Inhibition of Gap Junctional Intercellular Communication by Noncoplanar Polychlorinated Biphenyls: Inhibitory Potencies and Screening for Potential Mode(s) of Action

Abstract

Polychlorinated biphenyls (PCBs), a structurally diverse group of environmental pollutants, are effective promoters in two-stage cancer models, which implies that epigenetic mechanisms are involved. Inhibition of gap junctional intercellular communication (GJIC) belongs among critical epigenetic events of tumor promotion. We determined the relative potencies of a series of environmentally relevant PCB congeners to inhibit GJIC in vitro in a rat liver epithelial cell line with pluripotent oval cell characteristics. The nonplanar PCBs were potent inhibitors of GJIC, whereas the coplanar PCBs did not inhibit GJIC. We then compared the effects of the coplanar PCB 126 (3,3′,4,4′,5-pentachlorobiphenyl) and the noncoplanar PCB 153 (2,2′,4,4′,5,5′-hexachlorobiphenyl) with effects of two model GJIC inhibitors, a tumor promoter 12-O-tetradecanoylphorbol-13-acetate (TPA) and epidermal growth factor (EGF). In contrast to TPA or EGF, PCB 153 elicited a long-term downregulation of GJIC (up to 48 h). Using Western blot analysis with phospho-specific antibodies, it was found that PCB 153, and not PCB 126, activated mitogen-activated protein kinases ERK1/2; however in contrast to TPA and EGF, this activation was observed at the time points subsequent to GJIC inhibition. Moreover, blocking of ERK1/2 activation did not prevent the GJIC inhibition induced by PCB 153. Therefore, additional intracellular signaling pathways potentially involved in the downregulation of GJIC by PCBs were screened by using specific chemical probes inhibiting serine/threonine kinases, tyrosine kinases, and phospholipases. The inhibition of diacylglycerol lipase partially blocked and the selective inhibition of Src kinases and phosphatidylcholine-specific phospholipase C (PC-PLC) completely blocked the inhibitory effects of the noncoplanar PCB on GJIC, indicating that PC-PLC or sphingomyelinase and Src might be upstream regulators of noncoplanar PCB-induced inhibition of GJIC.

Article

Polychlorinated biphenyls (PCBs) are a group of structurally diverse and persistent environmental pollutants, widely distributed as complex mixtures. Non-ortho-substituted coplanar PCBs have been shown to elicit a set of adverse effects associated with the activation of the aryl hydrocarbon receptor (AhR), resulting in liver damage, thymus atrophy, skin lesions, a wasting syndrome, and tumor promotion (van den Berg et al., 1998). Di-ortho-substituted PCBs, which tend to be noncoplanar structures of the biphenyl molecule that does not significantly activate AhR, exhibit a different spectrum of toxic modes of action, such as modulation of steroid hormone and Ca2+-induced intracellular signaling, and have been linked to neurotoxicity, immunotoxicity, endocrine disruption, and tumor promotion (Brouwer et al., 1999; Dean et al., 2002; Hansen, 1998; Robertson and Hansen, 2001). Such toxic effects suggest interactions with different cellular components other than AhR; however, exact modes of action of noncoplanar PCBs still remain unclear. Mechanistic studies are extremely important because of an urgent need to estimate contributions of noncoplanar congeners to overall PCB toxicity (Hansen, 1998). In the currently accepted toxic equivalency factors approach, which is based exclusively on AhR-mediated toxicity data, the more prevalent di-ortho-substituted PCBs have zero toxic potencies (van den Berg et al., 1998), despite their known toxicity.

Widespread distribution and persistence of PCBs led to studies on their potential role in carcinogenesis (Silberhorn et al., 1990). Evidence that exposure to PCBs, as well as polybrominated biphenyls (PBBs), leads to mutations has been ambiguous and controversial (Stapleton et al., 2001; Tsushimoto et al., 1983), but numerous studies indicate that PCBs and PBBs act as tumor promoters (Jensen et al., 1982; Safe, 1989; Silberhorn et al., 1990; van der Plas et al., 2000). These results imply that nongenotoxic (epigenetic) mechanisms are involved in their carcinogenic potential. Although coplanar PCBs are supposed to play an important role in tumor-promoting effects of PCB mixtures, such as Aroclor 1260, through AhR activation, the noncoplanar fraction of this mixture also contributed significantly to its tumor-promotion potential (van der Plas et al., 2000). Therefore, understanding the underlying mechanisms of the nongenotoxic tumorigenic potential of the noncoplanar PCBs is essential in establishing a more accurate assessment of risk that environmentally relevant PCBs pose to human health.

