Cilostazol protects hepatocytes against alcohol-induced apoptosis via activation of AMPK pathway

Alcoholic liver disease (ALD) is a worldwide health problem and hepatocyte apoptosis has been associated with the development/progression of ALD. However, no definite effective pharmacotherapy for ALD is currently available. Cilostazol, a selective type III phosphodiesterase inhibitor has been shown to protect hepatocytes from ethanol-induced apoptosis. In the present study, the underlying mechanisms for the protective effects of cilostazol were examined. Primary rat hepatocytes were treated with ethanol in the presence or absence of cilostazol. Cell viability and intracellular cAMP were measured. Apoptosis was detected by Hoechst staining, TUNEL assay, and caspase-3 activity assay. The roles of cAMP and AMP-activated protein kinase (AMPK) pathways in the action of CTZ were explored using pharmacological inhibitors and siRNAs. Liver from mice received ethanol (5 g/kg body weight) by oral gavage following cilostazol treatment intraperitoneally was obtained for measurement of apoptosis and activation of AMPK pathway. Cilostazol inhibited ethanol-induced hepatocyte apoptosis and potentiated the increases in cAMP level induced by forskolin. However, the anti-apoptotic effect of cilostazol was not reversed by an inhibitor of adenylyl cyclase. Interestingly, cilostazol activated AMPK and increased the level of LC3-II, a marker of autophagy. The inhibition of AMPK abolished the effects of cilostazol on LC3-II expression and apoptosis. Moreover, the inhibition of LKB1 and CaMKK2, upstream kinases of AMPK, dampened cilostazol-inhibited apoptosis as well as AMPK activation. In conclusion, cilostazol protected hepatocytes from apoptosis induced by ethanol mainly via AMPK pathway which is regulated by both LKB1 and CaMKK2. Our results suggest that cilostazol may have potential as a promising therapeutic drug for treatment of ALD.


Introduction
Alcohol is an important risk factor for development of liver disease. Alcoholic liver disease (ALD) represents a spectrum of pathological conditions ranging from simple hepatic steatosis PLOS ONE | https://doi.org/10.1371/journal.pone.0211415 January 29, 2019 1 / 17 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 to alcoholic hepatitis, fibrosis and eventually to cirrhosis [1,2]. Among cellular pathogenesis of ALD, hepatocyte apoptosis is a prominent feature of alcoholic hepatitis and hepatic fibrosis [3,4]. The inhibition of hepatocellular apoptosis in various liver injury models has been shown to reduce liver damage and progression of liver diseases [5,6]. Therefore, apoptosis has been considered as a target for therapeutic management of ALD. Hepatocyte apoptosis by ethanol is mediated by various factors including ethanol metabolites, mitogen-activated protein kinases (MAPKs), reactive oxygen species (ROS) generation and TNFα production. It has been reported that cyclic AMP (cAMP) inhibits apoptotic process in hepatocytes via suppression of caspase activity and TNFα expression [7,8]. Moreover, chronic ethanol exposure has shown to reduce hepatic cAMP in animal model which is associated with liver injury [9].
Cilostazol, a selective phosphodiesterase III (PDE III) inhibitor, has been widely used in clinical trials as an anti-platelet drug for the treatment of peripheral vascular diseases [10,11]. In addition, cilostazol has shown protective effects in various liver injury models including hepatectomy [12], ischemia-reperfusion injury [13] and hepatic steatosis [14]. Very recently, it has been reported that cilostazol exerts protective effects on ethanol-induced hepatocyte damage through suppression of oxidative stress [15]. The pleiotropic effects of cilostazol have shown to be mediated by both cAMP-dependent and-independent pathways including antioxidant effect [16,17] and AMP-activated protein kinase (AMPK) pathway [18,19].
AMPK plays a critical role in controlling cellular energy homeostasis [20,21]. In addition to its metabolic functions, AMPK plays a key role in regulation of cell survival/death. Recent study has shown that metformin protected liver from TNFα-induced apoptotic injury via AMPK-mediated caspase-3 inhibition [22], indicating anti-apoptotic role of AMPK. Moreover, the increased AMPK activity has been reported to alleviate various detrimental responses induced by ethanol in liver [23,24]. Autophagy, a self-degradation of cellular components in lysosomes, has been reported to be an important homeostatic mechanism for liver function [25,26]. Ethanol suppressed AMPK activation has been accompanied by downregulation of autophagic activity, where the increase in AMPK activation restored hepatic autophagy and subsequently reduced hepatic injury [27,28]. These findings suggest possibility that the activation of AMPK might underlie the anti-apoptotic effects of cilostazol in liver exposed to ethanol. Indeed, in the present study, cilostazol suppressed hepatocyte apoptosis via AMPK pathway in in vitro and in vivo models. 100 μg/ml streptomycin, 2 mM L-glutamine, and 100 nM dexamethasone (10% WME). After 3 h, media were changed to WME containing 0.1% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine, and 10 nM dexamethasone (0.1% WME).

