Reversibility of Caspase Activation and Its Role during Glycochenodeoxycholate-induced Hepatocyte Apoptosis*

The accumulation of glycochenodeoxycholate (GCDC) induced hepatocyte apoptosis in cholestasis. However, many hepatocytes still survived GCDC-induced apoptosis. The molecular mechanism for the survival of hepatocytes remains unclear. In the present study, isolated rat hepatocytes were cultured in William's E medium and treated with 50 μm GCDC. DNA, RNA, cell lysate, and nuclear proteins were collected at different intervals for DNA fragmentation assay, reverse transcription PCR, Western blotting, and gel mobility shift assay, respectively. GCDC-induced active caspases were detected as early as 2 h by Western blotting and kinetic caspase assay, whereas hepatocyte apoptosis was found at 4 h by DNA fragmentation and terminal deoxynucleotidyl transferase-mediated dUPT nick-end labeling assay. When GCDC was removed, the increased caspases as well as NF-κB could be restored to control level. A1/Bfl-1 and inducible nitric oxide synthase (iNOS) were up-regulated in 2 h of GCDC stimulation. After GCDC was removed, hepatocytes decreased expression of A1/Bfl-1, but not iNOS, to the control level. NF-κB activation coincided with the change of A1/Bfl-1. Survivin, cIAP1, cIAP2, XIAP, and A1/Bfl-1, but not iNOS, were down-regulated by pan-caspase inhibitor benzyloxycarbonyl-VAD-fluoromethyl ketone. In addition, benzyloxycarbonyl-VAD-fluoromethyl ketone inhibited release of cytochrome c and suppressed NF-κB activation. Our data suggested that caspase pathway is an important regulatory factor during hepatocyte apoptosis. GCDC-induced caspase response is reversible, which may activate anti-apoptotic genes to protect hepatocytes from apoptosis.

Bile acids are accumulated in cholestasis (1). The major bile acids include glycocholate and taurocholate, which are crucial for digestion and absorption of fat (2). Pathophysiologically, ursodeoxycholate and its taurine-conjugated form are protective to hepatocytes by directly suppressing disruption of mitochondrial membrane structure (3), whereas glycochenodeoxycholate (GCDC) 1 is cytotoxic and contributes to hepatocellular injury (4). GCDC may cause hepatocyte damage by both acute necrosis and chronic apoptosis (5,6). GCDC-induced hepatocyte apoptosis includes death receptor and mitochondrial pathways (7,8). Moreover, two apoptotic pathways are mediated by caspase cascade activation (9). Caspases comprise a unique family of cysteine proteases involved in cytokine activation and in the execution of apoptosis (10).
Apoptosis, a cellular suicide program, is essential for correct development and homeostasis of multicellular organisms (11). Caspases are central initiators and executioners of apoptosis (10). Once activated, the caspases are responsible for cleavage of selective protein substrates that have been widely implicated in many models of apoptosis (11,12). In particular, caspase 8 has been found to be a prominent signaling caspase involved in initiation of apoptosis by the Fas, tumor necrosis factor type I, and DR3 receptors (13,14). Caspase 8 is abundant in the liver, allowing caspase to function as an initiator protease upstream of cathepsin B in bile salt-mediated hepatocyte apoptosis (15). Apoptosis is generally considered a fast process, because activation of caspases rapidly is followed by cell fragmentation and phagocytosis. However, different cells have distinct variances for caspase activation. Hepatocytes are sensitive to the activation of caspase to cause apoptosis, but neurons show a relative resistance that may allow a low level of caspase activity (16,17). There are many anti-apoptotic genes, such as cIAP 1 and 2, TRAFs 1 and 2, IEX-1L, and A1/Bfl-1 (18 -20). Some genes play the antiapoptotic role by direct activation, not by caspase pathway (21). The relationship between anti-apoptotic genes and caspase activation is not yet clearly understood.
