Inorganic and Dimethylated Arsenic Species Induce Cellular p53

Loma Linda University School of Medicine, Loma Linda, California 92354. Received August 26, 2002. Arsenic compounds are known for their ability both t...
0 downloads 0 Views 199KB Size
Download PDF
Chem. Res. Toxicol. 2003, 16, 423-431

423

Inorganic and Dimethylated Arsenic Species Induce Cellular p53 Maria Filippova and Penelope J. Duerksen-Hughes* Department of Biochemistry and Microbiology, Center for Molecular Biology and Gene Therapy, Loma Linda University School of Medicine, Loma Linda, California 92354 Received August 26, 2002

Arsenic compounds are known for their ability both to cause and to treat human cancers, although the molecular mechanisms underlying these actions are incompletely understood. The simplest explanation is that arsenic causes DNA damage that leads to mutations. However, the majority of scientific evidence indicates that arsenic is not a genotoxin or DNA-damaging agent. DNA damage typically leads to cellular responses designed to minimize the replication of damaged DNA, such as the induction of p53, and p53 induction has therefore been used as an indicator of DNA damage. Because this approach can be applied to human cells and does not rely on a specific, heritable mutation occurring at a particular site, it seemed possible that this method could detect DNA damage that was undetectable using other techniques. To examine the genotoxic potential of arsenic compounds, therefore, seven of these compounds (sodium arsenite, sodium arsenate, methyloxoarsine, iododimethylarsine, disodium methyl arsonate, dimethylarsinic acid, and arsenic trioxide) were tested for their ability to increase the cellular level of p53 as measured by ELISA. Of this group, arsenic trioxide was the strongest inducer of cellular p53, while dimethylarsinic acid, iododimethylarsine, and sodium arsenite also caused p53 induction in a dose- and time-dependent manner. Sodium arsenate, as well as the two monomethyl compounds tested, methyloxoarsine and disodium methyl arsonate, did not cause detectable increases in cellular p53. Our results indicate, therefore, that cells respond to several of these arsenic compounds as they do to chemicals that damage DNA, suggesting that exposure of cells to these compounds does in fact cause DNA damage. Such damage could then result in mutations and the observed development of cancer.

Introduction Arsenic compounds found in the environment have been convincingly shown to cause cancer in humans (1), although the mechanism(s) underlying this action are incompletely understood. In general, arsenic compounds are nonmutagenic in most single gene, in vitro assays that test for point mutations, such as the Ames test (for example, see refs 2, 3), and do not easily cause cancer in animal models. They do, however, cause cytogenic effects such as chromosomal aberrations, sister chromatid exchanges, and micronucleus formation. They have also been shown to interfere with DNA repair and may work synergistically with other mutagens (reviewed in ref 4). There are reports indicating that arsenic compounds can damage cellular DNA, thereby acting as genotoxins, as assessed by the comet or single cell gel assay. Some reports indicate that arsenite possesses this activity (57), while another indicates that although the methylated trivalent species are genotoxic, arsenite itself is not (8). Arsenic-induced damage has been proposed to occur by way of reactive oxygen species such as H2O2, as agents that reduce the level of reactive oxygen species (such as DMSO and the enzymes superoxide dismutase and catalase) reduce the cytotoxic and mutagenic potential of arsenic (9-14). Products associated with this damage may include 8-hydroxy-2′-deoxyguanosine (10), forma* To whom correspondence should be addressed. Tel: 909/558-4300 ext 81361. Fax: 909/558-0177. E-mail: [email protected].

midopyrimidine, and 8-oxoguanine (5, 14). Trivalent arsenic is highly reactive with sulfhydryl groups (15), resulting in modifications of various proteins and enzymes (16). One important target is glutathione, a cellular antioxidant. It may also be that the arsenicinduced production of nitric oxide also contributes to DNA damage (14, 17). One reason that understanding the mechanism of arsenic-induced carcinogenicity is important is that it impacts the estimation of the level of arsenic in drinking water likely to affect human health. In contrast to the harmful role of environmental arsenic on cancer in the human population, arsenic trioxide has been shown to be a useful therapeutic in the treatment of hematologic and other cancers (18, 19). Although the mechanism underlying its therapeutic usefulness is not understood in detail, it is likely due to its ability to cause apoptosis in many cell types (reviewed in ref 20). A number of approaches have been used to identify genotoxic compounds, those that cause damage to DNA. These assays include the Salmonella/mammalian microsome assay (Ames test), the metaphase analysis or micronucleus assay in rodent bone marrow assay, the Escherichia coli WP2 tryptophan reversion assay, the TK or HPRT forward mutation assay in clustered mammalian cells, the Drosophila sex-linked recessive lethal assay, assays for chromosomal aberrations, micronuclei and aneuploidy in Chinese hamster or human cells, mammalian DNA damage and repair assays, and mitotic

10.1021/tx025606a CCC: $25.00 © 2003 American Chemical Society Published on Web 02/22/2003

