TBK1 Participates in Glutaminolysis by Mediating the Phosphorylation of RIPK3 to Promote Endotoxin Tolerance

Background: TBK1 (TRAF-associated NF-κB activator (TANK)-binding kinase 1), a nonclassical IκB kinase (IKK), and its effect on inammation have not been entirely claried. This study aimed to determine how TBK1 participates in the catabolism of glutamine by mediating the phosphorylation of receptor-interacting protein kinase 3 (RIPK3) and promoting macrophage endotoxin tolerance (ET). Methods: Lipopolysaccharide (LPS)-induced ET mice models were used for in vivo studies. Hematoxylin and eosin staining, immunouorescence staining and immunohistochemical staining protocols were used to analyze histological changes in the liver, and the related key pathway proteins were analyzed by western blotting. Kupffer cells (KCs) and bone marrow-derived macrophages (BMDMs) isolated from C57BL/6 mice were used to establish ET models. ELISA was used to detect the levels of necrosis factor-α (TNF-α), interleukin-10 (IL-10) and α-ketoglutarate (α-KG). The differentiation of M1/M2 macrophages was assessed by ow cytometry for assessing the cell-surface markers, CD86 and F4/80. The related key proteins were analyzed by western blotting and immunouorescence staining. Results: The phosphorylation of TBK1 was higher in the ET group than in the NET group in vitro and in vivo. As TBK1 was overexpressedd by shRNA, the phosphorylation level of TBK1 was increased, and the protective effect of ET was also strengthened. The TBK1 protein directly interacts with the RIPK3 protein and mediates the phosphorylation of RIPK3 in macrophages. According to our co-IP results, the glutamate dehydrogenase 1 (GLUD1) and RIPK3 proteins interact in ET models. In addition, activated RIPK3 can directly bind to GLUD1, thereby improving its catalytic activity and increasing the production of α-ketoglutarate in α-KG generated from glutaminolysis promotes activation and restricts M1 polarization to induce the macrophage tolerance of LPS. These ndings reveal mechanism for the metabolic control of inammation and for the induction of ET by modulating glutamine metabolism.


Background
In ammation is the general protective host response to stimulation by invading pathogens or endogenous signals, such as damaged cells, the clearance of necrotic cells, and tissue repair (Netea et al. 2017). Toll-like receptor 4 (TLR4) is the major pattern recognition receptor (PRR) involved in the detection of gram-negative bacteria and their associated endotoxins (Biswas et al. 2007; Petrasek et al. 2013).However, uncontrolled in ammation leads to extensive tissue damage and manifestations of pathological states such as sepsis (Venet and Monneret 2018). Pathophysiological adaptations to regulate overexuberant in ammation serve as an important mechanism for host protection against endotoxin shock. One of these protective mechanisms is endotoxin tolerance (ET) (Biswas and Lopez-Collazo 2009). Long-term exposure to lipopolysaccharide (LPS) or injection of sublethal doses of LPS in animals can induce tolerance of endotoxin that reprograms the in ammatory response, resulting in cells or organisms entering into a transient unresponsive state where they are unable to respond to further challenges with endotoxin (Biswas and Lopez-Collazo 2009; Liu et al. 2019). In several studies that have addressed the molecular basis of ET, LPS unresponsiveness has been associated with decreased expression of nuclear factor-κB (NF-κB) (Del et al. 2009).
Although TBK1 has a domain composition similar to that of canonical IKKs, it appears that TBK1 does not generally target NF-κB signaling and that the role of TBK1 in NF-κB activation is highly dependent on cellular and signal-induced contexts ( According to reports, after TNF-α stimulates cells, activated RIPK3 can directly bind to glutamate dehydrogenase 1 (GLUD1), which is known as a critical enzyme for catalyzing glutamine decomposition, thus improving its catalytic activity and increasing the production of α-ketoglutarate (α-KG) in macrophages (Mastorodemos et al. 2009;He et al. 2020). After inhibiting the activation of RIPK3 induced by LPS, it can directly reduce the catalytic activity of GLUD1 (Nelson et al. 2018).
Macrophages have remarkable plasticity that allows them to e ciently respond to environmental signals and change their phenotype, and their physiology can be markedly altered by both innate and adaptive immune responses (Mosser and Edwards 2008). The phenotypic remodeling of macrophages must also be accompanied by a change in cell energy metabolism during in ammation induced by endotoxin.
Growing evidence suggests that macrophages can limit the invasion of bacteria and other microorganisms by increasing the catabolism of glutamine (Davies et al. 2017; Cruzat et al. 2018). More importantly, the glutamine metabolic pathway is also involved in the speci c procedures required for macrophage polarization (Davies et al. 2017;Nelson et al. 2018;Liu et al. 2017). It has been reported that α-KG generated from glutaminolysis can promote M2 activation via Jmjd3-dependent metabolic and epigenetic reprogramming (Liu et al. 2017). In contrast, α-KG impairs proin ammatory responses in M1 macrophages by suppressing the NF-κB pathway. While glutaminolysis restricts M1 polarization, α-KG produced from glutaminolysis during LPS stimulation has a crucial role in promoting LPS-induced ET in macrophages (Cruzat et al. 2018;Liu et al. 2017).
Overall, these studies have revealed the roles of TBK1 and RIPK3 in endotoxin-induced in ammation and suggested that RIPK3 and GLUD1 play essential roles in modulating the polarization of macrophages by regulating the production of α-KG. We report here that the TBK1-RIPK3 pathway enhances the catalytic activity of GLUD1 to accelerate the catabolism of glutamine to produce more α-KG, thereby participating in the regulation of phenotypic remodeling and inducing macrophage ET. These results highlight the mechanisms controlled by glutaminolysis to ne-tune macrophage activities and suggest that modulation of glutamine metabolism would be an attractive strategy for harnessing macrophagemediated immune response.

