The Inhibition of Activated Hepatic Stellate Cells Proliferation By

European Journal of Pharmacology 747 (2015) 71–87

Contents lists available at ScienceDirect

European Journal of Pharmacology

journal homepage: www.elsevier.com/locate/ejphar

Pulmonary, gastrointestinal and urogenital pharmacology The inhibition of activated hepatic stellate cells proliferation Kip1 by arctigenin through G0/G1 phase cell cycle arrest: Persistent p27 induction by interfering with PI3K/Akt/FOXO3a signaling pathway

Ao Li a,n,1, Jun Wang a,1, Mingjun Wu b,1, Xiaoxun Zhang a, Hongzhi Zhang a a College of Pharmacy and Bioengineering, Chongqing University of Technology, 69 Hong Guang Avenue, Chongqing 400054, PR China b Institute of Life Sciences, Chongqing Medical University, 1 Yi Xue Yuan Road, Chongqing 400016, PR China article info abstract

Article history: Proliferation of hepatic stellate cells (HSCs) is vital for the development of fibrosis during liver injury. In Received 14 July 2014 this study, we describe that arctigenin (ATG), a major bioactive component of Fructus Arctii, exhibited Received in revised form selective cytotoxic activity via inhibiting platelet-derived growth factor-BB (PDGF-BB)-activated HSCs 6 November 2014 proliferation and arrested cell cycle at G0/G1 phase, which could not be observed in normal human Accepted 20 November 2014 hepatocytes in vitro. The cyclin-dependent kinase (CDK) 4/6 activities could be strongly inhibited by ATG Available online 10 December 2014 through down-regulation of cyclin D1 and CDK4/6 expression in early G1 phase arrest. In the ATG- Chemical compounds studied in this article: treated HSCs, the expression level of p27Kip1 and the formation of CDK2-p27Kip1 complex were also Arctigenin (PubChem CID: 64981) increased. p27Kip1 silencing significantly attenuated the effect of ATG, including cell cycle arrest and Dimethyl sulfoxide (PubChem CID: 679) suppression of proliferation in activated HSCs. We also found that ATG suppressed PDGF-BB-induced LY294002 (PubChem CID: 3973) phosphorylation of Akt and its downstream transcription factor Forkhead box O 3a (FOXO3a), decreased SB203580 (PubChem CID: 176155) U0126 (PubChem CID: 3006531) binding of FOXO3a to 14-3-3 protein, and stimulated nuclear translocation of FOXO3a in activated HSCs. PHT-427 (PubChem CID: 44240850) Furthermore, knockdown of FOXO3a expression by FOXO3a siRNA attenuated ATG-induced up-regula- Kip1 Propidium iodide (PubChem CID: 104981) tion of p27 in activated HSCs. All the above findings suggested that ATG could increase the levels of 5-Bromo-2-deoxyuridine (PubChem CID: p27Kip1 protein through inhibition of Akt and improvement of FOXO3a activity, in turn inhibited the 6035) CDK2 kinase activity, and eventually caused an overall inhibition of HSCs proliferation. Phenylmethanesulfonyl fluoride (PubChem & 2014 Elsevier B.V. All rights reserved. CID: 4784) Sodium dodecyl sulfate (PubChem CID: 3423265)

Keywords: Hepatic stellate cell Arctigenin Forkhead box O (FOXO) p27Kip1 Proliferation Liver ﬁbrosis

1. Introduction fibrosis is a key stage of disease process from hepatic damage to cirrhosis or even to hepatocellular carcinoma (Bataller and Brenner, Liver fibrosis, a common outcome of chronic hepatocellular 2005; Tsukada et al., 2006). Hepatic stellate cells (HSCs) have been damage, could be induced by a variety of etiological factors, includ- considered as the primary source of ECM in liver injury (Friedman, ing viral infections, hepatic toxins, alcohol abuse and autoimmune 2010; Lotersztajn et al., 2005). According to the recent studies, the reactions and so on (Hui and Friedman, 2003). Liver fibrosis, whose HSCs proliferation and expansion of their pool were a key step pathological process is progressive, is characterized by increased during the fibrogenic process (Son et al., 2013; Sun et al., 2009). The deposition of extracellular matrix (ECM) proteins. Meanwhile, liver hepatic ECM accumulation resulted from both increased number and gene expression changes of activated HSCs. Hence, suppression of activated HSCs proliferation has been proposed as a therapeutic

n strategy for the treatment and prevention of liver fibrosis. Corresponding author. Tel.: þ86 23 62400251; fax: þ86 23 68906603. E-mail address: [email protected] (A. Li). The cell cycle is a common convergent point for the mitogenic 1 These authors contributed equally to this work. signaling cascades. In activated HSCs, every major checkpoint in http://dx.doi.org/10.1016/j.ejphar.2014.11.040 0014-2999/& 2014 Elsevier B.V. All rights reserved. 72 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 cell cycle has been controlled by multiple protein kinases, each of and all the other chemicals were purchased from Sigma (St. Louis, which contains a regulatory cyclin component and a catalytic MO, USA). cyclin-dependent kinase (CDK) (Peng et al., 2013). The kinase activities of CDK–cyclin complexes would be inhibited by two 2.2. Cell culture and treatment classes of cyclin-dependent kinase inhibitors (CKIs). One is the INK4 protein family consisting of p16INK4a,p15INK4b,p18INK4c and LX-2 was a well-characterized human HSCs line. Many features of p19INK4d, specifically inhibiting CDK4 and CDK6. The other is the activated HSCs phenotype could be recapitulated in LX-2, including the kinase inhibiting protein (KIP) family consisting of p21Cip1,p27Kip1 expression of the PDGFRβ and proliferation in response to PDGF (Xu and p57Kip2, which can inhibit several CDKs (Owa et al., 2001). et al., 2005). LX-2 cells were obtained from China Center for Type Specifically, p27Kip1 could negatively regulate the activity of Culture Collection (CCTCC, Wuhan, China) and maintained in Dulbecco's protein kinase complex cyclin E–CDK2, thereby blocking cell cycle modified Eagle's medium (DMEM; Hyclone Laboratories, Logan, UT, progression from G1 to S phase (Egozi et al., 2007). The aberrations USA) supplemented with 10% fetal bovine serum (FBS; Gibco, Grand in the G1 regulation following liver injury often lead to rapid Island, NY, USA) and antibiotics (100 U/ml penicillin G and 100 μg/ml proliferation of HSCs. Upon mitogenic stimuli, especially platelet- streptomycin)at371Cinhumidified air with 5% CO2.Humanembryo- derived growth factor-BB (PDGF-BB), the level of p27Kip1 is down- nic hepatocytes cell line L-02 was purchased from the Cell Bank of regulated, allowing cyclin E–CDK2 to drive HSCs into the S phase Chinese Academy of Sciences (Shanghai, China) and maintained in and promote rapid proliferation (Kossatz et al., 2004). RPMI 1640 medium (Hyclone). The culture conditions were the same as In our pervious work, herbs were screened for inhibiting the the conditions described above. The LX-2 and L-02 cells were prepared proliferation of HSCs. An ethanol extract from Fructus Arctii was with passages between 18 and 25. Both of two cells were collected for active in preventing cell growth of HSCs activated by PDGF-BB. As application in relevant experiments. When cells were 70% confluent, a main bioactive component from Fructus Arctii, arctigenin (ATG) they were cultured in medium without FBS for 24 h to synchronize, has been commonly applied in the clinical practice for reducing then the cells were pre-incubated for 4 h in medium containing 0.5% inflammation in diverse contexts, including treatment of anemo- FBS and either DMSO (vehicle control) or ATG. After pre-incubation, pyretic cold, sweeling of throat, cough, measles, and syphilis in human recombinant PDGF-BB (Peprotech, Rocky Hill, NJ, USA) with a China and other Asian countries (Cho et al., 2004; Lee et al., 2010). final 50 ng/ml concentration was added to the culture medium without However, few studies have focused on the therapeutic or chemo- wash and the mixture was incubated for different time intervals at preventive properties of ATG on hepatic fibrogenesis, both in vitro 37 1Cuntilfurtherassays. or in vivo. In the present study, we aimed to investigate the effects of ATG on the proliferation-related events of activated HSCs and to 2.3. Assay of cell proliferation clarify the underlying molecular mechanisms. LX-2 cells were seeded into 96-well plates at a density of 5 103 cells/well. The cell proliferation was analyzed with 5-bromo- 2-deoxyuridine (BrdU)-based Cell Proliferation ELISA kit (Roche 2. Materials and methods Applied Science, Mannheim, Germany). Briefly, cells were synchronized after serum starvation for 24 h. Then cells were transfected with 2.1. Materials either p27Kip1-specific siRNA or scrambled control siRNA and/or pretreated with ATG for certain time periods. All the cell groups were Arctigenin (ATG, 499% purity; MW: 372.41) was purchased from thenexposedto50ng/mlPDGF-BBfor12,24or48h.Inthelast4hof Nanjing Zelang Medical Technology Co. Ltd. (Nanjing, China). A stock treatment, the cells were labeled with 10 μMBrdU.Afterdisposingthe solution of ATG with the concentration of 16 mM was prepared in labeling medium, the genomic DNA of cells was fixed and denatured dimethyl sulfoxide (DMSO) and stored at 20 1C until use. The final with FixDenat solution. The BrdU incorporated in the cells was concentration of DMSO in all experiments did not exceed 0.1% (v/v); detected by a peroxidase-conjugated anti-BrdU antibody and quanti- this concentration did not cause any obvious deleterious effects on fied by measuring the absorbance (A) value at 370 nm with a universal cell viability (data not shown). Antibodies against the following microplate autoreader (Varioskan, Thermo Electron Co., Waltham, MA, proteins were used: phospho-(Thr183/Tyr185)-stress-activated USA; the reference wavelength was 492 nm). Absorbance values were protein kinase (SAPK)/c-Jun NH2 kinase (JNK), phospho-(Thr202/ directly correlated with the amount of synthesized DNA, as well as Tyr204)-extracellular signal-regulated kinase (ERK)1/2, phospho- indicating the number of proliferated cells. Each time point was (Thr180/Tyr182)-p38 mitogen-activated protein kinase (MAPK), conducted in triplicate and the experiment was repeated for six times. SAPK/JNK, ERK1/2, p38 MAPK (Cell Signaling Technology, Beverly, The cell proliferation inhibition rate (IR%) was calculated with the MA, USA); PDGF receptor-β (PDGFRβ) and phosphotyrosine (BD following formula: Biosciences, San Diego, CA, USA); phospho-(Thr308)-Akt, Akt, IR% ¼ðA A Þ=ðA A Þ100% CDK2, CDK4, CDK6, and lamin B1 (Santa Cruz Biotechnology, Santa control treated control blank Cruz,CA,USA);phospho-(Ser473)-Akt,14-3-3,cyclinD1,cyclinE1, where Atreated was the A value in wells treated with ATG before the Cip1 Kip1 p21 ,p27 , Forkhead box O 3a (FOXO3a), FOXO1, FOXO3a, exposure of PDGF-BB stimulation, Acontrol was the A value in wells FOXO4, and phospho-(Ser262)-FOXO4 (Epitomics, Burlingame, CA, treated with DMSO before the exposure of PDGF-BB stimulation, and USA); Fluorescein isothiocyanate (FITC)- or horseradish peroxidase Ablank was the A value in wells without cells but incubated with DMSO. (HRP)-conjugated goat anti-rabbit and goat anti-mouse IgG were The IC50 value was designated as the concentration that caused 50% obtained from Beijing Zhongshan Golden Bridge Biotechnology Co. inhibition of cells proliferation, and was calculated by SPSS 16.0 for Ltd (Beijing, China). LY294002 (phosphoinositide 3-kinase inhibitor; windows (IBM, Chicago, IL, USA). PI3K inhibitor), SB203580 (p38 MAPK inhibitor), normal rabbit IgG, and primary antibodies for phospho-(Thr24)-FOXO1/(Thr32)-FOXO3a 2.4. Assay of lactate dehydrogenase (LDH) release and β-actin were purchased from Beyotime Institute of Biotechnol- ogy (Jiangsu, China). U0126 (inhibitor of MAPK kinase; inhibitor of L-02 or LX-2 cells were seeded in 96-well plates at a density of MEK1/2) was provided by Cell Signaling Technology. PHT-427 5 103 cells/well. The cells were incubated in RPMI 1640 or DMEM (protein kinase B/Akt inhibitor; Sheng et al., 2011)waspurchased medium supplemented with 10% FBS for 24 h. Then cells were from Selleck Chemicals (Houston, TX, USA). Propidium iodide (PI) synchronized with serum starvation for another 24 h. After A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 73

