Experimental Procedures

Materials.

Male Sprague-Dawley rats, 200 to 300 g, from Harlan Sprague-Dawley Inc. (Indianapolis, IN) were used for hepatocyte isolation unless noted otherwise. Cell culture medium (Waymouth's MB 752/1), insulin, antibiotics, and other cell culture supplies were purchased from Life Technologies, Inc. (Bethesda, MD). CPT-cAMP, forskolin, dideoxyforskolin, and glucagon were purchased from Sigma (St. Louis, MO); R_p-cAMPS andS_p-cAMPS were purchased from Research Biochemicals International (Natick, MA) or from Biomol (Plymouth Meeting, PA); and (dB)-cAMP was purchased from Boehringer-Mannheim (Indianapolis, IN). 5,6-Dichloro-1-β-d-ribofuranosylbenzimidazole (DRB) was from Calbiochem (San Diego, CA). Enzyme immunoassay kits for cAMP were obtained from Amersham Life Science (Arlington Heights, IL). Matrigel was prepared as described previously (Chen et al., 1995).

Hepatocyte Isolation and Culture.

Isolation of rat hepatocytes was performed by in situ collagenase perfusion (Liddle et al., 1992). The viability of hepatocytes was between 70 and 85% (trypan blue exclusion) and the yield was 100 to 250 × 10⁶ viable cells/liver. The hepatocytes were plated in Waymouth's medium containing (unless otherwise stated) 185 nM insulin (Liddle et al., 1992) on 60-mm culture dishes (Falcon; Becton Dickinson, Lincoln Park, NJ) coated with 0.3 ml of Matrigel (7 mg/ml) at a density of 3 to 4 × 10⁶ live cells/dish. Medium was changed to remove the dead cells 4 h after plating, and every 48 h thereafter. All treatments of cells were initiated after 5 days in culture and cells were harvested at the indicated times as described previously (Liddle et al., 1992; Iber et al., 1997). DRB, forskolin, and dideoxyforskolin were dissolved in dimethyl sulfoxide; cAMP analogs and epinephrine were dissolved in culture medium. Glucagon was diluted from a 1 mM stock solution in glacial acetic acid. In all cases, control groups were treated with the appropriate concentration of vehicle.

Isolation of Total RNA, Northern Blots, and Slot-Blot Assays.

Total hepatocyte RNA was prepared by the acid phenol extraction method (Chomczynski and Sacchi, 1987). The relative abundance of CYP2C11 mRNA was measured by a slot-blot hybridization assay as described previously (Morgan, 1989) using a full-length CYP2C11 cDNA as a probe. Under these conditions CYP2C11 cDNA hybridizes with a single RNA band in Northern blots of RNA from male, but not female, rat liver or primary hepatocytes (Sewer and Morgan, 1997). The results were normalized to the contents of poly(A⁺) RNA in the samples, measured by probing slot-blots with an oligo(dT)₃₀ probe (Wright and Morgan, 1991). Bound, ³²P-labeled probes were detected by autoradiography and quantified by analysis on a Lynx video densitometer (Applied Imaging, Santa Clara, CA). The loads of total RNA used were previously determined to be in the linear range for the assay.

The expression of connexin32 mRNA in the samples was analyzed by Northern blotting. RNA was fractionated in the presence of 5% formaldehyde by electrophoresis on a 1% agarose gel (Sambrook et al., 1989), and transferred to nylon transfer membrane filters (MagnaGraph; Micron Separations, Inc., Westboro, MA) using a PosiBlot 30-30 pressure blotting system (Stratagene, La Jolla, CA). Following fixation of the blot by UV irradiation and baking, the blots were hybridized overnight at 40°C with an oligonucleotide complementary to nucleotides 75 to 114 (in the coding region) of the Connexin32 mRNA (GenBank X04070; 5′-ATGACGGACAGCCATACTCGGCCAATGGCTGTAGAATGCC-3′). The hybridization buffer was 10 mM sodium phosphate, pH 7.7, containing 0.18 M NaCl, 2 mM EDTA, 0.5% SDS, 0.1 mg/ml yeast tRNA, and 5× Denhardt's solution. Blots were washed twice for 3 min and twice for 30 min in 1× standard saline citrate (SSC) (15 mM sodium citrate pH 7.0, 0.15 M NaCl), 0.5% SDS, and twice for 30 min in 0.2× SSC, 1% SDS, all at room temperature. The final wash was for 30 min in 0.1× SSC, 0.5% SDS at 40°C. Under these conditions, the probe recognized a single RNA band on Northern blots of rat liver RNA, with an apparent size of 1.7 kb (relative to ribosomal RNAs: cf. 1.6 kb in Kojima et al., 1997).