Three critical epigenetic events are needed at the promotional stages of cancer: one being the removal of an initiated cell from the suppression of growth by neighboring cells through the intercellular transfer of signal transductants via gap junctions; the second being the activation of a mitogenic pathway such as the extracellular receptor kinase (ERK) class of mitogen-activated protein kinases (MAPK); and the third is the inhibition of apoptosis. The downregulation of gap junctional intercellular communication (GJIC) by tumor-promoting compounds is considered to be a critical step in the removal of an initiated cell from the growth suppression of neighboring cells (Ruch and Trosko, 2001; Yamasaki, 1996). Therefore, the inhibition of GJIC can be assumed to be a representative marker of tumor-promoting potency for a given compound (Rosenkrantz et al., 2000). PBBs were the first polyhalogenated biphenyls shown to inhibit GJIC (Trosko et al., 1981), and a subsequent study demonstrated that the chlorinated isomers of the biphenyls inhibited GJIC in a manner similar to their brominated analogues (Tsushimoto et al., 1983). Numerous studies have since demonstrated that several PCBs can inhibit GJIC, both in vivo (Bager et al., 1997; Dean et al., 2002; Haag-Grönlund et al., 1998; van der Plas et al., 2000) and in vitro in rat liver epithelial cells, mouse and rat hepatocytes, human keratinocytes, and normal human breast epithelial cells (Hemming et al., 1991; Kang et al., 1996; Ruch and Klaunig, 1986; Swierenga et al., 1990). Although relative potencies of a number of PCBs to inhibit GJIC have been reported (Hemming et al., 1991), there are still many PCB congeners occurring in the environment, where this important information is missing.

The mechanism by which PCBs inhibit GJIC is not known. In general, the intracellular signals involved in acute inhibition of GJIC are still poorly understood; only the mechanisms of prototypical GJIC inhibitors, such as 12-O-tetradecanoylphorbol-13-acetate (TPA) or epidermal growth factor (EGF), have been studied in some detail (Lau et al., 1992; Matesic et al., 1994; Rivedal and Opsahl, 2001; Warn-Cramer et al., 1998). EGF-induced inhibition of GJIC has been associated with the activation of the membrane-bound EGF receptor (EGFR)–Ras–Raf–MEK–ERK1/2 signal transduction cascade, and TPA-induced inhibition of GJIC has been associated with the activation of protein kinase C (PKC) followed by the consequent activation of ERK1/2. Pretreatment of cells with specific MAPK or PKC inhibitors has been demonstrated to prevent EGF or TPA-induced inhibition of GJIC (Husoy et al., 2001; Rivedal and Opsahl, 2001; Ruch et al., 2001). While the inhibition of GJIC by either TPA or EGF are MEK-dependent, the activation of MEK and ERK alone is insufficient for the inhibition of GJIC (Hossain et al., 1999; Rummel et al., 1999). Furthermore, not all compounds, such as polycyclic aromatic hydrocarbons, inhibit GJIC through MEK-dependent pathways, even though these compounds can activate MAPK (Rummel et al., 1999). Besides protein kinase activation, phospholipase C (PLC) (Upham et al., 2003) and phospholipase A2 (PLA2) (ref. Wang and Loch-Caruso, 2002) have been implicated in the mechanism of GJIC inhibition. Involvement of these signaling pathways in PCB-induced inhibition of GJIC has yet to be determined.

The aims of the present study were to determine relative potencies of a series of environmentally relevant PCB congeners to inhibit GJIC in vitro in a rat liver epithelial cell line with pluripotent oval cell characteristics. We then compared the effects of the coplanar PCB 126 (3,3′,4,4′,5-pentachlorobiphenyl) and the noncoplanar PCB 153 (2,2′,4,4′,5,5′-hexachlorobiphenyl) on the phosphorylation of ERK1/2. Finally, we used specific chemical probes that inhibit protein kinases and phospholipases to screen the intracellular signaling pathways potentially involved in the downregulation of GJIC by noncoplanar PCBs and compared these results with effects of prototypical GJIC inhibitors.

MATERIALS AND METHODS

Chemicals.

