TNFα-Mediated Necroptosis Aggravates Ischemia-Reperfusion Injury in the Fatty Liver by Regulating the Inflammatory Response

Nonalcoholic fatty liver disease (NAFLD) is more sensitive to ischemia and reperfusion injury (IRI), while there are no effective methods to alleviate IRI. Necroptosis, also known as “programmed necrosis,” incorporates features of necrosis and apoptosis. However, the role of necroptosis in IRI of the fatty liver remains largely unexplored. In the present study, we aimed to assess whether necroptosis was activated in the fatty liver and whether such activation accelerated IRI in the fatty liver. In this study, we found that the liver IRI was enhanced in HFD-fed mice with more release of TNFα. TNFα and supernatant of macrophages could induce necroptosis of hepatocytes in vitro. Necroptosis was activated in NAFLD, leading to more severe IRI, and such necroptosis could be inhibited by TN3-19.12, the neutralizing monoclonal antibody against TNFα. Pretreatment with Nec-1 and GSK′872, two inhibitors of necroptosis, significantly reduced the liver IRI and ROS production in HFD-fed mice. Moreover, the inhibition of necroptosis could decrease ROS production of hepatocytes in vitro. Inflammatory response was activated during IRI, and necroptosis inhibitors could suppress signaling pathways of inflammation and the soakage of inflammation cells. In conclusion, TNFα-induced necroptosis played an important role during IRI in the fatty liver. Our findings demonstrated that necroptosis might be a potential target to reduce the fatty liver-associated IRI.


Introduction
Hepatic ischemia-reperfusion injury (IRI) occurs under a variety of clinical conditions, such as liver transplantation and resection, as well as hemorrhagic shock [1,2]. IRI emerges due to not only the depletion of oxygen and ATP during hypoxia but also an excessive inflammatory response after reperfusion, leading to cell death, including apoptosis, necrosis, and ultimately organ dysfunction [3]. Although the nature of hepatic IRI has been widely studied, the molecular mechanism underlying the hepatocyte death remains largely unexplored. Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in Western countries, and it is predicted to become also the most frequent indication for liver transplantation by 2030 [4]. Both clinical studies and animal experiments have found that the steatotic liver is particularly susceptible to IRI [5,6]. Currently, as the main source of marginal donors, livers with greater than 30% of macrovesicular fat are considered unsuitable for transplantation due to their increased susceptibility to IRI and greater risk of early graft dysfunction [7]. The productions of proinflammatory cytokines, TNFα and IL1β, are increased during IRI, which are two critical mediators in the fatty liver [8]. TNFα plays a crucial role in almost all the pathogenic nodes of NAFLD, such as development of hepatic steatosis [9], hepatocyte death [10], and fibrosis [11]. However, whether proinflammatory cytokines, such as TNFα, are involved in regulating IRI in the fatty livers remains unknown.
Necroptosis is a novel mode of cell death, known as "programmed necrosis," which incorporates features of necrosis and apoptosis, and such type of cell death is controlled by two kinase receptor-interacting proteins (RIP1 and RIP3) [12]. At the functional level, the auto-and transphosphorylations of RIP1 and RIP3 are required for necrosome assembly and activation of necroptotic signaling [13]. RIP3 recruits and phosphorylates the mixed lineage kinase domain-like protein (MLKL), which in turn oligomerizes and causes irreversible cellular membrane damage, resulting in necrotic cell death [14]. It has been suggested that MLKL increases the production of mitochondrial reactive oxygen species (ROS) through mitochondrial targets. Accumulating evidence indicates that necroptosis plays a crucial role in the pathogenesis of inflammatory diseases, including NAFLD [15,16]. A study has found that necroptosis is best characterized in the setting of TNFα-induced cell death, which has high relevance for many types of liver diseases, but it may also occur under other conditions, including IRI [17]. It has already been shown that the inhibition of necroptosis attenuates necrotic cell death in cardiac, renal, and brain IRI as well as in the liver [18][19][20][21]. However, it remains unknown whether necroptosis can be activated by TNFα, and the role of necroptosis during IRI in the fatty liver is also unclear.
In the present study, we found that IRI and ROS production were more serious in the fatty liver compared with the normal liver. Macrophages stimulated with fatty acid expressed and released more TNFα during IRI both in vivo and in vitro. Moreover, necroptosis was activated in hepatocytes stimulated with TNFα or supernatant from palmitic acid-(PA-) treated macrophages followed by hypoxia-reoxygenation (H/R) injury. Necroptosis inhibitors necrostatin-1 (Nec-1) and GSK ′ 872 could protect livers from IRI in both CDand HFD-fed mice. In addition, Nec-1 and GSK ′ 872 reduced the ROS level induced by IRI. Furthermore, the inhibition of necroptosis could alleviate inflammatory reaction. Collectively, we, for the first time, investigated the roles of necroptosis during IRI in the fatty liver and provided a potential target to alleviate the fatty liver-associated IRI in liver surgery.

