Previously, we demonstrated that deoxycholic acid (DCA)-induced ERK1/2 and AKT signaling in primary hepatocytes is a protective response. In the present study, we examined the regulation of the phosphatidylinositol 3 (PI3) kinase/AKT/glycogen synthase (kinase) 3 (GSK3)/glycogen synthase (GS) pathway by bile acids. In primary hepatocytes, DCA activated ERBB1 (the epidermal growth factor receptor), ERBB2, and the insulin receptor, but not the insulin-like growth factor 1 (IGF-1) receptor. DCA-induced activation of the insulin receptor correlated with enhanced phosphorylation of insulin receptor substrate 1, effects that were both blocked by the insulin receptor inhibitor AG1024 and by expression of the dominant negative IGF-1 receptor (K1003R), which inhibited in trans. Expression of the dominant negative IGF-1 receptor (K1003R) also abolished DCA-induced AKT activation. Bile acid–induced activation of AKT and phosphorylation of GSK3 were blunted by the ERBB1 inhibitor AG1478 and abolished by AG1024. Bile acids caused activation of GS to a similar level induced by insulin (50 nM); both were blocked by inhibition of insulin receptor function and the PI3 kinase/AKT/GSK3 pathway. In conclusion, these findings suggest that bile acids and insulin may cooperate to regulate glucose storage in hepatocytes. (HEPATOLOGY 2004;39:456–463.)
Bile acids are detergent molecules—synthesized from cholesterol in the liver—that are released into the gut upon feeding and are essential for digestion.1 In the intestine, bile acids function in the solubilization and absorption of fats, certain vitamins, and cholesterol.2 After feeding, bile acids re-enter the liver by way of the portal vein together with digested nutrients and are recirculated into the gall bladder for use during the next feeding cycle.3 When retained within the liver because of impaired secretion into the bile canaliculi, bile acids are also known to have hepatocellular toxicity both in vivo and in vitro.4–8
Recently, we reported that treatment of primary rodent and human hepatocytes with physiologic concentrations of the bile acid deoxycholic acid (DCA), but not with the steroid-based detergent CHAPS (3-[3-cholamidopropyl)dimethyl-ammonio]-1-propane-sulfonate), could cause activation of ERBB1 (the epidermal growth factor receptor), which was responsible for activation of the ERK1/2 pathway.9–11 Similar data have been observed in other cell types.12 A blockade of DCA-induced ERK1/2 and AKT activation, with inhibitors of Ras, PI3 kinase, or MEK1/2, increased apoptosis tenfold within 6 hours of exposure. Apoptosis was dependent on bile acid–induced, ligand-independent activation of the FAS death receptor. Several other groups have also discovered that bile acids can activate ERBB1, the membrane associated tyrosine kinase SRC and the FAS receptor.13
The present study was designed initially to determine whether bile acids can activate other receptor and nonreceptor tyrosine kinases in primary hepatocytes, similar to effects induced by ionizing radiation in tumor cells. Based on the discovery that bile acids cause activation of the insulin receptor, further analyses then examined whether or not activation of the insulin receptor modulated hepatocyte GS activity by way of PI3 kinase/AKT/GSK3 signaling. Studies of various cell types, including hepatocytes, have linked the ability of insulin to regulate GS activity by way of the PI3 kinase/AKT/GSK3 pathway.14–16 The data in this article demonstrate that bile acids regulate glycogen synthase activity in primary rodent hepatocytes by way of the PI3 kinase/AKT pathway and suggest that bile acids could potentially cooperate with insulin to control glucose metabolism in the liver.
All bile acids and CHAPS were obtained from Sigma Chemical Co. (St. Louis, MO). Phospho-/total-ERK1/2, Phospho-/total-GSK3, Anti-S473 AKT, Anti-T308 AKT, and total AKT were purchased from Cell Signaling Technologies (Worcester, MA). All the secondary antibodies (anti–rabbit-HRP, anti–mouse-HRP, and anti–goat-HRP) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). PI3 kinase inhibitor (LY294002), ERBB1 inhibitor (AG1478), and insulin receptor/insulin-like growth factor 1 (IGF-1) receptor inhibitor (AG1024) were supplied by Calbiochem (San Diego, CA) as powder, dissolved in sterile dimethyl sulfoxide (DMSO), and stored frozen under light-protected conditions at −80°C. Enhanced chemiluminescence kits were purchased from Amersham Enhanced Chemi-Luminescence system (Bucks, England) and NEN Life Science Products (Boston, MA). Trypsin-EDTA, Williams Medium E, and penicillin-streptomycin were purchased from Gibco-BRL Life Technologies (Grand Island, NY). Other reagents were as in earlier studies.9–11
Primary Culture of Rodent Hepatocytes.