PCB congeners were purchased from Promochem, with exception of PCB18, which was a generous gift from Dr. Hansen (University of Illinois, Urbana, IL), and PCBs 70 and 74 obtained from Dr. Lehmler (University of Iowa, Iowa City, IA). Chemical inhibitors U0126, H-89, LY294002, AG 825, AG 879, D609, ET-18-0CH3, MJ-33, and MAFP were from Calbiochem (San Diego, CA); SB203,580, PP2, and AACOCF3 from Alexis (Carlsbad, CA); forskolin, RHC 80267, and BEL (bromoethanol lactone) from Biomol (Hamburg, Germany). TPA, EGF, lucifer yellow, dimethylsulfoxide (DMSO), formaldehyde, GF109203X (bis-indoleylmaleimide), and other chemical modulators of cell signaling were supplied by Sigma-Aldrich (Prague, Czech Republic). All chemicals used in the study were of the highest available purity.

Cell culture and treatment with chemicals.

WB-F344 rat liver epithelial cells (Tsao et al., 1984) were cultured in modified Eagle’s Minimum Essential Medium (Sigma-Aldrich, Prague, Czech Republic), supplemented with pyruvate (110 mg/l), 10 mM HEPES, and 5% fetal bovine serum (Sigma-Aldrich, Prague, Czech Republic). Confluent cells, grown in 24-well plates, were exposed to various concentrations of individual PCB congeners (up to 100 μM) or solvent (DMSO, 0.1%, v/v) for 30 min. For the study of acute and sustained downregulation of GJIC, prolonged 24- or 48-h exposure to TPA, coplanar PCB 126, or noncoplanar di-ortho-chlorinated PCB 153 were used.

Pretreatment with modulators of intracellular signal transduction pathways.

The model tumor promoter, TPA, and endogenous peptide, EGF, were used as prototypical inhibitors of GJIC. Modes of action of GJIC inhibitors were studied in the cells pretreated for 20 min (if not indicated otherwise) with selective inhibitors or activators of intracellular protein kinases or phospholipases followed by the 30-min exposure to a compound under study (PCB 153, TPA, or EGF). Tyrosine kinases were inhibited by 30-min pretreatment with 50 μM genistein, 10 μM AG1478, 50 μM AG829, 5 μM AG879, or 50 μM PP2, selective inhibitors of EGF receptor (EGFR), Neu/ErbB-2, nerve growth factor receptor, and Src kinases, respectively. Inhibition of MEK1/2, p38 MAPK, and phosphoinositol triphosphate kinase (PI3 K) were performed by a 30-minute preincubation with 20 μM U0126, 10 μM SB203,580, and 25 μM LY294002, respectively. Activation of PKCs was blocked by 30-min pretreatment with 5 μM GF109203X, a general inhibitor of PKCs, or by a 24-h pretreatment with 20 nM TPA, inducing PKC depletion. Protein kinase A (PKA) was inhibited by treating the WB cells with 10 μM H-89 and activated by 5 mM 8-bromo-cAMP or 40 μM forskolin (30-min pretreatment). Several PLA2 and PLC inhibitors were used to test the involvement of phospholipase(s) in inhibition of GJIC after exposure to PCB153. Methyl arachidonyl fluorophosphonate (MAFP) and arachidonyl trifluoromethylketone (AACOCF3), inhibitors of IV4A-PLA2 and VIA-PLA2, were used in 1.25 and 10 μM concentrations, respectively; 2.5 μM bromoenol lactone (BEL), a specific inhibitor of VI-PLA2, 10 μM MJ33 inhibiting preferentially acidic iPLA2, 5 μM p-bromophenacyl bromide (BrPhBr), which selectively inhibits secretory PLA2 isoenzymes, 2.5 μM U73122 and 40 μM ET-18-OCH3, two inhibitors of phosphatidylinositol-specific PLC (PI-PLC), 40 μM xanthogenate tricyclodecan-9-yl (D609), which blocks phosphatidylcholine-specific PLC (PC-PLC) and sphingomyelinase (SMase) activities, and 20 μM RHC 80267, an inhibitor of diacylglycerol lipase, were used to selectively inhibit major known intracellular protein kinase and phospholipase signal transduction segments.

Inhibitor concentrations were selected according to previously published data on the use of specific inhibitors in various in vitro systems, including rat liver oval cells. Several inhibitors, namely PP2, genistein, and H-89, were tested at least at three different concentrations. The chemical inhibitors themselves did not affect GJIC, with exception of AG879, U73122, BrPhBr, BEL, MJ33, and MAFP, which reduced GJIC at the lowest effective concentrations 15, 5, 10, 5, 20, and 5 μM, respectively.