Materials and Methods
2.1. Animals. Experiments were conducted using male C57BL/6J mice, which were purchased from the Animal Center of the Affiliated Drum Tower Hospital of Nanjing University Medical School and housed under specific pathogen-free conditions. The animal protocols were approved by the Institutional Animal Care and Use Committee of Nanjing University, China, based on the NIH Guide for the Care and Use of Laboratory Animals. All efforts were made to minimize suffering of animals.

Mouse
Hepatic IR Injury. Briefly, 70% hepatic warm ischemia of mice was induced as previously described [22]. After anesthesia, the hepatic artery, portal vein, and bile duct branches to the left and median liver lobes were clamped for 60 min. Mice were sacrificed after 6 h, and liver and serum samples were collected. Blood samples were analyzed immediately using an automatic analyzer (Fuji, Tokyo, Japan) for alanine aminotransferase (ALT) and aspartate aminotransferase (AST). The livers were cut into pieces and preserved in 4% formalin or snap frozen in liquid nitrogen.
2.3. Cell Culture. Kupffer cells (KCs) were isolated as previously described [22]. Briefly, livers were perfused in situ via the portal vein with CMF-HBSS, followed by 0.02% type IV collagenase in HBSS. Then, the liver was dissociated in 0.02% type IV collagenase and filtered through sterile nylon gauze to remove undigested tissue and connective tissue, followed by centrifugation at 35g for 3 min for two times to separate nonparenchymal cells (NPCs). NPCs were then suspended in HBSS and layered onto a 50/25% two-step Percoll gradient (Sigma-Aldrich, USA) in a 50 mL conical centrifuge tube and centrifuged at 1,800g for 15 min at 4°C. KCs in the middle layer were collected and allowed attaching onto cell culture plates in RPMI 1640 medium containing 1% penicillin/streptomycin and 10% fetal bovine serum (FBS).
For primary mouse hepatocytes, cell suspension was washed in William's medium supplemented with 100 nM dexamethasone, 2 mM L-glutamine, 1 μM insulin, 10% FBS, and 1% penicillin/streptomycin (attachment medium) twice. The hepatocyte suspension was plated on rat tail collagen Icoated six-well plates in the attachment medium. The cells were incubated for 4 h, and then they were washed and further incubated in William's medium supplemented with 2 mM L-glutamine, 10% FBS, and 1% penicillin/streptomycin. Primary hepatocytes were used for experiments within 2-3 days after isolation.
For evaluation of H/R injury in vitro, cells were treated with PA (Sigma-Aldrich, USA) for 16 h, followed by fluxing with 95% N 2 /5% CO 2 in the absence of FBS and incubation at 37°C for 16 h. For reoxygenation, cells were transferred to a 95% air/5% CO 2 gas mixture and 10% FBS was added.
2.6. Enzyme-Linked Immunosorbent Assay (ELISA). The levels of TNFα (eBioscience, USA) in mouse serum and cell culture supernatants were measured using commercially available ELISA kits according to the manufacturer's instructions.

2.7.
Histological and Immunohistochemical Analysis. Paraffin liver sections (5 μm) were stained with hematoxylin and eosin (HE) for histological evaluation of IRI based on standard pathology methods and visualized using a light microscope. Liver damage was evaluated using Suzuki's score by two independent pathologists. To identify macrophages and neutrophils, paraffin-embedded liver sections (5 μm) were stained with F4-80 (Abcam, USA) and myeloperoxidase (MPO, Abcam, USA) as previously described [22].

