Vagus Nerve Attenuates Hepatocyte Apoptosis upon Ischemia–Reperfusion via α7 Nicotinic Acetylcholine Receptor on Kupffer Cells in Mice

HEPATIC ischemia–reperfusion (HIR) injury occurs in many clinically important scenarios, including hepatic surgery, transplantation, trauma, and hemorrhagic shock.¹  HIR injury is characterized by necrosis and apoptosis of hepatocytes within several hours of reperfusion and is associated with the generation of reactive oxygen species (ROS), release of proinflammatory cytokines, and activation of the complement system.^2–5  Oxidative stress is the major initiator in activating signaling pathways that lead to necrosis and apoptosis during the HIR procedure.^6,7  Thus, the formation of ROS and oxidative stress are considered to be the disease mechanisms most commonly involved in HIR injury.^8,9  Hepatocyte apoptosis is one of the main causes of hepatic failure upon HIR injury²  or hepatic transplantation.^10,11  It has been reported that up to 50% of hepatocytes and sinusoidal endothelial cells undergo apoptosis within the first 24 h of reperfusion.^12,13  Inhibiting apoptosis with caspase inhibitor could protect the liver from ischemia–reperfusion injury.¹⁴ 

Kupffer cells (KCs) are the resident macrophages of the liver.¹⁵  A key event of the early phase of reperfusion is the activation of the KCs that are the primary source of extracellular ROS.⁸  It is well accepted that the soluble CD163 (sCD163) is a sensitive and specific marker of activated KCs.^16,17  Convincing evidence has shown that hepatocyte injury mediated by KCs is associated with the production of ROS in response to oxidative stress,^18,19  and inhibiting KC function could diminish the production of ROS and thus protect against HIR injury.^20,21 

Expression of α7 nicotinic acetylcholine receptor (α7nAChR) in macrophages was confirmed, and vagal stimulation can prevent inflammation by inhibiting the production of proinflammatory cytokines from macrophages through activation of α7nAChR by acetylcholine released from the vagus nerve.^22–24  Our previous studies found that the antishock effect of anisodamine, a muscarinic receptor antagonist, was linked to the α7nAChR-dependent anti-inflammatory pathway,²⁵  and dysfunction of vagus nerve contributed to end-organ damage via α7nAChR in hypertension.²⁶  Vagus nerve also plays a role in ameliorating carbon tetrachloride–induced hepatocellular necrosis.²⁷  Furthermore, activation of α7nAChR could inhibit apoptosis,^28,29  and it protects neurons from apoptosis induced by oxidative stress.^30,31 

In the current study, we tested the hypothesis that hepatic vagus nerve is protective during HIR-induced hepatocyte apoptosis by activating α7nAChR on KCs. Mainly, HIR-induced liver injury was examined in α7nAChR^−/− and hepatic vagotomy mice with versus without α7nAChR activation. Hypoxia/reoxygenation (HR)–induced apoptosis of hepatocytes with KC coculture was measured with versus without α7nAChR activation. HR-induced intracellular ROS and supernatant H₂O₂ in KCs were determined after activation of α7nAChR or treatment with catalase. Hepatocytes that suffered HR were incubated with the different supernatants from KCs, and the number of apoptotic cells was calculated. Finally, HIR-induced liver H₂O₂ production was determined in mice with versus without KC elimination, and KC activity was evaluated by sCD163 both in vivo and in vitro.

Male C57BL/6J mice (20 to 25 g) were purchased from Sino-British SIPPR/BK Laboratory Animal Ltd (China). The α7nAChR^−/− mice were purchased from Jackson Laboratory (USA) and were bred at the animal center of Tongji University, Shanghai, China. Male α7nAChR^−/− mice (20 to 25 g) were used. Animals were housed under a 12/12-h light/dark cycle (23° to 25°C) with free access to food and water and were randomly divided into experiment groups when used. The experimental procedures for the care and use of the animals were approved by the Ethics Committees of the Animal Center, Tongji University, Shanghai, China.

