Isoform- and Cell Type–Specific Roles of Glycogen Synthase Kinase 3 N

Isoform- and Cell Type−Specific Roles of Glycogen Synthase Kinase 3 N-Terminal Serine Phosphorylation in Liver Ischemia Reperfusion Injury This information is current as of September 25, 2021. Ming Ni, Haoming Zhou, Jing Zhang, Dan Jin, Tianfei Lu, Ronald W. Busuttil, Jerzy W. Kupiec-Weglinski, Xuehao Wang and Yuan Zhai J Immunol published online 17 July 2020 http://www.jimmunol.org/content/early/2020/07/16/jimmun Downloaded from ol.2000397

Supplementary http://www.jimmunol.org/content/suppl/2020/07/16/jimmunol.200039 http://www.jimmunol.org/ Material 7.DCSupplemental

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 25, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2020 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published July 17, 2020, doi:10.4049/jimmunol.2000397 The Journal of Immunology

Isoform- and Cell Type–Speciﬁc Roles of Glycogen Synthase Kinase 3 N-Terminal Serine Phosphorylation in Liver Ischemia Reperfusion Injury

Ming Ni,*,†,1 Haoming Zhou,*,†,1 Jing Zhang,* Dan Jin,*,‡ Tianfei Lu,*,x Ronald W. Busuttil,* Jerzy W. Kupiec-Weglinski,* Xuehao Wang,† and Yuan Zhai*

Glycogen synthase kinase 3 (Gsk3) a and b are both constitutively active and inhibited upon stimulation by N-terminal serine phosphorylation. Although roles of active Gsk3 in liver ischemia reperfusion injury (IRI) have been well appreciated, whether Gsk3 N-terminal serine phosphorylation has any functional signiﬁcance in the disease process remains unclear. In a murine liver partial warm ischemia model, we studied Gsk3 N-terminal serine mutant knock-in (KI) mice and showed that liver IRI was

decreased in Gsk3aS21A but increased in Gsk3bS9A mutant KI mice. Bone marrow chimeric experiments revealed that the Downloaded from Gsk3a, but not b, mutation in liver parenchyma protected from IRI, and both mutations in bone marrow–derived cells exacerbated liver injuries. Mechanistically, mutant Gsk3a protected hepatocytes from inﬂammatory (TNF-a) cell death by the activation of HIV-1 TAT-interactive protein 60 (TIP60)–mediated autophagy pathway. The pharmacological inhibition of TIP60 or autophagy diminished the protection of the Gsk3a mutant hepatocytes from inﬂammatory cell death in vitro and the Gsk3a mutant KI mice from liver IRI in vivo. Thus, Gsk3 N-terminal serine phosphorylation inhibits liver innate immune activation but

suppresses hepatocyte autophagy in response to inﬂammation. Gsk3 aS21, but not bS9, mutation is sufﬁcient to sustain Gsk4 http://www.jimmunol.org/ activities in hepatocytes and protect livers from IRI via TIP60 activation. The Journal of Immunology, 2020, 205: 000–000.

schemia reperfusion (IR) injury (IRI) remains one of the including Kupffer cells (KC)/macrophages, dendritic cells, and primary causes of liver dysfunction and failure after hepatic neutrophils (1, 2). These soluble mediators not only cause direct I tumor resection, liver transplantation, and trauma. Following hepatocellular damage but also recruit/activate more inflammatory the initial ischemic insult, reperfusion-triggered sterile inflam- cells from circulation into the tissue to amplify the immune re- mation is the major driving force of the disease pathogenesis. Host sponse and tissue damage. sentinel pattern recognition receptors are activated by danger- At the cellular level, TLR 4–mediated KC/macrophage activa- associated molecular pattern, such as high mobility group box 1, tion and TNF-a–mediated hepatocyte/endothelium death are by guest on September 25, 2021 derived from necrotic/stressed cells, resulting in the production critical in the pathophysiology of liver IRI (3–8). In the explora- of inflammatory cytokines/chemokines by innate immune cells, tion of regulatory mechanisms of these key processes, glycogen synthase kinase 3 (Gsk3) b has attracted much attention because *Dumont-UCLA Transplant Center, Division of Liver and Pancreas Transplantation, of its ubiquitous expression pattern and indispensable roles in Department of Surgery, David Geffen School of Medicine at University of California, cell metabolism, proliferation, differentiation, and death (9–12). † Los Angeles, Los Angeles, CA 90095; Department of Liver Surgery, First Affili- Gsk3b is a serine/threonine kinase and transduces signals down- ated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210008, China; ‡ a Department of Obstetrics and Gynecology, Shanghai Jiaotong University, Shanghai stream of both TLR4 and TNF- . Distinctive from MAPKs, 200025, China; and xLiver Surgery, Renji Hospital, Shanghai Jiaotong University, Gsk3b differentially regulates pro- and anti-inflammatory cyto- Shanghai 200025, China kine gene transcriptions (10). Gsk3b also determines the sensi- 1 M.N. and H.Z. contributed equally to this work. tivity of hepatocyte response to TNF-a that global Gsk3b ORCIDs: 0000-0003-1064-374X (M.N.); 0000-0001-8157-1531 (R.W.B.). knockout (KO) is lethal because of severe TNF-a toxicity in livers Received for publication April 10, 2020. Accepted for publication June 21, 2020. (13, 14). Additionally, Gsk3b has been found to regulate non– This work was supported by National Institutes of Health National Institute of receptor-mediated intrinsic cell death via the mitochondrial per- Allergy and Infectious Diseases Grant P01Al120944 (to Y.Z. and J.W.K.-W.), the meability transition pore (9, 15), which is critically involved in Dumont Research Foundation, and National Nature Science Foundation of China Grants 81600450, 81870448, and 3193002 (to H.Z. and X.W.). ischemia-induced cell death mechanism. a b Address correspondence and reprint requests to Dr. Xuehao Wang or Dr. Yuan Zhai, Gsk3 has two isoforms, and , which share extensive ho- Department of Liver Surgery, First Affiliated Hospital of Nanjing Medical University, mology in the kinase domain but are distinctive in their N- and 300 Guanzhou Road, Nanjing, People’s Republic of China (X.W.) or Dumont-UCLA C-terminals. The majority of Gsk3 research has been focused on Transplant Center, The David Geffen School of Medicine at University of California, b Los Angeles, 77-120 CHS, 10833 Le Conte Avenue, Los Angeles, CA 90095 (Y.Z.). the isoform using small molecule inhibitors. However, these E-mail addresses: [email protected] (X.W.) or [email protected] (Y.Z.) inhibitors actually do not discriminate between kinase domains of The online version of this article contains supplemental material. Gsk3 isoforms (16). Although Gsk3 isoforms have overlapping Abbreviations used in this article: EM, electron microscopy; ER, endoplasmic functions, isoform-specific properties have begun to be revealed in reticulum; Gsk3, glycogen synthase kinase 3; HIF, hypoxia-inducible factor; IR, recent studies. Whereas global Gsk3b KO mice are embryonic ischemia reperfusion; IRI, IR injury; KC, Kupffer cell; KI, knock-in; KI-WT, WT a mice that received mutant KI; KO, knockout; LDH, lactate dehydrogenase; 3-MA, lethal (13), Gsk3 KO mice are viable and “normal” (17). It has 3-methyladenine; NPC, nonparenchymal cell; sALT, serum alanine aminotransfer- been shown in myocardial disease models that Gsk3 a and b ase; Tg, thapsigargin; TIP60, HIV-1 TAT-interactive protein 60; WT, wild type; played distinctive or even opposite functions (18–22). WT-KI, mutant KI mice that received WT; WT-WT, WT mice that received WT. Gsk3 is a unique intracellular signaling kinase that is consti- Copyright Ó 2020 by The American Association of Immunologists, Inc. 0022-1767/20/$37.50 tutively active in resting cells and inhibited by phosphorylation in

