Contribution of dichloroacetate and trichloroacetate to liver tumor induction in mice by trichloroethylene

Toxicology and Applied Pharmacology 182, 55– 65 (2002)
doi:10.1006/taap.2002.9427
Contribution of Dichloroacetate and Trichloroacetate to Liver Tumor Richard J. Bull,*,†,2 Gayle A. Orner,* Rita S. Cheng,* Lisa Stillwell,* Anja J. Stauber,† Lyle B. Sasser,* Melissa K. Lingohr,† and Brian D. Thrall*,† *Molecular Biosciences Department, Pacific Northwest National Laboratory, Richland, Washington 99352; and †Pharmacology/Toxicology Program, Washington State University, Pullman, Washington 99163 Received August 13, 2001; accepted April 13, 2002 tumors were c-Jun؉, 13 were c-Jun؊, and 9 had a mixed pheno-
Contribution of Dichloroacetate and Trichloroacetate to Liver
type. Mutations of the H-ras protooncogene were also examined in
Tumor Induction in Mice by Trichloroethylene. Bull, R. J., Orner,
DCA-, TCA-, and TRI-induced tumors. The mutation frequency
G. A., Cheng, R. S., Stillwell, L., Stauber, A. J., Sasser, L. B.,
detected in tumors induced by TCA was significantly different
Lingohr, M. K., and Thrall, B. D. (2002). Toxicol. Appl. Pharma-
from that observed in TRI-induced tumors (0.44 vs 0.21, p < 0.05),
col. 182, 55– 65.
whereas that observed in DCA-induced tumors (0.33) was inter-
Determining the key events in the induction of liver cancer in
mediate between values obtained with TCA and TRI, but not
mice by trichloroethylene (TRI) is important in the determination
significantly different from TRI. No significant differences were
of how risks from this chemical should be treated at low doses. At
found in the mutation spectra of tumors produced by the three
least two metabolites can contribute to liver cancer in mice, di-
compounds. The presence of mutations in H-ras codon 61 ap-
chloroacetate (DCA) and trichloroacetate (TCA). TCA is pro-
peared to be a late event, but ras-dependent signaling pathways
duced from metabolism of TRI at systemic concentrations that can
were activated in all tumors. These data are not consistent with the
clearly contribute to this response. As a peroxisome proliferator
hypothesis that all liver tumors induced by TRI were produced by
and a species-specific carcinogen, TCA may not be important in
TCA. 2002 Elsevier Science (USA)
the induction of liver cancer in humans at the low doses of TRI
Key Words: trichloroethylene; dichloroacetate; trichloroacetate;
encountered in the environment. Because DCA is metabolized
trichloroethylene; mutation analysis; H-ras; c-Jun; liver tumors.
much more rapidly than TCA, it has not been possible to directly
determine whether it is produced at carcinogenic levels. Unlike
TCA, DCA is active as a carcinogen in both mice and rats. Its
Trichloroethylene (TRI) is a common contaminant of ground low-dose effects are not associated with peroxisome proliferation.
The present study examines whether biomarkers for DCA and
water as a result of poor disposal practices of the past. As a TCA can be used to determine if the liver tumor response to TRI
consequence, this solvent is the focus of many cleanup oper- seen in mice is completely attributable to TCA or if other metab-
ations of uncontrolled hazardous waste sites. TRI is carcino- olites, such as DCA, are involved. Previous work had shown that
genic in both mice and rats, but at different sites, the liver and DCA produces tumors in mice that display a diffuse immunore-
kidney, respectively (NCI, 1976; NTP, 1988; NTP, 1990).
activity to a c-Jun antibody (Santa Cruz Biotechnology, SC-45),
Liver tumor induction in mice has been the tumor most critical whereas TCA-induced tumors do not stain with this antibody. In
from the standpoint of environmental regulation (Bull, 2000).
the present study, we compared the c-Jun phenotype of tumors
Under the proposed cancer risk guidelines of the U.S. Envi- induced by DCA or TCA alone to those induced when they are
ronmental Protection Agency (EPA, 1996), identifying the given together in various combinations and to those induced by
TRI given in an aqueous vehicle. When given in various combi-
dose–response behavior of key events involved in carcinogenic nations, DCA and TCA produced a few tumors that were c-Jun؉,
responses can be used for developing alternative risk assess- many that were c-Jun؊, but a number with a mixed phenotype
that increased with the relative dose of DCA. Sixteen TRI-induced
A major difficulty in developing alternative approaches for assessing risk from TRI is the fact that three of its metabolites G. A. Orner was supported by Associated Western Universities, Inc., are capable of inducing liver cancer in mice (Bull et al., 1990; Northwest Division (AWU NW) under Grant DE-FG06-89ER-75522 or DE-FG06-92RL-12451 with the U.S. Department of Energy (DOE). Research was Daniel et al., 1992; DeAngelo et al., 1999; Pereria, 1996). Two funded by U.S. DOE contract number DE-AC06 –76RL0 1830. A portion of of these metabolites have distinct modes of action, dichloro- this work was presented at the 37th Annual Meeting of the Society of acetate (DCA) and trichloroacetate (TCA). The third metabo- Toxicology, March, 1998, Seattle, WA.
lite, chloral hydrate, is probably active as a result of its con- To whom correspondence should be addressed at Washington State Uni- version to one or both of these two metabolites. DCA inhibits versity–Tri-Cities, 2710 University Drive, Richland, WA 99352. Fax: (509)372-7552; E-mail: dbull@tricity.wsu.edu.
cell replication and apoptosis (Snyder et al., 1995) but stimu- lates cell division in tumor cells that are immunoreactive to Biosystems (Foster City, CA). All other chemicals were obtained from Sigma c-Jun antibodies (Stauber and Bull, 1997). On the other hand, TCA produces tumors that are uniformly negative with respectto the c-Jun phenotype. This phenotype is consistent with its activity as a peroxisome proliferator (DeAngelo et al., 1989; Male B6C3F1 mice were purchased from Charles River Laboratories (Ra- leigh, NC) at 31–35 days of age with experiments beginning at 6 weeks of age Ordinarily, the first approach to assigning causality to a for all experiments. Protocols and animal care were approved by the Institu- metabolite in tumorigenesis would be an attempt to measure its tional Animal Care and Use Committtee at Pacific Northwest National Labo- concentration in the body and associating that with tumorigenic ratory. Animals were housed four to six per cage in shoebox cages and had free concentrations observed when the compound is itself admin- access to NIH-07 rodent chow and drinking water at all times. Control, TCA,and DCA drinking water solutions were prepared from deionized water and istered. This can be done with relative ease with TCA (Fisher, adjusted with sodium hydroxide to pH 7.0 Ϯ 0.2. Animal rooms were main- 2000; Clewell et al., 2000). However, it has been shown that tained on a 12-h light– dark cycle with temperatures controlled to 22–24°C and DCA administered directly in drinking water induces liver cancer in mice where blood concentrations are below thedetection limit. (Kato-Weinstein et al. 1998; Merdink et al., Mutations in the ras protooncogene have been used to de- Experiment 1.
