Advertisement

The Slc25a47 locus is a novel determinant of hepatic mitochondrial function implicated in liver fibrosis

  • Author Footnotes
    # These authors contributed equally to this work.
    Nadia Bresciani
    Footnotes
    # These authors contributed equally to this work.
    Affiliations
    Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Author Footnotes
    # These authors contributed equally to this work.
    Hadrien Demagny
    Footnotes
    # These authors contributed equally to this work.
    Affiliations
    Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Vera Lemos
    Affiliations
    Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Author Footnotes
    $ These authors contributed equally to this work.
    Francesca Pontanari
    Footnotes
    $ These authors contributed equally to this work.
    Affiliations
    Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Author Footnotes
    $ These authors contributed equally to this work.
    Xiaoxu Li
    Footnotes
    $ These authors contributed equally to this work.
    Affiliations
    Laboratory of Integrative Systems Physiology, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Yu Sun
    Affiliations
    Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Hao Li
    Affiliations
    Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland

    Laboratory of Integrative Systems Physiology, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Alessia Perino
    Affiliations
    Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Johan Auwerx
    Affiliations
    Laboratory of Integrative Systems Physiology, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Kristina Schoonjans
    Correspondence
    Corresponding author. Address: Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland. Tel.: +41 21 693 18 91.
    Affiliations
    Laboratory of Metabolic Signaling, Institute of Bioengineering, School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland
    Search for articles by this author
  • Author Footnotes
    # These authors contributed equally to this work.
    $ These authors contributed equally to this work.
Open AccessPublished:June 14, 2022DOI:https://doi.org/10.1016/j.jhep.2022.05.040

      Highlights

      • SLC25A47 is a liver-specific transporter required to maintain mitochondrial homeostasis in hepatocytes.
      • Slc25a47hep-/- mice display impaired mitochondrial respiration, which leads to energy deficiency and fibrosis in hepatocytes.
      • Slc25a47hep-/- mice display a robust activation of the mitochondrial stress response (MSR) in hepatocytes associated with the secretion of the mitokine FGF21.
      • FGF21 drives a hypermetabolic phenotype in Slc25a47hep-/- mice.

      Background & Aims

      Transporters of the SLC25 mitochondrial carrier superfamily bridge cytoplasmic and mitochondrial metabolism by channeling metabolites across mitochondrial membranes and are pivotal for metabolic homeostasis. Despite their physiological relevance as gatekeepers of cellular metabolism, most of the SLC25 family members remain uncharacterized. We undertook a comprehensive tissue distribution analysis of all Slc25 family members across metabolic organs and identified SLC25A47 as a liver-specific mitochondrial carrier.

      Methods

      We used a murine loss-of-function model to unravel the role of this transporter in mitochondrial and hepatic homeostasis. We performed extensive metabolic phenotyping and molecular characterization of newly generated Slc25a47hep-/- and Slc25a47-Fgf21hep-/- mice.

      Results

      Slc25a47hep-/- mice displayed a wide variety of metabolic abnormalities, as a result of sustained energy deficiency in the liver originating from impaired mitochondrial respiration. This mitochondrial phenotype was associated with an activation of the mitochondrial stress response (MSR) in the liver, and the development of fibrosis, which was exacerbated upon feeding a high-fat high-sucrose diet. The MSR induced the secretion of several mitokines, amongst which FGF21 played a preponderant role on systemic physiology. To dissect the FGF21-dependent and -independent physiological changes induced in Slc25a47hep-/- mice, we generated a double Slc25a47-Fgf21hep-/- mouse model and demonstrated that several aspects of the hypermetabolic state were driven by hepatic secretion of FGF21. On the other hand, the metabolic fuel inflexibility observed in Slc25a47hep-/- mice could not be rescued with the genetic removal of Fgf21.

      Conclusion

      Collectively, our data place the Slc25a47 locus at the center of mitochondrial homeostasis, which upon dysfunction triggers robust liver-specific and systemic adaptive stress responses. The prominent role of the Slc25a47 locus in hepatic fibrosis identifies this carrier, or its transported metabolite, as a potential target for therapeutic intervention.

      Lay summary

      Herein, we report the importance of a locus containing a liver-specific gene coding for a mitochondrial transport protein called SLC25A47. Mitochondria are the powerhouses of cells. They are crucial for metabolism and energy generation. We show that mice with genetic disruption of the Slc25a47 locus cannot maintain mitochondrial homeostasis (balance), leading to wide-ranging problems in the liver that have far-reaching physiological consequences.

      Graphical abstract

      Keywords

      Introduction

      Compartmentalization plays a key role in metabolic regulation, enabling spatial separation of opposing anabolic and catabolic pathways in distinct organelles.
      • Hinzpeter F.
      • Gerland U.
      • Tostevin F.
      Optimal compartmentalization strategies for metabolic microcompartments.
      The functionality of compartmentalization relies on the expression of specific transporters that fine-tune the channeling of metabolites across subcellular compartments. Together with enzymes, such carriers define the metabolic identity of the cell and establish what a cell can and cannot do in terms of metabolism. Of all organelles, mitochondria are noticeable for their fundamental role in intermediary metabolism. All major types of nutrients pass through mitochondria as part of intermediary metabolism, including the degradation products of fats, sugars, proteins, as well as nucleotides, vitamins, and inorganic ions. Mitochondria are enveloped by a double membrane, of which the outer is relatively permeable to solutes due the presence of voltage-dependent anion channels. The inner mitochondrial membrane, however, is comparatively impermeable to maintain efficient oxidative phosphorylation.
      • Pfaff E.
      • Klingenberg M.
      • Ritt E.
      • Vogell W.
      Correlation of the unspecific permeable mitochondrial space with the "intermembrane space".
      To overcome this physical barrier, members of the mitochondrial inner membrane SLC25 (solute carrier family 25) family facilitate the transport of molecules involved in a variety of processes including metabolic cycles, oxidative phosphorylation, DNA maintenance, and iron metabolism.
      • Palmieri F.
      The mitochondrial transporter family SLC25: identification, properties and physiopathology.
      Metabolic flexibility encompasses the ability to rapidly modify the oxidation of nutrients according to their availability. The liver in general, and hepatic mitochondria in particular, play a central role in this process, allowing the organism to switch from fatty acid oxidation during the fasted state to enhanced glucose metabolism during the fed state.
      • Galgani J.E.
      • Moro C.
      • Ravussin E.
      Metabolic flexibility and insulin resistance.
      Consequently, impaired mitochondrial activity in the liver characterizes metabolic inflexibility, i.e. the inability to adapt to nutritional transitions frequently observed in obesity and type 2 diabetes.
      • Galgani J.E.
      • Moro C.
      • Ravussin E.
      Metabolic flexibility and insulin resistance.
      Thus, hepatic mitochondrial fitness is an important hallmark of metabolic flexibility and a better understanding of liver mitochondrial metabolism may reveal new approaches to manage metabolic disorders. While the role of mitochondrial enzymes in supporting hepatic metabolism has been well characterized,
      • Schimke R.T.
      • Doyle D.
      Control of enzyme levels in animal tissues.
      the regulatory function of mitochondrial carriers has been largely understudied. To fill this gap, we sought to determine the expression, distribution, and function of mitochondrial transporters in the liver. In this study, we report the critical role of the Slc25a47 locus, containing a poorly characterized liver-specific mitochondrial carrier, in hepatic and whole-body metabolic homeostasis.

      Materials and methods

      Animal studies

      C57BL/6 Slc25a47tm1a(EUCOMM)Hmgu mice were obtained from the European Mouse Mutant Cell Repository (EuMMCR); in this model, a LoxP- and FRT-flanked LacZ/neomycin resistance cassette was inserted upstream of exon 5 of Slc25a47. LoxP sites were inserted upstream of exon 5 and downstream of exon 6, thus flanking exons 5 and 6. Slc25a47tm1a(EUCOMM)Hmgu mice were initially crossed with CMV-Flp mice (The Jackson Laboratories, JAX#012930) to remove the LacZ/neomycin cassette leaving exons 5 and 6 floxed (Slc25a47tm1c(EUCOMM)Hmgu, also referred to as Slc25a47flox/flox mice). These 2 exons encode most of the protein (209/310 amino acids). Slc25a47tm1c(EUCOMM)Hmgu mice were then crossed with an Alb-Cre recombinase mouse line (The Jackson Laboratories, JAX stock #003574) to generate Slc25a47hep–/– mice harboring a hepatocyte-specific deletion of Slc25a47. Slc25a47-Fgf21hep–/– animals were obtained by crossing Slc25a47hep–/– with Fgf21flox/flox mice (The Jackson Laboratories, JAX stock #022361). Mice were housed with ad libitum access to water and food (normal chow, Safe 150 or high-fat high-sucrose [HFHS] diet, Envigo TD.08811) and kept under a 12-hour dark/12-hour light cycle (7pm/7am) with a temperature of 22°C±1°C and a humidity of 60%±20%. Only male mice were used in this study. For fasting protocols, mice were fasted overnight before euthanizing (from 6pm to 9am). For fed experiments, food was removed at 7am (beginning of the light phase) and mice were euthanized 2 hours later (9am). For fed electron microscopy (EM) analysis, food was not removed. For the HFHS diet study, 8-week-old mice were fed a HFHS diet for 14 weeks.