Fig. 1. Effects of ATG on LDH release from hepatocytes and PDGF-BB-activated HSCs; and on BrdU incorporation by activated HSCs. (A) LX-2 human HSCs or (B) L-02 human hepatocytes were pretreated with DMSO (vehicle) or various concentrations of ATG (0.125–16 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for 48 h. After treatment, an aliquot of conditioned medium was analyzed for LDH release. (C) Serum-starved LX-2 human HSCs were pretreated with DMSO (vehicle) or various concentrations of ATG (0.125–2 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for 12, 24 or 48 h. The HSCs proliferation was determined using BrdU incorporation ELISA. The results were quantiﬁed by differences in absorbance at wavelength 370 minus 492 nm. Data are shown as mean7S.E.M. from six independent experiments. #Po0.05 versus PDGF-BB-treated group at corresponding time intervals.

starvation, L-02 cells were treated with DMSO or ATG at certain 2.6. Preparation of cell lysates concentrations for 48 h, while LX-2 cells were pretreated with DMSO or ATG for 4 h followed by the exposure of PDGF-BB (50 ng/ml) for Cytoplasmic and nuclear extracts were obtained using the NE- 48 h. LDH activities in culture medium were investigated by measur- PER Nuclear and Cytoplasmic Extraction Kit (Pierce Biotechnology, ing the A value at 440 nm using a LDH release assay kit (Nanjing Rockford, IL, USA). Briefly, LX-2 cells were pretreated for 4 h with Jiancheng Bioengineering Institute, Nanjing, China) according to the DMSO or various concentrations (0.25, 0.5 or 1 μM) of ATG, manufacturer's instructions. followed by the exposure of PDGF-BB (50 ng/ml). At the end of the indicated time periods, cells were washed twice with ice-cold PBS (pH 7.4), detached and collected by centrifugation at 1000g for 2.5. Cell cycle analysis 5 min. The pellet was resuspended in cytoplasmic extraction reagent (CER) I, vortexed and incubated on ice for 10 min. Then LX-2 cells were seeded in 6-well plates at a density of CER II was added, and the extracts were further incubated on ice. 5 104 cells/well and synchronized with a serum starvation prior Thereafter, the extracts were vortexed again and centrifuged at to the experiment. After 24 h, the cells were transfected with 14,000g for 5 min at 4 1C. The supernatant (cytosolic extract) was either p27Kip1-specific siRNA or scrambled negative control siRNA transferred to a pre-chilled tube. The nuclear pellet was resus- and/or pretreated with ATG (0.5 μM) for certain time periods. pended in nuclear extraction reagent (NER) and incubated on ice Subsequently, cells were exposed to 50 ng/ml PDGF-BB for 24 or for 40 min, with vortexing every 10 min. After centrifuging the 48 h. The cells were harvested after PDGF-BB stimulation. The cell tube at 16,000g for 10 min, the supernatant fraction (nuclear were washed twice with ice-cold phosphate buffered saline (PBS; extract) was transferred to a clean tube. pH 7.4) and fixed in 70% ethanol in PBS overnight at 4 1C. After For isolating the whole cell extracts, LX-2 cells were seeded centrifugation at 1000g for 5 min, cells were re-suspended in into 6-well plates at a density of 2 105 cells/well. The cells 500 μl PBS containing 100 μg/ml RNase A (Roche) and 50 μg/ml PI. were synchronized with serum starvation for 24 h. After The cells were incubated and reacted in the dark at room starvation, the medium was switched to 0.5% serum-containing temperature for 30 min, and then analyzed using a FACSCalibur medium and cultures were pretreated with DMSO or ATG for 4 h, flow cytometer (BD Biosciences). A minimum of 1 104 cells/ followed by the treatment with 50 ng/ml PDGF-BB. At the end sample were counted. The percentage of cells in different phases of treatment, cells were harvested, washed twice with ice-cold of cell cycle was analyzed by CellQuest software (BD Biosciences). PBS, and lysed for 30 min in cell lysis buffer [20 mM Tris–HCl 74 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87

Fig. 2. ATG induces prominent G0/G1 phase arrest in PDGF-BB-activated HSCs. Serum-starved LX-2 human HSCs were pretreated with DMSO (vehicle) or ATG (0.5 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for 24 or 48 h. The distribution of HSCs in the cell cycle was determined by ﬂow cytometry with PI-stained nuclei as described in Section 2. (A) Representative ﬂow cytometric histograms are shown for HSCs pretreated with DMSO or ATG in the absence or presence of PDGF-BB. (B) The statistical data for the percentage of cells in each phase. Data are presented as mean7S.E.M. of three independent experiments. nPo0.05 versus vehicle-treated control group at corresponding time intervals; #Po0.05 versus PDGF-BB-treated group at corresponding time intervals.

(pH 7.5), 150 mM sodium chloride (NaCl), 1% (v/v) polyethylene 2.7. Western blotting analysis glycol octylphenol ether (Triton X-100), 2.5 mM sodium pyropho- sphate, 1 mM ethylene diamine tetraacetic acid (EDTA), 1% An aliquot of protein samples from cell lysates was boiled for (w/v) sodium orthovanadate, 0.5 μg/ml leupeptin, 1 mM phenyl- 5 min in a 5 sodium dodecyl sulfate (SDS) sample loading buffer methanesulfonyl fluoride (PMSF)]. Cell lysates were collected [250 mM Tris–HCl (pH 6.8), 500 mM dithiothreitol, 10% (w/v) SDS, by scraping from the plates and clarified by centrifugation at 50% (v/v) glycerol, 0.05% (w/v) bromophenol blue]. Next, proteins 14,000g at 4 1C for 5 min. Protein concentration in the super- were loaded at 50 μg/lane and separated on 8–12% SDS–polyacry- natants was determined using bicinchoninic acid (BCA) protein lamide gels. The separated proteins were electro-transferred to assay reagents (Beyotime) employing bovine serum albumin (BSA) polyvinylidene difluoride (PVDF) membrane (Millipore, Bedford, as the standard. All samples were stored at 80 1C until further MA, USA) in a glycine/methanol transfer buffer [25 mM Tris–HCl analysis. (pH 8.3), 192 mM glycine, 20% methanol] using the Trans-Blot SD A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 75

Fig. 3. Effects of ATG on cell cycle regulatory proteins in PDGF-BB-activated HSCs. Serum-starved LX-2 human HSCs were pretreated with DMSO (vehicle) or ATG (0.25, 0.5, or 1 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for 24 h. After treatment, whole cell lysates were collected and subjected to Western blotting analysis using antibodies specific for cyclin D1, cyclin E1, CDK2, CDK4, CDK6, p21Cip1 and p27Kip1. β-Actin level served as the loading control. (A) The experiments were repeated three times with similar results, and representative immunoblots are shown for each protein. (B)–(D) For quantitation of proteins of interest, band intensities were converted to arbitrary densitometric units, normalized by the value of β-actin and expressed relative to the level of the same protein in vehicle-treated cells (defined as one-fold). Data are presented as mean7S.E.M. of three independent experiments. nPo0.05 versus vehicle-treated control group; #Po0.05 versus PDGF-BB-treated group. semidry transfer cell system (Bio-Rad, Hercules, CA, USA). The relative protein level compared to that of in DMSO-treated cells membrane was blocked with 5% nonfat dried milk powder (w/v) (vehicle control), which was defined as one-fold. in Tris-buffered saline for 1 h at room temperature, and incubated overnight at 4 1C with a 1:1000 dilution of rabbit monoclonal 2.8. Immunoprecipitation and CDKs kinase activity assay antibodies against human phospho-(Thr183/Tyr185)-JNK, phospho-(Thr202/Tyr204)-ERK1/2, phospho-(Thr180/Tyr182)-p38 For the detection of phosphorylated PDGFRβ; or 14-3-3-phospho- MAPK, phospho-(Ser473)-Akt, SAPK/JNK, ERK1/2, p38 MAPK, FOXO3a or CKIs-CDKs binding affinity, LX-2 cells (2 105 cells/well) phospho-(Thr24)-FOXO1/(Thr32)-FOXO3a, phospho-(ser262)- were seeded into 6-well plates for 24 h in medium without FBS to FOXO4, FOXO1, FOXO3a, FOXO4, 14-3-3, cyclin D1, cyclin E1, synchronize, and then pre-incubated for 4 h in medium containing p21Cip1,orp27Kip1; or with rabbit polyclonal antibodies against 0.5% FBS and either DMSO or ATG, followed by the treatment with human phospho-(Thr308)-Akt, Akt, CDK2, CDK4, PDGFRβ, or lamin PDGF-BB (50 ng/ml). At the end of the indicated time points, cells B1 (1:250 dilution); or mouse monoclonal antibodies against were harvested and lysed in the ice-cold lysis buffer, as described in human CDK6, phosphotyrosine, or β-actin (1:1000 dilution). Sub- the preparation of whole cell lysates. For each sample, whole cell sequently, the membranes were incubated for 1 h at 25 1C with a lysates (200 μg) were incubated with protein A/G-plus agarose beads 1:3000 dilution of a secondary HRP-conjugated anti-rabbit or anti- (Santa Cruz) for 45 min at 4 1C to minimize the nonspecificbinding. mouse IgG. Enhanced chemiluminescence reagents (Millipore) PDGFRβ, 14-3-3, p21Cip1,orp27Kip1 were then immuno-precipitated were applied to visualize the immune-reaction bands, and densi- separately from the pre-cleared HSCs lysates with specificantibody tometric analysis was performed with Gel Doc/Chemi Doc Imaging (2 μg) by incubation overnight at 4 1C. Normal rabbit IgG (2 μg) was System and Quantity One software (Bio-Rad, Hercules, CA, USA). used as negative control. Next, 25 μl protein A/G-plus agarose beads The β-actin level served as an internal control. For the protein of was added and rocked at 4 1C for 2 h. The immunoprecipitates were interest, the band intensity was expressed as arbitrary densito- washed for three times, twice in ice-cold lysis buffer [150 mM NaCl, metric units. It was normalized by the value of β-actin as the 10 mM HEPES (pH 7.4), 0.2% (v/v) nonidet P40], and once in ice-cold 76 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87