After a 1-week acclimation period, animals were randomly assigned to treatment groups and given 0, 0.5, or 2 g/L DCA or 2 g/L TCA in termine if distinct patterns of DNA sequence alterations can their drinking water for the duration of the study. The doses of DCA were provide indications of the type of DNA damage that might be selected to bracket the highest and lowest doses that have produced hepatic produced by carcinogens. The presence of ras mutations in tumors without adversely affecting body weight gain. The lowest of these chemically induced tumors was originally suggested as a doses result in blood levels of DCA and TCA that roughly match (Kato- means of determining whether a chemical was genotoxic Weinstein et al., 1998) those produced by administering the lowest dose of TRIthat was previously shown to be carcinogenic (Merdink et al., 1998; NTP, (Reynolds et al., 1987). However, the discovery that sponta- 1990). A subset of mice was sampled after 52 weeks for high doses and neous tumors also contain this oncogene indicated that this associated control mice. Animals treated at 0.5 g/L were treated for an assumption was incorrect (Fox and Watanabe, 1985). Several additional 35 weeks and an equal number of control animals were maintained nongenotoxic carcinogens have been shown to produce liver for the same period. Times of euthanasia were selected based upon prior tumors in male B6C3F1 mice with a mutation frequency con- studies that established the appropriate latent periods for liver tumor develop-ment at these doses (Bull et al., 1990; Stauber and Bull, 1997). Only four siderably below those that result spontaneously. Among these animals were lost to miscellaneous causes in this experiment. There was no chemicals are peroxisome proliferators (Maronpot et al., evidence of treatment-related death. Tumors were identified at necropsy and 1995). DCA and TRI were found to induce tumors with similar their diameters were measured in two dimensions (longest and shortest) to mutation spectra (Anna et al., 1994), whereas only limited data obtain a mean diameter. Histopathological examination was limited to 15 have been available on TCA (Fereira-Gonzalez et al., 1995).
randomly selected tumors to ensure that nonneoplastic lesions were not beingmisclassified.
Schroeder et al. (1997) found H-ras to be rare in liver tumorsinduced by 3.5 g/L DCA in female B6C3F1 mice and raise Experiment 2.
DCA and TCA were administered as a mixture to male B6C3F1 mice in drinking water. Twenty animals were assigned to each of 10 questions about the role such mutations play in DCA-induced groups that received the following concentrations of DCA or TCA in their drinking water for 52 weeks: 0, 0.5 g/L TCA, 2 g/L TCA, 0.1 g/L DCA, 0.5 In the present study, we examined the question of whether g/L DCA, 2 g/L DCA, 0.1 g/L DCA ϩ 0.5 g/L TCA, 0.5 g/L DCA ϩ 0.5 g/L TRI (administered in an aqueous vehicle) induces tumors that TCA, 0.1 g/L DCA ϩ 2 g/L TCA, or 0.5 g/L DCA ϩ 2 g/L TCA. The animals carry the c-JunϪ phenotype, which would be predicted if TCA were euthanized at 52 weeks, the livers were examined for gross lesions asindicated above, and the histological sections were made for immunostaining were entirely responsible. We also investigated whether vari- and examination by a pathologist. Only two mice were lost during the course ations in this phenotype could be produced by treating animals of this experiment (one in the 2 g/L DCA group and another in the 0.5 g/L with various mixtures of DCA and TCA in drinking water. In addition, the mutation frequency and spectra in the H-ras Experiment 3.
TRI was administered once daily, 7 days per week by codon 61 in tumors produced by DCA, TCA, and TRI were gavage in a 5% Alkamuls in distilled water vehicle to a group of 50 mice for investigated to determine whether the hypothesis that TCA a period of 79 weeks. This mode of administration was used because Ͼ10,000 alone was responsible for the liver tumors could be supported mg/L TRI would be necessary to achieve a carcinogenic dose. Preliminaryexperiments established that such high concentrations would not be consumed by mice. The use of the Alkamuls vehicle also avoided some of the pharma-cokinetic complications associated with corn oil gavage. A control group of 15 MATERIALS AND METHODS
mice received an equivalent volume of the vehicle by the same method ofadministration. At the time of euthanasia, the livers were removed, tumors were identified, the size of the lesions was measured, and the tissues weresectioned for examination by a pathologist and for immunostaining. There DCA was obtained from Fluka Chemical Corporation (Ronkonkoma, NY) were six gavage-associated deaths during the course of this experiment among and TCA was purchased from Aldrich Chemical Company (Milwaukee, WI).
a total of 10 animals that died with TRI treatment. No animals were lost in the Reagents for PCR and sequencing were purchased from Perkin–Elmer Applied CONTRIBUTION OF DCA AND TCA TO TRI–INDUCED LIVER TUMORS Tumor Sampling, PCR Amplification, and Sequencing spectra were compared using a mutation analysis program described by Cari-ello et al. (1994). A p value Յ 0.05 was considered significant.
At the conclusion of each experiment, animals were euthanized, livers were removed, and macroscopically visible lesions (tumors) were identified, mea- sured, and separated from surrounding tissue. A portion of tissue was excisedfrom 25 tumors per treatment group (where available), frozen in liquid nitro- The serial liver sections were stained with the c-Jun antibody, SC-45 (Santa gen, and stored at Ϫ80°C until used for ras mutation analysis. Remaining Cruz Biotechnology) in a 1 to 25 dilution by methods previously described portions of the tumor were either snap-frozen in liquid nitrogen for use in (Stauber and Bull, 1997). This antibody was raised against the sequence Western blotting or were fixed in 10% neutral buffered formalin for 24 h and TPTPTQFLCPKNVTD, which includes Thr 91 and 93 of mouse c-Jun. These then were transferred into 70% ethanol until they were paraffin-embedded and residues are phosphorylated by JNK and this activity is thought to lead to examined histologically. Tumor samples (and samples from liver not including dephosphorylation of c-Jun in the DNA-binding domain (Nakano et al., 1994), tumors in Experiment 1) for H-ras analysis were digested overnight at 50°C in which leads to activation of c-Jun as a transcription factor. Therefore, this is DNA lysis buffer (50 mM Tris–Cl, pH 8.0; 20 mM NaCl; 1 mM EDTA; 1% potentially a deactivated form of c-Jun that accumulates in liver tumors of male SDS; and 1 mg/mL proteinase K), proteinase heat inactivated by boiling, mice (Kato-Weinstein et al., 2001).
diluted 1:10 in water, and 1– 4 ␮L was used as template for PCR amplification.
