Pentachlorophenol ( PCP ) produces liver oxidative stress and promotes but does not initiate hepatocarcinogenesis in B 6 C 3 F 1 mice

To elucidate the mechanism of hepatocarcinogenesis of pentachlorophenol (PCP) in mice, critical effects related to carcinogenicity were studied in the livers of B6C3F1 male mice administered PCP at concentrations of 600 and 1200 p.p.m. in the diet for 8 weeks. Oxidative stress was assessed by measurements of 8-oxodeoxyguanosine (8-oxodG) in the liver nuclear DNA and hepatocyte cell proliferation was quantified by bromodeoxyuridine incorporation. Also, initiation and promotion were assessed in a two-stage hepatocarcinogenesis model in which one group of mice was given PCP at concentrations of 600 and 1200 p.p.m. as initiator for the first 13 weeks with subsequent administration of phenobarbital (PB) as promoter at a concentration of 500 p.p.m. in the drinking water for 29 weeks. A second group was initiated with diethylnitrosamine (DEN) at 20 p.p.m. in the drinking water for the first 13 weeks followed after a 4 week recovery interval by PCP at concentrations of 300 and 600 p.p.m. in the diet for 25 weeks. Significant elevations in 8-oxodG levels and cell proliferation were observed in a dosedependent manner. Incidences and multiplicities of hepatocellular tumors in mice treated with PCP after DEN initiation were increased compared with those in mice given initiation only. In contrast, in mice given PCP as initiator followed by PB no enhancement of neoplastic lesions occurred. These findings are interpreted to demonstrate that PCP exerts a promoting action, but not an initiating effect on liver carcinogenesis and that the promoting action is related to oxidative stress and compensatory hepatocellular proliferation.

© Oxford University Press 1115 metabolism, it has been assumed that oxidative stress plays a key role in its carcinogenicity (5).In support of this hypothesis, we found that exposure of mice to PCP in the diet for 2 and 4 weeks was able to induce formation of 8-oxodeoxyguanosine (8-oxodG) in liver DNA (6).Simultaneously, an increase in hepatocyte cell proliferation was observed without overt hepatotoxicity, suggesting that oxidative stress might be responsible for the observed cell proliferation.A close relation between cell proliferation and tumor promotion in carcinogenesis (7) led us to hypothesize a potential of PCP to exert a promoting effect on tumor development.Reactive oxygen species are also believed to participate in the tumor promotion process (8,9).
8-oxodG adducts have been reported to cause misreading of the DNA sequence during replication, thereby inducing G:C→T:A transversion (10), and are also known to induce G→T transversion in the Ha-ras sequence (11,12).Introduction of this kind of oxidized nucleoside into DNA resulted in induction of mutagenesis in Escherichia coli and Salmonella typhimurium (13)(14)(15).Recently, the biological significance of repair enzymes for 8-oxodG adduct formation as against spontaneous mutation in bacteria has been reported (16).Thus, it has been assumed that 8-oxodG adducts may have a role in the carcinogenic process (17,18).
The present study examines potential initiating and promoting effects of PCP in mouse liver carcinogenesis and the role of oxidative stress and cell proliferation of hepatocytes in these processes.We report that PCP has promoting, but not initiating activity and that the promoting activity is related to oxidative stress and compensatory hepatocellular proliferation.

