation to neoplasia is confirmed because the transformation to neoplasia is confirmed because the transformants produce tumors when injected into nude mice. Excision and postreplication DNA repair were studied to determine whether the enhanced transformations were associated with either of these repair mechanisms. Independent of X-ray or of methyl methanesulfonate pretreatment, approximately 25% of the pyrimidine dimers are excised within 24 hr in cells irradiated with UV with 3 J/sq m. During this period, more than 70% of the genome of cells irradiated with UV has been replicated. Postreplication repair is measured by the time required to chase pulse-labeled nascent DNA strands to parental-sided DNA. Regardless of pretreatment, 1 and 3 hr are required for pluse-labeled DNA in control and irradiated (10 J/sq m) cells, respectively, to reach parental size. Therefore, no correlation is found between a change in the rate of excision or postreplication repair and enhancement of transformation. Relative to control cloning efficiency, the survival of HEC contain more than 10(5) pyrimidine dimers/genome. The level of survival is similar to the survival of human skin fibroblasts which excise pyrimidine dimers four to five times as efficiently. Moreover, postreplication repair cannot account for the ability of these cells to survive because it is three times slower than in human fibroblasts. Therefore, other repair mechanisms must be responsible for HEC survival and transformation.

Neoplastic transformation of guinea pig fetal cells in culture induced by chemical carcinogens.

Twenty-four cell strains derived from freshly isolated diploid strain 2 guinea pig fetal cells exposed in utero or directly in culture to a carcinogenic or noncarcinogenic chemical were evaluated during 4 to 24 months of continuous cultivation. Morphological alterations in carcinogen, i.e., benzo(a)pyrene-, 7,12-dimethylbenz(a)anthracene-, 3-methylcholanthrene-, N-methyl-N'-nitro-N-nitrosoguanidine-, diethylnitrosamine-, aflatoxin B1-, or N-acetoxyacetylaminofluorene-treated cultures were seen shortly after treatment; transformation, a loss of cell orientation, did not occur for 4 or more months, was not seen in controls and did not uniformly appear simultaneously with the capacity of transformed cells to grow as tumors in irradiated syngeneic newborn guinea pigs. Changes in plating efficiency, saturation density, doubling time, and chromosome alterations of transformed cells also did not correlate with progressive tumor growth. Formation of colonies in 0.35% agar appeared subsequent to or concomitantly with morphological transformation, and in every case they developed concurrently with the potential for neoplastic growth. Controls originally treated with a noncarcinogen (acetone, polycyclic hydocarbon, or aromatic amide) did not form agar colonies or tumors after inoculation of 108 cells. Studies with guinea pig cells indicate that persistent alterations in some growth parameters occur prior to expression of tumor growth. The ability of cells to form colonies in agar and to produce tumors may require 4 to 18 months of culturing. Of the cell properties examined, colony formation in agar was the best indication of the neoplastic state.

Chromosome alterations in Syrian hamster cells transformed in vitro by sodium bisulfite, a nonclastogenic carcinogen.

Sodium bisulfite, a nonmutagen at neutral pH, induces neoplastic transformation of cultured Syrian hamster fetal cells. Morphologically transformed fibroblast colonies were isolated, and derived cell lines formed anchorage-independent colonies in agarose and progressively growing s.c. fibrosarcomas in nu/nu mice. Five tumorigenic cell lines analyzed by G- and C-banding were chromosomally abnormal with numerical deviations and structural alterations. Three tumors that developed in nude mice had the chromosome constitution of the inoculated transformed cell as well as secondary changes associated with tumor progression. Transformed cell lines had either a predominantly near-diploid or a near-tetraploid population with consistent chromosome gain and loss. Monosomy of the chromosome 13 observed in three cell lines was a nonrandom numerical alteration. Four lines had abnormal chromosomes resulting from deletions, unbalanced translocations, or centric fusions, and one cell line had a chromosome with a homogeneously staining region. Changes of chromosomes 1 and X were observed in three lines. The breakpoints on X chromosome nonrandomly involved the region qa5 which is frequently affected in hamster cells transformed by other carcinogens and may result in loss of genes essential for the maintenance of a normal phenotype. The formation of abnormal chromosomes cannot be directly attributed to the initial DNA damage as bisulfite concentrations effective in causing neoplastic transformation induced a significant but minimal increase in sister chromatid exchanges and failed to cause chromosome aberrations. Bisulfite inhibition of DNA replication might be a contributing factor in the occurrence of abnormal chromosomes. This cytogenetic analysis provides the first evidence that neoplastically transformed cells by a nonclastogenic carcinogen exhibit persistent chromosome rearrangements, a genetic alteration essential to the process of malignant transformation.

Comparison of nude mice with the host species for evaluation of the tumorigenicity of guinea pig and hamster cells transformed in vitro by chemical carcinogens.

Nude mice (nu/nu) were compared with the species of origin for determination of the tumorigenicity of cells from chemical carcinogen-transformed and noncarcinogenic chemically treated, nontransformed guinea pig and Syrian hamster cultures. The incidence and time of appearance of progressively growing tumors were similar in the host species and in nude mice after injection of 10(7) transformed cells. Inoculation of 10(8) nontransformed cells routinely was nontumorigenic in the host species and in nude mice. The nude mouse has potential as a sensitive and reliable alternative host to the species of origin to elaluate the tumorigenicity of xenogeneic mammalian cells from cell culture model systems of carcinogenesis.

Integration of human papillomavirus 16 DNA and genomic rearrangements in immortalized human keratinocyte lines.