Ethanol and CTZ treatment in vivo
Seven-week old male C57BL/6 (18-22 g) mice were obtained from Central Lab. Animal Inc. (Seoul, Korea). After one week acclimatization, mice were received ethanol (5 g/kg body weight) by oral gavage and were sacrificed at different time (0~24 h). Ethanol was diluted with sterile water (32% w/v). For cilostazol treatment, mice were given an intraperitoneal injection of either vehicle (10% DMSO in PBS) or cilostazol (10 mg/kg/day) once a day for 4 days before ethanol administration. A single dose (5 g/kg body weight) of ethanol was given orally at one hour after the last treatment with cilostazol. Mice were sacrificed by anesthetization with a cocktail of tiletamine HCl (25 mg/kg), zolazepam (25 mg/kg) and xylazine HCl (10 mg/kg) and liver was collected 6 h after ethanol administration. The doses of cilostazol used in this study were selected based on previous reports by others [19].

Hoechst staining
Nuclear morphology was detected by staining DNA using Hoechst 33342. Cells grown on collagen coated glass coverslips were fixed in ice cold methanol/acetic acid and then stained with Hoechst 33342 (5 μg/ml). After washing with deionized water, cells were mounted in 50% glycerol containing 20 mM citric acid and 50 mM disodium orthophosphate and observed under a fluorescent microscope (U-LH 100-3, Olympus).

TUNEL assay
DNA fragmentation was detected with a terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-nick end-labeling (TUNEL) assay kit (DeadEnd TM fluorometric TUNEL system; Promega, Madison, WI, USA) according to the manufacturer's protocol. In brief, cells grown on collagen coated coverslips were fixed with a freshly prepared 4% paraformaldehyde for 25 min at 4˚C. Then, cells were washed two times with ice-cold PBS and permeabilized with 0.2% Triton X-100 in PBS on ice for 5 min. The TdT and fluorescein-12-dUTP reactions were performed for 1 h at 37˚C in a humidified box. TUNEL-positive cells were analyzed under a fluorescence microscope (U-LH 100-3, Olympus).

Western blotting
Equal amount of protein was separated by SDS-PAGE gel and transferred to nitrocellulose membrane. After blocking with 5% non-fat dry milk in Tris-buffered saline, the blots were incubated with the primary anti . The blots were incubated with secondary antibodies (1:3000 dilution into TBST/5% nonfat dry milk) conjugated to horseradish peroxidase. The protein bands were detected by the Super Signal (Pierce, Rockford, IL). The density of respective bands was analyzed by the Chemi-Doc XRS imaging system (Bio-Rad, Hercules, CA). The membranes were re-probed with anti-GAPDH antibody, which was used as loading control.

Measurement of intracellular cAMP
The concentration of cAMP was measured using a cAMP EIA kit (Cayman, Ann Arbor, MI, USA) according to the manufacturer's protocols. Briefly, cells were lysed in 1 ml of lysis buffer following exposure to indicated reagents and 50 μl of supernatant after centrifugation was used for assay and then absorbance was measured at 405 nm.

Statistical analysis
Data are expressed as mean ± S.E.M. Statistical analyses were performed using the Student's ttest for comparison of values between two groups or one way ANOVA followed by Tukey's post hoc test for comparison of values among more than three groups. A value of P < 0.05 was considered significant. Values were analyzed using GraphPad Prism 4.0 software.