Measurement of Cellular cAMP Concentration.

cAMP concentration was measured using the cAMP enzyme immunoassay system (Amersham Life Science). Following incubation of the cells with glucagon, hepatocytes were washed one time and harvested in 1 ml of cold phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 12 mM Na₂HPO₄, 1.7 mM KH₂PO₄ pH 7.4). To this was added 1.5 ml of ice-cold ethanol. The extracts were mixed, and centrifuged at 2000g for 15 min at 4°C and the supernatants were collected. The extracts were dried under vacuum, dissolved in 500 μl of assay buffer, and subsequently analyzed according to the manufacturer's protocol.

Presentation of Results and Statistical Analyses.

The RNA slot-blot assays used are semiquantitative only. Therefore, no attempt has been made to express measurements in any absolute units. Values for each experiment were calculated as a percentage of the mean value (arbitrary units) for an appropriate control group. One-way analysis of variance and Newman-Keuls test were used to test for significant differences between the means of different groups. All results are expressed as the mean ± S.E.M. for each group.

Results

Down-Regulation of CYP2C11 by Glucagon, Epinephrine, and Forskolin.

Treatment of primary hepatocytes with different concentrations of glucagon for 24 h caused suppression of CYP2C11 mRNA with an EC₅₀ of approximately 1 nM (Fig.1, A and B). Maximally effective concentrations of glucagon (≥30 nM) reduced expression of CYP2C11 mRNA to approximately 20% of control levels (Fig. 1, A and B). Epinephrine, which elevates cAMP via both the β₂- and the α₁-adrenergic receptors in hepatocytes (Morgan et al., 1983), also down-regulated CYP2C11 to a similar extent, with a maximum effect occurring at 10 μM the hormone (Fig. 1B). The adenylate cyclase activator forskolin mimicked the effects of glucagon and epinephrine (Fig. 1B). The EC₅₀ for forskolin, estimated from the experiment shown here and in other experiments not shown, was approximately 5 μM. To test the specificity of the forskolin effect, we examined the ability of the biologically inactive analog dideoxyforskolin (Seamon et al., 1984) to down-regulate CYP2C11. In contrast to forskolin, dideoxyforskolin at concentrations up to 100 μM failed to cause suppression of CYP2C11 expression (Fig. 1C).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Concentration-dependent suppression of CYP2C11 mRNA by adenylate cyclase activators. Hepatocytes were cultured on Matrigel for 5 days, and then incubated with glucagon, forskolin, dideoxyforskolin, or epinephrine at the indicated concentrations for 24 h. Cells were harvested and total RNA was prepared for slot-blot analysis of CYP2C11 expression. A, slot-blots of glucagon- or vehicle-treated cells probed with CYP2C11 cDNA (6 μg of RNA per slot) or oligo(dT) (0.5 μg of RNA per slot). Each band represents a single independent sample. Only four samples were present on the CYP2C11 blot of the 0.1 nM group due to sample loss during preparation. B, concentration-response curves for glucagon, forskolin, and epinephrine. C, cells were treated with 100 μM forskolin, dideoxyforskolin, or vehicle. CYP2C11 expression levels in B and C were normalized to the poly(A⁺) RNA contents of the samples. Data are expressed as the means ± S.E.M. for each group, and are expressed as a percentage of the mean value for the respective vehicle-treated (control) groups. Each data point represents the mean of four to five independent samples. *Significantly different from untreated group mean, P ≤ 0.05.

