Indolo[3,2-b]carbazole inhibits gap junctional intercellular communication in rat primary hepatocytes and acts as a potential tumor promoter


Indole-3-carbinol (I3C) is a naturally occurring substance that shows anti-carcinogenic properties in animal models. Besides its clear anti-carcinogenic effects, some studies indicate that I3C may sometimes act as a tumor promoter. Indolo[3,2-b]carbazole (ICZ), which is formed in the acidic environment of the stomach after intake of I3C, has a similar structure to, and shares biological effects with, the well-known tumor promoter 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Therefore, we hypothesized that ICZ could be responsible for the potential tumor-promoting activity of I3C. The aim of the present study was to investigate the effect of ICZ on gap junctional intercellular communication (GJIC) in primary cultured rat hepatocytes co-cultured with the rat liver epithelial cell line WB-F344. Indolo[3,2-b]carbazole inhibited GJIC in the rat hepatocytes in a dose- and time-dependent manner. Significant inhibition was observed after 8 and 12 h of treatment with 1 and 0.1 μM ICZ, respectively. Maximum GJIC inhibition (cell–cell communication only 5% of control values) was observed after 24–48 h of ICZ treatment. Continued exposure to 1 μM ICZ suppressed GJIC until ~120 h. Both ICZ and TCDD treatment reduced the Cx32 mRNA level as well as the plasma membrane Cx32 staining. Indolo[3,2-b]carbazole increased the Cyp1a1, Cyp1a2 and Cyp1b1 mRNA levels concurrently with an increase in 7-ethoxyresorufin O-deethylase (EROD) activities. Maximum EROD activity and Cyp1a1 mRNA levels were observed after ~12 h, whereas Cyp1a2 and Cyp1b1 mRNA levels peaked after 48 h. This study shows that ICZ may possess tumor promoter activity down-regulating GJIC by mechanisms, which seem to include activation of the Ah receptor and/or Cyp1 activity. Further studies are needed in order to clarify the anticarcinogenic/carcinogenic effects of I3C and ICZ before high doses of I3C may be recommended as a dietary supplement.



Several studies have shown that a diet rich in fruits and vegetables (especially cruciferous vegetables) reduces the risk of various forms of cancer (1). Indole-3-carbinol (I3C), which is formed in relatively high concentration during preparation and ingestion of cruciferous vegetables, reduces the level of DNA damage (2) and the incidence of chemically induced liver tumors (3) in rainbow trout fed the compound before initiation. In rats treated with aflatoxin B1, I3C was shown to inhibit both the initiation and the promotion step of the carcinogenic process (4). I3C has also been shown to act as a tumor promoter. In rats, I3C was recently shown to increase the hepatic foci formation (5). When I3C was fed to rainbow trout (3) or rats (6) after treatment with the tumor initiator 1,2-dimethylhydrazine, the incidence of liver tumors was increased. Furthermore, I3C both induces and reduces ornithine decarboxylase activity (7,8), suggesting that I3C may have epigenetic effects. An overview of the reported anti-carcinogenic and carcinogenic properties of indoles is given elsewhere (9,10).

The various biological effects observed for I3C in vivo are most probably caused by the condensation products of I3C formed in the acidic environment of the stomach. One of these acid-catalyzed condensation products is indolo[3,2-b]carbazole (ICZ), which has several biological and structural characteristics in common with 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). Indolo[3,2-b]carbazole has marginally lower binding affinity for the aryl hydrocarbon receptor (AhR) than TCDD (11). Both substances induce cytochrome P-450 subfamily 1 (Cyp1) enzyme activity in vitro and in vivo (12) via activation of the AhR. Therefore, we hypothesized that the observed liver tumor-promoting activity of I3C could be caused by ICZ produced during digestion.