Methods And Materials
Cell isolation RAW264.7 cells were obtained from the Laboratory of the Second A liated Hospital of Chongqing Medical University (Chongqing, China) and cultured in DMEM (Gibco Life Technologies) with 10% FBS (PAN-Biotechnology) and 1% penicillin-streptomycin (Beyotime Biotechnology) at 37°C, 95% humidity, and 5% CO 2 . The isolation methods of bone marrow-derived macrophages (BMDMs) and Kupffer cells (KCs) from C57BL/6 mice were performed according to the steps in the approach proposed by Li (Li et al. 2014). The cells were divided into three groups: the control group (CON) (treated with PBS), the endotoxin tolerance group (ET) (treated with 10 ng/mL LPS for 24 h, then 1000 ng/mL LPS for 12 h) and nonendotoxin tolerance group (NET) (treated with 1000 ng/mL LPS for 12 h).

Animals and protocols
Male C57BL/6 mice (8-10 weeks old, 22±2 g) were used for all experiments (Experiments Animal Center of Chongqing Medical University, Chongqing, China). Animals were fed standard rodent chow in a temperature-controlled environment with 50% humidity and 12 h light/dark cycles in cages of ve mice. To ensure the accuracy of the research, the mice were divided randomly into the following groups: normal saline group (CON) (mice were treated with intraperitoneal injections of sterile normal saline), endotoxin tolerance group (ET) (mice were pretreated with intraperitoneal injections of 50 µg/kg, 250 µg/kg and 500 µg/kg LPS for three consecutive days, and twelve hours after the last injection, mice were treated with intraperitoneal injections of 20 mg/kg LPS), nonendotoxin tolerance group (NET) (mice were intraperitoneally injected with 20 mg/kg LPS), control liposome group (mice were treated with an intraperitoneal injection of control liposomes) and neutral liposome group (mice were treated with an intraperitoneal injection of neutral liposomes).
Survival and weight changes were monitored once per 24 h period for 6 days, and then liver and blood samples were collected.

ELISA
The levels of TNF-α, interleukin-10 (IL-10) and α-KG in mouse serum and cell supernatants were determined by using ELISA kits from Jbswbio Biological Technology according to the manufacturer's instructions.