424

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

recombination assays in yeast and Drosophila (reviewed in ref 21). Each assay has its own limitations, and to date, no assay or combination of assays has been reliably shown to be capable of predicting carcinogenicity in rodents (for example, see refs 22-26). In addition, many of these assays require significant investments in time and resources. For this reason, there is a continuing need to develop accurate, inexpensive, and rapid ways to identify genotoxic chemicals. To address this need, we developed an assay for genotoxic compounds that is based on the normal cellular response to DNA damage. The tumor suppressor protein p53 is essential for maintaining the integrity of the genome (reviewed in refs 27-29), as evidenced in part by the fact that more than 50% of human tumors have lost or mutated their p53. When stabilized and activated, p53 regulates expression of a large number of genes, including those involved in DNA repair, cell cycle regulation, and the induction of apoptosis (30). p53 has earned the title of “universal sensor of genotoxic stress” because all examined forms of genotoxic damage, including those caused by exposing cells to ionizing radiation, UV, alkylating agents, DNA cross-linking agents, and reactive oxygen, can induce and activate the protein (29). p53 activation involves both a stabilization of the protein and a change in its ability to bind to DNA. These changes are in turn modulated by a variety of enzymes that posttranslationally modify p53 by phosphorylation and acetylation. Furthermore, the exact sites modified and the extent of the modification vary between the different forms of DNA damage. Modifications that stabilize p53 are thought to work by disrupting the interaction of p53 with its negative regulator MDM2, which mediates the rapid ubiquitination and degradation of p53 (28, 29). Although it is possible for p53 increases to be caused by events other than DNA damage (for example, by substances that block proteosome function (31, 32) or by changes in the methylation state of the promotor sequence), the overall evidence indicates that by far, the most likely explanation for an increase in p53 is DNA damage, caused by either direct interactions or indirectly (for example, through the generation of reactive oxygen species). In previous work, our laboratory has shown that compounds that damage DNA can be identified by their ability to induce p53 (33, 34). Cultured mammalian cells are treated with the test substance and then harvested, and their levels of p53 are analyzed by ELISA. Substances that increase the cellular level of p53 in a timeand dose-dependent manner are considered to be genotoxic, while those that do not are not. We also found that direct-acting genotoxins, compounds that interact directly with cellular DNA, induced p53 increases relatively rapidly (within 2-8 h), while indirect-acting genotoxins, which require metabolic activation or work through alternate targets before they can cause genotoxic damage, induced maximal p53 increases at later time points (more than 12 h). This assay possesses several advantages over most currently used approaches. It is rapid, inexpensive, and requires relatively little material. It appears to be at least as accurate in predicting DNA damage as the Ames test and the Syrian hamster embryo assay (33, 34, and unpublished data). It can be used with human cells, thereby eliminating the need for extrapolation between species, and does not rely on a specific, heritable mutation occurring at a particular site. Our previous work utilized mouse cells in order to compare our results with those

Filippova and Duerksen-Hughes

from rodent in vivo studies. We have now extended the assay to a human system and wished to examine the ability of arsenic compounds to induce p53 in these human cells. Positive results from our p53 induction assay would indicate that cells respond to arsenic compounds as they do to compounds known to damage DNA and provide evidence that they can, in fact, function as genotoxins.

Materials and Methods Cell Lines and Culture. LM (mouse fibroblast), U87 (human glioblastoma), U373 (human glioblastoma), and RKO (human colon carcinoma) cells were cultivated in Minimal Essential Media (MEM) (Invitrogen). SAOS-2 (human osteosarcoma), U2OS (human osteosarcoma), and HCT 116 (human colon carcinoma) cells were cultured in McCoy’s 5a media (CellGro). WI-38 (human lung fibroblast), JEG-3 (human breast carcinoma), and MCF-7 (human breast carcinoma) were cultured in EMEM (CellGro). ZR-75-1 (human breast carcinoma cells) was cultured in RPMI (InVitrogen), and LNZ308 (human glioblastoma) cells were cultured in DMEM (CellGro). All media were supplemented with 10% fetal bovine serum (FBS) (HyClone), with the exception of the media for SaOS-2, which was supplemented with 15% FBS. Medium for the MCF-7 cells was also supplemented by 0.01 mg/mL bovine insulin (Sigma). All cell lines were obtained from the ATCC (Manassas, VA). Chemicals. Mitomycin C (Roche) was dissolved in DMSO to yield a 5 mg/mL stock solution. Although mitomycin C is somewhat soluble in water, it dissolves more easily and remains in solution more completely when DMSO is used as the vehicle. Also, we have shown that DMSO causes no detectable increase in p53 (33) and routinely use DMSO-treated cells as our negative control whenever the test chemical is dissolved in it. K2Cr2O7 (Cr6+) and Cr2(SO4)3 (Cr3+) were obtained from the American Research Corporation of Virginia and dissolved in dH2O to form a stock solution. Arsenic trioxide (As2O3, Trisenox) was obtained from Cell Therapeutics (Seattle, WA). Sodium m-arsenite (iAsIII) and sodium arsenate (iAsV) were obtained from Sigma, and dimethylarsinic acid (DMAsV) and disodium methyl arsenate (MasV) were from Chem Service (West Chester, PA). Iododimethylarsine (DMAsIII) and methyloxoarsine (MAsIII) were kindly provided by Dr. M. J. Mass (EPA, Research Triangle Park, NC). Information regarding the chemical characteristics of the chromium and arsenic compounds was obtained from their respective Agency for Toxic Substances and Disease Registry toxicological profiles (35, 36) and references therein. All arsenic compounds were completely dissolved in sterile dH2O to form stock solutions, which were then diluted in PBS to obtain working stocks that were applied to cells. Cell Treatment and Lysate Preparation. The U2OS cells used for these experiments contain wild-type p53 and were used during their exponential growth phase. Cells (1 × 105 cells per well, in 1 mL media) were seeded in wells of a 24 well plate and allowed to adhere overnight. The indicated chemicals were added to the appropriate wells at the designated concentrations, and media remained at the appropriate pH (7.8 ( 0.1) following addition of the indicated compounds. Cells were then harvested at the appropriate times postexposure. Cells were first washed three times with PBS, and then, 100 µL of lysis buffer (50 mM TrisHCl, pH 7.2, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 5% glycerol, 1 mM PMSF, with one protease inhibitor tablet (Roche) added per 10 mL buffer) was added to each well, and the cells were allowed to lyse for 10 min at 4°C. Lysates were transferred to 1.5 mL microfuge tubes and cleared by centrifugation at 12 000 rpm for 10 min at 4 °C. A 25 µL amount of lysate was used to measure the protein concentration (BCA assay kit, Pierce), and 25 µL was used for the measurement of p53 by ELISA. For each arsenic compound, doses were selected and applied up to a level that led to significant cytotoxicity of the cells. At