Briefly, mice were anesthetized with sodium pentobarbital (45 mg/kg, intraperitoneal). A midline abdominal incision was made. The hepatic branch of the vagus nerve was identified, isolated, and transected under a surgical microscope. HIR was carried out 1 h later, and an additional dose of sodium pentobarbital (15 mg/kg, intraperitoneal) was injected to keep the mice under anesthesia.

KC depletion was performed using a commercially available liposome-clodronate reagent Clophosome-A (FormuMax Scientific, USA). Briefly, mice were given Clophosome-A (180 μl per mouse, intravenous; three times at 3-day intervals) or an equal volume of vehicle before HIR. The elimination of the KCs was confirmed by immunohistochemistry using a rat anti-mouse CD68 antibody (1:10; Serotec, USA).

One-hour ischemia followed by 6-h reperfusion was carried out in C57BL/6J, α7nAChR^−/−, hepatic vagotomized, and KC-eliminated mice with activation of α7nAChR by PNU-282987^26,30  (a selective α7nAChR agonist; 40 μg/kg, intraperitoneal; Sigma, USA) or in mice that received an equal volume of vehicle (0.4% DMSO in saline, intraperitoneal) as described previously.^9,32,33  Briefly, mice were anesthetized with sodium pentobarbital (45 mg/kg, intraperitoneal). A midline abdominal incision was made. All structures in the portal triad (hepatic artery, portal vein, and bile duct) to the left and median liver lobes were occluded with an atraumatic clamp. Thirty minutes before the occlusion, PNU-282987 was injected into the abdominal cavity (for mouse with hepatic vagotomy, it was injected 30 min after vagotomy but still 30 min before occlusion). The clamp was removed after 1 h to initiate a 6-h hepatic reperfusion period. Sham operation consisted of all procedures with the exception of vascular occlusion. Body temperature was maintained at 37°C using a heating pad (Physitemp Instruments, USA) during the experiment. At the end of 6-h reperfusion, the mice were anesthetized with inhaled methoxyflurane and were killed by exsanguination. Liver and blood samples were collected.

Blood samples were centrifuged at 3,000g for 10 min at 4°C. Serum alanine transaminase (ALT) was measured with a Synchron LX20 system (Beckman Coulter, USA), and serum H₂O₂ level was measured using a H₂O₂ assay kit (Beyotime Biotech, China) and a Tecan infinite M200 reader (Tecan, Switzerland) as described previously.³⁴ 

Liver specimen was made as described previously.^21,35  Briefly, liver samples were taken from a relatively fixed position, that is, from the left middle lobe about 5-mm distance from the portal vein, and then samples were fixed in 4% paraformaldehyde in phosphate buffer solution for 24 h, washed, dehydrated in a graded ethanol series, and embedded in paraffin. Sections (5 μm) were stained with hematoxylin and eosin for light microscopic investigation. For each liver sample, two sections were taken.

Apoptosis was measured using terminal deoxynucleotidyl transferase–mediated 2'-deoxyuridine 5'-triphosphate nick-end labeling (TUNEL) staining. Tissue sections (5 μm) were deparaffinized in xylene and hydrated in graded ethanol. Apoptotic cells were identified using a kit from Boehringer Mannheim (Germany) and quantified using an image analysis system from LEICA Imaging Systems (LEICA QUIPS; LEICA Imaging Systems, United Kingdom). Results are expressed as the percentage of the positively stained cells. For each specimen, two high-power fields were randomly selected and cells were counted.

The activity of caspase-3, -8, and -9 in liver lysate was measured using kits according to the user’s manual (MBL, Japan). Briefly, 100 mg liver sample was homogenized in 200 µl lysis buffer and centrifuged at 12,000g for 15 min at 4°C. Protein concentration was determined using a bicinchoninic acid assay method (Thermo Fisher Scientific, USA). The enzyme activity in the supernatant was measured according to the manufacturer’s protocol using a Tecan infinite M200 reader (Tecan).