      Tolerance tests

      For the pyruvate tolerance test (PTT) and the insulin tolerance test (ITT), food was removed from 8am until 2pm (6-hour mild fasting) and mice were single-caged before oral gavage of 2 g/kg sodium pyruvate or intraperitoneal injection of 0.5 IU insulin/kg, respectively. For the oral glucose tolerance test (OGTT), mice were fasted (following fasting protocol) and single-caged before the experiment; 2 g/kg of glucose was administered by oral gavage. In all the tolerance tests, glucose levels were measured with a glucometer at time: 0 (i.e. before injection/gavage) and 15, 30, 45, 60, 90 minutes after injection/gavage. For OGTT, glycemia levels were also measured 120 minutes after gavage and additional blood was collected with a capillary tube at time 0, 15, 30, 60 minutes.

      Indirect calorimetry

      The indirect calorimetry experiment was performed using the Promethion System (Sable Systems International). Each mouse was individually placed in a cage with normal bedding and free access to food and water during the test, which lasted approximately 48 hours (24 hours of acclimation and 24 hours of measurement in normal feeding conditions). The system was set to a 12/12-hour light/dark cycle. Data were normalized on metabolic body mass (body weight, BW0.75).

      Food intake recording

      Food intake was recorded at regular intervals (20 minutes) using a TSE PhenoMaster system (TSE Systems GmbH, Germany). Each mouse was placed individually in a cage with normal bedding and free access to food and water during the test, which lasted approximately 48 hours (24 hours of acclimation and 24 hours of measurement in normal feeding conditions). The system was set to a 12/12-hour light/dark cycle.

      Study approval

      All animal procedures were approved by the Swiss authorities (Canton of Vaud, animal protocols #3221, #3221.1) and performed in accordance with our institutional guidelines.

      Western blotting

      Liver tissues or mitochondrial pellets were homogenized with RIPA buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% NP-40, 2 mM EDTA and 50 mM NaF) supplemented with 1 mM PMSF and protease and phosphatase inhibitor cocktails (Roche). Before running SDS-PAGE, whole liver protein lysates were boiled for 5 min at 95°C. For mitochondrial protein immunoblot, lysates were not boiled. The following antibodies were used for immunoblotting: SLC25A47 (custom antibody manufactured by YenZym Antibodies LCC; Novus Biologicals, #NB100-57843), HSP70 (Abcam, #ab2787), VINCULIN (Abcam, #ab129002), HSP90 (BD Biosciences #610418), YME1L1 (Cusabio, #PA026267GA01HU), HSPA9 (Antibodies online, #ABIN361739), LONP1 (Sigma, #HPA002192) and GAPDH (Santa Cruz Biotechnology, #sc-365062). VINCULIN (Fig. 1B, S4D) and GAPDH (Fig. 4E) served as protein loading control; HSP90 (Fig. 2B) as mitochondrial protein loading control. All antibodies are listed in Table S2 and in the supplementary CTAT table.
      Figure thumbnail gr1
      Fig. 1SLC25A47 is a hepatocyte-specific mitochondrial carrier.
      (A) Heatmap representing mRNA levels (log2 CPM) of murine SLC25 carriers in different tissues. Average of 3 male mice (C57BL/6 J, 7/8-week-old). ENA: #ERP104395. SLC25s are ranked (top to bottom) according to their relative abundance in the liver. (B) Slc25a47 gene expression (n = 3) and protein levels (n = 2) in different tissues from 14-week-old C57BL/6 mice (L, liver; eWAT, epididymal white adipose tissue; K, kidney; Q, quadriceps; S, spleen; P, pancreas). (C) Single-cell RNA sequencing data of murine liver cell populations (3 female and 4 male C57BL/6 mice, 10-15-week-old) from Tabula Muris (GEO #109774, microfluidic droplet counting). (D) Heatmap representing the mRNA levels of murine SLC25 carriers across different hepatic layers. Data from GSE84498 (CV, central vein; PV, portal vein). (E) FLAG (green) and COXIV (red) immunofluorescence staining of HeLa cells overexpressing Empty-FLAG or FLAG-tagged SLC25A47. DAPI (blue) was used to stain nuclei. Scale bar: 50 μm. (F) Gene module association determination analysis of Slc25a47 in mouse liver datasets. The dashed lines indicate the applied threshold (0.26). Known modules connected to Slc25a47 are in red, and other modules over the threshold are in black. CPM, counts per million; FA, fatty acid; TCA, tricarboxylic acid; ETC, electron transport chain; IMM, inner mitochondrial membrane.
      Figure thumbnail gr2
      Fig. 2Slc25a47-deficient mice display hepatic metabolic dysregulation.
      (A) Slc25a47 transcript profiling in tissues from Slc25a47 floxed (Slc25a47hep+/+) and hepatocyte-specific Slc25a47 mutant (Slc25a47hep-/-) mice (9-week-old, n = 8 liver, n = 3 other tissues). (B) SLC25A47 immunoblot of liver mitochondrial extracts (n = 4) from mice described in panel A. (C) Representative pictures of Slc25a47hep+/+ and Slc25a47hep-/- mice. Ruler: 5 cm marks. (D-F) Body weight (D), length (E) and liver weight over body weight (F) of 8-week-old fed mice (n = 8-9). (G) Plasma levels of ALT and AST in fed mice (n = 7). (H) Representative images of livers and their staining with H&E and ORO from fed mice. Scale bars: 20 μm. (I) Representative EM images of livers from fed Slc25a47hep+/+and Slc25a47hep-/- mice. Magnified images (right panels) show glycogen (G), mitochondria (M), endoplasmic reticulum (ER), lipid droplets (LD) and nucleus (N). Scale bar: 5 μm. (J) Hepatic TG content in fed 8-week-old mice (n = 8-9). (K) Glycogen content in hepatic extracts from 8-week-old mice (n = 3-4). (L) Blood glucose levels in fasted 12-week-old mice (n = 8). (M) Plasma lactate in fed 8-week-old mice (n = 7). (N) Blood glucose levels during intraperitoneal PTT in 10-week-old mice (n = 8). Error bars represent mean ± SEM. ∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001 and ∗∗∗∗p <0.0001 relative to Slc25a47hep+/+ mice, as determined by unpaired Student’s t test (A, D-G, J-M) or two-way ANOVA and Bonferroni post hoc correction (N). EM, electron microscopy; ORO, oil red O; PTT, pyruvate tolerance test; TG, triglyceride.

      mRNA analysis

      RNA was extracted from the tissues of the mice using TRIZOL (Invitrogen) and purified with the RNeasy MinElute cleanup kit (Qiagen) following the manufacturer’s instructions. For quantitative reverse transcription PCR (RT-qPCR), cDNA was generated using the QuantiTect® Reverse Transcription Kit (Qiagen) and analyzed by RT-qPCR (SYBR Green chemistry) using a LightCycler® 480 Real-Time PCR System (Roche) and the primers listed in Table S1 and the supplementary CTAT table. Results were normalized to the mean of 36b4 or B2m (ΔΔCt method).

      Mitochondria isolation

      Mitochondria were isolated from livers or cells as previously described.
      • Frezza C.
      • Cipolat S.
      • Scorrano L.
      Organelle isolation: functional mitochondria from mouse liver, muscle and cultured fibroblasts.
      Pellets of mitochondria were resuspended with RIPA buffer for Western Blot (refer to Western Blotting section) or MiR05 for respiration assays (see below).

      Mitochondria functional assessment

      Oxygen consumption rate (OCR) was assessed in freshly isolated intact mitochondria or whole liver homogenate in respiration medium (MiR 05-kit, Oroboros Instruments) by high-resolution respirometry (Oxygraph 2k, Oroboros Instruments) following manufacturers’ instructions. Whole liver lysates were normalized on 16 S content or citrate synthase activity, while intact isolated mitochondria were normalized on (outer membrane) mitochondrial protein levels before performing the assay. For the respirometry assay compounds were added into the 2 ml-chamber. Compounds used: pyruvate (5 mM, Sigma), malate (2 mM, Sigma), glutamate (10 mM, Sigma), succinate (10 mM, Sigma), complex I-inhibitor rotenone (0.1 μM, Sigma) and ADP+Mg2+ (1.25 mM, Sigma) for complex II respiration, FCCP (titration of 1 μl each from 1 mM stock, Sigma) for maximal respiration, and oligomcyin (2 μg/ml Millipore) for uncoupled respiration. For fatty acid oxidation assay, palmitoyl-carnitine (50 μM, Sigma) and octanoyl-carnitine (0.2 mM, Tocris bioscience) were used in the presence of ADP. Malate was added to sustain fatty acid degradation and avoid acetyl-CoA accumulation. At the end of the protocol run, 2.5 μM Antimycin A was used to assess non-mitochondrial respiration. All the represented values were Antimycin A subtracted. To determine mitochondrial DNA content, genomic DNA was extracted from liver tissues using NucleoSpin Tissue (Macherey-Nagel), and RT-qPCR was performed for the expression analysis of 16 s (mitochondrial) normalized to Hk2 (nuclear gene). Citrate synthase activity was determined in a 96 well-plate using 50 μg of liver protein by adding 2 mg/ml DTNB (5,5-dithio-bis-2-nitrobenzoic acid, Sigma) and 15 mM acetyl-CoA (Sigma). Background absorbance at 412 nm was assessed using a spectrophotometer for 5 consecutive reads before addition of 10 mM oxaloacetate (Sigma). The reaction was measured by following the absorbance (412 nm) at 10 s intervals 25 times. ATP levels were measured using ATP assay kit (Abcam, for Fig. 3A) or Cell Titer Glo (Promega, for Fig. 6F, S1F) using ATP standard curve, according to the manufacturer’s instructions.
      Figure thumbnail gr3
      Fig. 3Genetic disruption of the Slc25a47 locus causes hepatic mitochondrial dysfunction.
      (A) Relative hepatic ATP levels from fed and fasted 8-week-old Slc25a47hep+/+ and Slc25a47hep-/- mice (n = 5-6). (B) Scheme representing the procedure to assess OCR. (C-F) Time course (C, E) and bar graph (D, F) representing OCR of freshly isolated mitochondria from fed (C, D) and fasted (E, F) 8-week-old Slc25a47hep+/+ and Slc25a47hep-/- mice (n = 5-8). CI-CII (with ADP): Complex I (pyruvate, malate, glutamate) and Complex II (pyruvate, malate, glutamate, succinate, rotenone) substrates in the presence of ADP. (G) OCR of freshly isolated liver mitochondria from mice described in panel E (n = 8). FAO: fatty acid oxidation substrates (C-POM) in the presence of ADP. Error bars represent mean ± SEM. ∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001 and ∗∗∗∗p <0.0001 relative to Slc25a47hep+/+ mice, as determined by unpaired Student’s t test (A, D, F, AUC in G) or two-way ANOVA and Bonferroni post hoc correction (C, E, G). C-POM, palmitoyl- and octanoyl-carnitine with malate; OCR, oxygen consumption rate.

