Inhibition of Mouse Hepatocyte Gap Junctional Intercellular Communication by Phenobarbital Correlates with Strain-Specific Hepatocarcinogenesis

Kristy A. Warner*, Martha J. Fernstrom* and Randall J. Ruch*,1

* Department of Pathology, Medical College of Ohio, 3055 Arlington Ave., Toledo, Ohio 43614

Received June 25, 2002; accepted October 29, 2002


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The inhibition of gap junctional intercellular communication (GJIC) is a common effect of nongenotoxic carcinogens and might be a biomarker for these agents. To further test this relationship, we hypothesized that phenobarbital would inhibit mouse hepatocyte GJIC and this would correlate with strain-specific hepatocarcinogenicity. Phenobarbital is a strong nongenotoxic hepatocarcinogen in B6C3F1 mice, but not in C57BL/6 mice. Hepatocytes were isolated from males of both strains, placed in coculture with rat liver epithelial cells, and treated with phenobarbital for up to 14 days. Male mice were also administered PB by single intraperitoneal injection (0.1 mg/kg), then sacrificed 24 h later, or given phenobarbital in the drinking water (500 ppm) for 14 days before sacrifice. GJIC was assayed in cocultures by fluorescent dye microinjection and in isolated liver tissue by fluorescent dye "cut-loading." Phenobarbital decreased GJIC only in cultured B6C3F1 hepatocytes; this was dose-responsive and temporary, because hepatocyte GJIC returned to control levels within 24 h of phenobarbital exposure. Administration of phenobarbital to mice for 14 days also decreased hepatocyte dye coupling in B6C3F1 liver, but this effect was not seen in C57BL/6 mice or observed after a single administration of the drug. Phenobarbital did not alter connexin32 and connexin26 expression, but increased hepatic Cyp2b1 expression and the liver weight:body weight ratio in both strains. In summary, phenobarbital inhibited mouse hepatocyte GJIC in vivo and in vitro and in correlation with strain-specific hepatocarcinogenicity. These data support the hypothesis that decreased GJIC is a biomarker for nongenotoxic carcinogens and involved in their carcinogenic mechanism.

Key Words: gap junctions; hepatocyte; cytochrome P450; tumor promotion; nongenotoxic carcinogens; phenobarbital; biomarkers.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The development of a cancer cell is due to the progressive accumulation of mutations and stable, nonmutational (epigenetic) alterations in gene expression and gene product (protein) function (MacPhee, 1998Go). Chemical carcinogens can be classified as genotoxic or nongenotoxic (Butterworth and Bogdanffy, 1999Go). Genotoxic carcinogens are mutagenic and may bind to DNA and form adducts that result in miscoding, deletion, and frame-shift mutations; induce DNA breaks and rearrangements; or disrupt the error-free replication of undamaged DNA by influencing DNA polymerase fidelity. In contrast, nongenotoxic carcinogens are not mutagenic but may stimulate cell proliferation, inhibit apoptosis, increase inflammation, and/or induce stable or transient epigenetic changes in critical genes in terminally differentiated or proliferating cells. The induction of cell proliferation and inhibition of apoptosis and terminal differentiation by carcinogens leads to the accumulation of cells with increasingly stable neoplastic changes and, ultimately, the appearance of a fully malignant cell.

Nongenotoxic carcinogens are difficult to identify. There are no widely accepted short-term tests or bioassays for nongenotoxic carcinogens; this contrasts with genotoxic carcinogens for which many short-term tests and biomarkers exist (Butterworth and Bogdanffy, 1999Go; Trosko et al., 1998Go). Presently, nongenotoxic carcinogens are identified in chronic animal bioassays. This is problematic for regulatory agencies, because humans are exposed to many nongenotoxic agents and most of these are not tested in chronic bioassays because of high costs. Thus, it would be beneficial to develop mechanistically relevant, short-term tests or biomarkers that identify nongenotoxic carcinogens that can be applied across species.

The inhibition of gap junctional intercellular communication (GJIC) is a mechanistically relevant effect common to most nongenotoxic carcinogens and might be useful as a biomarker for these agents (Budunova and Williams, 1994Go; Klaunig and Ruch, 1990Go; Rosenkranz et al., 2000Go; Trosko and Chang, 1988Go). Gap junctions are a type of cell–cell junction that consists of clusters of channels directly linking the interiors of adjacent cells and permitting the flow between cells of ions and molecules less than 1 kDa in diameter (Bruzzone et al., 1996Go). This exchange is known as gap junctional intercellular communication (GJIC), which is involved in cellular homeostasis, growth regulation, apoptosis, differentiation, and other functions (Ruch, 2000Go). The inhibition of GJIC by nongenotoxic carcinogens is, therefore, a likely factor in their ability to induce cell proliferation and inhibit apoptosis and terminal differentiation.

