Advertisement

Impaired Fasting-Induced Adaptive Lipid Droplet Biogenesis in Liver-Specific Atg5-Deficient Mouse Liver Is Mediated by Persistent Nuclear Factor-Like 2 Activation

  • Yuan Li
    Affiliations
    Department of Pharmacology, Toxicology and Therapeutics, The University of Kansas Medical Center, Kansas City, Kansas
    Search for articles by this author
  • Xiaojuan Chao
    Affiliations
    Department of Pharmacology, Toxicology and Therapeutics, The University of Kansas Medical Center, Kansas City, Kansas
    Search for articles by this author
  • Li Yang
    Affiliations
    Institute of Chinese Materia Medica, Shanghai University of Traditional Chinese Medicine, Shanghai, China
    Search for articles by this author
  • Qian Lu
    Affiliations
    Institute of Chinese Materia Medica, Shanghai University of Traditional Chinese Medicine, Shanghai, China
    Search for articles by this author
  • Tiangang Li
    Affiliations
    Department of Pharmacology, Toxicology and Therapeutics, The University of Kansas Medical Center, Kansas City, Kansas
    Search for articles by this author
  • Wen-Xing Ding
    Correspondence
    Address correspondence to Hong-Min Ni, M.D. or Wen-Xing Ding, Ph.D., Department of Pharmacology, Toxicology and Therapeutics, The University of Kansas Medical Center, MS 1018, 3901 Rainbow Blvd., Kansas City, KS 66160.
    Affiliations
    Department of Pharmacology, Toxicology and Therapeutics, The University of Kansas Medical Center, Kansas City, Kansas
    Search for articles by this author
  • Hong-Min Ni
    Correspondence
    Address correspondence to Hong-Min Ni, M.D. or Wen-Xing Ding, Ph.D., Department of Pharmacology, Toxicology and Therapeutics, The University of Kansas Medical Center, MS 1018, 3901 Rainbow Blvd., Kansas City, KS 66160.
    Affiliations
    Department of Pharmacology, Toxicology and Therapeutics, The University of Kansas Medical Center, Kansas City, Kansas
    Search for articles by this author
Open ArchivePublished:May 24, 2018DOI:https://doi.org/10.1016/j.ajpath.2018.04.015
      Lipid droplets (LDs) are intracellular organelles that store neutral lipids as energy reservoir. Recent studies suggest that autophagy is important in maintaining the homeostasis of intracellular LDs by either regulating the biogenesis of LDs, mobilization of fatty acids, or degradation of LDs in cultured cells. Increasing evidence also supports a role of autophagy in regulating glucose and lipid metabolism in vivo in mammals. In response to fasting/starvation, lipids are mobilized from the adipose tissue to the liver, which increases the number of intracellular LDs and stimulates fatty acid oxidation and ketogenesis. However, it is still controversial and unclear how impaired autophagy in hepatocytes affects the biogenesis of LDs in mouse livers. In the present study, it was demonstrated that hepatic autophagy-deficient (L-Atg)5 knockout mice had impaired adaptation to fasting-induced hepatic biogenesis of LDs. The maladaptation to fasting-induced hepatic biogenesis of LDs in L-Atg5 knockout mouse livers was not due to hepatic changes of de novo lipogenesis, secretion of very-low-density lipoprotein or fatty acid β-oxidation, but it was due to persistent nuclear factor-like 2 activation because biogenesis of LDs restored in L-Atg5/nuclear factor-like 2 double-knockout mice.
      Lipid droplets (LDs) are intracellular organelles that primarily store neutral lipids such as triglycerides (TGs) and sterol esters within cells. It is generally thought that LDs are originated from the endoplasmic reticulum, which bud off into the cytoplasm to form the LDs with the limiting phospholipid monolayer.
      • Pol A.
      • Gross S.P.
      • Parton R.G.
      Review: biogenesis of the multifunctional lipid droplet: lipids, proteins, and sites.
      • Yang H.
      • Galea A.
      • Sytnyk V.
      • Crossley M.
      Controlling the size of lipid droplets: lipid and protein factors.
      TGs are synthesized from free fatty acids (FFAs) and glycerol-3-phosphate by several key enzymes in the endoplasmic reticulum. First, glycerol-3-phosphate is acylated by glycerol-3-phosphate acyltransferases to generate lysophosphatidic acid. Second, lysophosphatidic acid is then converted to phosphatidic acid by adding an additional acyl chain via family members of 1-acylglycerol-3-phosphate acyltransferases. Third, diacylglycerol is synthesized from phosphatidic acid by the lipin proteins that have phosphatidate phosphatase activity in mammals. Finally, diacylglycerol O-acyltransferase 1 and 2 mediate the final step of esterifying diacylglycerol to produce TGs, which is then packed into LDs.
      • Pol A.
      • Gross S.P.
      • Parton R.G.
      Review: biogenesis of the multifunctional lipid droplet: lipids, proteins, and sites.
      • Yang H.
      • Galea A.
      • Sytnyk V.
      • Crossley M.
      Controlling the size of lipid droplets: lipid and protein factors.
      The formation of LDs serves as a reservoir of energy for the excess FFAs because TGs can be released when nutrient is scarce. Moreover, the formation of LDs can also help to protect cells against the cytotoxicity of FFAs, which is termed lipotoxicity.
      • Zechner R.
      • Madeo F.
      • Kratky D.
      Cytosolic lipolysis and lipophagy: two sides of the same coin.
      • Listenberger L.L.
      • Han X.
      • Lewis S.E.
      • Cases S.
      • Farese Jr., R.V.
      • Ory D.S.
      • Schaffer J.E.
      Triglyceride accumulation protects against fatty acid-induced lipotoxicity.
      Humans and animals have developed sophisticated adaptive mechanisms in response to the fluctuation of nutrient availability through evolution. In response to decreased energy and nutrient intake, the concentrations of circulating glucose and insulin are low, which stimulates the lipolysis of LDs at the adipose tissue.
      • Zechner R.
      • Strauss J.G.
      • Haemmerle G.
      • Lass A.
      • Zimmermann R.
      Lipolysis: pathway under construction.
      In the fasted state, the levels of catecholamines and cellular cAMP increase in adipose tissue that activate protein kinase A. Activated protein kinase A then phosphorylates a number of LD-associated proteins, including perilipin 1, hormone sensitive lipase, and comparative gene identification-58. Comparative gene identification-58 binds with patatin-like phospholipase domain containing 2/adipose TG lipase, which increases the catalytic activity of patatin-like phospholipase domain containing 2/adipose TG lipase. The stored TGs in LDs are sequentially hydrolyzed into diacylglycerol and monoacylglycerol and eventually into nonesterified FFAs and glycerol by patatin-like phospholipase domain containing 2/adipose TG lipase, hormone sensitive lipase, and monoacylglycerol lipase.
      • Zechner R.
      • Madeo F.
      • Kratky D.
      Cytosolic lipolysis and lipophagy: two sides of the same coin.
      Increased adipose tissue lipolysis then mobilizes FFAs to the liver, resulting in the increased hepatic accumulation of lipid, whereby they are oxidized to generate acetyl-CoA and ketone bodies to produce energy.
      • Badman M.K.
      • Pissios P.
      • Kennedy A.R.
      • Koukos G.
      • Flier J.S.
      • Maratos-Flier E.
      Hepatic fibroblast growth factor 21 is regulated by PPARalpha and is a key mediator of hepatic lipid metabolism in ketotic states.
      • Inagaki T.
      • Dutchak P.
      • Zhao G.
      • Ding X.
      • Gautron L.
      • Parameswara V.
      • Li Y.
      • Goetz R.
      • Mohammadi M.
      • Esser V.
      • Elmquist J.K.
      • Gerard R.D.
      • Burgess S.C.
      • Hammer R.E.
      • Mangelsdorf D.J.
      • Kliewer S.A.
      Endocrine regulation of the fasting response by PPARalpha-mediated induction of fibroblast growth factor 21.
      In addition to the neutral lipase-mediated hydrolysis of TGs in LDs, macroautophagy (hereafter referred to as autophagy) mediates lipid transport from LDs to the lysosomes for lipolysis.
      • Zechner R.
      • Madeo F.
      • Kratky D.
      Cytosolic lipolysis and lipophagy: two sides of the same coin.
      • Mizushima N.
      • Levine B.
      • Cuervo A.M.
      • Klionsky D.J.
      Autophagy fights disease through cellular self-digestion.
      • Madrigal-Matute J.
      • Cuervo A.M.
      Regulation of liver metabolism by autophagy.
      Autophagy is an intracellular degradation system via the formation of autophagosome and the later autolysosomes. Autophagy degrades long-lived proteins, organelles, and other cellular contents to provide nutrients for cell survival under starvation conditions.
      • Levine B.
      • Klionsky D.J.
      Development by self-digestion: molecular mechanisms and biological functions of autophagy.
      • Lum J.J.
      • DeBerardinis R.J.
      • Thompson C.B.
      Autophagy in metazoans: cell survival in the land of plenty.
      Recent evidence suggests that autophagy-related (Atg) proteins are associated with LDs and autophagy may also degrade LDs or at least a portion of LDs, a process that is termed as lipophagy.
      • Singh R.
      • Kaushik S.
      • Wang Y.
      • Xiang Y.
      • Novak I.
      • Komatsu M.
      • Tanaka K.
      • Cuervo A.M.
      • Czaja M.J.
      Autophagy regulates lipid metabolism.
      Impaired hepatic autophagy has been reported in nonalcoholic fatty liver disease and type 2 diabetes.
      • Yang L.
      • Li P.
      • Fu S.
      • Calay E.S.
      • Hotamisligil G.S.
      Defective hepatic autophagy in obesity promotes ER stress and causes insulin resistance.
      In addition, pharmacologic modulation of autophagy pathway improves both nonalcoholic and alcoholic steatosis in mice.
      • Czaja M.J.
      • Ding W.X.
      • Donohue Jr., T.M.
      • Friedman S.L.
      • Kim J.S.
      • Komatsu M.
      • Lemasters J.J.
      • Lemoine A.
      • Lin J.D.
      • Ou J.H.
      • Perlmutter D.H.
      • Randall G.
      • Ray R.B.
      • Tsung A.
      • Yin X.M.
      Functions of autophagy in normal and diseased liver.
      • Ding W.X.
      • Li M.
      • Chen X.
      • Ni H.M.
      • Lin C.W.
      • Gao W.
      • Lu B.
      • Stolz D.B.
      • Clemens D.L.
      • Yin X.M.
      Autophagy reduces acute ethanol-induced hepatotoxicity and steatosis in mice.
      • Lin C.W.
      • Zhang H.
      • Li M.
      • Xiong X.
      • Chen X.
      • Chen X.
      • Dong X.C.
      • Yin X.M.
      Pharmacological promotion of autophagy alleviates steatosis and injury in alcoholic and non-alcoholic fatty liver conditions in mice.
      These observations support an important role of autophagy/lipophagy in regulating lipid metabolism during the pathogenesis of nonalcoholic and alcoholic fatty liver diseases. Of interest, the numbers of LDs increased rather than decreased when autophagic flux increased under starvation conditions in several cell culture models.
      • Rambold A.S.
      • Cohen S.
      • Lippincott-Schwartz J.
      Fatty acid trafficking in starved cells: regulation by lipid droplet lipolysis, autophagy, and mitochondrial fusion dynamics.
      • Li Y.
      • Zong W.X.
      • Ding W.X.
      Recycling the danger via lipid droplet biogenesis after autophagy.
      • Nguyen T.B.
      • Louie S.M.
      • Daniele J.R.
      • Tran Q.
      • Dillin A.
      • Zoncu R.
      • Nomura D.K.
      • Olzmann J.A.
      DGAT1-dependent lipid droplet biogenesis protects mitochondrial function during starvation-induced autophagy.
      Moreover, pharmacologic inhibition of autophagy or genetic knockout (KO) of Atg5 in mouse embryonic fibroblasts (MEFs) resulted in decreased numbers of LDs after starvation.
      • Nguyen T.B.
      • Louie S.M.
      • Daniele J.R.
      • Tran Q.
      • Dillin A.
      • Zoncu R.
      • Nomura D.K.
      • Olzmann J.A.
      DGAT1-dependent lipid droplet biogenesis protects mitochondrial function during starvation-induced autophagy.
      Increased formation of LDs has been suggested to counteract the potential lipotoxicity due to the release of FFAs by autophagy/lipophagy-mediated TG hydrolysis in these cultured cells.
      • Nguyen T.B.
      • Louie S.M.
      • Daniele J.R.
      • Tran Q.
      • Dillin A.
      • Zoncu R.
      • Nomura D.K.
      • Olzmann J.A.
      DGAT1-dependent lipid droplet biogenesis protects mitochondrial function during starvation-induced autophagy.
      In line with these in vitro findings, liver-specific Atg7 or FIP200 KO mice also had decreased numbers of LDs in the liver after overnight fasting.
      • Ma D.
      • Molusky M.M.
      • Song J.
      • Hu C.R.
      • Fang F.
      • Rui C.
      • Mathew A.V.
      • Pennathur S.
      • Liu F.
      • Cheng J.X.
      • Guan J.L.
      • Lin J.D.
      Autophagy deficiency by hepatic FIP200 deletion uncouples steatosis from liver injury in NAFLD.
      • Kim K.H.
      • Jeong Y.T.
      • Oh H.
      • Kim S.H.
      • Cho J.M.
      • Kim Y.N.
      • Kim S.S.
      • Kim D.H.
      • Hur K.Y.
      • Kim H.K.
      • Ko T.
      • Han J.
      • Kim H.L.
      • Kim J.
      • Back S.H.
      • Komatsu M.
      • Chen H.
      • Chan D.C.
      • Konishi M.
      • Itoh N.
      • Choi C.S.
      • Lee M.S.
      Autophagy deficiency leads to protection from obesity and insulin resistance by inducing Fgf21 as a mitokine.
      The mechanisms of how impaired hepatic autophagies lead to decreased formation of LDs in mouse livers after fasting are not clear. Under starvation conditions, the physiological adaptive mechanisms in mice could be different from cultured cells. In response to starvation, there are active cross talks among different organs/tissues and changes of circulating endocrine factors in mice, which are all lacking in cultured cells. To maintain systemic energy balance during starvation, liver increases the uptake of FFAs derived from the increased lipolysis in white adipose tissues, resulting in rapid body fat mobilization and accumulation of LDs in the liver.
      • Zechner R.
      • Madeo F.
      • Kratky D.
      Cytosolic lipolysis and lipophagy: two sides of the same coin.
      • Zechner R.
      • Strauss J.G.
      • Haemmerle G.
      • Lass A.
      • Zimmermann R.
      Lipolysis: pathway under construction.
      FFAs activate peroxisome proliferator-activated receptor α (PPARα), which increases hepatic FA oxidation and ketogenesis. PPARα may also increase the production of hormone fibroblast growth factor (FGF) 21 from the liver during fasting to further stimulate hepatic FA oxidation and ketogenesis.
      • Inagaki T.
      • Dutchak P.
      • Zhao G.
      • Ding X.
      • Gautron L.
      • Parameswara V.
      • Li Y.
      • Goetz R.
      • Mohammadi M.
      • Esser V.
      • Elmquist J.K.
      • Gerard R.D.
      • Burgess S.C.
      • Hammer R.E.
      • Mangelsdorf D.J.
      • Kliewer S.A.
      Endocrine regulation of the fasting response by PPARalpha-mediated induction of fibroblast growth factor 21.
      Nuclear factor (erythroid-derived 2)-like 2 (NFE2L2 or Nrf2) is a transcription factor, which is activated to up-regulate the expression of cellular antioxidant genes in response to oxidative stress.
      • Taguchi K.
      • Motohashi H.
      • Yamamoto M.
      Molecular mechanisms of the Keap1-Nrf2 pathway in stress response and cancer evolution.
      • Lau A.
      • Villeneuve N.F.
      • Sun Z.
      • Wong P.K.
      • Zhang D.D.
      Dual roles of Nrf2 in cancer.
      Under normal conditions, Nrf2 binds with Kelch-like ECH-associated protein 1 (Keap1), which recruits Cullin 3-Ring Box Protein 1 E3 ubiquitin ligase complex and targets Nrf2 for degradation by the 26S proteasome. Oxidative/electrophilic stress modifies a key cysteine residue of Keap1, resulting in the disruption of Keap1-dependent degradation of Nrf2 that allows the nuclear translocation of Nrf2. Once it enters the nuclei, Nrf2 then forms a heterodimer with small musculoaponeurotic fibrosarcoma protein, which activates the target gene transcription by binding to the antioxidant response element. This above-mentioned Nrf2 regulation is termed canonical Nrf2 activation.
      • Dodson M.
      • Zhang D.D.
      Non-canonical activation of NRF2: new insights and its relevance to disease.
      In addition to canonical Nrf2 activation, Nrf2 can also be activated by the noncanonical pathway through SQSTM1/p62 (hereafter referred to as p62), a protein that generally accumulates in cells with impaired autophagy.
      • Bjorkoy G.
      • Lamark T.
      • Brech A.
      • Outzen H.
      • Perander M.
      • Overvatn A.
      • Stenmark H.
      • Johansen T.
      p62/SQSTM1 forms protein aggregates degraded by autophagy and has a protective effect on huntingtin-induced cell death.
      • Ni H.M.
      • Boggess N.
      • McGill M.R.
      • Lebofsky M.
      • Borude P.
      • Apte U.
      • Jaeschke H.
      • Ding W.X.
      Liver-specific loss of Atg5 causes persistent activation of Nrf2 and protects against acetaminophen-induced liver injury.
      • Ni H.M.
      • Woolbright B.L.
      • Williams J.
      • Copple B.
      • Cui W.
      • Luyendyk J.P.
      • Jaeschke H.
      • Ding W.X.
      Nrf2 promotes the development of fibrosis and tumorigenesis in mice with defective hepatic autophagy.
      p62 directly interacts with Keap1, resulting in Nrf2 accumulation and activation.
      • Komatsu M.
      • Kurokawa H.
      • Waguri S.
      • Taguchi K.
      • Kobayashi A.
      • Ichimura Y.
      • Sou Y.S.
      • Ueno I.
      • Sakamoto A.
      • Tong K.I.
      • Kim M.
      • Nishito Y.
      • Iemura S.
      • Natsume T.
      • Ueno T.
      • Kominami E.
      • Motohashi H.
      • Tanaka K.
      • Yamamoto M.
      The selective autophagy substrate p62 activates the stress responsive transcription factor Nrf2 through inactivation of Keap1.
      • Lau A.
      • Wang X.J.
      • Zhao F.
      • Villeneuve N.F.
      • Wu T.
      • Jiang T.
      • Sun Z.
      • White E.
      • Zhang D.D.
      A noncanonical mechanism of Nrf2 activation by autophagy deficiency: direct interaction between Keap1 and p62.
      We and others previously demonstrated that persistent activation of Nrf2 is responsible for the profound pathologic changes, including hepatocyte damage, liver inflammation, fibrosis, compensatory proliferation, and benign tumor formation, in liver-specific Atg5 (L-Atg5) and L-Atg7 KO mice.
      • Ni H.M.
      • Woolbright B.L.
      • Williams J.
      • Copple B.
      • Cui W.
      • Luyendyk J.P.
      • Jaeschke H.
      • Ding W.X.
      Nrf2 promotes the development of fibrosis and tumorigenesis in mice with defective hepatic autophagy.
      • Komatsu M.
      • Waguri S.
      • Koike M.
      • Sou Y.S.
      • Ueno T.
      • Hara T.
      • Mizushima N.
      • Iwata J.
      • Ezaki J.
      • Murata S.
      • Hamazaki J.
      • Nishito Y.
      • Iemura S.
      • Natsume T.
      • Yanagawa T.
      • Uwayama J.
      • Warabi E.
      • Yoshida H.
      • Ishii T.
      • Kobayashi A.
      • Yamamoto M.
      • Yue Z.
      • Uchiyama Y.
      • Kominami E.
      • Tanaka K.
      Homeostatic levels of p62 control cytoplasmic inclusion body formation in autophagy-deficient mice.
      • Inami Y.
      • Waguri S.
      • Sakamoto A.
      • Kouno T.
      • Nakada K.
      • Hino O.
      • Watanabe S.
      • Ando J.
      • Iwadate M.
      • Yamamoto M.
      • Lee M.S.
      • Tanaka K.
      • Komatsu M.
      Persistent activation of Nrf2 through p62 in hepatocellular carcinoma cells.
      Interestingly, accumulating evidence suggests that Nrf2 also regulates lipid metabolism and plays a role in fatty liver diseases although conflicting observations have been reported, likely due to different diets and mouse models that were used.
      • Xu J.
      • Donepudi A.C.
      • Moscovitz J.E.
      • Slitt A.L.
      Keap1-knockdown decreases fasting-induced fatty liver via altered lipid metabolism and decreased fatty acid mobilization from adipose tissue.
      • Chartoumpekis D.V.
      • Ziros P.G.
      • Psyrogiannis A.I.
      • Papavassiliou A.G.
      • Kyriazopoulou V.E.
      • Sykiotis G.P.
      • Habeos I.G.
      Nrf2 represses FGF21 during long-term high-fat diet-induced obesity in mice.
      • Ramadori P.
      • Drescher H.
      • Erschfeld S.
      • Schumacher F.
      • Berger C.
      • Fragoulis A.
      • Schenkel J.
      • Kensler T.W.
      • Wruck C.J.
      • Trautwein C.
      • Kroy D.C.
      • Streetz K.L.
      Hepatocyte-specific Keap1 deletion reduces liver steatosis but not inflammation during non-alcoholic steatohepatitis development.
      • Sugimoto H.
      • Okada K.
      • Shoda J.
      • Warabi E.
      • Ishige K.
      • Ueda T.
      • Taguchi K.
      • Yanagawa T.
      • Nakahara A.
      • Hyodo I.
      • Ishii T.
      • Yamamoto M.
      Deletion of nuclear factor-E2-related factor-2 leads to rapid onset and progression of nutritional steatohepatitis in mice.
      However, the role of Nrf2 activation in regulating hepatic lipid metabolism in autophagy-deficient mouse livers is unclear.
      In the present study, we demonstrated that hepatic autophagy-deficient mice (L-Atg5 KO mice) had impaired adaptation to fasting-induced hepatic accumulation of LDs. The maladaptation to fasting-induced hepatic accumulation of LDs was not due to changes of de novo lipogenesis, secretion of hepatic very-low-density lipoprotein (VLDL) or FA β-oxidation but was associated with persistent Nrf2 activation in L-Atg5 KO mouse livers.