Flow cytometry analysis
To determine the macrophage differentiation phenotypes, BMDMs were isolated from C57BL/6 mice and NET and/or ET models were established as previously reported. Then, adherent cells were collected for ow cytometry analysis of F4/80 and CD86. RAW264.7 cells were separately stained with a monoclonal antibody speci c for PE-CD86. Both cell cultures were then analyzed using a FACSVerse ow cytometer (BD Biosciences, San Jose, CA, USA).

Western blot
After washing with PBS for 3 times, the cells were lysed in a lysis solution containing a protease inhibitor, phosphatase inhibitors and PMSF (KeyGen, China) and then sonicated on ice. After centrifugation (15 min, 12000 × g, 4°C), the supernatant was mixed with SDS-PAGE loading buffer and heated to 100°C for 10 min. The concentration of protein was determined by an Enhanced BCA Protein Assay Kit (Beyotime, China). The proteins were separated by SDS-PAGE (Bio-Rad, Hercules, CA, USA) and electrotransferred to PVDF membranes (Millipore, Billerica, MA, USA). The membranes were blocked in 5% BSA (Solarbio, Shanghai, China) for 1 h at 37°C. The corresponding primary antibodies were incubated with the membranes used overnight at 4°C. After being washed, the membranes were incubated with an HRPconjugated anti-rabbit or anti-mouse secondary antibody (1:5000) for 1 h at 37°C. Finally, the membranes were washed and stained with the Ultra High Sensitivity ECL Kit (MCE, China). The stained target proteins were detected by the Quantity One gel scanning system (Bio-Rad, Hercules, CA, USA).

GLUD1 enzyme activity assay
The enzyme activities of GLUD1 in mouse livers and macrophages were determined using a micro GLUD1 assay kit (Solarbio, BC1465) according to the manufacturer's instructions.
Immuno uorescence staining Cells were xed by immersion in methanol for 10 min and permeabilized with 0.3% Triton X-100 at room temperature for 10 min. After blocking with 5% BSA at room temperature for 60 min, the cells were incubated at 4°C overnight with primary antibodies. The next day, after a brief wash with PBS, the cells were incubated with the secondary antibody (1:500) at room temperature in the dark for 1 h. Finally, the cells were incubated with DAPI in the dark at room temperature for 10 min. After sealing with a uorescent quenching agent, stained cells were observed by laser scanning confocal microscopy (LSCM).

Histological analysis
Liver tissues were xed in 10% neutral formalin, dehydrated, embedded in para n, and cut into 5-µmthick sections. The para n sections were stained with hematoxylin and eosin (H&E). The sections were stained using commercial kits (ZSGB-BIO, China) for immunohistochemical staining following the manufacturers' instructions. After dewaxing, hydration and antigen repair of the tissue sections, the tissue sections were incubated with an antibody at 4°C overnight. The corresponding IgG polymer, DAB and hematoxylin were used. After dehydration and sealing, images were observed and collected by microscopy.

Liver function analysis
The alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels in serum were measured using commercial kits (Solarbio, BC1550 and BC1560).

Coimmunoprecipitation
Cells were treated with cell lysis buffer (Beyotime, China) to obtain cell lysates and retain the input material. Agarose beads (Protein A Agarose, Beyotime, China) were incubated with an antibody at 4°C for 2 h to ensure that the agarose beads bound to the antibody. Agarose beads bound to the antibody were added to the cell lysate and shaken at 4°C overnight. The sample protein was obtained by repeated elution of nonspeci c binding proteins. The sample was mixed with SDS-PAGE loading buffer and heated to 100°C for 10 min. The results were detected by western blot. IPKine™ HRP, Mouse Anti-Rabbit IgG LCS (Abbkine, China) was used as the secondary antibody.