25 μl kinase buffer containing 20 μM cold ATP (Upstate Cell Signal- ing Solutions, Lake Placid, NY, USA), 5 μCi [γ-32P]ATP (3000 Ci/mmol, MP Biomedicals, Irvine, CA, USA) and 5 μg histone H1 (Upstate Cell Signaling Solutions) as a substrate for CDK2 assay, or 2 μg glu- tathione S-transferase-Rb C-terminal (Rb amino acids 769-921) fusion protein (Santa Cruz) as a substrate for CDK4 and CDK6 assay. The reaction was terminated by boiling the samples for 5 min in 2 Laemmli's loading buffer, and the samples were analyzed by 12% SDS–polyacrylamide gel electrophoresis. The gels were dried and subjected to autoradiography, and the bands of phosphorylated Rb and histone H1 were quantiﬁed with Quantity One software (Bio-Rad). The protein expression levels of CDK2, CDK4 and CDK6 determined by Western blotting were applied as loading controls. The kinase activity was normalized by the level of immunoprecipitation loading control as the relative activity compared to that of in DMSO-treated cells (vehicle control), which was deﬁned as one-fold.

2.9. Immunocytochemistry for FOXO1, FOXO3a, and FOXO4

LX-2 cells grown on coverslips in 24-well plates were pretreated with DMSO or ATG (0.5 μM) for 4 h before exposure to PDGF-BB (50 ng/ml) for 4 h. After washing three times with PBS, cells were ﬁxed in 3.7% paraformaldehyde at room temperature for 20 min, permeabilized with 0.5% (v/v) Triton X-100 at 37 1Cfor20min, washed three times with PBS, and blocked at room temperature with 10% normal sheep serum for 1 h. Subsequently, the samples were incubated at 4 1Covernightwithspeciﬁc primary antibodies against FOXO1 (1:250 dilution), FOXO3a (1:100) and FOXO4 (1:200). Cells were then rinsed with PBS three times for 10 min, and incubated with FITC-conjugated secondary antibodies (1:50) at 37 1Cfor1h.To identify nuclei, cells were counterstained with PI (red) for 1 min. The results were imaged with a Leica TCS-SP2 confocal laser scanning microscope (Leica Microsystems, Wetzlar, Germany).

2.10. RNA interference

Fig. 4. ATG inhibits kinase activities of CDK2 and CDK4/6 in PDGF-BB-activated The sequence of small interfering RNA (siRNA) oligonucleotides HSCs. Serum-starved LX-2 human HSCs were pretreated with DMSO (vehicle) or Kip1 ATG (0.5 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for targeting a scrambled negative control and p27 were designed 24 h. CDK complexes were isolated from equal amounts of cell lysates (200 μg) by according to previously published work (Han et al., 2009). The immunoprecipitation with anti-CDK2, anti-CDK4 or anti-CDK6 antibodies, and involved sequences of oligonucleotides were as follows: scrambled immunocomplex kinase assays were performed using histone H1 (for CDK2) or siRNA, 50-CCUACGCCACCAAUUUCGU (dTdT)-30;p27Kip1 siRNA, bacterially synthesized GST-Rb fusion protein (for CDK4/6) as substrates. CDK2 and 50-CGACGAUUCUUCUACUCAA (dTdT)-30. The sequences were synthe- CDK4/6 levels were used as loading controls. (A) Phosphorylated substrates were detected and visualized by radioactive image analysis. (B) The kinase activities in sized by GeneChem (Shanghai, China). FOXO3a siRNA and a scrambled each treatment were normalized by the value of immunoprecipitation loading control siRNA (catalog no. SR301619) were purchased from OriGene control and expressed relative to the activity in vehicle-treated cells (defined as (Rockville, MD, USA). LX-2 cells were either seeded into 96-well plates one-fold). Data are presented as mean7S.E.M. of three independent experiments. 3 n o # o at a density of 5 10 cells/well, or into 6-well plates at a density of P 0.05 versus vehicle-treated control group; P 0.05 versus PDGF-BB-treated 4 fi group. IP, immunoprecipitation; IB, immunoblotting. 5 10 cells/well. siRNA oligoduplexes were transfected at 10 nM nal concentration using opti-MEM plus Lipofectamine RNAiMAX Reagents 50 mM HEPES (pH 7.5). Samples were then boiled in 2 sample (Life Technologies, Inc., Carlsbad, CA, USA) according to the manufacturer's instructions. After transfection for 24 h, cells were further buffer for 5 min followed by centrifugation. The clear supernatants μ were obtained and subjected to SDS–polyacrylamide gel electrophor- treated with DMSO or ATG (0.5 M) for 4 h, followed by the stimula- esis on 10% gel. After transferring, the PVDF membranes were probed tion with PDGF-BB (50 ng/ml) for 24 or 48 h. At the end of treatment, with and visualized for phosphotyrosine, phospho-FOXO3a, CDK2, cells were assessed by Western blotting analysis, cell cycle analysis, or CDK4, and CDK6, as described for Western blotting analysis. proliferation assay, as described above. For the kinase activities, whole cell lysates (200 μg) from each sample were pre-cleared with protein A/G-plus agarose beads. 2.11. Statistical analysis Then, the CDK2, CDK4 and CDK6 proteins were immuno-precipitated with anti-CDK2, anti-CDK4 and anti-CDK6 antibodies All results are expressed as mean7standard error of the mean (2 μg) and protein A/G-plus agarose beads (Santa Cruz), respectively. (S.E.M.). Results shown were obtained from at least three inde- The immunoprecipitates were washed for five times: three times in pendent experiments. Inter-group differences were compared ice-cold lysis buffer [150 mM NaCl, 10 mM HEPES (pH 7.4), 0.2% (v/v) with one-way analysis of variance (ANOVA); when differences NP-40], and twice in kinase buffer [50 mM HEPES (pH 7.5), 10 mM were significant by this test (Po0.05), the differences were further magnesium chloride, 1 mM dithiothreitol, 10 mM β-glyceropho- tested using the Tukey post hoc test. The data were analyzed using sphate, 1 mM sodium fluoride, 0.1 mM sodium orthovanadate]. SPSS 16.0 for windows, and Po0.05 was considered to be Subsequently, the beads were incubated at 37 1C for 30 min in statistically significant. A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 77

Fig. 5. ATG increases the formation of CDK2-p27Kip1 complex and decreases the formation of CDK4-p21Cip1 and CDK6-p21Cip1 complexes in PDGF-BB-activated HSCs. Serum- starved LX-2 human HSCs were pretreated with DMSO (vehicle) or ATG (0.25, 0.5, or 1 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for 24 h. Equal amounts of cell lysates (200 μg) were immunoprecipitated (IP) with anti-p21Cip1 or anti-p27Kip1 antibodies, or normal rabbit IgG (negative control), and immunoblotted (IB) with antibodies against CDK2 or CDK4/6. The p21Cip1 and p27Kip1 levels were used as loading controls. (A) and (B) The experiments were repeated three times with similar results, and representative immunoblots are shown for each protein. (C) and (D) The protein levels of CDK2 and CDK4/6 were converted to arbitrary densitometric units, normalized by the value of the corresponding loading control and ﬁnally expressed according to the levels in vehicle-treated cells (deﬁned as one-fold). Data are presented as mean7S.E.M. of three independent experiments. nPo0.05 versus vehicle-treated control group; #Po0.05 versus PDGF-BB-treated group.

3. Results 3.2. ATG induces PDGF-BB-activated HSCs growth arrest in G0/G1 phase of cell cycle 3.1. ATG exhibits no cytotoxic effects on hepatocytes and PDGF-BB- activated HSCs, but it inhibits PDGF-BB-stimulated HSCs proliferation Flow cytometry using PI was performed to analyze the effect of ATG on cell cycle progression. After PDGF-BB stimulation, the flow Hepatocyte damage is generally considered as a triggering factor in cytometic analysis profiles at 24 and 48 h were a larger proportion of pathogenesis of liver fibrosis, antifibrotic agents should be able to cells in S phase and a smaller proportion of cells in G0/G1 phase protect hepatocytes without hepatotoxicity (Kisseleva and Brenner, (Fig. 2A and B), different from that in vehicle-treated condition. A

