We have previously shown that cAMP protects against bile acid-induced apoptosis in cultured rat hepatocytes in a phosphoinositide 3-kinase (PI3K)-dependent manner. In the present studies, we investigated the mechanisms involved in this anti-apoptotic effect. Hepatocyte apoptosis induced by glycodeoxycholate (GCDC) was associated with mitochondrial depolarization, activation of caspases, the release of cytochrome c from the mitochondria, and translocation of BAX from the cytosol to the mitochondria. cAMP inhibited GCDC-induced apoptosis, caspase 3 and caspase 9 activation, and cytochrome c release in a PI3K-dependent manner. cAMP activated PI3K in p85 immunoprecipitates and resulted in PI3K-dependent activation of the survival kinase Akt. Chemical inhibition of Akt phosphorylation with SB-203580 partially blocked the protective effect of cAMP. cAMP resulted in wortmannin-independent phosphorylation of BAD and was associated with translocation of BAD from the mitochondria to the cytosol. These results suggest that GCDC-induced apoptosis in cultured rat hepatocytes proceeds through a caspase-dependent intracellular stress pathway and that the survival effect of cAMP is mediated in part by PI3K-dependent Akt activation at the level of the mitochondria.
- phosphoinositide 3-kinase
- adenosine 3′,5′-cyclic monophosphate
several studies have verified the importance of apoptotic cell death in cholestatic hepatobiliary disorders (26, 54). Hepatocyte apoptosis in cholestatic diseases is the result of the retention of endogenous cytotoxic compounds, such as bile acids, which are normally excreted in bile. Bile acids induce hepatocyte apoptosis through death receptor activation and activation of intracellular stress signals (14, 25, 38, 47, 48, 50, 51, 59,61).
Apoptosis resulting from the intracellular stress pathway is mediated by translocation of pro-apoptotic proteins in the Bcl-2 family, BAX and/or BAK, to the mitochondria (22, 23). Insertion of these proteins into the mitochondrial membrane leads to the release of cytochrome c in the cytosol. Cytosolic cytochrome c binds to the apoptosis protease activator factor (APAF1), resulting in cleavage and activation of caspase 9. Activated caspase 9 cleaves effector caspases such as caspase 3 and 7, which result in apoptosis. In isolated rat hepatocytes and in rats fed a bile acid diet, the unconjugated bile acid, deoxycholate, induces apoptosis by translocating BAX from the cytosol to the mitochondria (51, 52).
Bile acids also induce ligand-independent activation of the Fas and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) death receptors (14, 25, 38, 47). In death receptor-mediated apoptosis, assembly of a death-initiating signal complex at the plasma membrane results in proteolysis and activation of caspase 8. In type 1 cells, caspase 8 directly cleaves effector caspases, leading to apoptosis. In type II cells, such as hepatocytes, caspase 8 activates a mitochondrial amplication cascade by cleaving the cytosolic proapoptotic protein BID. Truncated BID translocates to the mitochondria, resulting in the release of cytochrome c in the cytosol.
The mechanism whereby BAX or BID results in mitochondrial release of cytochrome c is highly controversial and may differ based on cell type and apoptotic trigger (2, 8, 12, 27, 35, 36, 41,42). One mechanism involves induction of the mitochondrial membrane permeability transition (MMPT). This transition is mediated by opening of a high-conductance pore on the inner mitochondrial membrane, which results in mitochondrial swelling and outer membrane disruption. Alternatively, BAX/BID may have direct effects on outer mitochondrial membrane permeability, perhaps by mediating pore formation. Bile acids can induce the MMPT in isolated rat liver mitochondria (3,52), and inhibitors of the MMPT can inhibit bile acid-induced cell death and mitochondrial depolarization in isolated hepatocytes (3, 62). The effect of cAMP on the MMPT in hepatocytes is unknown.
In several cell types, cAMP protects against apoptosis induced by death receptors or intracellular stress (19, 32, 39, 40, 56,63). In hepatocytes, cAMP protects against apoptosis due to bile acids or activation of the tumor necrosis factor-α (TNF-α) receptor or Fas receptor (16, 31, 59). The molecular basis for the anti-apoptotic action of cAMP has not been fully characterized. Classic cAMP signaling involves activation of protein kinase A (PKA) and the cytoprotective effect of cAMP in bile acid, and TNF-α receptor-mediated apoptosis in hepatocytes is partially PKA dependent (31, 59). Recent evidence suggests that cAMP may also transduce information through phosphoinositide 3-kinase (PI3K) in hepatocytes. The anti-apoptotic effect of cAMP in bile acid-induced hepatocyte apoptosis is reversed by PI3K inhibitors (59), and cAMP stimulates PI3K-dependent translocation of membrane transporters in hepatocytes (37,60).
