The final results of the MTT assay indicated that GA and analogues such as pyrogallol (P) and 5-Hydroxydopamine hydrochloride (H) demonstrated doseependent cytotoxic effects (25, 50, and 75 M) on aHSCs right after 24 hrs incubation (Fig. 1A), with an EC50 benefit of thirty.5.7, forty one.8.six, and 35.M, respectively, whilst other analogues did not. GA also held considerable antiproliferative effects (P<0.05) on aHSCs as determined by analyzing the newly synthesized DNA of dividing cells through the BrdU assay. ATP-polyamine-biotinA doseependent reduction (55.3.3 and 66.9.4%) in cell proliferation was observed after 24 hrs incubation at GA concentrations of 50 and 75 M, respectively (Fig. 1B). Notably, in contrast to aHSCs GA showed less cytotoxicity on quiescent HSCs (qHSCs) and no cytotoxicty on normal hepatocytes at GA concentrations of 25, 50, and 75 M after 24 hrs of incubation (Fig. 1C). The cell viability of qHSCs was almost 10 times higher than that of aHSCs (73.4% vs. 7.9%) at GA 75M after 24 hrs of incubation (S2 Fig.). Oxidative stress induced by GA and GA analogues was further investigated by analyzing ROS formation. The levels of hydrogen peroxide (H2O2) in aHSC culture medium was determined after treatment with GA and GA analogues at 25, 50, and 75 M. Increased levels of H2O2 were found in the GA, P, and H treated groups (Fig. 2A). GA also elevated intracellular content of H2O2 (Fig. 2B). Accumulated intracellular ROS (e.g. hydroxyl and peroxyl radicals) GA induces cytotoxic and anti-proliferative effects on aHSCs. (A) Cell viability was monitored by using an MTT assay. GA analogues were treated with various concentrations (0, 25, 50, 75 M) for 24 hrs. Gallic acid (GA) 2,3,4-Trihydroxybenzoic acid (TA) Protocatechuic acid (PA) -Resorcyclic acid (RA) 3,4-Cresotic acid (CA) Methyl-3,4,5-trihydroxybenzoate (M) Propyl gallate (PG) 4-Nitrocatechol (N) 5-Hydroxydopamine hydrochloride (H). (B) Cell proliferation was examined by using a BrdU assay. GA-treated cells demonstrated significant cell proliferative inhibitory effects compared to the control groups. (C) The cytotoxic effects of GA on aHSCs and hepatocytes (HC). Various concentrations of GA (0, 25, 50, and 75 M) were added to hepatocytes and aHSCs. The cell viability was measured by an MTT assay. Data were expresed as meanD from three different experiments. The asterisk () indicates a significant difference from control group ( P<0.05, P<0.01)determined by DCFDA cellular ROS detection assay was also observed in the GA, P, and H treated groups (Fig. 2C). In addition, GA-induced lipid peroxidation, and oxidative DNA in aHSCs were revealed, as evidenced by doseependent formation of MDA (Fig. 2D), lipid hydroperoxides (Fig. 2E), and 8-oxodG, (Fig. 2F), respectively. Intracellular GSH concentration was also decreased with the increase of GA (Fig. 2G). These results suggest that GA induces remarkable oxidative stress in aHSCs. Besides, GA analogues like P and H generated significant amount of ROS intracellularly and in culture medium, and induced remarkable cytotoxicity, whereas other analogues induced no or lower levels of ROS and cytotoxicity (Fig. 1A). These outcomes might suggest that the involvement of oxidative stress in cell demise was chemical structure specific. Moreover, significant cytotoxicity was observed in aHSCs but not in normal hepatocytes after the treatment of GA, which could be due to the decreased antioxidative activity in aHSCs [19].GA induces the formation of H2O2, ROS, DNA oxidation, and lipid peroxidation in aHSCs. (A) Induction of hydrogen peroxide by GA and GA analogues (0, 25, 50, and 75 M). (B) GA dose-dependently (0, 25, 50, and 75 M) increased the production of H2O2 both in the culture medium and the cytosol. (C) GA-induced ROS, was determined by DCFDA cellular