Expression of gap junction proteins is one marker of the degree of hepatocyte cell differentiation (Kojima et al., 1997). We found that treatment of hepatocytes with glucagon or forskolin for 24 h did not substantially affect the expression of the major hepatocyte gap junction protein connexin32 (Fig. 2), suggesting that these treatments do not cause significant changes in the differentiation state of the cells in this time frame.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Lack of effect of glucagon or forskolin on hepatocyte differentiation state. Hepatocytes were cultured on Matrigel for 5 days, and then incubated with 30 nM glucagon, 100 μM forskolin, or the respective vehicles for 24 h. Cells were harvested and total RNA was prepared for Northern blots analysis of connexin32 expression. A, Northern blots (20 μg of RNA per lane) probed with connexin32 oligonucleotide. B, quantitative analyses of a second experiment, in which the connexin32 mRNA content is normalized to the GAP mRNA contents of the samples. The data are the means ± S.E.M. of five independent samples for each group, and are expressed as a percentage of the mean value for the respective vehicle-treated (control) groups. Neither of the treatment groups was significantly different from its respective control.

Effects of cAMP Analogs.

To determine whether the suppression of CYP2C11 by glucagon, epinephrine, and forskolin is related to their abilities to elevate cAMP in the hepatocyte, we treated the cells with four different cell-permeable cAMP analogs. Exposure of hepatocytes to increasing concentrations of either (dB)-cAMP,S_p-cAMPS, or CPT-cAMP for 24 h caused a concentration-dependent suppression of CYP2C11 expression to about 20 to 40% of control levels in each case (Fig.3). CPT-cAMP was about 10- to 20-fold more potent, and S_p-cAMPS was 4 to 5 times more potent than (dB)-cAMP. Similar relative potencies of these agents have been observed for the cAMP antagonism of phenobarbital induction of P450 2B1 and 3A1 (Sidhu and Omiecinski, 1995).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Suppression of CYP2C11 mRNA by cAMP analogs. Hepatocytes were cultured on Matrigel for 5 days, and then treated with (dB)-cAMP, S_p-cAMPS,R_p-cAMPS or CPT-cAMP at the indicated concentrations for 24 h. CYP2C11 mRNA contents were analyzed by slot-blotting, and normalized to the poly(A⁺) RNA contents of the samples. The data are the means ± S.E.M. for each group, and are expressed as a percentage of the mean value for the respective vehicle-treated (control) groups. Each point represents the mean of four to five independent samples. *Significantly different from untreated group mean, P ≤ 0.05.

In contrast to the cAMP agonists,R_p-cAMPS did not affect CYP2C11 expression in the same concentration range (Fig. 3).R_p-cAMPS is a diastereomer ofS_p-cAMPS, and is a specific inhibitor of PKA activation that blocks cAMP binding to the regulatory subunit of the enzyme (Rothermel and Parker Botelho, 1988). Consistent with the involvement of PKA in CYP2C11 regulation,R_p-cAMPS was able to partially block the inhibitory effect of S_p-cAMPS on CYP2C11 expression (Fig. 4B). At the higher concentrations required to inhibit the effect ofS_p-cAMPS (100 μM),R_p-cAMPS itself produced a 40% suppression of CYP2C11 expression (Fig. 4B). This precluded experiments with higher concentrations of the antagonist.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Inhibition ofS_p-cAMPS-dependent CYP2C11 suppression byR_p-cAMPS. Hepatocytes were cultured on Matrigel for 5 days, and then treated with 30 μMS_p-cAMPS (S), 100 μMR_p-cAMPS (R), or both diastereomers for 16 h. Control cells (C) were treated with an equivalent volume of medium. A, slot-blots of total RNA from five independent samples per group, probed with CYP2C11 and oligo(dT) probes. B, quantitative analysis of the data. CYP2C11 mRNA contents were normalized to the poly(A⁺) RNA contents of the samples. The data are the means ± S.E.M. for each group, expressed as a percentage of the mean value for untreated control cells. *Significantly different from untreated group mean;^‡S_p-cAMPS-treated group mean; P < 0.05.