Accumulating evidence supports a role for PI3K as a survival kinase in hepatocytes. In rat hepatocytes, growth factor-mediated survival from bile acid or transforming growth factor-β (TGF-β) receptor-mediated apoptosis is PI3K dependent (13,61). In hepatoma cells, interleukin-6-, insulin-, and hepatocyte growth factor-mediated survival from TGF-β and Fas-mediated apoptosis is PI3K dependent (5, 6, 58). Nontoxic bile salts, such as ursodeoxycholate and taurochenodeoxycholate, activate PI3K survival pathways and can be converted to toxic bile salts by PI3K inhibition (48, 55).
In many tissues, Akt/protein kinase B is the downstream mediator of the survival effect of PI3K (10, 17). Several studies support a role for Akt in hepatocyte survival. Inhibition of growth factor-induced activation of Akt with PI3K inhibitors destroys the anti-apoptotic potential of these agents (13, 61). In hepatoma cells, transfection with constitutively active Akt prevents staurosporine and TFG-β-mediated apoptosis (6,55). Akt also protects mouse hepatocytes from TNF-α and FAS-mediated apoptosis (24). One of the survival substrates phosphorylated by Akt is the pro-apoptotic Bcl-2 protein BAD (17). Phosphorylation of BAD promotes survival, since it prevents BAD from heterodimerizing and inhibiting the anti-apoptotic protein Bcl-xL. Because cAMP stimulates PI3K-dependent Akt activation in hepatocytes (60) and is known to phosphorylate BAD in other tissues (23, 33, 64), the role of PI3K/Akt/BAD in the anti-apoptotic effect of cAMP warrants further characterization.
The goals of the present study were to verify the role of the intracellular stress pathway in apoptosis induced by hydrophobic conjugated bile acids in primary rat hepatocyte cultures, determine at what level in this stress-mediated model cAMP confers PI3K-dependent cytoprotection, and, finally, determine the role of Akt in the protective effect. We show that cAMP-mediated cytoprotection is associated with PI3K-dependent inhibition of bile acid-induced caspase 3 and 9 activation and mitochondrial release of cytochromec. cAMP phosphorylates Akt, and selective inhibition of Akt activation partially prevents the protective effect of cAMP. In addition, cAMP phosphorylates BAD and results in its translocation from the mitochondria to the cytosol, but these anti-apoptotic events are largely independent of PI3K activation.
MATERIALS AND METHODS
Cell culture and monitoring of apoptosis.
Hepatocytes were isolated from 170- to 200-g male Wistar rats by collagenase perfusion as previously described (57). Cells were plated on 0.01% type I collagen in MEM supplemented with 10% FCS and 100 ng/ml insulin, penicillin, and streptomycin. After 1 h, cultures were washed and placed in unsupplemented MEM for 3 h. Experiments to test the effect of modulators on bile acid-induced apoptosis were initiated at this point. Apoptosis was induced by the addition of 50 μM glycodeoxycholate (GCDC). Later (2 h), the amount of apoptosis was monitored morphologically with Hoechst 33258 staining, as previously described (59). Five hundred cells were counted from random fields, and the number of apoptotic cells was expressed as a percentage of the total number of cells counted. Modulators of apoptosis were added 30 min before the addition of GCDC. To observe the cytoprotective effect of cAMP on GCDC- induced apoptosis, the cell-permeable and phosphodiesterase-resistant cAMP analog 8-(4-chlorophenylthio)-cAMP (CPT-cAMP; Sigma-Aldrich Chemical, St. Louis, MO) was added 30 min before the addition of GCDC. Kinase inhibitor SB-203580 (Calbiochem, San Diego, CA), LY-294002 (40 μM), or wortmannin was added 30 min before CPT-cAMP. Appropriate concentrations of the inhibitor diluent dimethyl sulfoxide (<0.005%) were added to all tissue culture dishes.
Role of caspases.
Rat hepatocyte cultures treated for 2 h with GCDC were harvested in caspase lysis buffer [50 mM PIPES/KOH, pH = 6.5, 2 mM EDTA, 0.1% CHAPS supplemented with 5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 10 μg/ml leupeptin, aprotinin, and pepstatin], freeze-thawed three times to lyse the cells, and centrifuged at 14,000 g for 10 min. The supernatants were stored at −80°C. Protein was determined by the method of Lowry et al. (34). Lysate protein (100 μg) was separated on SDS-PAGE, and the proteins were electrophoretically transferred to nitrocellulose. Immunoblotting was performed with anti-caspase 3 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA). The presence of a cleavage product at 17 kDa was used as indication of proteolytic processing of caspase 3. Caspase activity in cells treated with GCDC for 2 h was determined with the fluorogenic substrates LEHD-7-amino-4-trifluoromethyl coumarin (AFC; for caspase 9-like activity), DEVD-AFC (for caspase 3-like activity), and IETD-AFC (for caspase 8-like activity). Cell lysates were prepared, and caspase assays were performed according to the manufacturer's instructions (Chemicon International, Temecula, CA). The assays were done at 37°C for 1 h using 50 μM substrate. For caspase inhibition studies, caspase inhibitors Z-VAD-fmk (Alexis Biochemical, San Diego, CA) or zinc sulfate were added to hepatocyte cultures 30 min before GCDC.