Animals and exposures
The protocol for this study was approved by the Animal Care and Utilization Committee of the National Institute of Health Sciences.Two hundred and fifteen male B6C3F 1 mice, 5 weeks old, were purchased from Japan SLC (Shizuoka, Japan).They were kept in polycarbonate cages (5 mice/cage) with hardwood chips for bedding and were housed in a conventional animal facility maintained under conditions of controlled temperature (23 Ϯ 2°C), humidity (60 Ϯ 5%), air change (12 per h) and lighting (12 h dark/light cycle).The animals were given free access to F-2 basal diet (Funabashi Farm Co., Chiba, Japan) and tapwater and were started on the study after a 1 week acclimation period.
The mice were randomly divided into 13 groups, as shown in Figure 1.For the measurements of 8-oxodG levels in liver DNA and BrdU labeling index of hepatocytes, five animals each in groups 1-3 were given PCP at concentrations of 0, 600 or 1200 p.p.m. in their diet for 8 weeks.They were given BrdU (20 mg/kg) by i.p. injection twice a day for the final 3 days of exposure and once on the day of termination at 2 h before death, as previously described (19).Animals were killed under ethyl ether anesthesia and the livers were immediately removed and slices taken from each lobe were fixed in buffered formalin for BrdU immunostaining.The remaining pieces of the livers were frozen with a dry ice/acetone mix and stored at -80°C until measurement of 8-oxodG in nuclear DNA.
Twenty animals each in groups 4 and 5 were untreated until 17 weeks after the start of the experiment and were then given PCP at concentrations of 0 or 600 p.p.m. in the diet for 25 weeks, respectively.Twenty animals in group 6 were untreated until 13 weeks after the start of the experiment and were then given PB at a concentration of 500 p.p.m. in the drinking water for 29 weeks.Twenty animals each in groups 7-10 were given DEN as an initiator at 20 p.p.m. in the drinking water for 13 weeks and then animals in groups 8-10 were given PB at a concentration of 500 p.p.m. in the drinking water and PCP at concentrations of 300 or 600 p.p.m., respectively, in their diet for 25 weeks after a 4 week recovery period to assess promoting activity.Animals in group 7 were untreated thereafter until the end of the experiment.Twenty animals each in groups 11-13 were given PCP at concentrations of 600, 1200 or 1200 p.p.m., respectively, for the first 13 weeks and then given PB as a promoter at 500 p.p.m. in the drinking water for 29 weeks, except group 13.The three doses of PCP used in the present study were 300 (low, L), 600 (moderate, M) and 1200 p.p.m. (high, H).The moderate dose (600 p.p.m.) was the same as the high dose in the NTP bioassay despite the impurities being different (2).The actual calculated doses of PCP in groups 9-12 were 1.4, 2.8, 2.5 and 5.2 mg/mouse/day, respectively, which were ~1000 times higher in comparison with the reported range of exposure levels in humans (20).At necropsy, mice were killed under the same conditions as described above at week 42 and the livers were removed and weighed.Slices taken from each lobe were fixed in buffered formalin and then routinely processed for embedding in paraffin and 4 µm sections were stained with hematoxylin and eosin (H&E) stain.Liver lesions were diagnosed as hepatocellular altered foci (HAF), adenomas and carcinomas as generally accepted (21).Especially, the definition of HAF conformed to the criteria proposed by Frith et al. (22).

Measurement of nuclear 8-oxodG
Nuclear fractions were obtained by centrifugation at 1000 g from 0.3 g liver tissue after homogenization in 2.7 ml of 20 mM EDTA, 5 mM phosphate buffer, pH 7.5.DNA extraction was performed with a Nucleic Acid Purification System (model 341; Applied Biosystem, Foster City, CA).The nuclear fractions were lysed with proteinase K and lysis buffer: the DNA was extracted with phenol/water/chloroform and then precipitated with isopropanol.All reagents in this process were supplied by Applied Biosystems.These DNA extraction procedures were automatically conducted in glass vessels filled with helium gas and in the dark in order to reduce auto-oxidation of 1116 deoxyguanosine (dG) (23).The DNA was digested to deoxynucleotides with nuclease P1 and alkaline phosphatase.8-OxodG levels (8-oxodG/10 5 dG) were assessed by HPLC with an electrochemical detection system (HPLC pump 420; Kontron AG, Zurich, Switzerland; model-5100A ESA; Coulochem, Bedford, MA).

Hepatocellular proliferation quantification
For each animal at least 3000 hepatocytes were counted.The replicating fraction (RF) was calculated as the percentage of cells positive for BrdU incorporation divided by the total number of cells counted.

