Cardiac-specific overexpression of perilipin 5 provokes severe cardiac steatosis via the formation of a lipolytic barrier.

Cardiac triacylglycerol (TG) catabolism critically depends on the TG hydrolytic activity of adipose triglyceride lipase (ATGL). Perilipin 5 (Plin5) is expressed in cardiac muscle (CM) and has been shown to interact with ATGL and its coactivator comparative gene identification-58 (CGI-58). Furthermore, ectopic Plin5 expression increases cellular TG content and Plin5-deficient mice exhibit reduced cardiac TG levels. In this study we show that mice with cardiac muscle-specific overexpression of perilipin 5 (CM-Plin5) massively accumulate TG in CM, which is accompanied by moderately reduced fatty acid (FA) oxidizing gene expression levels. Cardiac lipid droplet (LD) preparations from CM of CM-Plin5 mice showed reduced ATGL- and hormone-sensitive lipase-mediated TG mobilization implying that Plin5 overexpression restricts cardiac lipolysis via the formation of a lipolytic barrier. To test this hypothesis, we analyzed TG hydrolytic activities in preparations of Plin5-, ATGL-, and CGI-58-transfected cells. In vitro ATGL-mediated TG hydrolysis of an artificial micellar TG substrate was not inhibited by the presence of Plin5, whereas Plin5-coated LDs were resistant toward ATGL-mediated TG catabolism. These findings strongly suggest that Plin5 functions as a lipolytic barrier to protect the cardiac TG pool from uncontrolled TG mobilization and the excessive release of free FAs.

promoter ( Myh6 , GenBank accession number U71441) were generated by cloning full-length mouse Plin5 cDNA (amplifi ed from CM cDNA using the 5 ′ -ACA GGT CGA CAT GGA CCA GAG AGG TGA AGA CAC-3 ′ forward and 5 ′ -ACA GGT CGA CTC AAT GAT GAT GAT GAT GAT GGA AGT CCA GCT CTG GCA TCA-3 ′ reverse primers) in the ␣ -MHC promoter construct ( 31 ), kindly provided by J. Robbins as previously described ( 32 ). Characterized transgenic mice originated from a B6D2F2 background and were backcrossed four to fi ve times on a C57BL6 background. Littermates were used for phenotyping and mice with CM-Plin5 were hemizygous with respect to the integrated transgene. Animals were housed in a specifi c pathogen-free facility and maintained on a regular light-dark cycle (14 h light, 10 h dark) with ad libitum access to a standard laboratory chow diet (4.5% fat; ssniff Spezialdiäten, Germany) and water. For tissue collection, mice were euthanized by cervical dislocation and excised tissues were immediately snap-frozen. Maintenance, handling, and tissue collection from mice was approved by the Austrian Federal Ministry for Science and Research and by the ethics committee of the University of Graz.