Hepatocytes were isolated from adult male Sprague-Dawley rats by way of the two-step collagenase perfusion technique. The freshly isolated hepatocytes were plated on a rat-tail collagen (Vitrogen)-coated plate at a density of 2 × 105 cells/well and cultured in Williams E medium supplemented with 0.1 nM dexamethasone, 1 nM thyroxine, and 100 μg/ml of penicillin/streptomycin at 37°C in a humidified atmosphere containing 5% CO2. The initial medium change was performed 3 hours after cell seeding to minimize the contamination of dead or mechanically damaged cells. In some studies, where noted, hepatocytes were cultured in the presence of 50 nM insulin, 0.1 nM dexamethasone, 1 nM thyroxine, and 100 μg/ml of penicillin/streptomycin.
Recombinant Adenoviral Vectors: Generation and Infection In Vitro.
Two adenoviral technologies were used. Replication defective adenovirus was conjugated to poly-L-lysine and a cDNA plasmid construct as described in earlier studies.9–11 Second, recombinant adenoviruses were generated.17 Hepatocytes were transfected/infected with these adenoviruses at an approximate multiplicity of infection (MOI) of 200 and 30, respectively. Cells were further incubated for 24 hours to ensure adequate expression of transduced gene products.
SDS-PAGE and Western Blot Analysis.
At various time points after indicated treatment, hepatocytes were lysed in whole-cell lysis buffer (0.5 M Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 1% β-mercaptoethanol, 0.02% bromophenol blue) and the samples were boiled for 30 minutes. The boiled samples were loaded onto 14% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and electrophoresis was run overnight. Proteins were electrophoretically transferred onto 0.22 μm nitrocellulose and immunoblotted with various primary antibodies against different proteins. All immunoblots were visualized by way of enhanced chemiluminescence. For presentation, immunoblots were digitally scanned at 600 dpi using Adobe PhotoShop 5.0, and their color was removed; the figures were generated using Microsoft PowerPoint.
GS assays were performed essentially as described in earlier studies.18–20 After DCA treatment cells were washed with cold phosphate buffered saline (PBS) and scraped into 400 mL lysis buffer (50 mM Tris-HCl pH 7.8, 10 mM EDTA, 100 mM NaCl, 50 mM NaF, 1 μM Microcystin-LR, 1% (v/v) NP40, 1 mM PMSF, 40 μg/ml TPCK, 40 μg/ml TLCK). Fifty μl of cell lysate were added to an equal volume of GS assay buffer (50 mM Tris/HCl pH 7.8, 10 mM EDTA, 50 mM NaF, 1 μM Microcystin-LR containing UDP-[14C] glucose (0.5 μCi/mmol) and 15 mg/ml glycogen ± 10 mM G-6-P. After 15 minutes of incubation at 37°C, tubes were then chilled for 15 minutes on ice, after which the entire tube contents were spotted onto Whatman GF/A 2.4-cm filter papers (Maidstone, Kent, England). Spotted filter papers were immediately immersed in 25 ml of 70% (v/v) ethanol (4°C), and washed twice for 90 minutes each time. Filter papers were air-dried; radioactivity was incorporated into glycogen, determined by liquid scintillation spectroscopy.
Comparison of the effects of various treatments was performed using one-way ANOVA and a two-tailed t test. Differences with a P value less than .05 were considered statistically significant. Experiments shown are the means of multiple individual points (±SEM).
Initial studies characterized the activation of growth factor receptors by bile acids in primary hepatocytes. DCA—but not CHAPS—enhanced tyrosine phosphorylation of ERBB1, ERBB2, the insulin receptor and SHC proteins (Fig. 1A).9 Expression of the kinase inactive receptor ERBB3 was not detected in primary rodent hepatocytes under our culture conditions (data not shown). Surprisingly, DCA did not enhance tyrosine phosphorylation of the IGF-1 receptor (Fig. 1A). In contrast, insulin treatment of hepatocytes increased phosphorylation of the insulin receptor (Fig. 1B), which was greater than DCA-induced phosphorylation of the insulin receptor.