Statistical analysis
Statistical analysis of the data was performed using Prism software (GraphPad Prism version 8.0). The standard error of the mean was calculated from the average of at least 3 independent samples under a given treatment condition. Statistically signi cant differences between two groups were determined by a two-tailed Student's t test. One-way ANOVA was used for multiple group comparisons. And a p value of less than 0.05 was considered signi cant; *p 0.05, **p 0.01, ***p 0.001, ****p 0.0001.

A low dose of LPS induces ET in vitro and in vivo
LPS has the ability to incite a vigorous in ammatory response and result in all the physiological manifestations of septic shock (Su 2002). However, after exposure to low concentrations of LPS, cells or organisms can enter into a transient unresponsive state and are unable to respond to further challenges with LPS (Seeley and Ghosh 2017). This phenomenon is ET, one of the important mechanisms for host protection against endotoxin shock.
To explore the mechanism of ET, we stimulated C57BL/6 mice with low-dose LPS and established ET models in vivo and KCs in vitro. KCs isolated from C57BL/6 mice were randomly divided into three groups: the control (CON) group, the nonendotoxin tolerance (NET) group and the endotoxin tolerance (ET) group (NET: 1000 ng/ml; ET: 10 ng/ml+1000 ng/ml). The ET and NET groups were cultured with complete medium containing 1000 ng/ml LPS for 12 h, but before that, the ET group was pretreated with 10 ng/mL LPS for 24 h. The ET group had lower expression of proin ammatory cytokines including TNFα and iNOS (Fig. 1A). Additionally, the level of TNF-α in the cell supernatants of the NET group was signi cantly increased compared to that in the ET group, while the level of IL-10 in the NET group was lower than that in the ET group (Fig. 1B).
To study the effect of ET in vivo, mice in the ET group were pretreated with LPS (intraperitoneal injections of 50 µg/kg, 250 µg/kg, 500 µg/kg LPS for three consecutive days) and then the two groups (NET and ET) were restimulated with LPS (20 mg/kg, intraperitoneally) for 24 h. All of the mice were observed for 6 days, and we found that compared to NET group, mice in ET group had signi cantly improved survival ( Fig. 1C). No mice died during the observation period in the saline-treated group (CON), which was not re ected in the survival curve. The liver sections from each group showed less tissue damage and in ammatory cell in ltration in the ET group (Fig. 1D), and the ET group had reduced levels of proin ammatory cytokines, such as TNF-and iNOS (Fig. 1E). Consistent with the in vitro results, the serum levels of TNF-α in the NET group were higher than that in the ET group, and the levels of IL-10 in the NET group were decreased compared to that in the ET group (Fig. 1F). In addition, the serum AST and ALT levels showed that the ET group experienced lower in ammatory activation and tissue damage (Fig. 1G). The above results demonstrated that the ET model was successfully established in vitro and in vivo.