Hepatocytes and KCs were isolated from male mice as described previously.³⁶  Briefly, the liver was digested by 0.05% collagenase IV (Worthington, USA) combined with 15 U/ml DNase I (Roche, Germany) perfusion through the portal vein. The dispersed cells were filtrated and then centrifuged at 50g for 7 min. Parenchymal cells were collected from the pellet; nonparenchymal cells were concentrated in the supernatant. The pellet was washed twice by centrifuging at 50g for 5 min and cultured with the hepatocyte maintenance medium (Lonza, USA) containing 10% fetal bovine serum (FBS) before use. Nonparenchymal cells were centrifuged by density gradient centrifugation in a 70:30% Percoll (GE Healthcare Bio-Sciences, Sweden) gradient at 2,500g for 25 min at 4°C. KCs, concentrated at the interfaces between 70% Percoll and 30% Percoll, were carefully collected, washed twice, and cultured with the medium before use. Nonadherent cells were removed by replacing the medium after 4-h culture. Primary cells were shown to be alive (hepatocytes, greater than or equal to 80%; KCs, greater than or equal to 60%) by staining with 4% trypan blue. KCs were identified as F4/80-positive cells by flow cytometry, and the purity of them was greater than or equal to 90%. Morphologic examination indicated that greater than or equal to 90% cells were hepatocytes.

Hepatocytes were seeded onto a transwell plate in the hepatocyte maintenance medium with 10% FBS before use. KCs were seeded onto Transwell inserts (0.4-μm pore; 24-mm diameter, Corning Costar, USA) in Dulbecco Modified Eagle Medium (DMEM) with 10% FBS. After 24-h culture, the inserts together with KCs were transferred into the hepatocyte cultures and exposed to the medium with 0.1% dimethyl sulfoxide (DMSO) or PNU-282987 (30 µM) for 1 h, followed by 6-h hypoxia (94% N₂, 5% CO₂, and 1% O₂) and 2-h normoxia (74% N₂, 5% CO₂, and 21% O₂) in a trigas incubator (Thermo Fisher Scientific). At the end of reoxygenation, caspase-3, -8, and -9 activity in hepatocytes were measured, and apoptotic hepatocytes were determined using a dead cell apoptosis kit with annexin V/PI (Invitrogen, USA) and measured by flow cytometry (BD FACS Calibur Flow Cytometer; Becton Dickenson, USA).

KCs were cultured under hypoxia at 37°C for 6 h with PNU-282987 (30 µM) or catalase (200 U/ml) pretreatment (for 1 h) before 2-h reoxygenation (6-/2-h HR). Cells incubated under normoxia at 37°C for 8 h were included as a control. Intracellular ROS production was determined using the fluorescent probe 2′,7′-dichlorofluorescin diacetate (Beyotime Biotech) and measured by flow cytometry (BD FACS Calibur Flow Cytometer) as described previously.³⁷  Supernatant H₂O₂ production was measured using a H₂O₂ assay kit (Beyotime Biotech) and measured with a Tecan infinite M200 reader (Tecan) as described previously.³⁴ 

Hepatocytes that suffered 6-h hypoxia were cultured under normoxia together with the supernatants from KCs (conditioned treatment) for another 2 h, and then apoptosis was measured by flow cytometry.

sCD163 was measured with ELISA kit according to the user’s manual (MyBioSource, USA). Briefly, the samples were incubated with sCD163–HRP conjugate in precoated plate for 1 h. After incubation, the wells were decanted and washed five times and then incubated with a substrate for HRP enzyme. After that, the reaction was stopped and measured by using a Tecan infinite M200 reader at 450 nm.

To assess the role of hepatic vagus nerve in HIR-induced liver apoptosis, eight C57BL/6J mice were used per group based on previous work.³⁸  To determine the role of α7nAChR in HIR-induced liver apoptosis, eight wild-type (WT) mice and six α7nAChR^−/− mice were used.³⁹  To measure the effects of PNU-282987 in HIR-induced liver apoptosis in hepatic vagotomized mice, six mice were used for the vagotomized or vagotomized and PNU-282987 group, eight mice were used for the vagotomized and IR or vagotomized and IR and PNU-282987 group, and five liver samples from five different mice per group were used for TUNEL-positive cell counting; samples were coded by an independent colleague to ensure that the experimenter was blinded to the treatment group during cell counting.⁴⁰ In vitro, primary KCs or hepatocytes isolated from one mouse liver were considered one sample, three samples were used per group, and each experiment was repeated twice.⁴¹  No data were lost during observation in this work.