Strong associations have been observed among many chemicals between the inhibition of GJIC in cell cultures and nongenotoxic carcinogenic activity (Klaunig and Ruch, 1990Go; Budunova and Williams, 1994Go; Trosko and Chang, 1988Go). Some studies have used in vivo models and have also reported a similar correlation (Chaumontet et al., 1996Go; Ito et al., 1998Go; Kolaja et al., 2000Go; Sai et al., 2000Go; Smith et al., 2000Go), but there is less data for in vivo than for in vitro systems. In the present study, we hypothesized that the nongenotoxic rodent hepatocarcinogen phenobarbital (PB) would inhibit GJIC in hepatocytes of a mouse strain (B6C3F1) that was highly sensitive to PB hepatocarcinogenicity, and would have little effect on hepatocyte GJIC in a much less sensitive strain (C57BL/6) (Diwan et al., 1986Go; Weghorst et al., 1989Go). We previously reported that PB inhibited B6C3F1, but not C57BL/6 mouse hepatocyte GJIC in vitro, following limited exposure durations of less than 24 h (Klaunig and Ruch, 1987Go; Ruch and Klaunig, 1988Go). But we did not investigate whether this correlation occurred in vivo or following longer exposures in vitro. To address this, we have used fluorescent dye "cut-loading" to assess hepatocyte GJIC in vivo and fluorescent dye microinjection of mouse hepatocyte/rat liver epithelial cell cocultures to address the in vitro effects of PB.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Reagents and mice.
PB, dimethylsulfoxide (DMSO), dexamethasone, and fetal bovine serum (FBS) were purchased from Sigma Chemical Co. (St. Louis, MO). Other reagents were obtained from sources identified below. Male B6C3F1 and C57BL/6 mice (25–30 g) were purchased from Charles River Laboratories (Wilmington, MA). They were housed in polycarbonate cages and fed Purina Lab Chow (Ralston Purina Co., St. Louis, MO) and water ad libitum.

Hepatocyte isolation, culture, and treatment with PB.
Hepatocytes were isolated from adult male mice by two-stage perfusion of the liver with collagenase D (Boehringer Mannheim, Indianapolis, IN) (Ruch and Klaunig, 1988Go). Isolated hepatocytes were >90% viable and were plated (1 x 105 cells/cm2) onto 50% confluent cultures of WB-F344 rat liver epithelial cells (Ren and Ruch, 1996Go) in standard plastic tissue culture dishes. The culture medium was Richter’s minimal essential medium (Irvine Scientific Co., Santa Ana, CA) supplemented with 5% FBS, dexamethasone (1 µM), and gentamicin (Life Technologies, Gaithersburg, MD; 40 mg/ml). The cocultures were refed 2 and 24 h after plating of hepatocytes and were treated 48 h later with PB dissolved in DMSO or with DMSO (1 µl/ml; controls). The cocultures were refed and retreated every 3 days.

Assessment of cultured hepatocyte GJIC by fluorescent dye microinjection.
Hepatocyte GJIC was evaluated by microinjection of fluorescent Lucifer yellow CH dye (Sigma) (Ruch and Klaunig, 1988Go). Three dishes per treatment condition were sampled and 10 hepatocytes per dish were microinjected and filled with dye. All hepatocytes directly adjacent to microinjected ones were evaluated for dye uptake (dye-coupling) 5 min after microinjection. The data from three replicate cultures were pooled, and the percentage of dye-coupled cells for all injections was determined and is presented in the figures. Treatment groups were compared using the Chi-square test (Gad and Weill, 1986Go). GJIC in WB-F344 cells was also evaluated by dye microinjection.

Hepatocyte viability in cocultures.
Cocultured hepatocytes were counted to determine whether PB affected hepatocyte survival. Coculture medium was removed and replaced with trypan blue dye. Fifteen min later, the dye was removed and the cells were washed with Ca+2-Mg+2–free phosphate-buffered saline (PBS). Trypan blue dye-excluding (viable) hepatocytes in nine randomly selected fields on each dish were counted at x400 magnification. The total number of viable hepatocytes per dish was determined by multiplying the mean number of viable hepatocytes per field by the surface area of the 35 mm dish (962 mm2) and dividing that number by the area of the microscopic field (0.076 mm2). Three replicate dishes were evaluated per treatment group. Groups were compared statistically by Student’s t-test.

Treatment of mice with PB.
Mice were randomly divided into eight groups of 4–6 animals each (Table 1Go). They were administered PB by a single intraperitoneal injection (0.1 mg/kg in sterile 0.9% saline; groups 2 and 6) or given in drinking water (500 ppm, groups 4 and 8). Control mice were injected with saline (groups 1 and 5) or were given deionized drinking water (groups 3 and 7). Mice injected with PB or saline were euthanized 24 h after injection. Mice given PB in the drinking water were euthanized after 14 days of administration.


View this table:
[in this window]
[in a new window]
 
TABLE 1 Mouse Treatment Groups
 
Assessment of hepatocyte GJIC ex vivo by fluorescent dye "cut-loading."
The effect of PB on GJIC in intact mouse liver was evaluated by fluorescent dye cut-loading (Chaumontet et al., 1996Go; Ito et al., 1998Go; Kolaja et al., 2000Go; Sai et al., 2000Go; Smith et al., 2000Go). Livers were removed quickly from euthanized mice and weighed. Approximately 0.5 g portions of the right liver lobe were flash frozen in liquid nitrogen for connexin expression studies. Other portions of the right lobe were rinsed in PBS, submerged in a solution of 0.05% Lucifer yellow CH (Sigma) in PBS, then cut in half using a scalpel blade. The cut liver pieces were left immersed in the dye solution for 5 min to permit dye uptake and transfer, and then were rinsed in PBS and fixed by immersion in 10% phosphate-buffered formalin. After fixation for 4 days, the liver tissues were clarified by soaking in DMSO for 48 h. The tissues were embedded in paraffin and sectioned (5 mm) by standard methods. The tissues were sectioned transversely to the dye-loaded, cut surface.