www.jimmunol.org/cgi/doi/10.4049/jimmunol.2000397 2 ACTIVE Gsk3a PROMOTES HEPATOCYTE AUTOPHAGY response to various stimuli. The N-terminal phosphorylation of isoflurane and placed in a designed warm container (HTP-1500 Heat serine residues of Gsk3 a at Ser21 and b at Ser9 are the major Therapy Pump; Adroit Medical Systems, Loudon, TN) to keep the tem- mechanism of Gsk3 inactivation, which are induced in hearts and perature at 29˚C. After 90 min of partial hepatic warm ischemia, the clip was removed, initiating liver reperfusion. Mice were sacrificed after 1, 6, livers by ischemia reperfusion (IR) (23, 24). Multiple intracellular or 24 h postreperfusion. Sham controls underwent the same procedure but signaling pathways can act on the serine site of Gsk3s. Among without vascular occlusion. HIV-1 TAT-interactive protein 60 (TIP60) in- them, the PI3K–Akt pathway triggered by growth factors/insulin hibitor TH1834 (50 mg/kg; Axon Medchem, Reston, VA) or autophagy and inflammation and the AMPK–mTOR–S6K pathway induced inhibitor 3-methyladenine (3-MA, 30 mg/kg; Sigma, St. Louis, MO) or PBS (vehicle control) was administered i.p. 1 h prior to the onset of liver by nutrient/energy alterations are highly relevant to organ IRI. We ischemia. Serum alanine aminotransferase (sALT) levels were measured have shown in our previous studies that PI3K inhibitor reduced by IDEXX Laboratories (Westbrook, ME). Part of the liver specimens liver IR-induced Gsk3b phosphorylation and increased liver IRI were fixed in 10% buffered formalin and embedded in paraffin. Liver (24), whereas PTEN KO-induced PI3K activation had the opposite sections (4 mm) were stained with H&E. The severity of liver IRI was effect (25). Although experiments using Gsk3 inhibitors (24) or gene graded blindly using Suzuki’s criteria on a scale from 0 to 4. No necrosis or congestion/centrilobular ballooning is given a score of 0, whereas severe KO (26) have established their roles as active kinases, the functional congestion and .60% lobular necrosis is given a score of 4. significance of Gsk3 inhibitory phosphorylation in disease pathogenesis remains elusive. Indeed, studies in heart diseases have Primary hepatocytes isolation and cultures documented that active Gsk3 and Gsk3 N-terminal serine phos- Mouse primary hepatocytes were isolated as follows: livers were perfused in phorylation may play distinctive roles in different aspects of disease situ via the portal vein with calcium- and magnesium-free HBSS, followed mechanisms (22, 27, 28). Little is known about the functional by 0.27% collagenase IV (Sigma). Perfused livers were dissected and teased through 70-mM nylon mesh cell strainers (BD Biosciences, San Jose, CA). significance of Gsk3 inhibitory phosphorylation in liver diseases. Liver cells were suspended in 20 ml DMEM with 10% FBS and centri- Downloaded from In the current study, we took advantage of Gsk3 a or b fuged at 50 3 g for 2 min. The supernatant was decanted, and then 20 ml N-terminal serine phosphorylation-resistant mutant (S21A/S9A) 40% cold Percoll (P1644; Sigma, St. Louis, MO) was added to each tube knock-in (KI) mice (11) and their bone marrow chimeras to de- and mixed gently and then centrifuged at 150 3 g for 7 min. The super- termine the role of Gsk3 inhibitory phosphorylation in an isoform- natant was removed, and cells were washed with DMEM + 10% FBS (Thermo Fisher Scientific, Waltham, MA). Hepatocytes were suspended in and cell type–specific manner in sterile liver inflammatory tissue plating medium (Williams E Medium with Hepatocyte Thawing and injury post-IR. Plating Supplement Pack, Thermo Fisher Scientific) and plated in collagen http://www.jimmunol.org/ type I–coated plates. After 3-h culture, the plating medium was replaced with maintenance medium (Williams E Medium with Hepatocyte Main- Materials and Methods tenance Supplement Pack) and cultured overnight. Mice For cell death experiments, hepatocytes were isolated from WT, or GSK3 KI mice were pretreated with actinomycin D (50 ng/ml, Bio-Techne, a b Gsk3 S21A and S9A double mutant (alanine replaces serine at position Minneapolis, MN) for 30 min or thapsigargin (Tg) (1 mM, Bio-Techne) 21 or 9, resistant to inhibitory phosphorylation) KI were provided by for 1 h, followed by TNF-a (25 ng/ml, R&D Systems, Minneapolis, MN) Dr. D.R. Alessi of University of Dundee. We have rederived them by treatment for 12 h; culture medium was collected for further analysis. GSK3 crossing with wild type (WT) C57BL/6 mice (The Jackson Laboratory, inhibitor SB216763 (10 mM, Bio-Techne) and autophagy inhibitor 3-MA Bar Harbor, ME). Mice were housed in the University of California Los (5 mM, Bio-Techne) was added 30 min prior to the above treatment in some Angeles animal facility under specific pathogen-free conditions and re- experiments. For autophagy flux study, chloroquine (50 mM, Bio-Techne) was by guest on September 25, 2021 ceived humane care, according to the criteria outlined in the “Guide for the added 1 h prior to TNF-a stimulation. Cell death and survival were detected by Care and Use of Laboratory Animals” prepared by the National Academy lactate dehydrogenase (LDH) assay and CCK-8 assay (BioChain Institute, of Sciences and published by the National Institutes of Health (publication Hayward, CA), respectively, according to manufacturer’s instructions. 86-023 revised 1985). Genotyping of GSK3a21A and GSK3b9A mice was carried out Fluorescence/confocal microscopy of hepatocyte autophagy by PCR of genomic DNA isolated from mouse ear or tail samples as described (11). Primers for GSK3a21A were as follows: primer 1 (59- CYTO-ID Autophagy detection kit (Enzo Life Sciences, Farmingdale, NY) TTGAAGTGGCTGGTACTGG CTCTG-39) and primer 2 (59-GTGTG- was used in our experiments. In brief, primary hepatocytes were seeded CTCCAGAGTAGTACCTAGC-39), resulting in a 271-bp product from onto Nunc eight-well Permanox Plastic chamber slides (Thermo Fisher the WT allele and a 317-bp product from the mutant allele. Primers for Scientific) at 50–70% confluence and cultured overnight. Hepatocytes were GSK3b9A were as follows: primer 3 (59-TCACTGGTCTAGGGGTGG- treated with either Tg (1 mM), H202 (100 mM), or TNF-a (25 ng/ml; R&D TGGAAG-39) and primer 4 (59-GGAGTCAGTGACAACACTTAACTT-39), Systems) in the absence or presence of 10 mM TIP60 inhibitor NU9056 resulting in a 233-bp product from the WT allele and a 352-bp product from (Tocris Bioscience, Minneapolis, MN) for 6 h. The treated cells were the mutant allele. washed twice with 13 Assay Buffer, followed by the incubation with 100 ml of Microscopy Dual Detection Reagent for 30 min at 37˚C. The labeled Generation of chimeric mice cells were washed once with 13 Assay Buffer and fixed with 4% formaldehyde for 20 min. After being washed three times with 13 Assay Recipient mice were exposed to otherwise lethal 1000 cGy from a Cesium Buffer, a coverslip was placed on the microscope slide. Cells were an- 3 6 source 6 h prior to receiving 2.5 10 bone marrow cells by tail vein alyzed by wide-field confocal microscopy with 4003 magnification. A injection. The bone marrow cells were prepared under sterile conditions standard FITC filter set was used for imaging the autophagic signal, and from the tibia and femur bones of donor mice. Recipient mice were a DAPI filter was set for nuclear signal. monitored closely for the first 2 wk to ensure the successful bone marrow engraftment. Four groups of chimeric mice were generated: WT mice that Quantitative RT-PCR received WT (WT-WT), WT mice that received mutant KI (KI-WT) bone marrow, mutant KI mice that received WT (WT-KI), or mutant KI mice Total RNA (2 mg) was reverse-transcribed into cDNA using a SuperScript that received KI bone marrow. At 6–8 wk postreconstitution, liver non- III First-Strand Synthesis System (Invitrogen, Carlsbad, CA). Quantitative parenchymal cells (NPC) were isolated from WT or Gsk3a/b mutant KI or PCR was performed using DNA Engine with four detectors (MJ Research, WT-KI or KI-WT chimeric mice. Genomic DNA was isolated. WT and Waltham, MA). In a final reaction volume of 20 ml, the following were mutant Gsk3a/b fragments were analyzed by PCR. added: 13 SuperMix (Platinum SYBR Green qPCR Kit; Invitrogen, Carlsbad, CA), cDNA, and 0.5 mM of each primer. Amplification condi- Model of liver partial warm IRI tions were 50˚C for 2 min and 95˚C for 5 min, followed by 50 cycles of 95˚C for 15 s and then 60˚C for 30 s. Primers used to amplify specific A murine model of partial liver warm IRI was used, as described (29). Male mouse gene fragments were the same as described previously (30). mice at 6–8-wk-old were employed in our experiments. Briefly, after anes- thesia with 2.5% isoflurane, mice were injected with heparin (100 mg/kg). A Western blot midline incision was made to expose the abdominal cavity, and an atraumatic clip was then used to interrupt the arterial and portal venous blood supply to Liver tissue or cellular proteins were extracted with ice-cold lysis buffer the cephalad lobes of the liver. Mice were maintained anesthetized with (1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 10% glycerol, The Journal of Immunology 3

