Pharmacological inhibition of adipose tissue adipose triglyceride lipase by Atglistatin prevents catecholamine-induced myocardial damage 

Abstract

Aims

Heart failure (HF) is characterized by an overactivation of β-adrenergic signalling that directly contributes to impairment of myocardial function. Moreover, β-adrenergic overactivation induces adipose tissue lipolysis, which may further worsen the development of HF. Recently, we demonstrated that adipose tissue-specific deletion of adipose triglyceride lipase (ATGL) prevents pressure-mediated HF in mice. In this study, we investigated the cardioprotective effects of a new pharmacological inhibitor of ATGL, Atglistatin, predominantly targeting ATGL in adipose tissue, on catecholamine-induced cardiac damage.

Methods and results

Male 129/Sv mice received repeated injections of isoproterenol (ISO, 25 mg/kg BW) to induce cardiac damage. Five days prior to ISO application, oral Atglistatin (2 mmol/kg diet) or control treatment was started. Two and twelve days after the last ISO injection cardiac function was analysed by echocardiography. The myocardial deformation was evaluated using speckle-tracking-technique. Twelve days after the last ISO injection, echocardiographic analysis revealed a markedly impaired global longitudinal strain, which was significantly improved by the application of Atglistatin. No changes in ejection fraction were observed. Further studies included histological-, WB-, and RT-qPCR-based analysis of cardiac tissue, followed by cell culture experiments and mass spectrometry-based lipidome analysis. ISO application induced subendocardial fibrosis and a profound pro-apoptotic cardiac response, as demonstrated using an apoptosis-specific gene expression-array. Atglistatin treatment led to a dramatic reduction of these pro-fibrotic and pro-apoptotic processes. We then identified a specific set of fatty acids (FAs) liberated from adipocytes under ISO stimulation (palmitic acid, palmitoleic acid, and oleic acid), which induced pro-apoptotic effects in cardiomyocytes. Atglistatin significantly blocked this adipocytic FA secretion.

Conclusion

This study demonstrates cardioprotective effects of Atglistatin in a mouse model of catecholamine-induced cardiac damage/dysfunction, involving anti-apoptotic and anti-fibrotic actions. Notably, beneficial cardioprotective effects of Atglistatin are likely mediated by non-cardiac actions, supporting the concept that pharmacological targeting of adipose tissue may provide an effective way to treat cardiac dysfunction.

Graphical Abstract

Open in new tabDownload slide

1. Introduction

Increased adrenergic signalling is a recognized cause of some forms of heart failure (HF) and is universally seen with reduced cardiac output.¹ In the heart, the sympathetic nervous system exerts its actions mainly via binding of catecholamines to β-adrenergic receptors (β-ARs).¹ While β-AR stimulation initially may compensate for reduced cardiac output, chronic overactivation is a major contributor to HF development and progression as it promotes cardiomyocyte apoptosis, pro-inflammatory signalling, and myocardial fibrosis.^2,3 Despite the wealth of available drug therapy including β-AR antagonists, the prognosis of HF remains poor and new therapeutic approaches are urgently needed.^4,5

In adipose tissue, stimulation of β3-AR is a major driver for lipolysis, the metabolic pathway through which triacylglycerol (TAG) is hydrolysed into fatty acids (FAs) and glycerol.⁶ Lipolytic activity in adipocytes is mediated by adipose triglyceride lipase (ATGL), hormone-sensitive lipase (HSL), and monoacylglycerol lipase (MGL). ATGL, the rate-limiting enzyme of lipolysis, is ubiquitously expressed and strongly activated by catecholamines (norepinephrine and epinephrine) via β3-AR, as well as by natriuretic peptides.⁶ Consistently, serum levels of FAs are increased in subjects with HF, which has been linked to overactivation of the sympathetic nervous system.^7,8 Given that HF is associated with impaired metabolic processes worsening the course of the disease,⁹ targeting the crosstalk between the heart and non-cardiac tissues may represent a novel approach for HF treatment.¹⁰ Considering the chronic activation of adipose tissue lipolysis during HF, pharmacological modulation of this process could be a promising therapeutic concept.¹¹

Our previous results indicated that selective inhibition of ATGL activity in white adipose tissue (WAT), using adipose tissue-specific ATGL-deficient mice (atATGL-KO) led to cardioprotective effects in a murine model of pressure-mediated HF.¹² Recently, a highly specific, complete inhibitor vof adipose and hepatic ATGL (Atglistatin, ATGLi) has been developed without exerting inhibitory actions in cardiac tissue.^13,14 Atglistatin was shown to protect mice against HFD-induced insulin resistance, glucose intolerance, and hepatosteatosis.¹⁴ Furthermore, Parajuli et al.¹⁵ showed that Atglistatin attenuates the echocardiographic decline of transverse aortic constriction (TAC)-induced cardiac dysfunction. However, underlying mechanisms of cardioprotective Atglistatin effects and the impact of Atglistatin on the crosstalk between adipose tissue and the heart remain unknown.