      Materials and Methods

      Cell Culture and Primary Hepatocyte Isolation

      Atg5−/− (Atg5 KO) and Atg7−/− (Atg7 KO) MEFs and their matched wild-type (WT) MEFs were generously donated by Dr. Noboru Mizushima (University of Tokyo, Tokyo, Japan) and Dr. Masaaki Komatsu (Niigata University School of Medicine, Niigata, Japan), respectively. MEFs were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum (Gibco, Gaithersburg, MD), supplemented with l-glutamine and penicillin/streptomycin. All cell culture materials were obtained from Invitrogen (Carlsbad, CA), and all chemicals were obtained from Sigma-Aldrich (St. Louis, MO) if not specifically stated. Mouse hepatocytes were isolated by a retrograde, nonrecirculating perfusion of livers with 0.05% Collagenase Type IV (Sigma-Aldrich) as described previously.
      • Ding W.X.
      • Ni H.M.
      • DiFrancesca D.
      • Stolz D.B.
      • Yin X.M.
      Bid-dependent generation of oxygen radicals promotes death receptor activation-induced apoptosis in murine hepatocytes.
      Cells were cultured in William's medium E with 10% fetal bovine serum but no other supplements for 2 hours for attachment. Cells were then cultured in the same medium without serum overnight before they were loaded with oleic acid (OA; 250 μmol/L). Briefly, a 20 mmol/L solution of OA in 0.01 N NaOH was incubated at 70°C for 30 minutes, and FA soaps were then complexed with 5% bovine serum albumin (BSA) in phosphate-buffered saline at a 7:1 molar ratio of FAs to BSA as we described previously.
      • Mei S.
      • Ni H.M.
      • Manley S.
      • Bockus A.
      • Kassel K.M.
      • Luyendyk J.P.
      • Copple B.L.
      • Ding W.X.
      Differential roles of unsaturated and saturated fatty acids on autophagy and apoptosis in hepatocytes.
      The OA/BSA conjugates were administered to the cultured cells. Five percent BSA was used as a vehicle control. All cells were maintained in a 37°C incubator with 5% CO2.

