Carcinogenic Cr(VI) and the Nutritional Supplement Cr(III) Induce DNA Deletions in Yeast and Mice

Chromium most commonly exists in two stable oxidation states, hexavalent [Cr(VI)] and trivalent [Cr(III)]. Cr(VI) originates from industrial settings, whereas in nature, chromium exists in its trivalent form. Cr(III) is used as a micronutrient element in dietary supplements for weight loss and muscle gain (1). Widespread Cr(VI) environmental contamination and the popularity of Cr(III) dietary supplements make it imperative to examine the genotoxic and carcinogenic potential of both forms of chromium. Cr(III) has tested negative in the majority of genotoxicity tests (2) and scientists generally regard Cr(III) as nontoxic due to poor absorption. Cr(VI), however, is considered a pulmonary carcinogen (3) and has tested positive in a wide range of genotoxicity tests in numerous organisms (2). Environmental contamination has led to elevated levels of Cr(VI) in drinking water across the U.S. and in many parts of the world (3). The carcinogenic potential of Cr(VI) via the oral route has yet to be determined.

The scientific community has not established a molecular mechanism by which Cr(VI) elicits a carcinogenic effect in the pulmonary system. Evidence is accumulating that absorbed Cr(VI) is reduced by cellular reducing agents to Cr(III) via a number of reactive intermediates. Intracellular reduction of Cr(VI) can lead to DNA damage by two potential mechanisms: (a) oxidative DNA damage resulting from reactive intermediates and (b) Cr(III)-DNA interactions (4).

The lack of mechanistic knowledge and the insufficiency of epidemiologic and animal studies have led to divisions in expert opinions as to whether Cr(VI) ingestion can cause an increased risk of cancer. Such discord has led to doubts about current water quality standards and their ability to protect the public from the possible health risks associated with chromium. In response to these doubts, the National Toxicology Program (NTP) has launched a long-term animal carcinogenicity study to evaluate Cr(VI) in drinking water; the results of the study should be available later this year (5).

The NTP is also preparing a long-term cancer study to evaluate Cr(III). Although the current understanding of Cr(III) genotoxicity is limited, there is some evidence that Cr(III) can cause DNA damage because it could directly interact with DNA (6, 7). The binding of Cr(III) to DNA can result in a Cr(III)-mediated DNA interstrand crosslink, a Cr(III)-mediated protein-DNA crosslink or a Cr(III)-DNA mono-adduct (8). The significance of each of these lesions to the process of carcinogenesis is poorly understood. In 1995, Stearns et al. proposed that chromium supplementation might have an adverse long-term biological effect due to accumulation of Cr(III) in various tissues (9). However, no studies have thus far shown the carcinogenic potential of ingested Cr(III) in vivo.

The current study examined the ability of Cr(VI) and Cr(III) to induce DNA deletions in yeast and mice. Genomic rearrangements, which include DNA deletions, are known to be deleterious to genomic stability and are often found in human tumors (10). In addition, patients with cancer-prone genetic disorders, characterized by mutations in Atm, Trp53, and Wrn genes, have an elevated frequency of genomic rearrangements (11). The strong link between genomic rearrangements leading to DNA deletions and cancer makes the measure of DNA deletions a highly relevant gauge of carcinogenic potential.

The two systems used in this study to detect DNA deletions have accurately identified a wide range of carcinogens (12, 13). The RS112 tester strain of the yeast, Saccharomyces cerevisiae, contains a plasmid with an internal fragment of the HIS3 gene integrated at the genomic HIS3 locus, yielding an integrative disruption of the HIS3 gene (14). The disruption results in two copies of the HIS3 gene, each copy having one terminal deletion. Recombination between the two his3 deletion alleles reverts the tester cells to a HIS3⁺ phenotype. This recombination event leads to the deletion of the 6 kb of DNA comprising the integrated plasmid (Fig. 1). The system is, thus, named the deletion (DEL) assay (15). DNA deletion frequency measured by the DEL assay has proven to be a reliable indicator of carcinogenic potential. In a study of 60 compounds of known carcinogenic activity in animals, the yeast DEL assay was 86% accurate in identifying carcinogens (12).