ROS detection assay. GA and analogues GA (50 M) were treated with aHSCs the generation of DCF was measured fluorescently. (D) GA-induced lipid peroxidation products, malondialdehyde (MDA), were determined by HPLC or microplate reader after 24 hrs of incubation. (E) Lipid peroxidation products, lipid hydroperoxides, were determined with or without inhibitors of TNF- and RIP1, SPD-304 (2M) and Nec-1 (2g/mL), respectively. (F) Oxidized DNA (8-OH-dG), and (G) total GSH contents were determined. Data were expressed as meanD from three different experiments. The asterisk () indicates a significant difference from control group ( P<0.05, P<0.01).The accumulation of GAnduced H2O2 in aHSCs could be resulted from impaired intracellular antioxidant system. To further investigate this assumption, the effect of antioxidant system on cell survivability was then determined (Fig. 3). Reagents such as deferoxamine (DFX) (a ferric iron chelator to limit Fenton-like reaction), superoxide dismutase (SOD), and catalase (CAT) were used to reduce oxidative stress. DFX chelates ferric iron to retard Fenton's reaction and the subsequent radical generation. SOD catalyzes the dismutation of superoxide to oxygen and hydrogen peroxide. Catalase catalyzes the decomposition of hydrogen peroxide to water and oxygen. Activated HSCs were initially incubated with GA, followed by the addition of GA down-regulates the activity of catalase in aHSCs. (A) The effect of antioxidants on the aHSC mortality rate. Activated HSCs were pre-treated with 75 M of GA for 24 hrs, followed by treatements with DFX (100M), SOD (100 U/mL), and CAT (100 U/mL) at different time intervals (0, 0.5, 1, and 2 hrs) after GA treatment. The cell viability was determined by an MTT assay. (B) The effect of catalase activity on viability in GA treated aHSCs. The aHSCs were transfected with catalase genes and incubated with GA for 24 hrs, followed by measurement of the cell survivability by MTT assay. (C) GA inhibits the activity of catalase in aHSCs but not in hepatocytes. Cells were treated with GA (0, 25, and 50 M) for 24 hrs before the measurement of catalase activity. P<0.05, P<0.01 antioxidants at different time intervals (0, 0.5, 1, and 2 hrs) after GA treatment. After 24 hrs of incubation, the cell viability was determined. Fig. 3A indicates that group treated with catalase showed the greatest cell survival promoting effect compared to other antioxidants. Group treated with DFX showed reduced cytotoxic effect in the first two time periods (0 and 0.5 hr) presumably due to the suppression of hydroxyl radical production catalyzed by iron. However, at the late time period (1 and 2 hrs), the cytotoxcity of DFX and GA co-treatment group was similar to that of GA alone, suggesting the critical role of H2O2 in cytotoxicity. There were significant cytotoxicity and no rescuing effect observed in the groups treated with SOD probably because of the accumulation of H2O2 resulted by the catalyzation of superoxides. On the other hand, cell survival was significantly promoted in groups treated with catalase, indicating the involvement of H2O2 in cytotoxicity. Improved survivability of aHSCs at several levels of GA treatment (25, 50, and 75 M) was maintained by transducing the catalase genes (Fig. 3B). A significant 35.1% and 25.7% recovery (P<0.05) at GA concentrations of 50 and 75 M, respectively, was achieved. Furthermore, the inhibitory potency of GA on the catalase activity was studied. As displayed in Fig. 3C, hepatocytes possess higher catalase activity than that of aHSCs under normal conditions. With the addition of GA (25 and 50 M), the catalase activity of aHSCs was suppressed doseependently, whereas the activity of hepatocytes was promoted at higher GA concentrations. These findings suggest that catalase is