      Antibodies and Reagents

      Antibodies used in the study were CD36 (ab64014; Abcam, Cambridge, UK), rabbit anti-Atg5 antibody (PM050; MBL, Woburn, MA), p62 (H00008878-M01; Abnova, Taiwan), β-Actin (a5541; Sigma-Aldrich), and glyceraldehyde-3-phosphate dehydrogenase (2118; Cell Signaling Technology, Danvers MA). The rabbit polyclonal anti-LC3B antibody was made by using a peptide that represented the NH2-terminal 14 amino acids of human LC3B and an additional cysteine (PSEKTFKQRRTFEQC) as described before.
      • Ding W.X.
      • Ni H.M.
      • Gao W.
      • Chen X.
      • Kang J.H.
      • Stolz D.B.
      • Liu J.
      • Yin X.M.
      Oncogenic transformation confers a selective susceptibility to the combined suppression of the proteasome and autophagy.
      The secondary antibodies used in this study were horseradish peroxidase–conjugated goat anti-mouse (115-035-062; Jackson ImmunoResearch, West Grove, PA) or rabbit antibodies (111-035-045; Jackson ImmunoResearch).

      Animal Experiments

      Atg5flox/flox mice (C57BL/6/129) were generated by Dr. Noboru Mizushima as described previously
      • Hara T.
      • Nakamura K.
      • Matsui M.
      • Yamamoto A.
      • Nakahara Y.
      • Suzuki-Migishima R.
      • Yokoyama M.
      • Mishima K.
      • Saito I.
      • Okano H.
      • Mizushima N.
      Suppression of basal autophagy in neural cells causes neurodegenerative disease in mice.
      and were crossed with Albumin-Cre mice (Alb-Cre, C57BL/6J) (The Jackson Laboratory, Bar Harbor, ME) for at last eight generations to generate L-Atg5 KO mice. Nrf2−/− (C57BL/6J) mice were obtained from The Jackson Laboratory and were further crossed with Atg5 F/F, Alb Cre+ mice to generate Atg5 F/F, Alb Cre+/Nrf2−/− double-knockout (DKO) mice. Alb Cre-matched littermates were used as control WT mice in this study. All animals received humane care. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Kansas Medical Center. Two- to 3-month–old male L-Atg5 KO, L-Atg5 KO/Nrf2−/− (L-Atg5/Nrf2 DKO), Nrf2−/− (Nrf2 KO), and matched WT littermates were either on regular chow diet or fasted for overnight. After overnight fasting, some mice were refed with chow diet for another 6 hours before they were euthanized. Liver tissues were harvested, and hematoxylin and eosin staining as well as Oil Red O staining of liver sections were performed as described previously.
      • Williams J.A.
      • Ni H.M.
      • Ding Y.
      • Ding W.X.
      Parkin regulates mitophagy and mitochondrial function to protect against alcohol-induced liver injury and steatosis in mice.
      Liver TG and cholesterol were extracted and quantified as described previously.
      • Williams J.A.
      • Ni H.M.
      • Ding Y.
      • Ding W.X.
      Parkin regulates mitophagy and mitochondrial function to protect against alcohol-induced liver injury and steatosis in mice.
      Total liver lysates were prepared by using radioimmunoprecipitation assay buffer [1% NP40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl (lauryl) sulfate] with protease inhibitor.

      Immunoblot Assay

      Total liver lysates (20 μg per sample) were subjected to SDS-PAGE and transferred to polyvinylidene difluoride membranes. The membranes were incubated with primary antibodies, followed by secondary horseradish peroxidase–conjugated antibodies. SuperSignal West Pico chemiluminescent substrate (Pierce, Waltham, MA) was used for the membrane development.

      Serum Lipid and Other Biochemical Analysis

      Serum cholesterol (Pointe Scientific, Canton, MI), TG (Pointe Scientific), FFAs (BioVision, San Francisco, CA), and free glycerol (Sigma-Aldrich) as well as β-hydroxybutyrate (β-HB) (BioVision) were measured with colorimetric assay kits according to the manufactures' instructions. Pooled plasma samples (five to six mice) were used for fast protein liquid chromatographic analysis of lipoprotein profile, which was conducted by the Mouse Metabolic Phenotyping Center at the University of Cincinnati (Cincinnati, OH). Serum FGF21 levels were measured by using a commercial enzyme-linked immunosorbent assay kit (R&D Systems, Minneapolis, MN). All other reagents were either from Sigma-Aldrich or Invitrogen.

      Fluorescence Microscopy

      Cultured primary hepatocytes and MEFs were seeded in 12-well plates with microscope cover glasses (Thermo Fisher Scientific, Waltham, MA). Cells were loaded with 250 μmol/L OA or BSA control for 6 hours. After treatments, cells were fixed with 4% paraformaldehyde in phosphate-buffered saline. Cells were further stained with 1 μmol/L Bodipy 493/503 and 1 μmol/L Hoechst 33342. An inverted confocal microscope (Nikon, Melville, NY) or a Nikon Eclipse 200 fluorescence microscope was used for all imaging studies.

      Electron Microscopy

      Liver tissues were cut into small pieces and fixed with 2.5% glutaraldehyde in 0.1 mol/L phosphate buffer (pH 7.4), followed by 1% OsO4. Thin sections were stained with uranyl acetate and lead citrate after dehydration. Images were acquired digitally by using a JEM 1011CX electron microscope (JEOL, Peabody, MA). Randomly selected 15 to 20 cells under each condition were used to quantify the number of LDs and autophagosomes.

      Real-Time PCR

      Trizol (Invitrogen) was used to isolate liver RNA, and RNA was reverse transcribed into cDNA by RevertAid Reverse Transcriptase (Fermentas, Burlington, ON, Canada). Real-time quantitative PCR was used to quantify the genes for lipid metabolism and was performed on an Applied Biosystems Prism 7900HT real-time PCR instrument (ABI, Foster City, CA), using Maxima SYBR green/rox quantitative PCR reagents (Fermentas). The primer sequences are described in Table 1. Actb was used for normalization.
      Table 1List of Primers Used for Real-Time Quantitative PCR
      GeneForward primerReverse primer
      Abca15′-GCTTGTTGGCCTCAGTTAAGG-3′5′-GTAGCTCAGGCGTACAGAGAT-3′
      Abcg55′-AGGGCCTCACATCAACAGAG-3′5′-GCTGACGCTGTAGGACACAT-3′
      Abcg85′-CTGTGGAATGGGACTGTACTTC-3′5′-GTTGGACTGACCACTGTAGGT-3′
      Acaca5′-CTCCAGGACAGCACAGATCA-3′5′-TGACTGCCGAAACATCTCTG-3′
      Acox5′-CAGGAAGAGCAAGGAAGTGG-3′5′-CCTTTCTGGCTGATCCCATA-3′
      Actb5′-TGTTACCAACTGGGACGACA-3′5′-GGGGTGTTGAAGGTCTCAAA-3′
      Cd365′-ATGGGCTGTGATCGGAACTG-3′5′-AGCCAGGACTGCACCAATAAC-3′
      Cpt1a5′-CCAGGCTACAGTGGGACATT-3′5′-GAACTTGCCCATGTCCTTGT-3′
      Crot5′-TACTTTTACCACGGCCGAAC-3′5′-GACGGTCAAATCCTTTTCCA-3′
      Dgat15′-TCCGTCCAGGGTGGTAGTG-3′5′-TGAACAAAGAATCTTGCAGACGA-3′
      Dgat25′-GCGCTACTTCCGAGACTACTT-3′5′-GGGCCTTATGCCAGGAAACT-3′
      Fasn5′-TGGGTTCTAGCCAGCAGAGT-3′5′-ACCACCAGAGACCGTTATGC-3′
      Fgf215′-CTGGGGGTCTACCAAGCATA-3′5′-CACCCAGGATTTGAATGACC-3′
      Gyk5′-ATCCGCTGGCTAAGAGACAACC-3′5′-TGCACTGGGCTCCCAATAAGG-3′
      Hmgcr5′-TCTTGTGGAATGCCTTGTGA-3′5′-GCGACTATGAGCGTGAACAA-3′
      Insig15′-CACGACCACGTCTGGAACTAT-3′5′-TGAGAAGAGCACTAGGCTCCG-3′
      Ldlr5′-TGACTCAGACGAACAAGGCTG-3′5′-ATCTAGGCAATCTCGGTCTCC-3′
      Gapdhm5′-AGGTTTCAAGATCGCAATGG-3′5′-CTCCTTGGTGCTCCACTAGC-3′
      Acadm5′-CTCGCCATCGCCCTCACTG-3′5′-ACCGCTCACTCGCTCTTTGC-3′
      Mttp5′-CTCTTGGCAGTGCTTTTTCTCT-3′5′-GAGCTTGTATAGCCGCTCATT-3′
      Pgc1a5′-ATGTGTCGCCTTCTTGCTCT-3′5′-ATCTACTGCCTGGGGACCTT-3′
      Ppara5′-ATGCCAGTACTGCCGTTTTC-3′5′-GGCCTTGACCTTGTTCATGT-3′
      Pparg5′-TTTTCAAGGGTGCCAGTTTC-3′5′-AATCCTTGGCCCTCTGAGAT-3′
      Scarb15′-TTTGGAGTGGTAGTAAAAAGGGC-3′5′-TGACATCAGGGACTCAGAGTAG-3′
      Scd15′-TGCGATACACTCTGGTGCTC-3′5′-TAGTCGAAGGGGAAGGTGTG-3′
      Srebp1c5′-GATCAAAGAGGAGCCAGTGC-3′5′-TAGATGGTGGCTGCTGAGTG-3′
      Srebp25′-GCGTTCTGGAGACCATGGA-3′5′-ACAAAGTTGCTCTGAAAACAAATCA-3′

      Results

      Loss of Nrf2 Suppresses the Impaired LD Accumulation in L-Atg5 KO Mouse Livers after Fasting