TBK1 can mediate in ammatory regulation in ET
It was reported that promoting TBK1 phosphorylation can regulate the host innate immune response during infection (Chen 2021). To determine whether TBK1 was involved in ET, mice were treated as previously described to establish ET and NET models in vivo. In both the NET and ET groups, the phosphorylation of TBK1 increased; however, the phosphorylation of TBK1 was higher in the ET group than in the NET group. In addition, the ET group had fewer proin ammatory cytokines, such as TNF-and iNOS ( Fig. 2A). A higher level of TBK1 phosphorylation was also found in the ET group according to immuno uorescence staining (Fig. 2B). To clarify the relationship between TBK1 and ET in vitro, KCs isolated from normal C57BL/6 mice were treated as previously described. While total TBK1 was not affected in either the NET or ET groups (Fig. 2C), the phosphorylation of TBK1 in the ET group was increased to a greater extent (Fig. 2C-D), which suggested that TBK1 might be related to ET in vivo and in vitro.
To explore the role of TBK1 in ET, TBK1 expression was knocked down in RAW264.7 cells with siRNA (TBK1-siRNA). The TBK1-siRNA+ET group exhibited downregulated phosphorylation of TBK1 compared with the ET group, and the expression of proin ammatory cytokines, including TNF-and iNOS, was increased, which was the reverse of the previous results (Fig. 2E). As TBK1 was knocked down by siRNA, the phosphorylation level of TBK1 was decreased, and the protective effect of ET was also weakened. Correspondingly, the TBK1 gene was overexpressed in RAW264.7 cells by shRNA-TBK1 lentiviral transduction (LV-TBK1-OE). As predicted, the protein level of TBK1 increased signi cantly after lentiviral transduction for 72 h. The LV-TBK1-OE+ET group had the lower expression of proin ammatory cytokines (including TNF-and iNOS) and higher phosphorylation of TBK1 (Fig. 2F). The serum level of TNF-α in the TBK1-siRNA+ET group was signi cantly increased compared to the LV-TBK1-OE+ET group, while the serum level of IL-10 in the TBK1-siRNA+ET group was lower than that in the LV-TBK1-OE+ET group (Fig. 2G). These results revealed that the ET was partly dependent upon TBK1 in mediating the reduction of in ammation, and the phosphorylation level of TBK1 was closely related to the level of in ammation during ET: when TBK1 was knocked down and the phosphorylation level decreased, in ammation was aggravated and the secretion of in ammatory factors increased, while when TBK1 was overexpressed and the phosphorylation level increased, the in ammation was alleviated. The above results indicated that TBK1 could mediate in ammatory regulation in ET through protein phosphorylation.  (Fig. 3A). Immuno uorescence staining also con rmed that the phosphorylation level of RIPK3 was increased in the ET group (Fig. 3B). These results revealed that RIPK3 might be involved in ET. In addition, RIPK3 was found to interact with TBK1 by coimmunoprecipitation (co-IP) experiments established in KCs with coexpressed proteins in the ET group (Fig. 3C), which indicated that TBK1 and RIPK3 can interact during macrophage tolerance to LPS.
To further explore the relationship between TBK1 and RIPK3 in ET in macrophages, we rst knocked down the expression of TBK1 in RAW264.7 cells to observe whether the change in TBK1 expression would affect the expression of RIPK3. As the phosphorylation level of TBK1 decreased, the level of phosphorylated RIPK3 was decreased (Fig. 3D). As predicted, the phosphorylation levels of TBK1 and RIPK3 were both upregulated after increasing the expression of TBK1 by shRNA-TBK1 lentiviral transduction in the LV-TBK1-OE+ET group (Fig. 3E). These results revealed that TBK1 could affect the phosphorylation level of RIPK3 in ET, but the functional signi cance of the RIPK3 phosphorylation event and downstream substrates have not been fully clari ed ( Goossens et al. 1996), and growing evidence has shown that glutamine can transfer into mitochondria to function as an energy substrate (Zhang et al. 2009). The protein expression of GLUD1 and phosphorylated RIPK3 was higher in the ET group than in the NET group (Fig. 4A).
Immuno uorescence staining showed that the expression of GLUD1 increased in the ET group (Fig. 4B).