2007). In this study, the cell toxicity was evaluated with LDH release large proportion of the cells maintained in the G0/G1 phase at 24 assay in hepatocytes and PDGF-BB-activated HSCs. No significant and 48 h when the cells was treated with ATG (0.5 μM) before PDGF- LDH leakage was observed in the hepatocytes and activated HSCs BB stimulation. Compared with cells with PDGF-BB stimulation after treatment with different concentrations of ATG for 48 h alone, the cell population of S phase was significantly decreased in (Fig. 1A and B), suggesting that ATG exhibited no cytotoxic effects the cells treated with ATG (0.5 μM) before PDGF-BB stimulation on hepatocytes and activated HSCs in a wide range of doses (Po0.05). (o16 μM). Moreover, an appropriate dose range of ATG could be obtained for evaluating the inhibitory effects on activated HSCs. 3.3. The up-regulaition of p27Kip1, down-regulation of cyclin D1 and Next, the anti-proliferation potency of ATG on PDGF-BB-activated associated CDKs expression in PDGF-BB-activated HSCs with ATG HSCs was evaluated using Roche Cell Proliferation ELISA. HSCs were treatment pre-incubated with ATG (0.125–2 μM) and then treated with PDGF- BB (50 ng/ml) for 12, 24, or 48 h. The cell proliferation induced by The cell cycle progression is tightly regulated through a series of PDGF-BB was significantly inhibited by increasing concentrations of positive and negative cell cycle regulatory proteins, such as cyclins, ATG (Fig. 1C). The cell proliferation IR% ranged from less than 20% to CDKs and CKIs (Lee et al., 2003; Yu et al., 2004; Sherr and Roberts, more than 80%. The IC50 values was calculated by log-probit regres- 1995). Because ATG attenuated the DNA synthesis of HSCs and thus sion analysis using SPSS statistical software, which were approxi- the transition of G1 to S phase, we speculated ATG might affect the mately 0.291 μM [24 h; 95% confidence limits (CL): 0.219–0.386 μM] expression of cyclins, CDKs and CKIs. As mentioned above, the and 0.253 μM[48h;CL:0.116–0.553 μM]. Hence, the concentrations optimum conditions for G0/G1 cell cycle arrest in activated HSCs by of 0.25, 0.5 and 1 μM were selected to further elucidate possible ATG treatment was 24 h, with a relatively similar effect in 48 h. Thus, mechanisms on ATG's potent inhibitory effect on PDGF-BB-induced whole cell lysates were prepared after ATG pretreatment with HSCs proliferation. different concentrations (0.25, 0.5, or 1 μM) for 4 h and PDGF-BB 78 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87

(50 ng/ml) stimulation for 24 h. The levels of cyclin D1 protein were period. Compared with vehicle control, the expression of CDK4/6 was elevated after PDGF-BB stimulation (Fig. 3A and B), and pretreatment significantly increased in the PDGF-BB incubated HSCs, while the with ATG (0.25, 0.5, or 1 μM) considerably abolished this increase. expression of CDK2 protein was not significantly altered after PDGF-BB However, PDGF-BB did not significantly increase the expression level stimulation (Fig. 3A and C). Furthermore, the levels of CDK4/6 proteins of cyclin E1 relative to vehicle-treated cells during the stimulation induced by PDGF-BB were significantly inhibited by ATG in a A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 79

Kip1 concentration-dependent manner, while the expression of CDK2 was 3.5. p27 is the key regulator for ATG-caused G0/G1 phase arrest not influenced. and cell proliferation inhibition in PDGF-BB-activated HSCs Because the CDKs activities were not only controlled by cyclins Kip1 and CDKs but also CKIs, the levels of two known negative regulator To test the involvement of p27 in ATG-induced G0/G1 phase proteins (p21Cip1 and p27Kip1) were further examined in ATG and/ arrest and cell proliferation inhibition in PDGF-BB activated HSCs, or PDGF-BB-treated HSCs. Consistent with the onset of cell cycle the effects of p27Kip1 knockdown by specific siRNA were examined. arrest, the reduced p27Kip1 level after 24 h of PDGF-BB stimulation Regardless of whether they were pretreated with ATG and/or was elevated in a concentration-dependent manner in ATG- stimulated by PDGF-BB, cells transfected with a scrambled siRNA pretreated HSCs (Fig. 3A and D). However, the level of p21Cip1 showed no significant change in the expression of p27Kip1(Fig. 6A protein was strongly increased by pretreatment with ATG merely and B). Transfection of HSCs with p27Kip1 siRNA prevented the at a higher concentration (1 μM). ATG-induced increase of the p27Kip1 protein level. Accordingly, the

proportion of G0/G1 phase cells after PDGF-BB stimulation was significantly increased in the untransfected or scrambled siRNA transfected cells with ATG addition (Fig. 6C and D). Strikingly, 3.4. ATG inhibits CDKs kinase activities and increases formation of these effects were largely eliminated in cells transfected with Kip1 Kip1 CDK–CKI complexes in PDGF-BB-activated HSCs p27 siRNA, demonstrating that p27 expression is required for the ATG-induced G0/G1 phase arrest. Next, BrdU incorporation assay was performed to consolidate the In cell cycle, the transition from G1 to S phase was primarily fact that the cells coming out of G1 phase were actually participating determined by the sequential activation of CDK4/6 in early G1 in DNA replication and subsequent cell proliferation. The BrdU- phase and CDK2 in mid/late G1 phase (Obaya and Sedivy, 2002). Hence, the effects of ATG on the PDGF-BB-induced activities of positive S phase cells were reduced after pretreatment with ATG Kip1 CDK2 and CDK4/6 were further assessed. The CDKs complexes (Fig. 6E), while the blockage of p27 induction by siRNA transfec- fi were immuno-precipitated with specific anti-CDKs antibodies. The tion signi cantly inhibited the ATG-induced decrease in the propor- Kip1 activities of CDKs kinase were measured with histone H1 (for tion of cells in S phase, suggesting that p27 knockdown indeed CDK2) or bacterially synthesized GST-Rb fusion protein (for CDK4/ facilitated ATG-treated cells to enter S phase and participate in 6) as substrates. After HSCs were treated with PDGF-BB (50 ng/ml) subsequent cell proliferation. fi for 24 h, both CDK2 and CDK4/6 activities were elevated, and the Taken together, these ndings indicate that ATG induces G0/G1 PDGF-BB-mediated increase in CDK2 and CDK4/6 kinase activities phase arrest and the resulting cell proliferation inhibition in PDGF-BB- Kip1 were significantly reduced in ATG-pretreated HSCs (Fig. 4A and B). activated HSCs, at least in part, through the up-regulation of p27 . The inhibition of CDKs activities could result from the up- regulated level of CKIs protein and their binding with CDK–cyclin 3.6. The central role of PI3K/Akt pathway in ATG-induced up- complexes, in addition to decreased amount of cyclins and regulation of p27Kip1 in PDGF-BB-activated HSCs related CDKs in G1 phase (Pavletich, 1999; Roy et al., 2007). Thus, p21Cip1 and p27Kip1 proteins were immuno-precipitated from the It has been reported that PDGF-BB binding to PDGF receptor is whole cell lysates and the interaction of p21Cip1 and p27Kip1 proteins associated with dimerization, autophosphorylation of several tyrosine with CDK2, CDK4 and CDK6 was investigated. In ATG-pretreated residues of PDGF receptor (Hughes et al., 1996). Tyrosine phosphoryla- HSCs, the formation of CDK2-p27Kip1 complex, but not CDK2-p21Cip1, tion of PDGFRβ serves to mediate the binding of various downstream CDK4-p21Cip1,CDK4-p27Kip1,CDK6-p21Cip1 or CDK6-p27Kip1 com- signaling molecules and subsequently activates signal transduction plex, was increased (Fig. 5A–D). Intriguingly, a notable decrease in systems, including PI3K/Akt and MAPKs (Gäbele et al., 2005; Marra the formation of CDK4-p21Cip1 and CDK6-p21Cip1 complexes and et al., 1999; Reif et al., 2003; Si et al., 2008). To further delineate the unchanged formation of CDK2-p21Cip1 complex were observed after molecular mechanisms underlying ATG-induced up-regulation of pretreatment with higher concentration of ATG (1 μM), although p27Kip1 and cell proliferation inhibition, we next examined the effects ATG of identical concentration significantly increased the level of of ATG on downstream signaling cascades induced by PDGF-BB in p21Cip1 protein. From the current results, it seems reasonable to HSCs. PDGFRβ was immunoprecipitated from whole cell lysates, conclude that the repressing effects of ATG on CDK4/6 kinase followed by Western blotting analysis using antibodies against activities might result from the observed reduction in the levels of PDGFRβ and phosphotyrosine. The results showed incubation of cells cyclin D1 and CDK4/6 protein, and that the inhibiton of G1/S with different concentrations of ATG (0.25, 0.5, or 1 μM) did not affect phase transition-related CDK2 kinase activity by ATG treatment is PDGFRβ expression and PDGF-BB-stimulated autophosphorylation of not only attributed to increased formation of CDK2-cyclinE-p27Kip1 precipitated PDGFRβ (Fig. 7A and B). Meanwhile, it was found that complex, but also due to the up-regulation of p27Kip1 instead of PDGF-BB differentially induced ERK1/2, JNK, p38 MAPK, and Akt p21Cip1. phosphorylation and thus activation in HSCs in a time-dependent

Fig. 6. Transfection of p27Kip1 siRNA abolishes the inhibitory effect of ATG on PDGF-BB-induced cell cycle progression and proliferation in HSCs. Serum-starved LX-2 human HSCs were transiently transfected with scrambled siRNA or p27Kip1 siRNA (10 nM) for 24 h. After transfection, cells were treated in the presence or absence of DMSO (vehicle) or ATG (0.5 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for 24 or 48 h. After treatment for 24 h, (A) p27Kip1 knockdown efficiency was determined by Western blotting analysis, and a representative immunoblot shows the expression level of p27Kip1. β-actin served as a sample loading control. (B) Fold-change in relative levels of p27Kip1 protein is shown after normalizing by β-actin. After treatment for 48 h, (C) cell cycle distribution was determined by flow cytometry. Representative flow cytometric histograms demonstrating cell cycle distribution are shown. Control: without treatment; vehicle control: HSCs were incubated with DMSO alone; PDGF- BBþvehicle: HSCs were treated with DMSO before stimulated with PDGF-BB (50 ng/ml); PDGF-BBþvehicleþscrambled siRNA: HSCs were transiently transfected with scrambled siRNA (10 nM) and treated with DMSO before stimulated with PDGF-BB (50 ng/ml); PDGF-BBþvehicleþp27 siRNA: HSCs were transiently transfected with p27Kip1 siRNA (10 nM) and treated with DMSO before stimulated with PDGF-BB (50 ng/ml); PDGF-BBþATG: HSCs were pretreated with ATG (0.5 μM) before stimulated with PDGF-BB (50 ng/ml); PDGF-BBþATGþscrambled siRNA: HSCs were transiently transfected with scrambled siRNA (10 nM) and treated with ATG (0.5 μM) before stimulated with PDGF-BB (50 ng/ml); PDGF-BBþATGþp27 siRNA: HSCs were transiently transfected with p27Kip1 siRNA (10 nM) and treated with ATG (0.5 μM) before stimulated with