We took advantage of the fact that caspases were activated by GCDC to examine the apoptotic mechanism of caspase cascade activation (7). The present study was designed to investigate the role of caspases in GCDC-induced hepatocyte apoptosis. Pan-caspase inhibitor Z-VAD-FMK was utilized to further elucidate the direct relationship between antiapoptotic genes and GCDC-induced caspase response. Our data suggest GCDCinduced hepatocyte apoptosis is modulated by caspase cascade activation. GCDC-induced caspase response is reversible, which may activate anti-apoptotic genes to protect hepatocytes from apoptosis.

EXPERIMENTAL PROCEDURES
Materials-William's E medium and fetal calf serum were purchased from Invitrogen. Goat polyclonal antibody against Survivin, donkey anti-goat horseradish peroxidase, and mouse anti-␤-actin were obtained from Santa Cruz Biotechnology. Rabbit anti-mouse horseradish peroxidase was acquired from Cell Signaling Technology. All other chemicals were purchased from Sigma except as indicated.
Cell Culture-Primary hepatocytes were isolated from adult Sprague-Dawley rats by standard liver perfusion procedure (22). Dead cells were removed by Percoll (Sigma) and primary hepatocytes cultured in * This work was supported by a Falk Medical Research Foundation grant. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 1 The abbreviations used are: GCDC, glycochenodeoxycholate; Z-VAD-FMK, benzyloxycarbonyl-VAD-fluoromethyl ketone; RT, reverse transcription; TUNEL, terminal deoxynucleotidyl transferase-collagen-coated dishes (Falcon). Cells at a density of 5.5 ϫ 10 5 were incubated in William's E medium containing 10% fetal bovine serum for varying intervals.
Caspase Assay-Cells were grown in 35-mm dishes and harvested with cell lysis buffer. Protein concentration of cell lysate was determined with the bicinchoninic acid assay method (Pierce). 100 g of cell lysate was utilized to assay the activities of caspase-3 or caspase-8. The caspase assay kit was purchased from Calbiochem. The reaction system employed the colorimetric substrate IETD-pNA and calculated the activity as pmol/min.
Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) and Real-time Quantitative PCR-The RT-PCR kit was obtained from Qiagen. After reverse transcription, the cDNA product was amplified by PCR with Taq DNA polymerase using standard protocols. The amplified products (10 l) were separated on 2% agarose gels, stained with ethidium bromide, and photographed using ultraviolet illumination. The 5Ј forward and 3Ј reverse complement PCR primers for amplification of Survivin were CTGATTTGGCCCAGTGTTTT and TCATCTGA-CGTCCAGTTTCG, respectively. For cIAP2, PCR primers were ACAT-TTCCCCAGCTGCCCATTC and CTCCTGCTCCGTCTGCTCCTCT. For cIAP1, PCR primers were CCAGCCTGCCCTCAAACCCTCT and GG-GTCATCTCCGGGTTCCCAAC. For XIAP, PCR primers were CGCGA-GCGGGGTTTCTCTACAC and ACCAGGCACGGTCACAGGGTTC. For A1/Bfl-1, PCR primers were ATCCACTCCCTGGCTGAGAACT and AC-ATCCAGGCCAATCTGCTCTT. For iNOS, PCR primers were CGAGG-AGGCTGCCTGCAGACTGG and CTGGGAGGAGCTGATGGAGTA-GTA. For glyceraldehyde-3-phosphate dehydrogenase, PCR primers were CCATCACCATCTTCCAGGAG and CCTGCTCACCACCTTCTTG. All PCR primers were synthesized from Integrated DNA Technology. The relative mRNA levels of Survivin were confirmed by real-time PCR. The data were normalized to the expression level of 18 S rRNA.
TUNEL Assay-The TdT-FragEL TM DNA fragmentation detection kit was obtained from Oncogene Research Products, and instructions were followed for preparation of tissue slides. In this assay, terminal deoxynucleotidyl transferase (TdT) bound to exposed 3Ј-OH ends of DNA fragments generated in response to apoptotic signals and catalyzed the addition of biotin-labeled and unlabeled deoxynucleotides. Biotinylated nucleotides were detected using a streptavidin-horseradish peroxidase conjugate. Diaminobenzidine reacted with the labeled sample to generate an insoluble colored substrate at the site of DNA fragmentation. Counterstaining with methyl green aided in the morphological evaluation and characterization of normal and apoptotic cells.
