Phenobarbital at low dose exerts hormesis in rat hepatocarcinogenesis by reducing oxidative DNA damage, altering cell proliferation, apoptosis and gene expression

Abstract

Our recent research indicated that phenobarbital (PB) may inhibit the development of N-diethylnitrosamine (DEN)-initiated pre-neoplastic lesions at low doses in a rat liver medium-term bioassay (Ito test), while high doses exhibit promoting activity. This raises the question of whether treatment with low doses of PB might reduce cancer risk. For clarification, male 6-week-old F344 rats were treated with PB at doses of 0, 2, 15 and 500 p.p.m. in the diet for 10 or 33 weeks after initiation of hepatocarcinogenesis with DEN. In a second, short-term experiment, animals were given PB at doses of 2, 4, 15, 60 and 500 p.p.m. for 8 days. Formation of glutathione S-transferase placental form (GST-P) positive foci and liver tumors was inhibited at 2 p.p.m. Generation of oxidative DNA damage marker, 8-hydroxy-2′-deoxyguanosine (8-OHdG), cellular proliferation within the areas of GST-P positive foci and apoptosis in background liver parenchyma were suppressed. Suppression of 8-OHdG formation by PB at low dose might be related to the enhanced mRNA expression of 8-OHdG repair enzyme, oxoguanine glycosylase 1 (Ogg1). Moreover, as detected by cDNA microarray analysis, PB treatment at low dose enhanced mRNA expression of glutamic acid decarboxylase (GAD65), an enzyme involved in the synthesis of gamma-aminobutyric acid (GABA), and suppressed MAP kinase p38 and other intracellular kinases gene expression. On the contrary, when PB was applied at a high dose, GST-P positive foci numbers and areas, tumor multiplicity, hydroxyl radicals and 8-OHdG levels were greatly elevated with the increase in CYP2B1/2 and CYP3A2 mRNA, protein, activity and gene expression of GST, nuclear tyrosine phosphatase, NADPH- cytochrome P-450 reductase and guanine nucleotide binding protein G(O) alpha subunit. These results indicate that PB exhibits hormetic effect on rat hepatocarcinogenesis initiated with DEN by differentially altering cell proliferation, apoptosis and oxidative DNA damage at high and low doses.

Article

Introduction

It is generally believed that rodent carcinogens are human carcinogens, despite the existence of exceptions, the prevailing paradigm is that even tiny doses can induce cancer (1). However, many studies have shown benefits, not harm, from low-level exposure to toxicants, this phenomenon being known as hormesis (2). Furthermore, it has been observed that for some substances tested, carcinogens are similar to other toxicants in improving health at low doses, although, the mechanism of their action remains unclear (3).

Phenobarbital (PB), a sedative and anti-convulsant used as an anti-epilepsy drug in humans, is also a non-genotoxic carcinogen and a well-known promoter of hepatocarcinogenesis in vivo and in vitro (49). The promoting effect of PB at a high dose on hepatocarcinogenesis in rodents has been extensively studied, but reasons for its carcinogenic action have yet to be unequivocally clarified. Increased reactive oxygen species (ROS) due to the activity of detoxifying enzymes and induction of oxidative stress are suggested to be possible mechanisms by which non-genotoxic chemicals may exert carcinogenicity (10,11). Oxygen radicals attack DNA bases and deoxyribose residues, producing damaged bases and single strand breaks, or oxidize lipid and protein molecules, generating intermediates which can react with DNA and form adducts. Among the more abundant types of base modifications, 8-hydroxy-2′-deoxyguanosine (8-OHdG), produced by the oxidation of deoxyguanosine, is considered as the most sensitive and useful marker of oxidative DNA adducts (12). It has been shown that 8-OHdG is closely associated with various diseases, including cancer, and is produced by exposure to various carcinogens (12). It is known to cause mutations, predominantly G to T transversions (13,14).

Our previous research indicated that PB administered at doses from 60 to 500 p.p.m. dose-dependently increases the numbers and areas of glutathione S-transferase placental form (GST-P) positive foci, whereas, the dose 15–30 p.p.m. did not appear to have any effect in a medium-term rat liver bioassay (Ito test) (15). Interestingly, in the same study, inhibition activity of PB on the development of GST-P positive lesions was apparent for doses in the range of 1–7.5 p.p.m. with significance at 1 and 2 p.p.m. (15). In past studies a dose–response effect of PB on the increase of the hepatic cancer was observed, exhibiting a threshold at low doses (1618).