The sections were examined microscopically using a Nikon fluorescence microscope equipped with a fluorescein filter set. Dye-loaded hepatocytes along the cut edges were identified easily by their bright yellow fluorescence. GJIC was quantified by counting the number of dye-stained hepatocytes perpendicular to the cut edge. Dye-loaded cells along the cut edge were not counted. Four randomly selected points along each cut edge were evaluated. For each mouse, three liver pieces were "cut-loaded," sectioned, and evaluated in this way for a total of 12 counts per animal. The mean number of dye-coupled cells for each animal was determined from these 12 evaluations. From these means, treatment group means and standard deviations were determined, then compared statistically using Student’s t-test (Gad and Weill, 1986Go).

Analyses of Cx26 and Cx32 expression and localization.
Northern and Western blot assays and fluorescent immunohistochemistry were performed as described (Ren and Ruch, 1996Go) to investigate the effects of PB on mouse liver Cx26 and Cx32 mRNA and protein expression. For Northern analyses, total RNA was isolated from cocultures or from frozen liver pieces (~0.5 g). Frozen liver was first pulverized under liquid nitrogen. Total RNA was isolated from cultures or pulverized tissue using 1 ml of ice-cold Tri-Reagent (Molecular Research Center, Inc., Cincinnati, OH) according to the manufacturer’s protocol. The purified RNA was quantified spectrophotometrically, separated by electrophoresis in 1% agarose gels (10 µg total RNA per sample), transferred to Hybond N+ nylon membranes (Amersham-Pharmacia Biotech, Piscataway, NJ) by vacuum blotting, and cross-linked to the membranes by UV light exposure. The membranes were hybridized with Cx32, Cx26, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) radiolabeled probes prepared from full-length cDNAs. Membranes were exposed to X-ray film (Kodak Biomax MR; Fisher Scientific) with intensifying screens for 1–3 days. Films were developed in an automatic X-ray film developer.

For Western blot analyses of Cx32 and Cx26 protein, membrane-enriched extracts were prepared from cell cultures or pulverized frozen liver tissue (~0.5 g per sample). Cells were harvested from culture dishes by scraping into ice-cold PBS, then pelleted by centrifugation. One ml of ice-cold hypotonic lysis buffer (10 mM Tris, pH 7.5; 1 mM iodoacatamide; 2 mM phenylmethylsulfonylfluoride) was added to each cell pellet or tissue sample and these were disrupted by sonication on ice. The cell fragments were alkalinized by the addition of 1.5 µl of 40 mM NaOH, then pelleted by centrifugation (13,000 x g for 30 min). The pellets were dissolved in 2% sodium dodecylsulfate (SDS) and protein concentrations were determined using the Bio-Rad DC Protein Assay (Bio-Rad Corp., Richmond, CA). Proteins (40 µg per sample) were separated by SDS–PAGE on 12% polyacrylamide gels under reducing conditions. Proteins were transferred to Immobilon-P membranes and Cx26 and Cx32 were detected using mouse monoclonal anti-Cx26 and anti-Cx32 antibodies (Zymed, South San Francisco, CA).

The effects of PB on Cx26 and Cx32 localization were determined by immunohistochemistry. Frozen liver sections (5–6 µ) were prepared using a cryostat, mounted onto microscope slides coated with aminopropyl triethoxysilane (Sigma), fixed in 100% acetone, and stored at –80°C until use. For immunostaining, the sections were warmed to room temperature, rehydrated in PBS, and incubated sequentially in 3% normal goat serum (Jackson ImmunoResearch Laboratories, West Grove, PA) in blocking buffer (4% milk, 10% bovine serum albumin, and 0.3% Triton X-100) for 1 h, rabbit polyclonal anti-Cx26 or anti-Cx32 antibody (1:100 in PBS; Zymed) for 3 h, and fluorescein isothiocyanate-conjugated goat antirabbit antibody (1:100 in PBS; Jackson Immunoresearch) for 2 h. Sections were washed extensively in PBS after antibody incubations and mounted with Anti-Fade Solution (Molecular Probes, Eugene, OR).

Northern blot analyses of Cyp2b1 expression.
Northern blot analyses were performed as described above to determine the effects of PB on Cyp2b1 mRNA expression in the mouse liver. The probe was a single-stranded DNA oligonucleotide specific for Cyp2b1, 5'-GGTTGGTAGCCGGTGTGA-3' (Omiecinski et al., 1985Go) (Life Technologies) and was radiolabeled with [{gamma}32P]ATP (Amersham-Pharmacia Biotech) using T4 polynucleotide kinase (Life Technologies).


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Effects of PB on hepatocyte GJIC in vitro.
The dose- and time-dependent effects of PB on cocultured B6C3F1 and C57BL/6 hepatocyte GJIC were determined. PB decreased B6C3F1 hepatocyte GJIC after four-hour and one-day treatments, and the effect was maximal at approximately 0.50 mM PB (Fig. 1AGo). With longer treatments (7 and 14 days), no inhibition was observed. In contrast, PB did not affect C57Bl/6 hepatocyte GJIC at any sampling time (Fig. 1BGo). Similarly, PB did not decrease WB-F344 rat liver epithelial cell GJIC in B6C3F1 or C57BL/6 cocultures (Fig. 1CGo).