Gap junctional intercellular communication (GJIC) is generally lowered during the carcinogenic process, as documented by numerous investigations showing that connexin expression and GJIC is limited in tumors, and that tumor promoters inhibit GJIC. In vitro assays for GJIC show ~60% sensitivity for detecting known in vivo tumor promoters/epigenetic carcinogens (13), and many cell-growth inhibitors, cancer chemopreventive agents, anti-oncogenes and differentiation agents enhance GJIC (14). The inhibitory effect of TCDD on GJIC is dependent on the cell type. In WB-F344 cells or C3H/10T1/2 mouse fibroblasts, TCDD does not inhibit GJIC (15,16), whereas TCDD reduces GJIC in primary rat hepatocytes and mouse hepatoma cells (Hepa1c1c7) (17,18). It has therefore been proposed that the AhR activation is needed for TCDD to induce down-regulation of GJIC and that TCCD specifically causes a transcriptional down-regulation or reduced stability of the mRNA encoding for the gap junction protein connexin 32 (Cx32) (17). Cx32 is the major gap junction protein in primary hepatocytes. In contrast to hepatocytes, connexin 43 (Cx43) is the major gap junction protein in WB-F344 cells.

All experiments in the present study have been performed with co-cultures of primary rat hepatocytes and rat liver epithelial WB-F344 cells, as the literature indicate that the GJIC activity is more stable in the co-culture compared with the primary hepatocytes alone.

The aim of the present study was to investigate the effect of ICZ on GJIC in rat primary cultured hepatocytes and to investigate the possible role of AhR activation and/or Cyp1 enzyme activity levels in modulating the cell–cell communication.

Materials and methods

Cells and chemicals

WB-F344 rat liver epithelial cells were a gift from Prof. R.Ruch (Medical College of Ohio, OH). Indolo[3,2-b]carbazole was a gift from Prof. L.Bjeldanes (University of California, Berkeley, CA), and TCDD was a gift from Dr. M.van Iersel (TNO, Zeist, The Netherlands). Hank's balanced salt solution without Ca2+ and Mg2+ (HBSS), Dulbecco's minimum essential medium (DMEM), collagenase type II, hepatocyte attachment medium, fetal calf serum and gentamycin were purchased from Gibco BRL™ (TeckNunc, Roskilde, Denmark). Richter's improved minimum essential medium and modified Earle's salt solution were purchased from Irvine Scientific (Santa Ana, CA). Lucifer Yellow CH, dexamethasone, 7-ethoxyresorufin and resorufin were purchased from Sigma Chemical Co. (St Louis, MO). Standard TC Falcon culture dishes (35 and 100 mm dishes and 12 well plates) were purchased from Becton Dickinson (Broendby, Denmark).

Capillary tubes for microinjection (0.580–1.00 × 80 mm, with filament) were obtained from Modulohm A/S (Herlev, Denmark) and Eppendorf micro loaders from Radiometer Denmark (Roedovre, Denmark).

Normal goat serum, mouse monoclonal anti-Cx32 antibody and FITC-goat anti-mouse IgG antibody were purchased from Zymed (South San Francisco, CA) and mounting medium from Chemicon International (Temecula, CA).

Co-culture of primary hepatocytes with WB-F344 rat liver epithelial cells

Hepatocytes were isolated from male F344 rats (160–360 g; M&B A/S, Ry, Denmark) by a two-step collagenase perfusion method originally described by Gant et al. (19). In short, rats were anesthetized and the liver perfused through the portal vein with 0.5 l HBSS at a flow rate of 30 ml/min followed by 0.25 l DMEM containing 250 mg collagenase type II at a 25 ml/min flow rate. Liver cells were isolated and suspended in attachment medium. The suspension was filtered and washed with attachment medium, and viability determined using the trypan blue exclusion assay. Hepatocytes were plated at a density of 5×104 hepatocytes/cm2 on Falcon dishes with 30–40% confluent cultures of WB-F344 cells and incubated at 37°C (95% O2, 5% CO2, 100% humidity). After 2 h, attachment medium and dead/non-attached cells were removed and replaced with culture medium (Richter's MEM supplemented with 5% FBS, 1 mM dexamethasone and 50 mg/ml gentamycin). At day 1 of the experiment, the number of hepatocytes was 2.7 × 104 ± 0.2 × 104 hepatocytes/cm2 (mean ± SD). The number of living hepatocytes stayed constant for the duration of the experiment.