      Statistical analysis

      Data were represented and analyzed using GraphPad Prism (v9.2.0). Plots are expressed as means ± SEM. Comparison of differences between 2 groups was assessed using unpaired 2-tailed Student’s t tests. Multiple group comparisons were assessed by one-way or two-way ANOVA with Bonferroni post hoc correction or Tukey’s multiple comparison test. Differences below p <0.05 were considered statistically significant (∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001, ∗∗∗∗p <0.0001). Statistical methods and corresponding p values for the data shown in each figure are described in their respective legends.
      Detailed methods, bioinformatic analysis and additional assays are described in the supplementary methods section.

      Results

      SLC25A47 is a hepatocyte-specific mitochondrial carrier

      Despite their roles as gatekeepers of metabolic fluxes, mitochondrial carriers have been understudied.
      • Gutierrez-Aguilar M.
      • Baines C.P.
      Physiological and pathological roles of mitochondrial SLC25 carriers.
      To better characterize this family of transporters, we analyzed published RNA sequencing (RNA-seq) data
      • Sollner J.F.
      • Leparc G.
      • Hildebrandt T.
      • Klein H.
      • Thomas L.
      • Stupka E.
      • et al.
      An RNA-Seq atlas of gene expression in mouse and rat normal tissues.
      and undertook a comprehensive tissue distribution analysis of all SLC25 family members across 12 metabolic organs (Fig. 1A). This analysis revealed that members of the Slc25 family vary greatly in their tissue distribution with some members widely expressed across tissues such as the phosphate carrier, Slc25a3 (Fig. 1A), while others displayed a highly restricted pattern of expression such as Slc25a18, a glutamate transporter
      • Fiermonte G.
      • Palmieri L.
      • Todisco S.
      • Agrimi G.
      • Palmieri F.
      • Walker J.E.
      Identification of the mitochondrial glutamate transporter. Bacterial expression, reconstitution, functional characterization, and tissue distribution of two human isoforms.
      solely expressed in the brain (Fig. 1A). Only one carrier, Slc25a47, was strictly confined to the liver, where its expression levels were as high as some major mitochondrial carriers such as the ornithine transporter (Slc25a15, Fig. 1A). The liver-specific expression of Slc25a47 was further confirmed when we interrogated over 29 murine tissues available on published datasets
      • Su A.I.
      • Cooke M.P.
      • Ching K.A.
      • Hakak Y.
      • Walker J.R.
      • Wiltshire T.
      • Orth A.P.
      • et al.
      Large-scale analysis of the human and mouse transcriptomes.
      (Fig. S1A). RT-qPCR and immunoblotting on different mouse tissues corroborated the restricted hepatic expression of SLC25A47 (Fig. 1B).
      To gain further insight into the expression of this carrier at the cellular level, we analyzed published single-cell RNA-seq (scRNA-seq) data.
      • Tabula Muris C.
      • Overall c.
      • Logistical c.
      • Organ c.
      • processing Library p.
      • sequencing
      • et al.
      Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris.
      ,
      • Halpern K.B.
      • Shenhav R.
      • Matcovitch-Natan O.
      • Toth B.
      • Lemze D.
      • Golan M.
      • et al.
      Single-cell spatial reconstruction reveals global division of labour in the mammalian liver.
      These analyses revealed that Slc25a47 is uniquely expressed in hepatocytes (Fig. 1C) with the highest expression in periportal hepatocytes (Fig. 1D), a result confirmed by in situ hybridization (Fig. S1B). At the subcellular level, ectopic overexpression of SLC25A47-FLAG in HeLa cells revealed co-localization with the inner mitochondrial membrane protein, cytochrome c oxidase subunit 4 (Fig. 1E), confirming the putative mitochondrial localization of SLC25A47. To identify the metabolic pathways associated with Slc25a47, we performed gene module association determination (G-MAD) analysis using mouse liver data and gene/pathways annotations.
      • Li H.
      • Rukina D.
      • David F.P.A.
      • Li T.Y.
      • Oh C.M.
      • Gao A.W.
      • et al.
      Identifying gene function and module connections by the integration of multispecies expression compendia.
      This analysis showed positive correlations between Slc25a47 and the mitochondrial compartment, such as mitochondrial protein complexes, tricarboxylic acid cycle, electron transport chain (ETC), inner mitochondrial membrane, peroxisome proliferator-activated receptor (PPAR) signaling, lipid catabolism, and fatty acid oxidation (Fig. 1F). Conversely, Slc25a47 negatively correlated with genes involved in cellular proliferation and cholesterol biosynthesis (Fig. 1F). SLC25A47 has been reported to act as an uncoupler.
      • Jin X.
      • Yang Y.D.
      • Chen K.
      • Lv Z.Y.
      • Zheng L.
      • Liu Y.P.
      • et al.
      HDMCP uncouples yeast mitochondrial respiration and alleviates steatosis in L02 and hepG2 cells by decreasing ATP and H2O2 levels: a novel mechanism for NAFLD.
      ,
      • Tan M.G.
      • Ooi L.L.
      • Aw S.E.
      • Hui K.M.
      Cloning and identification of hepatocellular carcinoma down-regulated mitochondrial carrier protein, a novel liver-specific uncoupling protein.
      To address this point, we used a Crispr/Cas9 system to boost the endogenous expression of Slc25a47 in AML12 cells (Fig. S1C). A moderate, but significant, increase in respiration was observed upon induction of Slc25a47 transcript levels (Fig. S1D-E). In contrast to previous reports, induction of Slc25a47 stimulated coupled, but not uncoupled respiration (Fig. S1E), and increased ATP levels (Fig. S1F), suggesting that, in this experimental setup, SLC25A47 did not act as an uncoupler. Collectively, our data reveal that SLC25A47 is an inner mitochondrial membrane transporter highly and exclusively expressed in hepatocytes, suggesting an important role in hepatic metabolism.

      Slc25a47-deficient mice display hepatic metabolic dysregulation

      To understand the role of SLC25A47 in hepatic metabolism, we generated a mouse line carrying an hepatocyte-specific somatic deletion in the Slc25a47 locus (Slc25a47hep-/-) (Fig. S2A, B). RT-qPCR (Fig. 2A), immunoblotting (Fig. 2B), and RNAscope (Fig. S2C) analyses confirmed deletion of the carrier in Slc25a47hep-/- livers. Slc25a47hep-/- mice were distinctively shorter and lighter than their wild-type littermates (Fig. 2C-E), suggesting a broad impact on physiology. We noticed that 8-week-old Slc25a47hep-/- male mice displayed a reduced liver to body weight ratio both in the fed (Fig. 2F) and fasted state (Fig. S2D). Furthermore, plasma levels of aspartate aminotransferase and alanine aminotransferase were elevated, indicative of hepatic damage, which was more pronounced in the fasted state (Fig. S2E-F) than when animals were fed ad libitum (Fig. 2G). The livers of these Slc25a47hep-/- mice were also paler, which is suggestive of lipid accumulation (Fig. 2H). Accordingly, light microscopy (Fig. 2H; S2G), EM (Fig. 2I; S2H-I), and biochemical analyses (Fig. 2J; S2J-K) revealed abnormal lipid and cholesterol accumulation in the liver of fed and fasted Slc25a47hep-/- mice, consistent with G-MAD analysis (Fig. 1F). In the fed state, EM subcellular imaging and enzymatic assessment revealed substantial depletion of glycogen stores (Fig. 2I, K) in Slc25a47hep-/- livers suggestive of dysregulated glucose management in hepatocytes. This observation prompted us to further assess the role of the Slc25a47 locus in glucose homeostasis. We found that fasted Slc25a47hep-/- mice were markedly hypoglycemic (Fig. 2L) and had increased plasma lactate levels (Fig. S2L; 2M). To investigate whether this phenotype could arise from an impairment in de novo glucose production, we subjected the mice to a PTT. Glycemic excursion in response to pyruvate during PTT was significantly compromised in the Slc25a47hep-/- mice (Fig. 2N). In sum, Slc25a47hep-/- mice display a wide range of metabolic alterations ranging from altered gluconeogenesis and liver lipid metabolism to decreased glycogen storage.