Cytosolic and mitochondrial fractions were prepared from rat hepatocyte cultures treated for 2 h with GCDC, as previously described (51), with slight modifications. Briefly, cells were scraped in mannitol-sucrose buffer [200 mM mannitol, 70 mM sucrose, 1 mM EGTA, and 10 mM HEPES (pH = 7.5) supplemented with 200 mM dithiothreitol, 0.1 mM PMSF, and 10 μg/ml leupeptin, pepstatin, and aprotinin] and then dounce homogenized to lyse cells. The homogenate was spun at 400 g to remove nuclear debris. The supernatant was spun again at 10,000 g, and the pellet, which represented the mitochondrial fraction, was resuspended in mannitol-sucrose buffer containing 1% Nonidet P-40 (NP-40). The supernatant from the mitochondrial spin was centrifuged at 100,000g, and the supernatant from this high-speed spin was saved as the cytosolic fraction. Purity of the fractions was determined by immunoblotting with the mitochondrial protein anti-cytochromec oxidase (Molecular Probes, Eugene, OR) and by determination of maleate dehydrogenase activity, a mitochondrial marker enzyme (11). The mitochondrial release of cytochromec was monitored by immunoblotting. Cytosolic and mitochondrial proteins (30 μg) were separated on SDS-PAGE, transferred to nitrocellulose, and immunoblotted with anti-cytochromec antibody (Santa Cruz Biotechnology). The immunolocalization of Bcl-2-like proteins in cytosolic and mitochondrial fractions was monitored by immunoblotting with antibodies to BAX, BID (Santa Cruz Biotechnology), BAD, and Bcl-xL(Cell Signaling Technology, Beverly, MA)
Bile acid accumulation.
The 30-min accumulation of the radiolabeled bile acids [3H]taurocholate and [14C]glycocholate (Perkin-Elmer Life Science, Boston, MA) was determined in rat hepatocytes as previously described (60). To assess the effect of various modulators on bile acid accumulation, the compounds were added 30 min before the addition of the radiolabeled bile acid. Bile acid accumulation was expressed as nanomoles per milligram protein per 30 min.
Akt and p38 mitogen-activated protein kinase phosphorylation.
Rat hepatocyte cultures treated with modulators for 15 min were lysed in 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, and 50 mM Tris · HCl (pH = 7.5) supplemented with 100 nM okadaic acid, 1 mM sodium orthovanadate, 1 mM PMSF, and 10 μg/ml leupeptin, pepstatin, and aprotinin. Proteins (100 μg) were separated by SDS-PAGE, transferred to nitrocellulose, and then immunoblotted with antibodies that recognize the phosphorylated (active) form of Akt (phospho-Ser473) or p38 mitogen-activated protein kinase (MAPK; phospho-Thr180, Tyr182; Cell Signaling Technology). The blots were reacted with a secondary antibody linked to horseradish peroxidase and developed with enhanced chemiluminescence (Amersham). The gels were stripped and reprobed with Akt or p38 MAPK phosphorylation state-independent antibodies to verify equal protein loading.
Rat hepatocyte cultures were treated with 100 μM CPT-cAMP for 2, 5, 15, 30, and 60 min. Cells were lysed in 50 mM HEPES, 1% NP-40, 10% glycerol, and 100 mM EGTA (pH = 7.5) supplemented with 100 μM PMSF, 5 mM sodium orthovanadate, and 10 μg/ml leupeptin, aprotinin, and pepstatin; quick-frozen in liquid nitrogen; and stored at −80°C. PI3K was immunoprecipitated from 200 μg cell lysates using 10 μl of p85 antibody attached to agarose beads (Upstate Biotechnology, Lake Placid, NY) for 2 h. Immunoprecipitates were washed one time with lysis buffer and three times with PBS and resuspended in assay buffer (25 mM HEPES, 5 mM MgCl2, and 1 mM EGTA, pH = 7.0). The assay mixture was composed of 10 μg lipids, 1:1:1 phosphatidyinositol-phosphatidylinositol 4,5-bisphosphate-phosphoserine (Avanti Polar Lipids, Alabaster, AL), and 150 μM ATP with 25 μCi [γ-32P]ATP (Perkin-Elmer Life Science). After incubation for 20 min at 37°C, the reaction was stopped with 100 μl methanol-1 N hydrochloric acid (1:1), and lipids were extracted two times with 100 μl chloroform. The lipids were loaded on thin-layer chromatography plates (Whatman Silica Gel 60) precoated with oxalate and developed in H2O-n-propranol-acetic acid (34:65:1) overnight. The lipid product, phosphatidylinositol 3,4,5-trisphosphate, was quantified by liquid scintillation counting.
Assessment of mitochondrial integrity.