Quantitative analysis of mRNA expression levels
Gene expression analysis was performed by quantitative reverse transcriptase polymerase chain reaction (RT-qPCR ). Total RNA was extracted with the TRIzol reagent (Invitrogen) and treated with DNaseI (Invitrogen). For fi rst strand cDNA synthesis, 1 g of total RNA was reverse transcribed at 37°C for 1 h using random hexamer primer (Applied Biosystems) and Superscript II reverse transcriptase (Invitrogen). Primers used for RT-qPCR were designed to span exon-intron boundaries with an amplicon size of less than 150 bp and BLASTed for specifi city. RT-qPCR reactions (20 l) contained 8 ng of cDNA, 10 pM of each primer, and 10 l of SYBR Green master mix (Fermentas) and were carried out using the ABI-StepOnePlus™ detection system (Applied Biosystems). Relative mRNA levels were quantifi ed using the comparative ⌬ ⌬ CT method with ␤ -actin as reference gene. The following primer sequences were used for RT-PCR: ␤ -actin forward, 5 ′ -AGC CAT GTA CGT AGC CAT CCA-3 ′ , reverse, 5 ′ -TCT CCG GAG TCC ATC ACA ATG-3 ′ ; murine Plin5 forward, 5 ′ -AGG GGA CTA GAC AAA TTG G-3 ′ , reverse, 5 ′ -GCT TCT CCG ACT TGC C-3 ′ ; Cpt1b (carnitine palmitoyltransferase 1 ␤ ) forward, 5 ′ -GGC ACC TCT TCT GCC TTT AC-3 ′ , reverse, 5 ′ -TTT GGG TCA AAC ATG CAG AT-3 ′ ; Ppara (peroxisome proliferator-activated receptor ␣ ) forward, 5 ′ -GTA CCA CTA CGG AGT TCA CGC AT-3 ′ , reverse, 5 ′ -CGC CGA AAG AAG CCC TTA C-3 ′ ; Acox1 (acyl-CoA oxidase 1) forward, 5 ′ -AGA TTG GTA GAA ATT GCT GCA AAA-3 ′ , reverse, 5 ′ -ACG CCA CTT CCT TGC TCT TC-3 ′ ; Acadm (acyl-CoA dehydrogenase, medium chain) forward, 5 ′ -GAT GCA TCA CCC TCG TGT AAC-3 ′ , reverse, 5 ′ -AAG CCC TTT TCC CCT GAA-3 ′ ; Acadl (acyl-CoA dehydrogenase, long chain) forward, TG mobilization from white and brown adipose tissue is a relatively well characterized process involving perilipin 1 (Plin1), adipose triglyceride lipase (ATGL), and hormonesensitive lipase (HSL) among many other proteins and cofactors ( 11,12 ). In contrast, much less is known about LD TG catabolism in nonadipose tissues, including muscle and liver, in which Plin1 is replaced by other PAT family members ( 7,13 ). More recently, perilipin 5 (Plin5) was found to be highly present on LDs of oxidative tissues including cardiac, skeletal muscle, and liver ( 14,15 ). Ectopic expression of Plin5 increased cellular TG levels and reduced FA oxidation (FAO) suggesting a role for Plin5 in energy catabolism (16)(17)(18)(19). Interestingly, Plin5 colocalized with ATGL and its coactivator comparative gene indentifi cation-58 (CGI-58) ( 17,18 ), two critical players in the fi rst and rate-limiting step of TG catabolism ( 20,21 ). Furthermore, the interaction of Plin5 with ATGL decreased ATGL-mediated lipolysis implying that Plin5 participates in the regulation of ATGL activity ( 22 ). Mutations of both ATGL and CGI-58 are causative for the development of neutral lipid storage disease with a different clinical picture: ATGL mutations are exclusively linked to muscle TG accumulation and lethal cardiomyopathy in humans and mice ( 23,24 ), whereas mutated CGI-58 alleles are primarily involved in the development of hepatic steatosis and a defect in skin barrier formation (25)(26)(27). The distinct expression pattern of Plin1 and Plin5 (in adipose and oxidative tissue, respectively) suggests that these proteins interact with LD-associated proteins and the lipolytic machinery in an organ-specifi c manner. A recent study showed that Plin5 recruits mitochondria to the proximity of LDs and that the inhibitory effect of Plin5 on basal LD-mediated FA release was partially reversed under stimulated conditions ( 17 ). Considering that Plin5 expression is induced by peroxisome proliferator-activated receptor (PPAR) ␣ ( 16, 28 ), a nuclear hormone receptor that regulates the expression of numerous oxidative phosphorylation genes, it may suggest that Plin5 couples LD FA release to mitochondrial FAO. However, the global deletion of Plin5 provoked a relatively mild phenotype in mice ( 29 ), although LDs were virtually absent in cardiac muscle (CM), similar to the lack of LDs in transgenic mice with cardiac-specifi c ATGL overexpression ( 30 ). Thus, it appears feasible that Plin5 primarily shields LDs from uncontrolled TG mobilization but is not critical for ATGL-mediated lipolysis and FA channeling to mitochondria.
The aim of our study was to examine the consequences of cardiac muscle-specifi c overexpression of perilipin 5 (CM-Plin5) on heart lipid and energy metabolism. We found that Plin5 transgenic mice exhibit severe TG accumulation in CM and that Plin5-coated LDs are resistant to ATGLmediated TG hydrolysis, suggesting that Plin5 acts as a lipolytic barrier to prevent uncontrolled TG mobilization.

Animals
Transgenic mice expressing mouse Plin5 cDNA under the control of the cardiomyocyte-specifi c ␣ -myosin heavy chain (MHC) Stimulation of ATGL-mediated TG hydrolysis was performed by addition of purifi ed murine GST-tagged CGI-58 (dissolved in 0.01% NP40) or COS-7 cell lysates containing His-tagged murine CGI-58 ( 20 ). Samples in a total volume of 100 l buffer A were incubated with 100 l substrate in a water bath at 37°C for 1 h. The micellar TG substrate contained 330 M triolein, 3 H-triolein as tracer, 45 M phosphatidylcholine:phosphatidylinositol (PC:PI, 3:1), and was prepared by sonication (Virsonic 475; Virtis, Gardiner, NJ).

Labeling and preparation of LDs
COS-7 cells were transfected with beta -galactosidase (LacZ)-or Plin5-expression vectors as described before. To promote LD formation, one day posttransfection cells were incubated overnight in medium supplemented with 0.4 mM oleic acid complexed to essentially FA-free BSA in a molar ratio of 3:1 together with 4 mCi 3 H-9, 10-oleate/mmol as radioactive tracer. For isolation of LDs, cells were trypsinized, centrifuged, and washed three times with phosphatebuffered saline. Thereafter, cells were suspended in buffer A and disrupted by sonication (Virsonic 475; Virtis). Cell lysates were transferred to SW41 tubes, overlaid with buffer B (50 mM potassium phosphate pH 7.4, 100 mM KCl, 1 mM EDTA, 20 g/ml leupetine, 2 g/ml antipain, 1 g/ml pepstatin), and centrifuged in an SW41 rotor (Beckman, Fullerton, CA) (2 h, 40,000 rpm, 4°C). LDs (visible as a white layer on top of the tube) were collected, transferred to a new tube, and concentrated by centrifugation (20,000 g , 15 min, 4°C) and removal of the underlying fl uid. Subsequently, LDs were resuspended in buffer B by brief sonication.