All data were presented as mean ± SD and analyzed using GraphPad Prism Version 5.0 (GraphPad Software Inc., USA) and one-way ANOVA followed by Tukey test for multiple comparisons, and differences were considered significant at P value less than 0.05. Please also see Supplemental Digital Content, http://links.lww.com/ALN/B310.

Hepatic vagotomy alone did not change serum alanine transaminase (ALT) and activity of caspase-3, -8, and -9 in C57BL/6J mice. HIR resulted in significantly elevated serum ALT (fig. 1A). The activity of caspase-3, -8, and -9 in the liver was also increased by HIR (fig. 1, B to D). Importantly, HIR caused higher serum ALT (980 ± 39 vs. 773 ± 70; n = 8; P < 0.01; one-way ANOVA followed by Tukey test for multiple comparisons) and activity of caspase-3 and -8 in hepatic vagotomized mice than in intact mice (fig. 1). These data suggested that hepatic vagotomy aggravated liver injury and apoptosis induced by HIR.

The baseline values of serum ALT and liver caspase-3, -8, and -9 were similar between WT and α7nAChR^−/− mice. PNU-282987 (a selective α7nAChR agonist) pretreatment did not influence the baseline values. HIR induced significantly increased serum ALT and liver caspase activity in both α7nAChR^−/− and WT mice. However, HIR induced a higher level of ALT (1,147 ± 162 vs. 817 ± 76; P < 0.01) and caspase-3 (2.23 ± 0.31 vs. 1.72 ± 0.22; P < 0.05) and -8 activity in α7nAChR^−/− (n = 6) mice than in WT mice (n = 8). Pretreatment of PNU-282987 relieved ALT increase and reduced the response of activity of caspase-3, -8, and -9 caused by HIR in WT mice but not in α7nAChR^−/− mice (fig. 2). These data indicated that α7nAChR played an important role during HIR injury, and activating α7nAChR could protect against HIR damage.

Hepatic vagotomy alone did not induce changes histologically in C57Bl/6J mice. HIR resulted in extensive hepatocellular necrosis in the midzonal and pericentral areas of the liver. As a matter of fact, most areas of the liver tissue from PNU-282987-treated mice displayed structural features comparable to those from the control (fig. 3A). Serum ALT and H₂O₂ (210 ± 66 vs. 140 ± 21; n = 8; P < 0.01) were attenuated by activation of α7nAChR (fig. 3, B and C). TUNEL-positive cells and its percentage (58 ± 11 vs. 42 ± 9; n = 5; P < 0.05) and the activity of the caspase-3, -8, and -9 were substantially reduced by PNU-282987 pretreatment by HIR in hepatic vagotomized mice (fig. 4). These data demonstrated that PNU-282987 pretreatment could attenuate the damage caused by HIR in hepatic vagotomized mice, suggesting that activation of α7nAChR could compensate the loss of the hepatic vagus nerve in terms of the observed indices.

Primary hepatocytes and KCs were isolated from C57Bl/6J mice and cocultured in a transwell system. It was found that activation of α7nAChR by PNU-282987 pretreatment significantly reduced the percentage of early apoptotic hepatocytes (fig. 5A), caspase-3 (2.12 ± 0.18 vs. 1.54 ± 0.14; n = 3, repeated twice; P < 0.05), -8, and -9 activity in hepatocytes (fig. 5, B to D) in the cocultured system experienced 6-/2-h HR.

