If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Metformin-induced TTP mediates communication between Kupffer cells and hepatocytes to alleviate hepatic steatosis by regulating lipophagy and necroptosis
Department of Anatomy and Convergence Medical Science, Institute of Health Sciences, College of Medicine, Gyeongsang National University, Jinju 52727, Republic of Korea
Department of Anatomy and Convergence Medical Science, Institute of Health Sciences, College of Medicine, Gyeongsang National University, Jinju 52727, Republic of Korea
Department of Anatomy and Convergence Medical Science, Institute of Health Sciences, College of Medicine, Gyeongsang National University, Jinju 52727, Republic of Korea
Department of Molecular Medicine and Biopharmaceutical Sciences, Graduate School of Convergence Science and Technology, Seoul National University, Seoul 08826, Republic of KoreaCancer Research Institute, Seoul National University, Seoul 03080, Republic of Korea
Metformin induces TTP activation via the AMPK-Sirt1 axis.
•
TTP inhibits TNF-α production in Kupffer cells, to reduce hepatocyte necroptosis.
•
TTP activation increases lipophagy by destabilizing Rheb mRNA in hepatocytes.
•
Combined effects of TTP in hepatocytes and Kupffer cells reduce hepatic steatosis.
Abstract
Objective
Emerging evidence suggests that crosstalk between Kupffer cells (KCs) and hepatocytes protects against non-alcoholic fatty liver disease (NAFLD). However, the underlying mechanisms that lead to the reduction of steatosis in NAFLD remain obscure.
Methods
Ttp+/+ and Ttp−/− mice were fed with a high-fat diet. Hepatic steatosis was analyzed by Nile Red staining and measurement of inflammatory cytokines. Lipid accumulation and cell death were evaluated in co-culture systems with primary hepatocytes and KCs derived from either Ttp+/+ or Ttp−/− mice.
Results
Tristetraprolin (TTP), an mRNA binding protein, was essential for the protective effects of metformin in NAFLD. Metformin activated TTP via the AMPK-Sirt1 pathway in hepatocytes and KCs. TTP inhibited TNF-α production in KCs, which in turn decreased hepatocyte necroptosis. Downregulation of Rheb expression by TTP promoted hepatocyte lipophagy via mTORC1 inhibition and increased nuclear translocation of transcription factor-EB (TFEB). Consistently, TTP-deficient NAFLD mice failed to respond to metformin with respect to alleviation of hepatic steatosis, protection of hepatocyte necroptosis, or induction of lipophagy.
Conclusions
TTP, which is essential for the protective effects of metformin, may represent a novel primary therapeutic target in NAFLD.
Non-alcoholic fatty liver disease (NAFLD), a spectrum of conditions ranging in severity from simple steatosis to steatohepatitis, may progress to cirrhosis, and hepatocellular carcinoma (HCC) [
The diagnosis and management of non-alcoholic fatty liver disease: practice guideline by the American Association for the Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association.
]. The regulation of lipid metabolism by metformin, the first-line anti-diabetic drug, ameliorates excessive fat accumulation and the development of age related-metabolic disorders [
], and involves recruitment of receptor-interacting protein-3 (RIPK3) and phosphorylation of mixed lineage kinase-like protein (MLKL), resulting in cell swelling and membrane rupture [
Tristetraprolin (TTP, encoded by Ttp, also known as Zfp36) is an AU-rich element (ARE)-binding protein known to cause the degradation of its target genes by recruiting the deadenylase and the decapping complex [
Tristetraprolin (TTP)-14-3-3 complex formation protects TTP from dephosphorylation by protein phosphatase 2a and stabilizes tumor necrosis factor-alpha mRNA.
In this study, we sought to determine, using a mouse model of NAFLD and primary hepatocytes, whether metformin treatment, relative to control treatments, can alleviate hepatocyte LD accumulation and protect against hepatocyte cell death. We found that metformin can induce TTP activation via the AMPK/Sirt1 pathway. Using TTP deficient mice, we found that TTP was required to decrease LDs in hepatocytes, via activation of lipophagy. The mechanism by which metformin-induced TTP activates lipophagy was found to be dependent on mTORC1 inhibition via destabilization of Rheb, which promotes TFEB nuclear translocation. Furthermore, we demonstrate that TTP suppresses hepatocyte necroptosis via TNF-α degradation in Kupffer cells (KCs).
In conclusion our data demonstrate that TTP is essential for the reduction of hepatocyte necroptosis, the induction of lipophagy, and the reduction of hepatic steatosis in response to metformin, in models of NAFLD. We also propose a mechanism by which TTP-mediated downregulation of TNF-α production by KCs, and subsequent reduction of hepatocyte necroptosis, reduces the pathology of NAFLD.
2. Materials and methods
2.1 Reagents
Metformin, sodium oleate, chloroquine, compound C, and EX527 were purchased from Sigma-Aldrich (St Louis, MO, USA). Torin 1 was from Tocris Biotechne (Minneapolis, MN, USA).
2.2 Animals
Ttp−/− mice were provided by Dr. Perry J. Blackshear (Laboratory of Signal Transduction, National Institute of Environmental Health Sciences, USA). Animals were maintained in a specific pathogen-free facility with a 12 h light-dark cycle. Animal studies were approved by the University of Ulsan Animal Care and Use Committee (Reference number HTC-19-020). To construct a high fat diet (HFD; Research Diets, Inc., New Brunswick, NJ, USA)-induced liver disease model, Ttp+/+ (n = 5) and Ttp−/− (n = 5) mice were fed 60 % HFD for 8 weeks. Starting at 4 weeks of HFD, the mice were given metformin (200 mg/kg, oral administration) every day for 4 weeks (control group (normal diet); HFD group; HFD + met group; met group (normal diet+met). Then, these mice were used for various analyses. For the induction of non-alcoholic fatty liver disease (NAFLD), mice were fed with a methionine/choline-deficient (MCD) diet (Research Diets) for 2 weeks. Simultaneously, mice were treated daily with metformin (200 mg/kg, oral administration), and then, the liver and blood of mice were sampled ((control group (normal diet); MCD group; MCD + met group; met group (normal diet+Met), n = 5 per group).
2.3 Cell culture
The mouse liver cell line AML12 (ATCC, VA, USA) was cultured in DMEM/F12 (Gibco, Grand Island, USA) supplemented with 10 % fetal bovine serum (FBS), 1 % penicillin-streptomycin, insulin-transferrin‑sodium selenite media supplement, and dexamethasone (Sigma-Aldrich). Human embryonic kidney (HEK293) cells were maintained in DMEM medium containing 10 % FBS and 1 % penicillin-streptomycin. Cells were grown at 37 °C in humidified incubators containing an atmosphere of 5 % CO2.
2.4 Isolation of primary hepatocytes and primary Kupffer cells
The inferior vena cava was cannulated and perfused at 4 ml/min. The portal vein was sectioned to allow flow through the liver. The liver was first perfused with Ca2+ and Mg2+-free Hank's buffered salt solution (HBSS, Gibco, NY, USA), followed by perfusion with 0.02 % collagenase type IV in Williams' Medium E (Gibco, NY, USA) for 3 min at 4 ml/min. The liver was dissected, and hepatocytes were isolated by mechanical dissection, filtered through a sterile 100 μm nylon cell strainer (BD Falcon, CA, USA). Non-parenchymal cells (NPCs) were separated from hepatocytes by differential centrifugation at 50g for 3 min. The supernatant was centrifuged at 300g for 5 min to obtain NPCs. NPCs were then suspended in PBS and layered onto a 50/25 % two-step Percoll gradient and centrifuged at 1200g for 20 min at 4 °C. KCs in the middle layer were collected and allowed to attach onto cell culture plates in DMEM supplemented with 10 % FBS and 1 % penicillin-streptomycin. Hepatocytes were mixed using 40 % Percoll (GE healthcare, WI, USA), followed by centrifugation at 150g for 7 min. The pellet was washed with DMEM supplemented with 10 % FBS and 1 % penicillin-streptomycin, and then hepatocytes and KCs were seeded at a 2:1 ratio onto chambered coverglass 2-well and 6-well plates. After 4 h, the medium was replaced with fresh medium.