Determination of TG hydrolase activity using purifi ed LDs as substrate
LDs, prepared from COS-7 cells, were diluted to 0.05 mol TG/100 l (220 cpm/nmol) and 0.5% FA-free BSA in 100 mM potassium phosphate buffer (pH 7.0) was added. After incubation for 1 h, FA release from LDs was determined by extraction and determination of radioactivity essentially as described for the TG hydrolytic assay. LDs from cardiac tissue were isolated as described above and incubated with COS-7 cell lysates containing ATGL, HSL, CGI-58, or LacZ as control. LDs were diluted to a TG concentration of 0.4 mol/100 l and incubated in the presence of 0.5% FA-free BSA in 100 mM potassium phosphate buffer (pH 7.0) for 1 h at 37°C. The reaction was terminated by addition of 0.1% Triton X-100 followed by centrifugation at 20,000 g for 30 min. The lower phase was collected and FFAs were determined with a commercial kit (Wako Chemicals).

Analysis of TG levels of COS-7 cells expressing recombinant proteins
COS-7 cells were transfected with LacZ , Atgl , or Plin5 expression plasmids or cotransfected with both Atgl and Plin5 expression plasmids. To induce LD formation, COS-7 cells were loaded with 0.4 mM oleic acid and 4 mCi 3 H-9,10-oleic acid/mmol as tracer overnight (20 h). Total lipids were extracted and separated by thin-layer chromatography using hexane/diethyl ether/acetic acid (70:29:1) as solvent. TG-corresponding bands were excised and radioactivity was measured by liquid scintillation counting. For the analysis of the time-dependent incorporation of 3 H-labeled oleic acid into TG, LacZ and Plin5 expressing COS-7 cells were loaded with 0.4 mM oleic acid and 4 mCi 3 H-9,10-oleic acid/ mmol for time periods of 2, 4, 8, and 16 h. Lipids were extracted and analyzed as described above at the indicated time periods.

Plasma parameters
Blood samples were collected by retro-orbital puncture from isofl urane-anesthetized mice. Plasma parameters were analyzed with commercially available kits from Wako, Sigma, and Thermo Fisher Scientifi c and plasma glucose levels were determined using the Free-Style Freedom Lite® Blood Glucose Monitoring System (Abbott).

Tissue homogenization and lipid analysis
Snap-frozen hearts were homogenized in ice-cold lysis buffer A (0.25 M sucrose pH 7.0, 1 mM EDTA, 1 mM DTT, 20 g/ml leupetine, 2 g/ml antipain, 1 g/ml pepstatin) using an Ultra Turrax Homogenizer (IKA). The homogenates were centrifuged at 20,000 g for 30 min at 4°C and the infranatants were collected. Protein concentrations were determined using the Bio-Rad Protein Assay reagent (Bio-Rad Laboratories GmbH) and lipid extractions were performed according to the method of Folch ( 33 ). Aliquots of the organic phase were evaporated and the lipid extracts were resuspended in ice-cold 1% Triton X-100 by brief sonication. TG concentrations were then measured using a colorimetric kit (Infi nity TG reagent; Thermo Fisher Scientifi c).

TG hydrolase assay
TG hydrolase assays were performed as previously described ( 20 ). For the measurement of TG hydrolase activity cell lysates (1,000 g supernatant) or tissue extracts (10,000 g supernatant) were used. preparations ( Fig. 1E ) of CM from CM-Plin5 and wt mice, respectively. After ultracentrifugation, an LD layer was visible only in CM homogenates of Plin5 transgenic mice which was withdrawn for Western blot analysis. Plin5 protein signals were abundant in these LD fractions of Plin5 transgenic mice (Fig. 1E, outermost panel, right ). In contrast, Plin5 protein expression levels were similar in mitochondrial preparations of CM from both genotypes ( Fig. 1F ). As expected, Plin5 protein levels were unchanged in skeletal muscle (musculus quadriceps) of CM-Plin5 mice (supplementary Fig. I) implying that expression of the Plin5 cDNA under the control of the ␣ -MHC promoter was CM-specifi c.

Plasma parameters of CM-Plin5 mice are virtually unchanged
Next we examined the impact of cardiac Plin5 overexpression on plasma parameters ( Table 1 ). Plasma lipid and blood glucose levels of CM-Plin5 mice were similar compared with wt while plasma TG levels were moderately reduced ( Ϫ 17%) in nonfasted animals. These data implicate that cardiac Plin5 overexpression does not significantly affect systemic energy homeostasis.