Western Blotting-Samples were resolved by 10% SDS-polyacrylamide gel electrophoresis, transferred to Immobilon-P membrane (Millipore), and blotted with appropriate primary antibodies at a dilution of 1:1,000. Peroxidase-conjugated secondary antibodies were incubated at a dilution of 1:3,000. Bound antibody was visualized using chemiluminescent substrate (ECL; Amersham Biosciences) and exposed to Kodak X-Omat film. At least three independent experiments were performed.
Gel Mobility Shift Assay-Nuclear extract was prepared with the modified Dignam protocol. Concentration of nuclear protein was measured with the Bio-Rad protein assay. 4 g of nuclear proteins and 2 g of the nonspecific competitor poly(dI⅐dC) were incubated in binding buffer (100 mM Hepes, pH 7.6, 5 mM EDTA, 50 mM (NH4)2 S04, 5 mM dithiothreitol, Tween 20, 1% (w/v), 150 mM KCl) with 20 fmol/l doublestranded DNA oligonucleotide. The NF-B consensus binding sequence (5Ј-AGT TGA GGG GAC TTT CCC AGG C-3Ј) was labeled with digoxin. Binding reactions were performed by incubating the samples for 15 min at 22°C. Protein-DNA complexes were separated from the unbound DNA probe by electrophoresis through 6% native polyacrylamide gels containing 0.5ϫ Tris borate/EDTA. The gel was transferred to Nytran membrane and exposed to Kodak films. At least three independent experiments were performed.
Statistical Analysis-All data represent at least three experiments using cells or extract from a minimum of three separate isolations and are expressed as means Ϯ S.D. unless otherwise indicated. Differences between groups were compared using analysis of variance for repeated measures. All statistical analyses were performed with the software SPSS.

GCDC-induced Hepatocyte
Apoptosis-GCDC-induced hepatocyte apoptosis could be reflected by DNA fragmentation assay, caspase assay, and Western blotting (23)(24)(25). In the present study, rat hepatocytes were isolated and cultured in collagen-coated dishes with William's E medium containing 10% fetal bovine serum overnight. After stimulation with a concentration of 50 M GCDC at different intervals, hepatocytes were harvested. DNA, RNA, and cell lysate protein were separately collected for further analysis. DNA fragmentation demonstrated hepatocyte apoptosis after treatment by 50 M GCDC for 16 h (Fig. 1A). Caspases are divided into long prodomain caspases (caspase-2, -8, -9, and -10) and short prodomain caspases (caspase-3, -6, and -7) (26). In our study, caspase-8 and caspase-3 were selected as indicators to examine their involvement during GCDC-induced apoptosis. After 16 h, the activities of both caspase-3 and caspase-8 were increased. Moreover, the extent of activated caspase-3 was higher than that of activated caspase-8 (Fig. 1B). When the stimulation duration was decreased to 4 h in the presence of 50 M GCDC, hepatocyte apoptosis was still found by DNA fragmentation assay ( Fig. 2A). At 4 h, the anti-apoptotic gene Survivin was up-regulated (Fig. 2B). By Western blotting with crude cell lysate, the expression of caspase-3 and caspase-8 protein was initially increased at 2 h (Fig. 2C). Caspase activation was earlier than DNA fragmentation. The relative mRNA levels of Survivin were further confirmed by real-time PCR (Fig. 2D) and were identical to the level of Survivin protein by Western blotting (data not shown).
Reversibility of GCDC-induced Caspase Activation-With 50 M GCDC, caspase was activated at 2 h but DNA fragmentation was enhanced at 4 h. These results suggested GCDCinduced hepatocyte apoptosis was duration-dependent. To address this issue, the cultured primary hepatocytes were GCDC (Fig. 3, A and B). After removal of GCDC, no apoptosis was detected by DNA fragmentation. The mechanism is unknown (Fig. 3C). TUNEL assay showed no difference in apoptotic rate among the four groups (Fig. 3D). However, apoptotic rate was increased significantly when GCDC was not removed (Fig. 4). NF-B was activated after 2 h of GCDC stimulation, restored to control level after washing GCDC away, and reactivated by an additional 2 h of restimulation with 50 M GCDC (Fig. 3E). Survivin expression was unchanged (Fig. 3F). The data suggested that caspase response and NF-B activation were reversible in a short treatment (2 h) by 50 M GCDC. Under low concentration GCDC (50 M), the duration of GCDC treatment plays a crucial role in hepatocyte apoptosis.