Treatment of cells

Indolo[3,2-b]carbazole and TCDD were dissolved in DMSO, and added to the culture medium. The final DMSO concentration did not exceeded 0.1%. Two types of experiments were performed: continual treatment studies, and recovery studies. In continual treatment studies, the co-cultures were treated with DMSO, ICZ (0.1, 1 or 10 μM) or TCDD (1 nM) for varying lengths of time. The test medium was changed every 24 h. At the denoted time-points the GJIC and EROD activity were estimated and RNA isolated. To study the kinetics of reappearance of GJIC (recovery studies), the co-cultures were exposed to vehicle, 1 μM ICZ or 1 nM TCDD. After 24 h of exposure, the test medium was replaced by culture medium without further additions. At the denoted time-points the GJIC and EROD activity were estimated and RNA isolated.

Microinjection dye transfer (MI/DT) assay

Hepatocytes were co-cultured with WB-F344 cells in 35 mm culture dishes and GJIC was detected by MI/DT as described by Ruch and Klaunig (20) with slight modifications. In short, micropipettes were pulled on a PC-10 Micropipette puller (Narishige Co., Tokyo, Japan). The micropipette was mounted vertically in the injection setup [Micromanipulator (5171) and Transjector (5246) was from Eppendorf]. Lucifer Yellow (0.1 M in a 0.33 M LiCl solution) was then injected into a hepatocyte and the percentage of hepatocytes in direct contact with the microinjected cell to which the dye was transferred in 5 min. The hepatocytes were observed through a 31010 Luc Yel 606 barrier filter, on a DMIRB/E Leica inverted microscope from Leica Microsystems A/S (Glostrup, Denmark). Ten microinjections were performed in each dish and the experiments were performed in triplicate. Values are given as the percentage of dye-coupled adjacent cells (percent communication).

RNA expression

Hepatocytes were co-cultured with WB-F344 cells in 10 cm culture dishes. At the denoted time-points, the medium was removed and 1 ml Trizol Reagent (Gibco BRL) was added, and the cell homogenate transferred to a sterile microfuge tube. Isolation of total RNA was performed as described by the manufacturer.

For northern blotting, 20 μg total RNA from each treatment and time-point were analyzed by 1% agarose/0.2 M formaldehyde gel electrophoresis on the same gel and transferred by downward capillary blotting onto a nylon membrane. The membrane was then hybridized (Rapid-hyb buffer from Amersham Pharmacia Biotech, Hoersholm, Denmark) with [α-32P]dCTP labeled cDNA probes (rediprime, Amersham Pharmacia Biotech). Radioactivity was detected by phosphor imaging (Storm, Molecular Dynamics, Sunnyvale, CA), and relative band intensities were analyzed by the ImageQuaNT 5.2 software (Molecular Dynamics, 1999). The membrane was stripped and reprobed using the specific cDNA fragments. To adjust for unequal loading, data are presented as the ratio of specific mRNA of interest to the 18 S rRNA content. The transcript sizes were estimated by including a 0.24–9.5 kb RNA ladder (Gibco BRL).