The in vivo DNA deletion assay uses the C57BL/6J p^un/p^un mouse strain, which has a 70 kb tandem duplication at the pink-eyed dilution (p) locus (16, 17), termed pink-eyed unstable (p^un) allele (Fig. 2). The assay measures the deletion frequency of the 70 kb DNA duplication within the p^un locus in developing embryos. The somatic deletion reconstitutes the p gene (p^un reversion), which participates in black pigment assembly. Thus, DNA deletion events result in black-pigmented cells (eyespots) on the retinal pigment epithelium (RPE) in the offspring. The frequency of such deletions is elevated in mice treated with various carcinogens (18) as well as in mice with cancer-predisposing gene mutations such as Atm, Trp53, and Wrn (19–21). Thus, the measure of p^un reversion frequency provides a quantification of DNA damage that is highly relevant in the process of carcinogenesis.

This study used S. cerevisiae RS112 tester strain and the p^un/p^un mouse strain to determine the effect of Cr(VI) and Cr(III) on the frequency of DNA deletions. We exposed the mice to Cr(VI) and Cr(III) in drinking water.

Potassium dichromate (CAS 7778-80-9), chromium (III) nitrate (CAS 7789-02-8), and chromium chloride (CAS 10060-12-5) were purchased from Sigma-Aldrich, St. Louis, MO.

Media. Synthetic complete (SC) medium, SC medium lacking histidine (SC-HIS) and inoculation (-LEU) medium were prepared as previously described (22).

Recombination assay. The diploid strain RS112 carries the his3Δ3′-his3Δ5′ (HIS3 :: pRS6) recombination substrate (Fig. 1) on one homologue and a deletion of the entire region of homology to the recombination substrate on the other homologue (his3-Δ200).

The yeast DEL assay was done as previously described (12). In brief, single colonies of RS112 strain were grown overnight and then subcultured in the presence of K₂Cr₂O₇ [Cr(VI)] or Cr(NO₃)₃ [Cr(III)] for 17 hours at 30°C under constant shaking. Cells were plated on SC medium to determine the number of survivors and onto SC-HIS medium to score for DEL events.

The frequency of DEL recombination is expressed as the number of DEL recombination events versus the number of viable cells. Statistical significance between treatment group and control was measured using Student's t test.

C57BL/6Jp^un/p^un mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Mice were bred in the institutional specific pathogen–free animal facility under standard conditions with a 12-hour light/dark cycle, and were fed standard diet and water ad libitum. Pregnancy was timed by checking for vaginal plugs, with noon of the day of discovery counted as 0.5 days postcoitum. Similarly, the time of birth of a litter was timed with the noon of discovery counted as 0.5 days postpartum.

Pregnant dams were given free access to Cr-supplemented drinking water at 10.5 to 20.5 days postcoitum. Cr(III) (CrIII chloride salt) was used at either 1,875 or 3,750 mg/L concentration, which was calculated to yield an average dose of 375 or 750 mg of chemical per kg of body weight per day (mg/kg/d), respectively. Cr(VI) (potassium dichromate) was used at either 62.5 or 125.0 mg/L concentration yielding an average dose of 12.5 or 25 mg/kg/d, respectively. Control mice received regular (unsupplemented) drinking water. For each group, four to six dams were used. For determining DNA deletion frequency, 20-day-old offspring were harvested to visualize the eyespots (DNA deletions) in their RPE. For determining intracellular Cr concentration, mouse embryos were isolated at 17.5 days postcoitum.

Dissection of the RPE and scoring for DNA deletions. Offspring were sacrificed at 20 days of age, their eyes were dissected and whole mount RPE slides were prepared for microscopic analysis of eyespots. Eyes were processed to expose the RPE layer as previously described (13, 23). A pigmented cell or a group of adjacent pigmented cells separated from each other by no more than five unpigmented cells was considered as an eyespot that resulted from one deletion/p^un reversion event (13). The number of eyespots and number of cells comprising the eyespot was counted.

The intracellular chromium measurements reported in this study could be a combination of membrane-bound and intracellular concentrations. Although we extensively washed the cells with EDTA and metal-free water, the actual intracellular chromium concentration could be lower than reported here. This would be particularly true for Cr(III) molecules that have greater contact with the membrane due to passive diffusion. In any case, this would only mean that any genotoxic effects would occur at even lower intracellular concentrations than we report.