Impaired ATGL-and HSL-mediated FA release of LD preparations from CM of CM-Plin5 mice
To examine whether the marked TG accumulation in CM of Plin5 transgenic mice involves changes in cardiac lipolysis, we performed TG hydrolytic assays using an established micellar triolein substrate emulsifi ed with phospholipids. Interestingly, TG hydrolytic activities were markedly increased in CM tissue extracts derived from nonfasted and fasted CM-Plin5 mice (1.4-fold in the nonfasted and 1.8-fold in the fasted state, respectively) compared with wt tissue extracts ( Fig. 2A ). Addition of the ATGL coactivator CGI-58 signifi cantly increased TG hydrolytic activities in cardiac extracts of both genotypes independent of the nutritional state ( Fig. 2B ). Notably, the highest differences in TG hydrolytic activities were observed in cardiac tissue extracts of nonfasted CM-Plin5 mice compared with wt mice upon addition of recombinant CGI-58 (1.8-fold). Next, we examined whether the observed changes in cardiac TG hydrolytic activities of CM-Plin5 mice involve differences in ATGL and/or CGI-58 protein expression levels. Western blot analyses revealed a marked increase in ATGL and CGI-58 protein levels in cardiac homogenates of transgenic mice ( Fig. 2C ). This fi nding indicates that the increased TG hydrolytic activities measured in CM extracts of transgenic mice ( Fig. 2A ) are due to augmented protein expression of ATGL and its coactivator CGI-58. Because CM-Plin5 mice exhibited massive TG accumulation in CM, we assumed that the increased in vitro TG hydrolytic activities may not refl ect the in vivo situation, where Plin5 is abundantly present on the LD surface.
To address this issue, we prepared LDs from CM of transgenic mice and used these LD preparations as a substrate for TG hydrolysis and measured the FA release in the presence of ATGL, ATGL together with CGI-58, HSL, and LacZ as control. Because LDs are in extremely low abundance in CM of wt mice, we used LDs from CM of ATGL-knockout C (0.25 M sucrose, 5 mM HEPES pH 7.7, 0.25 mM EDTA, 20 g/ml leupetine, 2 g/ml antipain, 1 g/ml pepstatin). After centrifugation (1,300 g , 4°C, 15 min), the infranatant was collected and mitochondria were pelleted (11,000 g , 4°C, 20 min) and resuspended in buffer C. CPT-1 activities were measured according to an established protocol ( 34,35 ).

Tissue LPL activity
Tissue LPL activity was measured essentially as described in Ref. 36 .

Transmission electron microscopy
Mice were euthanized at the age of 10 weeks by an overdose of anesthetic (xylazine-ketamine; Sigma) and immediately perfused with 4% (wt/vol) paraformaldehyde in 0.1 M phosphate buffer pH 7.4, for 5 min. CM was dissected using a Zeiss OPI1 surgical microscope (Carl Zeiss). Small tissue fragments were fi xed in 2.5% (wt/vol) glutaraldehyde and 2% (wt/vol) paraformaldehyde in 0.1 M phosphate buffer pH 7.4, for 2 h, postfi xed in 2% (wt/vol) osmium tetroxide for 2 h at room temperature, dehydrated in graded series of acetone and embedded in a TAAB epoxy resin. Thin sections (70 nm thick) were contrasted with uranyl acetate and lead citrate. Images were taken using an FEI Tecnai G 2 20 transmission electron microscope (FEI Eindhoven) with a Gatan UltraScan 1000 charge-coupled device camera. Acceleration voltage was 120 kV.

Statistical analysis
Data are presented as mean ± SD . Statistical signifi cance was determined by the Student's unpaired two-tailed t-test. Group differences were considered signifi cant for * P < 0.05, ** P < 0.01, and *** P < 0.001.

Cardiac-specifi c Plin5 overexpression provokes severe cardiac steatosis
Cardiac-specifi c overexpression of Plin5 was achieved by cloning the murine Plin5 cDNA downstream of the ␣ -MHC promoter and microinjection of the transgene DNA construct ( Fig. 1A ) into the pronucleus of mouse embryos. Measurement of Plin5 mRNA expression levels in CM of two founders showed a 14.8-fold and 50.6-fold increase in cardiac Plin5 mRNA levels of CM-Plin5 transgenic line (Tg)26 and Tg32, respectively, compared with wild type (wt) ( Fig. 1B ). CM-specifi c Plin5 overexpression caused massive cardiac steatosis ( Fig. 1C , right) and the magnitude of TG accumulation ( Fig. 1C , left) correlated with the degree of Plin5 mRNA expression levels in the transgenic lines. We then focused our characterization on Tg32 which is hereafter designated as CM-Plin5. The massive TG accumulation was also refl ected by the relative increase of heart weight in relation to body weight of nonfasted and fasted CM-Plin5 mice (1.8-and 1.6-fold, respectively) compared with wt ( Fig. 1D ). Body weights were unchanged in CM-Plin5 mice compared with wt mice (22.1 ± 1.2 g vs. 21.8 ± 1.1 g in nonfasted mice and 20.8 ± 1.4 g vs. 20.2 ± 1.3 g in fasted mice, respectively). Next, we analyzed Plin5 protein levels in cytosolic, LD, and mitochondrial fractions of CM. Plin5 protein levels were similar in cytosolic tions from ATGL-defi cient mice. These differences in LD TG mobilization upon addition of exogenous lipases are a strong indication that Plin5 could function as a lipolytic barrier and block the access of lipases to the TG substrate of Plin5-enriched LDs.