Statistics
The significance of differences in the results for 8-oxodG levels and RFs was evaluated with the use of ANOVA, followed by Dunnett's multiple comparison test.Regarding the initiation-promotion bioassay, the significance of differences in the results for liver weight and the multiplicity of liver lesions was evaluated by Student's t-test and that for the incidence of liver lesions was evaluated by Fisher's exact probability test.

Results
The changes in 8-oxodG levels in liver nuclear DNA of mice exposed to PCP in the diet for 8 weeks are shown in Figure 2. 8-OxodG levels were elevated in a dose-dependent manner, the increase being statistically significant at 600 (1.54 Ϯ 0.30/10 5 dG, P Ͻ 0.01) and 1200 p.p.m. (2.31 Ϯ 0.57/10 5 dG, P Ͻ 0.01), as compared with the control value (0.61 Ϯ 0.18/10 5 dG).
The RFs of hepatocytes of mice exposed to PCP in the diet for 8 weeks are shown in Figure 3. RFs were also elevated in a dose-dependent manner, with significant increases at 600 (2.8 Ϯ 0.94%, P Ͻ 0.01) and 1200 p.p.m. (4.42 Ϯ 0.90%, P Ͻ 0.01), as compared with the control value (0.67 Ϯ 0.09 %).
At the termination of the initiation-promotion bioassay at 42 weeks, survival and body weights were decreased in groups that received DEN (Table I).Exposure to PCP at a concentration of 600 p.p.m. (group 5) or PB at a concentration of 500 p.p.m. (group 6) caused significant increases in liver weights (p Ͻ 0.01) as compared with the untreated control (group 4).The exposures to PCP or PB following DEN treatments (groups 8-10) caused hepatomegaly, the differences at a concentration of 600 p.p.m.PCP (group 10) being statistically significant (P Ͻ 0.01) as compared with the DEN alone group (group 7).Also, PB exposure following PCP treatment (groups 11 and 12) caused significant elevations of liver weight (P Ͻ 0.01) as compared with the PCP alone group (group 13).
The incidences and the multiplicities of liver neoplastic lesions in mice exposed to PCP in the two-stage hepatocarcinogenesis model are given in Table II.HAF were observed in all groups except the untreated control (Figure 4A).PCP exposure following DEN initiation (groups 7-10) increased the incidences and multiplicities, with the multiplicity of HAF at a concentration of 600 p.p.m.PCP (group 10) being significantly increased compared with the DEN alone group (group 7) (3.83 versus 0.67, P Ͻ 0.01).In contrast, among the groups that received PCP at concentrations of 0, 600 and 1200 p.p.m. as initiator followed by PB (groups 6, 11 and 12) there were no obvious differences.Hepatocellular adenomas were observed in all groups that received DEN (Figure 4B), hepatocellular carcinomas also being found in all groups  except the DEN alone group (Figure 4C).The incidences and multiplicities of neoplasms in mice given PCP as promoter at concentrations of 300 and 600 p.p.m. (groups 9 and 10) were 67% and 1.27 and 72% and 2.22, respectively.These were statistically significant (group 9, P Ͻ 0.05, P Ͻ 0.05; group 10, P Ͻ 0.01, P Ͻ 0.05) as compared with the values in mice administered DEN alone (group 7, 27% and 0.33).Twentyfive weeks exposure to PB following DEN exposure (group 8) resulted in the development of tumors resulting from DEN initiation, while PB promotion for 29 weeks failed to elicit tumors in the livers of mice exposed to PCP for 13 weeks (groups 11 and 12).