137 mM sodium chloride, and 20 mM Tris [pH 7.4]). Proteins (30 mg) were subjected to 12% SDS-PAGE electrophoresis and transferred to PVDF nitrocellulose membrane. Abs against Phospho-Gsk3a, Phospho-Gsk3b, Gsk3a, Gsk3b, LC3B, Phospho-TIP60 (Ser86), Phospho-AMPKa (Thr172), Phospho-p70 S6K (Thr389), and b-actin (Cell Signaling Technology, Danvers, MA) were used for Western blot analysis. ImageJ 1.47v soft- ware was used to quantitate Western blot bands. Electron microscope Fresh liver tissue blocks (1 mm3 in size) were fixed in 2% glutaraldehyde in PBS for 2 h and postfixed in 1% osmium tetroxide for 2 h. Samples were then dehydrated with an ascending series of alcohol before embedding. Ultrathin sections were cut and stained with lead citrate. Images were captured using a transmission electron microscope (HT7700; HITACHI). Statistical analysis Results are shown as mean 6 SD. Unpaired Student t test was used for pairwise statistical analyses, with a p value , 0.05 (two tailed) considered statistically significant.

Results Downloaded from The Gsk3 aS21 but not bS9 mutation protects livers from IRI We have shown previously that ischemia induced a transient downregulation of N-terminal serine 9 phosphorylation of Gsk3b, followed by upregulation of its phosphorylation during reperfusion in livers by Western blot analysis (24). A similar phosphorylation pattern was found in Gsk3a S21 in IR livers (Fig. 1). To determine FIGURE 1. Liver Gsk3 a and b N-terminal serine phosphorylation http://www.jimmunol.org/ the functional significance of Gsk3 N-terminal phosphorylation in profiles in response to IR. Liver tissues were harvested after 90-min is- liver IRI, we compared Gsk3 aS21A or bS9A single and double chemia and 0-, 1-, or 6-h reperfusion, as described in the Materials and mutant KI mice with their WT controls. Liver injury and inflam- Methods. Total tissue proteins were analyzed by Western blots with anti- a b a b b mation were measured at 6 and 24 h postreperfusion. The result phosphorylated Gsk3 S21, Gsk3 S9, total Gsk3 and , and -actin Abs a b showed that liver IRI was significantly decreased in a single and (upper panel). Average ratios of phosphorylated versus total Gsk3 and were plotted (lower panel). Representative results of two independent ab double mutant KI but increased in b single mutant KI mice as experiments. compared with WT controls. The sALT levels were significantly decreased, and liver histological architecture was preserved much better with lower Suzuki scores in the protected a mutant–containing severe liver IRI with higher levels of sALT, worse damaged liver by guest on September 25, 2021 KI mice, whereas opposite changes were observed in Gsk3b histological architecture (Suzuki scores), and upregulated intra- single mutant KI mice (Fig. 2A, 2B). Interestingly, liver expres- hepatic proinflammatory gene expression as compared with their sions of inflammatory genes, including TNF-a, IL-6, IL-1b, and WT-WT chimeric controls (Fig. 3). Similar results were observed CXCL10, were all upregulated in these Gsk3 a and b mutant KI in the bone marrow chimeras of Gsk3 aS21A/bS9A double mu- mice, regardless of liver injury levels (Fig. 2C). Thus, the Gsk3 a tant KI mice (data not shown). Additionally, KC/macrophage but not b N-terminal mutation protects livers from IRI, despite depletion by clodronate-encapsulated liposomes did not abro- that both of these mutations promote liver inflammatory activa- gate the liver protective phenotype in Gsk3a mutant KI mice tion. The observation that Gsk3 a and b double mutant KI mice (Supplemental Fig. 2), indicative of a parenchyma-dependent were protected from liver IRI, similar to Gsk3 a single mutant KI mechanism. In vitro TLR4 stimulation of either bone marrow– mice, indicates that the Gsk3 aS21A mutation is dominant in derived macrophages or KCs from these mutant KI mice resulted determining the liver susceptibility to IRI. in higher levels of TNF-a and lower levels of IL-10 productions as compared with their WT counterparts (Supplemental Fig. 3). Gsk3 aS21A mutation plays distinctive roles in parenchymal These results indicate that Gsk3aS21A mutation plays distinctive versus NPCs in liver IRI roles in different types of liver cells against IR; it protects livers The paradoxical effects of Gsk3aS21A mutation on liver in- from IRI in parenchyma but enhances liver inflammatory activa- flammatory response and tissue injury suggest that its distinctive tion and exacerbates hepatocellular damage in NPCs. roles in liver parenchymal versus NPC are upon IR stress. To a address this question, we created bone marrow chimeras by Gsk3 S21A mutation protects hepatocytes from a reconstituting lethally irradiated Gsk3a single mutant KI (WT-KI) TNF- cytotoxicity or WT (KI-WT) recipient mice with either WT or Gsk3a single As TNF-a is the major inflammatory mediator of hepatocytotox- mutant KI bone marrows, respectively. Thus, Gsk3a mutation was icity in liver IRI (6–8), we studied whether Gsk3 N-terminal serine present only in the parenchyma in WT-KI chimeras or in NPCs in phosphorylation regulated hepatocyte response against TNF-a KI-WT chimeras. The liver chimeric properties in reconstituted in vitro, primary hepatocytes from WT and Gsk3a or b mutant KI mice were confirmed by Gsk3a genotyping (Supplemental Fig. 1). mice were incubated with TNF-a for 24 h in the presence of ac- Liver IRI, measured at 6 h postreperfusion, was decreased in tinomycin D, and the cytotoxicity was quantitated by the LDH WT- KI chimeras with reduced sALT levels and better-preserved assay. Significantly lower levels of cell death were detected in liver histological architectures (lower Suzuki scores) as compared Gsk3a but not b mutant KI hepatocytes as compared with those in with WT-WT chimeric mice (Fig. 3A, 3B). Liver inflammatory their WT counterparts (Fig. 4). To confirm that active Gsk3 kinase gene levels were concomitantly lower in WT-KI chimeras (Fig. 3D). was responsible for the cytoprotection in Gsk3a mutant KI he- In contrast, KI-WT chimeric mice developed significantly more patocytes, we added Gsk3 inhibitor SB216763 in cell cultures. 4 ACTIVE Gsk3a PROMOTES HEPATOCYTE AUTOPHAGY Downloaded from http://www.jimmunol.org/ FIGURE 2. Liver IRI in Gsk3aS21A and/or Gsk3bS9A single and double mutant KI mice. WT and Gsk3 mutant KI mice were subjected to liver IR experiments, as described in the Materials and Methods. Liver IRI and inflammatory responses were evaluated at both 6 and 24 h postreperfusion. (A and C) Average sALT levels and Suzuki scores of different experimental groups. (B) Representative liver histological pictures (scale bar, 200 mM). (D) Average gene expression levels (ratios of target gene/HPRT) of different experiment groups. n = 4–6 per group. Representative results of two separate experiments. *p , 0.05.