Plin5 blocks lipase access to the LD TG moiety but does not specifi cally inhibit ATGL enzymatic activity
To confi rm our hypothesis that Plin5 functions as a lipolytic barrier to preserve the LD TG pool from unrestricted TG catabolism, we used a cell culture approach mice instead . The comparison of the FA release of LD preparations from CM-Plin5 and ATGL-defi cient mice is particularly intriguing with regard to the severe TG accumulation in CM of both genotypes. TG hydrolysis of LD preparations of both genotypes was induced by the addition of cell lysates containing ATGL, ATGL together with CGI-58, HSL, and LacZ which served as reference ( Fig. 2D ). Yet, FA release was in general markedly lower in LD preparations from CM of Plin5 mice upon addition of lysates containing ATGL ( Ϫ 81.4%), ATGL combined with CGI-58 ( Ϫ 63.5%), and HSL ( Ϫ 56.7%) compared with that of cardiac LD prepara- incubated LacZ-and Plin5-transfected COS-7 cells with oleic acid and 3 H-labeled oleic acid to induce TG synthesis and LD formation. Western blot experiments confi rmed that the recombinant proteins were present in the investigated cell lysates and that LDs isolated from Plin5-transfected COS-7 cells contained large amounts of Plin5 ( Fig. 3C ). Then, in vitro assays were performed using isolated control LDs or LDs coated with Plin5 as substrate after the addition of lysates containing various recombinant proteins. FAs were effi ciently released from control LDs (prepared from LacZ-transfected COS-7 cells incubated with 3 H-labeled oleic acid) when incubated with an ATGL-containing lysate (7.3-fold), lysates containing ATGL and CGI-58 (70.1-fold), and a lysate containing HSL (40.9-fold) as compared with FA release upon addition of LacZ-containing lysates ( Fig. 3D ). Notably, FA release from Plin5-coated LDs was markedly reduced compared with that of control LDs when incubated with lysates containing ATGL ( Ϫ 16.6%), ATGL and CGI-58 ( Ϫ 62.4%), or HSL ( Ϫ 65.9%) strongly suggesting that Plin5 represents a lipolytic barrier that hinders lipase access to the TG substrate.
To investigate whether Plin5 interferes with ATGL-mediated TG hydrolysis per se, we performed TG hydrolytic activity assays using again the artifi cial micellar 3 H-labeled TG substrate. We assumed that addition of control LDs would competitively lower TG hydrolysis of the artifi cial TG substrate, while if Plin5 shields the access of lipases to TG, addition of Plin5-coated LDs would have less or even no effect on TG lipolysis. In fact, addition of control (nonradioactively labeled) LD preparations to the artifi cial 3 H-triolein substrate lowered TG hydrolysis of lysates containing ATGL or ATGL and CGI-58 by about 50% ( Fig. 3E ). In contrast, TG hydrolytic activities of lysates containing ATGL or ATGL and CGI-58 were mildly affected upon addition of Plin5-coated LDs, indicating that Plin5 in fact substantially limits the access to the TG substrate.

Moderately decreased FAO gene expression levels and reduced mitochondrial CPT-1 activity in CM of CM-Plin5 mice
Because mice lacking ATGL exhibit severe cardiac TG accumulation linked to defective PPAR ␣ -activated FAO gene expression, we hypothesized that cardiac Plin5 overexpression may similarly affect FAO gene expression levels. To address this hypothesis, we measured mRNA and examined TG hydrolysis and homeostasis under various conditions. First, we examined the in vitro impact of Plin5 on basal and ATGL-mediated TG hydrolysis using again the artifi cial 3 H-labeled triolein substrate. TG hydrolytic activities were measured in a combination of cell lysates containing LacZ, ATGL, and Plin5. We also included COS-7 cell lysates containing G0S2, a known inhibitory protein of ATGL ( 37,38 ). TG hydrolytic activities substantially increased in cell preparations containing ATGL (up to 11.6-fold compared with the LacZ control) independent of whether ATGL-containing cell lysates were mixed with lysates containing LacZ or Plin5 ( Fig. 3A ). In contrast, ATGL TG hydrolytic activity was markedly inhibited ( Ϫ 58.5%) by the addition of G0S2 cell lysates. These fi ndings demonstrate that ATGL enzymatic activity per se is not affected by the presence of Plin5 in in vitro assays. To investigate the in vivo impact of Plin5 on ATGL-mediated TG hydrolysis, we cotransfected COS-7 cells with LacZ and ATGL-or Plin5-expressing plasmids, loaded cells with oleic acid and 3 H-labeled oleic acid as tracer, and analyzed the incorporation of radioactivity into TG ( Fig. 3B ). COS-7 cells transfected with ATGL showed reduced incorporation of radioactivity into the TG pool ( Ϫ 33.0%) compared with LacZ, due to increased TG mobilization. In contrast, Plin5-transfected cells showed moderately increased incorporation of radioactivity into cellular TG (1.3-fold) and this effect was unchanged even if these cells were cotransfected with ATGL (1.4-fold). The increased radioactivity present in the TG fraction of Plin5-expressing cells could originate from increased lipogenesis or impaired TG catabolism. To address this question, we loaded LacZ-and Plin5-transfected COS-7 cells again with oleic acid and 3 Hlabeled oleic acid and measured radioactivity levels in TG at several time points over a period of 16 h. As shown in supplementary Fig. II, the incorporation of radioactivity constantly increased to a similar extent in LacZ-and Plin5transfected cells indicating that Plin5 is not involved in the lipogenic pathway.
Taken together, these data suggest that Plin5 protects TG from ATGL-mediated hydrolysis in vivo presumably by limiting the access of lipases to TG stores. To validate this hypothesis, we tested whether Plin5-coated LDs are resistant toward in vitro ATGL-mediated TG hydrolysis. We also included HSL in the experiment to address whether Plin5 may represent a general lipolytic barrier or if it is specifi cally shielding/inhibiting ATGL. Therefore, we Various parameters were assayed with a glucometer and commercial kits in plasma samples obtained from nonfasted and fasted 12-week-old female mice (n у 4, * P < 0.05). Comparable values were obtained from male mice (data not shown). TC, total cholesterol. (+55%) were similar, or even increased, in transgenic mice compared with controls ( Fig. 4A ). To further examine the impact of cardiac Plin5 overexpression on mitochondrial energy metabolism, we measured CPT-1 protein levels ( Fig. 4B ) and CPT-1 activity ( Fig. 4C ) in mitochondrial preparations derived from CM-Plin5 transgenic and wt mice, respectively. Mitochondrial CPT-1 protein signals and activity levels were decreased ( Ϫ 28.4% and Ϫ 28.5% , respectively) in CM of transgenic mice suggesting decreased mitochondrial FA uptake.