      To study the role of autophagy and Nrf2 activation in fasting-induced adaptive accumulation of LD in the liver, WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were fasted overnight. Consistent with previous reports,
      • Inagaki T.
      • Dutchak P.
      • Zhao G.
      • Ding X.
      • Gautron L.
      • Parameswara V.
      • Li Y.
      • Goetz R.
      • Mohammadi M.
      • Esser V.
      • Elmquist J.K.
      • Gerard R.D.
      • Burgess S.C.
      • Hammer R.E.
      • Mangelsdorf D.J.
      • Kliewer S.A.
      Endocrine regulation of the fasting response by PPARalpha-mediated induction of fibroblast growth factor 21.
      • Ma D.
      • Molusky M.M.
      • Song J.
      • Hu C.R.
      • Fang F.
      • Rui C.
      • Mathew A.V.
      • Pennathur S.
      • Liu F.
      • Cheng J.X.
      • Guan J.L.
      • Lin J.D.
      Autophagy deficiency by hepatic FIP200 deletion uncouples steatosis from liver injury in NAFLD.
      • Xu J.
      • Donepudi A.C.
      • Moscovitz J.E.
      • Slitt A.L.
      Keap1-knockdown decreases fasting-induced fatty liver via altered lipid metabolism and decreased fatty acid mobilization from adipose tissue.
      • Shibata M.
      • Yoshimura K.
      • Furuya N.
      • Koike M.
      • Ueno T.
      • Komatsu M.
      • Arai H.
      • Tanaka K.
      • Kominami E.
      • Uchiyama Y.
      The MAP1-LC3 conjugation system is involved in lipid droplet formation.
      Oil Red O staining for neutral lipids showed increased hepatic steatosis in fasted WT mice as expected. In contrast, fasting-induced steatosis was almost absent in the L-Atg5 KO mouse livers (Figure 1A). Electron microscopic studies also showed increased accumulation of LDs in WT mouse livers after overnight fasting (Figure 1, B and C). L-Atg5 KO mice have slightly increased number of LDs, but most of these LDs were small and enwrapped in the aberrant multimembrane structures after fasting (Figure 1, B and C). L-Atg5 KO mice have persistent Nrf2 activation, which contributed to the liver pathogenesis, including cell death, inflammation, fibrosis, and tumorigenesis in L-Atg5 KO mice.
      • Ni H.M.
      • Boggess N.
      • McGill M.R.
      • Lebofsky M.
      • Borude P.
      • Apte U.
      • Jaeschke H.
      • Ding W.X.
      Liver-specific loss of Atg5 causes persistent activation of Nrf2 and protects against acetaminophen-induced liver injury.
      • Ni H.M.
      • Woolbright B.L.
      • Williams J.
      • Copple B.
      • Cui W.
      • Luyendyk J.P.
      • Jaeschke H.
      • Ding W.X.
      Nrf2 promotes the development of fibrosis and tumorigenesis in mice with defective hepatic autophagy.
      Interestingly, L-Atg5/Nrf2 DKO and Nrf2 KO mice had the normal hepatic fasting response as demonstrated by the increased hepatic accumulation of LDs after fasting (Figure 1, A–C). The levels of hepatic TGs and cholesterol levels in WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were further determined under the fed, fasting, and refed conditions. Consistent with the Oil Red O staining results, the levels of fasting hepatic TGs markedly increased (approximately sixfold) compared with the fed state, which declined almost to the levels of the fed state after refed in WT mice (Figure 2A). The levels of hepatic cholesterol, serum TGs, and cholesterol did not change significantly under the fed, fasting, and refed conditions in WT mice (Figure 2, B and C). Strikingly, the levels of hepatic TGs remained almost unchanged regardless of the fed, fasting, or refed condition in L-Atg5 KO mice. Similar to WT mice, the levels of hepatic cholesterol and serum TGs did not change in L-Atg5 KO mice under the fed, fasting, and refed conditions except that the serum TG levels were slightly higher in L-Atg5 KO mice than that of WT mice at the basal (fed) state. In addition, the serum TG levels in Nrf2 single-KO mice were also relatively higher than the other genotypes of mice under the fed, fasting, and refed conditions (Figure 2C). However, the serum cholesterol levels were significantly higher in L-Atg5 KO mice than their matched WT mice regardless of the fed, fasting, and refed conditions (Figure 2D), suggesting that L-Atg5 KO mice have hypercholesterolemia. Fast protein liquid chromatographic analysis confirmed that serum VLDL-TG levels were relatively higher in fasting Nrf2 KO mice (Figure 2E). Consistent with the increased total cholesterol levels in L-Atg5 KO mice, fast protein liquid chromatographic analysis also showed that fasting serum levels of LDL cholesterol and high-density lipoprotein cholesterol were much higher in L-Atg5 KO mice. However, L-Atg5/Nrf2 DKO and Nrf2 KO mice had almost the same levels of fasting serum cholesterol profiles with the WT mice (Figure 2, F and G). The body weight of age-matched WT and L-Atg5 KO mice was comparable from 2 to 9 months old (Supplemental Figure S1A). As reported previously,
      • Ni H.M.
      • Boggess N.
      • McGill M.R.
      • Lebofsky M.
      • Borude P.
      • Apte U.
      • Jaeschke H.
      • Ding W.X.
      Liver-specific loss of Atg5 causes persistent activation of Nrf2 and protects against acetaminophen-induced liver injury.
      the liver-to-body weight ratio was much higher in L-Atg5 KO mice than their matched WT mice, suggesting that L-Atg5 KO mice have hepatomegaly. In response to the fasting and refed conditions, the ratio of liver-to-body weight slightly decreased in WT and L-Atg5 KO mice. L-Atg5/Nrf2 DKO and Nrf2 KO mice had almost the same ratio of liver-to-body weight with the WT mice and did not change in response to the fasting and refed conditions (Supplemental Figure S1B). Taken together, these results indicated that fasting-induced hepatic LD accumulation was impaired in L-Atg5 KO mice, which was corrected by further deletion of Nrf2 in L-Atg5 KO mice.
      Figure thumbnail gr1
      Figure 1Fasting-induced hepatic accumulation of lipid droplets (LDs) is impaired in hepatic autophagy-deficient (L-Atg)5 knockout (KO) mice, which can be corrected by further deletion of nuclear factor-like (Nrf)2. A: WT, L-Atg5 KO, L-Atg5/Nrf2 double-knockout (DKO), and Nrf2 KO mice were either fed with a normal chow diet (Fed) or fasted overnight (Fasted). Liver tissues were harvested and cryo-liver sections were subjected to Oil Red O staining. Representative images are shown. B: Mice were treated as in panel A, and representative electron microscopic images are shown. The lower panel was an enlarged image from the boxed area in the upper panel image. Arrows indicate LDs. C: The number of LDs were quantified from panel B. Data are expressed as means ± SEM (C). n = 3 to 6 mice (A); n = 10 different images (C). P < 0.05 between different treatments, determined by t-test. Scale bar = 2 μm. N, nucleus.
      Figure thumbnail gr2
      Figure 2Hepatic and serum lipid changes in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5 KO/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice under the fed, fasting, and refed conditions. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed), fasted overnight (Fasted), or refed with chow diet for 6 hours (Refed). AD: Hepatic triglyceride (TG) (A) and cholesterol (B) levels as well as serum TG (C) and cholesterol (D) levels were quantified. Pooled plasma samples were used for fast phase liquid chromatographic (FPLC) analysis of lipoprotein profile. E: Serum TG profile. F: Serum cholesterol profile. G: Total serum cholesterol levels were quantified from the pooled FPLC analysis. Data are expressed as means ± SEM (AD). n = 3 to 6 mice (AD); n = 5 to 6 mice (EG). P < 0.05 (one-way analysis of variance analysis). HDL, high-density lipoprotein; LDL, low-density lipoprotein; VLDL, very-low-density lipoprotein.

      Autophagy and Nrf2 Signaling Changes in WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO Mice in Response to Nutrient Changes

      Increased hepatic protein levels of p62 and microtubule-associated protein 1 light chain 3 (LC3)-I form were found in L-Atg5 KO mice, indicating the defect of autophagy in these mouse livers (Supplemental Figure S2). The levels of p62 and LC3-I decreased in L-Atg5/Nrf2 DKO mice compared with L-Atg5 KO mice, suggesting Nrf2 may regulate p62 and LC3 at the transcription level as previously reported.
      • Jain A.
      • Lamark T.
      • Sjottem E.
      • Larsen K.B.
      • Awuh J.A.
      • Overvatn A.
      • McMahon M.
      • Hayes J.D.
      • Johansen T.
      p62/SQSTM1 is a target gene for transcription factor NRF2 and creates a positive feedback loop by inducing antioxidant response element-driven gene transcription.
      • Pajares M.
      • Jimenez-Moreno N.
      • Garcia-Yague A.J.
      • Escoll M.
      • de Ceballos M.L.
      • Van Leuven F.
      • Rabano A.
      • Yamamoto M.
      • Rojo A.I.
      • Cuadrado A.
      Transcription factor NFE2L2/NRF2 is a regulator of macroautophagy genes.
      In addition to increasing LD accumulation, fasting also increased the number of autophagosomes in WT mouse livers (Supplemental Figure S3, A and B). Some of the double membrane structures had enveloped mitochondrion (Supplemental Figure S3A). Close association of LDs was also found with intracellular vacuoles and some LDs intruded into the vacuoles (Supplemental Figure S3A), mimicking the lipophagy-like structures as reported previously,
      • Singh R.
      • Kaushik S.
      • Wang Y.
      • Xiang Y.
      • Novak I.
      • Komatsu M.
      • Tanaka K.
      • Cuervo A.M.
      • Czaja M.J.
      Autophagy regulates lipid metabolism.
      although these vacuoles appeared to be single-membrane structures. The nature of these single-membrane vacuoles remained to be determined in the future, which was beyond the scope of the present study. The hepatic protein levels of glutamate-cysteine ligase catalytic unit (GCLC), glutamate-cysteine ligase modifier unit (GCLM), and NAD(P)H quinone dehydrogenase 1 (NQO1), three well-known Nrf2 target genes, markedly increased in L-Atg5 KO mouse livers compared with WT mice, although the levels of GCLC, GCLM, and NQO1 seemed less affected by the fed, fasting, and refed conditions. The increased hepatic levels of GCLC, GCLM, and NQO1 in L-Atg5 KO mice was largely blunted in L-Atg5/Nrf2 DKO mice (Supplemental Figure S2). Taken together, these data confirmed that L-Atg5 KO mice have impaired hepatic autophagy and persistent hepatic Nrf2 activation.