Anticarcinogenic effects of PB might result from stimulation of hepatic drug-metabolizing enzymes, which detoxify carcinogens (19). In the post-initiation phase the mechanisms underlying modulation are not clear but hormesis-type responses may represent modest overcompensation to a disruption of homeostasis (20). Furthermore, a number of studies have revealed that biphasic dose–response appears to result from direct stimulation at low doses and inhibition at high doses without involvement of the above-mentioned compensatory responses (20).

In the present study, the effects of PB application at 2, 15 and 500 p.p.m. were investigated with reference to lesion development, oxidative stress, DNA damage, cellular proliferation and apoptosis in the rat liver. For this purpose the combination of high performance liquid chromatography (HPLC), electron spin resonance (ESR), competitive RT–PCR and double immunohistochemistry was employed. In addition, mRNA expression of 8-OHdG repair enzyme, oxoguanine DNA glycosylase 1 (Ogg1), was analysed by real-time PCR. We also made use of cDNA microarray technology to gain a global view of the pool of genes that might play important roles in inhibition of hepatocarcinogenesis by PB at low dose.

Materials and methods

Chemicals

PB sodium salt (CAS no. 57-30-7) (purity ≥98%) and other reagents were purchased from Wako Pure Chemicals Industries (Osaka, Japan). N-Diethylnitrosamine (DEN) was from Sakai Research Laboratory (Fukui, Japan). The spin trapping agent 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) was obtained from Labotec Co. (Tokyo, Japan).

Animals

A total of 168, 5-week-old male Fisher 344 rats (Charles River, Japan, Hino, Shiga, Japan) were quarantined for 1 week before the start of the experiment. They were housed in an animal facility maintained on a 12 h (07:00–19:00) light/dark cycle, at a constant temperature of 23 ± 1°C and relative humidity of 44 ± 5%, and were given free access to tap water and food (Oriental MF pellet diet, Oriental Yeast Co., Tokyo, Japan).

Experimental design

In the first experiment, 138 F344 rats were divided into six experimental groups. After 1 week on basal diet they underwent i.p. injection of DEN (100 mg/kg body wt) dissolved in saline to initiate hepatocarcinogenesis. This was performed three times, once per week. After 1 further week on basal MF diet, animals were administered diet containing PB at a dose of (control), 2, 15 or 500 p.p.m. for 10 (10 rats per group) or 33 (20 rats per group) weeks. Two groups were injected with saline instead of DEN and administered PB at doses of or 500 p.p.m. for 10 (five rats per group) and 33 (four rats per group) weeks. At experimental weeks 13 and 36 the animals were killed.

In the second, short-term experiment, animals were divided into six experimental groups, five rats each. After 1 week on basal diet they were administered diet containing PB at doses of (control), 2, 4, 15, 60 or 500 p.p.m. for 8 days. The duration of the treatment was chosen from the results of our previous study, in which PB administered to F344 rats at a dose of 500 p.p.m., showed the highest induction of hydroxyl radicals (OH) and 8-OHdG in the rat liver DNA after 8 days of application (21).

In experiment 1, immunohistochemical demonstration of GST-P positive foci and HE staining for assessment of the incidence, multiplicity and histopathological classification of liver tumors were carried out at weeks 13 and 36, respectively. Separate portions of normal-appearing liver tissue and tumors at week 36 were fixed in Bouin's solution and 10% phosphate-buffered formalin.

At death animals were anesthetized with diethyl ether. In experiments 1 (week 10) and 2, liver perfusion was performed in situ at 2 ml/min with ice-cold 1.15% KCl buffer (1.15% KCl, 1 mM EDTA, 0.25 mM PMSF) at room temperature (21). The perfusion was continued for 8 min.

After finishing the perfusion, livers were immediately removed. Separate portions were fixed in Bouin's solution and 10% phosphate-buffered formalin, partially frozen in liquid nitrogen and stored at −80°C for molecular assessment. The remainder of the liver tissue was immediately processed for microsomes isolation, as described elsewhere (22). The liver microsomal fraction was stored at −80°C for subsequent ESR examination of hydroxyl radical generation and cytochrome P-450 content, P-450 mediated hydroxylation of testosterone by HPLC for activity determination and western blotting (2326). SDS–PAGE was performed by the method of Laemmli (27). NADPH- cytochrome P-450 reductase (OR) activity in rat liver microsomes was measured using the artificial electron acceptor cytochrome C (28).