The cDNA templates used for probe synthesis were constructed using reverse transcribed rat liver total RNA and subsequent amplification by polymerase chain reaction (Advantage RT-for-PCR kit, Clontech Laboratories, Palo Alto, CA). Specific primers were designed from the known cDNA sequences published in Genbank of Cyp1a1 (accession number NM012540, nucleotides 623–942), Cyp1a2 (accession number NM012541, nucleotides 615–926), Cyp1b1 (accession number NM012940, nucleotides 1488–1934), Cx32 (accession number NM017251, nucleotides 595–1145) and Cx43 (accession number NM012567, nucleotides 1100–1631). All RT–PCR products were ligated into the pCRII vector (TA Cloning Kit Dual Promoter, Invitrogen) and sequenced (ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction Kit, Perkin Elmer) to confirm their identity.

EROD assay

Hepatocytes were cultured in 12 well plates. At the denoted time, the medium was removed, cells washed with PBS and enzyme assays initiated by addition of 1 ml of 5 μM 7-ethoxyresorufin in PBS. The cells were incubated with the substrate solution for 5–10 min at 37°C. The assay solution was removed and the amount of fluorescent resorufin formed during the incubation was determined using a Perking Elmer Luminescence Spectrometer (LS 50B), with a 530 nm excitation, and 585 nm emission wavelengths using a slit width of 10 nm. The amount of resorufin formed was calculated by comparison to a resorufin standard curve (0–500 nM). Results are given in pmol resorufin/min × cm2.

Cx32 immunocytochemical staining

Hepatocytes were cultured on sterile cover slips in 35 mm culture dishes. At the denoted time-points, cells were fixed with ice cold fixation solution (5% acetic acid in 95% methanol) for 15 min and stored at 4°C for no longer than 4 days before further processing. Cover slips were incubated with normal goat serum, diluted 1:100 in PBS, for 30 min at room temperature, followed by an overnight incubation at room temperature with a mouse anti-Cx32 antibody, diluted 1:100 in PBS. Subsequently, the cover slips were incubated with a FITC-goat anti-mouse IgG antibody, diluted 1:50 in PBS for 1 h at room temperature, and finally mounted on glass slides.


Results are expressed as means of three dishes ± SEM. Statistical differences on percentage data were tested using one-way ANOVA on Freeman-and Tukey-transformed data (21). Dunnett's post hoc test was used to test for differences to the controls at the specific time-point (P 0.05), using Systat version 10 (SSPS, Evanston, IL).


Effects of ICZ and TCDD on GJIC

No toxic effects of ICZ (0.1–10 μM) or 1 nM TCDD treatment were detected in the co-cultures (data not shown).

Indolo[3,2-b]carbazole inhibits GJIC in cultured rat hepatocytes co-cultured with the rat liver epithelial cell line WB-F344 dose- (0.1–10 μM, Figure 1) and time-dependently (Figure 2). One micromolar ICZ inhibited GJIC after 8 h of treatment, although not after 4 h of treatment. Four hours of treatment with 10 μM ICZ caused a significant GJIC inhibition, whereas a significant inhibition of GJIC with 0.1 μM ICZ could be observed after 12 h of treatment. Maximal inhibition of GJIC was induced after 24 h of treatment with 0.1 and 1 mM ICZ, corresponding to ~16 and 5%, respectively, of the GJIC in control cells. Continuous treatment with 1 μM ICZ sustained the GJIC at low levels (3–7% of the GJIC in control hepatocytes) until 120 h of treatment, after which the level of GJIC increased. GJIC was not completely restored after 192 h of treatment (Figure 2A). Continuous treatment with 0.1 μM ICZ only sustained GJIC inhibition for 24 h (data not shown). TCDD (1 nM) inhibited GJIC after 4 h of treatment (Figure 1) and low levels of GJIC (≤10% of GJIC in controls) were observed until 120 h (Figure 2A). After 192 h of treatment a slight increase in the GJIC level was also observed in the TCDD-treated hepatocytes. However, the effect of TCDD on GJIC was sustained longer than the effect of ICZ.

To identify the communication recovery rate after exposure to ICZ or TCDD, the co-cultures were exposed to 1 μM ICZ or 1 nM TCDD for 24 h and then replaced by media without the test compounds. In hepatocyte co-cultures pre-treated with ICZ, GJIC increased 12 h after replacement with fresh medium (Figure 2B). Already after 48 h of recovery, the level of GJIC was not significantly different from the level observed in DMSO controls.