2.5 Measurement of lipid by oil red O staining and BODIPY
Primary hepatocytes were grown in a 12-well plate. After reaching 70 % confluence, the cells were exposed to 200 μM sodium oleate for 18 h after pretreatment with or without 2 mM metformin for 12 h, followed by fixation with formalin solution, neutral buffered, 10 % (Sigma-Aldrich) for 30 min. Fixed cells were washed three times with PBS and stained with Oil Red O (Sigma-Aldrich) solution for 50 min at room temperature. Red-stained lipid droplets were subsequently observed with a light microscope. To quantify Oil Red O content levels, Oil Red O was eluted with 100 % isopropanol and the absorbance of samples were read at 520 nm on a spectrophotometer. To perform BODIPY (Thermo Scientific, Waltham, MA, USA) staining, AML12, primary hepatocytes, and co-cultured primary hepatocytes and KCs were grown in a confocal chamber and were incubated with 2 mM metformin for 12 h and then cultured in medium with 200 μM sodium oleate for an additional 18 h. Lipid droplets were stained with 2 μg/ml BODIPY for 1 h and then fixed with formalin solution for 30 min. The cells were washed three times with PBS followed by DAPI (Sigma-Aldrich) staining. Images of the cells were obtained using a confocal microscope (Olympus, Tokyo, Japan).
2.6 Flow cytometry
To measure purity of hepatocytes and Kupffer cells, freshly isolated hepatocytes were stained with APC anti-CD45 antibody and FITC anti-CD95 (Fas) antibody (BioLegend, CA, USA) for 20 min and then washed a time with ice cold FACS buffer (1 % BSA in PBS), and primary Kupffer cells were stained with APC anti-CD11b antibody and FITC anti-F4/80 antibody (BioLegend, CA, USA) for 20 min and then washed a time with ice cold FACS buffer. The fluorescence was assessed by using flow cytometry with a FACSCanto flow cytometry system (BD Bioscience, CA, USA), and data was analyzed by FlowJo V10 software.
2.7 RNA isolation and RT-PCR reverse transcription-polymerase chain reaction
Total RNA was extracted from AML12 cells, primary hepatocytes, and liver tissues using QIAzol Lysis reagent (QIAGEN, CA, USA). 2 μg of total RNA was used to synthesize cDNA by using M-MLV reverse transcriptase (Promega, Madison, WI, USA). The synthesized cDNA was subject to the PCR-based amplification using Taq polymerase (Bioneer, Daejeon, Korea). The following primers were used: mouse GAPDH (f-aggccggtgctgagtatgtc, r-tgcctgcttcaccttct), mouse TNF-α (f-agcccacgtcgtagcaaaccaccaa, r-acacccattcccttcacagagcaat), mouse IL-1β (f-ctgtgtctttcccgtggacc, r-cagctcatatgggtccgaca), mouse CXCL1 (f-gctgggattcacctcaagaa, r-gcacttcttttcgcacaaca), mouse TTP (f-ctctgccatctacgagagcc, r-gatggagtccgagtttatgttcc), mouse Rheb (f-atgcctcagtccaagtcccggaag, r-tcacatcaccgagcacgaaga), mouse Sirt1 (f-gcaacagcatcttgcctgat, r-gtgctactggtctcactt), and mouse 18S (f-cagtgaaactgcgaatggct, r-tgccttccttggatgtggta). qRT-PCR was conducted by SYBR Green Master Mix on an ABI 7500 Fast Real-Time PCR System (Applied Biosystems, CA, USA). The following qRT-PCR primers were mouse GAPDH (f-cggcctcaccccatttg, r-gggaagcccatcaccatct), mouse ATGL (f-gccacagcgctggtcact, r-cctccttggacacctcaataatg), mouse HSL (f-aggcctcagtgtgaccgcca, r-gccccacgcaactctgggtc), mouse MCOLN1 (f-gcgcctatgacaccatcaa, r-tatcctggactgctcgat), mouse TPP1 (f-aagccaggcctacatactcaga, r-ccaagtgcttcctgcagtttaga), mouse LAMP1 (f-taatggccagcttctctgcctcctt, r-aggctggggtcagaaacattttctt), mouse CatB (f-ttagcgctctcacttccactacc, r-tgcttgctaccttcctctggtta), mouse CatD (f-aactgctggacatcgcttgct, r-cattcttcacgtaggtgctgga), mouse TFEB (f-gcgagagctaacagatgctga, r-ccggtcattgatgttgaacc), mouse Rheb (f-aagtcccggaagatcgcca, r-ggttggatcgtaggaatcaacaa), mouse IL-1β (f-tcgctcagggtcacaagaaa, r-atcagaggcaaggaggaaacac), and mouse TNF-α (f-agaccctcacactcagatcatcttc, r-ttgctacgacgtgggctaca).
2.8 Western blot
Harvested tissues and cells were lysed in RIPA buffer (Thermo Scientific, Waltham, MA) containing phosphatase inhibitor and protease inhibitors (Sigma-Aldrich), and then the protein concentration was measured using the BCA protein assay reagent (Pierce Biotechnology, Rockford, IL, USA). Proteins were subjected to SDS-PAGE and transferred to polyvinylidene difluoride membrane (Millipore, Burlington, MA, USA). The membrane was incubated overnight with primary antibody against TTP (1:2000, Sigma-Aldrich), p-P70S6 kinase (1:1000, Cell Signaling, Danvers, MA, USA), P70S6 kinase (1:1000, Cell Signaling), p-S6 ribosomal protein (1:1000, Cell Signaling), S6 ribosomal protein (1:1000, Cell Signaling), Rheb (1:1000, Santa Cruz, Santa Cruz, CA, USA), TFEB (1:1000, Bethyl Laboratories, Montgomery, TX, USA), PARP (1:2000, Cell Signaling), LC3B (1:2000, Novus Biologicals, CO, USA), p62 (1:10000), LAMP1 (1:1000, Abcam, Cambridge, MA, USA), p-AMPK (1:1000, Cell Signaling), AMPK (1:1000, Cell Signaling), SIRT1 (1:2000, Millipore, Germany), Beclin1 (1:2000, Cell Signaling), ATG7 (1:2000, Cell Signaling), p-MLKL (1:1000, Cell Signaling, MA, USA), MLKL (1:1000, Cell Signaling, MA, USA), α-tubulin (1:2000, Cell Signaling), and β-actin (1:2500, Thermo Scientific). Antibody binding was visualized by ECL chemiluminescence (Pierce Biotechnology) using Azure Biosystems C300 analyzer (Azure Biosystems, Dublin, CA, USA). Optical intensity of the bands was analyzed using ImageJ software (NIH, Bethesda, MD, USA).
2.9 Immunofluorescence
To detect the primary KCs using anti-F4/80 antibody (Cell Signaling, MA, USA), cells were fixed with 10 % neutral-buffered formalin solution for 20 min, permeabilized with 0.1 % Triton X-100 for 5min and blocked with 3 % bovine serum albumin (BSA) in PBS for 30min. Cells were then incubated with anti-F4/80 primary antibody overnight at 4°C before being washed three times in PBS. Alexa Fluor 594-conjugated secondary antibody was added for 2h in the dark and washed three timesin PBS, before samples were stained with DAPI. Images of the cells were captured using a confocal microscope (Olympus, Tokyo, Japan).
2.10 Sytox green staining
Co-cultured primary hepatocytes and Kupffer cells were pretreated with 2 mM metformin for 12 h and then treated with 200 μM sodium oleate in the presence or absence of 100 ng/ml LPS stimulation for 18 h. Cells were then stained with 125 nM Sytox Green (Invitrogen, CA, USA) and DAPI in cell culture media for 15 min at 37 °C to determine necrotic cell death by confocal microscopy (Olympus, Tokyo, Japan).