Rat liver mitochondria were isolated, and induction of the MMPT was measured by a spectrophotometric assay, as described previously (3). In this assay, opening of the permeability transition pore is monitored by recording a decrease in light scattering of a mitochondrial suspension. This decrease in light scattering is indicative of the rapid high-amplitude mitochondrial swelling that accompanies the MMPT and can be monitored by a decrease in absorbance at 540 nm. Rat liver mitochondria (1 mg/ml) were treated with 1 mM glutamate and 1 mM maleate at time 0. Later (3 min), 5 μM rotenone was added. After this (2 min), varying concentrations of GCDC were added, and optical density was monitored for 10 min using a SpectraMax Plus spectrophotometer (Molecular Devices) at 25°C. Where indicated, mitochondrial suspensions were pretreated with 100 μM CPT-cAMP or 5 μM cyclosporine (CYA) for 5 min before the start of the assay.
Mitochondrial energization was determined by the retention of the dye 3,3′-dihexyloxacarbocyanine [DiOC6(3); Molecular Probes]. Hepatocyte cultures were treated sequentially with 100 μM CPT-cAMP or 5 μM CYA (30 min) and then with 50 μM GCDC for 2 h. During the last 30 min of incubation, cells were loaded with 100 nM DiOC6(3). After being washed with PBS, the cells were lysed in distilled water. The concentration of retained DiOC6(3) was measured with a Hitachi F-2000 fluorescence spectrophotometer at an excitation of 484 nm and emission of 500 nm.
Metabolic labeling of hepatocytes and immunoprecipitation of BAD.
Hepatocytes were grown in phosphate-free media and labeled for 1 h with [32P]orthophosphate (300 μCi/ml; Perkin-Elmer Life Sciences). After treatment with 100 μM CPT-cAMP for 15 min, whole cell lysates were prepared, and BAD was immunoprecipitated (Santa Cruz Biotechnology). After capture with protein A-agarose, the immunoprecipitates were washed, solubilized, and subjected to SDS-PAGE. The proteins were transferred to nitrocellulose, and the blot was directly exposed to radiographic film.
cAMP activates PI3K.
We have previously shown that cAMP protects cultured rat hepatocytes against GCDC-induced apoptosis and that this protection is PI3K dependent (59). Treatment of rat hepatocyte cultures with 100 μM CPT-cAMP resulted in a statistically significant increase in PI3K activity in p85 immunoprecipitates at 5, 15, 30, and 60 min (Fig.1).
cAMP prevents GCDC-induced caspase activaton.
To determine at what level cAMP exerts its PI3K-dependent protective effect, we evaluated the effect of cAMP on GCDC-induced caspase activation. GCDC-induced apoptosis was completely inhibited by the broad-spectrum caspase inhibitors Z-VAD-fluoromethylketone (fmk) or zinc (Fig.2 A). GCDC-induced apoptosis was associated with a time-dependent increase in caspase 3-like, caspase 9-like, and caspase 8-like activity, as determined using the substrates DEVD-AFC, LEHD-AFC, and IETD-AFC, respectively (Fig. 2 B). Pretreatment (30 min) with CPT-AMP partially inhibited GCDC-induced caspase 3-like and caspase 9-like activity but not caspase 8-like activity. cAMP-mediated inhibition of caspase 3 and 9 activity was prevented by the PI3K inhibitor wortmannin. We were also able to demonstrate catalytic processing of caspase 3 in immunoblots of crude cell lysates treated for 1 h with GCDC (Fig. 2 C). GCDC-induced proteolytic cleavage of caspase 3 was prevented by pretreatment with CPT-cAMP, and this inhibition of caspase processing was prevented by wortmannin. Collectively, these results show that GCDC-induced apoptosis is caspase dependent and that the protective effect of cAMP is the result of inhibition of an event upstream of caspase 3 and 9 activation but downstream of caspase 8 activation.
cAMP prevents GCDC-induced release of cytochrome c.
One of the events immediately upstream of caspase 9 activation in apoptosis is the release of cytochrome c from the mitochondria. We performed subcellular fractionation of hepatocytes treated with GCDC with and without pretreatment with CPT-cAMP and/or wortmannin. Immunoblots of cytosolic fractions showed that GCDC-induced apoptosis was associated with the release of cytochromec into the cytoplasm and that CPT-cAMP prevented the release (Fig. 3 A). Pretreatment with wortmannin prevented cAMP-induced inhibition of cytochrome crelease. These results suggest that cAMP exerts its protective effect before the mitochondrial release of cytochrome c. To verify the purity of our subcellular fractions, we performed two experiments. First we determined the amount of mitochondrial-specific maleate dehydrogenase activity in our preparations. Maleate dehydrogenase activity was enriched 11.8-fold in mitochondrial fractions and decreased to 0.13-fold in cytosolic preparations compared with activity in the original homogenates. Second, we immunoblotted cytosolic and mitochondria fractions with antibody against the mitochondrial-specific protein anti-cytochrome-c oxidase. Figure 3 Bshows that the expression of this protein was confined to the mitochondrial fractions. Even upon prolonged exposure (10 min) the cytosolic fractions failed to show any immunoreactivity to this protein (data not shown).