ESR

DMPO was employed as a spin trapping agent and signals were assessed with a quartz flat cell (inner size, 60 × 10 × 0.31 mm) using a JES-TE200 ESR Spectrometer (Japan Electronics Datum Co., Osaka, Japan) as described (21). After recording, the signal intensity of the spin adducts of DMPO and OH (DMPO–OH) was evaluated from the peak height of the third signal of the quartet and normalized as relative height against the standard signal intensity of the manganese oxide marker.

8-OHdG analysis

Rat liver DNA 8-OHdG levels were determined by an HPLC-ECD method (29). Furthermore, immunohistochemical staining of 8-OHdG was accomplished as described previously (21).

Immunohistochemistry for GST-P positive foci

Immunohistochemical assessment of GST-P was performed with the avidin–biotin complex method as described by Kitano et al. (15) using an anti-rat GST-P primary (1:2000) (IgG, 100 µg/ml) antibody diaminobenzidine tetrahydrochloride as the substrate to demonstrate sites of peroxidase binding. Quantitation of GST-P positive foci was done using two-dimensional evaluation (30). The numbers and areas of foci >0.2 mm in diameter, and total areas of liver sections, were measured using a color image processor (IPAP; Sumica Technos, Osaka, Japan) to give values per cm2 of liver section.

Double immunohistochemistry for proliferating cell nuclear antigen and GST-P

Double staining for GST-P positive foci and proliferating cell nuclear antigen (PCNA) was performed using the polyclonal rabbit anti-rat GST-P antibody at 1:2000 dilution and an anti-PCNA mouse monoclonal (PC-10, IgG2a; DAKO, Kyoto, Japan; 1:500) antibody. Immunohistochemical detection of GST-P was accomplished according to the protocol described above. The sites of peroxidase binding were demonstrated with alkaline phosphatase [Vectastain ABC-AP kit, Vector Red (SK-5100)] solution. Thereafter, sections were sequentially treated with 0.2 M glycine, pH 2.2, for 2 h to remove immune complexes. Immunohistochemistry for PCNA was essentially performed as described previously (21). GST-P positive foci were stained red, while blue staining of nuclei showed a positive immunoreaction of monoclonal primary antibody with PCNA. The PCNA indices were estimated for GST-P positive areas of a same size (foci consisting of 20–30 cells) and in background liver parenchyma as numbers of positive nuclei per 1000 cells.

Double immunohistochemistry for GST-P and apoptosis

Double immunohistochemistry for GST-P and apoptosis [single-stranded DNA (ssDNA)] was performed with Bouin-fixed sections as described above for GST-P and PCNA using alkaline phosphatase (Vectastain ABC-AP kit, Vector Red/Vector Blue) solutions for the immunohistochemical detection of GST-P and apoptosis (21,31). Blue-stained nuclei reflected binding of primary antibody against ssDNA, while GST-P positive foci were stained red. Apoptotic index was assessed as for PCNA.

RNA preparation

Total RNA was isolated from rat liver (pieces <5 mm in diameter) using ISOGEN (Nippon Gene, Toyama, Japan) (32,33). RNA was isopropanol precipitated, dissolved in DEPC-treated distilled water and kept at −80°C until use. RNA concentrations were determined with a spectrophotometer (Ultraspec 3000, UV/Visible Spectrophotometer; Pharmacia Biotech, Tokyo, Japan). Reverse transcription of 3 µg of total RNA was performed with Oligo-dT primer, and cDNA samples were stored at −20°C until assayed.