GJIC inhibition assay.

After exposure, a modified protocol of scrape-loading/dye transfer technique (Bláha et al., 2002; El Fouly et al., 1987) was used to assess in vitro modulations of GJIC. The cells were washed twice by phosphate-buffered saline solution (PBS), fluorescent dye was added (lucifer yellow, 0.05% w/v in PBS), and the cells were scraped using a surgical steel blade. After 2 min of the dye diffusion between the adjacent cells via gap junctions, the cells were washed by PBS and fixed with 4% (v/v) formaldehyde. The ratio of the gap-junctional dye transfer from the scrape line was measured with an epifluorescence microscope (Nikon Inc., Japan). At least three independent experiments were carried out in duplicates; at least three scrapes per well were evaluated. Cytotoxicity was measured by a conventional neutral red release assay (Balls et al., 1991). No apparent toxic effects of PCBs were observed within concentrations and exposure periods under study unless stated otherwise in the text.

Western blotting.

Cells were grown to the same confluency as for the SL/DT assay and then deprived of serum for 18–24 h to synchronize the cells and to reduce the background levels of ERK activity. PCB 153 inhibited GJIC at the same dose and time as in serum sufficient cells. Extraction and SDS–PAGE separation of proteins was performed according to the method of Rummel et al.(1999). The protein concentration was determined with BioRad DC protein kit (BioRad, Hercules, CA), and 15 μg of protein was loaded for each sample. The equal loading was verified by staining the blots with Ponceau S. Phosphorylated ERK 1 and ERK 2 were detected with a 1:2000 dilution of anti-phospho-ERK polyclonal antibodies for 24 h (New England Biolabs, Beverly, MA). The protein-primary antibody complex was probed with a 1:1000 dilution of HRP-conjugated anti-rabbit antibodies (Amersham Life Science Products, Arlington Heights, IL) for 1 h. The ERK protein bands were detected using the Super Signal chemiluminesence detection kit (Pierce Corp., Arlington Heights, IL) and ECL Hyperfine X-ray film.

Statistical data analysis.

The ratio of GJIC inhibition related to the negative control was evaluated and expressed in % (fraction of control, FOC). Nonparametric statistical methods were used for the data analyses. Kruskal-Wallis ANOVA followed by the Mann-Whitney test were used for the assessment of significance, and p values of less than 0.05 were considered statistically significant. Inhibition potency of a xenobiotic was expressed as a concentration causing 50% inhibition of GJIC (IC50); the IC50 values were determined from individual experiments by logit regression; relative error of estimate did not exceed 15%.

RESULTS

Acute Inhibition of GJIC by PCBs

The scrape load–dye transfer assay was used to measure GJIC in the WB-F344 rat epithelial cells, and the inhibitory effect of the noncoplanar PCB 153 and the lack of an inhibitory effect by the coplanar PCB 126 are shown in the fluorescent micrographs of Figure 1A. Inhibition of GJIC by PCB153 occurred within 15 min (Fig. 1B). The acute GJIC inhibitory potencies were determined for a series of 37 total environmentally occurring PCB congeners (Table 1). The IC50 values of the inhibiting compounds ranged within one order of magnitude (mostly between 10 and 25 μM), the result corresponding to previous findings with PCBs (Hemmings et al., 1991) or other organic environmental pollutants, e.g., DDT (Ren et al., 1998; Ruch et al., 1994; Wärngard et al., 1989), lindane (Leibold and Schwarz, 1993) or polycyclic aromatic hydrocarbons (Bláha et al., 2002; Upham et al., 1998). The noncoplanar tri- to hexachlorobiphenyls with chlorine substitutions at the ortho-position, such as PCB 153, were found to be potent inhibitors of GJIC. The most potent di-ortho-substituted PCB 47 elicited inhibiton of GJIC with IC50 being 10.1 μM. Also mono-ortho-chlorinated congeners showed similar potency to inhibit GJIC in the rat liver epithelial cells. High-molecular-weight hepta- and octachlorinated congeners and the non-ortho substituted PCBs elicited minimal or zero inhibition in the WB-F344 cells (Table 1, Figs. 1 and 2).