Arsenic Induces p53

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 425 Table 1. Identification of p53 Responsive Human Cell Lines

cell line

origin

p53 from untreated cells (ng/mg protein)

p53 from Mit C-treated cells (ng/mg protein)

ratio (treated/ untreated)

notes

LM LNZ 308 SAOS 2 U87 WI 38 ZR 75-1 MCF-7 U2OS RKO HCT 116 JEG 3 U373

mouse fibroblast human glioblastoma human osteosarcoma human glioblastoma human lung fibroblast human breast carcinoma human breast carcinoma human osteosarcoma human colon carcinoma human colon carcinoma human choriocarcinoma human glioblastoma

156.65 ( 8.0 110 ( 0.2 124.2 ( 0.5 188.3 ( 9l.8 214.2 ( 11.7 245.8 ( 1.88 324.6 ( 2.8 188.3 ( 9.8 486.5 ( 5.4 587.0 ( 0.2 1006.7 ( 39.9 1474.48 ( 203.5

766.0 ( 128.4 90.6 ( 8.5 125.2 ( 21.4 683.4 ( 96.1 639.7 ( 108.78 1253.7 ( 47.3 1657.9 ( 36.3 683.4 ( 96.1 793.5 ( 73.2 1039.3 ( 23.8 1939.3 ( 150.9 1678.0 ( 77.8

4.89 0.82 1.01 3.6 3.0 5.1 5.1 3.6 1.6 1.77 1.93 1.13

positive control p53 negative p53 negative wild-type p53 wild-type p53 wild-type p53 wild-type p53 wild-type p53 wild-type p53 wild-type p53 wild-type p53 mutant p53

these higher doses, breakdown of the cellular contents prevented us from obtaining an accurate measurement of p53. Doses at which cytotoxicity was observed are noted in the text. Several independent experiments were carried out for each compound tested, with at least three replicates per time point and per dose. p53 ELISA. The p53 ELISA was performed as described previously (33), with some modifications. Antibodies secreted by clone pAb122 (hybridoma obtained from the ATCC, antibodies purified from the culture medium using protein-A Sepharose) were used as monoclonal capture antibodies. This antibody has a broad specificity, binding to both human and mouse forms of p53 and to both wild-type and mutant forms of the protein (see catalog description of Chemicon MAB4195 and Table 1, this paper). These antibodies were absorbed (50 µL per well, 4 µg/ mL) onto the surfaces of a Nunc-immuno plate, MaxiSorp Surface (NalgeNunc International) by incubation overnight at 4 °C. The plates were then washed on an Auto Strip Washer, ELX50 (Bio-Tek Instruments, Inc.) using PBST (PBS plus 0.1% Tween 20) six times. Nonspecific binding sites were blocked by incubation with 200 µL per well of PBS including 10% calf serum (Invitrogen) (blocking buffer) for 2 h at room temperature, followed by washing as described above. Twenty-five microliters of each cell lysate to be tested was added to 75 µL of blocking buffer, and then, these 100 µL aliquots were then added to the coated wells and allowed to incubate overnight at 4 °C. After they were washed, 100 µL of a solution containing 4 µg/mL biotinylated anti-p53 antibodies (polyclonal, produced in sheep, Roche) diluted into blocking buffer was added to each well and allowed to incubate for 45 min at room temperature. The plates were washed, and then the avidin-peroxidase conjugate (Sigma; 100 µL/well, 2.5 µg/mL diluted into blocking buffer) was added and allowed to incubate for 30 min at room temperature. After it was washed, 100 µL of the substrate (0.3 mg/mL ABTS (2,2′azino-di-(3-ethylbenzthiazolin sulfonate) dissolved into 0.1 M citric acid, pH 4.35, with 1 µL/mL 30% H2O2 added just before use) was added to each well and allowed to incubate for approximately 30 min. The absorbance at 405 nm was read with a microplate reader (Dynex Technologies; MRX Revelation software). The p53-glutathione S-transferase fusion protein (Santa Cruz Biotechnology, Inc.) was used as a standard, and each point was measured in triplicate. The protein concentration of each lysate was also measured using the BCA method (Pierce) and used to normalize the measured p53 values for possible variations in the number of cells per well. Each p53 value (obtained from the ELISA assay) was divided by the protein concentration to obtain a normalized p53 value (ng p53 per mg total protein). The average and standard deviation of the replicates (a minimum of three) were calculated, normalized to the control, and used to prepare the graph. Statistical significance at the 95 or 99% confidence level was determined by comparing the value of the treated cells to that of the untreated cells using the Student’s one-tailed t-test.

Results Human Cells Respond to DNA Damage with Increases in Cellular p53. Our previously published

work was done using mouse LM or NCTC 929 cells. These cells have wild-type p53, a requirement for this assay, and their murine origin made it possible to directly compare our results with those obtained using in vivo experiments. We wished to extend our assay system to human cells, as they have the potential to more directly reflect the situation in humans. We therefore tested several human cell lines from a variety of tissues known to have wild-type p53 for their ability to respond to DNA damage by increasing their level of p53 (Table 1). Mitomycin C was used as the test chemical because our previous work (33, data not shown) indicated that this chemical consistently produced a robust p53 response in mouse cells. In this experiment, the LM mouse fibroblast cells represent a positive control. The LNZ308 and SAOS 2 cells are p53 negative and were expected to show low, background levels of p53 both with and without mitomycin C treatment, which they did. The U373 cells are known to express mutant p53 and were expected to have high levels of p53, both in the presence and in the absence of mitomycin C, which also proved to be the case. Of the remainder, each of the human cell lines tested displayed significant up-regulation of p53 following treatment with 30 µM mitomycin C for 16 h, with levels of induction ranging from 1.6 to 5.1. We chose the U2OS cell line for our future experiments, as these cells are easy to grow, divide rapidly, have a very low background level of p53, and demonstrate a robust increase in p53 following genotoxin treatment. In fact, while this experiment, which used a dose of 30 µM mitomycin C, showed an increase of 3.6-fold, later work using lower doses yielded increases of over 10-fold (Figure 1). U2OS cells treated with mitomycin C demonstrate a robust increase in cellular p53 that is both dose- (Figure 1A) and time- (Figure 1B) dependent. The induction is maximal following treatment for 16 h or more, consistent with its status as an indirect-acting genotoxin. These results are consistent with those obtained with the mouse cells and support the use of U2OS cells in the p53 induction assay. p53 Is Induced in Human U2OS Cells Following Treatment with Cr6+ but Not Cr3+. We then wished to examine the p53 response in U2OS cells following treatment with metal compounds known to cause or not to cause DNA damage. We therefore treated cells with either a trivalent chromium compound, Cr2(SO4)3 (Cr3+), or a hexavalent chromium compound, K2Cr2O7 (Cr6+). The hexavalent form of chromium is known to cause significant genotoxic damage, while the genotoxicity of the trivalent form is either much less or absent (37, 38, and references therein). This is likely due to a lack of