      Decreased Hepatic LD Accumulation in L-Atg5 KO Mice Is Not Due to Decreased Expression of de Novo Lipogenic Genes or Increased Expression of FA β-Oxidation Genes

      At least four factors can contribute to liver steatosis, including i) increased de novo FA synthesis, ii) decreased VLDL-TG secretion, iii) decreased FA β-oxidation, and iv) increased hepatic FA uptake. Quantitative PCR analysis was next performed to determine the mRNA expression of genes involved in lipid metabolism. The expression of many genes involved in de novo lipid synthesis, including Scd1 and Acaca, did not change significantly in response to fasting in both WT and L-Atg5 KO mice. The expression of hepatic Pparg did not change significantly in WT mice after fasting, although it increased in L-Atg5 KO mice after fasting and remained higher after the refed condition. The expression of hepatic FA synthase (Fasn) and sterol regulatory element-binding transcription factor 1c (Srebp1c) decreased after fasting in both WT and L-Atg5 KO mice. Interestingly, the expression levels of Fasn, Scd1, Acaca, and Srebp1c were significantly higher in L-Atg5/Nrf2 DKO and Nrf2 KO mice than WT and L-Atg5 KO mice under the fed condition but decreased significantly under the fasted and refed conditions (Figure 3A). Western blot analysis was then performed to determine the changes of some of the key de novo lipid synthesis proteins. The protein levels of FASn, acetyl-CoA carboxylase, and matured SREBP1 (mSREBP1) either decreased or remained unchanged under the fasted conditions but increased under the refed conditions in WT mice (Supplemental Figure S4). These results generally supported the notion that de novo lipid synthesis was not a main factor for fasting-induced hepatic steatosis. Compared with WT mice, L-Atg5 KO mice had relatively higher basal levels of FASn, acetyl-CoA carboxylase, and mSREBP1, which were all slightly decreased under the fasted conditions. L-Atg5/Nrf2 DKO had higher basal protein levels of FASn, acetyl-CoA carboxylase, and mSREBP1, which were further increased after fasting. In contrast, Nrf2 KO mice had higher basal protein levels of mSREBP1 but markedly decreased levels under the fasted and refed conditions. The full-length protein of FASn was not detected in Nrf2 KO mouse livers, but instead a relatively small molecular band of FASn was detected in Nrf2 KO mouse livers and the reasons for these observations were unclear (Supplemental Figure S4). The hepatic adipose TG lipase protein levels increased under the fasted conditions in all four genotypes of mice. The levels of phosphorylated hormone-sensitive lipase increased in WT and L-Atg5/Nrf2 DKO mice under the fasted conditions but did not change in L-Atg5 KO mice and even decreased in Nrf2 KO mice (Supplemental Figure S4). Taken together, these data suggested that de novo lipogenesis and lipolysis may not be the main factors for the decreased accumulation of LDs in L-Atg5 KO mice after fasting. These data also suggested that deletion of Nrf2 may have aberrant effects on lipid metabolism in mouse livers. Nrf2 deletion may affect the basal hepatic lipogenesis gene expression but it did not affect fasting-induced hepatic lipid accumulation in mice.
      Figure thumbnail gr3
      Figure 3Changes of hepatic lipogenesis gene expression in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed), fasted overnight (Fasted), or refed with chow diet for 6 hours (Refed). A and B: Hepatic mRNA was extracted followed by real-time quantitative PCR for fatty acid synthesis genes (A) and TG synthesis genes (B). Data are expressed as means ± SEM. n = 3 to 6 mice. P < 0.05 (one-way analysis of variance analysis).
      The expression of hepatic cholesterol metabolism genes, including Ldlr and Srebp2, did not change under the fed, fasted, and refed conditions in all four genotypes of mice. The expression of hepatic Hmgcr and Insig1 decreased in all four genotypes of mice under the fasted condition compared with the fed and refed conditions (Supplemental Figure S5). The expression of several hepatic transporters for sterol and cholesterol secretion, including Abcg5, Abcg8, and Abca1, increased after fasting and decreased after the refed condition in both WT and L-Atg5 KO mice. However, the expression levels of Abcg5 and Abcg8 were much higher in L-Atg5 KO mice than in WT, L-Atg5/Nrf2 DKO, and Nrf2 KO mice under all three conditions (fed, fasted, and refed). In contrast, the hepatic expression of Abcg5 and Abcg8 did not change after fasting but decreased under the refed condition in L-Atg5/Nrf2 DKO and Nrf2 KO mice (Supplemental Figure S6). In addition, the levels of Scarb1, a receptor for high-density lipoprotein, did not change significantly in all four genotypes of mice except it increased after fasting in L-Atg5/Nrf2 DKO mice (Supplemental Figure S6).
      The expression of hepatic TG synthesis genes was also determined under the fed, fasted, and refed conditions in WT, L-Atg5/Nrf2 DKO, and Nrf2 KO mice. The expression of Gyk, Gapdhm, Dgat1, and Dgat2 all increased in response to fasting and declined after the refed condition in all four genotypes of mice. These data suggested that hepatic TG synthesis may not be the factor that contributed to the decreased hepatic TGs in L-Atg5 mice after fasting (Figure 3B).
      To determine whether decreased accumulation of hepatic LDs would be due to increased secretion of VLDL-TG from hepatocytes in L-Atg5 KO mice after fasting, hepatic TG secretion assay was performed in mice. No difference was found for TG secretion between WT and L-Atg5 KO mice after fasting (Supplemental Figure S7A). The hepatic expression of microsomal triglyceride transfer protein, an important protein for hepatic lipoprotein assembly and secretion, increased in WT, L-Atg5/Nrf2 DKO, and Nrf2 KO mice but did not change in L-Atg5 KO mice after fasting (Supplemental Figure S7B).
      The hepatic expression of FA β-oxidation–related genes was next determined. The hepatic expression of Ppara, Cpt1a, Acox, Crot, and Acadm increased significantly in all four genotypes of mice under the fasted condition compared with the fed and refed conditions. However, the induction of these genes was much lower in L-Atg5 KO mice than in WT, L-Atg5/Nrf2 DKO, and Nrf2 KO mice (Figure 4A). It is well known that fasting increases serum levels of ketone bodies, including β-HB.
      • Inagaki T.
      • Dutchak P.
      • Zhao G.
      • Ding X.
      • Gautron L.
      • Parameswara V.
      • Li Y.
      • Goetz R.
      • Mohammadi M.
      • Esser V.
      • Elmquist J.K.
      • Gerard R.D.
      • Burgess S.C.
      • Hammer R.E.
      • Mangelsdorf D.J.
      • Kliewer S.A.
      Endocrine regulation of the fasting response by PPARalpha-mediated induction of fibroblast growth factor 21.
      Indeed, serum levels of β-HB increased in all four genotypes of mice after fasting. Although the levels of serum of β-HB were comparable after fasting among all four genotypes of mice, the basal fed levels of serum of β-HB were much higher in L-Atg5 KO, L-Atg5 KO/Nrf2 DKO, and Nrf2 KO mice. Such a high level of serum β-HB persisted even after the refed condition in L-Atg5 KO mice (Figure 4B). Taken together, these data suggested that the lack of accumulation of LDs in L-Atg5 KO mice after fasting was less likely due to the changes of de novo lipogenesis, secretion of VLDL-TG, and β-oxidation of FA.
      Figure thumbnail gr4
      Figure 4Changes of hepatic lipid metabolism gene expression in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed), fasted overnight (Fasted), or refed with chow diet for 6 hours (Refed). A: Hepatic mRNA was extracted, followed by real-time quantitative PCR for fatty acid β-oxidation synthesis genes. B: The serum β-hydroxybutyrate levels were measured. Data are expressed as means ± SEM. n = 3 to 6 mice. P < 0.05 (one-way analysis of variance analysis).

      Decreased Serum NEFAs after Fasting and Increased Hepatic CD36 in L-Atg5 KO Mice

      The serum levels of nonesterified FAs (NEFAs) increased almost fourfold compared with the fed or refed condition in WT mice. In response to fasting, the serum levels of NEFAs also increased in L-Atg5/Nrf2 DKO and Nrf2 KO mice, although the levels were lower than in WT mice. In contrast, the basal (fed) levels of serum NEFAs were relatively higher in L-Atg5 KO mice than in WT mice, and the serum NEFA levels did not change after fasting in L-Atg5 KO mice (Figure 5A). Similar to the changes of serum NEFA, the levels of serum FGF21 significantly increased under the fasted condition compared with the fed and refed conditions in WT, L-Atg5/Nrf2 DKO, and Nrf2 KO mice. In contrast, the basal (fed) levels of serum FGF21 in L-Atg5 KO mice was already significantly higher than either WT, L-Atg5/Nrf2 DKO, or Nrf2 KO mice, which remained almost at the same levels when the mice were either fasted or refed (Figure 5B). Consistent with the serum FGF21 changes, the hepatic mRNA levels of Fgf21 markedly increased under the fasted condition and declined under the refed condition in WT, L-Atg5/Nrf2 DKO, and Nrf2 KO mice. In contrast, the hepatic mRNA levels of Fgf21 were already higher at the fed (basal) condition in L-Atg5 KO mice compared with WT mice, which further increased significantly after fasting and declined after the refed condition (Supplemental Figure S8A). The hepatic mRNA levels of Pgc1a increased in all four genotypes of mice in response to fasting and declined to the levels of the fed condition after the mice were refed, except that the mRNA levels of Pgc1a were much higher at the fed (basal) condition of Nrf2 KO mice than that of WT, L-Atg5 KO, and L-Atg5/Nrf2 DKO mice (Supplemental Figure S8B). The serum levels of glycerol did not change in WT and L-Atg5/Nrf2 DKO mice under the fed, fasted, and refed conditions. The level of serum glycerol was almost doubled under the fed condition in L-Atg5 KO mice compared with WT mice, but it decreased to the same levels of WT mice under the fasted and refed conditions. Under the fasted condition, Nrf2 KO mice had higher serum glycerol levels (almost twofold higher) than WT, L-Atg5 KO, and L-Atg5/Nrf2 DKO mice (Figure 5C). The hepatic mRNA levels of Cd36, a gene that encodes an integral membrane protein that imports hepatic FAs into cells, increased in WT, L-Atg5/Nrf2 DKO, and Nrf2 KO mice under the fasted condition compared with the fed condition, which declined under the refed condition in WT and Nrf2 KO mice but not in L-Atg5/Nrf2 DKO mice. In contrast, the hepatic mRNA levels of Cd36 were approximately 10-fold higher in L-Atg5 KO mice than in WT mice under the fed condition. The levels of Cd36 expression further increased under the fasted condition and slightly declined under the refed condition in L-Atg5 KO mice (Figure 5D). Consistent with the mRNA data, Western blot analysis also revealed higher CD36 protein levels in L-Atg5 KO mice than in WT, L-Atg5/Nrf2 DKO, and Nrf2 KO mice (Figure 5E). Together these data indicated that L-Atg5 KO mice have increased basal levels of serum NEFAs, FGF21, and glycerol, which are not altered under the fasting conditions.
      Figure thumbnail gr5
      Figure 5The serum levels of nonesterified fatty acids (NEFAs), glycerol, and fibroblast growth factor (FGF)21 in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf) 2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed, FD), fasted overnight (Fasted, FT), or refed with chow diet for 6 hours (Refed, RF). AC: The serum levels of NEFAs (A), free glycerol (B), and FGF21 (C) were measured. D: Hepatic mRNA was extracted, followed by real-time quantitative PCR for Cd36 (D). E: Total liver lysates were prepared, and pooled liver lysates were subjected to Western blot analysis. Data are expressed as means ± SEM (AD). n = 3 to 6 mice. P < 0.05 (one-way analysis of variance analysis). GAPDH, glyceraldehyde-3-phosphate dehydrogenase; L-Exp, longer exposure; S-Exp, shorter exposure.