Sample preparation. To prepare cells for measurement of intracellular chromium concentration, a single colony was picked from YPAD medium and inoculated into 50 mL of −LEU medium and grown for 24 hours at 30°C under constant shaking (250 rpm). Cells were counted and divided into 20 to 300 mL subcultures. In order to achieve a similar cell density after exposure, cells were exposed in varying volumes of media while maintaining a constant initial cell concentration. The varying levels of cytotoxicity elicited by different concentrations of Cr(III) and Cr(VI) precluded the possibility of performing this experiment at constant culture volumes. Cells were exposed to either Cr(III) or Cr(VI) at different concentrations for 17 hours at 30°C under constant shaking. After the exposure, cells were pelleted in a clinical tabletop centrifuge, washed once with metal-free water, resuspended in metal-free water, and counted. An aliquot of cells from each sample was plated onto the appropriate medium to score for DEL recombination (as described above). The remaining cells were washed twice in 10 mmol/L EDTA (Sigma-Aldrich) followed by two washings in metal-free water. After the final washing, the cells were counted and pelleted.

Digestion and analysis. Cells were resuspended in 1 mL of 25% ultrapure nitric acid (Fisher, Pittsburgh, PA) and heated in closed containers at 80°C for 2 days. The clear solution of digested yeast cells was transferred to low-density polyethylene vials (Nalge Nunc International, Rochester, NY) and the volume was adjusted to 4 mL for each sample using metal-free water. The chromium concentration was measured at the University of California at Los Angeles ICP Facility in the Department of Chemistry and Biochemistry using the Thermo Jarrell Ash Iris 1000 inductively coupled plasma-atomic emission (ICP-AE) instrument.

One pregnant dam was selected from each exposure group and the embryos were harvested at 17.5 days postcoitum. Two pregnant dams were selected from the negative control group to determine the background chromium concentration in untreated animals. The embryos were harvested with plastic instruments to avoid chromium contamination from stainless steel. Embryos were then digested and analyzed for chromium concentration at the UCLA ICP Facility in the Department of Chemistry and Biochemistry on an Agilent 7500 c Quadrupole ICP-mass spectrometry (MS) equipped with an H₂/He Octapole reaction/collision cell. In brief, digestion was carried out in 10 mL of Ultrapure nitric acid (Optima, Fisher) at 90°C until no particulate or color change was observed. Matrix exchange was achieved by successive evaporation and dilution into a 5% nitric acid at 100°C to 110°C. ICP-MS analysis was done in accordance with Environmental Protection Agency method 200.8 using indium and bismuth as internal standards.

Effect of chromium exposure on the frequency of DNA deletions in yeast. The yeast DEL assay was used to determine the DNA deletion frequency in response to Cr(VI) and Cr(III). We chose chromium concentrations that range from no cytotoxic effect to a cytotoxic effect leading to <15% viable cells. Figure 3A depicts the dose-response curves for Cr(VI) and Cr(III) as the mean and SE of three independent experiments. Cr(VI) and Cr(III) induced DNA deletions in a dose-dependent manner. At the highest concentration tested, Cr(VI) and Cr(III) induced an average of 16.89 and 4.82 DNA deletion events per 10⁴ viable cells, respectively. Thus, both Cr(VI) and Cr(III) significantly induced DNA deletions in yeast.

Correlation of intracellular chromium concentration versus DNA deletion frequency in yeast. In order to compare the genotoxicity of Cr(VI) and Cr(III) in yeast without the effect of absorption, we measured intracellular chromium concentration after exposure to either Cr(VI) or Cr(III) and correlated these concentrations with DNA deletion frequency. After exposure to either Cr(VI) or Cr(III), we measured DNA deletions in samples of RS112 cells from each culture. ICP-AE spectroscopy analyses of the remaining cells provided intracellular chromium concentrations. Figure 3B depicts the linear regression analysis of DEL recombination frequency versus intracellular chromium concentration. The R² value for the Cr(VI) line of 0.89 and the Cr(III) line of 0.87, support a linear dose-response relationship between intracellular chromium concentration and DNA deletion frequency. Furthermore, the steeper slope of the Cr(III) line implies that at the same intracellular concentration Cr(III) exposure is a more potent inducer of DNA deletions than Cr(VI).