Five foreskin-derived keratinocyte lines, immortalized by transfection of human papillomavirus (HPV16) DNA, were cytogenetically abnormal, exhibiting numerical deviations and altered chromosomes due to translocations, deletions, achromatic lesions, or partial duplications. Furthermore, all lines had cells with either homogeneously staining regions or double minute chromosomes, alterations associated with malignancy or drug resistance. None of these lines were tumorigenic in nude mice, showing that such alterations which are a manifestation of DNA amplification also occur in nonneoplastic cells. By in situ chromosome hybridization, viral sequences were identified on abnormal chromosomes at the junction of chromosome translocations, at achromatic lesions and within homogeneously staining regions and duplicated chromosome segments. Thus, for the first time in an experimental system, HPV16 integration into the cellular genome was associated with the induction of a subset of chromosome alterations. HPV16 integration that frequently occurred at fragile sites and near protooncogenes may be a critical alteration which confers a selective growth advantage and an indefinite proliferative potential to HPV-transfected cells.

The susceptibility of guinea pig cells to the colony-inhibitory activity of lymphotoxin during carcinogenesis.

The time of susceptibility of cells to lymphotoxin during carcinogenesis was determined. At different stages of in vitro chemical carcinogen-induced neoplastic transformation, colony formation of guinea pig cells was evaluated with lymphotoxin obtained from syngeneic nonimmune leukocytes. Cells exhibiting sequential morphological alteration, morphological transformation, and neoplastic transformation had been preserved in liquid nitrogen and, after reintroduction in culture, were analyzed simultaneously for their susceptibility to lymphotoxin. Morphologically altered and morphologically transformed cells at subcultures prior to neoplastic transformation were resistant to lymphotoxin inhibition of colony formation. Cells neoplastically transformed by N-methyl-N'-nitro-N-nitrosoguanidine in vitro or by N-methyl-N'-nitro-N-nitrosoguanidine or diethylnitrosamine with the host-mediated, in vivo-in vitro method were susceptible and exhibited a quantitatively culture-specific degree of colony inhibition. The parental noncloned and cloned neoplastically transformed cells in each series, furthermore, exhibited similar degrees of colony inhibition, which indicates that lymphotoxin susceptibility developed concomitant with, or close to, the time of neoplastic transformation and remained stable during subsequent cell generations.

Cyclic nucleotide modulation of in vitro morphological transformation of Syrian hamster cells.

The cyclic nucleotides, cyclic adenosine 3':5'-monophosphate (cAMP) and cyclic guanosine 3':5'-monophosphate (cGMP) or their dibutyryl and monobrominated derivatives, may either increase or decrease morphological transformation of Syrian hamster embryo cells exposed to N-methyl-N'-nitro-N'-nitrosoguanidine (MNNG). The effect on transformation is primarily a function of the parent cyclic nucleotide and the duration of exposure to the nucleotides. At concentrations of 5 mM or larger for a minimum 24-hr exposure cAMP, cGMP, and their congeners reduced the colony-forming ability of nontransformed Syrian hamster embryo cells not exposed to MNNG; however, cGMP and its derivatives caused less toxicity than cAMP or its analogs. cAMP and its monobrominated and dibutyrylated derivatives decreased the transformation frequency associated with MNNG irrespective of whether the various adenylnucleotides were administered before or after MNNG. The greatest inhibitory effect of MNNG-induced transformation was obtained with N6,O2-dibutyryl cyclic adenosine 3':5'-monophosphate followed in order by 8-bromocyclic adenosine 3':5'-monophosphate and cAMP. At equimolar doses, the dibutyryl and brominated analogs of cGMP but not unsubstituted cGMP enhanced transformation when administered prior to exposure of the Syrian hamster embryo cells to MNNG but reduced the transformation frequency when added after MNNG. The enhancing and inhibitory effects of the guanine cyclic nucleotide-induced alteration of MNNG-associated transformation frequencies were dose and time dependent and occurred in the order N6,O2-dibutyryl cyclic guanosine 3':5'-monophosphate greater than 8-bromocyclic guanosine 3':5-'monophosphate much greater than cGMP. Butyric acid neither diminished nor increased MNNG-induced transformation frequency. The latter suggests that butyrate ions formed by metabolism of the cyclic nucleotide analogs were not a factor in the observed alterations of transformation frequency.

Enhancement of viral transformation for evaluation of the carcinogenic or mutagenic potential of inorganic metal salts.

Thirty-eight metal salts were tested for their capacity to enhance transformation of Syrian hamster embryo cells by a simian adenovirus, SA7. All of the metal salts with known carcinogenic potential in animals or mutagenic activity in microbial or mammalian cells increased the SA7 transformation frequency. Metals were classified into three groups according to the concentration necessary to produce significant enhancement. Those showing highest activity (positive at less than 0.05 mM) were the salts of antimony, arsenic, cadmium, chromium, and platinum. The second group (positive from 0.05 to 0.6 mM) included beryllium, cobalt, copper, lead, manganese, mercury, nickel, silver, thallium, and zinc. Iron salts were placed in a third group (only positive at concentrations greater than 0.9 mM). With the exception of ZnCl2 and ZnSO4, enhancement was demonstrated by both a relative increase in the viral transformation frequency and an absolute increase in the number of transformed foci among treated cells. The latter observation and the demonstration of enhancement in the absence of overt cell killing negate the possibility that enhancement resulted from the selection of transformation-sensitive cells.