Marked divergences in cardiomyocyte morphology and LD size in cardiac tissue of CM-Plin5 mice compared with that of ATGL-defi cient mice
The severe TG accumulation in CM of CM-Plin5 and ATGL-defi cient mice together with impaired FAO gene expression in both mouse models prompted us to examine cardiomyocyte morphology in WT, CM-Plin5, and ATGL-defi cient mice, respectively. Therefore, mice were euthanized with anesthetic and immediately perfused with paraformaldehyde. Thin sections from fi xed cardiac tissue were stained with uranylacetate and lead citrate and examined by transmission electron microscopy. At fi rst view, there were obvious differences in cardiomyocyte morphology of all examined genotypes. While LDs were not present in CM sections of wt tissue ( Fig. 5 A-C ), there was a pronounced increase of uniform and hypertrophied LDs which were homogenously dispersed in cardiomyocytes of CM-Plin5 mice ( Fig. 5 D-F). Notably, ATGL-defi cient cardiomyocytes contained LDs of all sizes including giant LDs, and the cellular architecture appeared atrophic ( Fig. 5 G-I). A higher resolution revealed that cardiac mitochondria of CM-Plin5 mice were tightly attached to LDs ( Fig. 5 E, F) and the increased number of LDs seemed not to affect cellular integrity as assumed for ATGL-defi cient cardiomyocytes. Furthermore, mitochondrial appearance and cristae structure were similar in CM-Plin5 tissue sections compared with WT mice. In contrast, ATGL-defi cient cardiomyocytes exhibited changes in mitochondrial shape and cristae structure (see Fig. 5 H, I). The marked differences in cardiomyocyte morphology of CM-Plin5 mice compared with ATGL-defi cient mice suggest that the cardiac phenotype of CM-Plin5 mice is not as severe as in the hearts of mice globally lacking ATGL.