When TCDD was removed from the hepatocyte culture medium, >48 h was required before an increase in GJIC was observed. In the TCDD-treated hepatocytes, GJIC remained lower than in DMSO controls for the duration of the experiment (P 0.001, Figure 2B).

Effects of ICZ and TCDD on Cx32 level

To elucidate the mechanism of GJIC down-regulation, Cx32 and Cx43 mRNAs were quantified. Cx32 was only expressed in hepatocytes whereas Cx43 was only expressed in WB-F344 cells and was not affected by ICZ or TCDD treatment (data not shown). A decrease in Cx32 mRNA levels was observed during the first 24 h of treatment, in both controls, ICZ- and TCDD-treated hepatocytes (Figure 3A). Beyond 24 h, both ICZ and TCDD caused a reduction in the Cx32 mRNA level compared with the control.

In the recovery study, the Cx32 mRNA level in ICZ-treated hepatocyte co-cultures slowly increased to the level of control hepatocytes whereas in TCDD-treated cells, the transcript levels did not return to the level of the controls within the duration of the experiment (Figure 3B).

Effects of ICZ and TCDD on Cx32 immunostaining

Both ICZ (1 μM) and TCDD (1 nM) decreased the Cx32 staining at the cell–cell interface (Figure 4). During the continuous treatment with ICZ, Cx32 reappears at the cell–cell interface at 192 h (Figure 4). Cx32 immunostaining was absent in hepatocyte co-cultures treated with TCDD from 24 h onwards (Figure 4).

Primary rat hepatocyte co-cultures treated with ICZ for 24 h and then allowed to recover in ICZ-free medium, showed a reappearance of Cx32 staining after 48h. After 120 and 192 h of recovery, Cx32 staining was at about the same level as observed in DMSO controls (Figure 5). In hepatocyte co-cultures allowed to recover after 24 h of TCDD treatment, staining was absent even after 120 h in TCDD-free medium. After 192 h of recovery, occasional immunopositive staining dots could be observed at the cell–cell interface (Figure 5).

Hepatocytes in control cultures showed Cx32 immunostaining at the cell–cell interface at all time-points.

Effects of ICZ and TCDD on EROD activity and Cyp1a1, Cyp1a2 and Cyp1b1 mRNA levels

To correlate the ICZ or TCDD-induced EROD activity to the observed GJIC inhibition, the EROD activity was measured both in the continuous exposure experiment and in the recovery experiment.

The EROD activity was determined after 4 and 8 h of treatment with 1 and 10 μM ICZ and 1 nM TCDD, and after 8, 12, 24, 48, 120 and 192 h treatment with 0.1 and 1 μM ICZ and 1 nM TCDD.

Both ICZ and TCDD induced EROD activity in rat hepatocyte co-cultures in a dose- (Figure 6) and time-dependent manner (Figure 7A). Induction was observed after 4 h of treatment with both 1 and 10 μM ICZ and 1 nM TCDD (Figure 6). Maximum EROD activity was observed in ICZ-treated cultures after 12–24 h of treatment. Continued treatment decreased EROD activity approaching the level found in controls after 120 h. The same pattern was observed for Cyp1 mRNAs peaking after 12–48 h of treatment and thereafter decreasing to about the half their maximum levels (Figure 7B).

When ICZ was removed from the co-cultures after 24 h of pre-treatment, the EROD activity decreased much faster (Figure 8A), than observed in co-cultures treated continuously with the compound (Figure 7A). After 120 h of recovery, EROD activity in the ICZ pre-treated cultures had decreased to the same level as in DMSO controls. Co-cultures pre-treated with TCDD did not fully recover within the 192 h time series. The ICZ-induced Cyp1A mRNA expression was reduced in the same way as the ICZ-induced EROD activity (Figure 8B).