View larger version (19K):
[in this window]
[in a new window]
 
FIG. 1. Dose- and time-dependent effects of phenobarbital (PB) on gap junctional intercellular communication (GJIC) in mouse hepatocyte/WB-F344 rat liver epithelial cell cocultures. GJIC was assayed by fluorescent dye microinjection and enumeration of dye-stained, directly adjacent, neighboring cells. The dye-coupling percentage is representative of >50 neighboring hepatocytes or WB-F344 cells evaluated from three cultures: (A) in male B6C3F1 mouse hepatocytes cocultured with WB-F344 cells; (B) in male C57BL/6 mouse hepatocytes cocultured with WB-F344 cells; and (C) in WB-F344 cells cocultured with male B6C3F1 hepatocytes or C57BL/6 hepatocytes. PB was dissolved in DMSO, and control cultures (0 mM dose) were treated with DMSO (1 µl/ml). Significant inhibition occurred in B6C3F1 hepatocytes treated with PB (all doses) for 4 h and for 1 day (p < 0.05).

 
Effects of PB on connexin expression in vitro.
Connexin protein expression was determined by Western blotting in PB-treated cocultures. After a 24-h exposure to 2 mM PB, there were no differences in Cx26, Cx32, or Cx43 protein content between DMSO control and PB-treated cocultures of both strains of mice (Fig. 2AGo). The multiple bands evident in the Cx43 blots are due to different phosphorylated forms of the protein (Ruch et al., 1994Go). In cocultures treated with PB for 14 days, however, increased levels of Cx26 and Cx32 protein and decreased content of Cx43 were evident (Fig. 2BGo). Similar changes in Cx26, Cx32, and Cx43 mRNA were observed by Northern blotting (data not shown).



View larger version (55K):
[in this window]
[in a new window]
 
FIG. 2. Effects of phenobarbital (PB) on connexin content in B6C3F1 and C57BL/6 mouse hepatocyte/WB-F344 rat liver epithelial cell cocultures. The cocultures were treated with 2 mM PB or 0.1% DMSO (controls) for 24 h (A) or 14 days (B). Western blot analyses were performed using mouse monoclonal antibodies to connexin26 (Cx26), connexin32 (Cx32), and connexin43 (Cx43)

 
The increase in hepatocyte connexin (Cx26 and Cx32) and decrease in WB-F344 connexin (Cx43) suggested there might be increased numbers of hepatocytes in the PB-treated cocultures. PB has been reported to improve the survival of primary cultured hepatocytes (Miyazaki et al., 1990Go). To determine whether hepatocytes were more prevalent in the PB-treated cocultures, viable (trypan blue dye-excluding) hepatocytes were counted in cocultures treated for 14 days. Figure 3Go illustrates that PB treatment for 14 days resulted in approximately a two-fold higher number of hepatocytes in the cocultures. Since mouse hepatocytes do not undergo mitosis in vitro when cultured in media containing fetal bovine serum, it is likely that PB increased the survival of the hepatocytes. It is noteworthy that an equivalent effect was observed with both strains of hepatocytes.



View larger version (17K):
[in this window]
[in a new window]
 
FIG. 3. Effects of phenobarbital (PB) on hepatocyte cell number in B6C3F1 and C57BL/6 mouse hepatocyte/WB-F344 rat liver epithelial cell cocultures. The cocultures were treated with 2 mM PB or 0.1% DMSO (controls) for 14 days. Viable hepatocytes were identified by trypan blue dye exclusion and counted microscopically (*p < 0.05 vs. DMSO group, mean ± SD, n = 3 cocultures).

 
Effects of in vivo PB treatment on hepatocyte dye-coupling and connexin expression.
To determine whether PB could inhibit GJIC in vivo, B6C3F1 and C57BL/6 mice were administered PB either as a single injection of PB or for 14 days in the drinking water. One day later, mice were euthanized and liver GJIC was assayed by "cut-loading." In control mouse liver of both strains, dye spread to approximately four interior hepatocytes (Fig. 4Go). A single intraperitoneal injection of PB (24 h treatment) did not affect GJIC in either strain of mouse. However, administration of PB for 14 d reduced GJIC in B6C3F1 mouse liver by approximately 50%, but had no effect in C57BL/6 liver. Figure 5Go illustrates dye coupling in cut-loaded control (Figs. 5AGo and 5BGo) and PB-treated (14 days, Figs. 5CGo and 5DGo) B6C3F1 mouse liver. Note the hepatocellular hypertrophy in the liver of the PB-treated mouse (Fig. 5CGo).



View larger version (25K):
[in this window]
[in a new window]
 
FIG. 4. Effects of in vivo administration of phenobarbital (PB) on B6C3F1 and C57BL/6 hepatocyte GJIC. Mice were given a single intraperitoneal injection of PB (0.1 mg/kg) and sampled 24 h later, or were administered PB in the drinking water (500 ppm) for 14 days before sampling. GJIC was assayed by fluorescent dye cut-loading and dye-stained cells perpendicular to the cut edge were enumerated microscopically (mean ± SD, n = 4 mice, *p < 0.05 vs. control group).

 


View larger version (109K):
[in this window]
[in a new window]
 
FIG. 5. Effects of in vivo phenobarbital treatment on dye coupling in B6C3F1 mouse liver. Mice were administered normal drinking water (control, panels A and B) or phenobarbital in the drinking water (500 ppm, C and D) for 14 days, then the mice were euthanized and GJIC was assayed by fluorescent dye "cut-loading." Photomicrographs depict liver sections stained with hematoxylin and eosin (A and C) or fluorescent dye transfer after cut-loading (B and D).