Effects of cilostazol on ethanol-reduced viability of hepatocytes
The protective effects of cilostazol on hepatocyte viability after ethanol exposure were determined. Treatment of cells with 100 mM of ethanol for 24 h reduced cell viability by about 20%. Pretreatment with cilostazol resulted in concentration-dependent improvement of cell viability and the effect was significant above 30 μM of cilostazol ( Fig 1A).

Effects of cilostazol on ethanol-induced apoptosis of hepatocytes
To determine the anti-apoptotic effect of cilostazol in hepatocytes, caspase-3 activity was assessed by activity assay and Western blotting. Ethanol significantly increased caspase-3 activity, which was suppressed by cilostazol in concentration-dependent manner ( Fig 1B). The inhibitory effect of cilostazol was significant above 10 μM. Consistent result was observed by Western blotting. Ethanol increased the levels of cleaved caspase-3, an active form of caspase-3, and cleaved poly (ADP-ribose) polymerase (PARP), one of the main cleavage targets of casapse-3, which were almost completely reduced by cilostazol ( Fig 1C).
Next, the apoptotic morphological changes were determined by Hoechst staining. Treatment with ethanol for 24 h increased the number of apoptotic cells with condensed chromatin and fragmented nuclei, which was reduced by pretreatment with cilostazol ( Fig 1D). Similarly, ethanol (100 and 200 mM) increased the number of TUNEL-positive cells, which was almost completely prevented by cilostazol ( Fig 1E).

The roles of cAMP in the anti-apoptotic effects of cilostazol in hepatocytes
Since many therapeutic effects of cilostazol are known to be attributed to the cAMP-dependent pathways, the role of cAMP in the anti-apoptotic effect of cilostazol was evaluated. Treatment with 10 μM of forskolin, an adenylyl cyclase activator, increased intracellular cAMP about four fold, which was almost completely blocked by SQ22536, an adenylyl cyclase inhibitor (Fig 2A). Although cilostazol alone slightly increased cAMP, cilostazol potentiated the cAMP elevation by forskolin (Fig 2A). Next, the role of cAMP in the inhibitory effect of cilostazol on the caspase-3 activity was measured. Treatment of cells with forskolin (1~10 μM) resulted in significant inhibition of ethanol-induced caspase-3 activity in concentrationdependent manner and the inhibitory effect of forskolin was reversed by SQ22536 ( Fig 2B). However, SQ22536 failed to dampen the anti-apoptotic effect of cilostazol, indicating that the anti-apoptotic effect of cilostazol in hepatocytes is mainly mediated by a cAMP-independent pathway.

The involvement of AMPK in the anti-apoptotic effects of cilostazol in hepatocytes
To determine whether the anti-apoptotic effects of cilostazol is mediated by AMPK pathway, the effect of cilostazol on AMPK signaling was examined. The activation of AMPK involves the phosphorylation of Thr 172 residue in α subunit of AMPK [30]. First, the activation of AMPK by cilostazol was examined by Western blotting using anti-phospho AMPKα Thr 172 antibody. Cilostazol increased phosphorylations of AMPK and acetyl-CoA carboxylase (ACC), a downstream target of AMPK, in a concentration-dependent manner. Phosphorylation of AMPK and ACC was effectively suppressed by compound C, an AMPK inhibitor ( Fig 3A).
Next, the role of AMPK in ethanol-induced apoptosis was examined. AICAR, an AMPK activator, and cilostazol substantially suppressed the increased cleaved caspase-3 level (Fig 3B). The inhibitory effects of cilostazol on cleaved-caspase-3 and -PARP were reversed by co-treatment with compound C. Compound C itself had no effect on the basal-and ethanol-induced caspase-3 activation (Fig 3C). Consistent results were observed by caspase-3 activity assay ( Fig  3D) and TUNEL assay (Fig 1E), where the anti-apoptotic effect of cilostazol was abolished by compound C. The involvement of AMPK in anti-apoptotic effect of cilostazol was further confirmed by using AMPK siRNA. As expected, AMPK siRNA significantly reduced AMPK activation and AMPK protein levels induced by cilostazol, whereas scramble siRNA had no effect ( Fig 3E). Consequently, AMPK siRNA suppressed the phosphorylation of ACC without affecting ACC protein level (Fig 3E). Similar to the effect of compound C, AMPK siRNA abolished the inhibitory effect of cilostazol on the caspase-3 activity induced by ethanol (Fig 3F). These results suggest that the anti-apoptotic effect of cilostazol is mainly mediated by AMPK pathway. Consistently, AMPK siRNA reduced the protective effect of cilostazol on cell viability reduced by ethanol (Fig 3G). Interestingly, AMPK siRNA treatment decreased basal cell viability by almost 50%. This suggests that AMPK activation plays a pivotal role in hepatocyte survival mechanisms including anti-apoptotic process.