DNA was amplified using a modification of the procedure described byManjanatha et al. (1996). The primer pairs GCCGCTGTAGAAGCTATGA and CTTGGTGTTGTTGATGGCAAATACA were used to amplify a 469-bpsection of H-ras containing the first and second exon. PCR reaction mixtures As previously reported (Bull et al., 1990), the percentage of contained 4 mM magnesium chloride, 10 mM Tris, 50 mM potassium chloride, body weight made up by liver (liver somatic index) was sig- 0.2 mM deoxynucleotides, 0.2 ␮M of each primer, and 0.01 U/␮L Thermusaquaticus DNA polymerase (Perkin–Elmer, Norwalk, CT). A second amplifi- nificantly elevated by DCA or TCA treatment at the doses and cation of an internal sequence containing codon 61 and adding M13 sequence periods of treatment that were examined in this study (Table 1).
to both ends was then conducted (primers: TGTAAAACGACGGCCAGTA- There were relatively minor effects on body weight. Only in CAGCCCAGGTCT-TGTA and CAGGAAACAGCTATGACCGTTGATG- the case of mixed treatments with DCA and TCA and with TRI GCAAATAC). After amplification, PCR products were purified using Micro- were there statistically significant effects on body weight. Mice con 100 filter units (Amicon, Beverly, MA) and sequenced on an automatedcycle sequencer (ABI 377 DNA sequencer). PCR products were sequenced in administered TRI for 79 weeks had body weights that were the forward direction using Perkin–Elmer Applied Biosystems dye primers and depressed 11% relative to concurrent controls. Administration confirmed by sequencing in the reverse direction using either dye primer of TRI in Alkamuls was done to obtain tumors independent of (Experiment 1) or D-rhodamine terminator (Experiment 2) cycle sequencing.
the corn oil vehicle that was used in previous studies. The DNA from several mutant tumors was reamplified and cloned using a parallel study using mixtures of DCA and TCA in drinking TOPO-TA Cloning Kit (Invitrogen, Carlsbad, CA), providing an additionaldegree of confidence in the detection of mutations.
water was done to determine whether the results obtained withTRI could be accounted for by these metabolites. The effects on body weight and somatic index of the liver in mice treatedwith TRI, the mixtures of metabolites, and their concurrent Tumors and adjacent noninvolved liver tissue were homogenized individu- controls are also provided in Table 1. The relatively small ally in ice-cold homogenization buffer (10 mM KPi, pH 7.5; 0.15 M KCl; 20%glycerol; 1 mM EDTA; 0.5 mM PMSF; 1.9 ␮g/mL aprotinin; 2 ␮g/mL tumor burden seen at 52 weeks had little effect on liver weight, leupeptin; and 200 ␮M sodium orthovanadate). Proteins were separated by so these data indicate that DCA and TCA produce a significant sodium dodecyl sulfate–polyacrylamide gel electrophoresis on 10% acryl- hepatomegaly. Effects seen on liver weight with longer treat- amide gels (Laemmili, 1970) and electrophoretically transferred to PVDF ment are complicated by increased tumor burdens.
membranes using a semidry blotter (Emprotech, Natick, MA). Blots were All treatments gave rise to tumorigenic responses in the liver probed with primary antibodies to Ras (Upstate Biotechnology, Lake Placid,NY), Mek (Transduction Laboratories, Lexington, KY), active Erk 1/2 (Pro- consistent with our past experience. The magnitude of the mega, Madison, WI), c-Fos, and insulin receptor ␤ (Santa Cruz Biotechnology, response to 2 g/L DCA and TCA was about 50% of that seen Santa Cruz, CA) followed by incubation with a horseradish peroxidase-linked in past experiments (Bull et al., 1990; Stauber and Bull, 1997; secondary antibody (Biorad, Hercules, CA) and the detection of immunoreac- DeAngelo et al., 1999). The tumor yield with TRI was slightly tive proteins by chemiluminescence (Amersham ECL kit, Amersham Corp., higher than was expected from prior NCI (1976) and NTP (1990) bioassays. All gross lesions in this experiment were examined by a pathologist and subjected to sequencing of theH-ras codon 61 and a random sample stained for c-Jun.
Sequencing chromatograms were compared using Sequencher software Tumor incidence and multiplicity (number of nodules, ade- (Ver. 3.0, Gene Codes Corporation, Ann Arbor, MI). Western blots were nomas, and carcinomas per animal) were significantly higher in scanned on a flatbed scanner and quantified on a Power Macintosh 7100 usingthe public domain software NIH Image, Version 1.57. Statistical analyses were animals treated with 2 g/L DCA or TCA for 52 weeks than in conducted using Sigma Stat Version 2.0 (Jandel Scientific, San Rafael, CA) control mice (Table 1). In addition, tumors were induced in except as otherwise noted. Animal weights were compared by one-way animals treated with 0.5 or 2 g/L DCA for 87 weeks. Mixtures ANOVA followed by the Tukey test for pairwise comparisons. Tumor inci- of DCA and TCA increased tumor induction in all combina- dence was compared using Fisher’s Exact test. Tumor size, multiplicity, and tions. TRI at a daily dose of 1 g/kg per day for 79 weeks also liver somatic indices were compared using the Kruskal–Wallis nonparametricone-way analysis of variance on ranks. Comparisons of mutation frequencies increased the gross tumor response significantly over the con- between treatments were tested for significance by ␹2 analysis and the mutation current vehicle control. The only treatment that did not in- Effect of Dichloroacetate and Trichloroacetate on Terminal Body Weight, Liver Somatic Index,
and Tumor Incidence, Number, and Mean Diameter
Note. Values are means Ϯ SEM.
a Total number of tumors divided by total number of animals.
b Drinking water vehicle.
c Significantly different from 52-week controls at p Ͻ 0.05.
d Significantly different from 79- or 87-week controls at p Ͻ 0.05.
e 5%Alkamuls in water vehicle.
crease tumor incidence or multiplicity was 0.1 g/L of DCA in nodules were identified. One of these was found to be a myelogenous lymphoid lesion. The remaining tumors found in The dose–response relationships between tumor responses the liver were hyperplastic nodules. TRI treatment significantly produced by DCA or TCA alone or in combination at 52 weeksof treatment are displayed in Fig. 1. In this figure, the results ofboth experiments using DCA given alone are combined asbeing more representative than found in the DCA alone treat-ment group in the mixture study. The experiments utilizingmixtures of DCA and TCA in drinking water produced re-sponses that were very close to additive when compared to alow dose of DCA (0.1 g/L) was combined with higher doselevels of TCA (0.5 and 2 g/L) when the chemicals wereadministered alone. However, DCA at 0.5 g/L failed to pro-duce a higher tumor incidence in combination with 2 g/L TCAthan was observed with TCA alone. Consequently, the onlyevidence of an interaction was a smaller number of tumors thanwould be expected in this latter group. However, the differencebetween the sum of the responses and that seen in the mixturewas not statistically significant.
