Verapamil inhibits early acute liver failure through suppressing the NLRP3 inflammasome pathway

Abstract Acute liver failure (ALF) is a rare disease characterized by the sudden onset of serious hepatic injury, as manifested by a profound liver dysfunction and hepatic encephalopathy in patients without prior liver disease. In this paper, we aim to investigate whether verapamil, an antagonist of TXNIP, inhibits early ALF through suppressing the NLRP3 inflammasome pathway. Firstly, an ALF mouse model was induced by lipopolysaccharide (LPS)/D‐galactosamine (GalN) treatment. The optimal concentration of verapamil in treating early ALF mice was determined followed by investigation on its mechanism in LPS/GalN‐induced liver injury. Western blot analysis and co‐immunoprecipitation were performed to determine the activation of the TXNIP/NLRP3 inflammasome pathway. Subsequently, overexpression of NLRP3 in mouse liver was induced by transfection with AAV‐NRLP3 in vivo and in vitro to identity whether verapamil inhibited early ALF through suppressing the activation of NLRP3 inflammasome. We found that ALF was induced by LPS/GalN in mice but was alleviated by verapamil through a mechanism that correlated with suppression of the NLRP3 inflammasome pathway. Oxidative stress and inflammatory response were induced by intraperitoneal injection of LPS/GalN, but alleviated with injection of verapamil. Overexpression of NLRP3 via AAV in mouse liver in vivo and in vitro reduced the therapeutic effect of verapamil on LPS/GalN‐induced ALF. Taken together, the TXNIP antagonist verapamil could inhibit activation of the NLRP3 inflammasome, inflammatory responses and oxidative stress to alleviate LPS/GalN‐induced ALF.

of new treatments or targets. In this regard, we note that there are two vital processes participating in the pathology of ALF. First, oxidative stress triggers a set of cysteine-aspartate proteases known as caspases and induces overwhelming apoptosis of hepatocytes, and second, varying degrees of resultant inflammation are a major factor in inducing cerebral oedema and multiorgan system failure. 6 Verapamil, a calcium channel blocker and P-glycoprotein (P-gp) blocker, has a definite therapeutic effect on ALF. 7 Excessive calcium influx is one of important signals of cell death, and the level of the drug transporter P-gp increased significantly in drug-injured liver, while remaining little changed in other organs. 7 Hence, verapamil can protect hepatocytes from dying in ALF. Meanwhile, verapamil is reported to have alleviated oxidative stress and exhibited hepatoprotective properties by reducing the level of the oxidation endproduct malondialdehyde (MDA). 8 However, it still has not been clarified how verapamil mediates these latter effects.
It is reported that verapamil can inhibit the expression of thioredoxin-interacting protein (TXNIP) and thus prevent islet cells from apoptosis. 9 TXNIP can bind to thioredoxin 1 (TRX1), resulting in increased oxidative stress and inflammation, and can also bind to nucleotide binding oligomerization domain-like receptors 3 (NLRP3), which promotes the release of IL-1β and subsequent inflammation.
During the treatment of chronic fatty hepatitis, verapamil plays a protective role through blocking the binding processes noted above. 10,11 Another study has revealed that verapamil can also treat acute hepatitis by up-regulating anti-inflammatory cytokine expression and downregulating pro-inflammatory cytokines through inhibition of NF-κB. 12 The NLRP3 inflammasome is a cytosolic signalling complex related to the pathogenesis of many diseases and mediating the activation of inflammatory factors such as interleukin-1 (IL-1). 13 The activity of the NLRP3 inflammasome can be perturbed by TXNIP silencing, which ultimately prevents inflammation and alters lipid metabolism. 14 Given the therapeutic effect of verapamil on ALF and significant role of TXNIP/NLRP3 in inflammation and oxidative stress, we were eager to establish whether the inhibitory effect of verapamil on ALF is related to inhibition of NLRP3 inflammatory pathway and oxidative stress.
In this study, a mouse model of liver failure was induced by intraperitoneal injection of lipopolysaccharide and D-galactosamine (LPS/ GalN), which is a procedure often used to study the mechanism of clinical liver diseases and search for potential treatments. 15 We next experimentally determined the optimal treatment dose of verapamil for ALF and investigated the mechanism by which verapamil functions to affect inflammatory and oxidative stress responses following ALF.