Indeed, Gsk3 inhibition increases hepatocyte cell death of the hepatocytes from inﬂammatory cell death, we added 3-MA in the mutant type and abrogates the differences between mutant KI culture media. Indeed, inhibition of autophagy increased cell and WT hepatocytes in their response to TNF-a. To test whether death of Gsk3a mutant KI but not WT hepatocytes in response to by guest on September 25, 2021 autophagy was involved in the protection of Gsk3a mutant KI TNF-a (Fig. 4). Thus, Gsk3a S21A but not b S9A mutation

FIGURE 3. Liver IRI in Gsk3aS21A mutant KI bone marrow chimeric mice. WT or Gsk3a mutant KI bone marrow chimeric mice, including WT-WT, WT-aKI, and aKI-WT, were generated and subjected to liver IR experiments, as described in the Materials and Methods. Liver injuries and inflammatory responses were evaluated at 6 h postreperfusion. (A) Average sALT levels and Suzuki scores of different experimental groups. (B) Representative liver histological pictures (scale bar, 200 mM). (C) Average levels of inflammatory cytokine/chemokine gene expressions (ratios of target gene/HPRT) in different experimental groups. n = 4–6 per group. Representative results of two separate experiments. *p , 0.05. The Journal of Immunology 5 Downloaded from FIGURE 4. TNF-a cytotoxicity in primary hepatocytes. Primary hepatocytes were isolated from WT and Gsk3a or b mutant KI mice, as described in the Materials and Methods. Cytotoxicity was induced by incubating cells with TNF-a in the presence of actinomycin D (ActD). Cell death was measured by LDH assay. Gsk3 inhibitor SB216763 or auto- FIGURE 5. Autophagy signaling pathways in primary hepatocytes in phagy inhibitor 3-MA were added prior to the addition of ActD/TNF-a. response to TNF-a. Primary hepatocytes were isolated from WT and Gsk3 Average percentages of cell death of different experimental groups were a and b mutant KI mice, as described in the Materials and Methods. Cells http://www.jimmunol.org/ plotted. Representative results of at least three experiments. *p , 0.05. were either unstimulated (2) or stimulated with TNF-a for 30 min and 2 and 6 h. Total cellular proteins were prepared and analyzed by Western blotting. Representative Western blots of phosphorylated forms and/or prevents the inactivation of Gsk3 in TNF-a–stimulated hepato- total of Gsk3a/b, TIP60, AMPK, S6K. and LC3B I/II. Representative cytes and protects them from inflammatory cell death by a pos- results of three independent experiments. sible autophagy-mediated mechanism. oxidative and endoplasmic reticulum (ER) stress in the mecha- Gsk3 aS21A mutation promotes hepatocyte autophagy in nism of hepatocellular damage in liver IRI. Thus, WT and Gsk3ab response to inflammation by guest on September 25, 2021 single or double mutant KI hepatocytes were incubated with either To further dissect the protective mechanisms of Gsk3a N-terminal Tg or H2O2 or TNF-a for 6 h, followed by the staining with serine mutation in liver parenchyma against IRI, we measured CYTO-ID and DAPI. Confocal microscopy of these stained cells hepatocyte autophagy in response to inflammation in vitro. Active showed autophagic vacuoles as green fluorescent particles sur- Gsk3 has been shown to mediate autophagy induction upon nutrient rounding blue cell nuclei (DAPI). Clearly, Gsk3 a but not b and growth factor deprivation by phosphorylation and activation of mutant KI hepatocytes had formed significantly more autophagic TIP60 (at serine 86), an acetyltransferase that activates ULK1 to vacuoles in response to TNF-a but not Tg or H2O2 (Fig. 6A). To initiate autophagy. The activation of autophagy signaling pathways determine whether TIP60 was responsible for the enhanced in primary hepatocytes upon TNF-a stimulation in vitro was an- autophagy induction in Gsk3a mutant KI hepatocytes, we added a alyzed by Western blotting. TNF-a triggered Gsk3 N-terminal chemical inhibitor of TIP60, NU9056 (31), in the cultures during phosphorylation in both a and b isoforms in WT hepatocytes TNF-a stimulation. Indeed, inhibition of TIP60 reduced numbers but not in mutant KI cells (Fig. 5). To measure Gsk3 kinase ac- of autophagic vacuoles in both the a mutant KI and WT cells and tivities, we analyzed glycogen synthase phosphorylation. The abrogated the difference between these two types of cells glycogen synthase phosphorylation was maintained at a constant (Fig. 6B). These results indicate that Gsk3a but not b N-terminal level in Gsk3a mutant KI cells but gradually decreased in WT and serine mutation facilitates autophagy induction in hepatocytes in the b mutant KI cells (Fig. 5), indicating that Gsk3a but not b response to the inflammatory stimulation. N-terminal serine mutation prevented the inactivation of Gsk3 a kinase activities in hepatocytes upon TNF-a stimulation. This Gsk3 S21A mutation promotes liver autophagy in response to resulted in significantly higher levels of LC3B II induction, TIP60 IR stress phosphorylation, AMPK phosphorylation (Thr172), and lower To determine whether Gsk3a mutation protected livers from IRI level S6K phosphorylation in Gsk3 aS21A mutant KI cells as by enhancing the autophagy induction, we analyzed autophagy- compared with those in WT and bS9A mutant KI cells (Fig. 5). related proteins in IR livers by Western blotting. Levels of auto- Autophagy flux differences in LC3B II levels in the absence and phagy biomarker LC3B II, as well as phosphorylated TIP60 and presence of chloroquine were also increased in Gsk3a mutant KI AMPKa, were significantly higher, whereas the level of phos- cells (data not shown). Thus, the N-terminal serine mutation of phorylated S6K (as a measurement of mTORC1 activities) was Gsk3a but not b enhanced the activation of autophagy signaling slightly lower in Gsk3a but not b mutant KI livers as compared pathways in hepatocytes upon TNF-a stimulation. with those in WT controls (Fig. 7). To directly visualize hepatocyte autophagy, we used a cationic To visualize hepatocyte autophagy in vivo in response to IR, we amphiphilic tracer (CYTO-ID), which specifically labeled auto- analyzed IR liver tissue sections by electron microscopy (EM). No phagic vacuoles in viable cells. We tested hepatocyte autophagy autophagosomes were detected in sham livers of both WT and induced by TNF-a, as well as H2O2 and Tg, which simulated Gsk3a KI mice. IR induced liver autophagy, which was further 6 ACTIVE Gsk3a PROMOTES HEPATOCYTE AUTOPHAGY Downloaded from http://www.jimmunol.org/