      No Difference in the Uptake of OA and Formation of LDs between Atg5 KO Hepatocytes and WT Hepatocytes

      A previous study reported that LC3-phosphatidylethanolamine (PE) conjugation was required for the biogenesis of LDs.
      • Shibata M.
      • Yoshimura K.
      • Furuya N.
      • Koike M.
      • Ueno T.
      • Komatsu M.
      • Arai H.
      • Tanaka K.
      • Kominami E.
      • Uchiyama Y.
      The MAP1-LC3 conjugation system is involved in lipid droplet formation.
      To further determine whether autophagy-deficient cells would have defects for uptake of FFAs and biogenesis of LDs, hepatocytes were isolated from L-Atg5 KO and WT mice, and these cells were loaded with OA. The number of LDs markedly increased in the cytoplasm of OA-loaded WT hepatocytes was based on the Bodipy staining and electron microscopic studies (Figure 6, A–C ). Atg5 KO hepatocytes already had a higher number of LDs than that of WT hepatocytes even when they were cultured in the regular medium without OA. Addition of OA further increased the number of LDs in Atg5 KO hepatocytes (Figure 6, A–C), suggesting that deletion of Atg5 in hepatocytes may not affect the uptake of OA and biogenesis of LDs. Western blot analysis showed the lack of LC3-II forms and increased p62 proteins in Atg5 KO hepatocytes (Figure 6D), supporting the lack of autophagy in Atg5 KO hepatocytes. Similar to primary hepatocytes, increased accumulation of LDs was also observed in WT, Atg5 KO, and Atg7 KO MEFs when they were loaded with OA (Supplemental Figure S9A). No LC3-II forms were detected in Atg5 or Atg7 KO MEFs (Supplemental Figure S9B). These data not only confirmed the lack of autophagy but also suggested that LC3-II was dispensable for the formation/biogenesis of LDs in Atg5 KO and Atg7 KO cells. Taken together, these data suggested that cells lacking essential autophagy proteins such as Atg5 or Atg7 did not compromise the uptake of FAs and the formation/biogenesis of LDs in these autophagy-deficient cells.
      Figure thumbnail gr6
      Figure 6Lipid droplet (LD) formation in primary cultured wild-type (WT) and hepatic autophagy-deficient (Atg)5 knockout (KO) hepatocytes. Primary cultured hepatocytes that were isolated from WT and L-Atg5 KO mice were loaded with either bovine serum albumin or 250 μmol/L oleic acid (OA) for 6 hours. Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline. Cells were further stained with 1 μmol/L Bodipy 493/503 and 1 μmol/L Hoechst 33342, followed by fluorescence microscopy. A: Representative images of Bodipy and Hoechst 33342 staining are shown. B: Some cells were fixed with 2.5% glutaraldehyde, followed by electron microscopic (EM) analysis. Representative EM images are shown. C: The number of LDs was quantified from panel B. D: Total cell lysates were extracted and subjected to Western blot analysis. Data are expressed as means ± SEM (C). n = 15 to 20 images (C). P < 0.05 (one-way analysis of variance analysis). Scale bars: 20 μm (A); 0.5 μm (B). LC3, light chain 3; M, mitochondria; N, nucleus.

      Discussion

      The main finding of the present study was that mice with defective hepatic autophagy (L-Atg5 KO mice) had impaired adaptation to fasting-induced hepatic accumulation of LDs, which could be rescued by further deletion of Nrf2. The maladaptation to fasting-induced hepatic accumulation of LDs was not due to changes of de novo lipogenesis, secretion of hepatic VLDL, or increase in FA β-oxidation.
      Previous reports show that the LC3-PE conjugated form is associated with LDs, and it has been suggested that LC3-PE conjugation is required for the biogenesis of LD.
      • Shibata M.
      • Yoshimura K.
      • Furuya N.
      • Koike M.
      • Ueno T.
      • Komatsu M.
      • Arai H.
      • Tanaka K.
      • Kominami E.
      • Uchiyama Y.
      The MAP1-LC3 conjugation system is involved in lipid droplet formation.
      However, this notion is largely based on the observation of decreased LD biogenesis in Atg7-deficient mouse livers after fasting. In the present study, both WT and Atg5 KO hepatocytes and MEFs increased the number of LD formation when they were loaded with OA in culture. Because both Atg5 KO hepatocytes and MEFs lack LC3-PE conjugation, these results suggest that LC3-PE conjugation is dispensable for the biogenesis of LD.
      In cultured cells, deprivation of amino acids, glucose, and serum increases the number of intracellular LDs accompanied with increased autophagy. Increased accumulation of LDs in these starved cells has been suggested to be due to autophagic degradation of intracellular membranes, which generates FFAs. Increased LD formation is to channel the FFAs as an adaptive protective mechanism against FFA accumulation and lipotoxicity.
      • Rambold A.S.
      • Cohen S.
      • Lippincott-Schwartz J.
      Fatty acid trafficking in starved cells: regulation by lipid droplet lipolysis, autophagy, and mitochondrial fusion dynamics.
      • Nguyen T.B.
      • Louie S.M.
      • Daniele J.R.
      • Tran Q.
      • Dillin A.
      • Zoncu R.
      • Nomura D.K.
      • Olzmann J.A.
      DGAT1-dependent lipid droplet biogenesis protects mitochondrial function during starvation-induced autophagy.
      Although increased numbers of AVs and lipophagy-like structures were found in fasted WT mouse livers, the number of LDs was markedly increased compared with fed mice. These results suggest that autophagy-mediated degradation of LDs (lipophagy) is likely not sufficient to counteract the formation/biogenesis of LDs that convert from FFAs under starvation/fasting conditions. In addition, although the phenotype of increased LD formation is quite similar in starved cells and fasted mouse livers, the mechanisms for the accumulation of the intracellular LDs could be different in vivo versus in vitro. As discussed in the introduction, in response to starvation, mammals, including humans, increase the mobilization of FFAs from adipose tissues to the liver that contributes to the accumulation of LDs in the fasted liver, which is different from increased autophagic degradation of membranes in cultured cells under starvation conditions.
      To determine the role of autophagy in hepatic biogenesis of LDs in response to fasting, mice that have impaired hepatic autophagy would be helpful tools. Indeed, L-Atg5 KO mice failed to increase the accumulation of hepatic LDs in response to overnight fasting compared with their matched WT mice. At first glance, these in vivo data from L-Atg5 KO mice would greatly support the notion derived from the cell culture studies that autophagy is important for the biogenesis of LDs. However, further deletion of Nrf2 in L-Atg5 KO mice almost completely corrected the lack of accumulation of hepatic LDs in fasted L-Atg5 mouse livers, suggesting that the lack of accumulation of LDs in fasted L-Atg5 KO mouse livers is most likely due to secondary effects in L-Atg5 KO mice. Mice with defective hepatic autophagy have increased liver cell death/injury, hepatomegaly, inflammation, and fibrosis and eventually develop benign adenoma in aged mice.
      • Ni H.M.
      • Boggess N.
      • McGill M.R.
      • Lebofsky M.
      • Borude P.
      • Apte U.
      • Jaeschke H.
      • Ding W.X.
      Liver-specific loss of Atg5 causes persistent activation of Nrf2 and protects against acetaminophen-induced liver injury.
      • Ni H.M.
      • Woolbright B.L.
      • Williams J.
      • Copple B.
      • Cui W.
      • Luyendyk J.P.
      • Jaeschke H.
      • Ding W.X.
      Nrf2 promotes the development of fibrosis and tumorigenesis in mice with defective hepatic autophagy.
      • Komatsu M.
      • Waguri S.
      • Koike M.
      • Sou Y.S.
      • Ueno T.
      • Hara T.
      • Mizushima N.
      • Iwata J.
      • Ezaki J.
      • Murata S.
      • Hamazaki J.
      • Nishito Y.
      • Iemura S.
      • Natsume T.
      • Yanagawa T.
      • Uwayama J.
      • Warabi E.
      • Yoshida H.
      • Ishii T.
      • Kobayashi A.
      • Yamamoto M.
      • Yue Z.
      • Uchiyama Y.
      • Kominami E.
      • Tanaka K.
      Homeostatic levels of p62 control cytoplasmic inclusion body formation in autophagy-deficient mice.
      It is highly likely that the pathologic changes in the L-Atg5 KO mice liver may influence hepatic lipid metabolism and likely affect the whole body metabolic status in L-Atg5 KO mice. Indeed, at the basal fed conditions, L-Atg5 KO mice had increased serum levels of FFAs, β-HB, and FGF21 compared with their matched WT mice. Because all of these markers generally increased under fasting/starvation conditions, it is unclear whether L-Atg5 KO mice may constantly be under a chronic starvation condition even when they are free to access chow diet. This would help to explain why L-Atg5 KO mice had diminished induction of hepatic PPARα and PPARα-mediated expression of FA oxidation genes. As a result, white adipose tissue lipolysis may also constantly increase, resulting in the high serum levels of FFAs to meet the need of the chronic starvation in L-Atg5 KO mice.
      Interestingly, accumulating evidence suggests that Nrf2 also regulates lipid metabolism and plays a role in fatty liver diseases although conflicting observations have been reported. Nrf2 KO mice have exacerbated, whereas Keap1 knockdown mice have less hepatic steatosis when these mice are fed with a methionine- and choline-deficient diet.
      • Sugimoto H.
      • Okada K.
      • Shoda J.
      • Warabi E.
      • Ishige K.
      • Ueda T.
      • Taguchi K.
      • Yanagawa T.
      • Nakahara A.
      • Hyodo I.
      • Ishii T.
      • Yamamoto M.
      Deletion of nuclear factor-E2-related factor-2 leads to rapid onset and progression of nutritional steatohepatitis in mice.
      • Zhang Y.K.
      • Yeager R.L.
      • Tanaka Y.
      • Klaassen C.D.
      Enhanced expression of Nrf2 in mice attenuates the fatty liver produced by a methionine- and choline-deficient diet.
      Moreover, pharmacologic activation of Nrf2 attenuates diet-induced hepatic steatosis.
      • Shin S.
      • Wakabayashi J.
      • Yates M.S.
      • Wakabayashi N.
      • Dolan P.M.
      • Aja S.
      • Liby K.T.
      • Sporn M.B.
      • Yamamoto M.
      • Kensler T.W.
      Role of Nrf2 in prevention of high-fat diet-induced obesity by synthetic triterpenoid CDDO-imidazolide.
      In contrast, Nrf2 KO mice are partially protected against diet-induced fatty liver and obesity likely due to the induction of FGF21 in the adipose and liver tissues, suggesting that Nrf2 is a negative regulator of FGF21.
      • Chartoumpekis D.V.
      • Ziros P.G.
      • Psyrogiannis A.I.
      • Papavassiliou A.G.
      • Kyriazopoulou V.E.
      • Sykiotis G.P.
      • Habeos I.G.
      Nrf2 represses FGF21 during long-term high-fat diet-induced obesity in mice.
      These controversial results are likely due to different diets and feeding periods in these studies, but it implies a very complex role of Nrf2 in regulating lipid metabolism. Interestingly, Keap1 knockdown mice show decreased hepatic accumulation of LDs when these mice are fasted for 24 hours, which is similar to the L-Atg5 KO mice in our present study. Both Keap1 knockdown mice and L-Atg5 KO mice all have persistent Nrf2 activation, implying that activation of Nrf2 may contribute to the lack of hepatic LD accumulation after fasting. However, there were also some fundamental differences among Keap1 knockdown mice and L-Atg5 KO mice in response to fasting. Xu et al
      • Xu J.
      • Donepudi A.C.
      • Moscovitz J.E.
      • Slitt A.L.
      Keap1-knockdown decreases fasting-induced fatty liver via altered lipid metabolism and decreased fatty acid mobilization from adipose tissue.
      found that the gene expression of de novo lipogenesis and FA transporters decreased in Keap1 knockdown mice. However, the expression of de novo lipogenesis genes decreased in both WT and L-Atg5 KO mice, suggesting that de novo lipogenesis may not be important for the accumulation of LDs in response to fasting. In contrast to the Keap1 knockdown mice, the hepatic levels of CD36 are much higher in the L-Atg5 KO mice even at the basal fed condition. Despite that both Keap1 knockdown and L-Atg5 KO mice have persistent Nrf2 activation, no obvious liver injury and other pathogenesis are observed in Keap1 knockdown/KO mice. Therefore, some of the differences between Keap1 knockdown and L-Atg5 KO mice may be due to the secondary effects in the L-Atg5 mouse livers as results of lack of autophagy in addition to Nrf2 activation.
      Notably, our results that decreased hepatic LD accumulation in L-Atg5 KO mice after fasting are consistent with some previous reports that used liver-specific Atg7 or FIP200 KO mice.
      • Ma D.
      • Molusky M.M.
      • Song J.
      • Hu C.R.
      • Fang F.
      • Rui C.
      • Mathew A.V.
      • Pennathur S.
      • Liu F.
      • Cheng J.X.
      • Guan J.L.
      • Lin J.D.
      Autophagy deficiency by hepatic FIP200 deletion uncouples steatosis from liver injury in NAFLD.
      • Kim K.H.
      • Jeong Y.T.
      • Oh H.
      • Kim S.H.
      • Cho J.M.
      • Kim Y.N.
      • Kim S.S.
      • Kim D.H.
      • Hur K.Y.
      • Kim H.K.
      • Ko T.
      • Han J.
      • Kim H.L.
      • Kim J.
      • Back S.H.
      • Komatsu M.
      • Chen H.
      • Chan D.C.
      • Konishi M.
      • Itoh N.
      • Choi C.S.
      • Lee M.S.
      Autophagy deficiency leads to protection from obesity and insulin resistance by inducing Fgf21 as a mitokine.
      Note that liver-specific Atg7 and FIP200 KO mice have similar liver pathogenesis, including hepatocyte death, inflammation, fibrosis, and benign liver tumors, similar to L-Atg5 KO mice. Although Nrf2 activation has not been examined in L-FIP200 KO mice, further deletion of Nrf2 also corrected the liver pathologic conditions in L-Atg7 KO mice. Therefore, the mechanisms that contribute to the lack of adaptive hepatic biogenesis of LD in L-Atg5 mice in response to fasting that we identified in the present study may be generally true for other mice that also have defective hepatic autophagy.