cDNA microarray analysis

One hundred micrograms of pooled aliquots of total RNA from 10 rats in each group were treated with DNase I and processed for PolyA+ RNA enrichment and generation of 32P-cDNA probes using the Atlas Pure RNA Labeling System (Clontech, Palo Alto, CA) according to the manufacturer's protocol. Radiolabeled cDNA was purified and eluted through a NucleoSpin Extraction Column (centrifuged 14 000 r.p.m.). Each Atlas Rat 1.2 Array (Clontech) was pre-hybridized with 5 ml of ExpressHyb buffer (Clontech) and 0.5 mg of denatured DNA from sheared salmon testes at 68°C for 1 h with continuous agitation. The probes synthesized from the treatment samples, and their corresponding controls were then added. After overnight hybridization and a high-stringency wash the arrays were scanned with a BAS-2500 Imager (Fujifilm Medical Systems, Stamford, CT) after 1 day exposure. Hybrid intensity was normalized to values for house keeping genes with Atlas Image Software. To minimize the effects of measurement variation introduced by artificial sources during the experiments, we only included spots with up-regulation or down-regulation at least 2-fold. CYP2B1, CYP3A2, CYP2C11 isoenzymes of P-450 and Ogg1 were not included in cDNA microarray analysis and their expression was investigated with the use of competitive or real-time RT–PCR. Microarray analysis was repeated twice to check the reproducibility of the data. Furthermore, results obtained by cDNA microarray for expression of GAD65 (glutamic acid decarboxylase) were checked by quantitative RT–PCR analysis using a set of primers (forward: GGCTCTGGCTTTTGGTCCTTC and reverse: TGCCAATTCCCAATTATACTCTTGA) as described previously (34).

Competitive RT–PCR

As CYP2B1/2, CYP3A2, CYP2C11 isoforms of cytochrome P-450 were not included in the microarray analysis, the determination of their mRNA expression and the house keeping gene cyclophillin was performed using a Rat Cytochrome P-450 Competitive RT–PCR set and a Takara RNA LA PCR TM Kit (Takara Biomedicals, Japan) according to previously reported protocols and the data were expressed as a relative density of CYP2B1/2, CYP3A2 or CYP2C11 bands as compared with the band density of RNA competitor (35).

Real-time LightCycler PCR

Using RNA extracted from the livers of rats without any chemical treatment as a template, OGG1 and GAPDH (glyceraldehyde-3-P dehydrogenase) were amplified by RT–PCR and subcloned into a pT-AdV vector plasmid using an AdvanTAge™ PCR cloning Kit (Clontech Laboratories, USA) as described previously (21). Serial dilutions ranging from 109 to 102 molecules were then prepared. Real-time quantitative LightCycler PCR (LC-PCR) for OGG1 and GAPDH was performed using a LightCycler™ DNA Master SYBR Green I FastStart Kit (Roche Molecular Biochemicals, Germany) and a set of primers for OGG1 and GAPDH as reported (21). Results were expressed relative to the number of GAPDH transcripts used as the internal control. Some amplification products were checked by electrophoresis on 3% ethidium bromide-stained agarose gels. All examinations were performed in quintuplicate.

Statistical analysis

Statistical analysis was performed with the Fisher's exact probability test for the incidence of liver tumors and the two-tailed Dunnet's test for the remainder of our data using the StatView-J 5.0 program (Berkeley, CA).

Results

General observations

All treatment diets were well tolerated and there were no significant differences among the groups with regard to food and water consumption or body weight gain (data not shown). Two rats died before the termination of the experiment without discernible cause. Final liver/body weight ratios were increased significantly only in the groups administered PB at a dose of 500 p.p.m. for 10 weeks (4.8 ± 0.6%, P < 0.0001) and 33 weeks (5.9 ± 1.6%, P < 0.0001) following DEN initia­tion as compared with values for the control groups, (3.5 ± 0.3%) and (3.1 ± 0.3%), respectively. No change in liver/body weight ratio was induced by administration of PB at a dose of 2 and 15 p.p.m. as compared with initiation control group.

Inhibitory and promoting effects of PB on formation of GST-P positive foci and liver tumors

Data on effects of PB application at different doses on formation of GST-P positive foci at week 13 and liver tumors at week 36 in experiment 1 are shown in Table I. After 10 weeks of continuous PB administration to rats initiated with DEN a significant decrease in number and area of GST-P positive foci was found in the group treated with PB at a dose of 2 p.p.m. as compared with the DEN initiation control group (Table I). In contrast, values were highly elevated in animals administered PB at a dose of 500 p.p.m.

Histological examination of liver tumors at week 36 in the experiment 1 demonstrated a significant reduction in the multiplicity of hepatocellular carcinomas (HCCs) and total tumors in the group treated with PB at a dose of 2 p.p.m. as compared with the DEN control group. Furthermore, a tendency for decreased incidences of adenomas and HCCs was also evident. In contrast, HCC and total tumor multiplicity values in the group treated with PB at 500 p.p.m. were highly elevated as compared with the controls. Most tumors developing in animals administered 500 p.p.m. of PB were well differentiated HCCs. However, in the 2 p.p.m. PB-treated group ∼70% of tumors were adenomas.