Inhibition Effects of PCB153 and PCB126 after Prolonged Exposure

We next investigated inhibition of GJIC by di-ortho-substituted noncoplanar PCB 153, coplanar PCB 126, or TPA during prolonged 24-h or 48-h exposure (Fig. 3). The transient inhibitory effect of the TPA and the recovery of GJIC after the prolonged exposure was observed as previously reported (Ren et al., 1998; Rivedal and Opsahl, 2001). On the other hand, PCB 153 induced a long-term downregulation of gap junction function between adjacent cells after 24 and 48 h within the same concentration range as was found after the acute 30-min treatment. Coplanar PCB 126 had no significant inhibitory effect on GJIC after 30 min, 24 h, or 48 h. Cytotoxicity was observed only after prolonged exposure to PCB 126 at the highest concentration (100 μM).

Coplanar versus Noncoplanar PCB Effects on MAPK

Two prototypical PCB congeners, Nos. 126 and 153, were selected for the study. The coplanar PCB 126 did not alter the phosphorylation of ERK1/2 as compared to the DMSO vehicle (Fig. 4), although slight differences in the phosphorylation status of ERK proteins were observed even with the vehicle control, probably due to high sensitivity of WB-F344 cells to manipulation (unpublished data). In contrast, the noncoplanar PCB 153 significantly activated ERK1/2 at 60 and 80 min (Fig. 4). Activation occurred after inhibition of GJIC, which is similar to the effects of polycyclic aromatic hydrocarbons (Rummel et al., 1999). In contrast to TPA, no apparent hyperphosphorylation of connexin43, which is a principal constituent of gap junction channels in WB-F344 cells, was observedafter PCB 153 treatment (data not shown). These and the above results indicate that inhibition of GJIC by PCB153 is independent of MAPK and differs from the reported mechanisms of TPA and EGF action (Hossain et al., 1999; Rivedal and Opsahl, 2001; Warn-Cramer et al., 1998).

Screening Mode(s) of Action of Inhibitors of GJIC

A combination of pretreatment with several selective inhibitors or other modulators of protein kinases and phospholipases, and exposure to EGF, TPA, or PCB 153 followed by the determination of GJIC was used to compare possible mode(s) of action of noncoplanar PCB with those of TPA and EGF (see Tables 2 and 3). The major components of intracellular signal transduction pathways involved in acute inhibition of prototypical compounds TPA or EGF have been previously reported. EGF blocks GJIC by the direct activation of EGF receptor and consequent activation of MEK1/2–ERK1/2 signaling pathway; TPA inhibits GJIC via the activation of PKC and by indirect activation of MEK1/2–ERK1/2 (Rivedal and Opsahl, 2001; Ruch et al., 2001). Our study confirmed these findings; the pretreatment of the cells with U0126, a specific inhibitor of MEK1/2, completely prevented the effect of both reference compounds. In addition, the pretreatment of the cells with AG1478, a selective inhibitor of membrane EGF receptor, prevented the inhibitory effect of the EGF on GJIC, while inhibition of PKC by GF109203X, as well as PKC depletion by prolonged 24-h pretreatment with TPA, prevented the downregulation of GJIC after TPA exposure (Table 2). However, neither U0126 nor GF109203X blocked PCB 153-induced inhibition of GJIC (Table 2). Therefore, effects of inhibitors/activators of several additional protein kinases were investigated. These included receptor tyrosine kinase ErbB-2 (reported to be a target of organochlorine pesticides in some cellular models; see Hatakeyama and Matsumura, 1999; Tessier and Matsumura, 2001), nerve growth factor receptor, src tyrosine kinase, phosphoinositol 3-kinase (PI3-K), p38 mitogen-activated protein kinase, or PKA.

Genistein, a nonspecific inhibitor of protein tyrosine kinases, significantly, although not completely, blocked the inhibition of GJIC (Fig. 5). None of other chemical inhibitors of protein kinases was able to block inhibition of GJIC after PCB 153 treatment with the exception of PP2, a selective inhibitor of Src, and H-89, a compound reported to be a potent inhibitor of PKA (Table 2). However, pretreatment of the WB-F344 cells with specific PKA activators, forskolin or 8-Br-cAMP, did not cause inhibition of GJIC; therefore, a possible role for PKA in signal transduction leading to downregulation of GJIC was not confirmed. These data seem to suggest that PCB congeners operate by a different mechanism than the prototypical GJIC inhibitors TPA and EGF. Neither the growth factor receptor tyrosine kinases under study (ERK1/2, p38, PI3-K, PKA) nor GF106203X-sensitive PKCs are probably involved in downregulation of GJIC observed after PCB 153 treatment in the WB-F344 cells.