critical to the survival of aHSCs insulted by GAnduced oxidative stress. It has been reported that restricted catalase activity shows in HSCs once being activated [31]. This could likely make aHSCs more vulnerable to oxidative stress than normal hepatocytes.The GA-induced cytotoxic effect on aHSCs was observed in dose-dependent manners (Fig. 1A). We then attempted to further reveal the molecular mechanisms by which GA mediated the death of aHSCs. Our cell cycle analysis showed that GA did not provoke significant apoptotic effects on aHSCs (Fig. 4A, S3 Fig.). The sub G1 phase showed slight change after GA treatment (25, 50, and 75 M). However, LDH release (P<0.05) appeared with the increase in GA concentrations (25, 50, and 75 M) (Fig. 4B). This dose-dependent LDH release implies the disruption of the plasma membrane and subcellular organelles. Thus, GA might likely mediate a programmed necrotic effect, necroptosis, on aHSCs. It is known that TNF- pathway has been suggested to be associated with necroptosis, and RIP1 is one of key factors of necroptosis. The TNF- antagonist, SPD-304, and RIP1 inhibitor, Nec-1, were then used to examine GA-induced programmed necrotic cell death. The addition of SPD-304 and Nec-1 significantly rescued the survivability of aHSCs (Fig. 4C), reduced the production of lipid hydroxides (Fig. 2E), and increased intracellular GSH (Fig. 2G), indicating the involvement of necroptosis in GA-induced programmed cell death. It is suggested that the activation of RIP3 and TRADD are critical elements of TNF signaling-mediated necroptosis [32]. As shown in Fig. 4D, GA induced substantial TNF- release from aHSCs, which could likely elicit the downstream activation of necroptosis. Additionally, RIP3, the trigger of necroptosis in the TNF- pathway, along with the up-regulated expression of TRADD and the blocked caspase-8 activity, engages the effector mechanisms of necroptosis [26]. The results of immunoblotting analysis revealed that with the whole lysates of aHSCs, GA significantly up-regulated TRADD and p-RIP3 (1.4 and 1.3-fold, respectively) and down-regulated the activation of caspase-8 (Fig. 4E). The co-treatment of GA (75 M) and SPD-304 (2 M), as expected, down-regulated TRADD almost 2-fold (w/o inhibitor vs. w/ inhibitor, 1.41 vs. 0.73) and p-RIP3 1.4-fold (1.32 vs. 0.99) compared to GA alone (Fig. 4E), and GA induces TNF- -mediated necroptosis in aHSCs. (A) GA induced low levels of sub-G1 population in aHSCs as analyzed by flow cytometry. (B) Plasma memebrane integrity of aHSCs after GA treatment at designated concentrations was evaulated by LDH assay. (C) Involvement of TNF- and RIP1 in GA-induced necroptosis. Increased cell viability of aHSCs was obtained through the co-incubation of GA at various concentrations and SPD304 (2M) or Nec-1 (2g/mL). (D) GA elicited substantial production of TNF- as determined by immunoblotting and RT-PCR analysis. Immunoblotting analysis of necroptosis-related factors at various GA concentrations (0, 25, 50, and 75 M) with or without (E) SPD-304, (F) BSO and TNF-, co-incubation for 24hrs. Representative immunoblots showed the levels of TRADD, caspase-8, p-RIP3, and RIP3. -actin was used as an internal control. P<0.05, P<0.01 promoted caspase 8 activation (1.14 vs. 0.8). These results indicate that GA induced a selective necroptosis in aHSCs by triggering TNF- signaling pathway. Based on these findings, GA could likely induce necroptosis partly through the actions of activation of TNF- pathway, suppression of pro-caspase 8 activation, and depletion of intracellular GSH. Buthionine sulphoximine (BSO), an inhibitor of -glutamylcysteine synthetase (GCS) to deplete intracellular GSH, was used in conjunction with TNF- to investigate whether the factors associated with necroptosis could be provoked. As