Furthermore, the enzyme activities of GLUD1, both in the liver and in BMDMs (the models were established as previously described), were signi cantly increased in the ET group compared with the NET group (Fig. 4C), while immunohistochemistry was also con rmed that the ET group had a higher expression of GLUD1 than the NET group (Fig. 4D). These results suggested that GLUD1 expression was indeed increased in ET, but how GLUD1 promotes LPS tolerance remains unclari ed.
Macrophages represent with distinct phenotypes with different functions according to the tissue microenvironment and different stimuli (Wynn et al. 2013; Murray and Wynn 2011). Liu reported that α-KG generated from glutaminolysis could promote M2 activation via Jmjd3-dependent metabolic and epigenetic reprogramming, while glutaminolysis restricted M1 polarization (Liu et al. 2017). Therefore, we carried out ELISAs to detect the contents of α-KG in the different groups. The levels of α-KG in serum and cell supernatants in the ET group were than those in the NET group (Fig. 4E). We then checked the ratios of M1-type and M2-type macrophages among BMDMs. As predicted, the ratio of M1-type macrophages in the ET group was signi cantly decreased (Fig. 4F). The expression of CD86 was increased in the NET group (Fig. 4G). As the enzyme activity of GLUD1 increased, the content of α-KG also increased restricting M1 polarization and promoting macrophage tolerance to LPS. Increasing the enzyme activity of GLUD1 could be an effective measure to induce ET in macrophages.
The regulatory effect of GLUD1 on ET was reversed after eliminating macrophages in mice Macrophages are critical to maintain the tissue homeostasis in innate and adaptive immune responses (Wynn et al. 2013). To determine whether the elimination of macrophages would have an effect on the induction of ET by GLUD1, we depleted macrophages in mice with clodronate (neutral) liposomes. Male C56BL/6J mice (8-10 weeks old, 22±2 g) were randomly divided into two groups: the control liposome group (injected peritoneally with 200 µl control liposomes) and the neutral liposome group (injected peritoneally with 200 µl neutral liposomes). The mice in the two groups were treated as previously described to establish ET in vivo. The control liposome group was divided into three groups: the control liposome control (C-CON) group, the control liposome NET (C-NET) group and the control liposome ET (C-ET) group, while the neutral liposome group was also divided into three groups: the neutral liposome CON (N-CON) group, the neutral liposome NET (N-NET) group and the neutral liposome ET (N-ET) group (Fig. 5A).
The KCs in the livers of mice in the control liposome group remained essentially unchanged, while the KCs in the neutral liposome group were basically eliminated (Fig. 5B). All of the mice were observed for 6 days, and we found that compared to the C-NET group, mice in the C-ET group had signi cantly improved survival. However, compared to the N-NET group, the survival of mice in the N-ET group was signi cantly decreased which was the reverse of the above results. Compared with the C-ET group, the percent survival of mice in the N-ET group was reduced by approximately 93%. Moreover, all mice in the N-ET group died before the sixth day of observation (Fig. 5C). H&E staining of the liver sections from each group showed that there was less tissue damage and in ammatory cell in ltration in the C-ET group than in the C-NET group, while there were no signi cant differences in tissue injury and in ammatory cell in ltration between the N-NET group and the N-ET group (Fig. 5D). Consistent with the H&E results: the serum levels of TNF-α and IL-10 were not signi cantly different between the N-NET and N-ET groups (Fig. 5E). Furthermore, the expression of GLUD1 was lower in the N-ET group than that in the C-ET group (Fig. 5F).
The enzyme activity of GLUD1 was decreased in the N-ET group compared with the C-ET group, resulting in a decrease in the content of α-KG in the liver (Fig. 5G-H). The in ammatory response in the neutral liposome group was still severe because the protective mechanism of ET was not induced, even if mice were stimulated with low dose LPS in advance. Taken together, these results indirectly proved that GLUD1 participated in ET by regulating macrophages and that the regulatory effect of GLUD1 on ET was reversed after eliminating the macrophages in mice.