PCBs or other organochlorines have been previously shown to activate PLA2 or PLC in rat neutrophils and several other cell types (Shin et al., 2002; Tithof et al., 1997; Wang and Loch-Caruso, 2002). Therefore, in this study a series of selective inhibitors of PLA2 isoenzymes, PI-PLC, and PC-PLC were investigated for potential effects on downregulation of GJIC by PCB 153 or TPA (Table 3). PLA2 inhibitors did not prevent but rather caused a slight increase in TPA- and PCB 153-induced inhibition of GJIC. Similarly, two PI-PLC inhibitors, U73122 and ET-18-OCH3, had no effect on GJIC inhibition. On the other hand, an inhibitor of PC-PLC, sphingomyelin synthase and SMase, D-609, prevented inhibition of GJIC. RHC 80267, a specific inhibitor of DAG lipase, an enzyme operating downstream of PLC isozymes, partially blocked the inhibition of GJIC by PCB 153 (Table 3). Interestingly, these compounds also blocked TPA inhibition of GJIC when concentrations of TPA was lower than 10 nM (Table 2).

DISCUSSION

Imbalance in tissue homeostasis due to disruption of cell-to-cell communication has been linked to growth and developmental diseases, such as cancer. Although there are many critical molecular events involved in maintaining homeostasis, considerable data accumulated over the last two decades indicate that intercellular communication through gap junctions plays an important role (Ruch and Trosko, 2001). Numerous studies have linked the interruption of GJIC, by either oncogenes or tumor promoters, with cancer, indicating that GJIC might be an important, albeit insufficient step of tumor promotion (Ruch et al., 2001). In addition to the removal of an initiated cell from growth suppression, mitogenic signal transduction pathways are also required for tumorigenesis. Therefore it is not surprising that PCBs were also shown to induce MAP kinase signal transduction pathways in our cell system, as well as activation of transcription factors such as AhR, NFkB, AP-1, or oxidative stress in other cell systems (reviewed in Glaubert et al., 2001), although their modes of mitogenic action are not sufficiently characterized. Both coplanar and noncoplanar PCB congeners have been shown to elicit hepatic tumor-promoting effects in rats (Buchmann et al., 1986; Dean et al., 2002; Glaubert et al., 2001), however at least one ortho-chlorine appeared to be essential for eliciting the inhibition of GJIC in rat liver epithelial cells (Hemming et al., 1991). Mono-ortho-substituted PCB114 and also PCB153 inhibited GJIC in human liver cells and keratinocytes, while coplanar PCB77 elicited no inhibition of GJIC (Swierenga et al., 1990). On the other hand, coplanar PCBs inhibited GJIC in mouse hepatoma cell line Hepa-1, and the AhR might be involved in this effect (De Haan et al., 1994).

Apparent relationships between the structure and the acute inhibition of GJIC by PCB congeners were also found in the present study. The noncoplanar tri- to hexachlorobiphenyls with chlorine substitutions at the ortho-position, such as PCB153, were found to be potent inhibitors of GJIC, and the mono-ortho-chlorinated congeners showed similar potency to inhibit GJIC in the rat liver epithelial cells (Table 1). Importantly, the inhibition of GJIC by prototypical noncoplanar PCB153 was not transient and lasted as long as 48 h (Fig. 2). The short time period of 15 min required for inhibition of GJIC indicated that posttranslational modification of the gap junctions was the most probable effect. On the other hand, high-molecular-weight hepta- and octachlorinated congeners elicited minimal or no inhibition. The non-ortho-substituted PCBs, which are potent AhR inducers, such as PCB126, had no inhibitory effect on GJIC in the WB-F344 cells, even after prolonged exposure (Fig. 2).

These findings confirmed previously reported effects of several noncoplanar, mono-ortho-substituted and coplanar PCBs in the same rat liver epithelial cell line (Hemming et al., 1991). We have identified in the present study a number of environmentally relevant PCB congeners that are also potent GJIC inhibitors, including PCBs 18, 47, 74, 114, 149, 163, 180, and 187 as well as other PCB congeners Nos. 110, 119, 123, 129, 157 occurring in the environment at significant concentrations (Rose et al., 2002; Hansen, 1998).