      Conclusions

      Here it was shown that hepatic autophagy-deficient mice (L-Atg5 KO mice) had impaired adaptation to fasting-induced hepatic steatosis. The secondary effects (liver injury and Nrf2 activation) rather than the direct lack of hepatic autophagy may be the key contributors for the decreased hepatic LD accumulation in autophagy-deficient livers in response to fasting.

      Acknowledgment

      We thank Dr. Noboru Mizushima (University of Tokyo, Tokyo, Japan) for providing Atg5–/– MEFs and Atg5flox/flox mice, and Dr. Masaaki Komatsu (Niigata University School of Medicine, Niigata, Japan) for providing Atg7−/− MEFs.

      Supplemental Data

      • Supplemental Figure S1

        The body weight and liver weight changes in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice. A: WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were fed with a normal chow diet, and the body weights were analyzed at 2, 4, 6, and 9 months. Two-month–old WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed), fasted overnight (Fasted), or refed with chow diet for 6 hours (Refed). B: The liver-to-body weight ratios were analyzed. Data are expressed as means ± SEM. n = 3 to 6 mice. P < 0.05 between different genotypes, determined by one-way analysis of variance analysis.

      • Supplemental Figure S2

        Changes of hepatic autophagy and nuclear factor-like (Nrf)2 signaling in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/Nrf2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (FD), fasted overnight (FT), or refed with chow diet for 6 hours (RF). Total liver lysates were prepared, and pooled liver lysates were subjected to Western blot analysis. n = 3 to 6 mice. GCLC, glutamate-cysteine ligase catalytic unit; GCLM, glutamate-cysteine ligase modifier unit; LC3, light chain 3; NQO1, NAD(P)H quinone dehydrogenase 1.

      • Supplemental Figure S3

        Increased hepatic autophagosome numbers in fasted wild-type (WT) mice. WT mice were fed with a normal chow diet (Fed) or fasted overnight. Mice were sacrificed after overnight fasting, and liver tissues were fixed and processed for electron microscopic (EM) studies. A: Representative EM images. The lower panels are enlarged images from the boxed areas in the upper panel images. Arrow denotes an autophagosome enwrapped with a mitochondrion. Arrowhead denotes the membrane of the vacuole. B: The number of autophagosomes (AVs) were quantified from panel A. Data are expressed as means ± SEM (B). n = 10 or more different images. P < 0.05, determined by t-test. Scale bar = 2 μm. LD, lipid droplet; M, mitochondria; N, nucleus; V, vacuole.

      • Supplemental Figure S4

        Changes of hepatic lipid metabolism proteins in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (FD), fasted overnight (FT), or refed with chow diet for 6 hours (RF). Total liver lysates were prepared and pooled liver lysates were subjected to Western blot analysis. n = 3 to 6 mice. ACC, acetyl-CoA carboxylase; ATGL, adipose triglyceride lipase; FASn, fatty acid synthase; pHSL, phosphorylated hormone-sensitive lipase; mSREBP1, matured sterol regulatory element–binding transcription factor 1; SREBP1, sterol regulatory element–binding transcription factor 1.

      • Supplemental Figure S5

        Changes of the expression of cholesterol synthesis genes in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed), fasted overnight (Fasted), or refed with chow diet for 6 hours (Refed). Hepatic mRNA was extracted, followed by real-time quantitative PCR for hepatic cholesterol synthesis genes. Data are expressed as means ± SEM. n = 3 to 6 mice. P < 0.05 between different treatments, determined by one-way analysis of variance analysis.

      • Supplemental Figure S6

        Changes of the expression of cholesterol transporter genes in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed), fasted overnight (Fasted), or refed with chow diet for 6 hours (Refed). Hepatic mRNA was extracted, followed by real-time quantitative PCR for hepatic cholesterol transporter genes. Data are expressed as means ± SEM. n = 3 to 6 mice. P < 0.05 between different treatments; P < 0.05 between different genotypes; determined by one-way analysis of variance analysis.

      • Supplemental Figure S7

        Changes of hepatic triglyceride (TG) secretion and the expression of MTTP in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. A: WT and L-Atg5 KO mice were fasted overnight and injected with 300 mg/kg tyloxapol via tail vein. Blood was collected for TG measurement at different time points. B: WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed), fasted overnight (Fasted), or refed with chow diet for 6 hours (Refed). Hepatic mRNA was extracted, followed by real-time quantitative PCR for hepatic expression of MTTP. Data are expressed as means ± SEM. n = 3 to 6 mice. P < 0.05 between different treatments; P < 0.05 between different genotypes; determined by one-way analysis of variance analysis.

      • Supplemental Figure S8

        Changes of the expression of Fgf21 and Pgc1a in wild-type (WT), hepatic autophagy-deficient (L-Atg)5 knockout (KO), L-Atg5/nuclear factor-like (Nrf)2 double-knockout (DKO), and Nrf2 KO mice in response to nutrient changes. WT, L-Atg5 KO, L-Atg5/Nrf2 DKO, and Nrf2 KO mice were either fed with a normal chow diet (Fed), fasted overnight (Fasted), or refed with chow diet for 6 hours (Refed). A and B: Hepatic mRNA was extracted, followed by real-time quantitative PCR for hepatic Fgf21 (A) and Pgc1a (B). Data are expressed as means ± SEM. n = 3 to 6 mice. P < 0.05 between different treatments; P < 0.05 between different genotypes; determined by one-way analysis of variance analysis.

      • Supplemental Figure S9

        Lipid droplet (LD) formation in cultured wild-type (WT) and autophagy (Atg)5 and Atg7 knockout (KO) mouse embryonic fibroblasts. WT, Atg5 KO, and Atg7 KO mice were loaded with either bovine serum albumin or 250 μmol/L oleic acid (OA) for 6 hours. Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline. Cells were further stained with 1 μmol/L Bodipy 493/503 and 1 μmol/L Hoechst 33342, followed by fluorescence microscopy. A: Representative images of Bodipy and Hoechst 33342 staining are shown. B: Total cell lysates were subjected to Western blot analysis. Scale bar = 20 μm.LC3, light chain 3.