The regulation of AMPK by LKB1 and CaMKK2 in cilostazol -treated hepatocytes
AMPK activation can be regulated by various upstream kinases and cAMP/PKA-dependent pathway. Since liver kinase B1 (LKB1) and calcium/calmodulin-dependent protein kinase 2 (CaMKK2) are well known upstream kinases responsible for the phosphorylation of α subunit of AMPK [31], their involvement in AMPK activation by cilostazol was examined. Cilostazol increased the phosphorylation of both LKB1 and CaMKK2 (Fig 4A). Both LKB1-and CaMKK2-siRNA reduced phosphorylation of AMPK and ACC induced by cilostazol ( Fig  4A), suggesting that both LKB1 and CaMKK2 are involved in cilostazol increased AMPK activation. In addition, both LKB1-and CaMKK2-siRNA abolished the inhibitory effect of cilostazol on the caspase3 activity increased by ethanol (Fig 4B). Similar to AMPK siRNA, LKB1and CaMKK2-specific siRNA significantly reduced both basal and cilostazol increased cell viability (Fig 4C).
The upstream regulators of AMPK were further confirmed with chemical inhibitors. STO-609, a CaMKK2 inhibitor decreased the phosphorylation of AMPK and ACC induced by cilostazol. However, KT5720, an inhibitor of cAMP-dependent protein kinase A, and SQ22536 had no effect on cilostazol-increased AMPK phosphorylation (Fig 4D). Consistently, the increased cell viability by cilostazol was revered by STO-609, but not by KT5720 (Fig 4E). These results indicate that cilostazol increased AMPK activation via LKB1-and CaMKK2-dependent pathways.

The increase in AMPK/mTOR-dependent autophagy by cilostazol
The dysregulation of autophagic function has been suggested to be an important hepatic pathology induced by ethanol [27,28,32,33]. Recently, it has been reported that cilostazol promoted autophagy via LKB1/AMPK/mTOR pathway, thereby increasing the degradation of β-amyloid in neuronal cells [34,35]. AMPK activation has shown to induce autophagy by suppressing mammalian target of rapamycin (mTOR) signaling [36,37]. Therefore, we examined whether the activation of AMPK signaling by cilostazol affects autophagy function in hepatocytes exposed to ethanol. Cilostazol significantly decreased the phosphorylation of mTOR and this inhibitory effect was abolished by AMPK siRNA (Fig 5A). As reported by others [27,28,33], ethanol reduced the level of Light Chain 3 (LC3)-II, a marker of autophagy, which was recovered by cilostazol in control cells treated with scramble siRNA. However, AMPK siRNA significantly reduced LC3-II level increased by cilostazol (Fig 5B). Consistently, compound C significantly reduced cilostazol effect on LC3-II expression (Fig 5C). These results suggest that cilostazol increased autophagy in hepatocytes via AMPK/mTOR dependent pathway.