PNU-282987 pretreatment significantly reduced intracellular ROS production (fig. 6A) and supernatant H₂O₂ levels (57 ± 1.0 vs. 32 ± 6.1; n = 3; repeated twice; P < 0.01; fig. 6B) in KCs induced by 6-/2-h HR. Catalase, an antioxidant enzyme, pretreatment led to similar results (fig. 6, A and B). The percentage of apoptotic cells was about 40% when hepatocytes, which just suffered 6-h hypoxia, were incubated with the supernatant from KCs experienced 8-h normoxia for another 2 h under normoxia. Supernatant from KCs that suffered 6-/2-h HR exacerbated apoptosis of hepatocytes (56 ± 2.0), while the supernatant from KCs that suffered 6-/2-h HR with PNU-282987 (34 ± 1.7; n = 3; repeated twice; P < 0.01) or catalase (21 ± 0.9; n = 3; repeated twice; P < 0.01) pretreatment completely prevented this exacerbation (fig. 6C). This suggested that preventing excessive ROS (e.g., H₂O₂) production in KCs could inhibit hepatocyte apoptosis.

It was found that KCs (CD68-positive cells) were eliminated in the liver in mice when treated with Clophosome-A (fig. 7A). Compared with the control, HIR-induced H₂O₂ production in the liver was significantly reduced in mice treated with Clophosome-A. In addition, H₂O₂ production was not changed between mice treated with Clophosome-A and vehicle without HIR induction. This suggested that excessive ROS in mice was released in KCs during HIR (fig. 7B).

Serum sCD163 was measured as the specific marker for KC activity. PNU-282987 itself did not change KC activity, while HIR induced a significant increase of serum sCD163 in mice, which was prevented by PNU-282987 pretreatment (446 ± 119 vs. 240 ± 34; n = 8; P < 0.01; fig. 8A). In vitro, the sCD163 level (5.60 ± 0.18) in the supernatant of cultured KCs was significantly increased after 6-/2-h HR, and both PNU-282987 (4.23 ± 0.06; n = 3; repeated twice; P < 0.01) and catalase pretreatment (3.76 ± 0.20; n = 3; repeated twice; P < 0.01) decreased sCD163 levels (fig. 8B).

In this study, we found that either hepatic vagotomy or α7nAChR^−/− caused higher levels of ALT and caspase activity by HIR; activating α7nAChR with the selective agonist PNU-282987 attenuated HIR injury (including hepatocyte apoptosis) in mice with or without hepatic vagotomy. In vitro experiments demonstrated that PNU-282987 pretreatment reduced hepatocyte apoptosis in a cocultured system with KCs during HR. Either PNU-282987 or catalase pretreatment reduced both ROS and H₂O₂ production in KCs, and the supernatant from PNU-282987 or catalase-treated KCs prevented hepatocyte apoptosis during HR. KC elimination reduced HIR-induced H₂O₂ production in mice. Activating α7nAChR attenuated sCD163 both in mice by HIR and in KCs by HR (fig. 9). These direct and indirect evidences suggest that hepatic vagus nerve may play an important role during HIR-induced liver apoptosis through α7nAChR on KCs.

Apoptosis is one of the mechanisms responsible for cell death upon HIR.⁷  Approximately half of the hepatocytes undergo apoptosis within 24 h of reperfusion.^12,14  Apoptosis also occurs extensively in hepatocytes during liver transplantation.^11,42  Inhibition of caspases could attenuate hepatic injury by inhibiting liver apoptosis during HIR.¹⁴  Proapoptotic proteins, such as caspase-3, -8, and -9, are activated during reperfusion. Caspase-8 can directly or indirectly activate procaspase-3 to caspase-3. The indirect pathway is mediated by caspase-9 in the mitochondria.^2,43–45  Caspase-3 activity initiates the final execution stages of apoptosis. Thus, in this study, we assessed apoptosis mainly by measuring the activity of caspase-3, -8, and -9, combined with TUNEL (for liver tissues) staining.