In agreement with activation of the insulin receptor, DCA also enhanced tyrosine phosphorylation of insulin receptor substrate 1 (IRS-1) (Fig. 1C). Phosphorylation of the insulin receptor and IRS-1 were significantly blunted by the insulin/IGF-1 receptor inhibitor AG1024 (Fig. 1C). The ERBB1 inhibitor AG1478 neither altered insulin receptor nor IRS-1 phosphorylation following DCA exposure (data not shown). Enhanced tyrosine phosphorylation of IRS-1 correlated with increased association of the p85 subunit of PI3 kinase with IRS-1 in IRS-1 immunoprecipitates, an association that was also blocked by the inhibitor AG1024 (Fig. 1C).
Previously, we have observed that the ERBB1 inhibitor AG1478, but not the ERBB2 inhibitor AG825, was capable of blunting DCA-induced activation of ERK1/2, AKT, and p70 S6 kinase.9–11 Because ERBB1 is thought to be a weaker potential activator of PI3 kinase than that of insulin and IGF-1 receptor signaling,21–23 we compared the abilities of AG1478 and AG1024 to modify DCA-induced AKT and GSK3 phosphorylation. DCA enhanced phosphorylation of AKT at T308 and at S473, which is indicative of kinase activation (Fig. 2A).24 Similar data were obtained with taurodeoxycholic acid (TDCA) and taurocholic acid (TCA) (Fig. 2B). Doses of DCA and TCA as low as 5 μM stimulated AKT phosphorylation (Fig. 2B). Incubation of hepatocytes with AG1024 abolished basal, DCA-stimulated, and DCA–insulin-stimulated AKT and GSK3 phosphorylation (Fig. 2C). In contrast, AG1478 blunted but did not abolish basal and DCA-stimulated AKT and GSK3 phosphorylation (Fig. 2D). Furthermore, in the presence of insulin, AG1478 only weakly modified DCA-induced AKT or GSK3 phosphorylation. In a similar manner to AG1024, inhibition of PI3 kinase signaling abolished the phosphorylation of AKT and GSK3, which is in agreement with other studies (see below).14–16, 21–23
The PI3 kinase/AKT pathway has been linked to the regulation of glycogen synthase (GS) by insulin, and further experiments tested whether bile acids could also regulate GS activity by way of insulin receptor/PI3 kinase/AKT signaling.14–16 DCA and the more hydrophilic bile acids TDCA and TCA enhanced GS activity by approximately 40% (Fig. 3A). This level of GS activity was similar to that observed in cells cultured in 50 nM insulin for 24 hours (Fig. 3A). To investigate whether or not DCA and insulin interact to cause GS activation, cells were treated either with one agent or with both agents. As individual agents, both insulin and DCA activated GS (Fig. 3B). The combination of insulin and DCA caused greater activation of GS than either agent alone that was greater than the additive combination of the individual treatments at the 30-minute time point. Furthermore, in hepatocytes cultured in the presence of 50 nM insulin for 24 hours, DCA also caused a further modest activation of GS (Fig. 3C); this is in general agreement with the enhanced AKT and GSK3 phosphorylation data presented in Fig. 2.
Because DCA activated the AKT/GSK3 pathway predominantly by way of the insulin receptor, we next investigated whether or not inhibition of insulin receptor signaling blunted DCA-induced GS activation. Incubation of cells with AG1024 did not significantly alter basal GS activity (Fig. 4A). However, AG1024 significantly reduced DCA-stimulated activation of GS. The ability of DCA to enhance GS activity was also blocked by expression of the PI3 kinase inhibitor LY294002 (Fig. 4B). To confirm by a molecular approach that insulin receptor signaling was causal in regulating DCA-induced GS activity, we expressed a mutant dominant negative IGF-1 receptor K1003R in hepatocytes.25, 26 Expression of this dominant negative molecule abolished DCA-induced phosphorylation of the insulin receptor, IRS-1, and AKT (Fig. 4C). Others have shown that dominant negative IGF-1 receptors can act intrans to also inhibit the function of the insulin receptor.26 In agreement with inhibition of insulin receptor function and IRS-1 and AKT phosphorylation, the dominant negative IGF-1 receptor K1003R also blocked DCA-induced GS activation (Fig. 4D).
Previous studies by this group have linked bile acid–induced ERBB1 activation to enhanced ERK1/2 and AKT signaling leading to a cytoprotective response versus bile acid–induced FAS receptor/caspase activation.9–11 Studies by other groups have shown similar phenomena for bile acid–induced activation of ERBB1 as well as the nonreceptor tyrosine kinase SRC in established hepatoma and colon carcinoma cells.12, 13 The studies described in this article were designed initially to determine whether bile acids induced activation of other growth factor receptors in primary rodent hepatocytes.