The effects of cilostazol on liver cell apoptosis in acute ethanol treated mice
To validate the physiological relevance of anti-apoptotic effects of cilostazol, caspase-3 activity in liver from acute ethanol (5 g/kg body weight) treated mice was measured by activity assay and immunohistochemistry. Ethanol treatment significantly increased caspase-3 activity after 3 h of ethanol feeding with maximum increase at 6 h (Fig 6A), which was reduced by pretreatment with cilostazol (10 mg/kg of body weight) (Fig 6B). Histologically, microsteatosis of hepatocytes was frequently identified in ethanol group (Fig 6Cb), which was improved in cilostazol treated group (Fig 6Cc). Sinusoidal or periportal hepatitis was not revealed in all three groups under these experimental conditions, but a significant increase in the expression of cleaved caspase-3 was observed in the cytoplasm of hepatocytes from ethanol-fed mice as compared to control mice (Fig 6Cd and 6Ce). The level of cleaved caspase-3 expression was substantially reduced in cilostazol treated group (Fig 6Cf). The involvement of AMPK signaling in the action of cilostazol in liver was determined by Western blotting. In consistent with previous reports showing the decreased AMPK phosphorylation by binge or chronic ethanol feeding [38], AMPK phosphorylation was decreased by ethanol feeding, which was recovered by cilostazol (Fig 6D). Interestingly, the protein level of AMPK was also reduced by ethanol. However, cilostazol had no effect on AMPK protein level. Although the levels of LKB1 and CaMKK2 phosphorylation were not affected by ethanol feeding, cilostazol significantly increased LKB1 and CaMKK2 phosphorylation compared to ethanol feeding (Fig 6D). These suggest that cilostazol may protect liver from ethanol induced steatosis and apoptosis, two most important causal factors for alcoholic liver injury, by recovering AMPK signaling through LKB1 and CaMKK2 activation. Discussion AMPK has recently emerged as an important regulator of various liver functions. Therefore, AMPK has been suggested to be a crucial therapeutic target to treat liver injury including steatosis, inflammation and apoptosis [39]. Cilostazol has been reported to have protective effects on diverse disease models including liver damage. The beneficial effects of cilostazol are mediated by multiple signaling pathways. However, the involvement of AMPK in the effects of cilostazol on hepatocyte apoptosis induced by ethanol has not been determined. In the present study, we have shown that cilostazol suppressed ethanol-induced hepatocyte apoptosis in in vitro primary hepatocyte and in vivo alcohol drinking mice. The pharmacological and gene knockdown approaches clearly demonstrated that the anti-apoptotic effect of cilostazol was mediated mainly through AMPK-dependent pathway. Of interest, the protective effect of cilostazol was cAMP/PKA independent although cAMP has shown an anti-apoptotic role. The activation of AMPK by cilostazol was down regulated by both LKB1-and CaMKK2-siRNA, which abrogated the anti-apoptotic effect of cilostazol. In addition, pharmacological inhibitor of CaMKK2, STO-609, but not inhibitors of cAMP/PKA pathway blunts AMPK activation, supporting the anti-apoptotic role of cilostazol via LKB1-CaMKK2/AMPK pathway. The protective effect of cilostazol on hepatocyte apoptosis is consistent with a recent report, in which cilostazol prevented hepatocyte apoptosis induced by ethanol by ameliorating ROS generation [15].
The increasing evidence suggests that the impairment of AMPK activity is associated with various liver diseases such as metabolic disorders, inflammation, and cancer. Ethanol treatment both in vitro and in vivo has shown to decrease hepatic AMPK activation [40], which was associated with development of steatosis, inflammation and apoptosis. Therefore, the identification of a reagent activating AMPK signaling can provide a good strategy to treat ALD. Consistently, our present study shows that the suppressive effect of cilostazol on hepatocyte apoptosis was abolished by AMPK inhibitor, compound C, and gene silencing of AMPKα, indicating that the anti-apoptotic effect of cilostazol is mediated by AMPK-dependent pathway.