 
PB did not affect Cx26 or Cx32 expression or localization in either strain of mouse. Northern and Western blotting revealed similar levels of Cx26 and Cx32 mRNA and protein in both strains of mouse liver after single or 14-day administration of PB (Fig. 6Go). The apparent reduction in Cx26 mRNA in the 4th lane of panel A of this figure was not reproducible. Likewise, PB did not alter the localization of Cx32 in the mice. As shown in Figure 7Go, punctate staining was evident at hepatocyte-hepatocyte interfaces and little cytoplasmic staining was present.



View larger version (78K):
[in this window]
[in a new window]
 
FIG. 6. Connexin expression in mouse liver after in vivo administration of phenobarbital (PB). B6C3F1 and C57BL/6 mice were given a single intraperitoneal injection of PB (0.1 mg/kg) and sampled 24 h later or were administered PB in the drinking water (500 ppm) for 14 days before sampling. Control mice (C) were injected with normal saline or given deionized drinking water. Northern (A) and Western (B) analyses were performed using specific cDNA probes or monoclonal antibodies for connexin26 (Cx26), connexin32 (Cx32), or glyceraldehyde-3-phosphate dehydrogenase (GAPDH, loading control).

 


View larger version (188K):
[in this window]
[in a new window]
 
FIG. 7. Effect of phenobarbital on connexin26 and connexin32 immunostaining in B6C3F1 mouse liver. Mice were administered normal drinking water (A and B) or 500 ppm PB in the drinking water (C and D) for 14 days. Connexin26 (A and C) and connexin32 (B and D) were detected using affinity-purified rabbit polyclonal antibodies followed by fluorescein isothiocyanate-conjugated goat antirabbit IgG secondary antibody.

 
Effects of PB on Cyp2b1 expression.
To investigate whether PB was active in both strains of mice, we evaluated the expression of Cyp2b1 and the liver weight/body weight ratios. By Northern analyses, Cyp2b1 mRNA was not detected in control mouse liver, but it was increased dramatically in livers of both strains of mice treated with PB for 14 days (Fig. 8Go). Similarly, the liver weight:body weight ratio was increased approximately 50% in both strains of PB-treated mice. These results suggest that PB induced Cyp2b1 expression and hepatomegaly equivalently in both strains of mice.



View larger version (36K):
[in this window]
[in a new window]
 
FIG. 8. Effects of phenobarbital (PB) on liver Cyp2b1 mRNA content and hepatomegaly in male B6C3F1 and C57BL/6 mice. Mice were given a single intraperitoneal injection of PB (0.1 mg/kg) and sampled 24 h later or were administered 500 ppm PB in the drinking water and sampled 14 days later. Control mice (C) were injected with normal saline or were given deionized drinking water. Northern analyses (A) were performed using specific probes for Cyp2b1 and glyceraldehyde-3-phosphate dehydrogenase (Gapdh). Liver and body weights (B) were determined at necropsy (mean ± SD, n = 4 mice, *p < 0.05 vs. control group).

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
In this study, PB inhibited GJIC in B6C3F1 mouse hepatocytes in primary coculture with rat liver epithelial cells and in vivo, but had no inhibitory effects on C57BL/6 hepatocytes. The inhibition by PB in vitro was transient and the hepatocytes recovered GJIC within 24 h of PB exposure. In contrast, inhibition of GJIC in vivo was not seen 24 h after a single intraperitoneal injection of the drug, but was observed after a 14-day administration of PB in the drinking water. No changes in hepatocyte connexin expression or localization were evident with PB treatment in vitro or in vivo, but expression of Cyp2b1 and hepatomegaly were induced in both strains of mice. This inhibition of GJIC in B6C3F1 hepatocytes, but not in C57BL/6 hepatocytes, correlates with the much stronger nongenotoxic hepatocarcinogenicity of PB in the former strain (Diwan et al., 1986Go; Weghorst et al., 1989Go).

Many groups have reported that PB and other nongenotoxic carcinogens impair GJIC and have observed a correlation between carcinogenicity and the impairment of GJIC (Budunova and Williams, 1994Go; Klaunig and Ruch, 1990Go; Trosko and Chang, 1988Go). In fact, the reduction of GJIC appears to be one of the most common effects of nongenotoxic (and some genotoxic) carcinogens described to date. This correlation has been observed across organs and species (including monkey and human) and correlative structure-activity relationships also have been observed (Baker et al., 1995Go; Pugh et al., 2000Go; Ren and Ruch, 1996Go; Upham et al., 1998Go). In fact, when a variety of in vitro tests for carcinogenicity were considered, the inhibition of GJIC correlated best with carcinogenicity (Rosenkranz et al., 2000Go).

These relationships suggest that the inhibition of GJIC is a biomarker for nongenotoxic carcinogens and might be exploited as a short-term bioassay for such agents. Presently, there are no satisfactory short-term tests that identify nongenotoxic carcinogens and two-year rodent bioassays are necessary to do so. Two-year carcinogenicity bioassays are time-consuming, expensive, and consequently not performed for most nongenotoxic agents to which humans are exposed. The development and validation of GJIC assays using a panel of primary cultured cell types or tissues, including those of human origin, might be highly sensitive and predictive of nongenotoxic carcinogenicity and would benefit risk assessment.