The difference between the effects of coplanar PCB congeners on the liver epithelial (oval-like) cells used in our system and hepatocyte-type cells (De Haan et al., 1994) might be related to the fact that the two cell types express different connexin proteins. While the oval cells express connexin 43, mature hepatocytes express connexin 26 and 32 as their predominant gap junction protein (Zhang and Thorgeirsson, 1994). AhR ligands, such as 2,3,7,8- tetrachlorodibenzo-p-dioxin (TCDD), are known to decrease GJIC in isolated hepatocytes, and this effect is probably related to AhR-controlled downregulation of connexin32 mRNA (Baker et al., 1995; Herrmann et al., 2002). Cell context has been also suggested to play a decisive role in effects of potential GJIC inhibitors, based on the fact that cell-specific, rather than connexin-specific inhibition by TPA, p,p′-DDT, or phenobarbital has been observed (Ren et al., 1998). Thus, while noncoplanar PCBs could inhibit GJIC in oval stem-like liver epithelial cells (this study), hepatocytes might be a target of inhibitory action of coplanar PCBs (De Haan et al., 1994). An importance of the cellular context for inhibition of GJIC by noncoplanar and coplanar PCBs is supported by the fact that coplanar PCBs did not affect GJIC in the rat liver epithelial cells WB-F344, although inducible AhR is probably present in these cells. The presence of functional AhR and P450-dependent monooxygenase activities in rat liver epithelial cells has been questioned (Herrmann et al., 2002; Schrenk et al., 1991). However, it has been shown that WB-F344 cells contain dioxin-inducible EROD activity (Köhle et al., 1999), and we have found that both TCDD and PAHs, known as AhR ligands, are able to induce CYP1A-dependent activity in this cellular model, which was inhibited by a specific inhibitor, α-naphthoflavone (unpublished data). This seems to suggest a presence of functional AhR in this cellular model. However, a possible role of chaperon complexes of AhR and other transcription factors such as CAR, which is known to be activated by noncoplanar PCBs, has not been currently studied in immediate intracellular events (Pascussi et al., 2003).

Although the inhibitory effects of noncoplanar PCBs on GJIC have been reported from various in vitro and in vivo systems, the mechanism(s) of their action is still unknown. The set of selective chemical inhibitors of protein kinases and phospholipases was used to identify the regulatory mechanisms of intracellular signaling pathways responsible for inhibition of GJIC. Noncoplanar PCB 153, as well as model prototypical inhibitors of GJIC, TPA, and EGF were selected for this comparative study. The participation of the EGF receptor–Ras–Raf–MEK1/2–ERK ½ signal transduction pathway in inhibition of GJIC by EGF is well documented (Kanemitsu and Lau, 1993; Rivedal and Opsahl, 2001; Ruch et al., 2001; Warn-Cramer et al., 1998), and our data were consistent with this model. It is believed that TPA blocks GJIC mainly by a direct action of PKC, but also partly through cross-talk with the ERK1/2 pathway (Oh et al., 1991; Ren et al., 1998; Rivedal and Opsahl, 2001). In this study, the specific inhibitor of EGFR (AG1478) blocked inhibition of GJIC after treatment with EGF; the general inhibitor of conventional and novel PKCs, GF109203X, prevented the GJIC inhibition after exposure to TPA. The MEK inhibitor U0126 prevented inhibition of GJIC caused by both TPA and EGF (Table 2).

From our data, it seems evident that a prototypical noncoplanar PCB 153 operates by a mode of action independent of ERK1/2 or PKC activation. Although PCB 153 activated ERK1/2, the phosphorylated forms of ERK1/2 were detected subsequently to inhibition of GJIC. This sequence of events is similar to the effects of polycyclic aromatic hydrocarbons (Rummel et al., 1999). Moreover, a specific inhibitor of ERK1/2 activation, U0126, did not affect the inhibitory potency of PCB 153. No apparent hyperphosphorylation of connexin43, which is a principal constituent of gap junction channels in WB-F344 cells and becomes phosphorylated after EGF or TPA treatment, was induced by PCB 153 (data not shown). These results indicate that inhibition of GJIC by PCB 153 is independent of ERK1/2. Similarly, inhibitors of other protein kinases, including PI3K, p38, and the general inhibitor of PKC, did not modulate the inhibition of GJIC by PCB 153. In contrast, the inhibitory effect of PCB 153 was partially blocked by genistein, a nonspecific tyrosine kinase inhibitor, and a strong prevention was found also after pretreatment with a high concentration of PP2, an inhibitor of Src tyrosine kinase. This seems to suggest an involvement of tyrosine kinase in effects of PCB 153, which remains to be specified. However, pretreatment with specific inhibitors of selected receptor tyrosine kinases, EGFR, ErbB2/Neu, and NGFR, did not prevent GJIC inhibition by PCB 153 (Table 2).