PI Staining. Cells and frozen liver sections (4 μm)
were fixed with 4% paraformaldehyde for 30 min and washed twice with PBS. After treatment with 10 mg/mL DNasefree RNase at 37°C for 30 min, cell nuclei were stained with 10 mg/mL propidium iodide (PI, KeyGEN BioTECH, China) at room temperature for 5 min in the dark, then counterstained with DAPI, and observed under a fluorescence microscope.  2.11. Immunocytofluorescence (ICF) Analysis. Immunofluorescence analysis was performed according to the previously described protocols. Briefly, frozen liver sections (4 μm) fixed with acetone were penetrated with 0.3% Triton for 15 min. Then, the slides were blocked with 10% fetal sheep serum, followed by incubation with primary antibodies overnight at 4°C. After washing, slides were incubated with the corresponding secondary antibodies, followed by incubation with DAPI. Representative images were observed under an immunofluorescence microscope. The following antibodies were used: anti-RIP3 (Santa Cruz Biotechnology, sc-374639) and goat anti-mouse IgG H&L (Alexa Fluor® 488) (Abcam, ab150117).
2.12. Statistical Analysis. Statistical analysis was performed using the GraphPad Prism software version 6.0. All data were expressed as mean ± standard error of the mean (SEM). Normally distributed data were tested by Student's t-test. P value less than 0.05 was considered statistically significant.

Expression and Secretion of TNFα
Are Increased in the Fatty Liver after IRI. Compared with the control group (fed with a control diet, CD), the mice fed with a HFD exhibited significantly worse IRI, evidenced by higher serum ALT and increased Suzuki's score in HFD-fed mice (Figures 1(a) and 1(b) and Table 1). From the pathological liver sections, there were more edema, sinusoidal congestion, and necrosis in HFD-fed mice (Figure 1(a)). The expression of TNFα at both the liver tissue and serum levels was higher in HFDfed mice compared with those fed with a CD (Figures 1(c)  and 1(d)). Macrophages are the major source of inflammatory factors, including TNFα. Therefore, we extracted KCs and treated with PA (50 μm) to simulate macrophages in the fatty liver. After 24 h of stimulation, lipid accumulation was found in the cytoplasm (Figure 1(e)). Then, we investigated the effect of IRI on KCs with steatosis using an in vitro model of IRI. After H/R stimulation, the expression of TNFα at the mRNA level in KCs and cellular supernatant was increased. Moreover, PA treatment enhanced the expression of TNFα (Figures 1(f) and 1(g)). In conclusion, macrophages in the fatty liver expressed and released more TNFα compared with the normal liver after IRI.

TNFα Induces Necroptosis In Vitro.
Studies have shown that necroptosis is best characterized in the setting of TNFα-induced cell death. To further verify whether The results are presented as the mean ± SEM of 6 to 8 animals per group. a Significant difference (P < 0 001) versus the CD-sham group. b Significant difference (P < 0 01) versus the CD-sham group. c Significant difference (P < 0 01) versus the CD-IRI group. d Significant difference (P < 0 01) versus the HFD-sham group. necroptosis was associated with TNFα, primary mouse hepatocytes were treated with TNFα for 24 h. Figure 2(a) shows that the expressions of RIP1, RIP3, and MLKL were significantly increased upon stimulation of TNFα and necroptosis was induced by TNFα in a concentrationdependent manner. Cellular supernatant of KCs treated with PA and H/R could also activate necroptosis of hepatocytes ( Figure 2(b)). The viability of hepatocytes was assessed by dual staining of DAPI and PI. A high proportion of PI + cells was found after TNFα treatment (Figures 2(c) and 2(d)). The number of PI + cells was also markedly increased when stimulated with supernatant of KCs treated with PA and H/R (Figures 2(e) and 2(f)). Therefore, these results showed that TNFα could induce necroptosis in vitro.

Necroptosis Is Found in NAFLD after IRI.
We showed TNFα-induced necroptosis in vitro, and then we detected necroptosis in the fatty liver following IRI. Figure 3(a) reveals that necroptosis was activated in livers of HFD-fed mice. After IRI, necroptosis was further activated, exhibiting the upregulation of necroptotic markers (RIP1, RIP3, and MLKL). This finding suggested that necroptosis was further activated by IRI (Figure 3(b)). Moreover, immunofluorescence staining of RIP3 further indicated that necroptosis was activated in NAFLD with or without IRI (Figure 3(c)). Taken together, necroptosis was activated in the fatty liver and further enhanced after IRI, which might contribute to the enhanced IRI in HFD-fed mice.