Time Course of CYP2C11 Suppression.

Forskolin and glucagon caused suppression of CYP2C11 mRNA with similar time courses. No significant effect of either agent was observed until 16 h after treatment of the cells, although the mean values started to decline within about 8 h (Fig. 5A). When hepatocytes were treated with the transcription inhibitor DRB, CYP2C11 mRNA declined with a half-life of 9.8 h (Fig. 5B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Time courses of suppression of CYP2C11 mRNA by glucagon, forskolin, and the transcription inhibitor DRB. Hepatocytes were cultured for 5 days, and then treated with 300 nM glucagon, 100 μM forskolin, or 157 μM DRB for the indicated amounts of time. The data are the means ± S.E.M. for each group, and are expressed as a percentage of the mean value for the respective vehicle-treated group at each time point. Each point represents the mean of four to five independent samples. *Significantly different from untreated group mean, P < 0.05. A, CYP2C11 mRNA levels in cells treated with glucagon or forskolin were analyzed by slot-blotting, and normalized to the poly(A⁺) RNA contents of the samples. B, CYP2C11 levels in cells treated with DRB were normalized to the rRNA contents of the samples, because DRB significantly decreased the poly(A⁺) contents during the course of the experiment. The slope of the decay curve was determined by exponential curve fitting using Kaleidagraph 3.5 (Synergy Software, Reading, PA). Hepatocytes from a male Fischer 344 rat were used for this experiment.

Influence of Insulin on cAMP Production and CYP2C11 Suppression in Response to Glucagon.

The intracellular cAMP concentration in liver is regulated by plasma insulin and glucagon concentrations, and their relative ratio (Hill and McCallum, 1991; Pilkis and Granner, 1992). To investigate the influence of insulin ratio on the suppression of CYP2C11 by glucagon, we cultured hepatocytes for 5 days in the presence of different insulin concentrations and measured the generation of cAMP and the suppression of CYP2C11 in response to varying concentrations of glucagon. As seen in Fig.6A, the basal levels of cAMP in hepatocytes were increased by culturing them for 5 days in the presence of increasing concentrations of insulin. cAMP levels in cells cultured in no insulin or low insulin were more sensitive to the effects of glucagon than cells cultured in high insulin (Fig. 6A). Basal expression of CYP2C11 is dependent on insulin, and we could not accurately measure the effects of glucagon on CYP2C11 levels in cells cultured at insulin concentrations of 46 nM or lower. When insulin concentrations were varied over a 4-fold range, the effects of insulin and glucagon on CYP2C11 expression appeared to be antagonistic (Fig.6B). Lowering the insulin concentration to 185 or 92.5 nM resulted in a decline in CYP2C11 expression to 89 and 75%, respectively, of the level measured at 370 nM insulin, and glucagon added at 0.1 or 1 nM suppressed the expression of CYP2C11 mRNA to 65 and 54% of the control level at 370 nM insulin (Fig. 6B). When cells were subjected to both lower insulin concentrations and glucagon treatment, CYP2C11 expression declined to levels lower than those observed with either condition alone. CYP2C11 mRNA levels in cells cultured in 185 nM insulin and treated with 0.1 or 1 nM glucagon was suppressed to 51 and 31%, respectively, compared with untreated cells cultured in high insulin. Thus, the effects of lowering the insulin concentration and raising the glucagon concentration were additive.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Effect of insulin concentration on suppression of CYP2C11 and elevation of cAMP by glucagon. Primary hepatocytes were cultured on Matrigel with the indicated insulin concentrations for 5 days, and then treated with the indicated glucagon and insulin concentrations for either 25 min (cAMP measurements) or 24 h (CYP2C11 mRNA measurements). A, for cAMP measurements, a representative of two experiments with n ≥ 3 for each group is shown. B, CYP2C11 measurements are from a single experiment withn > 5 for each group. CYP2C11 mRNA are expressed as a percentage of the mean value for cells cultured in the presence of 370 nM insulin and treated with medium only (0 nM glucagon). ND, not determined. *Significantly different from respective 0 nM glucagon group.