| LPS/GalN-induced ALF mice model
All the experimental animals fasted for 12 h before modelling. LPS (E. coli 055: b5) and GalN, purchased from Sigma-Aldrich (Sigma-Aldrich), were dissolved in phosphate-buffered saline (PBS) to appropriate concentrations for an intraperitoneal injection volume of 0.1 mL. LPS (30 μg/kg) and GalN (600 mg/kg) 16 were intraperitoneally injected into mice to induce (ALF). After 30 minutes, verapamil 10 mg/kg (0.1 mL) (Fengzhulin Chemistry Technology Ltd) was intraperitoneally injected into the mice. The mice were anaesthetized with intraperitoneal injection with pentobarbital sodium (50 mg/kg) and fixed on the heating pad at 3 hours after inducing ALF. Then, the mice were anaesthetized with inhalation of methoxyflurane and killed by exsanguination. The liver and blood samples were collected and stored at -80℃ for subsequent experiments.

| Experimental procedures and grouping
One hundred and eight C57BL/6 mice were purchased from

Laboratory Animal Center of Hubei Disease Prevention and Control
Center. 72 mice (12 mice for each group) were selected for survival analysis. The remaining 36 mice were randomly divided into 6 groups (6 mice per group), of which 5 groups received intraperitoneal (i.p) injection with LPS (30 μg/kg) and GalN (600 mg/kg) to induce ALF, followed by i.p. injection with different doses of verapamil (0, 5, 10 and 20 mg/ kg). The control group was injected with the same volume of saline.
Twenty-four C57BL/6 mice were randomly divided into 4 groups (6 mice per group), and each group received i.p. injection with LPS (30 μg/kg) and GalN (600 mg/kg) to induce ALF. Mice in the LPS/ GalN + verapamil group were intraperitoneally injected with verapamil (10 mg/kg) 30 minutes after inducing ALF. Mice in the verapamil along group were intraperitoneally injected with verapamil (10 mg/kg) without induction of ALF, and mice in the control group had no treatment.
The mice were killed 3 hours after inducing ALF, and liver and blood samples were collected for subsequent experiments.
Twenty-four C57BL/6 mice were randomly divided into 4 groups (6 mice per group) and were transfected with AAV-NLRP3 (adenoassociated virus construct encoding NLRP3) two weeks before ALF modelling. Mice transfected with AAV-8 empty vector served as a control group. ALF was induced after the expression of NLRP3 had been stably induced. Mice in the LPS/GalN group were injected i.p.
with LPS (30 μg/kg) and GalN (600 mg/kg) to induce ALF. Mice in the LPS/GalN + verapamil group were i.p. injected with verapamil (10 mg/kg) 30 minutes after inducing ALF, and mice in the verapamil group were i.p. injected with verapamil (10 mg/kg) without inducing ALF, whereas mice in the control group received no treatment. Mice were killed 3 hours after inducing ALF, and liver and blood samples were collected for subsequent experiments.
NCTC1469 hepatocytes were cultured in vitro and divided into control (no treatment), verapamil (hepatocytes treated with verapamil (10 mg/kg) without inducing LPS-induced liver injury), LPS (hepatocytes treated with LPS (1 μg/mL) within the culture medium to simulate the LPS-induced liver injury in vivo) and LPS + verapamil (hepatocytes treated with verapamil (10 mg/kg) 30 minutes after LPS-induced liver injury) groups, as in the animal experiment. Then, the cells were transfected with AAV-NLRP3 followed by the four groups of experiments as described above.

| In vitro model of LPS-induced liver injury
Murine liver cell NCTC1469, purchased from Bafeier Biotechnology Ltd, was cultured in Dulbecco's modified Eagle's medium (DMEM) with 10% (v/v) foetal bovine serum (FBS) in a humidified atmosphere with 5% CO 2 at 37°C. LPS (1 μg/mL) was added into the medium when cell confluency reached 75% to establish the in vitro model of LPS-induced liver injury. Six hours after induced injury, cells were collected for the subsequent experiments.

| AAV-NLRP3 construction and transfection
The AAV with encoding gene to overexpress NLRP3 was purchased from Genechem Ltd. The construction, recombination, amplification and purification of serotypes, specifically the transfection construction AAV-8 and AAV-NLRP3, were completed by Genechem Co., Ltd. AAV-NLRP3 was injected via the portal vein or into the liver parenchyma of mice 2 weeks before induction of ALF to obtain stable transfection with AAV-8. 17 Mice injected with AAV-8 empty vector served as control (AAV-NC group) to verify the efficiency of NLRP3 overexpression.
NCTC1469 cells were seeded at 10 5 cells per well in 6-well plate.