Dose–response relationships for tumor induction by treating male Table 2 displays the pathological characterization of the B6C3F1 mice with varying doses of dichloroacetate (DCA) and trichloroac- tumors that were produced in the TRI experiment. Of the etate (TCA) alone and in combination in drinking water for 52 weeks. The tumors identified in control mice for the TRI study, only one x-axis provides the concentration of either DCA or TCA in drinking water. Thelegend indicates the concentration of the second compound when provided in was a carcinoma and four were adenomas. These occurred in combination. Results are expressed as mean number of tumors per animal Ϯ three animals, giving an incidence of 0.20 at 79 weeks of SEM. See Table 2 for more explicit description of incidence data and number treatment (Table 2). In addition, four histocytic proliferative CONTRIBUTION OF DCA AND TCA TO TRI–INDUCED LIVER TUMORS Pathological Characterization of Tumors Produced by Trichloroethylene and Selected Tumors
Induced by Dichloroacetate or Trichloroacetate
Note. Values in parentheses are percentages.
increased the incidence of all three classes of hepatocellular produced by TRI, DCA, and TCA expressed c-Jun. As previ- lesions over that observed in control mice. Adenomas were ously shown, the c-Jun immunoreactivity found in tumor cells observed at an incidence of 0.64, carcinomas at 0.19, and their was distributed primarily in the cytosolic compartment when it combined incidence was 0.75. Several mice had more than one was present. Earlier work by Stauber and Bull (1997) indicated adenoma and/or carcinoma (Table 2). Histocytic proliferative that the occurrence of a c-JunϪ lesion was diagnostic of TCA, nodules were observed at the same incidence in TRI-treated whereas almost all lesions induced by DCA were c-Jun posi- and control mice and including these lesions in the total count tive. This result was generally borne out in the present study, raises the incidence to 0.83 in the TRI-treated mice.
but a higher proportion of c-Jun tumors were observed when The pathological characterization of hepatic tumors from DCA was administered alone. Nevertheless, approximately mice treated with DCA and TCA and mixtures of these two 50% of DCA-induced tumors were c-Junϩ. When mixtures of metabolites are also displayed in Table 2. The incidence ofhyperplastic nodules, hepatocellular adenomas, and hepatocel- lular carcinomas are reported and the total number of lesions of Frequency of Immunoreactivity of Tumors Produced by DCA,
each type is also provided. The pathology data presented have TCA, and TRI to c-Jun Antibody (Santa Cruz, SC-45)
been confined to those treatments with DCA that were runconcurrently with the mixture studies to avoid confusing the results with longer periods of treatment. As previously shown, both DCA and TCA produce dose-related increases in tumor incidence within a 52-week period at doses of 0.5 g/L of drinking water and above. In general, DCA induces a greater number of tumors at a given dose, but a higher fraction of the tumors produced by TCA are hepatocellular carcinomas. Tu- mor responses with mixed exposures of the two haloacetates were significantly increased in the mixed exposures. In terms of the total tumor yields, the effects appeared to be approxi- mately additive. However, it was notable that the addition of DCA to fixed doses of TCA substantially enhanced the yield of hepatocellular adenomas and produced little, if any, enhance- ment of the incidence or total numbers of hepatocellular car- Table 3 indicates the relative frequency at which tumors Note. Values in parentheses are frequencies.
104 weeks. However, the present study utilized lower dosesand was of a shorter duration than the Ferreira-Gonzalez study.
The mutation spectrum of the TCA-treated animals from thecurrent study is not significantly different from that of thehistorical controls and the AAA sequence at codon 61 was themost common mutation.
When the H-ras mutation frequencies for all studies of DCA-induced tumors that were conducted below the maxi-mally tolerated dose are analyzed, a clear increase in the H-rasmutation frequency in tumors is seen with time of treatment (orage) (see Maronpot et al., 1995 as well as the data presented inthis manuscript). Data from different time points were notavailable for the TCA and TRI, however.
The H-ras mutations also seem to be a late event in TRI- Mutation frequency and spectra for the H-ras codon 61 in mouse induced hepatic tumors. In a limited subset of tumors that were liver tumors induced by TRI (1 g/kg per day), TCA (0.5 and 2 g/L drinkingwater), and DCA (0.5 and 2 g/L of drinking water). The number of tumors both sequenced and classified histologically, only 8 of 34 sequenced for each treatment is indicated by n. The mutation frequency seen (24%) of the adenomas contained codon 61 mutations while 9 in TCA-induced tumors was significantly different than those induced by TRI of 15 (60%) carcinomas contained mutant H-ras at this codon.
(p Ͻ 0.05 by Fisher’s Exact test). The difference in mutation frequency in TRI- The percentage of mutant sequence within each tumor (as and DCA-induced tumors was not statistically significant.
judged by the ratio of mutant to wild-type peak heights) waslower than would be expected if the tumors were the result of DCA or TCA were administered, a large fraction of the tumors clonal expansion of cells bearing mutant ras. Most tumors displayed a mixed c-Jun phenotype. While half the lesions seen contained less than 50% mutant sequence and only one was with DCA alone were c-Junϩ, the mixtures gave rise to tumors completely without wild-type sequence. The average Ϯ SE that were c-Junϩ in one part of the tumor and c-JunϪ in other ranged from 38 Ϯ 5% in the 0.5 g/L DCA treatment group to portions. When TRI was administered, 42% of the lesions 49 Ϯ 13% in the 2 g/L DCA for the 87 weeks group. There displayed a c-Junϩ phenotype, only 34% exhibited a c-JunϪ were no significant differences in this measure classified by phenotype, and 24% were of the mixed phenotype.
Mutation frequencies and spectra from tumors from animals Proteins involved in the MAP kinase-signaling cascade were treated with TRI, DCA, and TCA are shown in Fig. 2. No examined in order to determine if the three common codon 61 mutations were detected in DNA from normal tissue of mutations of ras had different effects on downstream effectors.
B6C3F1 mice. Few tumors were observed in control mice Figure 3 contains Western blots of TCA tumor and nonin- sampled at the times examined in this study (nine control volved tissue probed with antibodies to ras, Mek, active Erk tumors were sequenced, of these only two contained a codon 1/2, and c-Fos. The antibody used in the current experiment 61 mutation). Therefore, for statistical purposes, mutation data recognized several proteins in mouse liver homogenates. A from this study were compared to historical control data 28-kDa protein was present as a doublet in nontumor tissue, (Maronpot et al., 1995). However, we have chosen to depict but in tumor tissue only the higher-molecular-weight protein mutation spectra in a way that includes the wild-type sequence was present (Fig. 3). Ras has been reported to migrate as a (CAA), as it provides a simultaneous view of the mutation doublet in Jurkat cells treated with lovastatin, where the faster migrating band was identified as farnesylated, membrane- The H-ras codon 61 mutation frequency in tumors of mice bound p21 and the slower migrating band was identified as differed significantly (p Ͻ 0.05) between TRI and TCA (Fig.