426

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

Filippova and Duerksen-Hughes

Figure 1. Human U2OS cells respond to mitomycin C treatment with a dose- and time-dependent increase in cellular p53. (A) U2OS cells were exposed to mitomycin C at the indicated doses for 16 h and then harvested, and the level of p53 was measured by ELISA. (B) U2OS cells were exposed to 6 µM mitomycin C for either 6, 16, or 40 h and then harvested, and the level of p53 was measured by ELISA. Error bars represent the standard deviation.

membrane carrier proteins for Cr3+, as Cr3+ does seem to function as the main genotoxic agent within the cell (39). We found that Cr3+ did not cause p53 induction at any of the doses tested (Figure 2A), while Cr6+ caused an increase in p53 at a dose of 0.7 µM, and nearly a 3-fold increase when added at a dose of 1.3 µM (Figure 2B). Doses of 6.7 and 13 µM were also tested but led to significant cytotoxicity. The p53 response to Cr6+ was also time-dependent (Figure 2C). These results are consistent with previous reports of the genotoxicity of these two compounds. iAsIII Induces p53. To examine the ability of inorganic arsenic compounds to induce p53, sodium arsenite (trivalent) (iAsIII) and sodium arsenate (pentavalent) (iAsV) were added to U2OS cells and the resultant cellular levels of p53 were measured by ELISA. The results for iASIII are shown in Figure 3 and demonstrate that trivalent inorganic arsenic causes an increase in cellular p53 that is both dose- (Figure 3A) and time(Figure 3B) dependent. For this compound, the maximal effect was obtained at a dose of 20 µM (16 h), with approximately a 2-fold induction. A dose of 100 µM was also tested but led to significant cytotoxicity that resulted in degradation and elimination of cellular p53. The maximal signal was obtained at 40 h (10 µM), consistent with the suggestion that arsenic acts in an indirect manner. The results obtained for pentavalent arsenic (iASV) do not support a statistically significant dose and time relationship (Figure 4). Doses above 200 µM were tested (1, 10, and 100 mM) and led to significant cytotoxicity. Dimethylated but Not Monomethylated Arsenic Compounds Induce p53. Previous reports have indi-

Figure 2. Cr+6 but not Cr3+ causes a dose- and time-dependent increase in cellular p53. (A) U2OS cells were exposed to the indicated concentrations of Cr3+ for 16 h, and the resulting level of cellular p53 was measured by ELISA. (B) U2OS cells were exposed to the indicated concentrations of Cr6+ for 16 h, and the resulting level of cellular p53 was measured by ELISA. (C) U2OS cells were exposed to either Cr3+ or Cr6+ for the indicated times, and the resulting level of p53 was measured by ELISA. Error bars represent the standard deviation. The Student’s onetailed t-test was used to determine statistical significance, with “*” representing a >0.95 level of confidence and “**” representing a >0.99 level of confidence.

cated that the monomethylated and dimethylated forms of arsenic may differ from the inorganic forms of arsenic with respect to their ability to cause genotoxic damage (8 and references therein). To examine this possibility in our system, we treated U2OS cells with both the monomethylated and the dimethylated forms of arsenic III and arsenic V and examined the resulting levels of cellular p53. The results for monomethylated arsenic III (MAsIII) (Figure 5A) indicate that this compound is unable to detectably induce p53 at any dose tested in our assay, suggesting that under these conditions, it is not genotoxic. We were unable to obtain p53 values for higher doses of this compound, as doses higher than 1 µM (10 and 100 µM) led to significant cytotoxicy of the cells. In contrast, the results from the dimethylated compound, DMAsIII (Figure 5B,C), do provide some evidence of p53 induction. The highest signals, of approximately a 70%

Arsenic Induces p53

Figure 3. iAsIII causes a dose- and time-dependent increase in cellular p53. (A) U2OS cells were exposed to the indicated concentrations of iAsIII for 16 h, and the resulting level of cellular p53 was measured by ELISA. (B) U2OS cells were treated with 10 µM iAsIII for the indicated times, and the resulting level of cellular p53 was measured by ELISA. Error bars represent the standard deviation. The Student’s one-tailed t-test was used to determine statistical significance, with “*” representing a >0.95 level of confidence and “**” representing a >0.99 level of confidence.

increase, were obtained at a dose of 20 µM for 16 h and at 40 h for a dose of 40 µM. Some cytotoxicity was apparent at the 100 µM dose (16 h). The results from the pentavalent compounds were similar. The monomethylated compound (MASV) was unable to detectably increase cellular p53 levels at any of the doses tested (from 1 µM to 10 mM) when assayed at either 16 or 40 h posttreatment (Figure 6A). A dose of 100 mM led to significant cytotoxicity. In contrast, the dimethylated compound (DMASV) did increase the level of p53 in a dose- and time-dependent manner, with maximal signals obtained at a dose of 500 µM for 16 h (Figure 6B) and at a time of 40 h (for 1 mM) (Figure 6C). Some cytotoxicity was apparent at the 1000 µM dose. Arsenic Trioxide Induces p53. Arsenic trioxide has been used in the treatment of hematological cancers such as promyelocytic leukemia (reviewed in 40). Although it is known to cause cell death, the molecular mechanisms underlying this activity have not been clearly defined. To see if this compound elicited the DNA damage response from treated cells, we treated U2OS cells with arsenic trioxide and measured the resulting p53 levels. The results shown in Figure 7 demonstrate that this compound produces a robust increase of p53 within 16 h, with maximal induction observed at a dose of 20 µg/ mL. This increase was even more pronounced by 40 h. A higher tested dose, 50 µM, led to significant cytotoxicity. These results indicate that arsenic trioxide is a potent genotoxic agent and may exert its therapeutic effects at least in part by triggering the p53-mediated apoptotic pathway in tumor cells.