DCA activated ERBB2 and the insulin receptor in primary hepatocytes with similar kinetics to that previously observed for the growth factor receptor ERBB1.11 In contrast, DCA did not enhance tyrosine phosphorylation of the IGF-1 receptor. CHAPS did not activate either the receptors or the signaling pathways.9–11 The findings in this manuscript with ERBB2 are expected based on the known ability of active ERBB1 to phosphorylate and transactivate ERBB2.11, 27 Our data, as well as those published by other groups with bile acids,28 are also similar to the pleiotropic effect ionizing radiation has on the activation of ERBB family and IGF-1 growth factor receptors in cancer cells.22, 23, 29, 30 A further similarity in the biology of radiation and bile acid–induced signaling effects is that quenching the generation of free radical species by both agents blocks both radiation- and DCA-induced growth factor receptor phosphorylation.11, 21, 23 The phosphorylation of ERBB1 and ERBB2 is negatively regulated by protein tyrosine phosphatases, which contain a free radical sensitive Cys in their active site; previously, we presented evidence that DCA can inhibit phosphatase activity against purified ERBB1 in vitro.11 Additional studies will be required to determine whether DCA-induced activation of ERBB2 is due to ERBB1 trans-phosphorylation or reduced phosphatase activity acting upon ERBB2.
Surprisingly, despite enhanced phosphorylation of the related insulin receptor, we found that tyrosine phosphorylation of the IGF-1 receptor was not observed. However, IGF-1 receptor phosphorylation was induced by its ligand, IGF-1, arguing that this was not due to a defect in receptor function (Studer and colleagues, unpublished data). The negative finding for IGF-1 receptor phosphorylation with DCA may be due to the relatively low expression levels of the IGF-1 receptor and lower amount of receptor activation induced by bile acid compared with natural ligand—hence our ability to detect small changes in the phosphorylation of a low-abundance protein.
The tyrphostin AG1024 has been shown to inhibit the catalytic kinase domain in both the insulin and IGF-1 receptor beta chains, and both receptors share a common downstream target: IRS-1.31 The insulin receptor beta chain tyrosine phosphorylation, but not phosphorylation of the IGF-1 receptor beta chain, was enhanced by DCA, and phosphorylation of IRS-1 was blocked by the inhibitor AG1024. Identical data to that with AG1024 were obtained when a dominant negative IGF-1 receptor was expressed.25, 26 Collectively, these findings demonstrate that IRS-1 phosphorylation in response to DCA exposure is largely dependent on signaling by the insulin receptor.
With a very similar biology to that previously reported for the ERK1/2 pathway,9 bile acids caused a dose-dependent increase in AKT activity. DCA and TCA concentrations as low as 5 μM enhanced S473 phosphorylation/AKT activation. AKT activation appeared to have reached a plateau at a bile acid dose of 25 μM. Many laboratories have treated liver-derived cells with bile acid concentrations in this range without any evidence of significant cellular toxicity.4, 12, 32, 33 Both TCA (a hydrophilic bile acid) and DCA (a hydrophobic bile acid) activated AKT to a similar extent, which was also reflected in a nearly equivalent ability of both bile acids to activate GS (see below). The physiological concentration of total serum bile acids can range from approximately 10 μM to 100 μM, depending on the animal species and feeding status.34–38 In the colon, bile acid concentrations are much higher, and also presumably immediately after feeding and bile acid re-absorption in the liver. Therefore, physiological concentrations of bile acids have the potential to help modulate insulin receptor activation and signaling pathways in hepatocytes that control glucose metabolism.
Insulin has been known for decades to promote the storage of glucose as glycogen in tissues, with the liver playing a key role in regulating plasma glucose homeostasis.39, 40 As DCA activated the insulin receptor, it was tempting to speculate that DCA would also regulate the activity of GS. Studies over the last 10 years have linked insulin receptor signaling to GS via IRS-1 and the activation of the PI3 kinase/AKT/GSK3 pathway.14–16 Phosphorylation and inactivation of GSK3 by AKT results in decreased phosphorylation of sites 3a, 3b, and 3c in GS, which in turn leads to activation of GS. DCA enhanced IRS-1 tyrosine phosphorylation that correlated with increased binding of the PI3 kinase p85 subunit with IRS-1, which was associated with elevated GS activity. DCA promoted activation of GS in hepatocytes to a similar extent as hepatocytes that were either cultured for 24 hours in 50 nM insulin or treated acutely with 50 nM insulin, suggesting that the modest 40% to 50% increases in GS activity caused by DCA were physiologic. Other groups, treating primary hepatocytes with insulin in the high pM to low nM range, have observed similarly modest (30% to 50%) increases in GS activity.18–20 The ability of DCA to cause GS activation was blocked by small molecule inhibitors of the insulin receptor and PI3 kinase as well as dominant negative GSK3 and PI3 kinase proteins that interfered with signaling from the insulin receptor to GS (unpublished observation). Furthermore, expression of a dominant negative IGF-1 receptor inhibited tyrosine phosphorylation of the insulin receptor and IRS-1 and blunted DCA-induced GS activation. These findings suggest that DCA and insulin use the same signaling pathway to regulate GS activity in hepatocytes.