2.11 Enzyme-Linked Immunosorbent Assays (ELISA)
Conditioned medium from primary hepatocytes, KCs, and co-cultured primary hepatocytes and KCs was analyzed for TNF-α according to the manufacturer's recommendations (BioLegend, CA, USA).
2.12 Histological examination
Mice were anesthetized with intramuscular Zoletil (5 mg/kg, Virbac Laboratories, Carros, France) and perfused with 4 % paraformaldehyde in 0.1 M phosphate-buffered saline for tissue analysis. After 6 h of fixation, mouse livers were embedded in paraffin and then sliced into 5-μm tissue sections. The liver sections were stained with hematoxylin and eosin (H&E, Sigma-Aldrich), and the images of the sections were captured using a BX51 microscope (Olympus, Tokyo, Japan). Each liver specimen was analyzed for NAFLD activity score (NAS score), defined as the sum of steatosis (0–3), hepatocellular ballooning (0–2) and lobular inflammation (0–2). To determine the intracellular lipid droplets in hepatic steatosis, fixed livers were sequentially immersed in 15 %, 20 %, and 30 % sucrose at 4 °C until they sank. Frozen liver sections (10 μm) were stained with Nile red (Sigma-Aldrich). Sections were visualized under a BX51 microscope (Olympus). Digital images were captured and saved.
2.13 Measurement of triglycerides
Triglycerides (TGs) contents of liver were measured with the TG colorimetric assay kit (Cayman Chemical, Ann Arbor, MI, USA). Briefly, extracts of liver tissue were prepared by homogenization in 200 μl diluted Standard Diluents. After centrifugation, the supernatant was used for the assay.
2.14 Activity of aspartate transaminase (AST) and alanine transaminase (ALT)
To evaluate liver damage, AST and ALT were measured in the serum using the EnzyChrom AST assay kit and EnzyChrom ALT assay kit from BioAssay Systems (Hayward, CA, USA).
2.15 Transfection
pEGFP-N1-TFEB, pmRFP-LC3, and pBABE-puro mCherry-EGFP-LC3B were purchased from Addgene (Cambridge, MA, USA). The 3′UTR cDNA of human Rheb was PCR amplified from the cDNA of HepG2 cells using Taq polymerase (Bioneer, Daejeon, Korea). The following primers were used CTCGAGTTCTGCTGCAAAGCCTGAGGA, GCGGCCGCTGCAAAGCAAAACTATATTTTAT. Plasmid psiCHECK2-Rheb 3′UTR was generated by inserting the PCR product into XhoI and NotI sites of the psiCHECK2 plasmid (Promega, Madison, WI, USA). Cells were transfected with pEGFP-N1-TFEB, pBABE-puro mCherry-EGFP-LC3B, and pmRFP-LC3 using Lipofectamine™ 2000 (Invitrogen-Thermo Fischer, Waltham, MA) in accordance with the manufacturer's protocol. To knockdown the genes of Tfeb, Sirt1, and Rheb, cells were transfected with scramble siRNA (scRNA) (Ambion, Austin, TX, USA), Tfeb, Sirt1, and Rheb siRNA (Santa Cruz) using Lipofectamine™ 2000 (Invitrogen). After 24-48 h, plasmid- or siRNA- transfected cells were treated with the indicated drug and then analyzed by RT-PCR, western blot, and confocal imaging.
2.16 Subcellular fractionation
The nucleus and cytosol fraction were extracted from cells using a nuclear/cytosol fractionation kit (Biovision, Milpitas, CA, USA). Subcellular fractionation by harvested cells was performed according to the manufacturer's instructions. Briefly, cell pellets were resuspended in cytosolic extraction buffer A (CEB-A) and CEB-B. After centrifugation, the supernatants were kept as the cytoplasmic fractions. The nuclear pellet was resuspended in nuclear extraction buffer (NEB) and vortexed. This step was repeated every 10 min for five times. The resuspended nuclear pellet was centrifuged, and the supernatant was kept as a nuclear fraction.
2.17 Lysosome quantification
The lysosomal intensity was assessed using LysoTracker Red (Invitrogen) according to the manufacturer's instructions. After treatment, cells were stained with LysoTracker Red for 1 h followed by fixation using 10 % formalin and then stained with DAPI. Red fluorescence was identified using an Olympus FV1200 confocal microscope (Tokyo, Japan).
2.18 Luciferase assay
HEK293 cells were transfected with psiCHECK2-Rheb 3′UTR for 36 h. After treatment, cell lysates were prepared in lysis buffer and mixed with luciferase assay reagent (Promega, Madison, WI, USA). The chemiluminescence was detected using a SpectraMax iD3 (Molecular Devices, CA, USA). Renilla luciferase of psiCHECK2-Rheb 3′-UTR was normalized to firefly luciferase in each sample.
2.19 RNA sequencing
Total RNA from primary hepatocytes was isolated using a RNeasy mini kit (QIAGEN, Valencia, CA, USA). The cDNA library from 1 μg total RNA was prepared using the TruSeq stranded RNA library preparation kit according to the manufacturer's instructions (Illumina, San Diego, CA, USA). Libraries were sequenced on an Illumina HiSeq 2500 system to acquire 150-bp paired-end reads. To observe the differential expression of genes, the TMM-normalized FPKM matrix (FPKM = total exon fragments/[mapped reads (millions) × exon length (kb)]) was used to generate heat maps in an R programming environment.
2.20 Transmission electron microscopy
Autophagosomes and autolysosomes were detected with a JEM 1011CX electron microscope (Hitachi, Tokyo, Japan). Liver specimens were fixed with 2 % glutaraldehyde and 2 % paraformaldehyde buffered with 0.05 M sodium cacodylate after perfusion using cold PBS. Fixed liver tissues were embedded in Spurr's resin (Electron Microscopy Sciences, Hatfield, PA, USA) and sectioned at a thickness of 80 nm. After dehydration, sections were stained with uranyl acetate and lead citrate and then acquired from a randomly selected pool of 5–8 fields under each group.
2.21 QuantSeq RNA-seq data anaylsis
Differentially expressed genes (DEGs) between Ttp+/+ primary hepatocytes and Ttp−/− primary hepatocytes without or with metformin were estimated using the ExDEGA v.2.5.0. To know the relation of DEGs with various biological pathway, KEGG (Kyoto Encyclopedia of Genes and Genomes Pathway) database (http://www.genome.jp/kegg) was used.
2.22 Antibody-mediated TNF-α neutralization
For neutralizing TNF-α, Ttp+/+and Ttp−/− primary hepatocytes were treated with 25 μg/ml anti-TNF-α (XT3.11) antibody or isotype control (BxCell, NH, USA), and were simultaneously incubated with the supernatant of oleic acid (OA)-treated Kupffer cells for 30 h.
2.23 Statistical analysis
All values are expressed as mean ± SD. A Student's t-test was used to evaluate differences between the samples of interest and the corresponding controls. Differences between groups were assessed by one-way ANOVA with Tukey post hoc test. For statistical differences between groups in Ttp+/+and Ttp−/− genotypes, as well as in siRNA-transfected cells, data were evaluated by two-way ANOVA with Bonferroni post-tests. Data were analyzed and presented with GraphPad Prism software (San Diego, CA, USA).
3. Results
3.1 TTP deficient mice on HFD do not display hepatic lipid accumulation, whereas TTP deficient primary hepatocytes accumulate lipids in response to oleic acid treatment
Exposure to HFD causes liver damage by leading to the accumulation of lipids and inflammation, which in turn can promote NAFLD progression. To understand the mechanisms by which metformin can prevent NAFLD, we initially sought to investigate the role of TTP, a known modulator of metabolic pathways, in NAFLD progression. Ttp+/+ and Ttp−/− mice were fed on HFD for 8 weeks. In Ttp+/+ mice fed with HFD, the liver tissue exhibited lipid accumulation as determined by Nile Red staining (Fig. 1A ). Unexpectedly, Ttp−/− mice fed HFD did not exhibit lipid accumulation in liver tissue in vivo (Fig. 1A). In wild type mice, hepatic lipid accumulation induced by HFD was reduced by metformin treatment. To confirm these results, we used an alternate liver injury model induced by methionine-choline deficient (MCD) diet. The MCD diet also induced lipid accumulation in the liver of Ttp+/+, but not in Ttp−/− mice (Fig. S1A). The effects of metformin on the reduction of lipid accumulation in wild type mice fed with MCD were similar to that observed in HFD mice.
