Caerulein (Cae; cat # C9026), L-arginine (L-Arg; cat # A5131), lipopolysaccharide (LPS; cat # L2880), cholecystokinin 8 (CCK 8, cat # C2175), 4-phenylbutyrate (4-PBA; cat # SML0309), and benzamidine hydrochloride (Ben; cat # 434760) were purchased from Sigma-Aldrich Chemical (St. Louis, MO, USA). CA074 Methyl ester (CA074Me, cat # S7420) was purchased from Selleck Chemicals (Houston, TX, USA). SR11302 (cat # 160162-42-5) was purchased from APExBIO (Houston, TX, USA). Antibodies against X-box–binding protein 1 (sXBP1, cat # sc-7160) and receptor interacting protein kinase 3 (RIP3, cat # sc-135171) were from Santa Cruz Biotechnology (Dallas, TX, USA). Antibodies against C/EBP homology protein (CHOP, cat # 2895), cathepsin B (CTSB, cat # 31718), phosphorylated protein kinase α (p-PKCα, cat # 9375), phosphorylated c-Jun NH2-terminal kinase (p-JNK, cat # 4668), phosphorylated extracellular signal-regulated kinas (p-ERK, cat # 4370), phosphorylated p38 mitogen activated protein kinases (p-p38MAPK, cat # 4511), phosphorylated cJun (p-cJun, cat # 3270) and IL1β(3A6) (cat # 12242) were from Cell Signaling Technology (Danvers, MA, USA). Antibodies against mixed lineage kinase domain-like (MLKL, cat # ab172868), phosphorylated mixed lineagekinase domain-like (pMLKL, cat # ab196436) and Ly6G (cat # ab25377) were from Abcam (Cambridge, MA, USA). Antibodies against glucose-related peptide 78 (GPR78, cat # 11587-1-AP) and TNFα (cat # 60291-1-Ig) were from Proteintech Biotechnology (Wuhan, China). Antibody against IL6 (cat # BS6419) was purchased from Bioworld Technology(St. Louis Park, MN, USA),and antibody against β-actin (cat # AF0003) was from Beyotime Biotechnology (Shanghai, China). Nuclear and cytoplasmic protein extraction kit was from Pierce (Rockford, IL, USA). The biotin-labeled probe containing the activating protein 1 (AP-1) binding site was purchased from Beyotime Biotechnology (Shanghai, China). The light shift chemilumines-cent EMSA kit was from Pierce (Rockford, IL, USA).

Balb/C mice (6 ~8 weeks, 20~22g, male) were purchased from Shanghai SLAC Laboratory Animal Co Ltd (Shanghai, China). All mice were kept in pathogen-free conditions in individually-ventilated cages (4 ~6 mice per cage) at 23 ± 2°C and a 12 h dark/light cycle with free access to water and standard rodent diet before experiment. All mice were allocated into groups in a completely randomized manner (n = 6 per group) to conduct the experiment. All experiments were approved by the Animal Ethics Committee of Shanghai Jiao Tong University School of Medicine (SYXK 2013-0050, Shanghai, China.).

Two AP models were built in vivo, both are widely used, rapidly induced and noninvasive (27). One is induced by injections of caerulein (100 μg/kg, i.p. with 1 h interval between injections, ten injections) and LPS (5 mg/kg, i.p. administered immediately after the last injection of caerulein) as reported before (28, 29). Controls received normal saline (NS) equivalent to caerulein. The first caerulein injection was defined as 0 h. The ER stress inhibitor (4-PBA, 4mg per mouse) or the CTSB inhibitor (CA074Me, 10 mg/kg) was injected intraperitoneally (i.p.) 0.5 h before the first caerulein injection and mice were sacrificed at 12 h. The other model is induced by L-Arg (4 g/kg, 8%, pH=7.0, with 1 h interval between injections, two injections) as previously described (30–33). Control mice received equal NS instead of L-Arg. The second L-Arg injection is defined as day 0. 4-PBA (4mg per mouse) was injected 0.5 h i.p. before the first L-Arg injection, and equivalent 4-PBA was added everyday in the following two days. Mice were sacrificed at day 3. Serum and pancreas were collected. Histological scoring of haematoxylin and eosin (H&E) sections were performed by two experienced pathologists (32–34).

Blood samples of each group were collected and centrifuged at 250 g for 20 min at 4°C. Serum amylase and lipase activities were measured by enzyme dynamics chemistry, according to the manufacturer’s instructions in a Roche/Hitachi Modular Analytics System (Roche, Basel, Switzerland).