Discussion
The major metabolite of PCP, tetrachlorohydroquinone (TCHQ), was reported to induce 8-oxodG formation in liver DNA of mice (24).Since further oxidation from TCHQ to tetrachlorobenzoquinone might give rise to superoxide anions (25), it has been accepted that formation of reactive oxygen radicals during PCP metabolism might be involved in its carcinogenicity.In a previous paper, we demonstrated that 8-oxodG levels in liver DNA and RFs of hepatocytes were concurrently increased in mice treated with PCP at concentrations of 300, 600 and 1200 p.p.m. in the diet for 2 and 4 weeks (6).The fact that even the concentration revealing no overt hepatotoxicity, i.e. 300 p.p.m., was nevertheless able to induce cell proliferation suggested the involvement of oxidative stress in PCP-induced mitogenesis.Moreover, we found prevention of the oxidative stress and cell proliferation due to PCP exposure by co-treatment with catechin, an antioxidant in green tea (unpublished data).
Exposure to PCP at concentrations of 300 and 600 p.p.m. in the diet for 25 weeks enhanced the incidence and multiplicity of HAF in a dose-dependent manner after DEN initiation.Moreover, carcinomas, which did not occur in mice given DEN alone, were observed in mice given DEN after PCP.However, PCP produced no initiating effect when followed by PB as a promoter.These findings establish that PCP can exert a promoting action in DEN carcinogenesis.Although the NTP bioassay was performed using two different grades of PCP, their impurities had no influence on the results (26).Regardless of the purity of the PCP in the present study being a little higher than those in the NTP bioassay, without doubt the observed promoting activity participated in PCP carcinogenesis.It has been reported that various events related to cell proliferation may result from cellular oxidation (27,28).Likewise, we have demonstrated that oxidative stress generated by a renal carcinogen is associated with induction of cell proliferation and 1118 associated promoting activity (29,30).Considering the present data showing sustained increases in cell proliferation concurrent with 8-oxodG formation, it is likely that oxidative stress due to PCP exposure may participate in its promoting action.8-OxodG lesions can be misread during DNA replication, leading to point mutations (10) and subsequent oncogene activation (11,12).Recently, in addition to cloning of the mutMST gene encoding 8-oxodG DNA glycosylase in Salmonella typhimurium, its human homolog, the hOGG1 gene, has been cloned (31,32).Interestingly, disruption of the mutMST gene in S.typhimurium results in an increase in mutation frequency and several mutations in the hOGG1 gene are found in human gastric, lung and kidney tumors (33,34).These data suggest the possibility that 8-oxodG adducts could have an initiating potential in the carcinogenic process.In the present study, 25 weeks exposure to PB following DEN initiation enhanced the incidence and multiplicity of hepatocellular adenomas due to DEN exposure, showing that PB treatment under these conditions was sufficient to facilitate initiated cells resulting from genotoxic events such as DEN exposure to develop into tumor cells.Nevertheless, no enhancement of neoplastic lesions was found in mice given PCP at concentrations of 600 and 1200 p.p.m. in the diet followed by PB treatment for 29 weeks.Thus, the 8-oxodG levels produced by PCP apparently had no initiating effect.This corresponds to the observations of Denda et al. (35) that induction of 8-oxodG by redox enzyme modulation did not effect initiation in rat liver.In a recent report, we provided evidence that persistent, but not transient, formation of 8-oxodG adducts might be required as a trigger for neoplastic transformation on account of the existence of repair mechanisms against oxidative stress (30).
While PCP was reported to be a liver carcinogen in mice (2), it failed to induce tumors in rats (36).Several explanations are possible.Among these, in the rat the metabolism of PCP may not produce the oxidative stress that it does in mouse liver.Also, oxidative stress in rat liver may have less of a mitogenic effect than in mouse liver.Finally, since many preexisting initiated cells are apparently available for promotion to overt neoplasms in the livers of mice, reflected in the high incidence of cryptogenic neoplasms (37), it may be that there is more substrate for promotion in the mouse liver.In any event, our findings implicate PCP as an epigenetic carcinogen of the promoter class.

Fig. 2 .
Fig. 2. Changes in 8-oxodG levels in the liver nuclear DNA of mice given PCP in the diet at concentrations of 600 and 1200 p.p.m. for 8 weeks.The values represent the means Ϯ SD of data for five mice.Significant differences from the control group fed the basal diet are shown, **P Ͻ 0.01.

Fig. 3 .
Fig. 3. Changes in RFs of hepatocytes in mice given PCP in the diet at concentrations of 600 and 1200 p.p.m. for 8 weeks.The values represent the means Ϯ SD of data for five mice.Significant differences from the control group fed the basal diet are shown, **P Ͻ 0.01.