Discussion

This work shows for the first time that CYP2C11 is down-regulated by glucagon and epinephrine. The effects of these agents are mimicked by the adenylate cyclase activator forskolin, as well as by cell-permeable cAMP analogs, suggesting that the hormones down-regulate CYP2C11 via elevation of the second messenger cAMP and subsequent activation of PKA. The effects of glucagon on CYP2C11 expression are antagonized by higher concentrations of insulin, suggesting that the plasma glucagon/insulin (G/I) ratio may play a role in physiological regulation of this enzyme in diabetes, fasting, and perhaps also inflammation.

Two main lines of evidence indicate that the effects of cAMP analogs used in this study were specifically due to activation of PKA. First, their relative potencies in down-regulation of CYP2C11 were similar to those used in other studies of cAMP-regulated signaling (Sandberg et al., 1991), and to those found for inhibition of CYP2B1 and 3A1 induction by phenobarbital (Sidhu and Omiecinski, 1995). Second,R_p-cAMPS, a diastereomer ofS_p-cAMPS that competes with cAMP agonists for binding to the regulatory subunit of PKA, but does not activate the enzyme (Rothermel et al., 1984; Rothermel and Parker Botelho, 1988), failed to down-regulate CYP2C11 expression and also attenuated the suppression of CYP2C11 byS_p-cAMPS. The specificity of action of forskolin in suppression of CYP2C11 via activation of adenylate cyclase is supported by the fact that dideoxyforskolin, which does not bind to or activate adenylate cyclase (Seamon et al., 1984), likewise failed to down-regulate CYP2C11.

The potency of glucagon in suppression of CYP2C11 is similar to that reported for its cAMP-dependent induction of hepatocyte glycogenolysis (Rothermel et al., 1984), and inhibition of phenobarbital-induced CYP2B expression (Sidhu and Omiecinski, 1995) in hepatocytes; higher concentrations of glucagon are apparently required for induction of CYP2E1 mRNA (Woodcroft and Novak, 1999). By the same token, the potency of epinephrine in CYP2C11 suppression is similar to that for the elevation of hepatocyte cAMP levels (Morgan et al., 1983). This, together with the ability of forskolin and cAMP analogs to mimic the effects of the hormones, suggests that the glucagon and epinephrine regulate CYP2C11 via cAMP. However, proof of this hypothesis awaits a clear demonstration that a specific PKA inhibitor can attenuate the suppression of CYP2C11 by glucagon or epinephrine. In one experiment, we obtained almost complete inhibition of glucagon-induced CYP2C11 suppression using only 3 μMR_p-cAMPS, without affecting suppression by the cytokine tumor growth factor-β (data not shown). However, in subsequent experiments we have been unable to consistently reproduce this observation. Inhibition of glucagon-stimulated glycogenolysis in freshly isolated hepatocytes by 10 μMR_p-cAMPS produces only a 6-fold decrease in glucagon potency (Rothermel et al., 1984), which would make it difficult to detect in our system given the level of variability inherent in CYP2C11 expression in cultured hepatocytes. It is also possible that CYP2C11 suppression (occurring over 8–16 h) by glucagon requires a more sustained cAMP signal than does stimulation of glycolysis (60 min); differences in the effects ofR_p-cAMPS on CYP2C11 suppression versus glycolysis could thus be related to the time-dependent metabolism of the antagonist in the cultures.