Mice pancreas specimens were fixed in 4% neutral paraformaldehyde for 24 ~48 h, embedded in paraffin, and cut into 4 μm sections for H&E staining by standard procedures. Endogenous peroxidase was neutralized by 3% hydrogen peroxide. Then sections were incubated overnight at 4°C with monoclonal antibody against Ly6G (1:100). After being rinsed in PBS for three times, sections were incubated with secondary antibody for 1 h at 37°C and then imaged by an ultrasensitive SP kit and a DAB kit (Fuzhou Maxin, China). Pancreatic tissue section were scored on a range of 0 to 3 (0 represented normal appearance and 3 meant severe), based on the presence of necrosis, edema and inflammation (35). The pathologists were blinded to the experiment groups.

Pancreatic acinar cells were isolated from Balb/C mice as described previously, using collagenase digestion with minor modifications (32, 33, 36). Primary pancreatic acinar cells were incubated at 37°C in Dulbecco’s modified Eagle’s medium/Ham (DMEM) F-12 medium containing 10% fetal bovine serum (FBS). Acinar cells were pre-treated with ER stress inhibitor 4-PBA (2.5 μM, 5 μM, 10 μM) or CTSB inhibitor (CA074Me; 50 μM) or trypsin inhibitor Ben (1 μM) or AP-1 inhibitor SR11302 (10 μM) for 30 min, and then stimulated by 200 nM CCK 8 for 30 min or 6 h. Cells were collected at the time points as indicated in the Figure legends.

Pancreatic tissue and pancreatic acinar cell extracts were used for western blotting analysis. Total amounts of protein were detected using the bicinchoninic acid assay method (Beyotime Biotechnology, China). Proteins (40 μg per lane) were separated by 10% SDS-PAGE at 120V and transferred to nitrocellulose membranes (Millipore, Mass, USA) for 30 ~ 60 min. Membranes were then incubated with primary antibodies against polyclonal GRP78 (1:1000), sXBP1 (1:400), RIP3 (1:400), MLKL (1:1000), IL-1β (1:1000), IL-6 (1:1000) and p-PKCα (1:400); monoclonal CHOP (1:1000), pMLKL (1:1000), CTSB (1:1000), p-JNK (1:1000), p-ERK (1:1000), p-p38MAPK (1:1000), p-cJun (1:1000), TNFα (1:1000) and β-actin (1:1000) overnight at 4°C. After being washed with PBS containing 0.1% Tween, membranes were probed by secondary antisera labelled with goat anti-mouse or goat anti-rabbit IR-Dye 700 or 800 cw for 1 h at 37°C. Membranes were scanned by an Odyssey Infra-red Scanner (LI-COR, Lincoln, NE, USA). Representative blot images were presented from three separate experiments. Relative expression of target proteins was expressed as fold changes compared to normal control after normalized to β-actin.

Cell survival assay was performed according to the manufacturer’s instructions of Cell Titer-GloLuminescent Cell Viability Assay kit (Promega, Madison,WI). Briefly, 50 μL ATP detection reagents were added into 100 μL cell suspension in a 96-well culture plate, and the levels of bioluminescence were recorded using a SpectraMax 190 system (Molecular Devices, San Jose, CA, USA).

Quantitative analysis of cytotoxicity can be achieved by detecting the activity of LDH released into the culture supernatant from injured cells using LDH Cytotoxicity Assay Kit (Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions. Briefly, pancreatic acinar cells in each group were collected and centrifuged at 100 g for 5 min at 4°C. A triplicate set of wells for the culture medium background with no cells were served as minimum (blank LDH), and untreated normal control cells lysed to yield LDH completely were used for maximum (total LDH), respectively. Then 120 μL of supernatant and 60 μL of LDH detection reagent were mixed in a 96-well plate for 30 min. A SpectraMax 190 system (Molecular Devices, San Jose, CA, USA) was used to detect the absorbance at 490 nm. The LDH release rate was calculated according to the following formula: LDH release rate (%) = [(sample LDH – blank LDH)/(total LDH – blank LDH)] × 100%.

Trypsinogen activation peptide (TAP) is a small peptide released from trypsinogen during its activation (37). Therefore, TAP concentration can indirectly reflect the levels of trypsinogen activation. TAP levels of pancreatic acinar cells were measured by enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s protocols (Westang Bio-Tech, Shanghai, China).

The binding activity of AP-1 was detected by EMSA. A biotin-labelled probe containing AP-1 binding site was incubated with 10 μg nuclear extracts for 30 min at room temperature. Protein-DNA complexes were separated by 5% polyacrylamide gel in 0.5x Tris Buffer EDTA at first and then transferred to a positively charged nylon membrane. The biotin-labelled DNA was examined by a chemiluminescent detection kit.