Fig. 1TTP deficiency impairs hepatocyte lipid accumulation in vivo in response to HFD, but not in cultured primary hepatocytes in vitro.
(A) Ttp+/+and Ttp−/− mice (n = 5) were fed 60 % HFD for 8 weeks or NCD with metformin (200 mg/kg). Lipid accumulation in liver tissues was analyzed by Nile Red staining. (B, C) Ttp+/+and Ttp−/− primary hepatocytes were exposed to 200 μM OA for 18 h after pretreatment with or without 2 mM metformin for 12 h, followed by staining with Oil Red O (B) and BODIPY (C). (D, E) Co-cultured cells were pretreated with 2 mM metformin for 12 h, followed by exposure to 200 μM OA for 18 h. After 30 h, cells were stained with BODIPY and were immunostained with anti-F4/80 antibody. Data are represented as mean ± SD; ***p < 0.001 and NS, not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
To understand the difference in the observed lipid accumulation between the livers of Ttp+/+ mice and Ttp−/− mice, we investigated lipid accumulation in primary hepatocytes and KCs derived from either Ttp+/+ or Ttp−/− mice. In contrast to the observations in vivo, lipid accumulation occurred in response to OA-treatment in primary hepatocytes (CD95+CD45−, see Fig. S1B for characterization) isolated from both Ttp+/+ and Ttp−/− mice (Fig. 1B and C) and in AML-12 cells and primary hepatocytes (Fig. S1D). In contrast, KCs (F4/80+CD11b+, see Fig. S1C for characterization) did not exhibit OA-induced lipid accumulation (Fig. S1E). Similar to the observed effects of metformin on lipid reduction in vivo, metformin suppressed lipid droplet (LD) formation by OA in wild type primary hepatocytes, however metformin failed to suppress OA-induced LD formation in Ttp−/− primary hepatocytes (Fig. 1B and C). TTP mRNA and protein levels in purified primary hepatocytes and Kupffer cells isolated from Ttp+/+and Ttp−/− mice are shown in Fig. S1F.
To understand the underlying mechanisms for altered hepatic lipid metabolism in the livers of Ttp−/− mice fed with HFD or MCD, and the absence of a lipid accumulation phenotype, we investigated OA-induced LD accumulation in co-cultures of hepatocytes and KCs. Similar to cultures of hepatocytes alone, when hepatocytes were co-cultured in the presence of wild type (Ttp+/+) KCs, OA treatment resulted in increased LD accumulation in either Ttp+/+ or Ttp−/− primary hepatocytes. In the presence of Ttp+/+ KCs, metformin treatment reduced OA-induced LD accumulation in Ttp+/+ hepatocytes, but not in Ttp−/− hepatocytes (Fig. 1D and E, left panels). These results suggest that Ttp+/+ KCs do not influence OA-induced LD accumulation and responsiveness to metformin in either Ttp+/+ or Ttp−/− hepatocytes. However, when co-cultured with Ttp−/− KCs, we could not detect OA-induced LD accumulation in either Ttp+/+ or Ttp−/− hepatocytes (Fig. 1D and E, right panels). These results recapitulate the in vivo phenotype observed in HFD- or MCD-fed Ttp−/− mice. Our results suggest that the interaction of KCs and hepatocytes may play a critical role in regulating lipid metabolism in the liver of mice fed with either HFD or MCD.
3.2 Ttp−/− Kupffer cells induce cell death in co-cultured OA-treated hepatocytes
Several reports suggest that KCs modulate hepatic lipid metabolism and hepatic steatosis [
]. We thus tested whether Ttp−/− KCs could induce cell death in co-cultured hepatocytes challenged with OA, by staining with Sytox green. Ttp+/+KCs did not induce cell death in either Ttp+/+ or Ttp−/− hepatocytes (Fig. 2A and B ). However, Ttp−/− KCs induced cell death when co-cultured with either OA-treated Ttp+/+ or Ttp−/− hepatocytes, which could not be rescued by metformin treatment (Fig. 2A and B). Similar results were obtained when KCs were treated with LPS and co-cultured with hepatocytes. In this case, cell death elicited by Ttp−/− KCs could not be reversed by metformin treatment, whereas, in the presence of LPS, wild type KCs slightly increased hepatocyte cell death in a metformin-sensitive fashion (Fig. 2C and D, S2A and S2B).
Fig. 2Co-cultured Ttp−/− Kupffer cells induce cell death in OA-treated hepatocytes.
(A, B) Co-cultured cells were pretreated with 2 mM metformin for 12 h, followed by exposure to 200 μM OA for 18 h. After 30 h, cells were stained with Sytox green. Ttp+/+ and Ttp−/− KCs were co-cultured with Ttp+/+ primary hepatocytes (A) or Ttp−/− primary hepatocytes (B). (C, D) Co-cultured primary hepatocytes and KCs were treated with 2 mM metformin prior to treatment with 200 μM OA and 100 ng/ml LPS. Cells were evaluated for the percentage of Sytox positive cells. (E) Primary hepatocytes and KCs were treated with 2 mM metformin before 200 μM OA. TNF-α levels were measured in cell culture supernatants using ELISA. (F, G) The levels of TNF-α in co-cultured hepatocytes and KCs were assessed by ELISA. (H, I) The levels of TNF-α were measured by ELISA in co-cultured primary hepatocytes and KCs treated with 2 mM metformin prior to treatment with 200 μM OA and 100 ng/ml LPS. Data are represented as mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001 and NS, not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
As a form of RCD, necroptosis can be initiated by TNF-α leading to MLKL phosphorylation and cell death (Sun et al., 1999). TTP deficiency is known to cause increases in TNF-α levels (Tayor et al., 1996). In single culture systems, OA treatment induced TNF-α secretion specifically from Ttp−/− KCs but not Ttp+/+ KCs. In contrast, OA treatment did not induce TNF-α secretion from Ttp+/+ or Ttp−/− hepatocytes, nor from Ttp+/+ KCs (Fig. 2E). In co-culture systems, high levels of TNF-α were detected from culture supernatants only when Ttp−/− KCs were co-cultured with either Ttp+/+ or Ttp−/− hepatocytes (Fig. 2F and G). Similar results were observed in OA-treated hepatocytes when co-cultured with KCs in the presence of LPS (Fig. 2H and I). In the presence of LPS, OA-induced TNF-α production was decreased by metformin in hepatocytes co-cultured with Ttp+/+ KCs, but not hepatocytes co-cultured with Ttp−/− KCs (Fig. 2H and I). These results are consistent with previous reports showing that metformin-induced TTP activation promotes TNF-α degradation [
] and that necroptosis is involved in a variety of acute and chronic types of liver disease (Nagy et al., 2016; Schwabe and Luedde, 2018). In addition, necroptosis is mediated by TNF-α [
]. To confirm whether hepatocyte cell death induced by Ttp−/− KCs is necroptosis, we measured the expression of the phosphorylated form of mixed lineage kinase domain-like protein (p-MLKL), a key mediator of necroptosis [
]. Ttp−/− KCs induced phosphorylation of MLKL in co-cultures of either OA-treated Ttp+/+ and Ttp−/− hepatocytes, while Ttp+/+ KCs did not (Fig. 3A ). Similar results were observed in OA-treated hepatocytes when co-cultured with LPS-stimulated KCs (Fig. 3B). Thus, we tested whether cytokines secreted from Ttp−/− KCs are responsible for necroptosis induction in hepatocytes. As shown in Fig. 3C, the culture supernatant from Ttp−/− KCs could induce phosphorylation of MLKL in both Ttp+/+ and Ttp−/− hepatocytes. The production of higher TNF-α levels in Ttp−/− mice prompted us to determine whether OA-induced TNF-α in Ttp−/− KCs is responsible for causing necroptosis in hepatocytes. We therefore pre-incubated the culture supernatant from Ttp−/− KCs with anti-TNF-α antibody to neutralize TNF-α before treating hepatocytes. The neutralization of TNF-α abolished the induction of MLKL phosphorylation in both Ttp+/+ and Ttp−/− hepatocytes (Fig. 3D). These results suggest that cell death of OA-challenged hepatocytes induced by Ttp−/− KCs can be characterized as TNF-α-dependent necroptosis (Fig. 3E).