PDGF-BB (50 ng/ml). (D) The percentage of the cell cycle within G0/G1, S and G2/M phases was determined by analysis with CellQuest software. (E) After treatment for 48 h, the HSCs proliferation was measured by a cell proliferation ELISA with BrdU. Data are presented as mean7S.E.M. of three independent experiments. nPo0.05 versus vehicle- treated control group; #Po0.05 versus PDGF-BB-treated group; †Po0.05 versus scrambled siRNAþATG þPDGF-BB-treated group. 80 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 manner (Fig. 7C).Attheindicatedtimepoints,ERK1/2,JNK,p38MAPK BB-stimulated Akt phosphorylation at Thr308 residue in a and Akt at Thr308 residue were fully phosphorylated at 1 h after concentration-dependent manner, while expression of total Akt was PDGF-BB stimulation, whereas Akt at Ser473 residue was fully not affected (Fig. 7D and E). Furthermore, PDGF-BB-induced ERK1/2 phosphorylated at 2 h. During the stimulation period, the total protein MAPK phosphorylation were not changed significantly at lower levels of all three MAPKs and Akt were not altered. As observed in concentrations of ATG (0.25 or 0.5 μM), but were dramatically HSCs, pretreatment with ATG for 4 h significantly inhibited PDGF-BB- decreased when the ATG concentration reached 1 μM(Fig. 7Dand induced Akt phosphorylation at Ser473 residue, and it inhibited PDGF- F). In contrast, the level of JNK phosphorylation remained unchanged A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 81 significantly by ATG pretreatment even at a concentration of 1 μM. On revealed that phosphorylated FOXO3a interacts with 14-3-3 under the other hand, PDGF-BB-induced increase in p38 MAPK phosphor- PDGF-BB-stimulated conditions in HSCs (Fig. 8D and E). In the presence ylation could be augmented by pretreatment with different concen- of various concentrations of ATG (0.5 or 1 μM), however, the interac- trations of ATG (0.5 or 1 μM). tion was substantially reduced. To determine whether ERK1/2, p38 MAPK, or PI3K/Akt are Next, we tested whether the effects of ATG on FOXO3a phosphor- involved in the ATG-mediated upreuglation of p27Kip1 in activated ylation coincide with changes in FOXO3a localization. Nuclear and HSCs through the down-regulation of PI3K/Akt and ERK1/2, or the cytoplasmic fractionation experiments indicated that PDGF-BB treat- up-regulation of p38 MAPK, cells were pretreated with DMSO or ment for 4 h would promote the FOXO3a translocation into cyto- ATG for 4 h, or p38 MAPK inhibitor SB203580, MEK1/2 inhibitor plasm, as reflected by a simultaneous increase in cytoplasmic FOXO3a U0126, PI3K inhibitor LY294002, or Akt inhibitor PHT-427 for 1 h, and decrease in nuclei (Fig. 8A and C). Pretreatment with ATG and then incubated with 50 ng/ml PDGF-BB for 48 h. Relative to (0.5 μM) for 4 h notably reduced this PDGF-BB-induced cytoplasmic PDGF-BB-induced control, pretreatment of HSCs with either ATG, translocation. To confirm the subcellular fractionation results, immu- or inhibitors of PI3K or Akt, but not inhibitor of MEK1/2, which nofluorescence staining was then used to directly visualize the operates upstream of ERK1/2 MAPK, effectively reversed PDGF-BB- subcellular distribution of FOXO3a. The results showed that endogen- induced downregulation of p27Kip1 (Fig. 7G and H). However, ous FOXO3a was detectable in both the cytoplasmic and nuclear coincubation p38 MAPK inhibitor with ATG did not show the compartments of vehicle-treated cells, and it resided almost exclu- inhibitory effect on ATG-mediated up-regulation of p27Kip1 in the sively in the cytosol upon PDGF-BB stimulation, whereas nuclear presence of PDGF-BB. Together, our results suggest that in acti- staining was negligible (Fig. 8G). However, pretreatment of cells with vated HSCs, ATG prevents PDGF-stimulated p27Kip1 down- ATG resulted in a translocation of FOXO3a into the nucleus, although regulation mainly by antagonizing PI3K/Akt pathway rather than there was still some punctate perinuclear FOXO3a staining remained MAPKs pathway. apparent. In contrast, the subcellular distribution of FOXO1 and FOXO4 after PDGF-BB stimulation was not influenced by the addition of ATG (Fig. 8F and H). In short, all of the above results support at least 3.7. ATG prevents PDGF-BB-induced phosphorylation of FOXO3a and in part the notion that nuclear translocation of FOXO3a may be a binding of FOXO3a to 14-3-3 protein, and results in FOXO3a nuclear necessary prerequisite for subsequent up-regulation of p27Kip1,in translocation in PDGF-BB-activated HSCs response to ATG treatment.

Forkhead box O (FOXO) transcription factors have been identiﬁed as direct targets of Akt and their transcriptional activities would be lost through phosphorylation by Akt kinase. Akt-mediated phosphorylation 3.8. ATG overcomes PI3K/Akt activities to inhibit PDGF-BB-induced of FOXO transcription factors leads to the release of them from the DNA FOXO3a phosphorylation in PDGF-BB-activated HSCs and translocation of those factors to the cytoplasm, where 14-3-3 chaperone protein binds to the phosphorylated FOXOs and retains The experiments described above suggest that ATG blocks them as inactive proteins, preventing them from initiating expression PDGF-BB-induced Akt phosphorylation. To determine whether of p27Kip1 (Dijkers et al., 2000; Fu and Tindall, 2008; Van der Horst and these effects are responsible for ATG suppressed FOXO3a phos- Burgering, 2007). To investigate the downstream events regulated by phorylation in PDGF-BB-activated HSCs, cells were incubated with ATG-mediated Akt dephosphorylation in activated HSCs, the phosphor- 50 ng/ml PDGF-BB in the presence or absence of PI3K/Akt signal- ylation levels of FOXOs and the association of FoxO3a with 14-3-3 ing inhibitors LY294002 and PHT-427. The two inhibitors nearly protein were examined. As illustrated in Fig. 8A and B, in parallel with completely inhibited PDGF-BB-induced phosphorylation of Akt Akt phosphorylation in HSCs, PDGF-BB (50 ng/ml) induced the phos- and its downstream target FOXO3a (Fig. 9A and B). In this study, phorylation of FOXOs at the three Akt-sensitive residues, namely we also combined ATG with LY294002 or PHT-427, and our results FOXO1 on Thr24, FOXO3a on Thr32 and FOXO4 on Ser262. However, showed that there was no statistical difference between PI3K/Akt ATG signiﬁcantly inhibited PDGF-BB-induced FOXO3a phosphorylation inhibitor and ATG in combination with PI3K/Akt inhibitor at the Thr32 site in a concentration-dependent manner, but neither (LY294002 or PHT-427) about the inhibition of PDGF-BB-induced FOXO1 nor FOXO4 were affected. Immunoprecipitation analysis phophorylation of Akt and FOXO3a.

Fig. 7. ATG induction of p27Kip1 expression is mediated through the inactivation of PI3K/Akt but not MAPKs pathway in PDGF-BB-activated HSCs. (A) Effect of ATG on PDGFRβ phosphorylation in activated HSCs. Serum-starved LX-2 human HSCs were pretreated with DMSO (vehicle) or ATG (0.25, 0.5, or 1 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for 0.5 h. Equal amounts of cell lysates (200 μg) were immunoprecipitated (IP) with anti-PDGFRβ antibody, and normal rabbit IgG (2 μg) was used as negative control. The immunoprecipitates were collected, and protein tyrosine phosphorylation (p-Tyr) and total PDGFRβ were determined by Western blotting analysis. The 180-kDa phosphotyrosine band was identified as tyrosine-phosphorylated PDGFRβ. Total protein level of PDGFRβ served as an internal control. The experiments were repeated three times with similar results and representative immunoblots are shown for each protein. (B) The protein level of phosphorylated PDGFRβ in whole cell lysates, was converted to arbitrary densitometric units, normalized by the value of total PDGFRβ and expressed relative to the phosphorylation ratio in vehicle- treated cells (defined as one-fold). (C) Time course of PDGF-BB-induced MAPKs and Akt activation. Serum-starved LX-2 human HSCs were exposed to 50 ng/ml PDGF-BB for the indicated times. After treatment, whole cell lysates were prepared and subjected to Western blotting analysis using antibodies specific for Akt, ERK1/2, JNK, p38 MAPK, and their phosphorylated forms. Total protein levels of Akt and MAPKs were used as internal controls. (D) Effects of ATG on the phosphorylation of Akt and MAPKs in PDGF- BB-activated HSCs. Serum-starved LX-2 human HSCs were pretreated with DMSO (vehicle) or the indicated concentrations of ATG (0.25, 0.5, or 1 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for either 1 h (to detect phosphorylation of MAPKs and Akt at Thr308 residue) or 2 h (to detect phosphorylation of Akt at Ser473 residue). Whole cell lysates were prepared and subjected to Western blotting analysis using antibodies against the phosphorylated forms or all forms of Akt and MAPKs. Total protein levels of Akt and MAPKs served as internal controls. The experiments were repeated three times with similar results and a representative blot was shown for each protein. (E) and (F) The protein levels of phosphorylated Akt and MAPKs were converted to arbitrary densitometric units, normalized by the value of the corresponding internal control and expressed relative to the phosphorylation ratio in vehicle-treated cells (defined as one-fold). (G) Serum-starved LX-2 human HSCs were pretreated with DMSO (vehicle) or ATG (0.5 μM) for 4 h, or p38 MAPK inhibitor SB203580 (20 μM), MEK1/2 inhibitor U0126 (20 μM), PI3K inhibitor LY294002 (10 μM), or Akt inhibitor PHT-427 (20 μM) for 1 h, and then incubated with or without 50 ng/ml PDGF-BB for 48 h. After treatment, aliquots of whole cell lysates were collected and subjected to Western blotting analysis. β-Actin served as an internal control for equal protein loading. The experiments were repeated three times with similar results and a representative immunoblot for p27Kip1 is shown. (H) Fold-change in relative level of p27Kip1 protein is shown after normalization with β-actin. Data are presented as mean7S.E.M. of three independent experiments. nPo0.05 versus vehicle-treated control group; #Po0.05 versus PDGF-BB-treated group. 82 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 83

3.9. Knockdown of FOXO3a expression by FOXO3a siRNA attenuates cell cycle analysis using flow cytometry. ATG treatment caused a Kip1 ATG-induced up-regulation of p27 in PDGF-BB-activated HSCs significant accumulation of cells in the G0/G1 phase and reduction in numbers of cell in the S phase. The experiments described above suggest that ATG inhibited Generally the cell cycle is regulated by the coordinated action PDGF-BB-induced FOXO3a phosphorylation and translocation into of CDKs in association with their specific regulatory cyclin proteins the cytoplasm. To provide more direct evidence that in ATG-treated (Sherr, 1996). Thus, functional activation of CDK–cyclin complexes