FIGURE 6. Hepatocyte autophagy induction by TNF-a, oxidative, and ER stress. Primary hepatocytes were isolated from WT and Gsk3a or b mutant KI mice. Autophagy was induced by ER stressor Tg or H2O2 or TNF-a, as described in the Materials and Methods. Autophagy was detected by labeling cells with CYTO-ID Green Detection Reagent. Cell nuclei were stained with DAPI. Labeled cells were analyzed by fluorescent microscopy. (A) Representative images of autophagic vacuoles in control and stimulated hepatocytes of WT and Gsk3a or b mutant. Mean fluorescent intensity (MFI) of different ex- by guest on September 25, 2021 perimental groups were plotted. (B) Hepatocytes were incubated with TNF-a in the absence or presence of a TIP60 inhibitor NU9056 (NU). Autophagy vacuoles and cell nuclei were detected as above. Representative images and MFI of different experimental groups were shown. Representative results of three separate experiments. Scale bar, 100 mM. *p , 0.05. enhanced by the Gsk3a mutation, as evidenced by the higher cell death and livers from IRI. The Gsk3a mutation enhanced the numbers of autophagosomes in IR livers of Gsk3a KI mutant mice activation of multiple autophagy induction pathways in hepatocytes than those of WT controls (Fig. 8). upon TNF-a stimulation that both TIP60- and AMPKa-activating To determine the functional significance of autophagy induction and TIP60 activation in Gsk3a mutant KI mice against IR, we treated the mice with 3-MA or a newly developed TIP60 inhibitor TH1834 (32, 33) prior to the onset of liver ischemia. These treatments did not affect the severity of liver IRI in WT mice (34). However, both inhibitors resulted in significant increases of liver IRI in Gsk3a mutant KI mice as evidenced by higher sALT levels and worse damaged liver histological architectures in the 3-MA– and TH1832- versus vehicle control–treated Gsk3a mutant KI mice (Fig. 9). Liver EM results showed that the 3-MA treatment diminished the numbers of autophagosomes in IR livers of Gsk3a KI mutant mice (Fig. 8). These data indicated that the Gsk3a mutation protected livers from IRI by enhancing TIP60 activation and autophagy induction in hepatocytes. Discussion By taking advantage of genetic tools, to our knowledge, we demonstrated for the first time in a clinically relevant liver sterile tissue inflammatory disease model that Gsk3 N-terminal serine FIGURE 7. Autophagy signaling pathways in liver IRI. WT and Gsk3a or phosphorylation was involved in both innate immune activation and b mutant KI mice were subjected to liver IR experiments, as described in the parenchymal cell death in an isoform- and cell type–specific man- Materials and Methods. Sham and IR (harvested at 6 h postreperfusion) ner. The N-terminal serine mutation of both Gsk3a and b enhanced livers were analyzed by Western blotting. Representative Western blots of macrophage proinflammatory activation, whereas Gsk3aS21A but phosphorylated forms of TIP60, AMPK, S6K, and total LC3B I/II and not Gsk3bS9A mutation protected hepatocytes from inflammatory b-actin. Representative results of three independent experiments. The Journal of Immunology 7

FIGURE 8. Liver autophagy induction by IR. Tissue sections were prepared from either sham or IR livers in WT or Gsk3a mutant KI mice treated with or without 3-MA, as described in the Mate- rials and Methods. Autophagosomes were detected by EM. Representative images of each experimental group were shown. Scale bar, 20 mM.

phosphorylation were increased with a simultaneous decrease lack of Gsk3 isoform specificities in vivo. We have tested a PI3 in the mTOR1 activity. Inhibition of autophagy or TIP60 re- kinase inhibitor, which was able to partially reduce Gsk3b S9 Downloaded from versed the protective phenotype in vitro and in vivo in Gsk3a phosphorylation in livers upon IR and showed that it did result in mutant KI cohorts. These results show that Gsk3 N-terminal the exaggeration of liver IRI but only under the mild ischemic serine phosphorylation plays active and important roles in the condition (24). In this study, we confirm that the N-terminal serine disease pathogenesis. phosphorylation of both Gsk3 isoforms inhibits innate immune Unlike MAPKs, Gsk3 is constitutively active in cells and activation. Interestingly, although local inflammatory activation a inhibited upon inflammatory/stress stimulation via the N-terminal by IR is enhanced in Gsk3 S21A mutant KI mice, livers are http://www.jimmunol.org/ serine phosphorylation. Multiple kinase/phosphatase pathways are protected against IRI because of the protection of mutant hepa- involved in the regulation of Gsk3 serine phosphorylation (35, 36). tocytes from inflammatory cell death. It suggests that the hepa- The PI3K-Akt–mediated Gsk3b S9 phosphorylation has been tocyte phenotype is dominant in determining the outcome of the shown to inhibit proinflammatory activation (24, 26, 37). However, inflammation. it has been difficult to precisely address the functional question of Gsk3 aS21A/bS9A double mutant KI mice were initially cre- N-terminal serine phosphorylation of Gsk3 by using chemical in- ated by McManus et al. (11) to study insulin and Wnt signaling, hibitors of its upstream kinases because of the off-target effects and and they have normal development and growth without signs of by guest on September 25, 2021

FIGURE 9. Inhibition of autophagy or TIP60 increases liver IRI in Gsk3a mutant KI but not WT mice. WT and Gsk3a mutant KI mice were pretreated with 3-MA or TIP60 inhibitor TH1834 (TIP60i) prior to the onset of liver IR, as described in the Materials and Methods. Liver IRI was evaluated at 6 h postreperfusion. (A and C) Average sALT levels and Suzuki scores and (B and D) representative liver histological pictures of different experimental groups were shown. Representative results of at least two experiments. Scale bar, 200 mM. *p , 0.05. 8 ACTIVE Gsk3a PROMOTES HEPATOCYTE AUTOPHAGY

Gsk3 inhibitors. Gsk3 has been shown to promote cell death caused by the mitochondrial intrinsic apoptotic pathway but inhibit the death receptor–mediated extrinsic apoptotic signaling pathway (9, 20, 21). Both these pathways are involved in the hepatocellular damage mechanism post-IR. Our data shows that active Gsk3a increase hepatocyte autophagy in response to TNF-a but not oxidative/ER stress. An obvious question for us to answer next is why the cyto- protective effect of N-terminal serine mutation was only associated with Gsk3a but not b isoform. We tested GS phosphorylation in these two types of mutant cells as an indication of Gsk3 kinase activities. Results show that Gsk3a, but not b, mutation sustained the kinase activity in hepatocytes upon TNF-a stimulation, indicating that Gsk3b was inactivated independent of S9 phosphorylation. Indeed, multiple types of posttranslational modiﬁcations have been identiﬁed to regulate Gsk3 b but not a activities (42), including inhibitory serine 389 phosphorylation (43) and ADP ribosylation (44). Additionally, GSK-3b can be citrullinated

within its N-terminal domain by protein arginine deiminase 4 to Downloaded from promote its nuclear accumulation (45) and ubiquitinated at lysine 63 by the E3 ligase TNFR-associated factor 6 for its assembly with TLR3 (46). Thus, the N-terminal serine mutation is only sufﬁcient for Gsk3a but not Gsk3b to avoid inactivation in TNF-a–stimulated hepatocytes.