GCDC induces the MMPT in isolated mitochondria and results in mitochondrial depolarization in intact cells.
One proposed mechanism to account for mitochondrial release of cytochrome c during apoptosis involves induction of the MMPT. Because bile acids are known inducers of the MMPT (3), we determined the role of the MMPT in cAMP-mediated protection from GCDC-induced apoptosis. In isolated mitochondria, GCDC caused a dose-dependent induction of the MMPT, as monitored by a decrease in light absorbance at 540 nm (Fig.4). Pretreatment with 5 μM CYA, a specific inhibitor of the MMPT, partially inhibited GCDC (250 μM)-induced mitochondrial swelling (42 ± 0.35% inhibition of GCDC-induced swelling), whereas pretreatment with 100 μM cAMP had no effect (Fig. 4). These results implied that cAMP had no direct effect on mitochondria to prevent GCDC MMPT induction but did not exclude the possibility that cAMP might act on a cytosolic factor that subsequently confers protection. We were unable to determine if coincubation of cAMP with a 100,000-g cytosolic fraction might confer protection against GCDC-induced mitochondrial swelling because addition of the cytosolic extract itself inhibited the GCDC-induced MMPT (data not shown). Thus, to assess the effect of cAMP on mitochondrial integrity in intact cells, we used retention of the potential dependent dye DiOC6(3). Because mitochondrial depolarization is a consequence of the MMPT in intact cells, loss of DiOC6(3) retention can be used as indirect evidence for the occurrence of the MMPT. We also used this system to determine if prolonged exposure to 50 μM GCDC, the dose used to induce hepatocyte apoptosis in these studies, might effect mitochondrial membrane potential in intact cells, since this dose did not appear to induce the MMPT, as monitored by mitochondrial swelling (Fig. 4). Permeability changes may occur in a nonsynchronous manner in mitochondrial populations so that permeability transition pores may open and then return to a closed state after a period of time. Such transient openings may not be detected by light-scattering techniques (42, 43). Treatment of cultured hepatocytes with 50 μM GCDC for 2 h caused a decrease in the retention of DiOC6(3) (56 ± 4.4% compared with control). Pretreatment with cAMP for 30 min partially prevented the GCDC-induced loss of DiOC6(3) retention (76.5 ± 1.6% of control values); however, pretreatment with CYA was without effect (51.8 ± 7.5% of control values). In accordance with the inability of CYA to prevent mitochondrial depolarization in intact cells, pretreatment with 5 μM CYA had no effect on GCDC-induced apoptosis (21.3 ± 2.5 and 17.6 ± 3.5% apoptosis with and without CYA pretreatment, respectively) and did prevent GCDC-induced mitochondrial release of cytochrome c (Fig. 3 A). Pretreatment with another MMPT inhibitor (5 μM decylubiquinone; see Ref. 27) for 30 min also failed to prevent GCDC-induced mitochondrial depolarization or apoptosis (data not shown). We were unable to document any effect of CYA or decylubiquinone on the accumulation of radiolabeled bile acid in cultured hepatocytes (data not shown).
Collectively these results suggest that GCDC-induced hepatocyte apoptosis, mitochondrial cytochrome c release, and mitochondrial depolarization may occur independently of the MMPT and that cAMP may protect against GCDC-induced apoptosis by preventing mitochondrial depolarization.
GCDC-induced apoptosis is associated with translocation of BAX to the mitochondria.
One of the proposed mechanisms to account for mitochondrial cytochromec release during apoptosis is translocation of BAX to the mitochondrial membrane. Previous studies in hepatocytes have shown that apoptosis induced by the unconjugated bile acid deoxycholate is associated with translocation of BAX to the mitochondria (51, 52). We immunoblotted mitochondrial fractions from cells treated with GCDC for 2 h for the expression of BAX. These studies demonstrated that GCDC induced the movement of BAX to the mitochondria (Fig. 5). Pretreatment with CPT-cAMP or CYA did not prevent this translocation. GCDC, cAMP, and/or wortmannin had no effect on the total BAX expression in whole cell lysates (data not shown). These results suggest that the protective effect of cAMP occurs at a step after BAX translocation or alternatively that cAMP alters the function of BAX in the mitochondrial membrane.
cAMP activates the prosurvival kinase Akt.