Although many groups have reported that PB inhibits hepatocyte GJIC, the mechanism remains unclear. We saw no changes in connexin gene expression at the mRNA or protein levels and did not observe changes in gap junction localization in sensitive B6C3F1 mouse hepatocytes. In rat liver, PB has been reported to alter connexin expression and gap junction localization, although the data are not consistent. Administration of PB for 4 and 11 weeks decreased Cx32 mRNA content in rat liver in one study (Neveu et al., 1994Go), but did not decrease Cx32 mRNA or protein levels in rat liver in other studies (Chaumontet et al., 1996Go; Krutovskikh et al., 1995Go). However, decreased dye coupling and Cx32 immunostaining of membrane gap junctions were seen. The decreased immunostaining was correlated with increased expression of CYP2B1 (Krutovskikh et al., 1995Go) and cytoplasmic localization of Cx32-containing gap junctions (Chaumontet et al., 1996Go). However, we saw no change in Cx32 or Cx26 immunostaining in mouse liver after PB administration (Fig. 7Go). These inconsistent results might reflect species or protocol differences.

The difference we observed between mouse strains, however, is clearly not the result of treatment protocol, but due to genetic differences between the B6C3F1 and C57BL/6 mice. One possibility is strain differences in PB metabolism and detoxification. PB is a potent inducer of Cyp2b1 in rodent liver and the inducible phenotype has been correlated with barbiturate liver carcinogenicity (Rice et al., 1994Go). The C57BL/6 mouse, however, is an exception. Diwan et al. (1986)Go reported that PB induced Cyp2b1 in C57BL/6 mice, and we observed that induction of this cytochrome and hepatomegaly were similar in this strain and B6C3F1 mice. Thus, induction of the cytochrome per se does not appear to account for the strain-specific differences in sensitivity of GJIC to PB. However, cytochrome P450 activity appears to be important in the inhibition of hepatocyte GJIC by PB. Inhibitors of cytochrome P450 activity (e.g., SKF-525A) prevented the blockage of hepatocyte GJIC by PB (Guppy et al., 1994Go). Induction of cytochrome P450 by PB also increased the production of reactive oxygen species due to uncoupling of the cytochrome P450 cycle (Venditti et al., 1998Go) and hepatocyte GJIC is sensitive to free radicals (Ruch and Klaunig, 1986Go). Therefore, the strain differences might be related to relative antioxidant capacity and ROS detoxification in the face of increased cytochrome P450 activity. Mouse strain susceptibility to liver tumor promotion by nongenotoxic agents has been related to oxidant stress and alterations in hepatic antioxidant function (Ahotupa et al., 1993Go; Stevenson et al., 1995Go). Thus, the inhibition of hepatocyte GJIC by PB could involve ROS that have direct or indirect effects on gap junctions. In fact, antioxidants prevented the inhibition of hepatocyte GJIC by PB (Ruch and Klaunig, 1986Go). Differences between B6C3F1 and C57BL/6 hepatocytes might be related to the extent of ROS generation and detoxification.

We also observed that the inhibition of B6C3F1 hepatocyte GJIC in vitro was transient, whereas that in vivo was not, at least after 14-day treatment. The transient inhibition in vitro by PB was reported earlier and was not due to depletion or degradation of the compound, because culture medium from refractory cells retained inhibitory activity towards naïve hepatocytes and treatment of refractory cells with fresh PB had no effect on GJIC (Ruch and Klaunig, 1988Go). Thus, the hepatocytes were truly refractory to PB. The mechanism is unclear but is not due to upregulated connexin expression. The refractory effect is reminiscent of the transient inhibition of GJIC induced by TPA that is due to the downregulation of protein kinase C (PKC) (Ren et al., 1998Go). It is unlikely, however, that PKC is involved in the PB effect, because potent inhibitors of the kinase and pretreatment of hepatocytes for 24 h with TPA did not prevent the reduction of hepatocyte GJIC by PB (Ren et al., 1998Go).

It is also interesting that refractoriness was only observed in vitro. This could be related to the much higher concentrations and more rapid onset of exposure of PB in cultured hepatocytes. These cells were treated instantaneously with 0.25–5.0 mM PB, whereas hepatocytes in vivo were exposed more gradually to the drug as the animals drank the PB-treated water and steady-state serum levels were achieved. These levels are much lower than the in vitro treatments. In male mice administered 500 ppm PB in the drinking water for 14 days, serum PB levels were approximately 10–25 mg/ml (0.04–0.1 mM) (Diwan et al., 1986Go). Thus, the rapid-onset exposure of hepatocytes in vitro to high concentrations of PB might activate a protective mechanism or cause the downregulation of the inhibitory mechanism. The lack of a refractory effect in vivo might be due to the gradual exposure and lower serum concentration of PB or another mechanism.

The maximal inhibition of GJIC by PB was approximately 50% in vivo and in vitro. This suggests some hepatocytes are resistant to PB. This might be related to the lobular heterogeneity of hepatocytes, although we did not investigate lobular position differences in the cut-loading studies. Position-dependent differences in hepatocyte drug sensitivity are due to nutrient, oxygen, and drug concentration gradients across the lobule; differences in Phase I and Phase II metabolic capacity; and hepatocyte age and ploidy (Moslen, 1996Go). Presumably, lobular position affects hepatocyte sensitivity to GJIC inhibitors like PB. Accordingly, Neveu et al. (1994)Go reported the induction of cytochrome P450 and reduction of Cx32 immunostaining was most prominent in centrilobular hepatocytes of rats fed dietary PB.