GCDC-induced Gene Expression-After 50 M GCDC stimulation, some cells died of apoptosis and floated in supernatant (Fig. 4A) but others attached to the dish and were alive (Fig.  4B). It was unclear whether anti-apoptotic genes were upregulated and accounted for cell survival. Therefore, RNA was isolated from cells and RT-PCR was performed to detect gene transcription. In the anti-apoptotic gene family we selected cIAP1, cIAP2, XIAP, A1/Bfl-1, iNOS, and Survivin. There were no changes among cIAP1, cIAP2, XIAP, and Survivin (Fig. 5A), but results were different for A1/Bfl-1 and iNOS. The A1/Bfl-1 gene was increased at 2 h and then decreased to the baseline level in another 4 h (without GCDC) of extended culture. iNOS showed higher amplification at 2 h but maintained at the same level in another 4 h (without GCDC) of culture. To demonstrate gene expression from bottom-attached cells, we washed away the top dead cells, isolated RNA from bottom cells, and then repeated RT-PCR. Expression of genes from dish-attached cells was completely identical (data not shown). NF-B activation was also increased at 2 h and then restored to the baseline level in another 4 h (without GCDC) of culture (Fig. 5B). Survivin expression by Western blotting was unchanged (data not shown). The relative expression levels of Survivin were measured by real-time PCR (Fig. 5C).
Gene Expression Affected by Caspase Inhibitor-Caspases mediate the intracellular activation of other caspases and selectively cleave distinct intracellular substrates, leading to dismantling of the architecture, signaling apparatus, and repair mechanisms of a cell. Eventually, caspases induce the activation of endonucleases that complete cellular suicide by inter- nucleosomal DNA fragmentation (27). Our study found that activities of caspase-3 and caspase-8 were increased at 2 h in the presence of 50 M GCDC. However, when GCDC stimulation was removed at 2 h, activities of caspases could be decreased and restored to control level by an unknown mechanism. GCDC-induced caspase activation is reversible. Next we investigated the correlation of caspase activation and antiapoptotic gene expression. Pan-caspase inhibitor (Alexis, Z-VAD-FMK, 260 020 M001) was utilized to treat hepatocytes. Caspase-3 and caspase-8 were significantly inhibited (Fig. 6). When 50 M GCDC was added into medium to stimulate hepatocytes for 4 h, activity of caspase-3 was increased to 11.40 pmol/min at 4 h compared with 3.98 pmol/min of control (without GCDC), whereas activity of caspase-8 was increasingly elevated from baseline 3.27 pmol/min to 8.53 pmol/min at 4 h. 50 M Z-VAD-FMK inhibited caspase-3 and caspase-8 60. 26 and 65.88%, respectively. Expression of Survivin, cIAP1, cIAP2, XIAP, A1/Bfl-1, but not iNOS, was down-regulated (Fig.  7A). The protein level of Survivin was consistent with Survivin mRNA (Fig. 7B).
Pathway for GCDC-induced Hepatocyte Apoptosis-GCDC induces hepatocyte apoptosis by death receptor or the mitochondrial pathway (28). GCDC causes direct activation of Fas, which stimulates FADD, caspase-8 activation, and subsequent apoptosis either through Bid cleavage and translocation to mitochondria or by direct activation of downstream effector caspases. In mitochondria there are actually two pathways. One directly stimulates generation of reactive oxygen species from the respiratory chain, which induces the permeability transition and cytochrome c release into cytosol. The other causes translocation of Bax to mitochondria, which function as channels for cytochrome release into cytosol and activation of downstream effector caspases (8). After stimulation with a concentration of 50 M GCDC, cytochrome c was increased even at 4 h (Fig. 8A). When 50 M Z-VAD-FMK was used, cytochrome c release was decreased. As hepatocytes were treated by 50 M GCDC, caspase 9 was increased at 4 h as well. However, under co-treatment of GCDC plus Z-VAD-FMK, the activity of caspase 9 was reduced (Fig. 8B). In addition, Z-VAD-FMK inhibited NF-B activation induced by 50 M GCDC (Fig.  8C). The role of cytochrome c and caspase 9 is worth further investigation. GCDC-induced hepatocyte apoptosis was a useful model to study the mechanism of apoptosis and pathogenesis of cholestatic liver injury. Although the roles of anti-apoptotic genes and NF-B activation has been incompletely understood, our study provides a clue to their protective function against GCDC-induced hepatocyte apoptosis.