Reduced mRNA expression levels of lipoprotein lipase, CD36, and PPAR ␥ in CM of Plin5 transgenic mice
Finally, we examined whether changes in cardiac FA uptake and lipogenesis contribute to TG accumulation in CM of CM-Plin5 mice. Therefore, we measured mRNA expression levels of genes implicated in FA uptake and lipogenesis. LPL and the FA transporter CD36 are critical players in cardiac FA uptake ( 39 ) and the expression of both genes is regulated by PPAR ␥ . As shown in Fig. 6 , mRNA expression levels of lpl ( Ϫ 31%), cd36 ( Ϫ 33%) and pparg1 ( Ϫ 68%) are signifi cantly reduced in CM of CM-Plin5 mice compared with levels found in wt mice. In contrast, LPL activities were similar in CM of fasted Plin5 transgenic mice compared with wt indicating that the marked increase in cardiac TG levels of CM-Plin5 mice does not involve changes in cardiac FA uptake and lipogenic pathways. expression levels of PPAR ␣ , PPAR ␣ (and PPAR ␤ / ␦ ) target genes, and peroxisome proliferator-activated receptor ␥ coactivator (PGC) -1 ␣ and PGC-1 ␤ in CM of overnight fasted mice ( Fig. 4A ). Cardiac mRNA expression levels of Ppara ( Ϫ 22%) and several established PPAR ␣ and PPAR ␤ / ␦ target genes including Cpt1b ( Ϫ 32%), Acox1 ( Ϫ 23%), Acadl ( Ϫ 32%), and Acadvl ( Ϫ 18%) were moderately reduced in CM of CM-Plin5 mice compared with levels found in wt mice. Furthermore, expression levels of Ppargc1a and Ppargc1b were signifi cantly reduced ( Ϫ 52% and Ϫ 54%, respectively) ( Fig. 4A ) in CM of CM-Plin5 mice compared with controls. Nonetheless, mRNA levels of Acadm and Pdhk4 Fig. 2. Measurement of tissue TG hydrolytic activities and FAs released from cardiac LDs. A: Cardiac TG hydrolase activities were increased in nonfasted and fasted 12-week-old Plin5 transgenic mice compared with the activity of wt tissues (n = 5; *** P < 0.001 versus wt; § § P < 0.01; § § § P < 0.001 versus nonfasted). B: Addition of recombinant GST-tagged CGI-58 signifi cantly increased TG hydrolytic activities in both genotypes (n = 5; ** P < 0.01; *** P < 0.001 versus wt; § P < 0.05; § § § P < 0.001 fasted versus nonfasted). C: Protein expression levels of ATGL and its coactivator CGI-58 were markedly elevated in CM of CM-Plin5 mice compared with wt (12-weekold mice). Analyzed samples contained 30 g protein. Specifi c signals for ATGL and CGI-58 are indicated as arrows. GAPDH served as protein loading control. D: FA release of LD preparations isolated from CM of Plin5 transgenic and ATGL-defi cient mice when incubated with COS-7 cell lysates containing LacZ, ATGL, ATGL + CGI-58, and HSL. The ATGL-and HSL-mediated FA release was signifi cantly reduced in LD preparations of Plin5 transgenic mice compared with that of LDs from ATGL-defi cient mice (n = 3). Data are shown as mean + SD. * P < 0.05; ** P < 0.01; and *** P < 0.001 versus LDs incubated with LacZ containing lysates. 22,40 ) suggesting a role for Plin5 in the regulation of cardiac lipolysis. This assumption was also corroborated by the phenotype of Plin5-defi cient mice which exhibited an almost complete depletion of cardiac LDs ( 29 ).
Here we show that mice with cardiac-specifi c overexpression of Plin5 showed massive cardiac TG accumulation, similar to the cardiac phenotype of ATGL-defi cient DISCUSSION Defective lipolysis caused by mutations in ATGL alleles strongly affected cardiac energy catabolism in humans and mice and led to severe cardiac steatosis and dysfunction ( 23,32 ). More recently, Plin5, a member of the PAT family, was shown to interact with ATGL and its coactivator CGI-58 ( 18,  lipolytic barrier on the LD surface, because: i ) ATGL-and HSL-mediated TG catabolism was impaired in Plin5-enriched LDs; ii ) ATGL TG hydrolytic activity, as determined in in vitro assays, was not affected by addition of Plin5 mice. This phenotype suggests that ATGL-mediated lipolysis is impaired by Plin5 overexpression and inspired us to carefully examine cardiac TG hydrolysis in this transgenic mouse model. Surprisingly, in in vitro assays, we measured increased TG hydrolytic activities in heart extracts of Plin5 transgenic mice using an artifi cial triolein substrate ( 41 ). This fi nding was even more challenging with respect to the report that mice with cardiac-specifi c ATGL overexpression exhibit a severe decline in cardiac TG levels ( 30 ) suggesting that results of our in vitro assay may not refl ect in vivo cardiac lipolysis of CM-Plin5 mice. In accordance with this assumption, we found that isolated LDs from CM of Plin5 transgenic mice are a less effective substrate for exogenously added ATGL as compared with LDs isolated from CM of ATGL-defi cient mice. This reduced ability of ATGL to hydrolyze TG of LDs derived from CM of Plin5 transgenic mice could be a consequence of Plin-5 directly inhibiting ATGL enzymatic activity, or from the more general formation of a lipolytic barrier that hinders lipase access to the TG substrate.