Dose-dependent production of oxygen radicals

The observed hydroxyl radical spectrum was in agreement with that reported earlier for the DMPO–OH adduct (23). Dose-dependent elevation of DMPO–OH spectra was found in the rat liver microsomal fraction at week 10 of PB administration (Table II). A slight but significant increase of DMPO–OH spectra was also evident in the 2 p.p.m. treated group after 8 days of PB application in experiment 2 (data not shown). Hydroxyl radical levels were highly elevated in the liver microsomal fraction by the treatment with PB at a dose of 500 p.p.m. in both long (Table II) and short-term (data not shown) studies.

Influence on 8-OHdG formation

HPLC analysis demonstrated 8-OHdG levels in rat liver DNA to be significantly decreased in the group administered PB at a dose of 2 p.p.m. in experiment 1 (Table II). Similar results were obtained in the short-term experiment (data not shown). Alteration of 8-OHdG showed a similar pattern with that of GST-P positive foci numbers, areas and incidence and multiplicities of liver tumors (Table I). Formation of oxidative base modifications was increased in the initiation control group as compared with vehicle controls and significantly enhanced by application of PB at a dose of 500 p.p.m. (Table II). The results of immunohistochemical examination and HPLC analysis for 8-OHdG showed good concordance (data not shown).

Increase of Ogg1 mRNA expression

In an attempt to provide an explanation for the observed inhibition of oxidative DNA damage by PB at low dose, we analyzed mRNA expression for the 8-OHdG repair enzyme, Ogg1, by real-time LC-PCR. Significant elevation was noted with PB application at 2 p.p.m., 15 and 500 p.p.m. in experiment 1 (Table II). Therefore, the difference in 8-OHdG levels observed in this experiment might be dependent on the balance between the production of ROS, formation of 8-OHdG and its elimination by Ogg1.

Alteration to cell proliferation

Double staining of GST-P and PCNA revealed an increase in number of positively stained cells within the area of GST-P positive foci in the 15 and 500 p.p.m. PB dose groups. Several hepatic foci, consisting of >20 cells, contained a large number of PCNA positive nuclei (Figure 1A and B). The biggest foci were strongly positive for PCNA. On the contrary, small foci were almost negative for PCNA in all groups. For that reason we estimated PCNA index within the areas of GST-P foci of the same size (20–30 cell foci). In the livers of rats treated with PB at a dose of 2 p.p.m., the PCNA index for GST-P positive foci was significantly lower than in the DEN control group (Table III and Figure 1C). Alterations to cell proliferation in the areas of hepatic foci and oxidative base modifications in the rat liver DNA were similar in the low dose group (Figure 2). Furthermore, cell proliferation was also slightly suppressed in the background liver parenchyma of 2 p.p.m. PB treated rats (Table III). Administration of PB at 500 p.p.m. without DEN initiation resulted in a non-significant decrease in PCNA positive cells in normal-appearing tissue (Table III).

Evaluation of apoptosis (ssDNA)

Analysis by double immunohistochemistry for GST-P and ssDNA demonstrated significant increase of apoptosis in the livers of rats treated with DEN alone and DEN followed by PB at 500 p.p.m., predominantly in the centralobular region. Hepatic foci were often localized in the areas positive for ssDNA (Figure 1D). In contrast, in the livers of rats given DEN and PB at a dose of 2 p.p.m. only a few apoptotic cells in the normal-appearing background liver tissue were found in line with the lowering of 8-OHdG (Figure 1E, Figure 2 and Table III). Immunohistochemistry for 8-OHdG and double staining for GST-P and apoptosis performed with serial sections of the liver tissue revealed ssDNA staining consistently in the regions positive for 8-OHdG (Figure 1D and F). No significant changes in apoptosis were apparent within areas of GST-P positive foci between initiation control, low, median and high dose PB groups. All GST-P positive foci were almost negative for apoptosis (Table III, Figure 1D and E).