      Hepatic Slc25a47 deficiency leads to impaired respiration

      We next interrogated if a chronic energy deficiency in Slc25a47hep-/- hepatocytes due to mitochondrial dysfunction could underlie the broad and diverse metabolic phenotypes observed. In line with this assumption, ATP levels were strongly diminished in the livers of fed and fasted 8-week-old Slc25a47hep-/- male mice (Fig. 3A) suggesting impaired oxidative phosphorylation. To assess the role of the Slc25a47 locus in mitochondrial function, we performed high-resolution respirometry assays on liver lysates and isolated mitochondria from fed and fasted Slc25a47hep-/- males. We used a combination of drugs and substrates to assess different respiration states (Fig. 3B). Compared to fed or fasted Slc25a47hep+/+ livers, mitochondria isolated from Slc25a47hep-/- livers respired significantly less, even in the presence of the complex I inhibitor, rotenone (Fig. 3C-F; S3A). The defect in respiration did not appear to be restricted to any given complex or limited to coupled or uncoupled respiration. In line with our cellular data (Fig. S1C-F), addition of the chemical uncoupling agent FCCP failed to rescue the respiratory phenotype observed in mitochondria isolated from Slc25a47hep-/- livers. We then extended these results by assessing the fatty acid oxidation potential of Slc25a47hep-/- mitochondria and found that palmitoyl- and octanoyl-carnitine-driven respiration was also severely impaired in mitochondria extracted from Slc25a47hep-/- livers (Fig. 3G). Taken together, our results demonstrate that Slc25a47hep-/- livers suffer from a shortage of cellular ATP as a consequence of impaired respiration most likely arising from a global mitochondrial defect.

      Slc25a47 deficiency leads to a strong induction of the mitochondrial stress response

      We then performed RNA-seq analysis on livers from fasted 9-week-old male Slc25a47hep-/- and Slc25a47hep+/+ mice. Principal component analysis revealed a strong separation between the 2 genotypes suggesting robust transcriptional changes induced by loss of Slc25a47 (Fig. 4A). Gene set enrichment analysis indicated that the mitochondrial stress response (MSR) was the most significantly enriched signature (Fig. 4B). The MSR is a highly conserved pathway triggered by various mitochondrial stressors that activate a wide range of factors to restore normal mitochondrial function.
      • Mottis A.
      • Herzig S.
      • Auwerx J.
      Mitocellular communication: shaping health and disease.
      Accordingly, the transcriptome of Slc25a47hep-/- livers was characterized by a strong upregulation of mitochondrial stress markers (Fig. 4C). The MSR encompasses and overlaps with the mitochondrial unfolded protein response (UPRmt), activated to restore protein homeostasis upon mitochondrial protein misfolding.
      • Mottis A.
      • Herzig S.
      • Auwerx J.
      Mitocellular communication: shaping health and disease.
      • Mouchiroud L.
      • Houtkooper R.H.
      • Moullan N.
      • Katsyuba E.
      • Ryu D.
      • Canto C.
      • et al.
      The NAD(+)/Sirtuin pathway modulates longevity through activation of mitochondrial UPR and FOXO signaling.
      • Houtkooper R.H.
      • Mouchiroud L.
      • Ryu D.
      • Moullan N.
      • Katsyuba E.
      • Knott G.
      • et al.
      Mitonuclear protein imbalance as a conserved longevity mechanism.
      RNA and protein levels of the prototypical UPRmt
      • Jovaisaite V.
      • Mouchiroud L.
      • Auwerx J.
      The mitochondrial unfolded protein response, a conserved stress response pathway with implications in health and disease.
      mediators, such as mitochondrial heat shock proteins (Hspa9, Hspd1, Hspa1a) and proteases (Yme1l1, Lonp1) were robustly induced in Slc25a47hep-/- livers (Fig. 4D-E). RNA-seq (Fig. S4A-B), qRT-PCR (Fig. S4C), and immunoblotting (Fig. S4D) analyses confirmed that activation of the MSR was also observed in fed animals. Mitochondrial stress does not only involve cell-autonomous communication from mitochondria to the nucleus, but is also typified by the expression, and secretion, of non-cell-autonomous factors that mediate inter-tissue communication to integrate the cellular mitochondrial stress response within a global metabolic reprogramming of the organism.
      • Mottis A.
      • Herzig S.
      • Auwerx J.
      Mitocellular communication: shaping health and disease.
      ,
      • Quiros P.M.
      • Mottis A.
      • Auwerx J.
      Mitonuclear communication in homeostasis and stress.
      ,
      • Kang S.G.
      • Choi M.J.
      • Jung S.B.
      • Chung H.K.
      • Chang J.Y.
      • Kim J.T.
      • et al.
      Differential roles of GDF15 and FGF21 in systemic metabolic adaptation to the mitochondrial integrated stress response.
      These factors are referred to as mitokines and were first described in invertebrates.
      • Durieux J.
      • Wolff S.
      • Dillin A.
      The cell-non-autonomous nature of electron transport chain-mediated longevity.
      ,
      • Copeland J.M.
      • Cho J.
      • Lo Jr., T.
      • Hur J.H.
      • Bahadorani S.
      • Arabyan T.
      • et al.
      Extension of Drosophila life span by RNAi of the mitochondrial respiratory chain.
      In mammals, fibroblast growth factor 21 (FGF21),
      • Kim K.H.
      • Jeong Y.T.
      • Oh H.
      • Kim S.H.
      • Cho J.M.
      • Kim Y.N.
      • et al.
      Autophagy deficiency leads to protection from obesity and insulin resistance by inducing Fgf21 as a mitokine.
      • Chau M.D.
      • Gao J.
      • Yang Q.
      • Wu Z.
      • Gromada J.
      Fibroblast growth factor 21 regulates energy metabolism by activating the AMPK-SIRT1-PGC-1alpha pathway.
      • Tyynismaa H.
      • Carroll C.J.
      • Raimundo N.
      • Ahola-Erkkila S.
      • Wenz T.
      • Ruhanen H.
      • et al.
      Mitochondrial myopathy induces a starvation-like response.
      • Kliewer S.A.
      • Mangelsdorf D.J.
      A dozen years of discovery: insights into the Physiology and pharmacology of FGF21.
      growth differentiation factor 15
      • Chung H.K.
      • Ryu D.
      • Kim K.S.
      • Chang J.Y.
      • Kim Y.K.
      • Yi H.S.
      • et al.
      Growth differentiation factor 15 is a myomitokine governing systemic energy homeostasis.
      and adrenomedullin 2
      • Lv Y.
      • Zhang S.Y.
      • Liang X.
      • Zhang H.
      • Xu Z.
      • Liu B.
      • et al.
      Adrenomedullin 2 enhances beiging in white adipose tissue directly in an adipocyte-autonomous manner and indirectly through activation of M2 macrophages.
      ,
      • Zhang S.Y.
      • Lv Y.
      • Zhang H.
      • Gao S.
      • Wang T.
      • Feng J.
      • et al.
      Adrenomedullin 2 improves early obesity-induced adipose insulin resistance by inhibiting the class II MHC in adipocytes.
      are the best characterized. All 3 mitokines were strongly induced in the livers of Slc25a47hep-/- mice (Fig. 4C; S4B) but FGF21 seemed to be the main secreted factor, as demonstrated by its plasma levels in both fasted (Fig. 4F) and fed (Fig. S4E) conditions.
      Figure thumbnail gr4
      Fig. 4Slc25a47 deficiency leads to a strong induction of the mitochondrial stress response.
      (A) PCA from RNA-seq of livers from fed and fasted 9-week-old Slc25a47hep+/+ and Slc25a47hep-/- mice (n = 4). (B) GSEA from RNA-seq data of fasted Slc25a47hep-/- (KO) and Slc25a47hep+/+ (WT) livers. ∗q value <0.05, ∗∗q value <0.01. (C) Volcano plot of significantly downregulated (blue) and upregulated (red) genes in liver from mice described in panel B. The names represent mitochondrial stress genes. Mitokine genes are in bold. (D, E) Relative gene expression (D, n = 10) and immunoblot (E, n = 4) from livers of fasted 8-week-old KO vs. WT. (F) Plasma quantification of FGF21, GDF15 and ADM2 in fasted 8-week-old mice (n = 7-8). Error bars represent mean ± SEM. ∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001 and ∗∗∗∗p <0.0001 relative to Slc25a47hep+/+ mice, as determined by unpaired Student’s t test. FC, fold change; GSEA, gene set-enrichment analysis; KO, knockout; NES, normalized enrichment score; PCA, principal component analysis; RNA-seq, RNA-sequencing; WT, wild-type.