The EROD activity and Cyp1 mRNA expression was identical for ICZ- and TCDD-treated cultures excluding the recovery studies, where EROD in cultures pre-treated with TCDD for 24 h did not fully reach the background activity even after 192 h recovery (Figure 8A). In WB-F344 rat liver epithelial cell cultured alone, the Cyp1b1 mRNA level was induced by ICZ, whereas the EROD activity and Cyp1a mRNA levels were not induced by ICZ or TCDD treatments (data not shown).


I3C has in various experiments been shown to inhibit tumor development (2224). Several possible mechanisms such as reduction of the levels of the activated genotoxic carcinogens by modulation of the metabolism (2527) and reduction in DNA-adduct levels (24) have been suggested. Beside these effects on the metabolism of endogenous and exogenous compounds, I3C has also been shown to (i) reduce tumor promoter induced ornithine decarboxylase activity (28), (ii) inhibit cellular growth (29) and (iii) induce apoptosis (30). I3C's anticarcinogenic properties may in part explain the reduced tumor incidence observed with large intakes of cruciferous vegetables, which support the general recommendation to increase the intake of fruit and vegetables in the human population. Furthermore, I3C is also available as a food supplement for humans in pure form and it is claimed by the commercial supplier that doses exceeding those normally found in cruciferous vegetables are needed to promote human health and to prevent cancer. However, the few reports showing a tumor-promoting effect of I3C must not be overlooked. A tumor-promoting effect of this compound has been observed in both rainbow trout (3) and rats (6) after treatment with aflatoxin B1 and 1,2-dimethylhydrazin, respectively. Several biomarkers related to tumor promotion are modulated by I3C, e.g. induction of ornithine decarboxylase activity (7). Various nitrosated and chlorinated indoles inhibit GJIC in Chinese hamster V79 cells (31,32), whereas indole-3-acetonitril, one of several degradation products of glucobrassicin, does not inhibit GJIC (31).

In the present study, we clearly show that ICZ down to 0.1 μM inhibits GJIC in rat hepatocytes. This indicates that ICZ may possess a tumor-promoting activity, and that the observed inhibitory effect of ICZ may explain some of the detrimental effects observed for I3C in vivo. Based on the structural and functional similarity between ICZ and the well-known tumor promoter TCDD (33), the tumor-promotive effect of ICZ is very likely. TCDD inhibits GJIC in primary cultured rat hepatocytes (17), mouse hepatoma cells (18), rat liver epithelial IAR 20 cells (34,35) and rat hippocampal primary cell culture (36).

An almost complete blockage of GJIC was observed after 24 h of treatment with either 1 μM ICZ or 1 nM TCDD, although it was a slower inhibition than observed for another CYP inducer, phenobarbital (2 h), in mouse hepatocytes (20). The more slowly occurring effects of TCDD and ICZ indicate that some other cellular events such as transcription of certain genes must take place before the ICZ-induced down-regulation of GJIC occurs. Furthermore, during prolonged exposure to ICZ and partly TCDD, the hepatocytes show decreased sensitivity to the inhibitory effect of these compounds indicating of a possible feedback mechanism. Development of refractoriness has been observed for phenobarbital as well, whereas continued exposure to lindane and DDT causes a sustained inhibition of GJIC (37). Also, the recovery of GJIC after removal of the test substance was slower for ICZ and TCDD than for phenobarbital, DDT and lindane (20,37,38). In the present study, the TCDD-induced GJIC down-regulation was sustained longer than the ICZ-induced GJIC down-regulation, which may be caused by a faster removal of ICZ than TCDD.