| Cell Counting Kit-8 (CCK-8) assay
The CCK-8 kit was purchased from Beyotime Biotechnology and used following manufacturer's instruction. Portions of cell suspension (100 μL) were seeded in a 96-well plate and incubated under 5% CO 2 at 37°C for 24 hours. Afterwards, cells were transfected with AAV followed by LPS-induced injury model establishment.
After modelling, 10 µL CCK8 solution was added into each well for incubation for 1 to 4 hours. The absorbance was read by spectrometry at 450 nm and cell viability was calculated.

| Annexin V apoptosis assay
Annexin V apoptosis assay was performed as follows. Falcon tubes were labelled according to the groups and washed with cold PBS twice. Then, cell suspension containing at 1 × 10 6 cells/mL was prepared with 1 × Binding Buffer. Each Falcon tube was added with 100 μL cells suspension, and then gently mixed with Annexin V and dye according to the kit protocol. The cell apoptosis rate was determined with flow cytometer. All reagents were purchased from Bafeier Biotechnology Co., Ltd.

| Terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) assay
The liver tissues were fixed with 4% paraformaldehyde, dehydrated

| Haematoxylin and eosin (HE) staining and histological score
The paraffin-embedded liver tissue was cut into 4 μm-thick sections for histological analysis by HE staining. The cross-section images were captured by a Leica Microsystems microscope (model: DM2000, CMS GmbH). Six views of each section were randomly selected to evaluate the liver injury. According to the standard of Suzuki, 18 three representative indexes of liver injury (congestion, vacuolation and necrosis) were scored from 0 to 4 according to their severity (0 for none; 1 for extremely mild; 2 for mild; 3 for moderate; and 4 for severe), thus to a maximum possible score of 12.

| Measurement of MDA, reactive oxygen species (ROS) and inflammatory cytokines
The liver tissue preserved in −80°C was homogenized with saline to prepare 10% w/v liver tissue homogenate (each 1 μg liver tissue was suspended in 9 μL saline). A bicinchoninc acid (BCA) kit was used to determine the protein concentration in the homogenate.
Afterwards, the sample was centrifuged for ten min at 14 000 g and the supernatant was collected. Then, the MDA and ROS contents in liver tissue were determined using corresponding kits purchased from Jiancheng Bioengineering Institute Ltd. TNFα, IL-6, IL-1β and IL-18 in liver tissue were detected using ELISA kits (Elabscience Biotechnology Co. Ltd) according to the manufacturer's protocol.

| Measurement of alanine aminotransferase (ALT) and aspartate transaminase (AST) in serum
Fresh blood (5 mL) was collected from mice and centrifuged at 1000 g to collect the serum. The ALT and AST levels released from liver tissue were measured using an AU5800 series automatic biochemical analysis system (Beckman Coulter Laboratory Systems Co Ltd) and ALT/AST-specific detection kits (Beckman Coulter Laboratory Systems). The result was used to determine the injury on hepatocytes.

| Immunohistochemistry (IHC)
The liver tissue was fixed with 4% paraformaldehyde and embed-

| Enzyme-linked immunosorbent assay (ELISA)
The ELISA kit (Jiangsu Enzyme Labeling Company) was taken from the refrigerator 1 hours in advance to warm to room temperature.
Then, the standard well and the sample well were set up, and 50 μL of the standard and 50 μL of the corresponding sample were added to the standard well and the sample well, respectively. A total of 100 μL of enzyme-labelled solution was added to all wells and incubated at 37°C for 60 minutes, and the plate was washed with washing solution for 4 times. Then, 300 μL of washing solution was added to each well followed by removal of the liquid after 30 seconds, and the plate was washed 5 times. Next, 100 μL of enzyme conjugate working solution was added to each well except the blank well. The reaction wells were sealed and incubated at 37°C for 30 minutes.
Afterwards, 50 μL of substrate A and B solutions was added to each well and incubated at 37°C for 15 minutes in the dark. Finally, 50 μL of stop solution was added to each well and the OD value was measured immediately after mixing.

| Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
The total RNA was extracted from the liver tissue preserved at −80℃ with Trizol reagent, followed by quantitation by Nano drop (Thermo Fisher Scientific Inc). Oligo (dT) primers and reverse transcriptase (Thermo Fisher Scientific Inc) were used to reverse transcribe the total RNA (1 μg) into cDNA. The expression of relative mRNA was analysed by the 2 -∆ ∆ Ct method, with GAPDH as the internal reference. Primers for GAPDH were (forward) 5'-GCTAACATCAAATGGGGTG-3' and (reverse) 5'-TTGCTGACAATCTTGAGGGAG-3'. PCR was performed using StepOnePlus (Applied Biosystems). The reaction proceeded for 1 cycle at 95°C for 15 minutes and 40 cycles at 95°C for 10 seconds and 60°C for 60 seconds.

| Western blot
The total protein was extracted from liver tissue samples using RIPA lysis buffer (the ratio of liver tissue to RIPA lysis buffer was 50 mg:

| Co-immunoprecipitation (Co-IP)
The liver tissue stored at −80°C was homogenized in RIPA lysis buffer (the ratio of liver tissue and RIPA lysis buffer was 50 mg: 1 mL), and the protein concentration of the extracted homogenate was determined by a BCA kit. According to the protein concentration, total protein lysate (1 mg) was used for Co-IP. The lysate was incubated with antibody (1 μg) against TXNIP or rabbit polyclonal IgG antibody as control. The mixture was shaken at 4°C for 4 hours to allow full binding to the corresponding antigens. Then, 20 μL re-

| Statistical analysis
Data are analysed using SPSS 21.0 software (IBM) and expressed as mean ± standard deviation. Unless otherwise noted, statistical comparisons were performed using an unpaired t test when only two groups were compared, or by Tukey's test-corrected one-way analysis of variance (ANOVA) with when more than two groups were compared. The Kaplan-Meier method was used to calculate the survival rate of mice, and log-rank test was used for univariate analysis. P <.05 indicates statistical significance.

| The optimal concentration of verapamil in treating early ALF
First, we established the optimal dose of verapamil in treating early ALF. LPS/GalN-induced ALF model is a mature technique used for studying the mechanisms of clinical liver disease and potential treatments. 15 The ALF model was induced in C57BL/6 mice by LPS/GalN treatment.

| Verapamil alleviates LPS/GalN-induced ALF
After confirming the optimal concentration of verapamil for ALF, we tried to further clarify the therapeutic effect of verapamil on ALF. It was evident that, compared with the control and verapamil groups, serum levels of ALT and AST were significantly increased in the LPS/GalN group (Figure 2A,B) and histological score was significantly increased ( Figure 2C), but was reduced by verapamil treatment. As reflected by TUNEL and HE staining, apoptosis, congestion, vacuolation and necrosis were significantly increased in liver of the LPS/GalN group, which could be reduced after addition of verapamil ( Figure 2D-F). In addition, the expression of C-caspase-3/caspase-3 and BAX/BCL-2 was significantly increased in the LPS/GalN group, which could be decreased after addition of verapamil ( Figure 2G,H). In conclusion, verapamil inhibits LPS/GalN-induced ALF.

| Verapamil alleviates LPS/GalN-induced inflammatory response and oxidative stress
Here, we explored further the alleviation by verapamil on the inflammatory response and oxidative stress in LPS/GalN-induced ALF, with the determination of plasma concentration of TNFα and IL-6 (the representative inflammatory cytokines), IL-1β and IL-18 (the cy-  Verapamil inhibited the expression of NLRP3 but had no effect on the expression of TXNIP and ASC after intraperitoneal injection of LPS/ GalN ( Figure 4A-C). Previous studies have shown that TXNIP inactively binds to Trx under physiological conditions, while TXNIP dissociates from Trx and binds to NLRP3 under oxidative stress, thereby activating the down-stream inflammatory pathway of NLRP3. 19 Therefore, the binding between TXNIP and NLRP3, TXNIP and Trx in mice liver of each group was further investigated using Co-IP. The results demonstrated that the formation of TXNIP/NLRP3 complex was increased and TXNIP/Trx complex was decreased after injection of LPS/GalN, but that verapamil treatment significantly inhibited the formation of the TXNIP/NLRP3 complex after intraperitoneal injection of LPS/ GalN, while having no effect on the TXNIP/Trx complex ( Figure 4D-F).