Fig. 3Ttp−/− Kupffer cells induce necroptosis in hepatocytes.
(A) Co-cultured cells were pretreated with 2 mM metformin for 12 h, followed by exposure to 200 μM OA for 18 h. After 30 h, p-MLKL and MLKL from co-cultured cells were measured by western immunoblotting. (B) Co-cultured primary hepatocytes and KCs were treated with 2 mM metformin prior to treatment with 200 μM OA and 100 ng/ml LPS. p-MLKL and MLKL were analyzed. (C) Primary hepatocytes (Ttp+/+, Ttp−/−) were cultured in the supernatant of OA-treated KCs from mice of the indicated genotypes. p-MLKL and MLKL from cells were measured by western immunoblotting. (D) Primary hepatocytes (Ttp+/+, Ttp−/−) were treated with neutralizing anti-mouse TNF-α antibody (25 μg/ml) or isotype control (IgG) antibody and cultured with the supernatant of OA-treated KCs from mouse of the indicated genotypes. The supernatant of untreated KCs was used as negative control. (E) Schematic model for OA-or HFD-induced TNF-α production in KCs occurs as hepatocyte necroptosis in Ttp−/− murine liver. (F) ATGL and HSL from Ttp+/+ and Ttp−/− primary hepatocytes were measured by qRT-PCR. (G) ATGL and HSL from Ttp+/+ and Ttp−/− liver of mice fed NCD or HFD were measured by qRT-PCR. (H) ATGL and HSL from Ttp+/+ primary hepatocytes co-cultured with Ttp+/+ KCs cells or Ttp−/− KCs and (I) from Ttp−/− primary hepatocytes co-cultured with Ttp+/+ KCs or Ttp−/− KCs were measured by qRT-PCR. Data represent mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001, and NS, not significant.
Prevention of lipid accumulation in hepatocytes by co-cultured Ttp−/− KCs prompted us to test whether Ttp−/− KCs sensitize hepatocytes to induce the expression of genes involved in lipolysis such as ATGL and HSL. Single cultures of hepatocytes showed similar expression patterns of ATGL and HSL. The expression levels of these genes were induced by co-treatment with OA and metformin in either Ttp+/+ or Ttp−/− hepatocytes (Fig. 3F and G), suggesting that TTP does not regulate their expression in isolated hepatocytes. In contrast, metformin treatment induced these genes in the liver tissue of Ttp+/+ but not Ttp−/− mice on HFD.
In co-culture systems, the expression levels of ATGL and HSL were induced by co-treatment with OA and metformin in either Ttp+/+ or Ttp−/− hepatocytes (Fig. 3H and I) when cultured in the presence of Ttp+/+KCs. Under these conditions, however, hepatocyte expression of these genes was diminished in the presence of Ttp−/− KCs. These results, taken together, suggest that the induction of lipolysis-related genes is not responsible for the inhibition of lipid accumulation in hepatocytes co-cultured in the presence of Ttp−/− KCs. Finally, the increase of TNF-α in Ttp−/− KCs by OA stimulation induces hepatocyte necroptosis. We conclude that these observations may account for the phenotype of Ttp−/− mice in vivo, with respect to their lack of hepatic lipid accumulation in response to HFD.
3.4 TTP is required for metformin-induced lipophagy and amelioration of fat accumulation in hepatocytes
To further understand the mechanism underlying metformin-mediated lipid reduction in the presence of TTP, we assessed the expression of genes related to autophagy. Because metformin is a mediator of autophagy via AMPK [
], metformin dose-dependently increased the expression of autophagy-related and lysosome-related genes, such as LC3B-II, Beclin1, ATG7, and LAMP1 (Figs. S3A and S3B), the content of lysosomes (Fig. S3C) and the expression of lysosome-related genes (Fig. S3D), in AML12 cells. In addition, metformin increased autophagic flux as determined using a chloroquine (CQ) assay (Fig. S3E). To investigate whether metformin can induce lipophagy, we assessed the co-localization of lysosomes or LC3B with LDs by staining with lysotracker or LC3-red fluorescence protein (RFP), respectively, and BODIPY. Consistent with the effects of metformin on autophagy and lipid accumulation, we found that metformin decreased lipid accumulation by inducing lipophagy, as determined by increased colocalization of Lysotracker or LC3B-RFP with BODIPY stain (Figs. S3F and S3G). Furthermore, metformin increased autophagy as detected by increased LC3B-II and LAMP1 expression, and decreased p62 levels (Fig. 4A ). TTP was required for these effects as Ttp−/− primary hepatocytes did not display metformin-induced autophagy, as assessed by reduced LC3B-II formation and lack of p62 turnover (Fig. 4A). The increase of lysosomal markers in response to metformin, including LAMP1, CatD, and MCOLN1 was dependent on TTP (Fig. 4B). We also verified that TTP was required for metformin-induced autophagic flux by using CQ treatment as an inhibitor of lysosomal activity (to observe LC3B-II turnover) (Fig. 4C); or by using mCherry-GFP-LC3B as a probe for autophagosome and autolysosome activity (Fig. 4D). To determine the role of TTP in the relationship between metformin-mediated lipid reduction and autophagy, we investigated the co-localization of LC3B and BODIPY in Ttp+/+ and Ttp−/− primary hepatocytes. Consistent with a previous study demonstrating that metformin can induce lipophagy [
], the co-localization of LC3B and BODIPY was increased by metformin in Ttp+/+ but not Ttp−/− primary hepatocytes (Fig. 4E). Taken together, these data indicate that TTP is required for reduction of LDs by metformin treatment via increasing lipophagy.
Fig. 4TTP is required for metformin-induced lipophagy and amelioration of fat accumulation in hepatocytes.
(A, B) Ttp+/+and Ttp−/− primary hepatocytes were treated with metformin for 12 h. The protein expression of LC3B-I/-II, LAMP1, and p62 was measured by western blotting (A). The RNA expression of lysosomal genes, LAMP1, CatD, and MCOLN1 were analyzed by qRT-PCR (B). (C) For analysis of autophagic flux, Ttp+/+ and Ttp−/− primary hepatocytes were treated with 2 mM metformin for 12 h in the presence or absence of chloroquine (CQ) (10 μM, 1 h). The levels of LC3B-I/-II and p62 were observed by western blotting. (D) Ttp+/+ and Ttp−/− primary hepatocytes were transfected with mCherry-GFP-LC3B and then pretreated with CQ (10 μM) prior to metformin (2 mM) treatment for 12 h. Cells were detected for fluorescence of both GFP and mCherry using confocal microscopy (left), scale bar: 20 μm. The number of autolysosomes (GFP−RFP+) and autophagosomes (GFP+RFP+) per cell in each group was quantified (right). (E) Ttp+/+and Ttp−/− primary hepatocytes were transfected with pmRFP-LC3 and then treated with metformin (2 mM, 12 h) before OA (200 μM, 18 h) treatment. Co-localization of LC3B (RFP) with lipid (BODIPY) was analyzed by confocal microscopy, scale bar: 10 μm. Data were presented as mean ± SD (n = 3); **p < 0.01, ***p < 0.001, and NS, not significant.