shown in Fig. 4F, BSO alone could significantly elicit the phosphorylation of RIP3 but could not upregulate other factors associated with necroptosis, e.g. TRADD. On the other hand, the combinatory effects of BSO and TNF- significantly promoted the activation of RIP3 and TRADD. These results might explain partly the necroptotic mechanisms that GA exerted on aHSCs indicated that the intracellular calcium level rose with the elevation of GA concentrations (25, 50, and 75 M) (P<0.05). The accumulation of calcium was suppressed by co- treatment with GA (75 M) and SPD-304 (GA alone vs. GA/SPD-304, 2.19 vs. 1.44 mg/dL, respectively). The addition of Nec-1 also suppressed GA-induced Ca2+ elevation (GA alone vs. GA/Nec-1, 2.19 vs. 1.25 mg/dL, respectively, at GA 75 M). These results indicate that GA- induced Ca2+ accumulation was through death receptor (DR)-elicited signaling. Molecules that associated with calcium-modulated necroptosis such as intracellular calcium concentration regulator, calmodulin (CaM), and calcium-activated neutral protease, calpain 1, were then examined. The active form of calpain executes lysosomal membrane permeabilization (LMP), which causes lysosome rupture and the spillage of acidic lysosomal contents to mediate cytoplasm acidification and degradation [33]. The results of the immunoblotting analysis indicate that GA remarkably up-regulated the expression of CaM and calpain 1, but the elevation was suppressed by the treatment of SPD-304 and Nec-1 (Fig. 5B), suggesting that GA triggers the process of necroptosis through the modulation of calcium signaling. Next, calpain-induced lysosomal membrane permeabilization (LMP) during GA-induced necroptosis was investigated by lysosomal staining with acridine orange. As shown in Fig. 5C, low level of orange fluorescence was observed in cells treated with GA alone, whereas increased orange fluorescence appeared upon the addition of SPD-304 and Nec-1, indicating the presence of intact acid organelle such as lysosome, after the treatment of inhibitors. These results indicate that either blocking TNF- signalling or RIP1 remarkably arrested the process of GAinduced LMP, which rescued the subsequent cell viability (Fig. 4B). Collectively, our data demonstrated that GA-induced TNF- -mediated necroptosis in aHSCs was elicited by triggering RIP1 and RIP3 necroptosome, followed by the modulation of Ca2+ signaling to execute LMP through calpain1 activation.It has been reported that necroptosis is reciprocal to apoptosis when the apoptotic signaling is blocked [34]. Therefore, we attempted to study whether blocking the GA-induced signals of necroptosis could divert cell death toward apoptosis. Various concentrations of GA were concurrently added with Nec-1 to aHSCs. Fig. 6 reveals that under apoptosis-competent GA-induced necroptosis is associated with Ca2+ signaling and lysosomal membrane permeabilization in aHSCs. (A) GA-induced calcium release was regulated by the TNF pathway. The cells were co-treated with GA (0, 25, 50, and 75 M), and SPD-304 (2 M) or Nec-1 (2 g/mL) for 24 hrs, followed by the analysis of cytosolic calcium contents. The results represent the meansD from three independent experiments. (B) Elevated cytosolic calcium levels triggered by GA upregulated the expression of CaM and calpain 1. 23646137The activated HSCs were treated with various concentrations of GA (0, 25,50, and 75 M) with or without SPD-304 (2 M) or Nec-1 (2 g/mL) for 24 hrs. Representative immunoblots showed the expression of CaM and calpain 1. actin was used as an internal control. P<0.05, P<0.01. (C) GA induces lysosomal membrane permeabilization in aHSCs. The effects of GA on lysosomal stability. The activated HSCs were treated with GA (25 and 50 M) and with or without SPD-304 (2 M) or Nec-1 (2 g/mL) for 24 hrs.