Activation of detoxifying enzymes

Similar to the results for DMPO–OH spectrum, dose-dependent elevation of cytochrome P-450 total content in the liver microsomal fraction was found in experiments 1 (Table II) and 2 (data not shown). Significant decrease of liver P-450 was found in the DEN control group as compared with the vehicle controls. Slight but significant increase of CYP3A2-2β (0.16 ± 0.01 nmol/mg/min), CYP3A2-6β (0.64 ± 036 nmol/mg/min), CYP2C11 (0.86 ± 0.39 nmol/mg/min) and OR (0.07 ± 0.003 nmol/mg/min) activities was evident after 8 days treatment with PB at 2 p.p.m. in the short-term experiment (Figure 3A). In the long-term study, PB administration at 2 p.p.m. after DEN initiation enhanced activities of CYP3A2 and CYP2C11 without significance (data not shown). However, no change in mRNA expression or protein levels of CYP2B1/2, CYP3A2 and CYP2C11 was found after treatment with the low dose of PB in experiments 1 (data not shown) and 2 (Figure 3B–D). In contrast, when animals were administered PB at a dose of 15 p.p.m. and higher, a significant increase in CYP2B1/2 and CYP3A2 mRNAs, proteins and activities was detected (Figure 3 and data not shown). In addition, CYP2C11 protein and activity levels were decreased after 8 days of PB administration at a dose of 500 p.p.m. The data were consistent with the previously published results (36). However, in experiment 1 CYP2C11 protein and mRNA were significantly enhanced after application of PB for 10 weeks, suggesting that short-term treatment results in CYP2C11 inhibition, however, continuous application leads to its elevation due to an increase of mRNA expression. Alterations of CYP3A2-2β, CYP3A2-6β and CYP2C11 activities were clearly apparent with that of OR, suggesting enzymes of the hepatic detoxification system to be activated by PB treatment at a dose of 2 p.p.m., but without any change in mRNA or protein levels of P-450 isoforms.

cDNA microarray analysis

The results of Atlas 1.2 Rat cDNA microarray analysis of differentially expressed genes, which were up- or down-regulated at least 2-fold in the livers of rats, obtained in experiment 1 after 10 weeks of PB administration are presented in Table IV. Ogg1, CYP2B1, CYP3A2 and CYP2C11 genes were not included in the analysis and their mRNA expression was detected by real-time and competitive RT–PCRs. Reproducibility and specificity of the data was confirmed by repeating microarray analysis two times. The results obtained from three determinations were compared with Atlas Image Software and the strong concordance of the results was found. Furthermore, the data obtained by cDNA microarray on the expression of GAD65 were checked by quantitative RT–PCR analysis, which showed the same pattern (data not shown). In the livers of animals in the DEN control group, not administered PB, over-expression of GST-P subunit (GST7-7), rac-beta serine/threonine kinase (rac-PK-beta), alcohol dehydrogenase class 1 (ADH1), plakoglobin, annexin V and interleukin-4-receptor was found. Furthermore, genes including GAD65, thymidylate synthase (TS), gamma aminobutyric acid (GABA-A) receptor beta subunit, guanylyl cyclase, proteasome subunit RC 10-II and orphan nuclear receptor TR4 were down-regulated. PCNA and others related to cell proliferation genes were slightly up-regulated, however, without significance.

Comparison between groups treated with DEN alone and DEN followed by PB at a dose of 2 p.p.m. revealed increased expression of GAD65, phospholipase C delta and chloride channel CIC-7 in the latter (Figure 4 and Table IV). Mitogen-activated protein kinase p38 (p38), Lyn tyrosine protein kinase, rac-beta serine/threonine kinase (rac-PK-beta), calponin and calcium/calmodulin-dependent protein kinase type IV genes were down-regulated. When animals were administered PB at a dose of 15 p.p.m., an enhancement of cytochrome P-450 isoenzyme 3A1, nuclear tyrosine phosphatase PRL-1, presenilin and reduction of plakoglobin gene expression were detected. Furthermore, results of cDNA microarray showed that treatment with PB at a dose of 500 p.p.m. after DEN initiation resulted in up-regulation of the hepatic detoxifying enzymes including cytochrome P-450 isoforms 3A1, 2C7 and 4A8, NADPH P-450 cytochrome reductase and GST-P and -Ya subunits, and nuclear tyrosine phosphatase PRL-1, guanine nucleotide binding protein G(O) alpha subunit, DOPA decarboxylase, DOPA/tyrosine phosphatase and presenilin 2 genes. Administration of PB at a dose of 500 p.p.m. without DEN initiation led to the increase in expression of the GST-Ya subunit, cytochrome P-450 isozymes 3A1 and B5, NADPH P-450 cytochrome reductase, guanine nucleotide binding protein G(O) alpha subunit and calponin.