]. Therefore, we first evaluated the effect of metformin treatment on the nuclear translocation of TFEB in hepatocytes. TFEB nuclear translocation was dose-dependently increased by metformin (Fig. S4A). Consistent with a previous report [
], HBSS promoted an increase of TFEB nuclear translocation. We also confirmed that hepatocytes overexpressing enhanced green fluorescent protein (EGFP)-TFEB clearly showed an increase of nuclear TFEB accumulation after treatment with metformin (Fig. S4B). The mTOR inhibitor, Torin1, was used as a positive control for TFEB nuclear translocation (Fig. S4B). To verify whether metformin-induced lipophagy was dependent on TFEB nuclear translocation, we transfected AML12 cells with scramble RNA (scRNA) or siRNA targeting TFEB (siTfeb) and then lipid clearance, and the expression levels of autophagy- or lysosome-related genes were evaluated after metformin treatment. TFEB mRNA and protein levels were significantly reduced in siTfeb-transfected cells (Fig. S4C). Metformin treatment decreased OA-induced LD accumulation in scRNA-transfected but not siTfeb-transfected hepatocytes (Fig. 5A ). In addition, metformin increased the levels of autophagy-related and lysosome-related genes in scRNA-transfected control cells, and these events were abolished in siTfeb-transfected cells (Fig. 5B and C). Collectively, our data demonstrate that TFEB plays a critical role in metformin-induced autophagy initiation and lysosomal biogenesis, and associated lipid clearance.
Fig. 5Metformin induces TTP-Rheb destabilization, leading to TFEB nuclear translocation in hepatocytes.
(A) After transfection with siTfeb, AML12 cells were pretreated with 2 mM metformin for 12 h and then exposed to 200 μM OA for 18 h. Lipid accumulation was analyzed by BODIPY staining. (B, C) siTfeb transfected-AML12 cells were treated with metformin for 12 h. (B) The protein expression of LC3B-I/-II, Atg7, and LAMP1 was measured by western blotting. (C) The mRNA expression of lysosomal genes, LAMP1, CatD, and MCOLN1 were analyzed by qRT-PCR. (D) Ttp+/+ and Ttp−/− primary hepatocytes were treated with metformin (2 mM, 6 h) or Torin1 (5 μM, 3 h). The protein expression of TFEB in the cytosol and nuclear fraction was observed by western blotting. (E) Ttp+/+and Ttp−/− primary hepatocytes were transfected with EGFP-TFEB and then treated with metformin (2 mM, 6 h) or Torin1 (5 μM, 3 h). TFEB localization was detected by confocal microscopy. (F) AML12 cells were treated with metformin (1, 2, 4, and 8 mM) for 6 h, Torin1 (5 μM) for 3 h, or starvation for 3 h. The protein expression of TTP, Rheb, p-S6K, S6K, p-S6, and S6 were measured by western blotting. (G) AML12 cells were treated with metformin (1, 2, 4, and 8 mM) for 6 h and 5 μM Torin1 for 3 h, or cells were starved with HBSS for 3 h. The mRNA expression of Ttp and Rheb were measured by RT-PCR. (H, I) Ttp+/+and Ttp−/− primary hepatocytes were treated with metformin for 6 h. The mRNA expression of Ttp and Rheb were measured by RT-PCR (H). The protein expression of TTP and Rheb were detected by western blotting (I). (J) HEK293 cells were transfected with psiCHECK2-Rheb-3′-UTR and then treated with metformin (1 and 2 mM) for 6 h. Degradation of the Rheb 3′-UTR was analyzed by luciferase activity. (K,L) AML12 cells (K) and Ttp+/+and Ttp−/− primary hepatocytes (L) were treated with 2 mM metformin for 6 h. Expression of Rheb mRNA was determined by qRT-PCR at the indicated times after treatment with 5 μg/ml actinomycin D. (M, N) Ttp+/+and Ttp−/− primary hepatocytes were transfected with siRheb and then treated with 2 mM metformin for 6 h. The nuclear translocation of TFEB was analyzed by western blotting in nuclear and cytosolic fractions (M). The protein expression of Rheb was measured by western blotting (N). (O) Schematic model; metformin-induced TTP activation induces TFEB nuclear translocation via Rheb-mediated mTORC1 inhibition, increasing ALP activation. Results are represented as mean ± SD; *p < 0.05, **p < 0.01, ***p < 0.001, and NS, not significant.
We next investigated the role of TTP in metformin-induced TFEB nuclear translocation and subsequent increase in autophagy. Consistent with our previous results [
], metformin increased the expression of TTP in a dose-dependent manner (Fig. S4D). To test whether TTP was required as the key mediator of TFEB nuclear translocation, we measured nuclear TFEB levels after metformin treatment in Ttp+/+ and Ttp−/− primary hepatocytes. As expected, we found that metformin increased the nuclear translocation of TFEB in wild type but not Ttp−/− primary hepatocytes (Fig. 5D) or by using EGFP-TFEB transfected hepatocytes (Fig. 5E). In contrast, in Torin1-treated cells, TFEB nuclear translocation was independent of TTP. These results suggest that TTP specifically mediates the metformin-induced nuclear translocation of TFEB in hepatocytes.
We next sought to determine how TTP enhances TFEB nuclear translocation in metformin-treated primary hepatocytes. Transcriptomic analysis of primary hepatocytes after 6 h exposure to metformin revealed a core group of 2364 regulated transcripts (Fig. S5A, Supplementary Dataset 1 for all RNA-seq data). TTP, encoded by the Zfp36 gene, was represented in the group of genes upregulated by metformin (Fig. S5A). To determine the role of TTP in autophagy, we conducted KEGG enrichment analysis of genes upregulated in Ttp−/− primary hepatocytes by metformin. mTOR signaling was one of the most prevalent KEGG pathways (Fig. S5B). Given that inhibition of mTORC signaling induced autophagy, we suggest that the products of upregulated genes in Ttp−/− hepatocytes could activate the mTOR signaling pathway. We previously found that TTP suppressed mTOR signaling via degrading Rheb [
]. Consistently, Rheb was represented in the genes upregulated by metformin in Ttp−/−vs. Ttp+/+ hepatocytes (Fig. S5C). We next investigated whether metformin-induced TTP affects Rheb expression in AML12 cells. We confirmed the increase of TTP expression and a decrease of Rheb protein expression in response to metformin treatment, in a dose-dependent manner (Fig. 5F). Rheb is known to be involved in the activation of mTOR, which suppresses autophagy [
]. Therefore, the degradation of Rheb by TTP potentially leads to inhibition of mTORC1 activity. Consistently, we found that metformin treatment dose-dependently reduced phosphorylation of the mTOR target, S6 kinase, in hepatocytes (Fig. 5F), suggesting that the degradation of Rheb by metformin-induced TTP inhibits mTOR activity. Ttp mRNA also was increased by metformin in a dose-dependent manner, whereas Rheb expression was decreased (Fig. 5G). We found that an increase of Ttp mRNA expression in response to metformin was associated with decreased Rheb mRNA and Rheb protein expression, in Ttp+/+, but not Ttp−/−, hepatocytes, suggesting that TTP is required for Rheb down-regulation (Fig. 5H and I). The expression of a luciferase reporter gene containing the Rheb-3′-UTR was decreased by metformin (Fig. 5J); and metformin enhanced Rheb mRNA degradation (Fig. 5K) in AML12 cells. These results suggest that Rheb-3′-UTR is responsible for Rheb down-regulation by metformin. While metformin induced Rheb mRNA degradation in Ttp+/+ hepatocytes, metformin did not exert this effect in Ttp−/− hepatocytes (Fig. 5L), indicating that TTP mediated the metformin-induced degradation of Rheb mRNA. Knockdown of Rheb resulted in maximal increase of TFEB nuclear translocation in both Ttp+/+and Ttp−/− primary hepatocytes even in the absence of metformin (Fig. 5M and N). These results confirm that metformin-induced TTP increases TFEB nuclear translocation through down-regulation of Rheb (Fig. 5O). Collectively, our data demonstrated that TTP was required for the decrease of Rheb expression and for inhibition of mTOR activity in response to metformin.