Effect of chromium ingestion on the frequency of DNA deletions in mice. In a second series of experiments pregnant dams, at 10.5 to 20.5 days postcoitum, ingested Cr(VI) or Cr(III) in drinking water. Cr (VI) exposure doses in this study overlap with the two lowest doses used by the NTP for a long-term carcinogenicity study (5). The Cr(III) doses in our study assume that animals absorb Cr(III) approximately 15 times less efficiently than Cr(VI) (24). Our goal was to achieve similar exposure doses of Cr(VI) and Cr(III) at the intracellular level. Mice treated with chromium did not exhibit any obvious signs of toxicity and the litter size did not differ from untreated controls for all treatment groups.

Control (untreated) mice displayed an average of 5.49 eyespots per RPE. Using concentrations of 62.5 and 125.0 mg/L, mice treated with Cr(VI) had 27% and 38% more eyespots as compared with untreated controls, respectively (P < 0.01; Fig. 4A). Similar increases are observed with Cr(III) but at much higher exposure doses. When compared with controls, the number of eyespots in the Cr(III) treatment group was 36% and 53% higher in 1,875 and 3,750 mg/L dose groups, respectively (P < 0.001; Fig. 4A). The treatment groups did not differ significantly. These data show that transplacental exposure to either Cr(III) or Cr(VI) results in elevated frequencies of DNA deletions in mice.

Correlation of tissue chromium concentration versus frequency of DNA deletions in mice. Chromium absorption by an animal and further passage through the placental barrier to the embryo is a critical issue in assessing an observed genotoxic effect. To address this issue, we measured the total chromium concentration in the embryos using ICP-MS. This allowed us to determine the chromium concentration that reached target cells after oral exposure of the dam. Figure 4B depicts the relationship between embryo chromium concentration and DNA deletion frequency as the mean and SE.

The background chromium concentration in the embryo was 6.19 ± 1.15 ng/g fresh tissue weight. For the 62.5 and 125.0 mg/L Cr(VI) oral exposure doses, the embryo chromium concentrations were 16.41 ± 0.33 and 27.39 ± 0.88 ng/g, respectively. For the 1,875 and 3,750 mg/L Cr(III) oral exposure doses, the embryo chromium concentrations were 8.72 ± 0.98 and 18.77 ± 3.62 ng/g, respectively. Transplacental animal exposure to Cr(III) leads to lower chromium accumulation in the embryo than does transplacental animal exposure to Cr(VI). At the cellular level, however, Cr(III) induces DNA deletions at a lower concentration than does Cr(VI). Cr(III) exposure led to 7.46 eyespots per RPE at a tissue chromium concentration of 8.72 ng/g, whereas Cr(VI) exposure lead to 7.57 eyespots/RPE at a tissue chromium concentration of 27.39 ng/g (Fig. 4B). Comparing the embryo chromium concentrations to DNA deletion frequency revealed that Cr(III) exposure leads to induction of DNA deletions at an ∼3-fold lower embryo chromium concentration than does exposure to Cr(VI). This trend is consistent with the results obtained in yeast.

The goal of this study was to determine the comparative effects of hexavalent and trivalent chromium on the induction of DNA deletions. We report that both Cr(VI) and Cr(III) induce DNA deletions in yeast and mice. Notably, the significantly elevated frequency of DNA deletions resulted in mice exposed to either forms of chromium in drinking water. It is particularly striking that Cr(III), which is regarded as a dietary micronutrient, causes large-scale irreversible genome deletions comparable to the industrial waste contaminant Cr(VI). Moreover, it seems that Cr(III) is more potent than Cr(VI) once the Cr(III) is absorbed. The poor absorption of Cr(III) through the gastrointestinal tract and cellular membranes has often been cited as a reason for the lack of genotoxicity associated with Cr(III). Some investigators maintain that the large capacity of body fluids to reduce Cr(VI) to the less permeable Cr(III) is sufficient to protect from any potential carcinogenicity of Cr(VI) ingestion (25, 26). Other studies have shown that rats injected with Cr(III) accumulated similar levels of chromium bound to chromatin as rats injected with Cr(VI) only the accumulation occurred at a much slower rate (27). The current study shows that even small amounts of absorbed Cr(III) are potentially dangerous. Cr(III) exposure leads to a similar level of DNA deletions as Cr(VI) but at an ∼3-fold lower embryo chromium concentration in mice and at an ∼4-fold lower intracellular chromium concentration in yeast. This observation has important implications for public health due to the widespread and unregulated consumption of Cr(III)-containing dietary supplements that are designed for efficient absorption.