Inhibition of Necroptosis Reduces IRI in NAFLD.
Necroptosis was activated during IRI, and the fatty liver further aggravated the activation. Therefore, we speculated whether the inhibition of necroptosis could attenuate IRI in the fatty liver. Nec-1 and GSK ′ 872, two necroptosis inhibitors, were administered by intraperitoneal injection 1 h before IRI. TN3-19.12, the neutralizing monoclonal antibody against TNFα [23], was also used to confirm whether TNFα was an effective trigger of necroptosis during liver IRI by intraperitoneal injection. We found that TN3-19.12, Nec-1, and GSK ′ 872 could significantly inhibit necroptosis in both CD-and HFDfed mice, showing increased expressions of RIP1, RIP3, and MLKL (Figures 4(a) and 4(b)). Consistent with our conjecture, Nec-1 and GSK′872 as well as TN3-19.12 reduced levels of ALT and AST in CD-and HFD-fed mice (Figures 4(c)-4(f)). HE staining and decreased Suzuki's scores revealed that liver injury was also reduced (Figure 4(g) and Tables 2 and 3). PI staining exhibited that hepatocyte necrosis, the direct result of IRI damage, was also alleviated after IRI (Figure 4(h)). We also studied the above-mentioned conjecture in vitro. Primary hepatocytes were stimulated with TNFα to induce IRI. Nec-1 and GSK ′ 872 could also reduce the expressions of RIP1, RIP3, and MLKL in TNFα-treated hepatocytes ( Figure 5(a)). Furthermore, Nec-1 and GSK ′ 872 also decreased the proportion of PI + hepatocytes ( Figure 5(b)). Taken together, the inhibition of necroptosis could reduce cell injury induced by TNFα during IRI in NAFLD.

Inhibition of Necroptosis Reduces ROS Production after IRI in NAFLD Both
In Vivo and In Vitro. The pathophysiology of hepatic IRI generally includes ROS production. In terms of entity of oxidative stress, the most relevant event is ROS production by activated inflammatory cells, while liver cells can also produce ROS by the uncoupled mitochondria due to oxygen readmission [24]. Moreover, it has been suggested that MLKL increases mitochondrial ROS production through mitochondrial targets [25]. Therefore, we detected whether Nec-1 and GSK ′ 872 could reduce the ROS production. Figure 6(a) reveals that ROS production was increased during IRI, and it was further enhanced in the fatty liver. Pretreatment with Nec-1 and GSK ′ 872 could significantly decrease the ROS level. The same results were also found by in vitro experiments (Figure 6(b)). Therefore, the inhibition of necroptosis could reduce the ROS level after IRI in NAFLD, and this might be another mechanism of alleviating IRI in the fatty liver.

Inhibition of Necroptosis Reduces the Inflammatory
Response after IRI in NAFLD. Danger-associated molecular patterns (DAMPs) are either passively released by necrotic cells and the damaged extracellular matrix or are actively secreted by stressed and injured cells [26]. Various types of DAMPs are released during liver IRI, and these DAMPs can interact with and activate toll-like receptor (TLRs). The TLRs are one of the components by which the innate immune system senses the invasion of pathogenic microorganisms or tissue damage by recognizing DAMPs [27]. Transcription factors have been shown to mediate TLR activation in liver IRI, including NF-κB, JNK, ERK, p38, and IKBα [28]. Therefore, we assessed the expressions of these transcription factors in the fatty livers. Figure 7(a) shows that compared  The results are presented as the mean ± SEM of 6 to 8 animals per group. a Significant difference (P < 0 01) versus the PBS-IRI group. b Significant difference (P < 0 05) versus the PBS-IRI group. The results are presented as the mean ± SEM of 6 to 8 animals per group. a Significant difference (P < 0 01) versus the PBS-IRI group. b Significant difference (P < 0 05) versus the PBS-IRI group.
with CD-fed mice, all the transcription factors were activated in HFD-fed mice and IRI increased the expressions of NF-κB, JNK, ERK, p38, and IKBα. Moreover, we found that the liver inflammatory response after IRI was inhibited by Nec-1 and GSK′872 in both CD-and HFD-fed mice (Figures 7(b) and 7(c)). In addition, the same findings were achieved by in vitro experiments (Figure 7(d)). There was also more infiltration of MPO-and F4-80-positive cells in HFD-fed mice after IRI. Inhibition of necroptosis could also decrease the soakage of inflammation cells (Figures 7(e) and 7(f)). In summary, the inhibition of necroptosis by Nec-1 and GSK ′ 872 could reduce the inflammatory response after IRI in NAFLD, which might be another mechanism protecting the fatty liver from IRI.