3.6 Metformin-TTP-Rheb destabilization is regulated by AMPK-Sirt1 activation
Since metformin promotes AMPK phosphorylation and Sirt1 activity [
], we investigated whether metformin-mediated TTP activation is dependent on the AMPK-Sirt1 pathway (Fig. 6A ). To determine the sequence of AMPK and Sirt1 activation by metformin, we investigated the levels of p-AMPK and Sirt1 after treatment with compound C (CC), an AMPK inhibitor, or EX527, a Sirt1 inhibitor. Whereas Sirt1 was suppressed by AMPK inhibition (Fig. 6B and C), the inhibition of Sirt1 did not affect AMPK activation (Fig. 6D). These results indicate that AMPK lies upstream of Sirt1 activation in response to metformin treatment. In addition, metformin-induced TTP expression was suppressed by CC (Fig. 6E). Conversely, the inhibition of Rheb by metformin was recovered by CC (Fig. 6B and E). The increase of TFEB nuclear translocation by metformin was inhibited by CC (Fig. 6F). Accordingly, CC suppressed metformin-induced LAMP1 and MCOLN1 expression levels (Fig. 6G). We also found that EX527 reduced the metformin-mediated increase in Ttp levels and decrease in Rheb expression levels (Fig. 6H and I). Similar to the results observed with CC treatment, EX527 suppressed metformin-induced TFEB activity and the expression of lysosome-related genes (Fig. 6J and K). To confirm the role of the AMPK-Sirt1 axis in metformin-mediated regulation of Rheb expression, a Rheb 3′-UTR assay and a Rheb mRNA stability assay were performed. Administration of CC or EX527 recovered the Rheb 3′-UTR dependent luciferase activity that was decreased by metformin treatment, and also reversed the decrease of Rheb mRNA stability by metformin (Fig. 6L and M). To verify whether the metformin-dependent decrease in Rheb levels was caused by Sirt1 activation, hepatocytes were transfected with siRNA against Sirt1. Metformin-decreased Rheb levels led to an induction of autophagy through mTORC1 inhibition, resulting in increased LC3B-II levels dependent on Sirt1 (Fig. 6N). Taken together, our results suggest that metformin increased autophagy via the AMPK-Sirt1-TTP-Rheb axis.
Fig. 6Metformin-TTP-Rheb destabilization is regulated by AMPK-Sirt1 activation.
(A) Schematic model for TTP activation via activation of metformin-AMPK-Sirt1 axis (B, C) AML12 cells were pretreated with compound C (CC, 10 μM) for 30 min and then treated with 2 mM metformin for 6 h. (B) Cells were subjected to western blotting using antibodies against p-AMPK, AMPK, SIRT1, and Rheb. (C) The mRNA expression of Sirt1 was measured by RT-PCR. (D) AML12 cells were treated with 2 mM metformin for 6 h in the presence or absence of SIRT1 inhibitor, EX527 (20 μM, 30 min). Immunoblot of cell lysates. (E-G) AML12 cells were pretreated with CC (10 μM) for 30 min and then treated with 2 mM metformin for another 6 h. (E) Measurement of Ttp and Rheb mRNA level was performed by RT-PCR (left) and qRT-PCR (right). (F) Measurement of TFEB activity was performed by western blotting in nuclear and cytoplasmic extracts. (G) The mRNA expression of LAMP1 and MCOLN1 was determined by qRT-PCR. (H–K) AML12 cells were pretreated with 20 μM EX527 for 30 min, and then treated with metformin (2 mM) for 6 h. (H, I) The expression of Ttp and Rheb mRNA was analyzed by RT-PCR and qRT-PCR. (J) The protein expression of TFEB in the cytosol and nuclear fraction was detected by western blotting. (K) The mRNA expression of lysosomal genes was detected by qRT-PCR. (L) HEK293 cells were transfected with psiCHECK2-Rheb-3′-UTR and then treated with CC or EX527 followed by treatment of metformin. Degradation of Rheb 3′-UTR was measured by luciferase activity. (M) For measurement of Rheb mRNA stability, AML12 cells were pretreated with CC or EX527 and then treated with 2 mM metformin for 6 h. The expression of Rheb mRNA was determined by qRT-PCR at the indicated times after the treatment of 5 μg/ml actinomycin D. (N) AML12 cells were transfected with siSirt1 and then treated with 2 mM metformin for 6 h. Immunoblot of cell lysates. Data represent mean ± SD; *p < 0.05, **p < 0.01, and ***p < 0.001.
3.7 TTP activation by metformin ameliorates NAFLD via reducing necroptosis and increasing lipophagy
We investigated whether TTP plays a regulatory role in reducing HFD- or MCD-induced NAFLD in response to metformin treatment. In Ttp+/+ mice, HFD increased histological score for steatosis, ballooning and inflammation (Fig. 7A ), NAFLD activity score (Fig. 7B), liver TG (Fig. 7C), and serum AST and ALT levels (Fig. 7D), which were reduced by metformin treatment. Interestingly, these indicators were not reduced by metformin treatment in Ttp−/− mice fed with HFD. The same phenomena were observed in Ttp+/+ and Ttp−/− mice fed with MCD (Figs. S6A-S6E). While HFD or MCD did not show any effects of lipid accumulation in the liver of Ttp−/− mice, despite increased inflammation, and the latter was not affected by metformin (Figs. 7A and S6A). These results clearly suggest that TTP is required for the protective effect of metformin on HFD- or MCD-induced liver damage.
Fig. 7TTP activation by metformin ameliorates NAFLD via reducing necroptosis and increasing lipophagy.
(A-H) Ttp+/+ and Ttp−/− mice (n = 5) were fed 60 % HFD for 8 weeks or NCD with metformin (200 mg/kg) (A) H&E staining of liver tissues. Scale bar: 100 μm (left). Histology scores for steatosis, hepatocyte ballooning, and inflammation (right). (B) NAFLD activity score (NAS). (C) Liver triglycerides and levels of AST and ALT (D) in serum were measured. (E) The mRNA expression levels of TNF-α, IL-1β, and CXCL1 was detected with RT-PCR in liver tissues of HFD. (F, G) The mRNA expression of IL-1β (F) and TNF-α (G) was detected with qRT-PCR in liver tissues of HFD. (H) p-MLKL and MLKL from liver tissues of HFD-fed mice were measured by western immunoblotting. (I) The protein levels of TTP, Rheb, LAMP1, and LC3B-II in liver tissues of MCD-fed mice was determined by western blotting. (J-L) The mRNA expression of Rheb (J), LAMP1 (K), and MCOLN1 (L) in liver tissues of MCD-fed mice was detected with qRT-PCR. (M) Autophagosomes and autolysosomes were examined by transmission electron microscopy (TEM). Arrowheads indicate autophagosomes, and arrows indicate autolysosomes. Scale bar: 1 and 2 μm. Data are represented as mean ± SD; **p < 0.01, ***p < 0.001, and NS, not significant.
Although our results revealed the mechanism by which metformin-induced TTP prevents lipid accumulation in hepatocytes in vitro, we next sought to determine the underlying mechanism for reduced hepatic lipid accumulation by metformin but enhanced liver damage in Ttp−/− mice fed with HFD or MCD. First, we demonstrated that the livers of Ttp−/− mice fed with HFD or MCD expressed higher levels of TNF-α, IL-1β, and CXCL1 than those of Ttp+/+ mice (Figs. 7E-7G, S6F, and S6G). In addition, metformin suppressed HFD- or MCD-induced TNF-α, IL-1β, and CXCL1 only in Ttp+/+ mice but not in Ttp−/− mice. Previous reports suggest that hepatocyte necroptosis is activated in patients with NAFLD and experimental NASH [
] and that necroptosis is involved in a variety of acute and chronic types of liver disease (Nagy et al., 2016; Schwabe and Luedde, 2018). Consistent with our in vitro results, the phosphorylation of MLKL was increased in the liver of HFD-fed Ttp−/− mice (Fig. 7H and S6H). These results suggest that hepatocytes of HFD-fed Ttp−/− mice may undergo necroptosis in vivo.