The half-life of CYP2C11 RNA in unstimulated cells, measured using the transcriptional inhibitor DRB, was found to be approximately 10 h, similar to that reported for the closely related CYP2C12 mRNA based on its decay in the presence of actinomycin D (Tollet et al., 1990). The kinetics of suppression of CYP2C11 mRNA by glucagon and forskolin were similar to those observed in the presence of DRB, and we therefore conclude that the major mechanism for suppression of CYP2C11 by glucagon and forskolin is likely to be transcriptional. These kinetics are also similar to those for suppression of the same mRNA by interleukin-1 (Chen et al., 1995), and we have found that interleukin-1 inhibits the expression of CYP2C11 via a transcriptional suppression (Chen et al., 1995). On the other hand, our results cannot exclude a possible post-transcriptional component to the glucagon-evoked suppression of CYP2C11. It has been reported that the decrease in CYP2C11 mRNA that occurs after hepatocytes are isolated is due to stimulated RNA degradation, providing a precedent for such a mechanism (Wang et al., 1997).

Expression of CYP2C11 and other P450s in hepatocytes cultured on Matrigel occurs because culture on this substratum maintains the cells in a highly differentiated state (Schuetz et al., 1988). We therefore considered whether down-regulation of CYP2C11 by cAMP-elevating agents could be a consequence of a reduced state of hepatocyte differentiation. Our results showed that the expression of the mRNA for connexin32, the gap junction protein whose expression in hepatocytes is also dependent on a highly differentiated state of the cell (Kojima et al., 1997), was unaffected by concentrations of glucagon and forskolin that down-regulated CYP2C11. Glucagon treatment actually helps to suppress proliferation and to maintain expression of connexin32 in hepatocytes cultured on collagen (Kojima et al., 1997). We conclude, therefore, that the effects of glucagon or forskolin on CYP2C11 expression are not due to induced dedifferentiation of the hepatocytes.

It was shown previously that the intracellular cAMP concentration in liver depends on the relative G/I ratio in blood (Hill and McCallum, 1991). Glucagon activates cAMP synthesis, and uses the nucleotide as a second messenger to transduce its signal (Exton, 1987), whereas insulin stimulates cAMP degradation by activating a phosphodiesterase (Pilkis and Granner, 1992). Our results show that the G/I ratio influences CYP2C11 mRNA expression. The degree of suppression of CYP2C11 by glucagon is attenuated in the presence of higher insulin concentrations, and this correlates with the G/I dependence of cAMP. Thus, changing the G/I ratio gives the organism a more sensitive tool to influence multiple cellular parameters, including CYP2C11 expression, than does modulating either of these hormones alone.

The observation that glucagon can down-regulate CYP2C11 expression is novel, and suggests that this hormone contributes to the suppression of CYP2C11 in diabetes (Donahue et al., 1991). A substantial portion of the decrease in diabetic rats can be attributed to a change in the male-specific pattern of GH secretion that is required for maximal CYP2C11 expression (Donahue et al., 1991). However, there is a GH-independent component to the decrease, because the down-regulation is still seen in diabetic hypophysectomized animals with or without GH supplementation (Donahue et al., 1991). The fact that glucagon down-regulates CYP2C11 expression, and that the magnitude of the effect is dependent on the G/I ratio, provides a possible hormonal basis for the GH-independent component.

Since the plasma glucagon/insulin ratio is also increased in starvation, it may contribute to the CYP2C11 down-regulation that occurs in that state also (Imaoka et al., 1990). Although the down-regulation of CYP2C11 during inflammation and infection can be mimicked by inflammatory cytokines, it is also possible that the increased plasma glucagon and epinephrine concentrations that occur in these pathophysiological states (Hill and McCallum, 1991; Lynch et al., 1997) could also play a role. Correspondingly, although physiological levels of glucocorticoids can suppress CYP2C11 in vitro (Iber et al., 1997) and in vivo (Murray, 2000), our finding that epinephrine can down-regulate CYP2C11 in hepatocytes may indicate that release of catecholamines could also contribute to the in vivo suppression of CYP2C11 caused by stress (Merrill et al., 1992).

In conclusion, the discovery that CYP2C11 expression is regulated by glucagon, which is affected in several physiological and pathophysiological states, and that this in turn regulates CYP2C11 through cAMP, a cellular second messenger used by many different hormones, provides a possible explanation why this major rat P450 is down-regulated in many models of stress and metabolic and inflammatory disease. It remains to be determined whether other constitutively expressed P450s will be regulated by this mechanism.