In this study, we show that Atglistatin prevents myocardial injury in response to overstimulation of β-AR by reducing myocardial fibrosis and apoptosis. In cell culture experiments, we demonstrate that Atglistatin modulates the lipolytic response to β-adrenergic signalling in adipocytes resulting in a reduced secretion of specific FAs that are responsible for cardiomyocyte apoptosis. Given that the role of adipose tissue and cardiomyocytes ATGL has been delineated, the novelty of our study is to demonstrate that pharmacological inhibition of ATGL-activity solely in adipose tissue improves cardiac function, in particular diastolic parameters, during catecholamine-induced cardiac damage. These beneficial actions are likely mediated by advantageous lipidomic changes leading to anti-apoptotic cardiac actions and subsequent prevention of cardiac fibrosis.

2. Methods

All data supporting the findings of this study are available within the article and data supplement or from the corresponding author upon request.

2.1 In vivo studies

All in vivo experiments were performed according to the German animal welfare act and were approved by the local authorities (Landesamt für Gesundheit und Soziales Berlin, Germany), and were conducted in accordance with the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals. Cardiac injury was induced by application of ISO as previously described.^16,17 Briefly, 6–9 weeks old male 129/Sv mice (Janvier Labs, France) received four repetitive subcutaneous injections of ISO hydrochloride (25 mg/kg BW/day, Sigma-Aldrich, Germany), saline served as vehicle (VEH). Atglistatin application was performed analogue to published protocol.¹⁴ Briefly, Atglistatin (kindly provided by R.B.) was mixed with control diet (M-Z extrudate, V1126; Ssniff Spezialdiäten, Germany) in a final concentration of 2 mmol kg⁻¹. Atglistatin application was started 5 days prior to first ISO injection. Mice were pair-fed depending on food intake of Atglistatin-fed mice.¹⁴ Animals were randomized by body weight into four groups (CTRL/VEH, CTRL/ISO, ATGLi/VEH, and ATGLi/ISO) and allocated to two different study protocols. For investigation of short-term treatment effects, a first subgroup of mice [n = 10 (CTRL/VEH), n = 11 (CTRL/ISO), n = 8 (ATGLi/VEH), n = 9 (ATGLi/ISO)] was sacrificed 2 days after the final ISO application. To analyse long-term treatment effects, a second group of mice [n = 7 (CTRL/VEH), n = 12 (CTRL/ISO), n = 13 (ATGLi/VEH), n = 10 (ATGLi/ISO)] was sacrificed 12 days after the last ISO injection. The blood serum and tissue samples were collected for further analysis. In vivo cardiac function was assessed using echocardiography one day before the animals were sacrificed.

An additional group of 129/Sv mice was first treated with VEH/ISO for 4 consecutive days to establish ISO-induced cardiac dysfunction and treated with Atglistatin after the last ISO injection [ATGLi/VEH (n = 10); ATGLi/ISO (n = 12)]. Analysis was performed similar to the second group of mice above and results were compared to controls (CTRL) of the second group. In addition, adipocyte-specific ATGL knockout mice (atATGL-KO)¹⁸ or wild-type littermates (WT) were repetitively injected with ISO (5 mg/kg BW) or saline as vehicle (VEH) for 2 weeks [atATGL-KO/VEH (n = 11), atATGL-KO/ISO (n = 15), WT/VEH (n = 9), and WT/ISO (n = 12)]. Mice were sacrificed and analysed 7 days after final ISO-injection, time-matched to experiments performed on 129/Sv mice. A small group (n = 3) of 4–6 weeks old male C57BL/6 mice (Charles River, Germany) was fed a 60 kcal% fat high-fat diet (HFD, Altromin Spezialfutter, Germany)¹⁹ for 15 weeks. Since 129/Sv mice are known to be resistant to HFD-induced obesity,²⁰ we used C57BL/6 mice for these experiments. An additional control group of C57BL/6 (n = 6), age-matched to 129/Sv mice, received a chow diet. Both groups were euthanized and heart samples were harvested for further TAG quantification. All animals were sacrificed by cervical dislocation under inhalative isoflurane anaesthesia (3% isoflurane, for ∼1 min).