The activation of AMPK composed of catalytic α subunit and regulatory β and γ subunits is regulated by several independent mechanisms. It is generally known that AMPK activity is regulated by intracellular energy status. In response to energy depletion, the increased AMP/ATP ratio increased AMPK activation via binding of AMP to AMPKγ subunit which leads to its conformational change [41]. The best known as an AMPK activator, AICAR is converted to ZMP within cells and then activates AMPK via mechanism mimicking AMP [42]. Similarly, metformin and thiazolidinediones (TZDs), which are used to treat diabetes in clinical situation, have shown to increase AMP/ATP ratio by inhibiting Complex I of the respiratory chain and subsequently activate AMPK [43]. However, metformin has also shown to activate AMPK without altering AMP/APT ratio [44]. It has been reported that ethanol treatment reduced ATP level in hepatocyte which was recovered by cilostazol [15]. This suggests that cilostazol may activate AMPK via AMP-independent pathway. The activation of AMPK is increased by upstream kinases including LKB1 and CaMKK2. Depending on cell types and stimuli, AMPK is phosphorylated by specific upstream kinases. In the present study, AMPK phosphorylation induced by cilostazol was abolished by both LKB1-and CaMKK2-siRNA. Although LKB1 is mainly involved in AMPK activation, CaMKK2 also acts as an upstream kinase of AMPK in response to the increased intracellular calcium [31]. Interestingly, cilostazol has shown to increase intracellular calcium in endothelial cells [45]. However, the mechanisms responsible for cilostazol regulated LKB1 and CaMKK2 in hepatocytes are not completely understood and further investigations are needed in the future study.
In contrast to our finding, the increase of intracellular cAMP underlies the various pharmacological effects of cilostazol. Moreover, cilostazol induced AMPK phosphorylation in neuronal cells was abolished by KT5720, a PKA inhibitor [35]. Intracellular cAMP is produced by activation of adenylyl cyclase and degraded by PDE. cAMP, a second messenger molecule, is known to be involved in a wide range of cellular responses including cell proliferation, differentiation, inflammation, and apoptosis. In liver, the involvement of the cAMP signaling pathway in both cell survival and apoptosis depending on stimuli has been demonstrated. It has been reported that chronic ethanol caused a decrease of cAMP and activation of adenylyl cyclase prevented ethanol-induced liver injury [8,46], indicating the protective role of cAMP. In agreement with these reports, our results show that forskolin significantly increased intracellular cAMP level and suppressed caspase-3 activation increased by ethanol (Fig 2A and 2B). Although cilostazol alone marginally increased cAMP level under this condition, CTZ potentiated forskolin effect on cAMP elevation. However, adenylyl cyclase inhibitor SQ22536 failed to abolish the anti-apoptotic effect of cilostazol. Furthermore, neither SQ22536 nor KT5720 had effect on cilostazol-induced AMPK activation. These suggest that the protective effect of cilostazol on ethanol-induced apoptosis is mainly mediated by cAMP-independent pathways. Similarly, previous reports have shown that cilostazol did not affect intracellular cAMP level and elicited its beneficial effects via cAMP-independent pathways [47,48].
Autophagy is considered as an important cellular mechanism for the maintenance of liver functions by lysosomal degradation and recycling of cellular components [25,26], where AMPK activation plays a critical role in autophagy induction by inhibiting mTOR signaling [36,37]. Acute or chronic ethanol treatment induced a decrease in hepatic autophagy, and restoration of the autophagic activity ameliorated the resulting pathogenesis including steatosis, oxidative stress, and apoptosis [27,28,32,33]. In the present study, cilostazol decreased mTOR phosphorylation and restored the level of LC3-II expression decreased by ethanol, which was AMPK dependent. This suggests that cilostazol improved autophagy function in hepatocytes via AMPK/mTOR dependent pathway, which may be responsible in part for the protective effect of cilostazol from hepatic fat accumulation and apoptosis induced by ethanol.
Oxidative stress is one of important contributing factors for ethanol-induced liver injury and numerous anti-oxidants have shown beneficial effects on ALD. We have also overserved that trolox, an antioxidant, increased hepatocyte viability reduced by ethanol (S1A Fig). In recent study, cilostazol prevented ethanol-induced hepatocyte apoptosis via ameliorating ROS generation [15]. Consistently, we have also observed that cilostazol completely blocked ROS generation induced by ethanol in hepatocytes (S1B Fig). However, this anti-oxidant effect of cilostazol was not reversed by inhibition of either cAMP/PKA pathway or AMPK pathway (S1B Fig). Therefore, cilostazol may protect liver cells from ethanol via AMPK-dependent and-independent pathways including antioxidant effect.
Taken together, cilostazol protected hepatocytes from apoptosis induced by ethanol and its effect is mainly mediated by activation of AMPK pathway and dampening ROS generation. The activation of AMPK is regulated by upstream kinases LKB1 and CaMKK2. Considering that cilostazol has been safely used in various clinical situations for the treatment of peripheral vascular diseases, cilostazol may be a promising a novel pharmacological therapy for ALD.