Genotoxicity and carcinogenicity of Cr(VI). A limited number of studies have examined the genotoxicity and carcinogenicity of Cr(VI) via ingestion. The only lifetime study of chronic drinking water Cr(VI) exposure showed that Cr(VI) ingestion induced malignant forestomach tumors in mice (28). However, a scientific review panel later refuted this study due to technical problems (29). A more recent study showed that exposure to chromate [Cr(VI)] in drinking water caused an increase in susceptibility to UV-induced skin tumors in hairless mice (30). Other studies have shown that Cr(VI) ingestion via drinking water could lead to fetotoxicity and embryotoxicity (31), and teratogenicity (32) in rats. An increase in genomic DNA fragmentation resulted in the liver and brain tissues of C57BL/6Ntac mice when exposed to 0.10 and 0.5 LD₅₀ acute oral doses of Cr(VI) (33). Nonetheless, there is a lack of sufficient epidemiologic and animal studies that link ingestion of Cr(VI) contaminated drinking water with genotoxicity or cancer. Accordingly, the NTP is currently evaluating Cr(VI) in a 2-year cancer study in rodents. The NTP study used exposure doses that range from 62.5 to 1,000 mg/L (5). The exposure doses in our study were 62.5 and 125 mg/L, which represents the lower dose range of the NTP study. These doses increased the frequency of 70 kb DNA deletions by 27% and 38%, respectively. Thus, our finding suggests that long-term Cr(VI) oral exposure might lead to the development of cancer.

Genotoxicity and carcinogenicity of Cr(III). The ability of Cr(III) to elicit a genotoxic response in vivo was an unexpected result in this study. Although the current understanding of Cr(III) genotoxicity is very limited, there is some evidence that Cr(III) could cause DNA damage because it can interact directly with DNA (6, 7), which can result in either a Cr(III)-mediated DNA interstrand crosslink, a Cr(III)-mediated protein-DNA crosslink, or a Cr(III)-DNA mono-adduct (8). However, it is not known whether the Cr(III)-DNA interaction leads to irreversible genome damage and carcinogenesis. In fact, Cr(III) does not cause point mutations. For example, Cr(III)-treated plasmids induced a very weak response in shuttle-vector mutagenesis experiments (34) and lesions produced by Cr(III) were not mutagenic in the hypoxanthine-Gua-phosphoribosyl-transferase assay (35). The absence of mutagenicity, however, does not rule out the presence of carcinogenic activity. Many animal and human carcinogens test negative in short-term tests that measure point mutations. The Salmonella assay, for example, has been reported to detect ∼50% of carcinogens (36, 37). On the other hand, the yeast DEL assay detects 86% of carcinogens (12, 13) including nonmutagenic ones (38). In this study, we found that Cr(III) markedly increased the frequency of homology-mediated deletions in both yeast and mouse assays.

A recent study showed that conditions of Cr(VI) reduction that led to the production of Cr(III)-mediated DNA interstrand crosslinks resulted in DNA polymerase arresting lesions (39). DNA polymerase arresting lesions are subject to repair by homologous recombination and sister chromatid exchange (40–42). Furthermore, a yeast mutant deficient in homologous recombination repair was more sensitive to Cr(III)-induced cytotoxicity, which implies that repair of a Cr(III)-mediated DNA lesion is dependent on homologous recombination in yeast (43). Therefore, the Cr(III)-mediated interstrand crosslink could be the lesion leading to DNA deletions in response to Cr(III) exposure observed in the current study.

In summary, the results of this study clearly indicate that both valance states of chromium cause large-scale irreversible genome damage that may further lead to carcinogenesis.

Grant support: Center for Occupational and Environmental Health Sciences, UCLA (R.H. Schiestl), postdoctoral research fellowships of the UC Toxic Substances Research and Teaching Program and the Lymphoma Research Foundation Elizabeth Banks Jacobs and Byron Wade Strunk Memorial Fellowship (R. Reliene), and a UCLA Chancellor's Fellowship, a research fellowship of the UC Toxic Substances Research and Teaching Lead Campus Program, and Environmental Protection Agency STAR fellowship (Z. Kirpnick-Sobol).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Thanks are due to Nicholas Carls for expert technical assistance and Dr. Arlene Russel for insightful editing of the manuscript. Dr. Amir Liba, director of the UCLA ICP facility, performed ICP analyses.