      FGF21-dependent systemic phenotypes upon loss of Slc25a47

      Because FGF21 levels were high in Slc25a47hep-/- plasma, we examined the molecular link between its induction and the genetic disruption of the Slc25a47 locus. An in silico analysis of the murine Fgf21 promoter region found putative binding sites for over 30 transcription factors (Fig. S5A). Correlation analyses between their mRNA expression levels and Fgf21 indicated that the stress-induced proteins activating transcription factor (ATF)4 and ATF5 are strongly correlated with Fgf21, suggesting a causative link between these transcription factors and the induction of this mitokine (Fig. 5A). This was confirmed by chromatin immunoprecipitation assays, which showed robust recruitment of both ATF4 and ATF5 to the Fgf21 promoter specifically in Slc25a47hep-/- livers (Fig. 5B). Genetic recombination within the Slc25a47 locus thus leads to an induction of the ATF-FGF21 stress pathway.
      Figure thumbnail gr5
      Fig. 5FGF21-dependent systemic phenotypes upon genetic recombination wihtin the Slc25a47 locus.
      (A) Pearson correlation between transcription factors and Fgf21 obtained using RNA-seq data from livers of fed and fasted Slc25a47hep+/+ and Slc25a47hep-/- mice (n = 8). (B) ChIP assays for ATF4 and ATF5, followed by qPCR on 2 binding sites (Fgf21_1 and Fgf21_2) within the Fgf21 promoter, in livers of Slc25a47hep+/+ and Slc25a47hep-/- mice (n = 8). (C, D) Body weight (C, n = 7-10) and length (D, n = 3-7) of 8-week-old fasted control (Ctrl), Slc25a47hep-/-, Fgf21hep-/- and Slc25a47-Fgf21hep-/- mice. (E) Cumulative food intake (24 h) of 17-week-old Ctrl, Slc25a47hep-/-, Fgf21hep-/- and Slc25a47-Fgf21hep-/- mice (n = 8). The rectangle represents the dark phase. (F, G) H&E staining (F) and relative transcript levels of beiging markers (G, n = 5) in scWAT from mice described in panel C. Scale bar: 20 μm. (H) Energy expenditure in 13-week-old Ctrl, Slc25a47hep-/-, Fgf21hep-/- and Slc25a47-Fgf21hep-/- mice (n = 8). The rectangle represents the dark phase. Error bars represent mean ± SEM. ∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001 and ∗∗∗∗p <0.0001, as determined by one-way ANOVA and Tukey’s multiple comparison test (B-D, G) or two-way ANOVA and Bonferroni post hoc correction (E, H). Asterisks in (E, H) represent Slc25a47hep-/- vs. Ctrl. scWAT, subcutaneous white adipose tissue.
      FGF21 mediates a multitude of metabolic actions,
      • Kharitonenkov A.
      • Shiyanova T.L.
      • Koester A.
      • Ford A.M.
      • Micanovic R.
      • Galbreath E.J.
      • et al.
      FGF-21 as a novel metabolic regulator.
      ,
      • Xu J.
      • Lloyd D.J.
      • Hale C.
      • Stanislaus S.
      • Chen M.
      • Sivits G.
      • et al.
      Fibroblast growth factor 21 reverses hepatic steatosis, increases energy expenditure, and improves insulin sensitivity in diet-induced obese mice.
      and its chronic upregulation is known to induce a profound and persistent hypermetabolic state.
      • Inagaki T.
      • Lin V.Y.
      • Goetz R.
      • Mohammadi M.
      • Mangelsdorf D.J.
      • Kliewer S.A.
      Inhibition of growth hormone signaling by the fasting-induced hormone FGF21.
      ,
      • Owen B.M.
      • Ding X.
      • Morgan D.A.
      • Coate K.C.
      • Bookout A.L.
      • Rahmouni K.
      • et al.
      FGF21 acts centrally to induce sympathetic nerve activity, energy expenditure, and weight loss.
      We hence generated a new mouse line in which both Slc25a47 and Fgf21 were genetically ablated in hepatocytes (Slc25a47-Fgf21hep-/-); we then phenotyped these mice (Fig. S5B-D). Most of the morphological and systemic phenotypes of Slc25a47hep-/- mice were fully rescued in double-mutant mice. In particular, body weight (Fig. 5C) and body length (Fig. 5D; S5E) were normal in 8-week-old male control and Slc25a47-Fgf21hep-/- animals but were abnormally reduced in Slc25a47hep-/- mice. Food intake (Fig. 5E) on the other hand was markedly increased in Slc25a47hep-/- animals but normalized upon double ablation of Slc25a47 and Fgf21. Slc25a47hep-/- mice showed marked beiging of subcutaneous white adipose tissue (scWAT, Fig. 5F-G) associated with a strong increase in energy expenditure both during the day and night (Fig. 5H) which was not linked to increased movement activity (Fig. S5F). In contrast, scWAT beiging and energy expenditure were normal in Slc25a47-Fgf21hep-/- animals (Fig. 5F-H). Altogether, these results support a model in which the hepatic mitochondrial stress induced in Slc25a47hep-/- livers triggers an FGF21-driven hypermetabolic phenotype in Slc25a47hep-/- mice.

      FGF21-independent hepatic phenotypes upon loss of Slc25a47

      Contrary to what we observed at the systemic level, the liver-specific phenotypes of Slc25a47hep-/- animals remained after the genetic deletion of Fgf21. In particular, 8-week-old male single Slc25a47hep-/- and double Slc25a47-Fgf21hep-/- mice were similarly affected by a decrease in the liver to body weight ratio (Fig. 6A) and displayed comparable changes in liver triglycerides (Fig. 6B; S6A) and alanine aminotransferase/aspartate aminotransferase (Fig. 6C) levels upon fasting. All phenotypes associated with glucose homeostasis were also independent of FGF21 expression and its induction of scWAT beiging. Indeed, single Slc25a47hep-/- and double Slc25a47-Fgf21hep-/- mice displayed similarly improved glucose clearance rates during OGTT (Fig. 6D). ITT after mild (6-hour) fasting of the experimental cohorts indicated that, prior to insulin injection, single Slc25a47hep-/- and double Slc25a47-Fgf21hep-/- mice were already hypoglycemic (Fig. 6E; S6D) but responded similarly to the actions of insulin (Fig. S6B), suggesting normal insulin signaling in Slc25a47-deficient animals. We then explored whether the impaired mitochondrial activity in SLC25A47-depleted livers forces hepatocytes to rely more on glycolytic metabolism. In line with this hypothesis, ATP levels were robustly reduced in both single Slc25a47hep-/- and double Slc25a47-Fgf21hep-/- livers (Fig. 6F). Moreover, ablation of Fgf21 in Slc25a47hep-/- mice failed to rescue the defects in respiration (Fig. 6G-H). Both mouse lines also displayed elevated plasma lactate levels during OGTT (Fig. S6C). Additionally, the glycemic excursion after oral pyruvate administration (PTT, Fig. 6I) was no longer observed in single Slc25a47hep-/- and double Slc25a47-Fgf21hep-/- mice, suggesting that ATP depletion in the liver of Slc25a47hep-/- mice impairs normal gluconeogenesis, a hypothesis in line with the fasting-induced hypoglycemia observed in both Slc25a47hep-/- and Slc25a47-Fgf21hep-/- mice (Fig. S6D). Collectively, these data demonstrate that deletion of Slc25a47 has a broad metabolic impact not only in the liver, but also on the whole organism. While several of the hypermetabolic phenotypes could be attributed to chronic FGF21 signaling, those involved in glycemic control remain defective, even after removal of FGF21.
      Figure thumbnail gr6
      Fig. 6Slc25a47-Fgf21hep-/- mice present FGF21-independent phenotypes.
      (A) Liver weight over body weight in fasted 8-week-old Ctrl, Slc25a47hep-/-, Fgf21hep-/- and Slc25a47-Fgf21hep-/- mice (n = 7-9). (B) Hepatic TG content in mice described in panel A (n = 4). (C) Plasma levels of ALT and AST in 22-week-old fasted mice (n = 6). (D-E) Blood glucose levels during OGTT (D) and ITT (E) in 12- and 15-week-old mice (n = 8). (F) Relative ATP levels from fasted livers described in panel A (n = 6-7). (G, H) Time course (G) and bar graphs (H) representing OCR of freshly isolated mitochondria from fasted livers described in panel A (n = 5-6). (I) Blood glucose levels during intraperitoneal PTT in 20-week-old mice (n = 8). Error bars represent mean ± SEM. A-C, F, H: ∗p <0.05, ∗∗p <0.01 and ∗∗∗p <0.001 as determined by one-way ANOVA and Tukey’s multiple comparison test. D, E, G, I: ∗p <0.05, ∗∗p <0.01and ∗∗∗p <0.001 for Slc25a47hep-/- vs. Ctrl; &p <0.05, &&p <0.01 and &&&p <0.001 for Slc25a47-Fgf21hep-/- vs. Ctrl as determined by two-way ANOVA and Bonferroni post hoc correction. ITT, insulin tolerance test; OCR, oxygen consumption rate; OGTT, oral glucose tolerance test; PTT, pyruvate tolerance test; TG, triglyceride.

      Slc25a47hep-/- mice develop liver fibrosis

      To gain insights into the mechanism leading to liver injury in Slc25a47hep-/- mice (Fig. 2G; S2E-F), we performed in silico cell type enrichment analysis with our RNA-seq data (Fig. 7A). More inflammatory cells were present in Slc25a47hep-/- livers along with an upregulation of stellate cells (Fig. 7A) suggesting a predisposition to fibrosis. We hence performed Sirius Red staining on livers of young (8 weeks), mid-aged (32 weeks), and old (80 weeks) Slc25a47hep-/- mice and control littermates. Collagen deposition was significantly increased upon aging in Slc25a47hep-/- livers, as shown by Sirius Red staining (Fig. S7A). We then challenged Slc25a47hep-/- animals with a HFHS diet to further exacerbate the metabolic pressure in the liver. While Slc25a47hep-/- mice were significantly protected against HFHS-induced obesity (Fig. S7B) and diabetes (Fig. 7D-F, S7C) as a result of their increased energy expenditure (Fig. 5H) and improved glycemic profile (Fig. 6), their livers were much more inflamed than control littermates and displayed fibrosis (Fig. 7B-C; S7D). These molecular and histological markers indicate that Slc25a47hep-/- mice are predisposed to develop non-alcoholic steatohepatitis and liver fibrosis.
      Figure thumbnail gr7
      Fig. 7Slc25a47hep-/- mice develop liver fibrosis.
      (A) Cell type enrichment analysis from fed and fasted 8-week-old Slc25a47hep-/- (KO) and Slc25a47hep+/+ (WT) livers. (B) Representative images of hepatic tissue stained for F4/80 and CD45 positive cells, Masson’s trichrome, and Sirius red in HFHS diet fed Slc25a47hep+/+ and Slc25a47hep-/- mice (22- week-old). Mice were fed HFHS diet for 14 weeks. Scale bars: 50 μm. Arrows indicate F4/80 and CD45 positive cells. (C) Relative transcript levels of fibrosis and inflammation markers in 22-week-old HFHS diet fed Slc25a47hep+/+ and Slc25a47hep-/- livers (n = 6-8). (D-F) Blood glucose levels during OGTT (D), ITT (E) and PTT (F) in 12-, 15-, 20-week-old Slc25a47hep+/+ and Slc25a47hep-/- mice (n = 8). Error bars represent mean ± SEM. ∗p/q<0.05, ∗∗p/q<0.01, ∗∗∗p/q<0.001 and ∗∗∗∗p/q<0.0001, as determined by unpaired Student’s t test (C) or two-way ANOVA and Bonferroni post hoc correction (D-F). HFHS, high-fat high-sucrose; ITT, insulin tolerance test; KO, knockout; OGTT, oral glucose tolerance test; PTT, pyruvate tolerance test; WT, wild-type.