      References

        • Pol A.
        • Gross S.P.
        • Parton R.G.
        Review: biogenesis of the multifunctional lipid droplet: lipids, proteins, and sites.
        J Cell Biol. 2014; 204: 635-646
        • Yang H.
        • Galea A.
        • Sytnyk V.
        • Crossley M.
        Controlling the size of lipid droplets: lipid and protein factors.
        Curr Opin Cell Biol. 2012; 24: 509-516
        • Zechner R.
        • Madeo F.
        • Kratky D.
        Cytosolic lipolysis and lipophagy: two sides of the same coin.
        Nat Rev Mol Cell Biol. 2017; 18: 671-684
        • Listenberger L.L.
        • Han X.
        • Lewis S.E.
        • Cases S.
        • Farese Jr., R.V.
        • Ory D.S.
        • Schaffer J.E.
        Triglyceride accumulation protects against fatty acid-induced lipotoxicity.
        Proc Natl Acad Sci U S A. 2003; 100: 3077-3082
        • Zechner R.
        • Strauss J.G.
        • Haemmerle G.
        • Lass A.
        • Zimmermann R.
        Lipolysis: pathway under construction.
        Curr Opin Lipidol. 2005; 16: 333-340
        • Badman M.K.
        • Pissios P.
        • Kennedy A.R.
        • Koukos G.
        • Flier J.S.
        • Maratos-Flier E.
        Hepatic fibroblast growth factor 21 is regulated by PPARalpha and is a key mediator of hepatic lipid metabolism in ketotic states.
        Cell Metab. 2007; 5: 426-437
        • Inagaki T.
        • Dutchak P.
        • Zhao G.
        • Ding X.
        • Gautron L.
        • Parameswara V.
        • Li Y.
        • Goetz R.
        • Mohammadi M.
        • Esser V.
        • Elmquist J.K.
        • Gerard R.D.
        • Burgess S.C.
        • Hammer R.E.
        • Mangelsdorf D.J.
        • Kliewer S.A.
        Endocrine regulation of the fasting response by PPARalpha-mediated induction of fibroblast growth factor 21.
        Cell Metab. 2007; 5: 415-425
        • Mizushima N.
        • Levine B.
        • Cuervo A.M.
        • Klionsky D.J.
        Autophagy fights disease through cellular self-digestion.
        Nature. 2008; 451: 1069-1075
        • Madrigal-Matute J.
        • Cuervo A.M.
        Regulation of liver metabolism by autophagy.
        Gastroenterology. 2016; 150: 328-339
        • Levine B.
        • Klionsky D.J.
        Development by self-digestion: molecular mechanisms and biological functions of autophagy.
        Dev Cell. 2004; 6: 463-477
        • Lum J.J.
        • DeBerardinis R.J.
        • Thompson C.B.
        Autophagy in metazoans: cell survival in the land of plenty.
        Nat Rev Mol Cell Biol. 2005; 6: 439-448
        • Singh R.
        • Kaushik S.
        • Wang Y.
        • Xiang Y.
        • Novak I.
        • Komatsu M.
        • Tanaka K.
        • Cuervo A.M.
        • Czaja M.J.
        Autophagy regulates lipid metabolism.
        Nature. 2009; 458: 1131-1135
        • Yang L.
        • Li P.
        • Fu S.
        • Calay E.S.
        • Hotamisligil G.S.
        Defective hepatic autophagy in obesity promotes ER stress and causes insulin resistance.
        Cell Metab. 2010; 11: 467-478
        • Czaja M.J.
        • Ding W.X.
        • Donohue Jr., T.M.
        • Friedman S.L.
        • Kim J.S.
        • Komatsu M.
        • Lemasters J.J.
        • Lemoine A.
        • Lin J.D.
        • Ou J.H.
        • Perlmutter D.H.
        • Randall G.
        • Ray R.B.
        • Tsung A.
        • Yin X.M.
        Functions of autophagy in normal and diseased liver.
        Autophagy. 2013; 9: 1131-1158
        • Ding W.X.
        • Li M.
        • Chen X.
        • Ni H.M.
        • Lin C.W.
        • Gao W.
        • Lu B.
        • Stolz D.B.
        • Clemens D.L.
        • Yin X.M.
        Autophagy reduces acute ethanol-induced hepatotoxicity and steatosis in mice.
        Gastroenterology. 2010; 139: 1740-1752
        • Lin C.W.
        • Zhang H.
        • Li M.
        • Xiong X.
        • Chen X.
        • Chen X.
        • Dong X.C.
        • Yin X.M.
        Pharmacological promotion of autophagy alleviates steatosis and injury in alcoholic and non-alcoholic fatty liver conditions in mice.
        J Hepatol. 2013; 58: 993-999
        • Rambold A.S.
        • Cohen S.
        • Lippincott-Schwartz J.
        Fatty acid trafficking in starved cells: regulation by lipid droplet lipolysis, autophagy, and mitochondrial fusion dynamics.
        Dev Cell. 2015; 32: 678-692
        • Li Y.
        • Zong W.X.
        • Ding W.X.
        Recycling the danger via lipid droplet biogenesis after autophagy.
        Autophagy. 2017; 13: 1995-1997
        • Nguyen T.B.
        • Louie S.M.
        • Daniele J.R.
        • Tran Q.
        • Dillin A.
        • Zoncu R.
        • Nomura D.K.
        • Olzmann J.A.
        DGAT1-dependent lipid droplet biogenesis protects mitochondrial function during starvation-induced autophagy.
        Dev Cell. 2017; 42: 9-21.e5
        • Ma D.
        • Molusky M.M.
        • Song J.
        • Hu C.R.
        • Fang F.
        • Rui C.
        • Mathew A.V.
        • Pennathur S.
        • Liu F.
        • Cheng J.X.
        • Guan J.L.
        • Lin J.D.
        Autophagy deficiency by hepatic FIP200 deletion uncouples steatosis from liver injury in NAFLD.
        Mol Endocrinol. 2013; 27: 1643-1654
        • Kim K.H.
        • Jeong Y.T.
        • Oh H.
        • Kim S.H.
        • Cho J.M.
        • Kim Y.N.
        • Kim S.S.
        • Kim D.H.
        • Hur K.Y.
        • Kim H.K.
        • Ko T.
        • Han J.
        • Kim H.L.
        • Kim J.
        • Back S.H.
        • Komatsu M.
        • Chen H.
        • Chan D.C.
        • Konishi M.
        • Itoh N.
        • Choi C.S.
        • Lee M.S.
        Autophagy deficiency leads to protection from obesity and insulin resistance by inducing Fgf21 as a mitokine.
        Nat Med. 2013; 19: 83-92
        • Taguchi K.
        • Motohashi H.
        • Yamamoto M.
        Molecular mechanisms of the Keap1-Nrf2 pathway in stress response and cancer evolution.
        Genes Cells. 2011; 16: 123-140
        • Lau A.
        • Villeneuve N.F.
        • Sun Z.
        • Wong P.K.
        • Zhang D.D.
        Dual roles of Nrf2 in cancer.
        Pharmacol Res. 2008; 58: 262-270
        • Dodson M.
        • Zhang D.D.
        Non-canonical activation of NRF2: new insights and its relevance to disease.
        Curr Pathobiol Rep. 2017; 5: 171-176
        • Bjorkoy G.
        • Lamark T.
        • Brech A.
        • Outzen H.
        • Perander M.
        • Overvatn A.
        • Stenmark H.
        • Johansen T.
        p62/SQSTM1 forms protein aggregates degraded by autophagy and has a protective effect on huntingtin-induced cell death.
        J Cell Biol. 2005; 171: 603-614
        • Ni H.M.
        • Boggess N.
        • McGill M.R.
        • Lebofsky M.
        • Borude P.
        • Apte U.
        • Jaeschke H.
        • Ding W.X.
        Liver-specific loss of Atg5 causes persistent activation of Nrf2 and protects against acetaminophen-induced liver injury.
        Toxicol Sci. 2012; 127: 438-450
        • Ni H.M.
        • Woolbright B.L.
        • Williams J.
        • Copple B.
        • Cui W.
        • Luyendyk J.P.
        • Jaeschke H.
        • Ding W.X.
        Nrf2 promotes the development of fibrosis and tumorigenesis in mice with defective hepatic autophagy.
        J Hepatol. 2014; 61: 617-625
        • Komatsu M.
        • Kurokawa H.
        • Waguri S.
        • Taguchi K.
        • Kobayashi A.
        • Ichimura Y.
        • Sou Y.S.
        • Ueno I.
        • Sakamoto A.
        • Tong K.I.
        • Kim M.
        • Nishito Y.
        • Iemura S.
        • Natsume T.
        • Ueno T.
        • Kominami E.
        • Motohashi H.
        • Tanaka K.
        • Yamamoto M.
        The selective autophagy substrate p62 activates the stress responsive transcription factor Nrf2 through inactivation of Keap1.
        Nat Cell Biol. 2010; 12: 213-223
        • Lau A.
        • Wang X.J.
        • Zhao F.
        • Villeneuve N.F.
        • Wu T.
        • Jiang T.
        • Sun Z.
        • White E.
        • Zhang D.D.
        A noncanonical mechanism of Nrf2 activation by autophagy deficiency: direct interaction between Keap1 and p62.
        Mol Cell Biol. 2010; 30: 3275-3285
        • Komatsu M.
        • Waguri S.
        • Koike M.
        • Sou Y.S.
        • Ueno T.
        • Hara T.
        • Mizushima N.
        • Iwata J.
        • Ezaki J.
        • Murata S.
        • Hamazaki J.
        • Nishito Y.
        • Iemura S.
        • Natsume T.
        • Yanagawa T.
        • Uwayama J.
        • Warabi E.
        • Yoshida H.
        • Ishii T.
        • Kobayashi A.
        • Yamamoto M.
        • Yue Z.
        • Uchiyama Y.
        • Kominami E.
        • Tanaka K.
        Homeostatic levels of p62 control cytoplasmic inclusion body formation in autophagy-deficient mice.
        Cell. 2007; 131: 1149-1163
        • Inami Y.
        • Waguri S.
        • Sakamoto A.
        • Kouno T.
        • Nakada K.
        • Hino O.
        • Watanabe S.
        • Ando J.
        • Iwadate M.
        • Yamamoto M.
        • Lee M.S.
        • Tanaka K.
        • Komatsu M.
        Persistent activation of Nrf2 through p62 in hepatocellular carcinoma cells.
        J Cell Biol. 2011; 193: 275-284
        • Xu J.
        • Donepudi A.C.
        • Moscovitz J.E.
        • Slitt A.L.
        Keap1-knockdown decreases fasting-induced fatty liver via altered lipid metabolism and decreased fatty acid mobilization from adipose tissue.
        PLoS One. 2013; 8: e79841
        • Chartoumpekis D.V.
        • Ziros P.G.
        • Psyrogiannis A.I.
        • Papavassiliou A.G.
        • Kyriazopoulou V.E.
        • Sykiotis G.P.
        • Habeos I.G.
        Nrf2 represses FGF21 during long-term high-fat diet-induced obesity in mice.
        Diabetes. 2011; 60: 2465-2473
        • Ramadori P.
        • Drescher H.
        • Erschfeld S.
        • Schumacher F.
        • Berger C.
        • Fragoulis A.
        • Schenkel J.
        • Kensler T.W.
        • Wruck C.J.
        • Trautwein C.
        • Kroy D.C.
        • Streetz K.L.
        Hepatocyte-specific Keap1 deletion reduces liver steatosis but not inflammation during non-alcoholic steatohepatitis development.
        Free Radic Biol Med. 2016; 91: 114-126
        • Sugimoto H.
        • Okada K.
        • Shoda J.
        • Warabi E.
        • Ishige K.
        • Ueda T.
        • Taguchi K.
        • Yanagawa T.
        • Nakahara A.
        • Hyodo I.
        • Ishii T.
        • Yamamoto M.
        Deletion of nuclear factor-E2-related factor-2 leads to rapid onset and progression of nutritional steatohepatitis in mice.
        Am J Physiol Gastrointest Liver Physiol. 2010; 298: G283-G294
        • Ding W.X.
        • Ni H.M.
        • DiFrancesca D.
        • Stolz D.B.
        • Yin X.M.
        Bid-dependent generation of oxygen radicals promotes death receptor activation-induced apoptosis in murine hepatocytes.
        Hepatology. 2004; 40: 403-413
        • Mei S.
        • Ni H.M.
        • Manley S.
        • Bockus A.
        • Kassel K.M.
        • Luyendyk J.P.
        • Copple B.L.
        • Ding W.X.
        Differential roles of unsaturated and saturated fatty acids on autophagy and apoptosis in hepatocytes.
        J Pharmacol Exp Ther. 2011; 339: 487-498
        • Ding W.X.
        • Ni H.M.
        • Gao W.
        • Chen X.
        • Kang J.H.
        • Stolz D.B.
        • Liu J.
        • Yin X.M.
        Oncogenic transformation confers a selective susceptibility to the combined suppression of the proteasome and autophagy.
        Mol Cancer Ther. 2009; 8: 2036-2045
        • Hara T.
        • Nakamura K.
        • Matsui M.
        • Yamamoto A.
        • Nakahara Y.
        • Suzuki-Migishima R.
        • Yokoyama M.
        • Mishima K.
        • Saito I.
        • Okano H.
        • Mizushima N.
        Suppression of basal autophagy in neural cells causes neurodegenerative disease in mice.
        Nature. 2006; 441: 885-889
        • Williams J.A.
        • Ni H.M.
        • Ding Y.
        • Ding W.X.
        Parkin regulates mitophagy and mitochondrial function to protect against alcohol-induced liver injury and steatosis in mice.
        Am J Physiol Gastrointest Liver Physiol. 2015; 309: G324-G340
        • Shibata M.
        • Yoshimura K.
        • Furuya N.
        • Koike M.
        • Ueno T.
        • Komatsu M.
        • Arai H.
        • Tanaka K.
        • Kominami E.
        • Uchiyama Y.
        The MAP1-LC3 conjugation system is involved in lipid droplet formation.
        Biochem Biophys Res Commun. 2009; 382: 419-423
        • Jain A.
        • Lamark T.
        • Sjottem E.
        • Larsen K.B.
        • Awuh J.A.
        • Overvatn A.
        • McMahon M.
        • Hayes J.D.
        • Johansen T.
        p62/SQSTM1 is a target gene for transcription factor NRF2 and creates a positive feedback loop by inducing antioxidant response element-driven gene transcription.
        J Biol Chem. 2010; 285: 22576-22591
        • Pajares M.
        • Jimenez-Moreno N.
        • Garcia-Yague A.J.
        • Escoll M.
        • de Ceballos M.L.
        • Van Leuven F.
        • Rabano A.
        • Yamamoto M.
        • Rojo A.I.
        • Cuadrado A.
        Transcription factor NFE2L2/NRF2 is a regulator of macroautophagy genes.
        Autophagy. 2016; 12: 1902-1916
        • Zhang Y.K.
        • Yeager R.L.
        • Tanaka Y.
        • Klaassen C.D.
        Enhanced expression of Nrf2 in mice attenuates the fatty liver produced by a methionine- and choline-deficient diet.
        Toxicol Appl Pharmacol. 2010; 245: 326-334
        • Shin S.
        • Wakabayashi J.
        • Yates M.S.
        • Wakabayashi N.
        • Dolan P.M.
        • Aja S.
        • Liby K.T.
        • Sporn M.B.
        • Yamamoto M.
        • Kensler T.W.
        Role of Nrf2 in prevention of high-fat diet-induced obesity by synthetic triterpenoid CDDO-imidazolide.
        Eur J Pharmacol. 2009; 620: 138-144