In summary, PB decreased B6C3F1 mouse hepatocyte GJIC in vivo and in vitro and had no effect on C57BL/6 mouse hepatocyte GJIC. The decrease in GJIC was not due to altered expression or abnormal localization of Cx26 and Cx32, but to a post-translational mechanism. The inhibition of GJIC correlated with strain-specific sensitivities to PB hepatocarcinogenicity and suggests that the inhibition is involved in the carcinogenic mechanism of PB. These data and those from other groups (Chaumontet et al., 1996Go; Ito et al., 1998Go; Kolaja et al., 2000Go; Sai et al., 2000Go; Smith et al., 2000Go) also suggest that in vivo and/or in vitro GJIC assays are mechanism-based biomarkers for nongenotoxic carcinogenicity.


    NOTES
 
1 To whom correspondence should be addressed. Fax: (419) 383-3089. E-mail: rruch{at}mco.edu. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Ahotupa, M., Bereziat, J. C., Mantyla, E., and Bartsch, H. (1993). Dietary fat and phenobarbital-induced alterations in hepatic antioxidant functions of mice. Carcinogenesis 14, 1225–1228.[Abstract]

Baker, T. K., Bachowski, S., Stevenson, D. E., Walborg, E. F., Jr., and Klaunig, J. E. (1995). Modulation of gap junctional intercellular communication in rodent, monkey, and human hepatocytes by nongenotoxic compounds. Prog. Clin. Biol. Res. 391, 71–80.[Medline]

Bruzzone, R., White, T. W., and Paul, D. (1996). Connections with connexins: The molecular basis of direct intercellular signaling. Eur. J. Biochem. 238, 1–27.[Abstract]

Budunova, I. V., and Williams, G. M. (1994). Cell culture assays for chemicals with tumor-promoting or tumor-inhibiting activity based on modulation of intercellular communication. Cell Biol. Toxicol. 10, 71–116.[ISI][Medline]

Butterworth, B. E., and Bogdanffy, M. S. (1999). A comprehensive approach for integration of toxicity and cancer risk assessments. Regul. Toxicol. Pharmacol. 29, 23–36.[ISI][Medline]

Chaumontet, C., Suschetet, M., Honikman-Leban, E., Krutovskikh, V. A., Berges, R., Le Bon, A. M., Heberden, C., Shahin, M. M., Yamasaki, H., and Martel, P. (1996). Lack of tumor-promoting effects of flavonoids: Studies on rat liver preneoplastic foci and on in vivo and in vitro gap junctional intercellular communication. Nutr. Cancer 26, 251–263.[ISI][Medline]

Diwan, B. A., Rice, J. M., Ohshima, M., and Ward, J. M. (1986). Interstrain differences in susceptibility to liver carcinogenesis initiated by N-nitrosodiethylamine and its promotion by phenobarbital in C57BL6/6NCr, C3H/HeNCRMTV–, and DBA/2NCr mice. Carcinogenesis 7, 215–220.[Abstract]

Gad, S., and Weill, C. S. (1986). Statistics and Experimental Design for Toxicologists. Telford Press, Caldwell, NJ.

Guppy, M. J., Wilton, J. C., Sharma, R., Coleman, R., and Chipman, J. K. (1994). Modulation of phenobarbitone-induced loss of gap junctional intercellular communication in hepatocyte couplets. Carcinogenesis 15,1917–1921.[Abstract]

Ito, S., Tsuda, M., Yoshitake, A., Yanai, T., and Masegi, T. (1998). Effect of phenobarbital on hepatic gap junctional intercellular communication in rats. Toxicol. Pathol. 26, 253–259.[ISI][Medline]

Klaunig, J. E., and Ruch, R. J. (1987). Strain and species effects on the inhibition of hepatocyte intercellular communication by liver tumor promoters. Cancer Lett. 36, 161–168.[ISI][Medline]

Klaunig, J. E., and Ruch, R. J. (1990). Role of intercellular communication in nongenotoxic carcinogenesis. Lab. Invest. 62, 135–146.[ISI][Medline]

Kolaja, K. L., Engelken, D. T., and Klaassen, C. D. (2000). Inhibition of gap-junctional-intercellular communication in intact rat liver by nongenotoxic hepatocarcinogens. Toxicology 146, 15–22.[ISI][Medline]

Krutovskikh, V. A., Mesnil, M., Mazzoleni, G., and Yamasaki, H. (1995). Inhibition of rat liver gap junction intercellular communication by tumor-promoting agents in vivo. Lab. Invest. 72, 571–577.[ISI][Medline]

MacPhee, D. G. (1998). Epigenetics and epimutagens: Some new perspectives on cancer, germ line effects, and endocrine disrupters. Mutat. Res. 400, 369–379.[ISI][Medline]

Miyazaki, M., Bai, L., and Sato, J. (1990). Influence of liver tumor promoters and structurally related chemicals on survival of normal adult rat hepatocytes in primary culture. J. Cancer Res. Clin. Oncol. 116, 259–263.[ISI][Medline]

Moslen, M. T. (1996). Toxic responses of the liver. In Cassarett and Doull’s Toxicology: The Basic Science of Poisons. (C. D. Klaassen, Ed.), pp. 403–416. McGraw-Hill, New York.

