ATGL-mediated fat catabolism regulates cardiac mitochondrial function via PPAR-α and PGC-1

Abstract

Peroxisome proliferator-activated receptors (PPARs) are nuclear hormone receptors that regulate genes involved in energy metabolism and inflammation. For biological activity, PPARs require cognate lipid ligands, heterodimerization with retinoic X receptors, and coactivation by PPAR-γ coactivator-1α or PPAR-γ coactivator-1β (PGC-1α or PGC-1β, encoded by Ppargc1a and Ppargc1b, respectively). Here we show that lipolysis of cellular triglycerides by adipose triglyceride lipase (patatin-like phospholipase domain containing protein 2, encoded by Pnpla2; hereafter referred to as Atgl) generates essential mediator(s) involved in the generation of lipid ligands for PPAR activation. Atgl deficiency in mice decreases mRNA levels of PPAR-α and PPAR-δ target genes. In the heart, this leads to decreased PGC-1α and PGC-1β expression and severely disrupted mitochondrial substrate oxidation and respiration; this is followed by excessive lipid accumulation, cardiac insufficiency and lethal cardiomyopathy. Reconstituting normal PPAR target gene expression by pharmacological treatment of Atgl-deficient mice with PPAR-α agonists completely reverses the mitochondrial defects, restores normal heart function and prevents premature death. These findings reveal a potential treatment for the excessive cardiac lipid accumulation and often-lethal cardiomyopathy in people with neutral lipid storage disease, a disease marked by reduced or absent ATGL activity.

Figure 1

Expression of PPAR-α and PPAR-δ target genes and PGC-1α and PGC-1β in Atg/KO, Hs/KO, and wild-type tissues. mRNA expression levels for selected PPAR-α and PPAR-δ target genes and PGC-1α and PGC-1β were determined by RT-qPCR analysis. (a,b) Cardiac (a) and hepatic (b) mRNA expression of PPAR-α and PPAR-δ target genes were markedly decreased in fasted 8- to 10-week-old female Atg/KO mice compared to age-matched Hs/KO and wild-type mice. (c,d) mRNA levels of genes encoding PGC-1α and PGC-1β mRNA were also reduced in cardiac muscle (c) but increased in the liver (d) of fasted Atg/KO mice compared to wild-type mice. n = 4. Error bars show means ± s.d. *P < 0.05, **P < 0.01 and ***P < 0.001.

Figure 2

Morphology, glycogen content, mitochondria size and mitochondrial DNA content in cardiac muscle of wild-type and Atg/KO mice. (a) Cardiac muscle glycogen content (measured as glucose after hydrolysis) of 10-week-old female wild-type and Atg/KO mice (n = 9). (b) Transmission electron microscopy of cardiac muscle sections from 10-weeks old female mice. Top images, wild-type cardiac muscle sections show a typical intermyofibrillar network containing mitochondria (M), glycogen (*Gly) and lipid droplets (LD). In Atg/KO cardiac muscle (lower panels) lipid droplet size and the number of glycogen granules embedded within the intermyofibrillar network are increased. VE, vessel. Scale bars, 1 μm for upper and lower left images; 0.5 μm for upper and lower right images. (c,d) Morphometric (c) and cytofluorimetric (d) analyses of mitochondria from cardiac muscle of wild-type and Atg/KO mice. Size was either determined from sections of 100 randomly selected mitochondria per genotype or from isolated mitochondria (fluorescence-activated cell sorting (FACS) analysis, n = 4). AU, arbitrary units. (e) Relative mitochondrial DNA (mtDNA) content (normalized to the single-copy nuclear gene Ndufv1) in cardiac muscle of 10-week-old female wild-type and Atg/KO mice (n = 5). Error bars are means ± s.d. **P < 0.01.

Figure 3

Mitochondrial OXPHOS function and oxidative stress in cardiac muscle of wild-type and Atg/KO mice. (a,b) Oxygen consumption, an indicator for mitochondrial respiration, in Atg/KO cardiac homogenates of 4-week-old (a) and 8-week-old (b) male mice in the presence of glucose (n = 6). (c) Triglyceride (TG) content in cardiac muscle of wild-type and Atg/KO mice. (d,e) Oxygen flux of mitochondria isolated from cardiac tissue of 8- to 9-week-old male wild-type and Atg/KO mice. ADP-driven (state 3) and uncoupled (state U) oxygen flow was measured in the presence of pyruvate (d) and palmitoyl-CoA (e) in subsarcolemmal (SS) and in intramyofibrillar (IMF) mitochondria (n = 6). (f) Western blotting analysis of mitochondrial respiratory chain proteins NDUFA9 of complex I and SDHA of complex II in mitochondrial preparations of Atg/KO mice and wild-type mice. MTCO1, a marker of complex IV, served as loading control. (g) Mitochondrial membrane potential (tetramethyl-rhodaminemethylester perchlorate (TMRM) staining) in isolated cardiac mitochondria of 8- to 9-week-old female Atg/KO compared to wild-type mice (n = 4). (h) Relative concentrations of non-oxidized (free) thiol groups in isolated mitochondria of 8- to 9-week-old female Atg/KO mice compared to wild-type mice (n = 4). Error bars are means ± s.d. *P < 0.05, **P < 0.01 and ***P < 0.001.