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 427

Figure 4. iAsV does not cause a significant increase in cellular p53. (A) U2OS cells were exposed to the indicated concentrations of iAsV for 16 h, and the resulting level of cellular p53 was measured by ELISA. (B) U2OS cells were treated with 200 µM iAsV for the indicated times, and the resulting level of cellular p53 was measured by ELISA. Error bars represent the standard deviation.

Discussion Our previous work has shown that p53 induction can serve as a marker of DNA damage and thus be used to identify genotoxic compounds (33, 34). One attractive feature of the p53 induction assay is that it can be used with human cells, thus eliminating the need for species to species extrapolation. Another is that it can detect DNA damage that is repaired before DNA replication occurs; that is, it does not require that the damage be inherited (as opposed to assays such as the Ames test). The p53 assay also does not require that the damage occur at a specific locus. It therefore has the potential to detect DNA damage that would not be detected using other methods. For example, diethylstilbesterol causes DNA damage that is quickly repaired (41, 42); it yields a negative result on the Ames test (43) but a positive result from our p53 induction assay (33). The determination of whether arsenic compounds can function as genotoxins is important in its risk assessment. Traditional risk assessment approaches are based on the principle that for nongenotoxic substances, there is a “safe” dose, below which no detectable harm is likely to occur. This assumption is not made for genotoxic substances, as they cause DNA damage that can accumulate over time. In theory, damage caused by such a substance to just one DNA base could ultimately lead, with some low probability (and assuming a series of other events also occur), to the development of cancer. Therefore, there may not be a safe level of exposure to genotoxic substances, and any regulated level is likely to be lower than for a nongenotoxic counterpart. In the case of arsenic compounds, however, risk assessment becomes difficult because its effects do not fit

428

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

Figure 5. DMAsIII but not MAsIII causes a dose- and timedependent increase in cellular p53. (A) U2OS cells were exposed to the indicated concentration of MAsIII for 16 h, and the resulting level of cellular p53 was measured by ELISA. (B) U2OS cells were exposed to the indicated concentration of DMAsIII for 16 h, and the resulting level of cellular p53 was measured by ELISA. (C) U2OS cells were exposed to 4 µM DMAsIII for the indicated times, and the resulting level of cellular p53 was measured by ELISA. Error bars represent the standard deviation. The Student’s one-tailed t-test was used to determine statistical significance, with “*” representing a >0.95 level of confidence and “**” representing a >0.99 level of confidence.

the traditional profile of either a genotoxic or a nongenotoxic compound. While they clearly can cause cancer in humans, arsenic compounds are generally acknowledged to be nongenotoxic and nonmutagenic as judged by traditional assays. In particular, they are negative in the Ames assay and do not easily cause cancer in experimental animals. Some arsenic compounds are known to cause DNA damage such as single-strand breakage, with some studies attributing this effect to an inhibition of DNA repair. Arsenic compounds can also cause cytogenic effects such as chromosomal aberrations, sister chromatid exchange, and micronucleus formation, and there is evidence that arsenic can work as a comutagen (reviewed in ref 4). A complicating factor is the fact that arsenic compounds are metabolized to various forms. Arsenate, for example, can be reduced, monomethylated and dimethylated (44). There is also a lack of agreement regarding the relative toxicity of the various forms of

Filippova and Duerksen-Hughes

Figure 6. DMAsV but not MAsV causes a dose- and timedependent increase in cellular p53. (A) U2OS cells were exposed to the indicated concentrations of MasV for either 16 or 40 h, and the resulting level of cellular p53 was measured by ELISA. (B) U2OS cells were exposed to the indicated concentrations of DMAsV for 16 h, and the resulting level of cellular p53 was measured by ELISA. (C) U2OS cells were exposed to 1 mM DMAsV for 6, 16, or 40 h, and the resulting level of cellular p53 was measured by ELISA. Error bars represent the standard deviation. The Student’s one-tailed t-test was used to determine statistical significance, with “**” representing a >0.99 level of confidence.

Figure 7. Trisenox causes a dose- and time-dependent increase in cellular p53. (A) U2OS cells were exposed to the indicated concentrations of trisenox for 16 or 40 h, and the resulting level of cellular p53 was measured by ELISA. Error bars represent the standard deviation. The Student’s one-tailed t-test was used to determine statistical significance, with “*” representing a >0.95 level of confidence and “**” representing a >0.99 level of confidence.

arsenic, with some studies indicating decreased toxicity for the methylated compounds, while other studies suggest an increased toxicity following methylation (20 and references therein). This extensive collection of data,

Arsenic Induces p53

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 429

Table 2. Induction of p53 by Chromium and Arsenic Compounds compd

concentrations

doseresponse

timedependent

p53 inductiona

Mit C Cr3+ Cr6+ IAsIII IAsV MAsIII DMAsIII MasV DMAsV As2O3

1.2-12 µM 0.25-50 µM 0.7-3.3 µM 0.4-20 µM 4-200 µM 40 nM-1 µM 0.1-100 µM 1 µM-10 mM 40 µM-1 mM 0.5-20 µM

yes no yes yes no no yes no yes yes

yes no yes yes no no yes no yes yes

++++ ++ ++ + + +++

a Ranking system based on highest observed response: -, less than 150% of untreated; +, 150-250%; ++, 250-350%; +++, 350-450%; ++++, greater than 450%.