Consistent with the positive effect of metformin-induced TTP on lipophagy in vitro, metformin treatment in HFD or MCD-fed Ttp+/+ mice led to an increase in TTP expression, subsequent to a decrease in Rheb expression, and also to increased LAMP1 expression, and LC3B-II conversion (Figs. 7I, S6I, 7J, S6J, and S6K). However, in the Ttp−/− mice, due to Rheb stabilization, autophagic flux was impaired with enhanced LC3B-II levels observed in both the absence or presence of metformin (Figs. 7I, S6I, 7J, S6J, and S6K). In addition, metformin increased LAMP1, and MCOLN1 mRNA expression, and the number of autophagosomes by EM analysis, in MCD-fed Ttp+/+ mice, but not in Ttp−/− mice (Figs. 7K-7M). Taken together, our data demonstrate that metformin-induced lipophagy ameliorated hepatic steatosis in a TTP-dependent manner. Finally, we suggest that recovery for NAFLD by metformin could be considered as dependent on the cooperative roles of KCs and hepatocytes, which are mediated by the presence of TTP.
4. Discussion
The liver is a central organ for fat-storage, which renders it particularly susceptible to steatosis and subsequent inflammation [
]. While metformin has been studied for decades for its potential to ameliorate hepatic steatosis in metabolic disease, its therapeutic targets remain enigmatic. We propose that TTP has a crucial role in alleviating hepatic steatosis in response to metformin treatment. Consequently, TTP activation with metformin treatment leads to the suppression of inflammation and necroptosis via degradation of TNF-α and to a decrease of lipid accumulation by promoting lipophagy. Hepatic cell death prompts the development of chronic liver injury under inflammatory responses that lead to hepatic fibrosis and cirrhosis [
In this study, we found that metformin reduces lipid accumulation and hepatic dysfunction in HFD- or MCD-fed wild type mice. However, TTP deficient mice did not display HFD or MCD-increased LDs accumulation nor the beneficial effect of metformin on liver injury markers (Figs. 1 and S1). Here, we suggest two possible mechanisms by which HFD or MCD exposure fail to cause LDs accumulation in Ttp−/− liver, which differ from our in vitro observations using Ttp−/− hepatocytes alone subject to OA treatment. First, based on previous findings of increased lipolysis by inducers of inflammation [
], HFD did not increase lipid accumulation in TTP deficient mice due to severe inflammation (Fig. 1). Namely, Ttp−/− mice displayed higher TNF-α and IL-1β levels than TTP wild type mice under HFD. This is attributed to the fact that TTP can regulate the mRNA stability of numerous inflammatory mediators, resulting in a decrease of various inflammatory cytokine levels [
]. Second, we conclude that excessive TNF-α production from OA-treated Ttp−/− KCs induces hepatocyte necroptosis, which may account for the lack of accumulation of LDs in Ttp−/− mouse liver. Among the pro-inflammatory cytokines, TNF-α can induce necroptosis in hepatocytes [
]. However, we demonstrate here that the regulation of lipogenesis or lipolysis-related gene expression with metformin was not dependent on TTP activation.
Another important feature of our work is the observation that TFEB ablation blocks autophagy induction by metformin. Metformin increases autophagy in various disease models [
Metformin alleviates hepatic steatosis and insulin resistance in a mouse model of high-fat diet-induced nonalcoholic fatty liver disease by promoting transcription factor EB-dependent autophagy.
], but the mechanism remains unknown. TFEB regulates autophagic flux by promoting the biogenesis of lysosomes and the formation of autophagosomes, thereby facilitating substrate clearance [
]. We demonstrated that metformin increases TFEB nuclear translocation, which is dependent on TTP, and subsequently increases autophagy. Dephosphorylated TFEB translocates to the nucleus to increase the expression of its target genes [
]. In the RNA-seq analysis, we confirm the increase of Rheb expression in TTP deficient mice. In agreement with previous findings, metformin-induced TTP activation in turn decreases Rheb expression. Thus, we suggest that metformin inhibits mTORC1 activity through the TTP-dependent degradation of Rheb, leading to increased TFEB nuclear translocation. Metformin acts as an activator of AMPK and Sirt1 [
]. Thus, our results demonstrate that AMPK is upstream of Sirt1 in the pathway underlying metformin-induced autophagy activation.
Consistent with our in vitro results, metformin-induced TTP activation leads to decreased Rheb mRNA and protein levels, which in turn enhances autophagy and lysosome-related gene expression, resulting in alleviation of hepatic LD accumulation. Consistent with a previous study [
], HFD or MCD increase Rheb expression in the liver of vehicle-treated wild type mice. Moreover, mTORC1, and its downstream targets, can promote lipid synthesis and storage in response to nutrients [
]. Although further studies will be needed to understand the molecular basis for the increase of Rheb levels under HFD or MCD conditions, based on previous findings, there may be a potential association between the Rheb-mTORC1 axis and lipid metabolism.
Finally, this study provides insight into the molecular processes involved in the degradation of LDs by metformin. In conclusion, decrease of TNF-α production in KCs by metformin-induced TTP activation prevents hepatocyte necroptosis; and TTP-mediated Rheb destabilization by metformin increases lipophagy in primary hepatocytes and liver. Therefore, to protect against liver damage, it is essential for communication of KCs and hepatocytes, which is dependent on the presence of TTP. Thus, we propose that TTP may serve as a potential therapeutic target for counteracting the development of hepatosteatosis.
A limitation of the current study is that supraphysiological concentrations of metformin were used to study mechanisms of action. Further dose-titration and mechanistic studies will be needed to extrapolate the findings to human therapeutic use of metformin. Despite these limitations, the results of the current support a promising possibility of repurposing metformin, an approved anti-diabetic drug, for use as a therapy in NAFLD. This will require further validation of mechanisms of action of metformin, as well as further human clinical studies.
CRediT authorship contribution statement
Jeongmin Park, Hun Taeg Chung, and Yeonsoo Joe conceived this study. Jeongmin Park, So-Young Rah, Hyeong Seok An, Jong Youl Lee, and Yeonsoo Joe designed and performed experiments. Gu Seob Roh, Stefan W. Ryter, Jeong Woo Park, Chae Ha Yang, Young-Joon Surh, and Uh-Hyun Kim analyzed data and prepared the manuscript and critically discussed the data. Jeongmin Park, Stefan W. Ryter, Hun Taeg Chung, and Yeonsoo Joe revised the article. All authors read and approved the final manuscript.
Data availability
The data that supports the findings of this study are available in the manuscript and supplementary material of this article.
Declaration of competing interest
The authors declare no conflicts of interest.
Acknowledgments
This work was supported by the Priority Research Centers Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2014R1A6A1030318), NRF-2020R1A2C1009192 to H.T.C., NRF-2020R1I1A1A01073488 to J.P., NRF-2020R1A2C1006470 to Y. J., and 2018R1A5A2025272 to Yang.
The diagnosis and management of non-alcoholic fatty liver disease: practice guideline by the American Association for the Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association.
Tristetraprolin (TTP)-14-3-3 complex formation protects TTP from dephosphorylation by protein phosphatase 2a and stabilizes tumor necrosis factor-alpha mRNA.
Metformin alleviates hepatic steatosis and insulin resistance in a mouse model of high-fat diet-induced nonalcoholic fatty liver disease by promoting transcription factor EB-dependent autophagy.
00119-1&id=ga1.jpg){kind=link}
00119-1&id=gr1.jpg){kind=link}
00119-1&id=gr2.jpg){kind=link}
00119-1&id=gr3.jpg){kind=link}
00119-1&id=gr4.jpg){kind=link}
00119-1&id=gr5.jpg){kind=link}
00119-1&id=gr6.jpg){kind=link}
00119-1&id=gr7.jpg){kind=link}