HSCs, the induction of FOXO3a nuclear translocation correlated with kinase activities is required for the G1 to S phase transition. In Kip1 up-regulation of p27 , we transiently transfected cells with FOXO3a mammalian cells, the key G1 phase CDKs are CDK4 (and its siRNA and scrambled control siRNA, and then pretreated them with functional counterpart, CDK6) and CDK2. In early G1 phase, growth ATG. After stimulating the cells with PDGF-BB, we measured levels of factor signaling exerts a positive effect on the synthesis of cyclin D. p27Kip1 and nuclear FOXO3a proteins by Western blotting. The Upon stimulation, cyclin D bind and activate CDK proteins CDK4/6, expression level of nuclear FOXO3a was decreased at 24 h after the and the cyclin D–CDK4/6 complexes mediate progression of the transfection of FOXO3a siRNA at 10 nM in ATG-pretreated and/or cell cycle in early G1 phase (Obaya and Sedivy, 2002). In our study, PDGF-BB-activated HSCs, while a reduction of nuclear FOXO3a was the kinase activities of cyclin D–CDK4/6 complexes were inhibited not observed by transfection of scrambled siRNA (Fig. 10A). Mean- by ATG pretrement in PDGF-BB-activeated HSCs. Meanwhile, ATG- while, knockdown of FOXO3a by siRNA transfection strikingly atte- pretreatment resulted in a considerable decrease in the levels of nuated ATG-mediated increase in the level of p27Kip1 in PDGF-BB- cyclin D1 and CDK4/6 protein, suggesting reduced activities of activated HSCs, compared with that of cells treated only with PDGF- CDK4/6 by ATG are associated with down-regulation of cyclin D1

BB by transfecting the cells with scrambled siRNA (Fig. 10BandC).In and CDK4/6 expression during early G1 phase. It has been shown contrast, cells transfected with scrambled siRNA showed similar levels cyclin E, a transcriptional target dependent on cyclin D–CDK4/6 Kip1 of p27 protein after PDGF-BB stimulation, irrespective of whether activation, in turn activates cyclin E/CDK2 complex during late G1 they were pretreated with ATG. These results suggest that ATG up- phase, and is required for S phase entry (Lundberg and Weinberg, regulated expression of p27Kip1, at least in part, by increasing nuclear 1998; Ohtsubo et al., 1995). In this study, the activity of cyclin accumulation of FOXO3a in PDGF-BB-activated HSCs. E–CDK2 complex was also significantly decreased by ATG pretreatment. Since there was little change in the levels of cyclin E1 and CDK2 proteins after 24 h of ATG exposure, it seems that the 4. Discussion decrease in the activity of cyclin E–CDK2 complex in ATG-treated HSCs is not due to changes in the levels of these proteins. During the progression of liver fibrosis, proliferation of HSCs The interaction between the cyclin–CDK complexes and CKIs is plays an important role in synthesis of ECM. The inflammatory another important mechanism for the regulation of cyclin–CDK response in the injured liver has been considered to activate HSCs activity. In the progression of the cell, the CKIs p21Cip1 and p27Kip1 proliferation (King and Goodell, 2011; Schwabe et al., 2003). As a bind to cyclin–CDK complexes to inhibit their catalytic activity and potent anti-inflammatory agent, ATG could exert its effects on induce cell cycle arrest (Kossatz and Malek, 2007; Xiong et al., lipopolysaccharide (LPS)-stimulated murine macrophages in vitro, 1993). In the current study, the levels of p27Kip1 levels were colitis in mice and experimental glomerulonephritis in rats (Cho markedly increased by ATG pretreatment. Because the cyclin et al., 2002; Hyam et al., 2013; Wu et al., 2009; Zhao et al., 2009). E–CDK2 kinase activity is essential for driving the cells through Kip1 In the present study, the effects of ATG on the proliferation of G1-S transition, concomitant with p27 induction, we also activated HSCs and underlying mechanisms were explored. observed an increase in CDK2-p27Kip1 binding together with an Cell proliferation relies on DNA synthesis and replication. In inhibition of cyclin E–CDK2 kinase activity in ATG-treated cells. this study, a colorimetric BrdU cell proliferation assay was used to Similarly, the protein level of p21Cip1, another universal inhibitor measure DNA synthesis, and the results showed ATG treatment of CDKs, was also increased by ATG pretreatment for 24 h, but at a decreased the number of activated HSCs that incorporated BrdU in higher concentration (1 μM). Under the same conditions, a sig- a concentration- and time-dependent manner, indicating that nificant decreased formation of CDK4-p21Cip1 and CDK6-p21Cip1 PDGF-BB-induced HSCs proliferation could be inhibited by ATG. complexes and unchanged formation of CDK2-p21Cip1 complex Furthermore, pharmacological inhibition of cell proliferation may were also observed. These results indicate that ATG-mediated be attributed to cytotoxicity-associated cellular injuries (Gomez- induction of p27Kip1, instead of p21Cip1, was responsible for the Lechon et al., 2008). However, LDH release assay proved that the decrease in the activity of cyclin E–CDK2 complex and for the anti-proliferative effect of ATG did not result from cytoclasis or cell subsequent cell cycle arrest. disruption, even up to the concentration of 16 μM. This caused us PDGF acts as a potent mitogen for HSCs proliferation through to question whether the proliferation inhibition induced by ATG the interaction with PDGF receptor (Pinzani et al., 1989). The may be mediated through cell cycle events. This was supported by binding of PDGF to its cellular receptor, a 180-kDa transmembrane

Fig. 8. ATG inhibits PDGF-BB-induced FOXO3a phosphorylation and binding of FOXO3a to 14-3-3 protein, leading to nuclear localization of FOXO3a protein in PDGF-BB- activated HSCs. Serum-starved LX-2 human HSCs were pretreated with DMSO (vehicle) or the indicated concentrations of ATG (0.25, 0.5, or 1 μM) for 4 h, and then incubated with or without 50 ng/ml PDGF-BB for 4 h. After treatment, whole cell lysates were collected and subjected to Western blotting analysis using antibodies specific for FOXO1, FOXO3a and FOXO4, and their phosphorylated forms. Total protein levels of FOXOs served as internal controls. Meanwhile, cytoplasmic and nuclear fractions were collected and analyzed by Western blotting to detect FOXO3a. β-Actin and lamin B1 protein levels were used as internal controls, respectively, for cytoplasmic extracts (CE) and nuclear extracts (NE). (A) The experiments were repeated three times with similar results, and representative immunoblots are shown for each protein. (B) The protein levels of phosphorylated FOXOs in whole cell lysates, and (C) of FOXO3a in cytoplasmic and nuclear extracts, were converted to arbitrary densitometric units, normalized by the value of the corresponding internal control and expressed relative to the phosphorylation ratio or to the protein level in vehicle-treated cells (defined as one-fold). (D) Equal amounts of cytoplasmic fractions (200 μg) were immunoprecipitated (IP) with anti-14-3-3 antibody, or normal rabbit IgG (negative control), and immunoblotted (IB) with antibody specific for phosphorylated FOXO3a. The protein level of 14-3-3 was used as an internal control. The experiments were repeated three times with similar results and a representative immunoblot for phosphorylated FOXO3a is shown. (E) Fold-change in relative level of phosphorylated FOXO3a protein is shown after normalizing by 14- 3-3. (F)–(H) Immunofluorescence staining was performed to detect the distribution and expression of FOXO1, FOXO3a and FOXO4 in the cytoplasm and nuclei of cells. After treatment, cells were incubated with primary antibodies as indicated, followed by FITC-conjugated goat anti-rabbit IgG (green). Nuclei were visualized with the nuclear dye PI (red). Finally, cells were visualized and photographed using a Leica TCS-SP2 confocal laser scanning microscope at 400 magnification. Scale Bar, 20 μm. Data are presented as mean7S.E.M. of three independent experiments. nPo0.05 versus vehicle-treated control group; #Po0.05 versus PDGF-BB-treated group. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 84 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87

Fig. 9. The inhibition of PI3K/Akt pathway is involved in ATG-induced FOXO3a dephosphorylation in PDGF-BB-activated HSCs. Serum-starved LX-2 human HSCs were pretreated with or without DMSO (vehicle) or ATG (0.5 μM) for 4 h, in the Fig. 10. Effect of knockdown of p27Kip1 expression by p27Kip1 siRNA on ATG- absence or presence of PI3K inhibitor LY294002 (20 μM) or Akt inhibitor PHT-427 induced up-regulation of p27Kip1 in activated HSCs. Serum-starved LX-2 human (20 μM), and then incubated with or without 50 ng/ml PDGF-BB for 4 h. After HSCs were transiently transfected with scrambled control siRNA or FOXO3a siRNA treatment, aliquots of whole cell lysates were collected and subjected to Western (10 nM). After 24 h transfection, cells were pretreated with or without DMSO blotting analysis. (A) The experiments were repeated three times with similar (vehicle) or ATG (0.5 μM) for 4 h, and then incubated in the presence or absence of results and a representative blot was shown for each protein. Total protein levels of 50 ng/ml PDGF-BB for 24 or 48 h. After treatment for 24 h, (A) nuclear fractions Akt and FOXO3a served as internal controls. (B) Fold-changes in relative levels of were collected and subjected to Western blotting analysis to detect FOXO3a. Lamin Akt and FOXO3a phosphorylation ratio are shown after normalizing to the B1 protein level was used as an internal control for nuclear extracts (NE). After corresponding internal control. Data are presented as mean7S.E.M. of three treatment for 48 h, whole cell lysates were subjected to Western blotting analysis independent experiments. nPo0.05 versus vehicle-treated control group; with anti-p27Kip1 antibody. (B) The experiments were repeated three times with #Po0.05 versus PDGF-BB-treated group. similar results and a representative immunoblot for p27Kip1 is shown. β-Actin served as a loading control. (C) Fold-change in relative levels of p27Kip1 protein is shown after normalizing with β-actin. Data are presented as mean7S.E.M. of three glycoprotein, initiates a cascade of events beginning with rapid independent experiments. nPo0.05 versus vehicle-treated control group; #Po0.05 † activation of the protein-tyrosine kinase activity of the PDGF versus PDGF-BB-treated group; Po0.05 versus scrambled siRNAþATG þPDGF-BB- receptor, and autophosphorylation of the receptor on tyrosine treated group. residues, which in turn recruits and phosphorylates downstream signaling molecules, including PI3K/Akt and MAPKs. PI3K/Akt is PDGF-BB-stimulated phosphorylation and thus activation of Akt known as one of the central downstream signal transduction and ERK1/2, whereas it did not affect the phosphorylation of JNK. pathways for HSCs proliferation and fibrosis progression (Gäbele In contrast, the PDGF-BB-stimulated p38 MAPK phosphorylation et al., 2005; Reif et al., 2003). Among MAPK family, ERK1/2 has was further enhanced by pretreatment of ATG. This is in agree- been implicated in the proliferation and migration of HSCs (Marra ment with the previous report showing that activation of the p38 et al., 1999; Si et al., 2008). Upon PDGF stimulation, activated MAPK cascade led to a significant decrease in cyclin D1 transcrip- ERK1/2 is translocated into nuclei, where it phosphorylates several tion and protein level and, conversely, inhibition of this signaling cell proliferation-associated transcription factors, such as Elk-1 pathway by the specific inhibitor SB203580 had an opposite and SAP (Pagès et al., 1993). JNK and p38 MAPK seem to exert enhancing effect in lung fibroblast cells (Lavoie et al., 1996). These divergent effects on HSCs proliferation. JNK stimulated HSCs results suggest that ATG inhibits PDGF-BB-activated Akt and only proliferation, while p38 MAPK inhibited the proliferation rate of certain MAPK family members. HSCs (Schnabl et al., 2001). We showed that ATG had no detect- Using RNA interference-mediated knockdown of p27Kip1,we able effect on PDGFRβ expression and PDGF-BB-induced PDGFRβ have found that the increase in p27Kip1 accumulation is involved in autophosphorylation, suggesting that PDGF receptor may not be G0/G1 phase arrest and suppression of proliferation induced by the direct target of ATG for inhibition of HSCs proliferation. ATG in activated HSCs. Specific chemical inhibitors were then used Furthermore, our results showed ATG significantly inhibited to delineate the signaling pathways involved in ATG-mediated A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 85