unfortunately, has not readily led to the development of a generally accepted mechanism underlying the basis for cancer development in humans. In turn, this lack of a mechanistic understanding has hindered the development and acceptance of safe levels for the presence of these substances in drinking water. Most of the previous studies regarding the effects of arsenic compounds examined the effects of only one or two compounds, often with a very limited sampling of time points and administered doses. The p53 induction assay, which measures a cell’s response to DNA damage in a quantifiable manner, seemed a promising way to develop a more detailed understanding of the cellular events occurring following exposure to potentially DNAdamaging arsenic compounds and therefore was applied to seven such compounds. We found that the inorganic compound sodium arsenate induced p53, that the monomethyl derivatives methyloxoarsine and disodium methyl arsonate did not, that the dimethylated derivatives (iododimethylarsine and dimethylarsinic acid) also did so (although to a lesser extent), and that the chemotherapy agent, arsenic trioxide, was the strongest inducer of p53 within this group (see summary in Table 2). The time course of p53 induction for each of these compounds was most consistent with an indirect mode of DNA damage, as maximal induction occurred later than 12 h postexposure. Earlier reports (5, 14, 17, 45) have shown that arsenite (iAs3+) can induce DNA damage through oxidative damage 1-4 h postexposure, with the cells losing viability within 72 h to 7 days, depending on the cell type. p53 induction in human fibroblasts following arsenite treatment has been reported to occur either as early as 1 h (45) or as late as 18 h or even 14 days (depending on the dose) postexposure (46). Our data with arsenite showed a small increase in p53 within 6 h that became greater and more statistically significant with time (Figure 3B). A model in which DNA damage leads to p53 induction, which could then lead to cell death, would predict that p53 increases would be observed later than the actual DNA damage and before cell death, consistent with our actual observations. These results indicate that cells respond to arsenic exposures as they would to DNA-damaging agents and provide evidence that several arsenic compounds can function as genotoxins under at least some conditions. Previous work that has examined the ability of inorganic arsenic compounds to induce p53 has yielded conflicting results. For example, arsenite at doses ranging from 0.1 to 50 µM was shown to increase p53 protein

expression in Jurkat, HeLa, LCL-EBV, WI38, and human fibroblast cells (45-49). These cell lines differed with respect to the dose required for the maximum effect and the time of peak induction. In the case of the human fibroblast cells, events associated with p53 induction, such as phosphorylation of p53 and the accumulation of the p53-regulated proteins p21 and mdm-2 as well as a kinase known to phosphorylate p53, ATM (ataxia telangiectasia muated kinase), were also reported (45, 48, 49). Arsenic trioxide, the compound used to treat cancer, has been shown to up-regulate p53 in AGS and MKN-28 cells (50). In this study, the ability of the compound to create DNA strand breaks was confirmed by the comet assay. On the other hand, other studies show either a decrease (51) or no effect (52) on the level or transcriptional activity of p53. One group demonstrated that treatment with either arsenite or arsenate can induce cell death in both p53+ and p53- cells, indicating that at least one apoptotic pathway other than the p53-dependent pathway can operate in these cells, although this pathway was not identified (53, 54). Our work, which examined a broader spectrum of arsenic compounds than used in previous studies, as well as a relatively wide range of doses and times tested, provides evidence that the ability of arsenic to induce p53 may vary significantly depending on the exact chemical species. For example, four different arsenic species with a valency of +3 were tested as follows: iAsIII, the methylated version (MAsIII), the dimethylated version (DMAsIII), and arsenic trioxide. Of these, arsenic trioxide was the most potent, followed by iAsIII, DMAsIII, and DMAsV. These results suggest that a number of factors in addition to valency and the presence or absence of methyl groups are relevant in eliciting the DNA damage response in a cell. One factor that may be important is the ionic status of a compound; this could, for example, affect transport into and within the cell. Arsenic trioxide has been shown to elicit apoptosis in a number of systems (reviewed in ref 20). Our data is certainly consistent with its doing so in a p53-mediated manner, although it does not rule out the possibility that p53-independent pathways could also be used. It is perhaps noteworthy that the compound with the greatest ability to induce p53 is the one with a demonstrated effectiveness in cancer treatment, while those with a lower ability to induce p53 are associated rather with the causation of cancer. It may be that this low level of p53 induction is insufficient to trigger apoptosis in many cells, making it possible for mutations to become fixed into the genome, and, in the case of long-term exposure, for cells to accumulate damage such that neoplasms can develop. Other mechanisms for the cancer-causing ability of these compounds have been suggested, such as inhibiting DNA repair and working as a comutagen, and such mechanisms might well work together with any DNA-damaging effects exerted. One recent study that did find evidence of the genotoxicity of arsenic compounds examined the ability of six species, iAsIII, iAsV, MAsIII, DMAsIII, MAsV, and DMAsV, to damage naked DNA (assayed by nicking of φX174) and to degrade cellular DNA (as measured by the comet assay) (8). These researchers found that both the mono- and the dimethylated trivalent arsenic species could do so, although little or no activity was detected with the other substances examined. Consistent with these results, we find that DMAsIII induces cellular p53

430

Chem. Res. Toxicol., Vol. 16, No. 3, 2003

but find in addition that iAsIII and DMAsV also do so. Conversely, while the previous study found genotoxic activity for the monomethylated trivalent compound (MAsIII), we did not find evidence that this compound induced p53. We expect the differences in these results to reflect the respective natures of the assay systems. The DNA cleavage assay is a very direct system for measuring DNA damage and will detect only those compounds immediately capable of damaging DNA. The comet assay captures the ability of the test compound to enter a cell in addition to its ability to cleave DNA. In contrast, the p53 induction assay measures the induction of a cell’s typical response to DNA damage by the combination of all events following exposure. Indeed, the time course we noted for p53 induction (effect maximal after 12 or more hours) argues that indirect mechanisms, rather than direct DNA damage mechanisms, provide the primary driving force. A number of cellular events can occur that follow exposure and precede DNA damage and p53 induction, and the influence of variable cellular environments on any or all of these factors could underlie the variable results noted in the literature. The compounds can, of course, be metabolized into products that are more or less genotoxic than the original parent compound. One or more of these products could then work either by damaging DNA or by causing cellular events that lead less directly to DNA damage. One suggested possibility is that these compounds act by altering the cellular oxidative state (9-12). Ultimately, the influence of arsenic compounds on human health will be mediated by many factors, including the ability of the various metabolites to cause DNA damage. Our results demonstrate that cells respond to several arsenic compounds by increasing their level of p53, indicating that exposure to these substances triggers the DNA damage sensing and response pathways and providing evidence that these compounds can indeed act as genotoxins, most likely by one or more indirect mechanisms.