| Overexpression of NLRP3 eliminates therapeutical effect of verapamil on ALF
As indicated by the above-mentioned findings, NLRP3 may play a critical role in verapamil-induced alleviation on ALF. In the following study, we explored further whether verapamil regulated the TXNIP/ NLRP3 inflammasome pathway to alleviate ALF through overexpression of NLRP3 obtained by AAV-NLRP3 treatment in mice.
First, the efficiency of NLRP3 overexpression using AAV via portal vein injection was verified by Western blot, which showed that the expression of NLRP3 protein was significantly promoted after portal vein injection of AAV-NLRP3 in mice, compared with the AAV-NC group ( Figure 5A). We next investigated whether the overexpression of NLRP3 modulated the effect of verapamil on ALF.
We found that, compared with the control and verapamil groups, ALT and AST levels, as well as histological scores, were increased significantly in the LPS/GalN group ( Figure 5B-D). TUNEL staining showed that apoptosis in the LPS/GalN group was conspicuously increased ( Figure 5E-G). However, verapamil did not reduce liver injury after overexpression of NLRP3, as evidenced by the lack of significant improvement in histology scores ( Figure 5F) and no significant decrease in apoptosis ( Figure 5G). and IL-18, which could not be inhibited by verapamil ( Figure 6E-6I).

Moreover, in vitro experiments
The results suggested that NLRP3 played a critical role in the treatment of ALF with verapamil.

| D ISCUSS I ON
Verapamil is a commonly used calcium channel blocker and Pglycoprotein inhibitor, which can play therapeutic roles in cardiovascular diseases, 21-23 diabetes 11,24,25 and certain liver diseases including liver fibrosis, 26 liver ischaemia-reperfusion injury 27 and   10 Based on its protective effects, verapamil may treat liver disease by inhibiting oxidative stress and inflammatory response, which is the early pathological manifestation of ALF. In addition, Yumoto et al proved that ALF can be treated or arrested with verapamil. 7 We first identified the optimal dose of verapamil (10 mg/kg) for the treatment of early ALF, which was not hither addressed. Strikingly, we then identified the mechanism that verapamil can reduce TXNIP/NLRP3 inflammatory pathway, so as to reduce oxidative stress and inflammatory responses to alleviate early ALF (Figure 7). that NLRP3 inflammasome activation could be impaired by TXNIP deficiency. 33 TXNIP, an endogenous inhibitor of TRX, is reported to induce acute ischaemic stroke through inflammasome activation and redox imbalance. 34 Besides, exosomal miR-17 could target TXNIP and inhibit the activation of inflammasome in liver macrophages to treat ALF. 35 Additionally, verapamil can treat non-alcoholic fatty liver disease by inhibiting the TXNIP/NLRP3 pathway and reducing the level of IL-1β and IL-18 to attenuate hepatic metaflammation. 10

F I G U R E 6
Overexpression of NLRP3 eliminates the verapamil-induced inhibitory effect on the TXNP/NLRP3 inflammasome pathway. Mice were induced by LPS/GalN or treated with verapamil. A, The mRNA levels of IL-1β in mouse liver tissue in each group determined with RT-qPCR; B, the mRNA levels of TNFα in mouse liver tissue in each group determined with RT-qPCR; C, the mRNA levels of IL-6 in mouse liver tissue in each group determined with RT-qPCR; D, the mRNA levels of IL-18 in mouse liver tissue in each group determined with RT-qPCR; E, the expression of pro-caspase-1/caspase-1, pro-IL-1β/IL-1β and pro-IL-18/IL-18 proteins in mouse liver tissue in each groups determined with Western blot; F, the relative expression of caspase-1 normalized to GAPDH. G, the relative expression of IL-1β normalized to GAPDH; H, the relative expression of IL-18 normalized to GAPDH; I, the activity of caspase-1, IL-1β and IL-18 measured by ELISA. *P <.05. All values are expressed as mean ± standard deviation. Statistical comparisons are performed by Tukey's test-corrected one-way analysis of variance (ANOVA) when more than two groups were compared, n = 6 ing the TXNIP/NLRP3 pathway. 37 Another study has revealed that overexpression of TXNIP and NLRP3 is associated with the downregulated antioxidant genes such as catalase and MnSOD, 38 which was consistent with an anti-oxidative effect of verapamil reported in this study.
In conclusion, we identified the optimum verapamil dose (10 mg/kg) for treating the ALF mouse model and showed that verapamil can alleviate early ALF by inhibiting the TXNIP/NLRP3 pathway, which was otherwise associated with inflammation and oxidative stress. The relationship among the NLRP3 pathway, inflammatory responses and oxidative stress is complex. Although we have not fully articulated this complex mechanism, our findings have laid a foundation for the more appropriate application of verapamil for liver disease in clinical practice.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets generated during the current study are available.