      Discussion

      The liver is one of the first organs to respond to dietary factors
      • Kraegen E.W.
      • Clark P.W.
      • Jenkins A.B.
      • Daley E.A.
      • Chisholm D.J.
      • Storlien L.H.
      Development of muscle insulin resistance after liver insulin resistance in high-fat–fed rats.
      and its mitochondria demonstrate high metabolic flexibility to coordinate the changes in bioenergetic demand.
      • Koliaki C.
      • Roden M.
      Alterations of mitochondrial function and insulin sensitivity in human obesity and diabetes mellitus.
      We observed that loss-of-function of the Slc25a47 locus impaired hepatocyte mitochondrial function and, in turn, rendered Slc25a47hep-/- mice metabolically inflexible as demonstrated by their blunted ability to switch to lipid oxidation in the fasted state. The chronic energy deficit also led to impaired gluconeogenesis and faster depletion of the glycogen stores in the fed state. Overall, these phenotypes underlie the observed hypoglycemia, even after mild fasting. Moreover, Slc25a47hep-/- mice suffered from liver damage preponderantly in the fasted state suggesting that Slc25a47-deficient hepatocytes fail to cope with the additional metabolic pressure induced by fasting or HFHS feeding. While our study used a design based on a genetic and liver-specific recombination of the Slc25a47 locus which, surely, maximizes the negative effects on hepatic homeostasis, a milder inhibition of SLC25A47 could prove interesting from a therapeutic perspective. Our data show that one of the main consequences of SLC25A47 downregulation is an inhibition of the hepatic ETC, which results in impaired gluconeogenesis. This mechanism of action is reminiscent of that of metformin, the most frequently prescribed antidiabetic treatment,
      • Pernicova I.
      • Korbonits M.
      Metformin--mode of action and clinical implications for diabetes and cancer.
      which reduces hepatic gluconeogenesis through moderate inhibition of complex I.
      • Pernicova I.
      • Korbonits M.
      Metformin--mode of action and clinical implications for diabetes and cancer.
      Theoretically, a mild inhibition of SLC25A47 could hamper the ETC specifically in hepatocytes and lower glycemia by blunting de novo glucose production. In line with this hypothesis, we observed that heterozygous Slc25a47hep+/- mice did not show signs of liver damage or metabolic stress but secreted less glucose in response to pyruvate (Fig. S2M). This putative therapeutic application remains to be tested under several conditions of metabolic pressure, but is an example of how a better understanding of mitochondrial carriers, and of their specific tissue distribution, could be leveraged to treat metabolic disorders.
      • Palmieri F.
      The mitochondrial transporter family SLC25: identification, properties and physiopathology.
      It should be noted, however, that SLC25A47 was first identified as a gene downregulated during hepatocellular carcinoma development.
      • Tan M.G.
      • Ooi L.L.
      • Aw S.E.
      • Hui K.M.
      Cloning and identification of hepatocellular carcinoma down-regulated mitochondrial carrier protein, a novel liver-specific uncoupling protein.
      This finding calls for a cautious monitoring of SLC25A47 inhibition with regard to cancer progression.
      Our data show that the Slc25a47 locus is essential to maintain hepatic mitochondrial homeostasis as its disruption results in the activation of a profound and persistent MSR. Mitochondrial stress induces mitochondrial quality control mechanisms that are essential to maintain mitochondrial function.
      • Mottis A.
      • Herzig S.
      • Auwerx J.
      Mitocellular communication: shaping health and disease.
      Retrograde signals from the mitochondria to the nucleus have been well-documented in multiple species
      • Cagin U.
      • Enriquez J.A.
      The complex crosstalk between mitochondria and the nucleus: what goes in between?.
      ,
      • Quiros P.M.
      • Prado M.A.
      • Zamboni N.
      • D'Amico D.
      • Williams R.W.
      • Finley D.
      • et al.
      Multi-omics analysis identifies ATF4 as a key regulator of the mitochondrial stress response in mammals.
      and involve the induction of transcription factors, including ATF4,
      • Quiros P.M.
      • Prado M.A.
      • Zamboni N.
      • D'Amico D.
      • Williams R.W.
      • Finley D.
      • et al.
      Multi-omics analysis identifies ATF4 as a key regulator of the mitochondrial stress response in mammals.
      • Bao X.R.
      • Ong S.-E.
      • Goldberger O.
      • Peng J.
      • Sharma R.
      • Thompson D.A.
      • et al.
      Mitochondrial dysfunction remodels one-carbon metabolism in human cells.
      • Münch C.
      • Harper J.W.
      Mitochondrial unfolded protein response controls matrix pre-RNA processing and translation.
      ATF5
      • Fiorese C.J.
      • Schulz A.M.
      • Lin Y.-F.
      • Rosin N.
      • Pellegrino M.W.
      • Haynes C.M.
      The transcription factor ATF5 mediates a mammalian mitochondrial UPR.
      CHOP (C/EBP homologous protein
      • Aldridge J.E.
      • Horibe T.
      • Hoogenraad N.J.
      Discovery of genes activated by the mitochondrial unfolded protein response (mtUPR) and cognate promoter elements.
      ,
      • Zhao Q.
      • Wang J.
      • Levichkin I.V.
      • Stasinopoulos S.
      • Ryan M.T.
      • Hoogenraad N.J.
      A mitochondrial specific stress response in mammalian cells.
      ), and C/EBPβ.
      • Aldridge J.E.
      • Horibe T.
      • Hoogenraad N.J.
      Discovery of genes activated by the mitochondrial unfolded protein response (mtUPR) and cognate promoter elements.
      ,
      • Zhao Q.
      • Wang J.
      • Levichkin I.V.
      • Stasinopoulos S.
      • Ryan M.T.
      • Hoogenraad N.J.
      A mitochondrial specific stress response in mammalian cells.
      Alongside these cell-autonomous effectors, non-cell-autonomous factors are also secreted, enabling the MSR to rewire organismal metabolic homeostasis.
      • Mottis A.
      • Herzig S.
      • Auwerx J.
      Mitocellular communication: shaping health and disease.
      ,
      • Quiros P.M.
      • Mottis A.
      • Auwerx J.
      Mitonuclear communication in homeostasis and stress.
      • Kang S.G.
      • Choi M.J.
      • Jung S.B.
      • Chung H.K.
      • Chang J.Y.
      • Kim J.T.
      • et al.
      Differential roles of GDF15 and FGF21 in systemic metabolic adaptation to the mitochondrial integrated stress response.
      • Durieux J.
      • Wolff S.
      • Dillin A.
      The cell-non-autonomous nature of electron transport chain-mediated longevity.
      ,
      • Kim K.H.
      • Jeong Y.T.
      • Oh H.
      • Kim S.H.
      • Cho J.M.
      • Kim Y.N.
      • et al.
      Autophagy deficiency leads to protection from obesity and insulin resistance by inducing Fgf21 as a mitokine.
      ,
      • Forsstrom S.
      • Jackson C.B.
      • Carroll C.J.
      • Kuronen M.
      • Pirinen E.
      • Pradhan S.
      • et al.
      Fibroblast growth factor 21 drives dynamics of local and systemic stress responses in mitochondrial myopathy with mtDNA deletions.
      In this study, we demonstrated that FGF21 is a key hepatic mitokine that modulates systemic energy metabolism in Slc25a47hep-/- mice. Persistent activation of the ATF family most likely contributes to the marked increase of plasma FGF21 levels in Slc25a47hep-/- mice.
      • De Sousa-Coelho A.L.
      • Marrero P.F.
      • Haro D.
      Activating transcription factor 4-dependent induction of FGF21 during amino acid deprivation.
      We demonstrated that this induction mediates a profound hypermetabolic state, driving increased scWAT beiging, energy expenditure, hyperphagia, and blunted growth, the latter most likely resulting from the chronic hypermetabolism observed in Slc25a47hep-/- mice.
      The MSR is normally activated by defects in mitonuclear protein imbalance, such as seen with defects in mtDNA replication,
      • Forsstrom S.
      • Jackson C.B.
      • Carroll C.J.
      • Kuronen M.
      • Pirinen E.
      • Pradhan S.
      • et al.
      Fibroblast growth factor 21 drives dynamics of local and systemic stress responses in mitochondrial myopathy with mtDNA deletions.
      ,
      • Martinus R.D.
      • Garth G.P.
      • Webster T.L.
      • Cartwright P.
      • Naylor D.J.
      • Hoj P.B.
      • et al.
      Selective induction of mitochondrial chaperones in response to loss of the mitochondrial genome.
      ,
      • Khan N.A.
      • Nikkanen J.
      • Yatsuga S.
      • Jackson C.
      • Wang L.
      • Pradhan S.
      • et al.
      mTORC1 regulates mitochondrial integrated stress response and mitochondrial myopathy progression.
      in the ETC,
      • Durieux J.
      • Wolff S.
      • Dillin A.
      The cell-non-autonomous nature of electron transport chain-mediated longevity.
      ,
      • Quiros P.M.
      • Prado M.A.
      • Zamboni N.
      • D'Amico D.
      • Williams R.W.
      • Finley D.
      • et al.
      Multi-omics analysis identifies ATF4 as a key regulator of the mitochondrial stress response in mammals.
      in mito-ribosomal translation
      • Houtkooper R.H.
      • Mouchiroud L.
      • Ryu D.
      • Moullan N.
      • Katsyuba E.
      • Knott G.
      • et al.
      Mitonuclear protein imbalance as a conserved longevity mechanism.
      ,
      • Kang S.G.
      • Choi M.J.
      • Jung S.B.
      • Chung H.K.
      • Chang J.Y.
      • Kim J.T.
      • et al.
      Differential roles of GDF15 and FGF21 in systemic metabolic adaptation to the mitochondrial integrated stress response.
      or upon mitochondrial biogenesis.
      • Mouchiroud L.
      • Houtkooper R.H.
      • Moullan N.
      • Katsyuba E.
      • Ryu D.
      • Canto C.
      • et al.
      The NAD(+)/Sirtuin pathway modulates longevity through activation of mitochondrial UPR and FOXO signaling.
      ,
      • Zhang H.
      • Ryu D.
      • Wu Y.
      • Gariani K.
      • Wang X.
      • Luan P.
      • et al.
      NAD(+) repletion improves mitochondrial and stem cell function and enhances life span in mice.
      Here we linked the activation of the MSR with the loss of a mitochondrial carrier. Involvement of mitochondrial carriers in mitochondrial stress, however, is not unprecedented, as an early study attempting to define the MSR in Arabidopsis thaliana already reported that the SLC25 family was over-represented amongst stress-responsive genes.
      • Van Aken O.
      • Zhang B.
      • Carrie C.
      • Uggalla V.
      • Paynter E.
      • Giraud E.
      • et al.
      Defining the mitochondrial stress response in Arabidopsis thaliana.
      While this study suggests that mitochondrial stress resolution involves a rewiring of metabolite trafficking across mitochondrial membranes,
      • Van Aken O.
      • Zhang B.
      • Carrie C.
      • Uggalla V.
      • Paynter E.
      • Giraud E.
      • et al.
      Defining the mitochondrial stress response in Arabidopsis thaliana.
      it will be interesting to determine to what extent depletion or accumulation of given metabolites can trigger mitochondrial stress. Our work suggests that SLC25A47 controls the levels of a mitochondrial metabolite, the nature of which is currently unknown and requires further investigation given its potential impact on hepatocyte mitochondrial function, liver fibrosis, and organismal health.
      While genetic disruption of the Slc25a47 locus improves the metabolic profile of mice on an obesogenic, diabetogenic diet, this is achieved at the expense of liver health and integrity, as exemplified by the pronounced fibrotic phenotype induced progressively upon aging or acutely under the metabolic pressure of an HFHS diet. Fibrosis is a common pathological feature of most end-stage organ diseases and growing evidence indicates that mitochondrial dysfunction contributes to the development and progression of fibrosis in a number of organs (reviewed in
      • Li X.
      • Zhang W.
      • Cao Q.
      • Wang Z.
      • Zhao M.
      • Xu L.
      • et al.
      Mitochondrial dysfunction in fibrotic diseases.
      ). Our observations indicate that SLC25A47 could be a promising target to mitigate this process in the liver. Further studies will be needed to further explore in-depth its role in liver fibrogenesis.