All the data are presented as Mean ± SEM, from at least three replicates with six mice per group. The distribution of data was assessed by Kolmogorov–Smirnov test at first. If data followed a Gaussian distribution, parametric tests were carried out: Student’s t test for two groups and one-way ANOVA for three or more groups. If data were not normally distributed, non-parametric tests were performed: Mann–Whitney test for two groups and Kruskal–Wallis test for three or more groups. All analyses were performed using GraphPadPrism (version 7.00 for Windows, GraphPad Software, La Jolla California USA, www.graphpad.com). p value <0.05 was considered as statistically significant.

It is reported that several pathological events such as ER stress and necroptosis are involved in AP (38). To investigate the relationship of ER stress and necroptosis in AP, pancreatic acinar cells were stimulated by 200 nM CCK in vitro, ER stress indicators such as GRP78, sXBP1 and CHOP, and necroptosis markers such as RIP3 and pMLKL were detected by western blotting. Data showed that GRP78, sXBP1 and CHOP were elevated significantly at 0.5 ~1 h ( Figure 1A ); RIP3 and pMLKL were markedly increased at 6 h ( Figure 1B ). That is to say, ER stress was activated in the early stage of CCK stimulation in acinar cells, while necroptosis activated in the late stage. Besides, trypsinogen was activated at 0.5 h ( Figure 1C ). Then we induced ER stress in acinar cells by 2 μM thapsigargin (Tg) for 6 h. As shown in Supplementary Figure 1 , RIP3 level was increased at 12 ~24 h and MLKL was phosphorylated at 18 ~24 h in a dose-dependent manner. However, ATP depletion was only slightly elevated (about 5 ~10%), and trypsinogen cannot be activated by 2 μM Tg. Therefore, ER stress may regulate necroptosis, independent of trypsinogen activation.

In order to explore the relationship between ER stress and necroptosis, 4-PBA (2.5 μM, 5 μM, 10 μM), a specific inhibitor of ER stress, was used 30 min in advance and then stimulated by 200 nM CCK in vitro. Data indicated that GRP78, sXBP1 and CHOP levels were dose-dependently inhibited by 4-PBA in CCK-stimulated acinar cells ( Figure 2A ). Furthermore, we found that 4-PBA significantly blocked necroptosis in a dose-dependent manner in CCK-stimulated acinar cells in vitro, manifested in the decreased RIP3 and pMLKL levels. 4-PBA also significantly reduced CCK-induced ATP depletion, and LDH release rate in CCK-stimulated pancreatic acinar cells ( Figure 2C, D ). In addition, TAP levels were elevated after CCK stimulation, which were remarkably reduced after 4-PBA treatment ( Figure 2E ). These data showed that ER stress inhibition by 4-PBA significantly alleviated pancreatic acinar cell necroptosis during AP in vitro.

In vivo, caerulein and LPS-induced AP model in Balb/C mice were built with or without 4-PBA treatment at 0.5 h in advance. Histological examination showed that the extent of pancreatic injury in 4-PBA-treated mice was less severe than that in AP group at 12 h after the first caerulein injection ( Figure 3A, B ). Amylase and lipase activities were markedly decreased in the sera of 4-PBA-treated mice compared to AP groups ( Figure 3C ). Furthermore, we examined the infiltration of Ly6G⁺ neutrophils in the pancreatic tissue by immunohistochemistry staining. Data showed that 4-PBA reduced Ly6G⁺ neutrophil infiltration in pancreas when compared to AP groups ( Figure 3D ). 4-PBA treatment blocked the upregulation of ER stress indicators such as GRP78, sXBP1 and CHOP, and necroptosis markers such as RIP3 and pMLKL during AP ( Figure 3E, F ). In addition, the serum TNFα, IL1β and IL6 levels were also markedly decreased in 4-PBA pre-treated mice compared to AP group ( Figure 3G ). L-Arg-induced AP model were also built with or without 4-PBA treatment. The similar results were obtained at day 3 after the second L-Arg injection, as shown in Figure 4 . We found that 4-PBA significantly alleviated pancreatic injury assessing by histological examination and pathological score analysis; reduced serum amylase and lipase levels; and markedly decreased GRP78, sXBP1, CHOP, RIP3 and pMLKL levels in pancreatic tissue during L-Arg-induced AP ( Figure 4A-E ). Collectively, these results suggested that inhibition of ER stress by 4-PBA alleviated pancreatic necroptosis during AP both in vitro and in vivo.