Acknowledgment. This work was supported by the National Medical Technology Testbed (NMTB) and Loma Linda University. The project or effort depicted was sponsored by the Department of the Army under Cooperative Agreement DAMD17-97-2-7016. The content of the information does not necessarily reflect the position or the policy of the government or NMTB, and no official endorsement should be inferred. We especially thank Dr. Marc Mass for samples of iododimethylarsine (DMAsIII) and methyloxoarsine (MAsIII).

References (1) Jager, J. W., and Ostosky-Wegman, P. (1997) Arsenic: a paradoxical human carcinogen. Mutat. Res. 386, 181-184. (2) Shelby, M. D. (1988) The genetic toxicity of human carcinogens and its implications. Mutat. Res. 204, 3-15. (3) Anderson, K. J., Leighty, E. G., and Takahashi, M. T. (1972) Evaluation of herbicides for possible mutagenic properties. J. Agric. Food Chem. 20, 649-656. (4) Basu, A., Mahata, J., Gupta, S., and Giri, A. K. (2001) Genetic toxicology of a paradoxical human carcinogen, arsenic: a review. Mutat. Res. 448, 171-194. (5) Wang, T.-S., Hsu, T.-Y., Chung, C.-H., Wang, A. S. S., and et al. (2001) Arsenite induces oxidative DNA adducts and DNA-protein cross-links in mammalian cells. Free Radical Biol. Med. 31 (3), 321-330. (6) Schaumloffel, N., and Gebel, T. (1998) Heterogeneity of the DNA

Filippova and Duerksen-Hughes

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16) (17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26) (27)

(28)

(29)

(30)

damage provoked by antiomony and arsenic. Mutagenesis 13 (3), 281-286. Hartmann, A., and Speit, G. (1994) Comparative investigations of the genotoxic effects of metals in the single cells gel (SCG) assay and the sister chromatid exchange (SCE) test. Environ. Mol. Mutagen. 23 (4), 299-9305. Mass, M. J., Tennant, A., Roop, B. C., Cullen, W. R., and et al. (2001) Methylated trivalent arsenic species are genotoxic. Chem. Res. Toxicol. 14, 355-361. Jing, Y., Dai, J., Chalmers-Redman, R. M. E., Tatton, W. G., and et al. (1999) Arsenic trioxide selectively induces acute promyelocyticleukemia cell apoptosis via a hydrogen peroxide-dependent pathway. Blood 94 (6), 2102-2111. Kessel, M., Liu, S. W., Xu, A., Santella, R., and et al. (2002) Arsenic induces oxidative DNA damage in mammalian cells. Mol. Cell. Biochem. 234-235 (1-2), 301-308. Liu, S. W., Athar, M., Lippai, I., Waldren, C., and et al. (2001) Induction of oxyradicals by arsenic: Implication for mechanism of genotoxicity. Proc. Natl. Acad. Sci. U.S.A. 98 (4), 1643-1648. Hei, T. K., Liu, S. X., and Waldren, C. (1998) Mutagenicity of arsenic in mammalian cells: Role of reactive oxygen species. Proc. Natl. Acad. Sci. U.S.A. 95, 8103-8107. Wang, H.-G., Rikitake, Y., Carter, M. C., Yaciuk, P., and et al. (1993) Identification of specific adenovirus E1A N-terminal residues critical to the binding of clelular proteins and to the control of cell growth. J. Virol. 67, 476-488. Liu, F., and Jan, K.-Y. (2000) DNA damage in arsenite- and cadmium-treated bovine aortic endothelial cells. Free Radical Biol. Med. 28 (1), 55-63. Scott, N., Hatleid, K. M., MacKenzie, N. E., and Carter, D. E. (1993) Reactions of arsenic(III) and arsenic(V) species with glutathione. Chem. Res. Toxicol. 6, 102-106. Gebel, T. (1997) Arsenic and antimony: comparative approach on mechanistic toxicology. Chem. Biol. Interact. 107, 131-144. Lynn, S. N., Shiung, J.-N., Gurr, J.-R., and Jan, K. Y. (1998) Arsenite stimulates poly(ADP-ribosylation) by generation of nitric oxide. Free Radical Biol. Med. 24 (3), 442-449. Slack, J. L., Waxman, S., Tricot, G., Tallman, M. S., and et al. (2002) Advances in the management of acute promyelocytic leukemia and other hematologic malignancies with arsenic trioxide. Oncologist 7, 1-13. Murgo, A. J. (2001) Clinical trials of arsenic trioxide in hematologic and solid tumors: overview of the National Cancer Institute Cooperative Research and Development Studies. Oncologist 6, 22-28. Bode, A. M., and Dong, Z. (2002) The paradox of arsenic: molecular mechanisms of cell transformation and chemotherapeutic effects. Crit. Rev. Oncol./Hemat. 42, 5-24. Preston, R. J., and Hoffmann, G. R. (2001) Genetic Toxicology. In Casarett & Doull’s Toxicology (Klaassen, C. D., Ed.) pp 331342, McGraw-Hill, New York. Zeiger, E., Haseman, J. K., Shelby, M. D., Margolin, H., and et al. (1990) Evaluation of four in vitro genetic toxicity tests for predicting rodent carcinogenicity: Confirmation of early results with 41 additional chemicals. Environ. Mol. Mutagen. 16, 1-14. Tennant, R. W., Margolin, B. H., Shelby, M. D., Zeiger, E., and et al. (1987) Prediction of chemical carcinogenicity in rodents from in vitro genetic toxicity assays. Science 236, 933-941. Tennant, R. W., Spalding, J. W., Stasiewicz, S., Caspary, W. D., and et al. (1987) Comparative evaluation of genetic toxicity patterns of carcinogens and noncarcinogenes: strategies for predictive use of short-term assays. Environ. Health Perspect. 75, 87-95. Tennant, R. W. (1991) The Genetic Toxicology Database of the National Toxicology Program: Evaluation of the relationships between genetic toxicology and carcinogenicity. Environ. Health Perspect. 96, 47-51. Wachsman, J. T. (1993) Predicting chemical carcinogenesis in rodents. Environ. Health Perspect. 101, 444-445. Liu, Y., and Kulesz-Martin, M. (2001) p53 protein at the hub of cellular DNA damage response pathways through sequencespecific and nonsequence-specific DNA binding. Carcingenesis 22, 851-860. Colman, M. S., Afshari, C. A., and Barrett, J. C. (2000) Regulation of p53 stability and activity in response to genotoxic stress. Mutat. Res. 462, 179-188. Appella, E., and Anderson, C. W. (2001) Posttranslational modifications and activation of p53 by genotoxic stresses. Eur. J. Biochem. 268, 2764-2772. El Deiry, W. S. (1988) Regulation of p53 downstream genes. Semin. Cancer Biol. 8, 345-357.