      Abbreviations

      ATF, activating transcription factor; EM, electron microscopy; ETC, electron transport chain; FC, fold change; FGF21, fibroblast growth factor 21; HFHS, high-fat high-sucrose, ITT, insulin tolerance test; MSR, mitochondrial stress response; OCR, oxygen consumption rate; OGTT, oral glucose tolerance test; PTT, pyruvate tolerance test; RT-qPCR, reverse-transcription quantitative PCR; scWAT: subcutaneous white adipose tissue; UPRmt, mitochondrial unfolded protein response.

      Financial support

      This work was supported by the Ecole Polytechnique Fédérale de Lausanne (EPFL to K.S. & J.A.), and grants from the Swiss National Science Foundation, Switzerland (SNSF 31003 A_166695 to K.S., SNSF 31003A-179435 to J.A.), the European Research Council (ERC-AdG-787702 to J.A.), the Swiss Cancer League (KFS-4226-08-2017 to K.S.), and the Global Research Laboratory (GRL) National Research Foundation of Korea (NRF 2017K1A1A2013124 to J.A.). H.D. received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No 846001.

      Authors’ contributions

      K.S. conceived, designed, and supervised the project. N.B., H.D., V.L., F.P. and A.P. performed animal experiments. N.B. performed in vitro experiments. X.L. analyzed the RNA-seq data. H.L. analyzed the scRNA-seq data. Y.S. performed ChIP experiments. J.A. provided expertise and funding. N.B., H.D., and K.S. wrote the manuscript.

      Data availability statement

      The data associated with this paper are available upon request to the corresponding author. Reagents, antibodies and resources are listed in the material and methods section, supplementary materials and methods and in the CTAT table.

      Conflict of interest

      The authors declare no conflicts of interest that pertain to this work.
      Please refer to the accompanying ICMJE disclosure forms for further details.

      Acknowledgments

      We thank the Schoonjans and Auwerx lab members, in particular Sabrina Bichet, Thibaud Clerc, Jéromine Imbach, Kim Borany, Fabiana Fraga, Marie Janod, Carla Mendes Ferreira, Penelope Stefanelli, as well as the EPFL-SV animal facility (UDP), the histology facility (HCF) and the electron microscopy facility (BioEM) for technical assistance.

      Supplementary data

      The following are the supplementary data to this article:

      References

        • Hinzpeter F.
        • Gerland U.
        • Tostevin F.
        Optimal compartmentalization strategies for metabolic microcompartments.
        Biophys J. 2017; 112: 767-779
        • Pfaff E.
        • Klingenberg M.
        • Ritt E.
        • Vogell W.
        Correlation of the unspecific permeable mitochondrial space with the "intermembrane space".
        Eur J Biochem. 1968; 5: 222-232
        • Palmieri F.
        The mitochondrial transporter family SLC25: identification, properties and physiopathology.
        Mol Aspects Med. 2013; 34: 465-484
        • Galgani J.E.
        • Moro C.
        • Ravussin E.
        Metabolic flexibility and insulin resistance.
        Am J Physiol Endocrinol Metab. 2008; 295: E1009-E1017
        • Schimke R.T.
        • Doyle D.
        Control of enzyme levels in animal tissues.
        Annu Rev Biochem. 1970; 39: 929-976
        • Frezza C.
        • Cipolat S.
        • Scorrano L.
        Organelle isolation: functional mitochondria from mouse liver, muscle and cultured fibroblasts.
        Nat Protoc. 2007; 2: 287-295
        • Gutierrez-Aguilar M.
        • Baines C.P.
        Physiological and pathological roles of mitochondrial SLC25 carriers.
        Biochem J. 2013; 454: 371-386
        • Sollner J.F.
        • Leparc G.
        • Hildebrandt T.
        • Klein H.
        • Thomas L.
        • Stupka E.
        • et al.
        An RNA-Seq atlas of gene expression in mouse and rat normal tissues.
        Sci Data. 2017; 4170185
        • Fiermonte G.
        • Palmieri L.
        • Todisco S.
        • Agrimi G.
        • Palmieri F.
        • Walker J.E.
        Identification of the mitochondrial glutamate transporter. Bacterial expression, reconstitution, functional characterization, and tissue distribution of two human isoforms.
        J Biol Chem. 2002; 277: 19289-19294
        • Su A.I.
        • Cooke M.P.
        • Ching K.A.
        • Hakak Y.
        • Walker J.R.
        • Wiltshire T.
        • Orth A.P.
        • et al.
        Large-scale analysis of the human and mouse transcriptomes.
        Proc Natl Acad Sci U S A. 2002; 99: 4465-4470
        • Tabula Muris C.
        • Overall c.
        • Logistical c.
        • Organ c.
        • processing Library p.
        • sequencing
        • et al.
        Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris.
        Nature. 2018; 562: 367-372
        • Halpern K.B.
        • Shenhav R.
        • Matcovitch-Natan O.
        • Toth B.
        • Lemze D.
        • Golan M.
        • et al.
        Single-cell spatial reconstruction reveals global division of labour in the mammalian liver.
        Nature. 2017; 542: 352-356
        • Li H.
        • Rukina D.
        • David F.P.A.
        • Li T.Y.
        • Oh C.M.
        • Gao A.W.
        • et al.
        Identifying gene function and module connections by the integration of multispecies expression compendia.
        Genome Res. 2019; 29: 2034-2045
        • Jin X.
        • Yang Y.D.
        • Chen K.
        • Lv Z.Y.
        • Zheng L.
        • Liu Y.P.
        • et al.
        HDMCP uncouples yeast mitochondrial respiration and alleviates steatosis in L02 and hepG2 cells by decreasing ATP and H2O2 levels: a novel mechanism for NAFLD.
        J Hepatol. 2009; 50: 1019-1028
        • Tan M.G.
        • Ooi L.L.
        • Aw S.E.
        • Hui K.M.
        Cloning and identification of hepatocellular carcinoma down-regulated mitochondrial carrier protein, a novel liver-specific uncoupling protein.
        J Biol Chem. 2004; 279: 45235-45244
        • Mottis A.
        • Herzig S.
        • Auwerx J.
        Mitocellular communication: shaping health and disease.
        Science. 2019; 366: 827-832
        • Mouchiroud L.
        • Houtkooper R.H.
        • Moullan N.
        • Katsyuba E.
        • Ryu D.
        • Canto C.
        • et al.
        The NAD(+)/Sirtuin pathway modulates longevity through activation of mitochondrial UPR and FOXO signaling.
        Cell. 2013; 154: 430-441
        • Houtkooper R.H.
        • Mouchiroud L.
        • Ryu D.
        • Moullan N.
        • Katsyuba E.
        • Knott G.
        • et al.
        Mitonuclear protein imbalance as a conserved longevity mechanism.
        Nature. 2013; 497: 451-457
        • Jovaisaite V.
        • Mouchiroud L.
        • Auwerx J.
        The mitochondrial unfolded protein response, a conserved stress response pathway with implications in health and disease.
        J Exp Biol. 2014; 217: 137-143
        • Quiros P.M.
        • Mottis A.
        • Auwerx J.
        Mitonuclear communication in homeostasis and stress.
        Nat Rev Mol Cell Biol. 2016; 17: 213-226
        • Kang S.G.
        • Choi M.J.
        • Jung S.B.
        • Chung H.K.
        • Chang J.Y.
        • Kim J.T.
        • et al.
        Differential roles of GDF15 and FGF21 in systemic metabolic adaptation to the mitochondrial integrated stress response.
        iScience. 2021; 24102181
        • Durieux J.
        • Wolff S.
        • Dillin A.
        The cell-non-autonomous nature of electron transport chain-mediated longevity.
        Cell. 2011; 144: 79-91
        • Copeland J.M.
        • Cho J.
        • Lo Jr., T.
        • Hur J.H.
        • Bahadorani S.
        • Arabyan T.
        • et al.
        Extension of Drosophila life span by RNAi of the mitochondrial respiratory chain.
        Curr Biol. 2009; 19: 1591-1598
        • Kim K.H.
        • Jeong Y.T.
        • Oh H.
        • Kim S.H.
        • Cho J.M.
        • Kim Y.N.
        • et al.
        Autophagy deficiency leads to protection from obesity and insulin resistance by inducing Fgf21 as a mitokine.
        Nat Med. 2013; 19: 83-92
        • Chau M.D.
        • Gao J.
        • Yang Q.
        • Wu Z.
        • Gromada J.
        Fibroblast growth factor 21 regulates energy metabolism by activating the AMPK-SIRT1-PGC-1alpha pathway.
        Proc Natl Acad Sci U S A. 2010; 107: 12553-12558
        • Tyynismaa H.
        • Carroll C.J.
        • Raimundo N.
        • Ahola-Erkkila S.
        • Wenz T.
        • Ruhanen H.
        • et al.
        Mitochondrial myopathy induces a starvation-like response.
        Hum Mol Genet. 2010; 19: 3948-3958
        • Kliewer S.A.
        • Mangelsdorf D.J.
        A dozen years of discovery: insights into the Physiology and pharmacology of FGF21.
        Cell Metab. 2019; 29: 246-253
        • Chung H.K.
        • Ryu D.
        • Kim K.S.
        • Chang J.Y.
        • Kim Y.K.
        • Yi H.S.
        • et al.
        Growth differentiation factor 15 is a myomitokine governing systemic energy homeostasis.
        J Cell Biol. 2017; 216: 149-165
        • Lv Y.
        • Zhang S.Y.
        • Liang X.
        • Zhang H.
        • Xu Z.
        • Liu B.
        • et al.
        Adrenomedullin 2 enhances beiging in white adipose tissue directly in an adipocyte-autonomous manner and indirectly through activation of M2 macrophages.
        J Biol Chem. 2016; 291: 23390-23402
        • Zhang S.Y.
        • Lv Y.
        • Zhang H.
        • Gao S.
        • Wang T.
        • Feng J.
        • et al.
        Adrenomedullin 2 improves early obesity-induced adipose insulin resistance by inhibiting the class II MHC in adipocytes.
        Diabetes. 2016; 65: 2342-2355
        • Kharitonenkov A.
        • Shiyanova T.L.
        • Koester A.
        • Ford A.M.
        • Micanovic R.
        • Galbreath E.J.
        • et al.
        FGF-21 as a novel metabolic regulator.
        J Clin Invest. 2005; 115: 1627-1635
        • Xu J.
        • Lloyd D.J.
        • Hale C.
        • Stanislaus S.
        • Chen M.
        • Sivits G.
        • et al.
        Fibroblast growth factor 21 reverses hepatic steatosis, increases energy expenditure, and improves insulin sensitivity in diet-induced obese mice.
        Diabetes. 2009; 58: 250-259
        • Inagaki T.
        • Lin V.Y.
        • Goetz R.
        • Mohammadi M.
        • Mangelsdorf D.J.
        • Kliewer S.A.
        Inhibition of growth hormone signaling by the fasting-induced hormone FGF21.
        Cell Metab. 2008; 8: 77-83
        • Owen B.M.
        • Ding X.
        • Morgan D.A.
        • Coate K.C.
        • Bookout A.L.
        • Rahmouni K.
        • et al.
        FGF21 acts centrally to induce sympathetic nerve activity, energy expenditure, and weight loss.
        Cell Metab. 2014; 20: 670-677
        • Kraegen E.W.
        • Clark P.W.
        • Jenkins A.B.
        • Daley E.A.
        • Chisholm D.J.
        • Storlien L.H.
        Development of muscle insulin resistance after liver insulin resistance in high-fat–fed rats.
        Diabetes. 1991; 40: 1397-1403
        • Koliaki C.
        • Roden M.
        Alterations of mitochondrial function and insulin sensitivity in human obesity and diabetes mellitus.
        Annu Rev Nutr. 2016; 36: 337-367
        • Pernicova I.
        • Korbonits M.
        Metformin--mode of action and clinical implications for diabetes and cancer.
        Nat Rev Endocrinol. 2014; 10: 143-156
        • Cagin U.
        • Enriquez J.A.
        The complex crosstalk between mitochondria and the nucleus: what goes in between?.
        Int J Biochem Cell Biol. 2015; 63: 10-15
        • Quiros P.M.
        • Prado M.A.
        • Zamboni N.
        • D'Amico D.
        • Williams R.W.
        • Finley D.
        • et al.
        Multi-omics analysis identifies ATF4 as a key regulator of the mitochondrial stress response in mammals.
        J Cell Biol. 2017; 216: 2027-2045
        • Bao X.R.
        • Ong S.-E.
        • Goldberger O.
        • Peng J.
        • Sharma R.
        • Thompson D.A.
        • et al.
        Mitochondrial dysfunction remodels one-carbon metabolism in human cells.
        elife. 2016; 5e10575
        • Münch C.
        • Harper J.W.
        Mitochondrial unfolded protein response controls matrix pre-RNA processing and translation.
        Nature. 2016; 534: 710-713
        • Fiorese C.J.
        • Schulz A.M.
        • Lin Y.-F.
        • Rosin N.
        • Pellegrino M.W.
        • Haynes C.M.
        The transcription factor ATF5 mediates a mammalian mitochondrial UPR.
        Curr Biol. 2016; 26: 2037-2043
        • Aldridge J.E.
        • Horibe T.
        • Hoogenraad N.J.
        Discovery of genes activated by the mitochondrial unfolded protein response (mtUPR) and cognate promoter elements.
        PloS one. 2007; 2: e874
        • Zhao Q.
        • Wang J.
        • Levichkin I.V.
        • Stasinopoulos S.
        • Ryan M.T.
        • Hoogenraad N.J.
        A mitochondrial specific stress response in mammalian cells.
        EMBO J. 2002; 21: 4411-4419
        • Forsstrom S.
        • Jackson C.B.
        • Carroll C.J.
        • Kuronen M.
        • Pirinen E.
        • Pradhan S.
        • et al.
        Fibroblast growth factor 21 drives dynamics of local and systemic stress responses in mitochondrial myopathy with mtDNA deletions.
        Cell Metab. 2019; 30: 1040-1054 e1047
        • De Sousa-Coelho A.L.
        • Marrero P.F.
        • Haro D.
        Activating transcription factor 4-dependent induction of FGF21 during amino acid deprivation.
        Biochem J. 2012; 443: 165-171
        • Martinus R.D.
        • Garth G.P.
        • Webster T.L.
        • Cartwright P.
        • Naylor D.J.
        • Hoj P.B.
        • et al.
        Selective induction of mitochondrial chaperones in response to loss of the mitochondrial genome.
        Eur J Biochem. 1996; 240: 98-103
        • Khan N.A.
        • Nikkanen J.
        • Yatsuga S.
        • Jackson C.
        • Wang L.
        • Pradhan S.
        • et al.
        mTORC1 regulates mitochondrial integrated stress response and mitochondrial myopathy progression.
        Cell Metab. 2017; 26: 419-428 e415
        • Zhang H.
        • Ryu D.
        • Wu Y.
        • Gariani K.
        • Wang X.
        • Luan P.
        • et al.
        NAD(+) repletion improves mitochondrial and stem cell function and enhances life span in mice.
        Science. 2016; 352: 1436-1443
        • Van Aken O.
        • Zhang B.
        • Carrie C.
        • Uggalla V.
        • Paynter E.
        • Giraud E.
        • et al.
        Defining the mitochondrial stress response in Arabidopsis thaliana.
        Mol Plant. 2009; 2: 1310-1324
        • Li X.
        • Zhang W.
        • Cao Q.
        • Wang Z.
        • Zhao M.
        • Xu L.
        • et al.
        Mitochondrial dysfunction in fibrotic diseases.
        Cell Death Discov. 2020; 6: 80