Discussion
IRI results from a prolonged ischemic insult, followed by the restoration of blood perfusion. Hepatic IRI can lead to severe liver injury, and it is a major cause for the failure of liver transplantation. However, the fatty liver is more sensitive to IRI, leading to more severe outcomes of patients. Moreover, in the past two decades, urbanization has led to sedentary lifestyle and overnutrition, setting the stage for the epidemic of obesity and NAFLD, which is currently estimated to be 24% worldwide [29,30]. Therefore, it is urgently necessary to prevent and attenuate IRI [22,31]. However, there are no available effective and simple methods available to reduce IRI in the fatty liver. In the present study, we also found that IRI in the fatty liver was more severe compared with the normal liver and the TNFα level was increased in serum and liver of NASH animals, which was in agreement with the previous report [11]. As a newly defined type of programmed cell death, necroptosis is tightly controlled by the multiprotein complex of RIP1 and RIP3, known as the necrosome. Accumulating evidence indicates that MLKL and the protein kinases (RIPK1 and RIPK3) contribute to inflammatory processes through both the induction of necroptotic cell death and other cellular changes [32,33]. Necroptosis has been shown to be involved in various ischemic, inflammatory, and neurodegenerative human disorders [34]. Necroptosis has been identified as a mechanism of cell death in renal, cardiac, and retinal IRI [18,35,36]. A recent study has found that necroptosis contributes to hepatic damage during IR, which induces autophagy via ERK activation [21]. However, another study has found that necroptotic molecules are not increased in the necrosis-dominant hepatic IRI model, and antinecroptosis does not have an overall protective effect on necrosis-dominant hepatic IRI [37]. Therefore, the role of necroptosis or even whether necroptosis is activated in liver IRI remains largely unexplored. In the present study, we found that TNFα was upregulated in the fatty liver and its level was further increased after IRI. Early studies have found that TNFα is the best characterized activator to induce necroptosis. Therefore, we stimulated hepatocytes with TNFα. As expected, necroptosis was significantly activated by TNFα as well as supernatant of KCs treated with PA and H/R. Moreover, we tested the expressions of RIP1, RIP3, and MLKL in liver tissues suffering IRI, and all three markers of necroptosis were upregulated. In addition, the activation of necroptosis was much more intensive in the fatty liver, which was consistent with the level of TNFα. Furthermore, Nec-1 and GSK ′ 872 could significantly reduce necroptosis and protect the liver from IRI in both CD-and HFD-fed mice (Figure 8). To the best of our knowledge, we, for the first time, demonstrated that necroptosis was activated during IRI in the fatty liver, and inhibition of necroptosis could reduce IRI in NAFLD.
Immune cells, such as macrophages, are activated during the ischemic phase and even more during reperfusion. Once activated, they produce proinflammatory cytokines, including TNFα [38]. Cytokines play critical roles by stimulating hepatocytes to produce ROS, greatly contributing to their damage [39]. In the present study, we also found that ROS production was increased in hepatocytes of the fatty liver after IRI as well as hepatocytes stimulated with TNFα. Inhibition of necroptosis could reduce the level of ROS. Therefore, necroptosis contributed to the ROS production, which might aggravate the IRI in the fatty liver. DAMPs released during liver IRI bind to a group of receptors termed pattern recognition receptors (PRRs) to induce the inflammatory response [40]. Transcription factors, including NF-κB, JNK, ERK, p38, and IKBα, participate in the activation of the inflammatory response. In this study, we found that transcription factors, as well as the soakage of inflammation cells, were all significantly upregulated during IRI in the fatty liver, while the inhibition of necroptosis could reduce the inflammatory response (Figure 8).

Conclusions
In the present study, we found a new mechanism, which could explain why the fatty liver was more susceptible to IRI, and demonstrated the mechanism underlying the necroptosis of the fatty liver. Our findings provided a potential target to reduce the fatty liver-associated IRI in liver transplantation.  Figure 8: TNFα-mediated necroptosis aggravates IRI in the fatty liver by regulating the inflammatory response. Macrophages were activated during IRI, which secreted more TNFα in the fatty liver compared with the normal liver. Necroptosis induced by TNFα was activated in NAFLD, leading to more severe IRI. Necroptosis promoted the production of DAMPs, which increased signaling pathways of inflammation and the soakage of inflammation cells to aggravate IRI. Inhibition of necroptosis with Nec-1 and GSK′872 could decrease the signaling pathway of inflammation and ROS production to protect the liver from IRI.