The role of autophagy in IRI is both complex and controversial http://www.jimmunol.org/ (47, 48). Our own study in the same murine liver IRI model has shown that mild ischemia (30 min) triggers autophagy to protect livers from IRI, whereas prolonged ischemia (90 min) inhibits FIGURE 10. Gsk3 regulates multiple intracellular signaling pathways in autophagy to facilitate liver IRI (34). Our current results reveal a hepatocytes in response to IR. Hypoxia and inflammatory stimuli trigger novel regulatory pathway in hepatocytes in response to TNF-a the N-terminal inhibitory phosphorylation of Gsk3a/b during liver IR, that Gsk3 inhibitory phosphorylation downregulate autophagy, in leading to decreased activation of TIP60 (S86) and AMPKa (T172), in- part, by inactivating the TIP60–ULK1 pathway. Our results also creased activation of mTORC1, and possibly higher levels of HIF-1a. reveal the potential involvement of AMPK and mTORC1 in These pathways are involved in autophagy and pathogenesis of liver IRI. protecting hepatocytes downstream of active Gsk3a (Fig. 10). It by guest on September 25, 2021 Gsk3a S21A mutation spares the kinase from this inhibitory phos- has been found in Gsk3a KO mice that mTORC1 activities are phorylation and sustains its kinase activity, leading to higher TIP60, AMPK, and lower mTORC1 activation. Gsk3b S9A mutation, in con- significantly increased, leading to impaired autophagy in the heart trast, is not sufficient to sustain its kinase activity in stimulation cells and fibroblasts (49). because of other inactivation mechanisms (e.g., S398 phosphorylation, Gsk3 can phosphorylate hypoxia-inducible factor (HIF) 1a di- ADP ribosylation, etc.). rectly (50), leading to its ubiquitinylation and degradation via F-box and WD protein Fbw7, independent of von Hippel–Lindau (51). Thus, Gsk3 N-terminal mutation may result in lower levels metabolic disorder and insulin resistance. In the myocardial dis- of HIF-1a in response to IR. It will be interesting whether HIF-1a ease model, Gsk3ab double mutant KI has been shown to protect or its other isoforms are involved downstream of Gsk3 a or b to against pathological hypertrophy induced by chronic adrenergic regulate liver IRI (Fig. 10). HIFs are accumulated/upregulated in stimulation, maintaining cardiac function and attenuating inter- response to not only hypoxia but also inflammatory stimuli, such stitial fibrosis (38). However, its effects on chronic pathological as LPS and TNF-a (52–54). Although they are generally thought hypertrophy after regional infarction was marginal, suggesting its to promote autophagy/survival in cells in their adaptation to ox- disease context-specific roles (39). Gsk3bS9A single mutant KI ygen and nutrient deprivation, the in vivo effect of HIF inhibition mice exhibited attenuated, although Gsk3aS21A single mutant KI or genetic inactivation on the disease process is context and iso- mice showed exacerbated, hypertrophy and heart failure in re- form dependent. In a HepG2 tumor spheroid model, knock-down sponse to pressure overload (28). In cardiomyocyte-specific and of one HIF isoform (1a or 2a) resulted in the upregulation of the -inducible Gsk3b KO mice, Gsk3b regulated post–myocardial other, leading to higher levels of autophagy and survival advan- injury remodeling but played a minor role in the hypertrophic tages in tumor cells (55). Although the genetic KO of prolyl hy- response to pressure overload (18). Thus, Gsk3 regulates myo- droxylase domain enzyme 1, which regulates the stability of HIFs, cardial injury and remodeling in isoform-specific, molecular protects livers from IRI in vivo and hepatocytes from simulated IR form–specific (as active kinase or via its inhibitory phosphoryla- in vitro, knock-down of HIF-1a or 2a fails to increase hepatocyte tion process) and disease context–dependent manner. death against IR (56). Additionally, HIF has been identified as a Gsk3 inhibitors have been shown to alleviate tissue damages in in key regulator of macrophage activation/functions in tissue inflam- the models of myocardial infarction, stroke, and liver IRI (15, 40, mation via its ability to metabolically reprogram the cells (53, 57). 41). We have shown previously, in the same murine liver partial HIF isoforms may play distinctive roles in either facilitating or reg- warm ischemia model, that Gsk3 inhibitor SB216763 protected ulating macrophage proinflammatory responses. One recent study livers from inflammatory tissue injury via an IL-10–mediated documents that HIF-2a protects acute liver injury by promoting IL-6 immune regulatory mechanism (24). Results in this study raise an production in macrophages (58). Clearly, multiple immune regulatory obvious question regarding the potential detrimental effect of genes regulated by HIFs, such as ecto-59-nucleotidase (CD73) (59), The Journal of Immunology 9 can be involved in the regulation of tissue inflammatory response 20. Juhaszova, M., D. B. Zorov, S. H. Kim, S. Pepe, Q. Fu, K. W. Fishbein, further downstream of the Gsk3–HIF pathway. B. D. Ziman, S. Wang, K. Ytrehus, C. L. Antos, et al. 2004. Glycogen synthase kinase-3beta mediates convergence of protection signaling to inhibit the mito- In summary, our results demonstrate in vivo the functional chondrial permeability transition pore. J. Clin. Invest. 113: 1535–1549. significance of Gsk3a and b N-terminal serine phosphorylation in 21. Juhaszova, M., D. B. Zorov, Y. Yaniv, H. B. Nuss, S. Wang, and S. J. Sollott. 2009. Role of glycogen synthase kinase-3beta in cardioprotection. Circ. Res. sterile inflammatory tissue injury. The novel regulatory role of 104: 1240–1252. active Gsk3 in hepatocyte autophagy in response to inflammatory 22. Lal, H., F. Ahmad, J. Woodgett, and T. Force. 2015. The GSK-3 family as stimuli sheds a new light on the disease mechanism that Gsk3 therapeutic target for myocardial diseases. Circ. Res. 116: 138–149. 23. Vigneron, F., P. Dos Santos, S. Lemoine, M. Bonnet, L. Tariosse, T. Couffinhal, N-terminal phosphorylation contributes to liver IRI by inhibiting C. Duplaa`, and B. Jaspard-Vinassa. 2011. GSK-3b at the crossroads in the TIP60-initiated hepatocyte autophagy. Thus, Gsk3 functions both signalling of heart preconditioning: implication of mTOR and Wnt pathways. as an active kinase and via its inhibitory phosphorylation in the Cardiovasc. Res. 90: 49–56. 24. Ren, F., Z. Duan, Q. Cheng, X. Shen, F. Gao, L. Bai, J. Liu, R. W. Busuttil, pathogenesis of liver IRI. We must take into consideration its J. W. Kupiec-Weglinski, and Y. Zhai. 2011. Inhibition of glycogen synthase kinase complex roles in the isoform- and cell type–specific manner to 3 beta ameliorates liver ischemia reperfusion injury by way of an interleukin- design safe and effective Gsk3-targeted therapy. 10-mediated immune regulatory mechanism. Hepatology 54: 687–696. 25. Yue, S., J. Rao, J. Zhu, R. W. Busuttil, J. W. Kupiec-Weglinski, L. Lu, X. Wang, and Y. Zhai. 2014. Myeloid PTEN deficiency protects livers from ischemia Disclosures reperfusion injury by facilitating M2 macrophage differentiation. J. Immunol. 192: 5343–5353. The authors have no financial conflicts of interest. 26. Zhou, H., H. Wang, M. Ni, S. Yue, Y. Xia, R. W. Busuttil, J. W. Kupiec- Weglinski, L. Lu, X. Wang, and Y. Zhai. 2018. Glycogen synthase kinase 3b promotes liver innate immune activation by restraining AMP-activated protein

References kinase activation. J. Hepatol. 69: 99–109. Downloaded from 1. Kaczorowski, D. J., A. Tsung, and T. R. Billiar. 2009. Innate immune mecha- 27. Antos, C. L., T. A. McKinsey, N. Frey, W. Kutschke, J. McAnally, J. M. Shelton, nisms in ischemia/reperfusion. Front. Biosci. (Elite Ed.) 1: 91–98. J. A. Richardson, J. A. Hill, and E. N. Olson. 2002. Activated glycogen synthase- 2. Zhai, Y., R. W. Busuttil, and J. W. Kupiec-Weglinski. 2011. Liver ischemia and 3 beta suppresses cardiac hypertrophy in vivo. Proc. Natl. Acad. Sci. USA 99: reperfusion injury: new insights into mechanisms of innate-adaptive immune- 907–912. mediated tissue inﬂammation. Am. J. Transplant. 11: 1563–1569. 28. Matsuda, T., P. Zhai, Y. Maejima, C. Hong, S. Gao, B. Tian, K. Goto, H. Takagi, 3. Tsung, A., R. A. Hoffman, K. Izuishi, N. D. Critchlow, A. Nakao, M. H. Chan, M. Tamamori-Adachi, S. Kitajima, and J. Sadoshima. 2008. Distinct roles of M. T. Lotze, D. A. Geller, and T. R. Billiar. 2005. Hepatic ischemia/reperfusion GSK-3alpha and GSK-3beta phosphorylation in the heart under pressure over- J. Immunol. load. Proc. Natl. Acad. Sci. USA 105: 20900–20905.