The relative abilities of the tyrphostin receptor inhibitors AG1478 (ERBB1) and AG1024 (insulin receptor (INS-R), IGF-1 receptor) to block AKT phosphorylation and GS activation were different. ERBB2 signaling has also been linked to enhanced PI3 kinase activity; however, the ERBB2 inhibitor AG825 weakly modified DCA-induced AKT activity, similar to our previous observations for ERK1/2 (Han and Dent, unpublished observations).11 The weaker ability of the ERBB1 inhibitor AG1478 to block GS activation compared with the insulin receptor inhibitor AG1024 suggests that DCA-induced signaling from growth factor receptors to AKT and to GS in hepatocytes is predominantly by way of the insulin receptor and IRS-1.
Despite the fact that insulin caused considerably more phosphorylation of the insulin receptor than DCA, 50 nM insulin was only marginally better than 50 μM DCA or 50 μM TCA at activating GS in a side-by-side assay. The relative activation of GS by all agents is similar to that reported by other groups using primary hepatocytes.17–20 It should be noted, however, that 50 nM insulin is approximately 200 times higher than normal plasma insulin levels40 but is within the range of insulin concentrations that are frequently used by laboratories for the culture of primary hepatocytes and other responsive cells, or for the in vitro activation of GS.17, 18, 32, 33, 41 One potential explanation for the similarity of GS activation by insulin and bile acids in primary hepatocytes may be that the amplification of signaling occurs downstream of the receptor in such a way that a relatively small upstream activation of a receptor may cause a large activation of a downstream intracellular effector (e.g., AKT in Fig. 2C and D).
Recently, De Fabiani et al. reported that bile acids could potentially inhibit hepatocyte and hepatic gluconeogenesis by reducing expression of the rate-limiting enzyme in this process: phosphoenolpyruvate carboxykinase.42 Insulin is also known to suppress both phosphoenolpyruvate carboxykinase and glucose-6-phosphatase gene expression by the activation of the phosphatidyl inositol 3 kinase (PI3K) pathway, although PI3K-independent pathways have also been shown to lead to the inhibition of gluconeogenic enzyme activities.43 Insulin can also suppress expression of cholesterol 7 alpha-hydroxylase in hepatocytes44: previous studies by our group have shown that DCA-induced activation of the FAS receptor/c-Jun NH2-terminal kinase pathway is responsible for the acute inhibition of cholesterol 7 alpha-hydroxylase levels in hepatocytes.45, 46 Our studies, combined with those of other groups, suggest that bile acid–induced activation of the INS-R/PI3K module in hepatocytes may promote both GS activation, inhibition of gluconeogenesis, and regulation of bile acid synthesis in a coordinated fashion. Additional studies will be required to prove or refute this possibility.
Thus, based on the ability of DCA and insulin to cooperate in enhancing GS activation, it is possible that bile acids may aid the liver in the storage of glucose. As digested food enters the liver by way of the portal vein, the pancreas releases insulin; this promotes catabolic processes, including glucose storage.47, 48 Bile acids such as DCA, TDCA, and TCA also re-enter the liver with the digested food. It is tempting to speculate that bile acids may thus be able to assist insulin as regulatory molecules in the control of plasma glucose-homeostatic control by the liver. The GS activating effect of a hydrophobic bile acid (DCA) was similar to that of a hydrophilic bile acid (TCA), suggesting that this effect will not be affected by bile acid binding to plasma proteins. Further studies will be required to prove whether or not bile acids and insulin cooperate to regulate GS in vivo.
The recombinant adenovirus to express dominant negative GSK3 S9A was kindly provided by Dr. M. Birnbaum (Howard Hughes Medical Institute, University of Pennsylvania School of Medicine, Philadelphia, PA). P.D. thanks Dr. M. Pandak for discussions pertaining to bile acid and insulin levels in the liver.