unprocessed, cytosolic p21 (Goldman et al., 1996). It is pos- 2). Only 21% of the tumors from TRI-treated animals had a sible that the ras-immunoreactive protein accumulating in mutation in codon 61, whereas 44% of the tumors from TCA- TCA and DCA tumors is the unprocessed, inactive form. This treated animals had mutations. This difference in mutation is supported by our observation (not shown) that this protein is frequency was statistically different (p Ͻ 0.05) by Fisher’s primarily expressed in the cytosol of tumors. Conversely, our exact test. Tumors from DCA-treated mice were intermediate finding that levels of Mek and the dually phosphorylated (ac- in their frequency of H-ras mutations (33%), but this was not tive) forms of MAPK (p44Erk1 and p42Erk2) are strongly significantly different from tumors induced by TRI. In all induced in tumor tissue would not support this view. Each pair cases, these frequencies are below the frequency of codon 61 (T/NT) is from tumor vs adjacent, noninvolved tissue from the mutations in spontaneous tumors in this strain of mice (56%) same liver. A ras-immunoreactive protein is present as a single (Maronpot et al., 1995). The mutation frequency reported by intense band in tumors and as a doublet in the noninvolved Ferreira-Gonzalez et al. (1995) was 55% in 11 hepatocellular tissue (both identified by the single arrow in Fig. 3). Mek levels carcinomas of male B6C3F1 mice treated with 4.5 g/L TCA for are increased in tumor relative to adjacent tissue, as is phos- CONTRIBUTION OF DCA AND TCA TO TRI–INDUCED LIVER TUMORS Western blots of tissue extracts of tumors carrying different mutations in codon 61 of the H-ras gene and surrounding normal tissue for Ras, Mek, active Erk 1/2, and c-Fos. The high and low MW bands for Ras referred to in the text are both in the doublet identified by the arrow. Differences in proteinexpression were consistently detected when normal tissue was compared to tumors from the same animal. The expression of proteins within tumors did notdepend upon the sequence in H-ras codon 61.
phorylation of the MAP kinases Erk 1/2. These differences are DISCUSSION
also observed in tumors from control and DCA-treated ani-mals. Interestingly, there was no difference in the levels of any These results strongly suggest that TCA is not solely respon- of these signaling proteins between different mutation types or sible for liver tumors when mice are treated with TRI. The liver even between codon 61 mutant and wild-type tumors.
tumor response to TRI has been exclusively attributed to TCA Figure 4 provides information about the cell-signaling envi- in the past (Elcombe, 1985). However, the c-Junϩ phenotype ronment within DCA-induced tumors relative to surrounding is not produced when TCA is administered alone to B6C3F1 tissue and in tissue from control mice. These data indicate a mice (Stauber and Bull, 1997) nor by other peroxisome pro- consistent upregulation of the insulin receptor, ras activation, liferators (Nakano et al., 1994). Furthermore, the metabolites and ERK 1/2 activity in tumors relative to normal tissue. It is display the same proclivity for producing a particular pheno- of note that these same changes were seen in tumors that arose type when they stimulate the growth of colonies from suspen- spontaneously in control animals as well.
sions of primary mouse hepatocytes on soft agar (Stauber et al., Figure 5 provides more direct comparison of insulin receptor 1998). In the TRI-treated mice, two-thirds of the tumors that expression in the tumor and nontumor portions of liver from were immunohistochemically examined expressed the c-Junϩ mice treated with TRI, TCA, and two concentrations of DCA.
phenotype exclusively or in part. A caution must be expressed It appears that insulin receptor upregulation occurs in most that this result does not necessarily implicate c-Jun in the liver tumors induced in the male B6C3F1 mouse relative to development of the tumor, because the antibody used in this normal portions of the liver. Therefore, this parameter does not study recognizes a form of the transcription factor that islargely located in the cytosol that surrounds the nucleus, rather distinguish between tumors induced by any of these three than within the nucleus. Other antibodies gave different pat- terns to the staining because they interacted differently tophosphorylated and nonphosphorylated forms of the antibodyat different epitopes (Kato-Weinstein et al., 2001).
The detection of a mixed phenotypic expression of c-Junϩ cells with TRI treatment clearly indicates that TCA is notentirely responsible for the liver tumors produced by TRI.
When given alone, DCA induced 45–50% c-Junϩ lesionswhile c-Junϩ lesions have never been observed with TCA.
Even higher percentages of tumors were found to express thec-Junϩ phenotype with prior studies (Stauber and Bull, 1997;Richmond et al., 1991). Since the tumor response seen at 2 g/LDCA in the present study was lower than that seen previous Differences in insulin-receptor protein (IR), Ras, and phosphory- experiments, the greater uniformity of DCA tumors could have lated forms of Erk 1/2 (designated P44 and P42) in normal liver and tumorsfrom control mice and mice that were treated with dichloroacetate (DCA).
been related to the higher tumor yields at this high dose in past Insulin receptor (IR) expression in normal liver (N) and tumors (T) of male B6C3F1 mice induced by trichloroethylene (TRI), trichloroacetate (TCA), and the two indicated dose levels of dichloroacetate (DCA). The dose of TRI was 1 g/L daily for 87 weeks and that of TCA was 2 g/L of drinking water for52 weeks.
studies. Inconsistent findings can be anticipated at the interme- within the tumor? Alternatively, they could represent the fu- diate dose range because blood levels are seen to increase from sion of two distinct lesions that arose independently. The Ͻ1 ␮M to 300 to 500 ␮M when the concentrations in drinking former possibility suggests that exploration of the differentwater are increased from 0.5 to 2 g/L (Kato-Weinstein et al., portions of these lesions for minimal differences at the 1998). The disproportionate increases in blood levels arise genomic level could provide important insights into the ques- from the fact that DCA is a suicide inhibitor of its own tion of liver tumor progression. These two compounds may be metabolism in mice in this dose range (Gonzalez-Leon et al., very convenient models since neither appears to be mutagenic 1999). Thus, the tumor response would be expected to vary in at concentrations that are active in vivo (discussed more fully this dose region because relatively small variations in dose (i.e., by drinking more or less water) could result in large The H-ras codon 61 mutation frequencies and spectra pro- variations in the systemic concentrations of DCA.
vided a second test of the hypothesis that TCA was a primary, An unanticipated finding was that the pattern of expression if not exclusive, contributor to TRI-induced liver tumors in in tumors from mice treated with varying combinations of mice. This was based on the simple hypothesis that there DCA and TCA did not give rise to tumors with distinct should be congruence between the mutation frequency and phenotypes but to tumors that displayed the two phenotypes in spectra if TCA were solely responsible for the tumors. Clearly, different parts of the same tumor. While some tumors produced this was not the case, as the H-ras codon mutation frequency by TRI were found to have a mixed phenotype, a substantial was significantly lower in TRI-induced tumors than in TCA- fraction of the tumors that resulted expressed only the c-Junϩ induced tumors. The mutation frequency seen in DCA-induced phenotype. The concentrations of DCA and TCA selected for tumors was not significantly different from that of TRI, al- the mixture study should produce approximately the same in though examination of the data also makes it difficult to vivo ratio and individual peak concentrations of DCA and TCA attribute the tumors exclusively to DCA. Therefore, these data that would be predicted to arise from the metabolism of a support the hypothesis that neither metabolite is the exclusive 1-g/kg dose of TRI (Merdink et al., 1998). These data rule out cause of liver tumors in TRI-treated mice.