Employing recombinant proteins and cell experiments, we examined the direct effect of Plin5 and Plin5-coated LDs on ATGL activity in in vitro assays as well as in living cells. We found that Plin5 apparently acts as a general Fig. 4. Measurement of cardiac mRNA levels of established PPAR ␣ and PPAR ␤ / ␦ target genes and mitochondrial CPT-1 activity. A: mRNA expression levels of PPAR ␣ , selected PPAR ␣ and PPAR ␤ / ␦ target genes, PGC-1 ␣ and PGC-1 ␤ were determined by RT-qPCR in CM RNA prepared from 12-week-old fasted Plin5 transgenic and wt mice, respectively (n у 5). * P < 0.05 and ** P < 0.01 versus wt samples. B: Western blot and densitometric analysis of CPT-1 protein expression levels in mitochondria preparations of 15-week-old fasted mice. COXIV served as loading control (** P < 0.01). Ten micrograms of mitochondrial protein were separated by SDS-PAGE prior to blotting. C: CM-specifi c Plin5 overexpression signifi cantly decreased CPT-1 activity in mitochondria preparations from CM of CM-Plin5 mice compared with that of controls (n у 5; ** P < 0.01 versus WT mice).   6. mRNA expression levels of LPL, CD36, and PPAR ␥ were reduced in CM of Plin5 transgenic mice whereas LPL activities were comparable to that of WT mice. A: mRNA expression levels were measured by RT-qPCR in cardiac tissue RNA derived from 12-week-old Plin5 transgenic and wt mice, respectively (n у 4). ** P < 0.01 and *** P < 0.001 versus WT mice. B: For the measurement of cardiac LPL-activities, hearts from overnight-fasted mice were surgically removed and minced in a medium containing heparin. LPL-activity of the supernatant was measured in duplicates. Values are shown as mean + SD of tissue samples from 4 mice of each genotype. TG substrate. Finally, Plin5 protein concentrations may generally differ among LDs and thus variably affect TG mobilization leading to increased TG accumulation of a distinct LD population thereby not generally impairing TG catabolism. In line with such an assumption, Plin5coated LDs did not affect the TG hydrolytic activity of the micellar TG substrate.
Plin5 protein is also present on mitochondria where the protein promotes the tight association of mitochondria and LDs thereby spatially connecting LD TG mobilization and FA release to mitochondrial FA uptake and oxidation ( 15,17,43 ). Notably, virtually every LD present in CM sections of Plin5 transgenic mice was in close proximity to one or more mitochondria strongly supporting the assumption that Plin5 recruits and tightly attaches mitochondria to the LD surface. Whether this observed phenomenon has an adverse or even benefi cial impact on TG mobilization and/or mitochondrial function in cardiomyocytes of Plin5 transgenic mice is currently unknown and an important question to be addressed in further studies.
In summary, the present study reveals an important role of Plin5 in cardiac TG homeostasis via the formation of a lipolytic barrier. Accordingly, mutations linked to increased Plin5 expression may also be involved in the development of cardiac steatosis and dysfunction in humans.
containing lysates alone; iii ) ATGL-mediated TG hydrolysis of an artifi cial triolein substrate was not affected by the addition of Plin5-coated LDs whereas control LDs competed with the hydrolysis of the artifi cial substrate; and iv ) in living cells, overexpression of Plin5 led to increased TG accumulation even if ATGL was coexpressed. To summarize, these data suggest that Plin5 forms a lipolytic barrier at the LD surface thereby impairing the access of ATGL and HSL to the TG substrate. Whether this Plin5-mediated barrier function mechanistically involves the inhibition of ATGL enzymatic activity per se when Plin5 and ATGL colocalize and interact on the LD surface needs further experimental clarifi cation.
Cardiac steatosis of ATGL-defi cient mice was mainly caused by a severe defect in PPAR ␣ -activated gene expression and the TG accumulation could be reversed by treatment with a PPAR ␣ agonist ( 32 ). Interestingly, the mRNA expression levels of selected PPAR ␣ and PPAR ␤ / ␦ target genes which have been implicated in mitochondrial FA uptake and oxidation, were moderately decreased in CM of CM-Plin5 mice when compared with the pronounced defect in PPAR ␣ -activated gene expression in CM of AT-GL-defi cient mice ( 32 ). Furthermore, cardiomyocyte morphology strongly differed in CM of Plin5 transgenic mice compared with the cellular architecture of ATGL-defi cient cardiomyocytes. Cardiac Plin5 overexpression provoked the accumulation of hypertrophied LDs of similar sizes which were homogenously distributed throughout the cytosol in tight association with mitochondria. In contrast, ATGL-defi cient cardiomyocytes appeared atrophic and showed a severe accumulation of heterogenous LDs, including giant droplets which strongly interfered with the myofi brillar architecture and mitochondrial shape. These divergences in cardiac morphology strongly imply that Plin5 overexpression does not severely impair ATGL-mediated TG catabolism and hence may moderately affect cardiac function. In accordance with this assumption, we actually have no evidence that Plin5 overexpression provokes an early lethal heart dysfunction reported for mice lacking ATGL (data not shown).
Ectopic Plin5 expression variably increased TG levels in several cell lines and affected mRNA expression levels of genes implicated in FAO ( 16,19 ). Notably, the induction of Plin5 expression particularly impaired TG mobilization under nonstimulated conditions and this effect could be partially reversed via forskolin-induced protein kinase A stimulation which was paralleled by increased Plin5 phosphorylation ( 22 ). Accordingly, stimulation of cardiac lipolysis may lead to Plin5 phosphorylation/modifi cation and a relaxation of the lipolytic barrier. Given that cardiac Plin5 overexpression mainly impairs lipolysis under nonfasted conditions, a more moderate cardiac phenotype would be predicted compared with the severe cardiac dysfunction of mice globally lacking ATGL. Along that line, we cannot exclude that changes in LD protein composition ( 42 ) and/or Plin5 interaction partners including CGI-58 and ATGL in response to increased energy demands possibly affect Plin5 modifi cation and its binding to interaction partners thereby allowing lipase access to the