Arsenic Induces p53 (31) Lopes, U. G., Erhardt, P., Yao, R., and Cooper, G. M. (1997) p53dependent induction of apoptosis by proteasome inhibitors. J. Biol. Chem. 372, 12893-12896. (32) Shinohara, K., Tomioka, M., Nakano, H., Tone, S., and et al. (1996) Apoptosis induction resulting from proteasome inhibition. Biochem. J. 317, 385-388. (33) Yang, J., and Duerksen-Hughes, P. (1998) A new approach to identifying genotoxic carcinogens: p53 induction as an indicator of genotoxic damage. Carcinogenesis 19, 1117-1125. (34) Duerksen-Hughes, P. J., Yang, J., and Ozcan, O. (1999) p53 induction as a genotoxic test for twenty-five chemicals undergoing in vivo carcinogenicity testing. Environ. Health Perspect. 107 (10), 805-812. (35) ATSDR. (2000) Toxicological Profile for Arsenic (Update), U.S. Department of Health and Human Services. (36) ATSDR. (2000) Toxicological Profile for Chromium. U.S. Department of Health and Human Services. (37) De Flora, S., Bagnasco, M., Serra, D., and Zanacchi, P. (1990) Genotoxicity of chromium compounds. A review. Mutat. Res. 238, 99-172. (38) Katz, S. A., and Salem, H. (1993) The toxicology of chromium with respect to its chemical speciation: a review. J. Appl. Toxicol. 13 (3), 217-224. (39) Bianchi, V., and Levis, A. G. (1988) Review of genetic effects and mechanisms of action of chromium compounds. Sci. Total Environ. 71 (3), 351-355. (40) Novick, S. C., and Warrell, R. P. (2000) Arsenicals in hematologic cancers. Semin. Oncol. 27 (5), 495-501. (41) Gladek, A., and Liehr, J. G. (1989) Mechanism of genotoxicity of diethylstilbestrol in vivo. J. Biol. Chem. 264, 16847-16852. (42) Liehr, J. G. (1990) Genotoxic effects of estrogens. Mutat. Res. 238, 269-276. (43) Glatt, H. R., Metzler, M., and Oesch, F. (1979) Diethylstilbestrol and 11 derivatives: a mutagenicity study with Salmonella typhimurium. Mutat. Res. 67 (2), 113-121. (44) Sampayo-Reyes, A., Zakharyan, R. A., Healy, S. M., and Aposhian, H. V. (2000) Monomethylarsonic acid reductase and monomethyl-

Chem. Res. Toxicol., Vol. 16, No. 3, 2003 431

(45) (46)

(47) (48) (49) (50)

(51)

(52) (53) (54)

arsonous acid in hamster tissue. Chem. Res. Toxicol. 13, 693697. Yih, L. H., and Lee, T. C. (2000) Arsenite induces p53 accumulation through an ATM-dependent pathway in human fibroblasts. Cancer Res. 60, 6346-6352. Vogt, B. L., and Rossman, T. G. (2001) Effects of arsenite on p53-, p21 and cyclin D expression in normal human fibroblasts - - a possible mechanism for arsenite’s comutagenicity. Mutat. Res. 478, 159-168. Salazar, A. M., Ostrosky-Wegman, P., Menendez, D., Miranda, E., and et al. (1997) Induction of p53 protein expression by sodium arsenite. Mutat. Res. 381, 259-265. Yih, L. H., Ho, I. C., and Lee, T. C. (1997) Sodium arsenite disturbs mitosis and induces chromosome loss in human fibroblasts. Cancer Res. 57, 5051-5059. Yih, L. H., and Lee, T. C. (1999) Effects of exposure protocols on induction of kinetochore-plus and -minus micronuclei by arsenite in diploid human fibroblasts. Mutat. Res. 440, 75-82. Jiang, X. H., Wong, B. C., Yuen, S. T., Jiang, S. H., and et al. (2001) Arsenic trioxide induces apoptosis in human gastric cancer cells through up-regulation of p53 and activation of caspase-3. Int. J. Cancer 91, 173-179. Hamadeh, H. K., Vargas, M., Lee, E., and Menzel, D. B. (1999) Arsenic disrupts cellular levels of p53 and mdm2: a potential mechanism of carcinogenesis. Biochem. Biophys. Res. Commun. 263, 446-449. Huang, C., Ma, W. Y., Li, J., and Dong, Z. (1999) Arsenic induces apoptosis through a c-Jun NH2-terminal kinase-dependent, p53independent pathway. Cancer Res. 59, 3053-3058. Harvey, M., Sands, A. T., Weiss, R. S., and et al. (1993) In vitro growth characteristics of embryo fibroblasts isolated from p53deficient mice. Oncogene 8, 2456-2467. Huang, C., Ma, W. Y., Li, J., Hecht, S. S., and et al. (1998) Essential role of p53 in phenethyl isothiocyanate-induced apoptosis. Cancer Res. 58, 4102-4106.

TX025606A