Discussion

The present study demonstrated a hormetic effect of PB treatment on rat hepatocarcinogenesis, inhibited development of GST-P positive foci and liver tumors by PB at a dose of 2 p.p.m. being evident after administration for 10 and 33 weeks, respectively. The number and area of GST-P positive foci in the low dose group were decreased by 25% and 31%, respectively. Furthermore, an impressive inhibition in the multiplicity of HCC was observed, although the multiplicity of adenoma and incidences of HCC and adenoma were decreased without significance. In contrast, the chemical administered to rats at a dose of 500 p.p.m. resulted in promotion in line with the literature (49).

8-OHdG, a marker of oxidative DNA damage, potentially involved in carcinogenesis in various experimental models, is known to cause mutations (1214,37,38). Induction of a significant and steady elevation of 8-OHdG is thought to play an integral role in carcinogenesis (39). Significant increase of 8-OHdG levels in the DEN initiation group over the vehicle control observed in the present study supported this concept. Therefore, the suppression of GST-P positive foci and tumor development by PB at low dose might be related to the observed reduction in 8-OHdG level in the DNA of hepatocytes. Conspicuous increase of GST-P positive foci numbers and areas after 10 weeks of PB administration at a dose of 500 p.p.m., on the other hand, was associated with elevation of intracellular OH and 8-OHdG levels in the rat liver DNA. Enhanced protein and activity levels of P-450 isoenzymes CYP2B1/2 and CYP3A2 might be related to accumulation of OH and 8-OHdG with the high dose (21).

The induction of OGG1 mRNA expression observed in this study suggested that a reason for the 8-OHdG decrease in the background liver parenchyma in the low dose group might have been activation of 8-OHdG repair. This has been shown to be elevated in response to DNA damage generated by hydroxyl radicals and is attributable to glycosylases, endonucleases and lyase activity (40,41). Our results revealed the enhancement of Ogg1 mRNA expression after treatment at 2 p.p.m. as well as 15 and 500 p.p.m. dose of PB. However, the levels of hydroxyl radicals and oxidative DNA base modifications differed between the low and high dose groups. OH increased dose-dependently, while 8-OHdG generation was suppressed by PB at a dose of 2 p.p.m. and markedly enhanced by the high dose treatment. It is thus possible that Ogg1, induced by generation of OH, might swing the balance between generation and elimination of 8-OHdG in DNA. Hydroxyl radicals and 8-OHdG levels were presumably very high due to the increase of cytochrome P-450 CYP2B1/2 and CYP3A2 in the livers of rats treated with PB at a dose of 500 p.p.m., and the repair of oxidative base modifications was insufficient to prevent the elevation of 8-OHdG.

The present data demonstrated an enhancement of cell proliferation particularly in the foci consisting of more than 20 cells in the groups treated with 15 and 500 p.p.m. dose of PB. However, the PCNA index was significantly suppressed with the low dose. As detected by cDNA microarray analysis, PB treatment at low dose enhanced gene expression of GAD65, an enzyme involved in the synthesis of GABA. Furthermore, the expression of GAD65 was reduced in the livers of animals of the initiation group as compared with the vehicle control, but no significant differences in GAD65 expression were observed in 15 and 500 p.p.m. dose groups. Therefore, the induction of GAD65 might be specific for PB activity at a dose of 2 p.p.m. Recently, it has been reported about the negative correlation between the expression of GABA-A receptors in hepatocytes and thymidine incorporation in liver specimens but without evidence of causal relationship, and GABA-B receptor subtype involvement in mechanisms of hepatocyte DNA synthesis and mediation of growth stimulation (4244). Therefore, the increased mRNA expression of GAD65 by the low dose of PB observed in the present study may reflect something interesting; however, the role of GABA and the subtypes of GABA receptors in liver proliferation is not at all clear and needs further investigation. Furthermore, suppression of gene expression of signal transduction modulators, such as MAP kinase p38, Lyn tyrosine kinase, calcium-calmodulin-dependent protein kinase and rac-beta serine/threonine kinase detected with the use of cDNA microarray analysis might have been related to the inhibitory effect of PB on cell proliferation. To support this idea, the analysis of the levels of the active (phosphorylated) forms of kinases will be performed in our future investigations.