In several cell types, PI3K-dependent activation of Akt promotes survival. We have previously shown that cAMP can activate Akt in isolated hepatocytes (59). We extend these observations to show that CPT-cAMP activates Akt in primary cultures of hepatocytes and show that this activation is inhibited by two mechanistically different inhibitors of PI3K (wortmannin and LY-294002; Fig.6 A). Activation is maximal at 15 min and persists for up to 120 min (data not shown). To determine if Akt represented the wortmannin-dependent step in mediating cAMP protection from GCDC-induced apoptosis, we used SB-203580 as an inhibitor of Akt. SB-203580 was originally identified as an inhibitor of p38 MAPK, but studies have shown that, at higher concentrations (>10 μM), this inhibitor blocks Akt phosphorylation (30). We have been able to verify these results in cultured hepatocytes. At a concentration of 1 μM, SB-203580 inhibits basal p38 MAPK phosphorylation (Fig. 6 C) but is without effect on Akt (Fig. 6 B). At 50 μM, SB-203580 inhibits both basal p38 MAPK and Akt activity in rat hepatocytes. CPT-cAMP results in a small but significant increase in p38 MAPK phosphorylation that is inhibited completely by 1 μM SB-203508 (Fig. 6 C). CPT-cAMP-induced phosphorylation of Akt is unaffected by 1 μM SB-203580 but is completely abolished by a concentration of 50 μM (Fig. 6 B).
We next used SB-203580 to selectively inhibit p38 MAPK (1 μM) or Akt (50 μM) in hepatocyte cultures to determine the role of these kinases in GCDC-induced apoptosis and the anti-apoptotic effect of cAMP (Fig. 7). At 1 μM, SB-203580 had no effect on GCDC-induced apoptosis or the cytoprotective effect of cAMP. At 50 μM, SB-203580 mildly increased GCDC-induced apoptosis and partially reversed the protective effect of CPT-cAMP. These results suggest that cAMP-mediated activation of Akt is necessary for the anti-apoptotic effect of cAMP and are consistent with the role of Akt as a survival kinase in hepatocytes.
Previous studies have shown that SB-203580-induced inhibition of Akt is associated with inhibition of the upstream kinase PDK1 and not with a direct effect on PI3K activity (30). We have verified that SB-203580 (50 μM) has no effect on PI3K activity in hepatocytes (data not shown). In addition, SB-203580 (1 or 50 μM) does not influence bile acid-induced apoptosis by altering bile acid uptake in hepatocytes, since this inhibitor had no effect on the accumulation of radiolabeled bile acids in hepatocytes. The 30-min accumulation of [3H]taurocholate in control hepatocytes and hepatocytes treated with 1 and 50 μM SB-203580 was 10.0 ± 1.9 nmol/mg protein, 10.3 ± 0.35 nmol/mg, and 11.0 ± 1.3 pmol/mg protein, respectively.
Effect of cAMP on BAD and Bcl-xL.
BAD is a proapoptotic member of the Bcl-2 family that localizes to the outer mitochondrial membrane by phosphorylation-dependent heterodimerization with the anti-apoptotic proteins Bcl-xL or Bcl-2. When phosphorylated, BAD is unable to bind to and inhibit the function of these proteins but instead translocates to the cytosol and complexes with the cytosolic-binding protein 14–3-3 (23). The survival effect of Akt has been linked to BAD phosphorylation (10). We sought to determine if PI3K- and/or Akt-dependent phosphorylation of BAD might be involved in cAMP cytoprotection. In metabolically labeled cells, treatment with cAMP resulted in phosphorylation of BAD (Fig.8 A). Pretreatment with wortmannin, however, only marginally inhibited cAMP-stimulated BAD phosphorylation. To evaluate whether cAMP-induced phosphorylation was associated with translocation of BAD from the mitochondria to the cytosol, we performed immunoblots on mitochondrial and cytosolic fractions (Fig. 8 B). In control cells, BAD was primarily found in the mitochondria, and there was little change in localization of the protein after treatment of hepatocytes with GCDC for 2 h. Pretreatment with cAMP, however, resulted in wortmannin-independent translocation of BAD to the cytosol.
We next looked at the subcellular distribution of the anti-apoptotic protein Bcl-xL. In control hepatocytes, Bcl-xL was primarily localized to the mitochondria (Fig.8 C). Upon GCDC treatment, there was a slight decrease in the amount of mitochondrial Bcl-xL with a concomitant increase in the amount of this protein in cytosolic preparations. Pretreatment with cAMP resulted in a further decrease in mitochondrial expression and an increase in cytosolic expression of Bcl-xL. cAMP caused no change in total BAD or Bcl-xL expression in whole cell lysates (data not shown).
cAMP cytoprotection does not require new protein synthesis and is not associated with interference with bile acid uptake.
To verify that the observed effects of cAMP and wortmannin on apoptosis were not influenced by alterations in bile acid uptake, we determined the accumulation of radiolabeled [14C]glycocholate and [3H]taurocholate in cells treated with cAMP alone or in combination with wortmannin (Fig.9). cAMP and wortmannin had no effect on the 30-min accumulation of taurocholate or glycocholate in rat hepatocytes.