Figure 4

Changes in PPAR-α and PPAR-δ activated gene expression and OXPHOS in mice lacking or overexpressing Atgl in cardiac muscle. (a) Cardiac triglyceride content in wild-type and conditional knockout mice lacking Atgl in cardiac and skeletal muscle (muscleAtg/KO mice) demonstrating a drastic cardiac steatosis in muscleAtg/KO mice (n = 5). Scale bars, 5 mm. (b) mRNA expression levels of PPAR-α and PPAR-δ target genes and of the gene encoding PGC-1α in cardiac muscle of muscleAtg/KO mice compared to wild-type mice (n = 5). (c–e) Heart weight (c), cardiac muscle triglyceride (TG) content (d), and white and brown adipose tissue (WAT and BAT) weight (e) of wild-type, Atg/KO and Atg/KO-cmAtg/TG mice expressing an Atgl transgene on an Atg/KO background (n = 6). (f) mRNA expression levels of PPAR-α and PPAR-δ target genes and genes encoding PGC-1α and PGC-1β in cardiac muscle of wild-type, Atg/KO and Atg/KO-cmAtg/TG mice (n = 4). (g) Oxygen consumption in cardiac homogenates prepared from 8- to 9-week-old female wild-type and Atg/KO-cmAtg/TG mice (n = 6). (h) Relative luciferase activities in lysates of HepG2 cells transfected with a PPRE-luciferase reporter plasmid and a PPAR-α expression vector. The additional expression of Atgl increases luciferase activity in the absence or presence of exogenously added linoleic acid (LA). Transfection of the bacterial β-galactosidase gene (lacZ)-containing plasmid and colorimetric determination of β-galactosidase (β-gal) enzyme activity was used for normalization of transfection efficiency. Error bars show means ± s.d. *P < 0.05, **P < 0.01 and ***P < 0.001.

Figure 5

Triglyceride content, oxygen consumption and cardiac function in Atg/KO mice treated with PPAR-α agonists. (a) Cardiac and hepatic triglyceride content in 6-week-old female Atg/KO mice on chow diet with or without 0.1% Wy14643 for 3 weeks (n = 5). (b) Cardiac and hepatic triglyceride content in 6-week-old female Atg/KO mice on chow diet with or without 0.2% fenofibrate for 10 weeks (n = 4–5). (c) mRNA expression levels of PPAR-α and PPAR-δ target genes and genes encoding PGC-1α and PGC-1β in cardiac muscle of female wild-type and AtglKO mice fed a chow diet with or without 0.1% Wy14643 for 3 weeks (n = 5). (d) Oxygen consumption in cardiac muscle preparations under both basal conditions and succinate-stimulated conditions of 9-week-old male wild-type, Atg/KO mice and Atg/KO mice fed a chow diet with 0.1% (wt/wt) Wy14643 for 3 weeks (n = 5). (e) Representative echocardiographic images (M- and B-Mode) of a 9-week-old female wild-type and Atg/KO mouse on chow diet and a 9-week-old female Atg/KO mouse fed a chow diet containing 0.1% (wt/wt) Wy14643 for 3 weeks. We measured interventricular septum (IVS) and posterior wall (PW) thickness from original tracings. We measured left ventricular end-systolic dimensions (ESD) and left ventricular end-diastolic dimensions (EDD) from original tracings according to the leading edge convention of the American Society of Echocardiography. (f,g) Left ventricular fractional shortening (LVFS) (f) and left ventricular (LV) mass (g), calculated from the echocardiographic tracings as previously described (n = 5). Error bars show means ± s.d. *P < 0.05, **P < 0.01 and ***P < 0.001.

Figure 6

Life span, tissue triglyceride content and energy substrate utilization in wild-type and Atg/KO mice treated with the PPAR-α agonist Wy14643. (a,b) Treatment of 8-week-old Atg/KO mice on chow diet containing 0.1% WY14643 for 12 weeks prevented cardiac death (a) and lowered tissue triglyceride (TG) content (b), including in cardiac muscle and liver, compared to that observed in wild-type animals (n = 4). (c) Relative whole-body oxygen consumption of 8- to 9-week-old female wild-type and Atg/KO mice housed in metabolic cages (n = 5). (d) Respiratory quotients (calculated from the ratio of carbon dioxide elimination versus oxygen consumption) in Atg/KO mice compared to wild-type during the light period and in the fasted state indicating preferential glucose utilization as oxidative fuel (n = 5). Error bars show means ± s.d. *P < 0.05, **P < 0.01 and ***P < 0.001. (e) Scheme of the integration of Atgl-mediated lipolysis in PPAR signaling. Fatty acids from exogenous or endogenous sources are not available as ligands for nuclear receptor signaling but instead are activated to acyl-CoAs and subsequently oxidized or esterified to triglycerides. Atgl-mediated lipolysis of triglyceride stores preferentially generates ligands or precursors of ligands for nuclear receptors controlling mitochondrial function and OXPHOS. CD36, cluster of differentiation 36; Fatp, fatty acid transport protein; FFA, free fatty acid; TGRLP, triglyceride-rich lipoproteins.