RIPK3 induces ET by promoting the enzyme activity of GLUD1
The interaction of endogenous RIPK3 with GLUD1 increases after TNF-α stimulation, and in cells that have su cient RIPK3 expression, the gateway to glutamine use is readily opened (Zhang et al. 2009). In addition, it could directly reduce the catalytic activity of GLUD1 after inhibiting the activation of RIPK3 induced by LPS (Nelson et al. 2018). Based on the facts of previous studies, we tried to explore the roles of RIPK3 and GLUD1 in ET. According to our co-IP results, the GLUD1 and RIPK3 proteins interact in ET models in vivo and in vitro (Fig. 6A-B).
To further determine the regulatory effect of RIPK3 on GLUD1, we rst knocked down RIPK3 in RAW264.7 cells using siRNA and established ET and NET models (RIPK3-siRNA). The protein levels of RIPK3 and phosphorylated RIPK3 decreased signi cantly in the RIPK3-siRNA+ET group. As shown in Fig. 6C, with the decrease in RIPK3, the expression of GLUD1 in the RIPK3-siRNA+ET group was signi cantly lower than that in the ET group, while the expression of TNF-, and iNOS was increased. In addition, the enzyme activity of GLUD1 in the RIPK3-siRNA+ET group was much lower than that in the ET group (Fig. 6D). As the level of α-KG was decreased in the RIPK3-siRNA+ET group compared with the ET group (Fig. 6E), the ratio of M1 macrophages in the RIPK3-siRNA+ET group was increased (Fig. 6F). After knocking down the expression of RIPK3, the enzyme activity of GLUD1 decreased, and the production of α-KG decreased, resulting in a reduction in macrophage tolerance to LPS.
Correspondingly, when the RIPK3 gene was overexpressed in RAW264.7 cells by shRNA-RIPK3 lentiviral transduction (LV-RIPK3-OE), GLUD1 expression increased, while the enzyme activity of GLUD1 also increased ( Fig. 6G-H), which led to a higher level of α-KG in the LV-RIPK3-OE+ET group (Fig. 6I). These results indicated that increasing RIPK3 can promote the enzyme activity of GLUD1 to increase the level of α-KG. The expression of iNOS and TNF-in the LV-RIPK3-OE+ET group was signi cantly decreased after increasing the expression of RIPK3, which showed that overexpression of RIPK3 can promote the enzyme activity of GLUD1, thus strengthening the effect of α-KG on promoting ET and alleviating the in ammatory reaction (Fig. 6G). Taken together, RIPK3 can mediate the in ammatory response induced by LPS and enhance ET through glutaminolysis. TBK1 participates in glutaminolysis by mediating the phosphorylation of RIPK3 to promote ET TBK1 is required in ET both in vivo and in vitro; however, the speci c mechanism of how TBK1 participate in ET is still unclear. As TBK1 promoted the phosphorylation of RIPK3, we next sought to determine whether TBK1 participated in the glutamine metabolic pathway to promote ET. We knocked down the expression of TBK1 by using siRNA and found that the expression of GLUD1 in the TBK1-siRNA+ET group was lower than that in the ET group (Fig. 7A), as well as the enzyme activity of GLUD1 (Fig. 7B). In addition, the level of α-KG was also decreased in the TBK1-siRNA+ET group compared with the ET group (Fig. 7C). Flow cytometry showed that the ratio of M1 macrophages was lower in the TBK1-siRNA+ET group than in the ET group (Fig. 7D). Taken together, these ndings demonstrated that the decrease in TBK1 expression could reduce the expression of phosphorylated RIPK3 and then affect the enzyme activity of GLUD1, resulting in a decrease in the production of α-KG. Finally, shRNA-TBK1-OE lentiviral particles were used to increase TBK1 expression in RAW264.7 cells. Compared with the ET group, overexpression of TBK1 could successfully increased the expression of GLUD1 (Fig. 7E). Additionally, the enzyme activity of GLUD1 increased, and the level of α-KG increased in the LV-TBK1-OE+ET group compared with the ET group ( Fig. 7F-G). All of these ndings indicated that the expression of TBK1 can mediate the phosphorylation of RIPK3, thereby inducing GLUD1 to decompose glutamine to produce α-KG. These results suggested that the modulation of glutamine metabolism, which plays a vital role in macrophage ET, would be an attractive strategy for harnessing macrophage-mediated immune responses.