Kitano et al. (15) have reported that development of transforming growth factor-α (TGF-α) positive foci in the livers of rats treated with PB at different doses is basically similar to that of GST-P positive foci. Hence, the suppression of cell proliferation detected by double immunohistochemical staining in the areas of hepatic foci by the treatment with low dose of the chemical and its elevation in the groups treated with 15 and 500 p.p.m. suggests a link to the previously observed effect of PB administration on formation of TGF-α positive foci. Rise of the PCNA labeling index within the GST-P areas observed in our study 10 weeks post-treatment with 15 and 500 p.p.m. dose of PB is generally in line with the previously published results regarding hepatocarcinogenesis (45). Furthermore, as detected by cDNA microarray analysis and competitive RT–PCR, PB treatment at high dose enhanced gene expression of enzymes involved in xenobiotic metabolism in the liver consistently with previously reported data (46). Its administration at 15 and 500 p.p.m. after DEN initiation increased expression of nuclear tyrosine phosphatase PRL-1, affecting the cell growth. Moreover, PB at 500 p.p.m. enhanced gene expression of guanine nucleotide binding protein G(O) alpha subunit, which may result in activation of GTPase signal transduction pathway and induction of cell proliferation (47).

A reduction of apoptosis in the normal-appearing liver tissue surrounding the GST-P positive foci due to the inhibition of oxidative DNA damage after treatment with low dose PB might suppress their enlargement. Apoptosis of the normal-appearing tissue might be one of the factors regulating the size of foci, as the enlargement of GST-P positive foci probably is more difficult in the case of non-apoptotic surrounding tissue. The results of the immunohistochemical examination of GST-P, ssDNA and 8-OHdG in serial sections showed that 8-OHdG and ssDNA positive hepatocytes were located in the same area. Furthermore, in our previous study, double immunohistochemistry for 8-OHdG and apoptosis in the livers of rats treated for 8 days with PB at a dose of 500 p.p.m. revealed high levels of concordance between induction of 8-OHdG and apoptosis (ssDNA) (21). Thus, the data of the present experiment are in agreement with our previous results. Moreover, recently, depression of apoptosis by PB has been explained on the basis of its ability to inhibit p53 (48), p21WAF1/Cip1 (49), and enhance bcl-2 gene family expression (50). PB may also co-operate with c-myc and TGF-α in the selective inhibition of apoptosis through diverse molecular pathways (51).

A decrease of P-450 expression in liver nodules induced by DEN has been found as compared with the normal tissue (52). The data are consistent with the results of the present experiment showing a reduction of P-450 total content in the liver microsomal fraction of the DEN control animals. Furthermore, it has been reported that suppressive effect of PB on the development of pre-neoplastic lesions might be due to the stimulation of hepatic drug-metabolizing enzymes, which detoxify carcinogens (17). Therefore, activation of P-450 isoenzymes CYP2C11, CYP3A2 and OR in the liver microsomes observed in the present study after administration of PB at low dose, if not accompanied by induction of their gene expression leading to the tremendous elevation of OH, might have a protective effect. No significant differences in mRNA or protein levels of CYP3A2, CYP2B1/2 and CYP2C11 were detected in the low-dose group in both experiments, suggesting that those enzymes of the detoxification system were activated, but their gene expression was not altered. In contrast, P-450 isoenzymes mRNA and protein levels were conspicuously elevated in the high dose-treated groups, and the data were in line with the previously noted results (11). In addition, an inhibitory effect of PB on development of GST-P positive foci in a medium-term rat liver bioassay was reported previously to correlate with changes in CYP3A2 isoenzyme protein content, which was suppressed by the low and enhanced by high doses of PB (15). However, in this study we did not observe any effect of PB at low dose either on the levels of CYP3A2 mRNA or protein expression, possibly due to differences in experimental models.

In conclusion, PB, a promoter of hepatocarcinogenesis in rodents at high dose, at low dose inhibited the formation of GST-P positive foci and liver tumors. Our data demonstrate the presence of a threshold for the promoting effects of PB and the lack of linearity in the low-dose area of the dose–response curve, rather providing evidence of a J-shape. Suppression of 8-OHdG formation and apoptosis in background normal-appearing liver parenchyma, cell proliferation in the areas of hepatic foci, and alterations in the expression of genes related to control of cell proliferation, might explain in part the hormetic inhibition of hepatocarcinogenesis by PB at low dose.

Acknowledgments

We thank Emi Kawakami, Chizu Imazato, Miyoko Yamanaka and Kaori Touma for their technical assistance, and Mari Dokoh and Yuko Onishi for their help during preparation of this manuscript. This research was supported by a grant from the Japan Science and Technology Corporation, included into the Project of Core Research for Evolutional Science and Technology (CREST) in Japan.

Footnotes

References