TCA as the sole cause of liver tumors induced by TRI. Con- The contribution of DCA to the liver tumor response in mice versely, these data do not prove that the c-Junϩ tumors are has been somewhat controversial primarily because of the very entirely the result of metabolism of TRI to DCA or arise from low levels of DCA that are produced in the metabolism of TRI an interaction between the two metabolites. The possibility of (Merdink et al., 1998). Minimally effective systemic concen- a third factor cannot be ruled out. Because of the complexity of trations of DCA are less than 1 ␮M in the blood of mice the metabolism of TRI, there are several possibilities that could (Barton et al., 1999). The present study suggests that low levels account for this change. One possible candidate is dichlorovi- of DCA superimposed on high levels of TCA are additive, at nyl cysteine (DCVC). It is possible that the genotoxic effects of least at low doses. The total tumor response seen at 0.1 g/L this metabolite could contribute to the noncongruence between DCA and 2 g/L of TCA was not significantly greater than findings with TRI and mixtures of DCA and TCA. However, would be expected by the sum of their individual responses.
DCVC has not been subjected to cancer bioassay in animals.
Increasing the concentrations of DCA to 0.5 and TCA to 2 g/L The expression of both phenotypes in the same tumors produced a tumor response that was no greater than produced produced by TRI or combinations of DCA and TCA is of by 2 g/L TCA alone but remained within the confidence limits interest. Do these lesions arise as clones and then the second of the sum of the responses to the independent treatments at metabolite “encourages” the growth of mutant cells that arise CONTRIBUTION OF DCA AND TCA TO TRI–INDUCED LIVER TUMORS Mutation spectra have been useful for backtracking mecha- ing as the tumor grows in size. This supposition is supported by nistically to the adducts to DNA bases produced by genotoxic the observation that the major effect of DCA in liver tumori- carcinogens. The utility of mutation spectra for the study of genesis is on tumor growth rates (Miller et al., 2000).
nongenotoxic chemicals is not as clear. The H-ras mutation Bacterial mutagenicity assays of DCA and TCA have gen- frequency and spectrum of the tumors from male B6C3F1 mice erally been negative (Bull, 2000). Recently DCA was found to treated with 2 g/L TCA resembled those of historical controls.
be weakly, but clearly mutagenic in the mouse lymphoma This is comparable to what Ferreira-Gonzalez et al. (1995) assay and cytogenetic analyses revealed an increase in chro- observed in a limited number of tumors from mice given a mosome aberrations (Harrington-Brock et al., 1998). However, higher dose (3.5 g/L) of TCA for 104 weeks. Tumors induced these effects occurred at concentrations in the test systems that by most other nongenotoxic carcinogens, including phenobar- are several orders of magnitude greater than required for DCA bital, chloroform, and the peroxisome proliferators ciprofibrate induction of liver cancer (Kato-Weinstein et al., 1998; Merdink and methyl clofenapate, have lower ras mutation frequencies et al., 1998). In the Big Blue transgenic mouse mutagenesis than spontaneous tumors (Fox et al., 1990; Hegi et al., 1993; assay, an increased frequency of mutations in the bacterial lacI Stanley et al., 1994). One notable exception occurred in lesions target gene occurred in mice treated with DCA at 3.5 g/L for of female B6C3F1 mice treated with the peroxisome prolifera- 60, but not for 4 or 10 weeks nor with any lower doses (Leavitt tor LY171883, in which the H-ras mutation frequency was et al., 1997). High doses of DCA (Ն2 g/L in drinking water) 64% (Helvering et al., 1994). Although LY171883 is a perox- cause substantial increases in cell replication within cells with isome proliferator, it activates a different member of the per- a c-Junϩ phenotype and many small foci of such cells are oxisome proliferator receptor family (PPAR-␥) than does cip- detected in the liver of mice treated with these high concen- rofibrate (Kliewer et al., 1994). TCA has been shown to trations of DCA (Stauber and Bull, 1997). Therefore, there are activate both PPAR-␣ and PPAR-␥ (Maloney and Waxman, alternative explanations for this observation based upon mea- surement of DCA’s effects on tumor growth rates (Miller et al., The H-ras mutation frequency of the DCA-induced tumors in this study was lower than reported previously (Anna et al., Expression and activity of proteins involved in the MAP 1994; Ferreira-Gonzalez et al., 1995). A number of factors kinase signaling cascade were measured in order to determine could account for this, including differences in mutation de- if there were differential effects on this pathway in tumors tection methodologies, treatment duration, and/or dose. Our possessing codon 61 mutations in H-ras. Activation of the analysis of the H-ras mutation frequency in DCA-induced MAP kinase cascade occurred in all tumors examined and was tumors over time suggests that codon 61 mutations may be late independent of H-ras codon 61 mutation status. This is con- events in male B6C3F1 mice. This is supported by the obser- sistent with a recent study by Kalkuhl et al. (1998) in which vation that when the time of euthanasia is considered over all Raf activity was found to be activated in mouse liver tumors available studies with the frequency of codon 61 H-ras muta- without regard to the H- or K-ras mutation status. Levels of tions increases with time (Maronpot et al., 1995). It should be H-ras mRNA have been found to be elevated in tumors of noted that treatments of 5 g/L DCA (Fereira-Gonzalez et al.,1995), which substantially exceeds the maximally tolerated DCA- or TCA-treated mice (Nelson et al., 1990; Richmond et dose in that it results in very substantial decreases in water al., 1991). Nesnow and DeAngelo (1995) also reported that consumption and losses in body weight does result in a higher DCA-induced tumors lacking H-ras mutations overexpressed mutation frequency in a shorter time period. Moreover, it would be anticipated that a large portion of this group did not Examination of insulin receptor expression revealed that this survive the treatment to the 76 weeks duration of the study.
did not distinguish tumors produced by DCA, TCA, or TRI. In Therefore, it is likely that these data represent a censored all cases, insulin receptor protein expression was increased in population. In a limited number of tumors that were both tumor regardless of cause. In the case of DCA treatment, sequenced and classified histologically, we observed that the Lingohr et al. (2001) have shown that there is a consistent mutation frequency was higher in hepatocellular carcinomas downregulation of the insulin receptor in normal hepatocytes, than in adenomas. Anna et al. (1994) also reported a higher an effect that is reproducible in isolated hepatocytes.
frequency of H-ras mutations in carcinomas compared to ad- In summary, the occurrence of two distinct c-Jun phenotypes enomas in their TRI- and DCA-treated animals, and Nelson et of liver tumors in mice treated with TRI is inconsistent with al. (1990) noted that the expression of H-ras mRNA was TCA being solely responsible for these tumors. It is probable highly correlated with malignancy. This is in contrast to studies that DCA contributes to the development of liver tumors in of strong DNA-alkylating agents for which ras activation ap- mice treated with TRI, but mixtures of DCA and TCA tend to pears to be an early event (Wiseman et al., 1986; Watson et al., produce tumors with a mixed phenotype, whereas TRI pro- 1995). Therefore, these studies indicate that the effects of DCA duced some hepatic tumors that appeared uniformly Junϩ.
and TRI are not typical of genotoxic agents. A more likely role H-ras codon 61 mutation frequencies and spectra support these is increasing clonal expansion with DNA damage accumulat- conclusions. However, activation of ras-dependent signaling pathways are activated in all hepatic tumors examined irre- with genotoxic and nongenotoxic hepatocarcinogens. Cancer Res. 50,
Fox, T. R., and Watanabe, P. G. (1985). Detection of a cellular oncogene in spontaneous liver tumors of B6C3F1 mice. Science 228, 596 –597.