Discussion
ET serves as a protective mechanism to reduce the overproduction of proin ammatory cytokines in response to infection. Defects in the establishment of ET lead to a higher incidence of septic shock and mortality in individuals with infection. However, individuals with ET become immunocompromised (Nathan and Ding 2010). Therefore, understanding the mechanisms controlling ET is important to the design of interventions to ne tune immune responses. Here, we present a novel function by which TBK1 participates in glutaminolysis by mediating the phosphorylation of RIPK3 to promote macrophage tolerance to endotoxin. The TBK1 protein directly interacts with the RIPK3 protein and mediates the phosphorylation of RIPK3 in macrophages. Then, activated RIPK3 directly improves the enzyme activity of GLUD1, which increases the production of α-KG to restrict M1 polarization, resulting in the promotion of macrophage tolerance to LPS (Fig. 8).
Macrophages are one of the major cell types involved in ET, and LPS, a TLR4 ligand, is most commonly . Therefore, we speculate that TBK1 is involved in the polarization of macrophage and plays a role in macrophage ET. In our study, we found that the phosphorylation of TBK1 in KCs in the ET group was higher than that in KCs in the NET group, and the effect of inducing ET was weakened after the expression of TBK1 was inhibited by siRNA. When the phosphorylation level of TBK1 increased, the tolerance of macrophages to endotoxin was also strengthened. Consequently, TBK1 is a target molecule that regulates the tolerance of macrophages to endotoxin. Chan 2013). Therefore, the biological function of RIPK3 in in ammation is largely unknown (Newton 2015). In our study, the TBK1 protein directly interacted with the RIPK3 protein in KCs in co-IP experiments. Interestingly, RIPK3 also had a higher phosphorylation level in the ET group than in the NET group. Meanwhile, we noticed that the level of phosphorylated RIPK3 was decreased after knocking down the expression of TBK1. After increasing the expression of TBK1, the level of RIPK3 phosphorylation was also increased. Based on these data, we speculated that TBK1 could affect the phosphorylation level of RIPK3 in macrophages. GLUD1, a metabolic enzyme that converts glutamine to α-KG, was identi ed in the RIPK3 complex in treated cells (Zhang et al. 2009). Our study showed that the GLUD1 and RIPK3 proteins interacted in ET models in vivo and in vitro, and the expression of GLUD1 and phosphorylated RIPK3 was higher in the ET group than in the NET group. Furthermore, the enzyme activity of GLUD1 and the serum and cell supernatant levels of α-KG were increased in the ET group compared with the NET group. The ratio of M1type macrophages in the ET group signi cantly decreased. Growing evidence has shown that α-KG generated from glutaminolysis promotes M2 activation and restricts M1 polarization, resulting in a crucial role for this metabolite in promoting tolerance induced by LPS (Cruzat et al. 2018;Liu et al. 2017).
Interestingly, the regulatory effect of GLUD1 and α-KG on ET was reversed after eliminating macrophages with neutral liposomes: the percent survival of mice in the N-ET group was reduced by approximately 93% compared with that in the C-ET group; there were no signi cant differences in tissue injury or in ammatory cell in ltration between the N-NET group and the N-ET group; and the enzyme activity of GLUD1 was lower in the N-NET group than in the C-NET group, resulting in a decrease in α-KG. Taken together, GLUD1 and α-KG promote ET mainly by mediating the polarization of macrophages.
With the decrease in RIPK3, the expression of GLUD1 was signi cantly lower in the RIPK3-siRNA+ET group than that in the ET group. In addition, GLUD1 enzyme activity and α-KG levels were also decreased in the RIPK3-siRNA+ET group compared with the ET group, which resulted in an increase in the ratio of M1 macrophages in the RIPK3-siRNA+ET group. Correspondingly, when RIPK3 expression was increased, GLUD1 expression increased as well, which led to higher levels of α-KG.

Conclusion
Our study indicated that overexpression of RIPK3 can promote GLUD1 enzyme activity to increase the production of α-KG. Interestingly, we found that after knocking down the expression of TBK1, the expression of GLUD1 and enzyme activity of GLUD1 were decreased. Additionally, the level of α-KG was also decreased in the RIPK3-siRNA+ET group compared with the ET group, while the ratio of M1 macrophages was increased. In contrast, overexpression of TBK1 could mediate phosphorylation of RIPK3, thereby increasing the expression and enzyme activity of GLUD1 to achieve macrophage ET by enhancing the production of α-KG. Collectively, our study provides a novel pathway by which TBK1 participates in glutaminolysis by mediating the phosphorylation of RIPK3 to promote macrophage tolerance to endotoxin and unveils the potential for therapeutic targeting of the metabolic status of macrophages. Data are presented as the means ± SD of at least three independent experiments; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.      KG in cell supernatants was measured by ELISA (n=3). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.