Fig. 11. The scheme of proposed molecular mechanisms underlying the effect of ATG on PI3K/Akt/FOXO3a/p27Kip1 pathway and regulation of PDGF-BB-induced HSCs proliferation. ATG exerted an anti-proliferative effect on activated HSCs and caused G0/G1 cell cycle arrest. CDK4/6 kinase activities were inhibited by ATG at non-cytotoxic Cip1 concentrations in early G1 cell cycle that was independent of p21 induction, but instead it could be attributed to the decrease in cyclin D1 and CDK4/6 protein levels. Meanwhile, ATG treatment inhibited PDGF-BB-induced phosphorylation of Akt and downstream transcription factor FOXO3a, decreased binding of FOXO3a to 14-3-3 protein, promoted nuclear accumulation of FOXO3a, and then up-regulated p27Kip expression. This was accompanied by elevated the formation of p27Kip1-cyclin E–CDK2 complex, which in turn inhibited CDK2 kinase activity, and eventually interrupted the progress of the cell cycle. Priority areas for future research were indicated with purple fonts and question marks. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) up-regulation of p27Kip1. Our molecular evidence showed that and combining ATG with PI3K/Akt inhibitor, implying that the either PI3K inhibitor LY294002 or Akt inhibitor PHT-427, but not inhibitory effect of ATG on PDGF-BB-induced phosphorylation of MEK inhibitor U0126, could independently mimic the ATG effects. FOXO3a might be attributed mainly to its capacity to suppress Subsequent evidence that pharmacological blockade of p38 MAPK PI3K/Akt activation. Third, ATG treatment of the activated HSCs pathway by p38 MAPK inhibitor SB203580 had no effect on up- also resulted in 14-3-3 dissociation from FOXO3a and re- regulation of p27Kip1 in ATG-pretreated HSCs, indicating that ATG localization of FOXO3a to the nucleus, thereby strongly implicating induction of p27Kip1 was associated with disruption of PDGF ATG in the regulation of FOXO3a nuclear localization. Fourth, receptor-mediated PI3K/Akt pathway rather than MAPK pathways. silence of FOXO3a efficiently reversed ATG-elevated p27Kip1 These findings were consistent with observations from earl- expression, suggesting that FOXO3a played an important role in ier studies, for example, PI3K/Akt inhibition by tangeretin was ATG up-regulation of p27Kip1 in activated HSCs. Nevertheless, the reported to suppress PGDF-BB-induced proliferation of aortic up-regulaed level of p27Kip1 by ATG was partially rescued by smooth muscle cells (Seo et al., 2011). FOXO3a silencing compared with PDGF-BB stimulation alone. The best-characterized nuclear substrates of Akt are transcrip- Previous studies have shown that p27Kip1 promoter activity could tion factors of the FOXO family: FOXO1/FKHR, FOXO3a/FKHRL1, be regulated by other transcription factors, such as Krüppel like and FOXO4/AFX (Tzivion et al., 2011). In the presence of survival factor 4 (KLF-4), peroxisome proliferator-activated receptor factors, including PDGF-BB, Akt becomes activated, phosphorylates δ (PPARδ) and CREB-binding protein (CBP) (Kee et al., 2011; Sue FOXOs at several regulatory sites. Phosphorylation of FOXOs causes et al., 2009). Thus, the failure to obtain complete rescue can be nuclear exclusion and their association with 14-3-3, thereby explained by presumption that other factors other than FOXO3a inhibiting FOXOs-dependent transcription (Dijkers et al., 2000; may also be associated with ATG-enhanced p27Kip1 expression. Medema et al., 2000). When Akt is inactive, FOXOs are depho- sphorylated and localized in the nucleus, where they activate transcription target genes such as p27Kip1 and triggers cell cycle 5. Conclusions arrest. We identified four vital functions of ATG that inactivate PI3K/Akt signaling and can employ Forkhead family member In summary, we have shown here that ATG suppresses PDGF- Kip1 FOXO3a to initiate transcription of the p27 gene. First, we BB-induced HSCs proliferation and arrested cell cycle at G0/G1 demonstrated that the inhibition of PI3K/Akt by ATG led to phase, and occurs in parallel with increased the protein levels of decreased phosphorylation of FOXO3a at Akt sites, but not FOXO1 p21Cip1 and p27Kip1, and decreased cyclin D1 and CDK4/6, resulting and FOXO4. Second, in order to see if there is a separate effect of in decreased activities of cyclin E–CDK2 and cyclin D–CDK4/6 ATG on pathways other than PI3K/Akt, we also observed the effect complexes. We also found that ATG decreased PDGF-BB-induced of combining ATG with LY294002 or PHT-427 on PDGF-BB- phosphorylation of Akt and FOXO3a, increased nuclear accumula- induced phosphorylation of Akt and FOXO3a. Our results showed tion, and activation of FOXO3a, and then elevated the expression of that there was no statistical difference between PI3K/Akt inhibitor p27Kip1, subsequently inhibiting the proliferation of HSCs. 86 A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87

Treatment with the PI3K/Akt inhibitor had a similar effect on up- Kossatz, U., Malek, N.P., 2007. p27: tumor suppressor and oncogene…? Cell Res. 17, regulation of p27Kip1 in HSCs. These results suggest that ATG may 832–833. Lavoie, J.N., L’Allemain, G., Brunet, A., Müller, R., Pouysségur, J., 1996. Cyclin D1 act as PI3K/Akt pathway blocker and contribute to the elevation of expression is regulated positively by the p42/p44MAPK and negatively by the Kip1 p27 and the reduction in proliferation of HSCs (Fig. 11). Thus, p38/HOGMAPK pathway. J. Biol. Chem. 271, 20608–20616. our findings suggest an underlying molecular mechanism respon- Lee, J.Y., Cho, B.J., Park, T.W., Park, B.E., Kim, S.J., Sim, S.S., Kim, C.J., 2010. Dibenzylbutyrolactone lignans from Forsythia koreana fruits attenuate sible for the anti-proliferative effect of ATG in HSCs and may be lipopolysaccharide-induced inducible nitric oxide synthetase and useful for the development of new prophylactic strategies for the cyclooxygenase-2 expressions through activation of nuclear factor-κB and prevention of liver fibrosis. mitogen-activated protein kinase in RAW264. 7 cells. Biol. Pharm. Bull. 33, 1847–1853. Lee, W.S., Liu, C.W., Juan, S.H., Liang, Y.C., Ho, P.Y., Lee, Y.H., 2003. Molecular mechanism of progesterone-induced antiproliferation in rat aortic smooth Acknowledgments muscle cells. Endocrinology 144, 2785–2790. Lotersztajn, S., Julien, B., Teixeira-Clerc, F., Grenard, P., Mallat, A., 2005. Hepatic fibrosis: molecular mechanisms and drug targets. Annu. Rev. Pharmacol. This work was supported by the National Natural Science Toxicol. 45, 605–628. Foundation of China (No. 81001675), the Chongqing Natural Lundberg, A.S., Weinberg, R.A., 1998. Functional inactivation of the retinoblastoma fi Science Foundation (No. cstc2014jcyjA10030), the Scientific and protein requires sequential modi cation by at least two distinct cyclin-cdk complexes. Mol. Cell. Biol. 18, 753–761. Technological Research Program of Chongqing Municipal Educa- Marra, F., Arrighi, M.C., Fazi, M., Caligiuri, A., Pinzani, M., Romanelli, R.G., Efsen, E., tion Commission (No. KJ120820) and the Undergraduate Training Laffi, G., Gentilini, P., 1999. Extracellular signal-regulated kinase activation Programs for Innovation and Entrepreneurship in Chongqing (No. differentially regulates platelet-derived growth factor's actions in hepatic stellate cells, and is induced by in vivo liver injury in the rat. Hepatology 30, 201311660024). 951–958. Medema, R.H., Kops, G.J., Bos, J.L., Burgering, B.M., 2000. AFX-like Forkhead transcription factors mediate cell-cycle regulation by Ras and PKB through p27kip1. Nature 404, 782–787. Appendix A. Supplementary information Obaya, A., Sedivy, J., 2002. Regulation of cyclin-Cdk activity in mammalian cells. Cell. Mol. Life Sci. 59, 126–142. Supplementary data associated with this article can be found in Ohtsubo, M., Theodoras, A.M., Schumacher, J., Roberts, J.M., Pagano, M., 1995. the online version at http://dx.doi.org/10.1016/j.ejphar.2014.11.040. Human cyclin E, a nuclear protein essential for the G1-to-S phase transition. Mol. Cell. Biol. 15, 2612–2624.