ACKNOWLEDGMENTS
Goldman, F., Hohl, R. J., Crabtree, J., Lewistibesar, K., and Koretzky, G.
(1996). Lovastatin inhibits T-cell antigen receptor signaling independent of The authors thank the following individuals for their assistance: Drs. Rod- its effects on ras. Blood 88, 4611– 4619.
ney A. Miller and Gerald Dagle for pathology; Teresa Luders, Barb A. Meyers,Kristine G. Studniski, and Freddie L. Wallace for animal care; Jessica A.
Gonzalez-Leon, A., Merdink, J. L., Bull, R. J., and Schultz, I. R. (1999). Effect Malone for technical assistance; and Faith Barker and Junko Kato-Weinstein of pre-treatment with dichloroacetic acid or trichloroacetic acid in drinking water on the pharmacokinetics of a subsequent dose in B6C3F1 mice.
Chem.—Biol. Interact. 123, 239 –253.
Harrington-Brock, K., Doerr, C. L., and Moore, M. M. (1998). Mutagenicity of REFERENCES
three disinfection by-products: Di- and trichloroacetic acid, and chloral
hydrate in L5178Y/TKϩ/Ϫ-3.7. LC mouse lymphoma cells. Mutat. Res. 413,
Anna, C. H., Maronpot, R. R., Pereira, M. A., Foley, J. F., Malarkey, D. E., and Anderson, M. W. (1994). ras proto-oncogene activation in dichloroacetic-, Hegi, M. E., Fox, R. R., Belinsky, S. A., Devereux, T. R., and Anderson, trichloroethylene- and tetrachloroethylene-induced liver tumors in B6C3F1 M. W. (1993). Analysis of activated protooncogenes in B6C3F1 mouse liver mice. Carcinogenesis 15, 2255–2261.
tumors induced by ciprofibrate, a potent peroxisome proliferators. Carcino- Barton, H.A., Bull, R., Schultz, I., and Anderson, M. E. (1999). Dichloro- genesis 14, 145–149.
acetate (DCA) dosimetry: Interpreting DCA-induced liver cancer dose re-sponse and the potential for DCA to contribute to trichloroethylene-induced Helvering, L. M., Richardson, F. C., Horn, D. M., Rexroat, M. A., Engelhardt, liver cancer. Toxicol. Lett. 106, 9 –21.
J. A., and Richardson, K. K. (1994). H-ras 61st codon activation in archivalproliferative hepatic lesions isolated from female B6C3F1 mice exposed to Bull, R. J. (2000). Mode of action of liver tumor induction by trichloroethylene the leukotriene D4-antagonist, LY171883. Carcinogenesis 15, 331–333.
and its metabolites, trichloracetate and dichloroacetate. Environ. Health
Perspect. 108(Suppl. 2), 241–259.
Kalkuhl, A., Troppmair, J., Buchmann, A., Stinchcombe, S., Buenemann, C. L., Rapp, U. R., Kaestner, I. K., and Schwarz, M. (1998). p21 (Ras) Bull, R. J., Sanchez, I. M., Nelson, M. A., Larson, J. L., and Lansing, A. J.
downstream effectors are increased in activity or expression in mouse liver (1990). Liver tumor induction in B6C3F1 mice by dichloroacetate and tumors but do not differentiate between RAS-mutated and RAS-wild-type trichloroacetate. Toxicology 63, 341–359.
lesions. Hepatology 27, 1081–1088.
Cariello, N. F, Piegorsch, W. W., Adams, W. T., and Skopek, T. R. (1994).
Kato-Weinstein, J., Lingohr, M. K., Thrall, B. D., and Bull, R. J. (1998).
Computer program for the analysis of mutational spectra: Application to p53 Effects of dichloroacetate-treatment on carbohydrate metabolism in B6C3F1 mutations. Carcinogenesis 15, 2281–2285.
mice. Toxicology 130, 141–154.
Clewell III, H. J., Gentry, P. R., Covington, T. R., and Gearhart, J. M. (2000).
Kato-Weinstein, J., Orner, G. A., Thrall, B. D., and Bull, R. J. (2001).
Development of a physiologically based pharmacokinetic model of trichlo- Differences in the detection of c-Jun/ubiquitin immunoreactive proteins by roethylene and its metabolites for use in risk assessment. Environ. Health different c-Jun antibodies. Toxicol. Methods 11, 1–19.
Perspect. 108(Suppl. 2), 283–305.
Kliewer, S. A., Forman, B. M., Blumberg, B., Ong, E. S., Borgmeyer, U., Daniel, F. B., DeAngelo, A. B., Stober, J. A., Olson, G. R., and Page, N. P.
Mangelsdorf, D. J., Umesono, K., and Evens, R. M. (1994). Differential (1992). Hepatocarcinogenicity of chloral hydrate, 2-chloroacetaldehyde, and expression and activation of a family of murine peroxisome proliferator- dichloroacetic acid in male B6C3F1 mouse. Fundam. Appl. Toxicol. 19,
activated receptors. Proc. Natl. Acad. Sci. USA 91, 7355–7359.
Laemmili, U. K. (1970). Cleavage of structural proteins during the assembly of DeAngelo, A. B., Daniel, F. B., McMillan, L., Wernsing, P., and Savage, R. E., the head of bacteriophage T4. Nature 227, 680 – 685.
Jr. (1989). Species and strain sensitivity to the induction of peroxisome
proliferation by chloroacetic acids. Toxicol. Appl. Pharmacol. 101, 285–
Leavitt, S. A., DeAngelo, A. B., George, M. H., and Ross, J. A. (1997).
Assessment of the mutagenicity of dichloroacetic acid in lacI transgenic
B6C3F1 mouse liver. Carcinogenesis 18, 2101–2106.
DeAngelo, A. B., George, M. H., and House, D. E. (1999). Hepatocarcinoge- nicity in the male B6C3F1 mouse following a lifetime exposure to dichlo- Lingohr, M. K., Thrall, B. D., and Bull, R. J. (2001). Effects of dichloroacetate roacetic acid in the drinking water: Dose–response determination and modes (DCA) on serum insulin levels and insulin-controlled signaling proteins in of action. J. Toxicol. Environ. Health 58, 485–507.
livers of male B6C3F1 mice. Toxicol. Sci. 59, 178 –184.
Elcombe, C. R. (1985). Species differences in carcinogenicity and peroxisome Maloney, E. K., and Waxman, D. J. (1999). trans-Activation of PPAR␣ and proliferation due to trichloroethylene: A biochemical human hazard assess- PPAR␥ by structurally diverse environmental chemicals. Toxicol. Appl. ment. Arch. Toxicol. Suppl. 8, 6 –17.
Pharmacol. 16, 209 –218.