To determine the need for macromolecular synthesis in the survival effect of cAMP, we pretreated cultures with 20 μM cycloheximide 30 min before sequential treatment with CPT-cAMP and GCDC and determined the amount of apoptosis 2 h later. Inhibition of protein synthesis had no effect on GCDC-induced apoptosis or on the anti-apoptotic effect of cAMP (data not shown). In addition, cyclohexamide had no effect on the uptake of radiolabeled bile acids in hepatocytes (Fig. 9).
In this study, we demonstrate that hepatocyte apoptosis induced by toxic conjugated bile salts is associated with activation of caspases, MMPT-independent release of cytochromec from the mitochondria, and translocation of BAX from the cytosol to the mitochondria. cAMP mediates PI3K-dependent protection against bile acid-induced apoptosis by phosphorylating Akt and inhibiting GCDC-induced caspase 3 and 9 activation and cytochrome c release but has no effect on GCDC-induced mitochondrial BAX translocation or caspase 8 activation. Collectively, these observations imply that the survival effect of cAMP is mediated in part by PI3K-dependent Akt phosphorylation at the level of the mitochondria. We also provide evidence that cAMP may transduce PI3K-independent survival signals by regulating the phosphorylation and intracellular localization of the pro-apoptotic Bcl-2 protein BAD.
Results of the present study suggest that cAMP-mediated activation of PI3K/Akt signaling is necessary for its anti-apoptotic action. This is supported by our observations that cAMP produces a sustained activation of PI3K in hepatocytes and that inhibition of PI3K or specific inhibition of Akt inhibits the anti-apoptotic effect of cAMP. cAMP has previously been shown to result in the PI3K-dependent activation of Akt in hepatocytes (58), COS-7 cells (15), and thyroid cells (49), but this is the first report to link cAMP-stimulated Akt phosphorylation to cell survival. Growth factor stimulation of the PI3K-Akt pathway, however, is strongly correlated with cell survival from a variety of apoptotic stimuli in hepatocytes (5, 6, 13, 58, 61). Our inability to detect cAMP-mediated activation of PI3K in an earlier report (59) may be because of the less-sensitive whole cell lysate assay technique used in these former studies. This technique requires dilution of detergent-solubilized whole cell lysates and permits only nanogram amounts of protein to be evaluated. In comparison with the immune complex assay used in this present report, PI3K activity was determined in a microgram quantity of whole cell lysate.
Our findings that bile acids induce the release of cytochromec from hepatic mitochondria and that prevention of this release with cAMP protects against apoptosis confirm the importance of cytochrome c release in the pathogenesis of bile acid-induced hepatocyte apoptosis. The exact mechanisms that control the release of cytochrome c from the mitochondria are not fully understood. Experimental evidence suggests that mitochondrial release of cytochrome c may occur either as a consequence of osmotic rupture of the outer mitochondrial membrane rupture secondary to the occurrence of the MMPT (27, 41,42) or through channels formed in the outer mitochondrial membrane by proapoptotic Bcl-2 proteins, such as BAX, BID, or BAK (2, 12, 22, 23, 35, 36, 45).
Bile acids are known inducers of the MMPT (3, 50, 52). Our results in cultured rat hepatocytes suggest that, although GCDC can induce the MMPT in isolated mitochondria, it is unlikely that the MMPT plays a major role in GCDC-induced apoptosis in intact cells. This conclusion is supported by our findings that MMPT inhibitors, such as CYA, were unable to inhibit GCDC-induced apoptosis or cytochrome c release in intact cells. In addition, we were unable to demonstrate that exposure of isolated mitochondria to GCDC at concentrations used to induce apoptosis in intact cells (50 μM) was able to cause significant mitochondrial swelling. We were able to demonstrate that longer exposures to 50 μM GCDC in intact cells resulted in mitochondrial depolarization. This depolarization was not prevented by MMPT inhibitors and therefore is unlikely to be the result of the MMPT. cAMP, which did not prevent the MMPT in isolated mitochondria, however, was capable of preventing GCDC-induced mitochondrial depolarization.
Our observations on the inability of CYA to protect against bile acid-induced cell death disagree with previous studies (3,62). This discrepancy may be explained by the different cell systems studies. Previous studies in which CYA inhibited GCDC-induced necrosis or apoptosis were done in suspensions of rat hepatocytes. Our studies were done on cultured rat hepatocytes attached to type 1 collagen. It is well accepted that cell attachment to extracellular matrixes has profound implications on the control of gene expression and cell survival (4, 20).
Our studies suggest that cAMP does not produce its anti-apoptotic effect by preventing GCDC-induced translocation of BAX to the mitochondria but provide additional support for the importance of this facet of the intracellular stress signaling pathway in bile acid hepatotoxicity. Previous studies have shown that apoptosis induced by bile acids in isolated hepatocytes or by bile acid feeding of rats is accompanied by BAX translocation to the mitochondria (51, 52). In addition, in bile duct-ligated Fas-deficientlpr mice, chronic exposure to biliary toxins results in apoptotic cell death that is associated with mitochondria translocation of BAX (38). BAX-induced apoptosis in these mice may represent an alternate apoptotic pathway, since bile acids would not be capable of inducing Fas-mediated apoptosis. Because primary rat hepatocyte cultures are known to be relatively resistant to Fas-mediated apoptosis (16) as well, it may be that BAX-mediated intracellular stress pathways predominate over death receptor pathways in our system.