injury involves functional TLR4 signaling in nonparenchymal cells. http://www.jimmunol.org/ 175: 7661–7668. 29. Shen, X. D., B. Ke, Y. Zhai, F. Amersi, F. Gao, D. M. Anselmo, R. W. Busuttil, 4. Zhai, Y., X. D. Shen, R. O’Connell, F. Gao, C. Lassman, R. W. Busuttil, and J. W. Kupiec-Weglinski. 2002. CD154-CD40 T-cell costimulation pathway G. Cheng, and J. W. Kupiec-Weglinski. 2004. Cutting edge: TLR4 activation is required in the mechanism of hepatic ischemia/reperfusion injury, and its mediates liver ischemia/reperfusion inﬂammatory response via IFN regulatory blockade facilitates and depends on heme oxygenase-1 mediated cytoprotection. factor 3-dependent MyD88-independent pathway. J. Immunol. 173: 7115–7119. Transplantation 74: 315–319. 5. Wu, H. S., J. X. Zhang, L. Wang, Y. Tian, H. Wang, and O. Rotstein. 2004. Toll- 30. Zhai, Y., X. D. Shen, F. Gao, A. Zhao, M. C. Freitas, C. Lassman, A. D. Luster, like receptor 4 involvement in hepatic ischemia/reperfusion injury in mice. R. W. Busuttil, and J. W. Kupiec-Weglinski. 2008. CXCL10 regulates liver in- HBPD INT 3: 250–253. nate immune response against ischemia and reperfusion injury. Hepatology 47: 6. Colletti, L. M., D. G. Remick, G. D. Burtch, S. L. Kunkel, R. M. Strieter, and 207–214. D. A. Campbell, Jr. 1990. Role of tumor necrosis factor-alpha in the pathophysi- 31. Coffey, K., T. J. Blackburn, S. Cook, B. T. Golding, R. J. Grifﬁn, ologic alterations after hepatic ischemia/reperfusion injury in the rat. J. Clin. Invest. I. R. Hardcastle, L. Hewitt, K. Huberman, H. V. McNeill, D. R. Newell, et al.

85: 1936–1943. 2012. Characterisation of a Tip60 specific inhibitor, NU9056, in prostate cancer. by guest on September 25, 2021 7. Teoh, N., J. Field, J. Sutton, and G. Farrell. 2004. Dual role of tumor necrosis PLoS One 7: e45539. factor-alpha in hepatic ischemia-reperfusion injury: studies in tumor necrosis 32. Idrissou, M., G. Judes, M. Daures, A. Sanchez, D. El Ouardi, S. Besse, factor-alpha gene knockout mice. Hepatology 39: 412–421. F. Degoul, F. Penault-Llorca, Y. J. Bignon, and D. Bernard-Gallon. 2019. TIP60 8. Ru¨diger, H. A., and P. A. Clavien. 2002. Tumor necrosis factor alpha, but not inhibitor TH1834 reduces breast cancer progression in xenografts in mice. Fas, mediates hepatocellular apoptosis in the murine ischemic liver. Gastroen- OMICS 23: 457–459. terology 122: 202–210. 33. Gao, C., E. Bourke, M. Scobie, M. A. Famme, T. Koolmeister, T. Helleday, 9. Beurel, E., and R. S. Jope. 2006. The paradoxical pro- and anti-apoptotic actions L. A. Eriksson, N. F. Lowndes, and J. A. L. Brown. 2014. Rational design and of GSK3 in the intrinsic and extrinsic apoptosis signaling pathways. Prog. validation of a Tip60 histone acetyltransferase inhibitor. Sci. Rep. 4: 5372. Neurobiol. 79: 173–189. 34. Zhou, H., J. Zhu, S. Yue, L. Lu, R. W. Busuttil, J. W. Kupiec-Weglinski, 10. Beurel, E., S. M. Michalek, and R. S. Jope. 2010. Innate and adaptive immune X. Wang, and Y. Zhai. 2016. The dichotomy of endoplasmic reticulum stress responses regulated by glycogen synthase kinase-3 (GSK3). Trends Immunol. 31: response in liver ischemia-reperfusion injury. Transplantation 100: 365–372. 24–31. 35. Kaidanovich-Beilin, O., and J. R. Woodgett. 2011. GSK-3: functional insights 11. McManus, E. J., K. Sakamoto, L. J. Armit, L. Ronaldson, N. Shpiro, R. Marquez, from cell biology and animal models. Front. Mol. Neurosci. 4: 40. and D. R. Alessi. 2005. Role that phosphorylation of GSK3 plays in insulin and 36. Beurel, E., S. F. Grieco, and R. S. Jope. 2015. Glycogen synthase kinase-3 Wnt signalling defined by knockin analysis. EMBO J. 24: 1571–1583. (GSK3): regulation, actions, and diseases. Pharmacol. Ther. 148: 114–131. 12. Cohen, P., and M. Goedert. 2004. GSK3 inhibitors: development and therapeutic 37. Martin, M., K. Rehani, R. S. Jope, and S. M. Michalek. 2005. Toll-like receptor- potential. Nat. Rev. Drug Discov. 3: 479–487. mediated cytokine production is differentially regulated by glycogen synthase 13. Hoeflich, K. P., J. Luo, E. A. Rubie, M. S. Tsao, O. Jin, and J. R. Woodgett. 2000. kinase 3. Nat. Immunol. 6: 777–784. Requirement for glycogen synthase kinase-3beta in cell survival and NF-kappaB 38. Webb, I. G., Y. Nishino, J. E. Clark, C. Murdoch, S. J. Walker, M. R. Makowski, activation. Nature 406: 86–90. R. M. Botnar, S. R. Redwood, A. M. Shah, and M. S. Marber. 2010. Constitutive 14. Schwabe, R. F., and D. A. Brenner. 2002. Role of glycogen synthase kinase-3 glycogen synthase kinase-3alpha/beta activity protects against chronic beta- in TNF-alpha-induced NF-kappaB activation and apoptosis in hepatocytes. Am. adrenergic remodelling of the heart. Cardiovasc. Res. 87: 494–503. J. Physiol. Gastrointest. Liver Physiol. 283: G204–G211. 39. Webb, I. G., P. Sicard, J. E. Clark, S. Redwood, and M. S. Marber. 2010. 15. Gomez, L., M. Paillard, H. Thibault, G. Derumeaux, and M. Ovize. 2008. In- Myocardial stress remodelling after regional infarction is independent of gly- hibition of GSK3beta by postconditioning is required to prevent opening of the cogen synthase kinase-3 inactivation. J. Mol. Cell. Cardiol. 49: 897–900. mitochondrial permeability transition pore during reperfusion. Circulation 117: 40. Ren, M., V. V. Senatorov, R. W. Chen, and D. M. Chuang. 2003. Postinsult 2761–2768. treatment with lithium reduces brain damage and facilitates neurological re- 16. Avrahami, L., A. Licht-Murava, M. Eisenstein, and H. Eldar-Finkelman. 2013. covery in a rat ischemia/reperfusion model. Proc. Natl. Acad. Sci. USA 100: GSK-3 inhibition: achieving moderate efficacy with high selectivity. Biochim. 6210–6215. Biophys. Acta 1834: 1410–1414. 41. Fu, H., H. Xu, H. Chen, Y. Li, W. Li, Q. Zhu, Q. Zhang, H. Yuan, F. Liu, 17. MacAulay, K., B. W. Doble, S. Patel, T. Hansotia, E. M. Sinclair, D. J. Drucker, Q. Wang, et al. 2014. Inhibition of glycogen synthase kinase 3 ameliorates liver A. Nagy, and J. R. Woodgett. 2007. Glycogen synthase kinase 3alpha-specific ischemia/reperfusion injury via an energy-dependent mitochondrial mechanism. regulation of murine hepatic glycogen metabolism. Cell Metab. 6: 329–337. J. Hepatol. 61: 816–824. 18. Woulfe, K. C., E. Gao, H. Lal, D. Harris, Q. Fan, R. Vagnozzi, M. DeCaul, 42. Cormier, K. W., and J. R. Woodgett. 2017. Recent advances in understanding the X. Shang, S. Patel, J. R. Woodgett, et al. 2010. Glycogen synthase kinase-3beta cellular roles of GSK-3. F1000Res. 6: F1000 Faculty Rev-167. regulates post-myocardial infarction remodeling and stress-induced car- 43. Thornton, T. M., P. Delgado, L. Chen, B. Salas, D. Krementsov, M. Fernandez, diomyocyte proliferation in vivo. Circ. Res. 106: 1635–1645. S. Vernia, R. J. Davis, R. Heimann, C. Teuscher, et al. 2016. Inactivation of 19. Lal, H., J. Zhou, F. Ahmad, R. Zaka, R. J. Vagnozzi, M. Decaul, J. Woodgett, nuclear GSK3b by Ser(389) phosphorylation promotes lymphocyte fitness dur- E. Gao, and T. Force. 2012. Glycogen synthase kinase-3a limits ischemic injury, ing DNA double-strand break response. Nat. Commun. 7: 10553. cardiac rupture, post-myocardial infarction remodeling and death. Circulation 44. Feijs, K. L., H. Kleine, A. Braczynski, A. H. Forst, N. Herzog, P. Verheugd, 125: 65–75. U. Linzen, E. Kremmer, and B. Lu¨scher. 2013. ARTD10 substrate identification 10 ACTIVE Gsk3a PROMOTES HEPATOCYTE AUTOPHAGY