EPA (1996). U. S. Environmental Protection Agency: Proposed Guidelines for Manjanatha, M. G., Li, E., Fu, P. P., and Heflich, R. H. (1996). H- and K-ras carcinogen risk assessment; notice. Fed. Reg. 61, 17960 –18011.
mutational profiles in chemically induced liver tumors from B6C3F1 and Ferreira-Gonzalez, A., DeAngelo, A. B., Nasim, S., and Garett, C. T. (1995).
CD-1 mice. J. Toxicol. Environ. Health 47, 195–208.
Ras oncogene activation during hepatocarcinogenesis in B6C3F1 mice by Maronpot, R. R., Fox, T., Malarkey, D. E., and Goldsworthy, T. L. (1995).
dichloroacetic and trichloroacetic acids. Carcinogenesis 16, 495–500.
Mutations in the ras proto-oncogene: Clues to etiology and molecular Fisher, J. W. (2000). Physiologically based pharmacokinetic models for tri- pathogenesis of mouse liver tumors. Toxicology 101, 125–156.
chloroethylene and its oxidative metabolites. Environ. Health Perspect. Merdink, J. L., Gonzalez-Leon, A., Bull, R. J., and Schultz, I. R. (1998). The 108(Suppl. 2), 265–273.
extent of dichloroacetate formation from trichloroethylene, chloral hydrate, Fox, T. R., Schumann, A. M., Watanabe, P. G., Yano, B. L., Maher, V. M., and trichloroacetate, and trichloroethanol in B6C3F1 mice. Toxicol. Sci. 45,
McCormick, J. J. (1990). Mutational analysis of the H-ras oncogene in spontaneous C57BL/6 ϫ C3H/He mouse liver tumors and tumors induced Miller, J. H., Minard, K. M., Wind, R. A., Orner, G. A., and Bull, R. J. (2000).
CONTRIBUTION OF DCA AND TCA TO TRI–INDUCED LIVER TUMORS In vivo MRI measurements of tumor growth induced by dichloroacetate: roacetic acid in the liver of female B6C3F1 mice. Fundam. Appl. Toxicol. Implications for mode of action. Toxicology 145, 115–125.
31, 192–199.
Nakano, H., Hatayama, I., Satoh, K., Suzuki, S., Sato, K., and Tsuchida, S.
Reynolds, S. H., Stowers, S. J., Patterson, R. M., Maronpot, R. R., Aaronson, (1994). c-Jun expression in single cells and preneoplastic foci induced by S. A., and Anderson, M. W. (1987). Activated oncogenes in B6C3F1 mouse diethylnitrosamine in B6C3F1 mice: Comparison with the expression of liver tumors: Implications for risk assessment. Science 237, 1309 –1316.
pi-class glutathione S-transferase. Carcinogenesis 15, 1853–1857.
Richmond, R. E., DeAngelo, A. B., Potter, C. L., and Daniel, F. B. (1991). The National Cancer Institute (NCI) (1976). Carcinogenesis Bioassay of Trichlo- role of hyperplastic nodules in dichloroacetic acid-induced hepatocarcino- roethylene (CAS No. 79-01-6). NCI-CG-TR-2, DHEW publication no.
genesis in B6C3F1 male mice. Carcinogenesis 12, 1383–1387.
(NIH)76-802. National Cancer Institute, Division of Cancer Cause and Schroeder, M., DeAngelo, A. B., and Mass M. J. (1997). Dichloroacetic acid Prevention, Carcinogenesis Program, Carcinogen Bioassay and Program reduces Ha-ras codon 61 mutations in liver tumors from female B6C3F1 mice. Carcinogenesis 18, 1675–1678.
National Toxicology Program (NTP) (1988). Toxicology and Carcinogenesis Snyder, R. D., Pullman, J., Carter, J. H., Carter, H. W., and DeAngelo, A. B.
Studies of Trichloroethylene (CAS No. 79-01-6) in Four Strains of Rats (1995). In vivo administration of dichloroacetic acid suppresses spontaneous (ACI, August, Marshall, Osborne-Mendel) (Gavage Studies). NTP Techni- apoptosis in murine hepatocytes. Cancer Res. 55, 3702–3705.
cal Report 273. NIH publication no. 88-2525. U.S. Department of Healthand Human Services, Public Health Service, National Institutes of Health, Stanley, L. A., Blackburn, D. R., Devereaux, S., Foley, J., Lord, P. G., Maronpot, R. R., Orton, T. C., and Anderson, M. W. (1994). Ras mutationsin methylclofenapate-induced B6C3F1 and C57BL/10J mouse liver tu- National Toxicology Program (NTP) (1990). Carcinogenesis Studies of Tri- mours. Carcinogenesis 15, 1125–1131.
chloroethylene (Without Epichlorohydrin)(CAS No. 79-01-6) in Fischer-344/N Rats and B6C3F1 Mice (Gavage Studies). NTP Technical Report Stauber, A. J., and Bull, R. J. (1997). Differences in phenotype and cell 243. NIH publication no. 90-1799. U.S. Department of Health and Human replicative behavior of hepatic tumors inducted by dichloroacetate (DCA) Services, Public Health Service, National Institutes of Health, Research and trichloroacetate (TCA). Toxicol. Appl. Pharmacol. 144, 235–246.
Stauber, A. J., Bull, R. J., and Thrall, B. D. (1998). Dichloroacetate and Nelson, M. A., Sanchez, I. M., Bull, R. J., and Sylvester, S. R. (1990).
trichloroacetate promote clonal expansion of anchorage-independent hepa- Increased expression of c-myc, and c-H-ras in dichloroacetate- and trichlo- tocytes in vivo and in vitro. Toxicol. Appl. Pharmacol. 150, 287–294.
roacetate-induced liver tumors in B6C3F1 mice. Toxicology 64, 47–57.
Watson, M. A., Devereux, T. R., Malarkey, D. E., Anderson, M. W., and Nesnow, S., and DeAngelo, A. B. (1995). Overview on the mechanisms of Maronpot, R. R. (1995). H-ras oncogene mutation spectra in B6C3F1 and carcinogenic action of dichloroacetic acid in the male B6C3F1 mouse: C57BL/6 mouse liver tumors provide evidence for TCDD promotion of Application to the construction of a biologically-based dose–response spontaneous and vinyl carbamate-initiated liver cells. Carcinogenesis 16,
model. In Disinfection By-Products in Drinking Water: Critical Issues in Health Effects Research. Summary report of a workshop organized by ILSI Wiseman, R. W., Stowers, S. J., Miller, E. C., Anderson, M. W., and Miller, Health and Environmental Science Insititute. Chapel Hill, NC. Oct. 23–25, J. A. (1986). Activating mutations of the c-Ha-ras protooncogene in chem- ically induced hepatomas of the male B6C3F1 mouse. Proc. Natl. Acad. Sci. Pereira, M. A. (1996). Carcinogenic activity of dichloroacetic acid and trichlo- USA 83, 5825–5829.

Source: http://glejak.pl/forum/files/DCA-2.pdf