Owa, T., Yoshino, H., Yoshimatsu, K., Nagasu, T., 2001. Cell cycle regulation in the G1 phase: a promising target for the development of new chemotherapeutic References anticancer agents. Curr. Med. Chem. 8, 1487–1503. Pagès, G., Lenormand, P., L’Allemain, G., Chambard, J.C., Meloche, S., Pouysségur, J., Bataller, R., Brenner, D.A., 2005. Liver fibrosis. J. Clin. Investig. 115, 209–218. 1993. Mitogen-activated protein kinases p42mapk and p44mapk are required fi – Cho, M.K., Jang, Y.P., Kim, Y.C., Kim, S.G., 2004. Arctigenin, a phenylpropanoid for broblast proliferation. Proc. Natl. Acad. Sci. USA 90, 8319 8323. dibenzylbutyrolactone lignan, inhibits MAP kinases and AP-1 activation via Pavletich, N.P., 1999. Mechanisms of cyclin-dependent kinase regulation: structures potent MKK inhibition: the role in TNF-α inhibition. Int. Immunopharmacol. 4, of cdks, their cyclin activators, and cip and INK4 inhibitors. J. Mol. Biol. 287, – 1419–1429. 821 828. Cho, M.K., Park, J.W., Jang, Y.P., Kim, Y.C., Kim, S.G., 2002. Potent inhibition of Peng, Y., Yang, H., Wang, N., Ouyang, Y., Yi, Y., Liao, L., Shen, H., Hu, G., Wang, Z., Tao, fi lipopolysaccharide-inducible nitric oxide synthase expression by dibenzylbu- L., 2013. Fluorofenidone attenuates hepatic brosis by suppressing the prolif- – tyrolactone lignans through inhibition of I-κBα phosphorylation and of p65 eration and activation of hepatic stellate cells. Am. J. Physiol. Gastrointest. – nuclear translocation in macrophages. Int. Immunopharmacol. 2, 105–116. Liver Physiol. 306, G253 G263. Dijkers, P.F., Medema, R.H., Pals, C., Banerji, L., Thomas, N.S.B., Lam, E.W.F., Pinzani, M., Gesualdo, L., Sabbah, G.M., Abboud, H., 1989. Effects of platelet-derived Burgering, B.M., Raaijmakers, J.A., Lammers, J.W.J., Koenderman, L., 2000. growth factor and other polypeptide mitogens on DNA synthesis and growth of Forkhead transcription factor FKHR-L1 modulates cytokine-dependent tran- cultured rat liver fat-storing cells. J. Clin. Investig. 84, 1786. scriptional regulation of p27KIP1. Mol. Cell. Biol. 20, 9138–9148. Reif, S., Lang, A., Lindquist, J.N., Yata, Y., Gäbele, E., Scanga, A., Brenner, D.A., Rippe, Egozi, D., Shapira, M., Paor, G., Ben-Izhak, O., Skorecki, K., Hershko, D.D., 2007. R.A., 2003. The role of focal adhesion kinase-phosphatidylinositol 3-kinase-akt Regulation of the cell cycle inhibitor p27 and its ubiquitin ligase Skp2 in signaling in hepatic stellate cell proliferation and type I collagen expression. – differentiation of human embryonic stem cells. FASEB J. 21, 2807–2817. J. Biol. Chem. 278, 8083 8090. Friedman, S.L., 2010. Evolving challenges in hepatic fibrosis. Nat. Rev. Gastroenterol. Roy, S., Kaur, M., Agarwal, C., Tecklenburg, M., Sclafani, R.A., Agarwal, R., 2007. P21 Hepatol. 7, 425–436. and p27 induction by silibinin is essential for its cell cycle arrest effect in – Fu, Z., Tindall, D., 2008. FOXOs, cancer and regulation of apoptosis. Oncogene 27, prostate carcinoma cells. Mol. Cancer Ther. 6, 2696 2707. 2312–2319. Schnabl, B., Bradham, C.A., Bennett, B.L., Manning, A.M., Stefanovic, B., Brenner, D.A., Gäbele, E., Reif, S., Tsukada, S., Bataller, R., Yata, Y., Morris, T., Schrum, L.W., Brenner, 2001. TAK1/JNK and p38 have opposite effects on rat hepatic stellate cells. – D.A., Rippe, R.A., 2005. The role of p70S6K in hepatic stellate cell collagen gene Hepatology 34, 953 963. expression and cell proliferation. J. Biol. Chem. 280, 13374–13382. Schwabe, R.F., Bataller, R., Brenner, D.A., 2003. Human hepatic stellate cells express – Gomez-Lechon, M.J., OConnor, J.E., Lahoz, A., Castell, J.V., Donato, M.T., 2008. CCR5 and RANTES to induce proliferation and migration. Am. J. Physiol. – Identification of apoptotic drugs: multiparametric evaluation in cultured Gastrointest. Liver Physiol. 285, G949 G958. hepatocytes. Curr. Med. Chem. 15, 2071–2085. Seo, J., Lee, H.S., Ryoo, S., Seo, J.H., Min, B.S., Lee, J.H., 2011. Tangeretin, a citrus fl Han, Y.H., Moon, H.J., You, B.R., Kim, S.Z., Kim, S.H., Park, W.H., 2009. Effects of avonoid, inhibits PGDF-BB-induced proliferation and migration of aortic – carbonyl cyanide p-(trifluoromethoxy) phenylhydrazone on the growth inhibi- smooth muscle cells by blocking AKT activation. Eur. J. Pharmacol. 673, 56 64. tion in human pulmonary adenocarcinoma Calu-6 cells. Toxicology 265, Sheng, Z., Ma, L., Sun, J.E., Zhu, L.J., Green, M.R., 2011. BCR-ABL suppresses 101–107. autophagy through ATF5-mediated regulation of mTOR transcription. Blood – Hughes, A., Clunn, G., Refson, J., Demoliou-Mason, C., 1996. Platelet-derived growth 118, 2840 2848. – factor (PDGF): actions and mechanisms in vascular smooth muscle. Gen. Sherr, C.J., 1996. Cancer cell cycles. Science 274, 1672 1677. Pharmacol. – Vasc. Syst. 27, 1079–1089. Sherr, C.J., Roberts, J.M., 1995. Inhibitors of mammalian G1 cyclin-dependent – Hui, A.Y., Friedman, S.L., 2003. Molecular basis of hepatic fibrosis. Expert Rev. Mol. kinases. Gene Dev. 9, 1149 1163. fl Med. 5, 1–23. Si, H., Lv, X., Guo, A., Jiang, H., Li, J., 2008. Suppressive effect of le unomide on rat Hyam,S.R.,Lee,I.A.,Gu,W.,Kim,K.A.,Jeong,J.J.,Jang,S.E.,Han,M.J.,Kim,D.H.,2013. hepatic stellate cell proliferation involves on PDGF-BB-elicited activation of – Arctigenin ameliorates inflammation in vitro and in vivo by inhibiting the PI3K/AKT three mitogen-activated protein kinases. Cytokine 42, 24 31. pathway and polarizing M1 macrophages to M2-like macrophages. Eur. J. Pharma- Son, M.K., Ryu, Y.L., Jung, K.H., Lee, H., Lee, H.S., Yan, H.H., Park, H.J., Ryu, J.K., Suh, J. col. 708, 21–29. K., Hong, S., 2013. HS-173, a novel PI3K inhibitor, attenuates the activation of fi Kee, H.J., Kwon, J.S., Shin, S., Ahn, Y., Jeong, M.H., Kook, H., 2011. Trichostatin A hepatic stellate cells in liver brosis. Sci. Rep. 3, 3470. Sue, Y.M., Chung, C.P., Lin, H., Chou, Y., Jen, C.Y., Li, H.F., Chang, C.C., Juan, S.H., 2009. prevents neointimal hyperplasia via activation of Krüppel like factor 4. Vasc. δ Pharmacol. 55, 127–134. PPAR -mediated p21/p27 induction via increased CREB-binding protein King, K.Y., Goodell, M.A., 2011. Inflammatory modulation of HSCs: viewing the HSC nuclear translocation in beraprost-induced antiproliferation of murine aortic – – as a foundation for the immune response. Nat. Rev. Immunol. 11, 685–692. smooth muscle cells. Am. J. Physiol. Cell Physiol. 297, C321 C329. Kisseleva, T., Brenner, D.A., 2007. Role of hepatic stellate cells in fibrogenesis and Sun, X., Zhang, X.D., Cheng, G., Hu, Y.H., Wang, H.Y., 2009. Inhibition of hepatic the reversal of fibrosis. J. Gastroenterol. Hepatol. 22, S73–S78. stellate cell proliferation by heat shock protein 90 inhibitors in vitro. Mol. Cell. – Kossatz, U., Dietrich, N., Zender, L., Buer, J., Manns, M.P., Malek, N.P., 2004. Skp2- Biochem. 330, 181 185. fi dependent degradation of p27kip1 is essential for cell cycle progression. Gene Tsukada, S., Parsons, C.J., Rippe, R.A., 2006. Mechanisms of liver brosis. Clin. Chim. – Dev. 18, 2602–2607. Acta 364, 33 60. A. Li et al. / European Journal of Pharmacology 747 (2015) 71–87 87

Tzivion, G., Dobson, M., Ramakrishnan, G., 2011. FoxO transcription factors; Xu, L., Hui, A., Albanis, E., Arthur, M., O’Byrne, S., Blaner, W., Mukherjee, P., Regulation by AKT and 14-3-3 proteins. BBA Mol. Cell Res. 1813, 1938– Friedman, S., Eng, F., 2005. Human hepatic stellate cell lines, LX-1 and LX-2: 1945. new tools for analysis of hepatic ﬁbrosis. Gut 54, 142–151. Van der Horst, A., Burgering, B.M., 2007. Stressing the role of FoxO proteins in Yu, C.H., Wu, J., Su, Y.F., Ho, P.Y., Liang, Y.C., Sheu, M.T., Lee, W.S., 2004. Anti- lifespan and disease. Nat. Rev. Mol. Cell Biol. 8, 440–450. proliferation effect of 3-amino-2-imino-3, 4-dihydro-2H-1, 3-benzothiazin-4-

Wu, J.G., Wu, J.Z., Sun, L.N., Han, T., Du, J., Ye, Q., Zhang, H., Zhang, Y.G., 2009. one (BJ-601) on human vascular endothelial cells: G0/G1 p21-associated cell Ameliorative effects of arctiin from Arctium lappa on experimental glomerulo- cycle arrest. Biochem. Pharmacol. 67, 1907–1916. nephritis in rats. Phytomedicine 16, 1033–1041. Zhao, F., Wang, L., Liu, K., 2009. In vitro anti-inﬂammatory effects of arctigenin, a Xiong, Y., Hannon, G.J., Zhang, H., Casso, D., Kobayashi, R., Beach, D., 1993. p21 is a lignan from Arctium lappa L., through inhibition on iNOS pathway. J. Ethno- universal inhibitor of cyclin kinases. Nature 366, 701–704. pharmacol. 122, 457–462.