The cholinergic antiinflammatory pathway indicates that the vagus nerve (which innervates most of the peripheral organs) can control inflammation by stimulating the α7nAChR on resident tissue macrophages through releasing its neurotransmitter acetylcholine.^22,24,46  Down-regulation of α7nAChR in tissue/organs innervated by the vagus nerve contributed to organ damage.²⁶  In a rat model of HIR, acetylcholine receptor agonists protected early phase of hepatic injury by inhibiting cytokine production,⁴⁷  and in a Fas-induced hepatitis mouse model, the hepatic vagus nerve could attenuate liver apoptosis possibly through α7nAChR.³⁸  In this study, HIR-induced liver apoptosis was more severe in hepatic vagotomized mice than in mice with intact vagus nerve. In α7nAChR^−/− mice, the activity of caspase-3 and -8 was significantly higher after HIR than that in WT control. Pretreatment with PNU-282987, a selective α7nAChR agonist, before ischemia, significantly attenuated the liver caspase-3, -8, and -9 activity caused by ischemia–reperfusion in WT mice but failed in α7nAChR^−/− mice. These findings indicate that vagus nerve and α7nAChR are essential in liver apoptosis caused by HIR, and activating α7nAChR could attenuate liver apoptosis induced by HIR.

To confirm the role of activating α7nAChR in HIR apoptosis and explore whether activation of α7nAChR could compensate the loss of hepatic vagotomy, we investigated the effects of PNU-282987 pretreatment on HIR injury in mice models with hepatic vagotomy. It was found that activation of α7nAChR largely inhibited HIR-induced serum H₂O₂ and liver caspase-3, -8, and -9 activity and reduced the number of apoptotic hepatocytes in hepatic vagotomy mice. These findings indicate that the hepatic vagus nerve–mediated protection is dependent on activation of α7nAChR in HIR injury.

KCs are resident macrophages of the liver, and sCD163 is a sensitive and specific marker for KC activity.^16,17  The fact that α7nAChR is expressed in macrophages is well accepted.^22,24  The study has shown that after reperfusion, KCs generated a large amount of ROS, a major initiator that leads to necrosis and apoptosis during the HIR injury.^6,19  Functional inactivation or depletion of KCs attenuates injury during early and late reperfusion.^20,21  Activation of α7nAChR protects neurons against apoptosis by augmenting the capability of cells to scavenge ROS.³⁰  H₂O₂, a major component of ROS, is considered a signaling molecule because of high stability and the ability to cross biologic membranes, and it contributes to oxidative stress and tissue damage pathologically.⁴⁸  Results from our in vitro experiments showed that activation of α7nAChR inhibited apoptosis of hepatocytes when cocultured with KCs during HR. To further investigate whether the role of α7nAChR on KCs on hepatocyte apoptosis during HR was modulated by ROS (e.g., H₂O₂), we treated hepatocytes with supernatants from KCs pretreated with PNU-282987 (a selective α7nAChR agonist) or with catalase (an antioxidant enzyme that catalyzes the decomposition of H₂O₂ to water and oxygen) before HR. In this work, we found that either PNU-282987 or catalase pretreatment reduced ROS and H₂O₂ production induced by HR in KCs, and these supernatants from KCs prevented hepatocyte apoptosis caused by HR. The fact that HIR-induced H₂O₂ production was reduced in KC-eliminated mice and activating α7nAChR attenuated sCD163 both in mice by HIR and in cultured primary KCs by HR indicated that KCs are involved in the protective mechanism. These findings suggest that activation of α7nAChR could prevent excessive ROS production in KCs, which subsequently blocks the initiation of hepatocyte apoptosis.

In summary, by using α7nAChR^−/− and hepatic vagotomy mice with versus without α7nAChR activation and KC-eliminated mice models in vivo, and by coculturing hepatocyte and KCs with versus without α7nAChR activation or catalase treatment in vitro, our results suggest that the hepatic vagus nerve attenuates hepatocyte apoptosis upon ischemia–reperfusion via α7nAChR on KCs in mice. The findings of the current study could have clinical significance. Specifically, preservation of hepatic vagus nerve should be emphasized in liver surgery, and α7nAChR on KCs could represent a target for attenuating HIR injury both in liver surgery and transplantation.

Supported by grants 81370558 and 81300081 from the National Natural Science Foundation of China (Beijing, China).

The authors declare no competing interests.