on protein microarrays: regulation of GSK3b by mono-ADP-ribosylation. Cell 52. Lee, P., N. S. Chandel, and M. C. Simon. 2020. Cellular adaptation to hypoxia Commun. Signal. 11: 5. through hypoxia inducible factors and beyond. Nat. Rev. Mol. Cell Biol. 21: 45. Stadler, S. C., C. T. Vincent, V. D. Fedorov, A. Patsialou, B. D. Cherrington, 268–283. J. J. Wakshlag, S. Mohanan, B. M. Zee, X. Zhang, B. A. Garcia, et al. 2013. 53. Eltzschig, H. K., D. L. Bratton, and S. P. Colgan. 2014. Targeting hypoxia Dysregulation of PAD4-mediated citrullination of nuclear GSK3b activates signalling for the treatment of ischaemic and inflammatory diseases. Nat. Rev. Drug Discov. 13: 852–869. TGF-b signaling and induces epithelial-to-mesenchymal transition in breast 54. Hellwig-Bu¨rgel, T., K. Rutkowski, E. Metzen, J. Fandrey, and W. Jelkmann. cancer cells. [Published erratum appears in 2013 Proc. Natl. Acad. Sci. USA 110: 1999. Interleukin-1beta and tumor necrosis factor-alpha stimulate DNA binding 16283.] Proc. Natl. Acad. Sci. USA 110: 11851–11856. of hypoxia-inducible factor-1. Blood 94: 1561–1567. 46. Ko, R., J. H. Park, H. Ha, Y. Choi, and S. Y. Lee. 2015. Glycogen synthase kinase 55. Menrad, H., C. Werno, T. Schmid, E. Copanaki, T. Deller, N. Dehne, and 3b ubiquitination by TRAF6 regulates TLR3-mediated pro-inflammatory cyto- B. Bru¨ne. 2010. Roles of hypoxia-inducible factor-1alpha (HIF-1alpha) versus kine production. Nat. Commun. 6: 6765. HIF-2alpha in the survival of hepatocellular tumor spheroids. Hepatology 51: 47. Przyklenk, K., Y. Dong, V. V. Undyala, and P. Whittaker. 2012. Autophagy as a 2183–2192. therapeutic target for ischaemia /reperfusion injury? Concepts, controversies, 56. Schneider, M., K. Van Geyte, P. Fraisl, J. Kiss, J. Aragone´s, M. Mazzone, and challenges. Cardiovasc. Res. 94: 197–205. H. Mairbaürl, K. De Bock, N. H. Jeoung, M. Mollenhauer, et al. 2010. Loss or 48. Gottlieb, R. A., and R. M. Mentzer. 2010. Autophagy during cardiac stress: joys silencing of the PHD1 prolyl hydroxylase protects livers of mice against and frustrations of autophagy. Annu. Rev. Physiol. 72: 45–59. ischemia/reperfusion injury. Gastroenterology 138: 1143–1154.e1–e2. 49. Zhou, J., T. A. Freeman, F. Ahmad, X. Shang, E. Mangano, E. Gao, J. Farber, 57. Lin, N., and M. C. Simon. 2016. Hypoxia-inducible factors: key regulators of Y. Wang, X. L. Ma, J. Woodgett, et al. 2013. GSK-3a is a central regulator of myeloid cells during inflammation. J. Clin. Invest. 126: 3661–3671. 58. Gao, R. Y., M. Wang, Q. Liu, D. Feng, Y. Wen, Y. Xia, S. P. Colgan, age-related pathologies in mice. J. Clin. Invest. 123: 1821–1832. H. K. Eltzschig, and C. Ju. 2020. Hypoxia-inducible factor-2a reprograms liver 50. Flugel, D., A. Go¨rlach, C. Michiels, and T. Kietzmann. 2007. Glycogen synthase ¨ macrophages to protect against acute liver injury through the production of in- kinase 3 phosphorylates hypoxia-inducible factor 1alpha and mediates its de- terleukin-6. Hepatology 71: 2105–2117. stabilization in a VHL-independent manner. Mol. Cell. Biol. 27: 3253–3265. 59. Hart, M. L., C. Much, I. C. Gorzolla, J. Schittenhelm, D. Kloor, G. L. Stahl, and 51. Flu¨gel, D., A. Go¨rlach, and T. Kietzmann. 2012. GSK-3b regulates cell growth, H. K. Eltzschig. 2008. Extracellular adenosine production by ecto-59-nucleotidase migration, and angiogenesis via Fbw7 and USP28-dependent degradation of protects during murine hepatic ischemic preconditioning. Gastroenterology 135: Downloaded from HIF-1a. Blood 119: 1292–1301. 1739–1750.e3. http://www.jimmunol.org/ by guest on September 25, 2021 Supplements

Figure 1: Gsk3 genotyping of chimeric mice. Liver NPCs were isolated from WT or Gsk3a mutant KI, or WT-KI or KI-WT chimeric mice. Genomic DNA was isolated. WT and mutant Gsk3a fragments were analyzed by PCR.

Figure 2: Gsk3  mutant KI protects livers from IRI in the absence of KCs. WT and mutant KI mice were treated with either blank- (Ctl) or clodronate-liposomes prior to the onset of liver ischemia, as described in the Materials and Methods. Liver IRI was evaluated at 6h post reperfusion. (a) Average sALT levels, (b) Representative liver histology pictures, (c) Liver inflammatory gene expressions (qRT-PCR), of different experimental groups were plotted/shown. n=3-4/group. Representative results of at least 2 experiments. *p<0.05

Fig.3. Macrophage responses in vitro. BMMs and KCs were isolated from either WT or Gsk3  or  mutant KI mice and stimulated in vitro with PBS (-) or LPS (100ng/ml) for 6h and 24h. Cytokine levels in culture supernatants were measured by ELISA. n=3/group. *p<0.05