Neveu, M. J., Babcock, K. L., Hertzberg, E. L., Paul, D. L., Nicholson, B. J., and Pitot, H. C. (1994). Colocalized alterations in connexin32 and cytochrome P450IIB1/2 by phenobarbital and related liver tumor promoters. Cancer Res. 54, 3145–3152.[Abstract]

Omiecinski, C., Wals, F., and Vlasuk, G. (1985). Phenobarbital induction of rat liver cytochromes P-450b and P-450e. J. Biol. Chem. 260, 3247–3250.[Abstract]

Pugh, G., Jr., Isenberg, J. S., Kamendulis, L. M., Ackley, D. C., Clare, L. J., Brown, R., Lington, A. W., Smith, J. H., and Klaunig, J. E. (2000). Effects of di-isononyl phthalate, di-2-ethylhexyl phthalate, and clofibrate in cynomolgus monkeys. Toxicol. Sci. 56, 181–188.[Abstract/Free Full Text]

Ren, P., Mehta, P. P., and Ruch, R. J. (1998). Inhibition of gap junctional intercellular communication by tumor promoters in connexin43 and connexin32-expressing liver cells: Cell specificity and role of protein kinase C. Carcinogenesis 19, 169–175.[Abstract]

Ren, P., and Ruch, R. J. (1996). Inhibition of gap junctional intercellular communication by barbiturates in long-term primary cultured rat hepatocytes is correlated with liver tumor promoting activity. Carcinogenesis 17, 2119–2124.[Abstract]

Rice, J. M., Diwan, B. A., Hu, H., Ward, J. M., Nims, R. W., and Lubet, R. A. (1994). Enhancement of hepatocarcinogenesis and induction of specific cytochrome P450-dependent monooxygenase activities by the barbiturates allobarbital, aprobarbital, pentobarbital, secobarbital and 5-phenyl- and 5-ethylbarbituric acids. Carcinogenesis 15, 395–402.[Abstract]

Rosenkranz, H. S., Pollack, N., and Cunningham, A. R. (2000). Exploring the relationship between the inhibition of gap junctional intercellular communication and other biological phenomena. Carcinogenesis 21, 1007–1011.[Abstract/Free Full Text]

Ruch, R. J. (2000). Role of gap junctions in cellular growth control and neoplasia: Evidence and mechanisms. Curr. Top. Membr. 49, 535–554.[ISI]

Ruch, R. J., Bonney, W. A., Sigler, K., Guan, X., Matesic, D., Schafer, L. D., Dupont, E. and Trosko, J. E. (1994). Loss of gap junctions from DDT-treated rat liver epithelial cells. Carcinogenesis 15, 301–306.[Abstract]

Ruch, R. J., and Klaunig, J. E. (1986). Antioxidant prevention of tumor promoter induced inhibition of mouse hapatocyte intercellular communication. Cancer Lett. 33, 137–150.[ISI][Medline]

Ruch, R. J., and Klaunig, J. E. (1988). Kinetics of phenobarbital inhibition of intercellular communication in mouse hepatocytes. Cancer Res. 48, 2519–2523.[Abstract]

Sai, K., Kanno, J., Hasegawa, R., Trosko, J. E., and Inoue, T. (2000). Prevention of the down-regulation of gap junctional intercellular communication by green tea in the liver of mice fed pentachlorophenol. Carcinogenesis 21, 1671–1676.[Abstract/Free Full Text]

Smith, J. H., Isenberg, J. S., Pugh, G., Jr., Kamendulis, L. M., Ackley, D., Lington, A. W., and Klaunig, J. E. (2000). Comparative in vivo hepatic effects of di-isononyl phthalate (DINP) and related C7–C11 dialkyl phthalates on gap junctional intercellular communication (GJIC), peroxisomal ß-oxidation (PBOX), and DNA synthesis in rat and mouse liver. Toxicol. Sci. 54, 312–321.[Abstract/Free Full Text]

Stevenson, D. E., Kehrer, J. P., Kolaja, K. L., Walborg, E. F., Jr., and Klaunig, J. E. (1995). Effect of dietary antioxidants on dieldrin-induced hepatotoxicity in mice. Toxicol. Lett. 75, 177–183.[ISI][Medline]

Trosko, J. E., and Chang, C. C. (1988). Nongenotoxic mechanisms in carcinogenensis: Role of inhibited intercellular communication. In Banbury Report 31: Carcinogen Risk Assessment: New Directions in Qualitative and Quantitative Aspects. (R.W. Hart and F.D. Hoerger, Eds.), pp. 139–170. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.

Trosko, J. E., Chang, C. C., Upham, B., and Wilson, M. (1998). Epigenetic toxicology as toxicant-induced changes in intracellular signaling leading to altered gap junctional intercellular communication. Toxicol. Lett. 102–103, 71–78.

Upham, B. L., Weis, L. M., and Trosko, J. E. (1998). Modulated gap junctional intercellular communication as a biomarker of PAH epigenetic toxicity: Structure-function relationship. Environ. Health Perspect. 106(Suppl. 4), 975–981.[ISI][Medline]

Venditti, P., Daniele, C. M., De Leo, T., and Di Meo, S. (1998). Effect of phenobarbital treatment on characteristics determining susceptibility to oxidants of homogenates, mitochondria, and microsomes from rat liver. Cell Physiol. Biochem. 8, 328–338.[ISI][Medline]

Weghorst, C. M., Pereira, M. A., and Klaunig, J. E. (1989). Strain differences in hepatic tumor promotion by phenobarbital in diethylnitrosamine- and dimethylnitrosamine-initiated infant male mice. Carcinogenesis 10, 1409–1412.[Abstract]