The observations that cAMP inhibits GCDC-induced mitochondrial cytochrome c release and mitochondrial depolarization but does not prevent BAX translocation or caspase 8 activation suggest that cAMP protection occurs at the level of the mitochondria. Studies in neurons (46), fibroblasts (63), and hepatocytes (13) have demonstrated that cAMP can inhibit the mitochondrial release of cytochrome c after growth factor withdrawal, protein synthesis inhibition, or exposure to TNF-α, respectively. In the neuronal model, similar to our results in hepatocytes, cAMP failed to prevent the translocation of BAX (46). These studies suggest that the maintenance of mitochondrial integrity by cAMP may be a general mechanism whereby cAMP exerts its survival effect in many cells in response to a variety of apoptotic stimuli.
Accumulating evidence suggests that PI3K/Akt prevents mitochondrial cytochrome c release by indirect effects on cellular metabolism that maintain mitochondrial integrity (28). Akt can phosphorylate proteins involved in intermediary metabolism, including glycogen synthase kinase-3β (GSK-β) and 6-phosphofructo-2-kinase (10). Overexpression of active GSK-β can induce apoptosis, whereas inhibition ameliorates apoptosis (7, 9). The survival effect of Akt has also been linked to increased glycolytic activity, glucose transporter expression, and Akt-stimulated mitochondrial hexokinase activity (21, 44). Study of the link between cAMP control of intermediary metabolism in hepatocytes and mitochondrial integrity is ongoing in our laboratory.
We have demonstrated that cAMP results in phosphorylation and translocation of BAD from the mitochondria to the cytosol. BAD is a pro-apoptotic member of the Bcl-2 family that can heterodimerize with and inhibit the function of anti-apoptotic Bcl-2 members, such as Bcl-xL. Although cAMP causes a slight decrease in the mitochondrial expression of Bcl-xL, the net effect of cAMP treatment in our studies is a dramatic decrease in the mitochondrial ratio BAD/Bcl-xL. Because Bcl-xL is known to prevent BAX-mediated cytochrome c release (23), such a shift would promote maintenance of mitochondrial integrity. cAMP-mediated BAD phosphorylation occurs in other cells and is mediated by PKA-dependent phosphorylation at Ser155 (33,64), although BAD can also be phosphorylated at Ser136 by PI3K/Akt (10, 17). We find that wortmannin resulted in only minor inhibition of cAMP-stimulated BAD phosphorylation and had no effect on BAD translocation. These results suggest that the effects of cAMP on BAD are largely PI3K independent in hepatocytes. The mechanisms whereby cAMP controls BAD phosphorylation and the implications of this phosphorylation on bile acid-induced apoptosis will require further characterization.
A previous study in primary rat hepatocyte cultures failed to demonstrate a role for Akt in the survival effect of cAMP in TNF-α-mediated apoptosis (31). In that study, cAMP stimulation of Akt was PI3K dependent, but wortmannin had no effect on the cytoprotective effect of cAMP. The difference between this study and our present work may be explained by two factors. The TNF-α study used 24-h hepatocyte cultures, and it is well known that hepatocytes dedifferentiate with time in culture. Because the effects of PI3K/Akt may be cell stage specific (53), the differing results may represent the varying metabolic state of 4-h cultures compared with 24-h cultures. Alternatively, the difference may reflect the relative contribution of PKA and PI3K. Because cAMP cytoprotection is both PKA and PI3K dependent, it may be that PKA effects predominate at longer culture times.
In summary, our studies show that the anti-apoptotic effect of cAMP in GCDC-induced hepatocyte apoptosis is mediated by stabilization of mitochondrial function through PI3K- and/or Akt-dependent inhibition of caspase activation and release of mitochondrial cytochrome c and is likely promoted by phosphorylation and cytosolic translocation of the mitochondrial pro-apoptotic protein BAD. These results could have important implications in the development of therapeutic strategies to target specific cAMP-mediated survival signaling events in the treatment of hepatic disorders. In vivo results in rodent models have already suggested that cAMP-elevating agents can suppress toxic liver damage (1, 18).
We thank Irwin Arias, Lyuba Varticovsky, and Tatehiro Kagawa for helpful discussions.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-33436 (M. S. Anwer) and DK-02721 (C. R. L. Webster).
Address for reprint requests and other correspondence: C. R. L. Webster, Tufts Univ., 200 Westboro Rd., North Grafton, MA 01536 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
May 10, 2002;10.1152/ajpgi.00410.2001
- Copyright © 2002 the American Physiological Society