It has been suggested that TCDD only inhibits GJIC in cells expressing functional AhR (17,18). Beside a direct regulation by the activated AhR, it has also been shown that down-stream cellular effects of AhR activation are involved in inhibition of GJIC [e.g. Cyp inhibitors, antioxidants (39,40), cAMP (17) and protein kinase inhibitors (39)]. Cyp1a1, Cyp1a2 and Cyp1b1 genes are parts of the AhR gene battery. Therefore, the immediate increase in EROD activity, and in Cyp1a1, Cyp1a2 and Cyp1b1 mRNA levels, and down-regulation after removal of ICZ, which correlated well with the inhibition of GJIC, indicates that AhR activation and/or Cyp activities are involved in the observed inhibition of GJIC. Preliminary studies performed in our laboratory using the Cyp1a inhibitors (SKF-525A and α-naphthoflavone) indicate a direct role of the Cyp1a activity on the down-regulation of GJIC. In contrast to the good correlation between the immediate rise in EROD activity and inhibition of GJIC, no clear correlation was found during the later stages of declining EROD activity and increasing GJIC. In the present experiments, the WB-F344 cells alone had no induction of Cyp1a1 mRNA or EROD, indicating that these cells do not show Ah-receptor activity, as found by one other group (41).

The level of GJIC measured agrees well with the Cx32 immunoreactive staining at the cell–cell interface. Exposures to ICZ or TCDD lowered the Cx32 staining and GJIC, and the re-establishment of GJIC correlated well with the reappearance of Cx32 immunostaining at the cell–cell interface. Small dots of staining were observed in the area over the nucleus in hepatocytes with reduced or no Cx32 staining at the cell–cell interface. The Cx32 mRNA level was reduced when treated with ICZ and TCDD during the total treatment period when compared with controls. This indicates that the initial inhibition of GJIC is at least partly caused by a down-regulation of the transcription of Cx32, but it is not possible to exclude any inhibitory effects directly on the functional Cx32. The continuously low level of Cx32 mRNA throughout the experiment, also when GJIC and the Cx32 plaques are partly recovered, point out a second mechanism of regulation at post-transcriptional level.

The present results show that ICZ may possess tumor-promoter activity, although various factors may modify this effect in vivo. The GJIC inhibitory effects of ICZ at concentration below 0.1 μM are not known and only small amounts of ICZ are formed in the human body during ingestion of cruciferous vegetables. Only 0.0002% ICZ is formed during a 10 min acid-catalyzed conversion of I3C (11) but in rats the level of ICZ increased dramatically in the liver following a single intake of 500 μmol I3C/kg body wt to 0.06 nM (42) and to 1.6 nM following a 7 days treatment with 880 mmol I3C/kg body wt/day (43). Furthermore, other I3C-derived products may also act as ligands for the AhR and hence add to the possible cancer-promoting activity of ICZ.

In conclusion, the results presented in this report demonstrate that ICZ inhibits GJIC in rat primary hepatocytes by preventing the synthesis and/or the fusion of Cx32 proteins with the cell membrane. Furthermore, ICZ induces the EROD activity concomitant with a reduction in Cx32 immunostaining at the hepatocyte–hepatocyte interface. The effects, however, are transient and the level of GJIC and Cx32 staining returns to normal levels after about a week of continuous treatment. The studies reported herein do indicate that ICZ reduces hepatocyte GJIC possibly through AhR activation and/or increased Cyp1 activity. However, the present results do not provide final evidence to a role of the AhR and Cyp1 activity in ICZ- and TCDD-induced GJIC inhibition and more studies are clearly needed.

The inhibition of GJIC by ICZ may indicate that ICZ has tumor-promoting activity. These results call for studies focusing on the tumor-promoting activity of ICZ and perhaps other degradation products of I3C, before I3C supplement recommendations should be made to healthy individuals.


We thank Ms Mette Nielsen and Prof. Randall Ruch (Medical College of Ohio, OH) for the introduction to the GJIC analysis and Anne Lise Maarup for technical assistance in performing the EROD analysis. The work was partly supported by Rockwooll Foundation.