DISCUSSION
Accumulation of bile acids induces hepatocyte apoptosis and necrosis (29). Our study demonstrated that low dosages (50 M) of GCDC caused hepatocyte apoptosis, not necrosis. The increase of caspases could be restored to control level after the low dosage of GCDC was removed. The activities of caspases were reversible. GCDC also induced expression of anti-apoptotic genes, such as Survivin, A1/Bfl-1, and iNOS. Moreover, activities of caspases were correlated to those genes because the pan-caspase inhibitor Z-VAD-FMK may inhibit expression of Survivin, cIAP1, cIAP2, XIAP, A1/Bfl-1, but not iNOS. At present we cannot completely understand the role of those genes during GCDC-induced hepatocyte apoptosis.
In inflammatory conditions hepatocyte survival is dependent on the protective function of anti-apoptotic genes (30). The expression of anti-apoptotic genes is mediated by caspase response, which can be inhibited by Z-VAD-FMK (31). Many important anti-apoptotic genes such as the inhibitor of apoptosis proteins (IAP) family, A1/Bfl-1, and iNOS have been identified that have anti-apoptotic roles in mature hepatocytes by controlling the transcription of specific survival genes (32). Although our preliminary results did not indicate the altered levels of cIAP1, cIAP2, and XIAP in 2 h of stimulation, the potential role for Survivin, A1/Bfl-1, and iNOS had been investigated. Survivin inhibits apoptosis via a pathway independent of bcl-2 (33). Survivin modulates caspase activation as well as Fas-mediated hepatic apoptosis, which is regulated via the mitochondrial pathway (34). Mitochondrial A1/Bfl-1, a member of the Bcl-2 family, is expressed in various tissues during development and adult life as well as in cancer cell lines. A1/Bfl-1 was shown to protect cells from apoptosis (35). A1/ Bfl-1 induced by cytokines suppresses the release of cytochrome c (20). Nitric oxide (NO) inhibits caspase-3 activity in hepatocytes (37). NO protects hepatocytes from tumor necrosis factor-␣/ActD-induced apoptosis via the interruption of mitochondrial apoptotic signaling through S-nitrosylation of caspase-8 (38). NO production depends on the expression of iNOS in response to inflammatory cytokines (39,40). It is known that iNOS gene transfer could suppress hepatocyte apoptosis (41). iNOS is heterogeneously distributed in liver (42,43), which may represent a mechanism through which hepatocytes control the degree of apoptosis in the liver (44,45).
NF-B is the prototypic transcription factor in eukaryotic cells known to play a pivotal role in transactivation of promoters for genes involved in inflammation, immune responses, and anti-apoptotic mechanism (46). The anti-apoptotic molecules regulated by NF-B include cIAP 1 and 2, TRAFs 1 and 2, IEX-1L, and A20 (32,33). The IAP family has been suggested to act as direct inhibitors of caspases. NF-B activation suppresses mitochondrial release of cytochrome c through the activation of a Bcl-2 homologue A1/Bfl-1 (20). The expression of iNOS can be regulated at the level of transcription, and NF-B activation is essential for its expression (47). NF-B activates endogenous iNOS via IKK␤ and provides protection from apoptosis (36).
In summary, GCDC-induced hepatocyte apoptosis is mediated by caspase activation, which is a highly regulated process. Caspase cascade may activate anti-apoptotic genes to prevent hepatocytes from apoptosis. GCDC-induced caspase response is reversible, which may be a useful approach for blocking GCDCinduced hepatocyte apoptosis.