The detailed mechanisms of ER stress mediating necroptosis remains unclear. CTSB is well-known as a protease to activate intrapancreatic trypsinogen and initiate the onset of AP (21, 23). Recent studies suggest that excessive CTSB released from lysosomes into the cytosol can shift the cell death pathway towards necroptosis (23). TNFα is a ligand binding to death receptor to motivate necroptosis, and AP-1 is a transcription factor encoding the expression of inflammatory factors, such as TNFα (25, 26). To determine the detailed mechanisms of ER stress mediating necroptosis, we investigated the role of CTSB maturation and AP-1 activation during ER stress both in pancreatic acinar cells in vitro and in pancreatic tissue of experimental AP model in vivo. Firstly, western blotting analysis showed that 4-PBA significantly reduced the expression of mature CTSB in a dose-dependent manner in CCK-stimulated acinar cells in vitro ( Figure 5A ). Secondly, phosphorylation levels of PKCα, JNK and cJun were markedly decreased by 4-PBA, while ERK and p38MAPK phosphorylation levels showed no significant difference after 4-PBA treatment in CCK-stimulated acinar cells in vitro ( Figure 5B ). Next, further experiments showed that 4-PBA then reduced AP-1 binding activity in CCK-stimulated acinar cells ( Figure 5C ), as previous studies have shown that PKC-MAPKs signaling pathway can activate AP-1 (24). Lastly, we confirmed these results both in caerulein and LPS-induced and L-Arg-induced AP models in vivo. Consistent with the results in vitro, we also found that 4-PBA significantly inhibited the phosphorylation levels of PKCα, JNK and cJun, but not ERK and p38MAPK phosphorylation levels both in two experimental AP models in vivo ( Figure 5D, E ). To sum up, ER stress blockade by 4-PBA not only inhibited CTSB maturation but also suppressed AP-1 activation via PKCα-JNK-cJun pathway during AP both in vitro and in vivo.

Next, in order to clarify the relationship between CTSB and AP-1, CTSB inhibitor CA074Me was administered to block CTSB at 30 min before caerulein and LPS or CCK challenge, and then AP-1 activation was determined. In vitro, blockade of CTSB by CA074Me significantly inhibited the phosphorylation levels of PKCα, JNK and cJun, AP-1 activation and TNFα levels in CCK-stimulated pancreatic acinar cells ( Figure 6A-C ). It was reported that CTSB can activate trypsinogen (21, 23). In order to exclude the influence of trypsinogen activation, we used benzamidine hydrochloride to inhibit trypsin activity and observed changes of PKCα-JNK-cJun pathway, TNFα level and necroptosis. Notably, our data showed that benzamidine hydrochloride had no effect on phosphorylation levels of PKCα, JNK and cJun, TNFα levels, or necroptosis markers including RIP3 and pMLKL in CCK-stimulated acinar cells ( Figure 6D, E ), although total ATP depletion and LDH release rate significantly reduced after benzamidine hydrochloride treatment ( Figure 6F, G ). In summary, CTSB participated in AP through two pathways: one is to activate AP-1 via PKCα-JNK-cJun signaling and induce necroptosis, which is independent of trypsin activity; and the other way is to activate trypsinogen and induce necrosis. Similarly, we observed the effect of CA074Me on caerulein and LPS-induced AP model in vivo. As expected, CA074Me led to a significant decrease of PKCα, JNK, cJun phosphorylation levels and TNFα levels in pancreatic tissue ( Figure 6H, I ). That is to say, CTSB enhanced AP-1 activation and TNFα levels to induce necroptosis via PKCα-JNK-cJun pathway in caerulein and LPS-induced AP model. In a word, CTSB induced AP-1 activation and necroptosis via PKCα-JNK-cJun pathway during AP, independent of trypsin activation.

AP-1, as an important transcription factor, is related to the transcription of TNFα (24, 26), which has been demonstrated to trigger necroptosis in many cell types (6, 7). In order to determine the role of AP-1 on necroptosis during AP, we used SR11302 30 min in advance to inhibit AP-1 in CCK-stimulated pancreatic acinar cells. We found that SR11302 significantly lowered the TNFα, RIP3 and pMLKL levels ( Figure 7A ). Furthermore, treatment with SR11302 also resulted in reduction of CCK-induced pancreatic acinar cells necrosis, manifested as ATP depletion and LDH release rate were reduced markedly by SR11302 in CCK-stimulated pancreatic acinar cells ( Figure 7B, C ). This phenomenon indicated that AP-1 caused acinar cell necroptosis by promoting TNFα autocrine secretion. Taken together, all these results demonstrated that ER stress promoted pancreatic acinar cell necroptosis through CTSB maturation, thus induced AP-1 activation and TNFα autocrine secretion via PKCα-JNK-cJun pathway during AP, not related with trypsin activity ( Figure 7D ).