H-89, a widely used inhibitor of intracellular PKA, blocked inhibition of GJIC after treatment with PCB 153. However in our study, the activation of PKA by forskolin or the brominated analog of cyclic adenosine monophosphate (8-Br-cAMP) was not associated with inhibition of GJIC (Table 2). Therefore, we can conclude either that PKA was not involved in inhibition of GJIC after treatment with PCB 153 and prevention by H-89 was probably due to a broader inhibition specificity of H-89, or that cAMP alone is insufficient to activate the appropriate combination of signaling pathways needed to inhibit GJIC. Specificities of many inhibitors of protein kinases and probably also phospholipases have not been tested sufficiently (Davies et al., 2000), and it is necessary to be careful when the data from inhibitory studies are interpreted. For example, H-89 has been reported to selectively antagonize beta adrenergic receptors and other G protein-coupled receptors (Penn et al., 1999). Notably, G protein-coupled receptor agonists, such as lysophosphatidic acid, rapidly disrupt GJIC by activation of Src tyrosine kinase (Giepmans et al., 2001; Postma et al., 1998) or Ras-Raf-MEK1/2-ERK1/2 pathway (Warn-Cramer et al., 1998).

Another important group of enzymes involved in intracellular signaling are phospholipases and lipases, such as PLA2 isozymes, PI-PLC, PC-PLC or diacylglycerol (DAG)-lipase (Nozawa, 2002). PCBs have been reported to activate PLA2 or PLC in various cell types (Shin et al., 2002; Tithof et al., 1997; Wang and Loch-Caruso, 2002). In the present study, the inhibitor of PC-PLC, D609, strongly blocked downregulation of GJIC, while a series of PLA2 and PI-PLC inhibitors failed to prevent inhibition of GJIC by PCB 153 (Table 3). In addition, DAG lipase inhibitor RHC 80267 also partially blocked GJIC inhibition induced by PCB 153, which suggests a potential role for DAG lipase products in the inhibitory mode(s) of action of noncoplanar PCB 153.

Unlike some bacterial analogues, the mammalian PC-PLC has not yet been characterized at the molecular and regulatory levels. Therefore, the exact role of PC-PLC in signal transduction remains to be defined, and all the published studies are based on the use of D609, its putative inhibitor (Nozawa, 2002). D609 is considered to be a specific inhibitor of PC-PLC activity, since it has been reported not to interfere with activities of PI-PLC, PLA2, or PLD (Amtmann, 1996). D609 has been reported to also inhibit sphingomyelin synthase activity, which modulates DAG/ceramide ratio with resulting significant effects on proliferation and apoptosis (Luberto and Hannun, 1998) and SMase activity. It is currently not known whether sphingomyelin synthase and PC-PLC activities are due to the same or different enzymes (Ohanian and Ohanian, 2001). Interestingly, a partial inhibition of GJIC by low TPA concentrations was also prevented by H-89, PP2, D609, or RHC 80267, suggesting that low doses of TPA might operate through a slightly different mechanism than solely by the previously reported activation of phorbol ester-dependent PKC isoenzymes (Rivedal and Opsahl, 2001; Ruch et al., 2001).

In conclusion, a series of environmentally relevant mono-ortho- and di-ortho-substituted PCBs caused the inhibition of GJIC in the rat liver epithelial (oval-like) cells; this cell line was not a direct target of disruption of intercellular communication by coplanar PCB congeners. In contrast to prototypical inhibitors of GJIC, TPA, and EGF, noncoplanar PCB153 elicited a potent inhibition of GJIC by a different mode of action, which probably does not involve activation of PKC or ERK1/2. Using selective chemical inhibitors of signal transduction pathways revealed that PC-PLC, Src, and DAG lipase might be involved in the GJIC inhibitory effect of PCBs.

Acknowledgments

This research was supported by the Czech Ministry of Agriculture (grant No. QC0194), the Grant Agency of the Czech Republic (No. 525/00/D101) to L. B., and the National Institute of Environmental Health Science Superfund (grant #P42 ES04911-07) to J.E.T.

Footnotes

REFERENCES