2.2 Echocardiographic analyses

Echocardiography was performed according to a standard operation protocol.¹⁶ Briefly, mice were examined on a Vevo 3100 Imaging System equipped with a 30 MHz linear transducer (MX400; FUJIFILM VisualSonics Inc., Canada). Anaesthesia was induced by 3% isoflurane (in 80% oxygen). For image acquisition, isoflurane concentration was reduced to 1–1.5%, and adjusted to maintain comparable heart rates. B-Mode and M-Mode images were obtained, as described in the Supplementary material online. All images were analysed using Vevo LAB analysis software (FUJIFILM VisualSonics Inc., Canada). Global myocardial peak strain (rate) was semi-automatically assessed in B-mode images acquired from parasternal long-axis (longitudinal dimension) or short-axis view (radial and circumferential dimension) using Vevo Strain Software with integrated two-dimensional speckle-tracking algorithm. Three images with three cardiac cycles each were analysed for every image adjustment in M-Mode and strain analyses and subsequently averaged. Reported tissue velocities were measured in the apical four-chamber view at the septal mitral annulus. Transmitral flow patterns were recorded in the apical four-chamber view by pulsed-wave Doppler after guidance with colour Doppler, as described before.¹⁷

2.3 In vitro TAG hydrolase activity in cardiac tissue

TAG hydrolase activity was determined in cardiac tissue of CTRL/ISO- and ISO/ATGLi-treated mice, as described in Supplementary material online.

2.4 TAG quantification in heart tissue

TAGs were measured in heart tissue harvested from 129/Sv mice 12 days after last ISO injection, from age-matched C57BL/6 mice, and from C57BL/6 mice fed with a high-fat diet (HFD) for 15 weeks, using Folch’s protocol. Liquid chromatography/mass spectrometry (LC/MS) analysis was performed using Agilent 1290 HPLC coupled with a 6470 Triplequadrupole mass spectrometer (Agilent Technologies, Santa Clara, CA, USA) in positive electrospray mode, as described in Supplementary material online.

2.5 Histology

Paraffin-embedded cardiac cross-sections were deparaffinized and stained with haematoxylin for 8 min at room temperature. Collagen fibres were stained with Picrosirius-Red, as described in Supplementary material online. Digital images of sections were captured using a pathological slide scanner (Aperio CS2, Leica Biosystems, Germany). An image classifier algorithm (Aperio GENIE, Leica Biosystems) was trained to detect red-stained collagen fibres. Subendocardial and subepicardial collagen content was defined as proportion of collagen fibres in whole cardiac tissue.

Immunohistological analysis of macrophages marker MAC-3 and T cells marker CD3 was performed, as described in the Supplementary material online.

2.6 Ex vivo lipolysis of isolated adipose tissue explants

Ex vivo lipolysis assay was performed, as described previously.^19,21 Gonadal adipose tissue of 129/Sv mice (n = 6) was surgically removed and washed with PBS (GIBCO, USA). Tissue explants of each animal (∼20 mg, n = 16–32) were randomized into four groups (CTRL/VEH, CTRL/ISO, ATGLi/VEH, and ATGLi/ISO), and pre-incubated in pre-warmed DMEM with 4.5 g l⁻¹ glucose and L-glutamine (GIBCO, USA), containing 40 µM Atglistatin or DMSO (Sigma-Aldrich) for 8 h in a humified atmosphere of 37°C, 5% CO₂. Thereafter, fat explants were maintained in DMEM containing 2% FA-free BSA (Carl Roth, Germany) and 40 µM Atglistatin or DMSO in the presence of 100 µM ISO (Sigma-Aldrich) or PBS (control) and incubated for another 2 h. For measurement of adipose tissue lipolysis in CTRL/ISO- and ATGLi/ISO-treated mice (n = 6 per group) gonadal adipose tissue was removed 12 days after final ISO injection (second group). Tissue explants of each animal (n = 4) were pre-incubated in DMEM containing 2% FA-free BSA for 1 h. Subsequently, fat explants were maintained in DMEM containing 2% FA-free BSA in the presence of 10 µM forskolin or DMSO for 1 h. Nonesterified fatty acids (NEFA) and glycerol release were analysed using commercial kits, according to manufactures’ recommendations (HR-NEFA, Wako Diagnostics; Free Glycerol Reagent, Sigma-Aldrich). In addition, protein content of each fat pad was determined, as described in Supplementary material online. Total NEFA and glycerol content were calculated as a ratio of NEFA and glycerol to cell protein content. NEFA and glycerol release of each group were averaged per animal and then used as experimental unit.

2.7 3T3-L1 and HL-1 cell culture experiments

Experiments are specifically described in Supplementary material online. Three independent biological replicates were performed for every cell culture experiment.

2.8 Oil-Red-O staining

For TAG detection 3T3-L1-adipocytes and preadipocytes were washed with PBS and stained with Oil-Red-O solution [0.3% Oil-Red-O (SERVA, Germany) in 60% isopropanol] for 4 h at room temperature. Afterwards, the cells were washed in PBS and visualized using microscope.

2.9 FA-profiling

FA-profiling, performed of conditioned medium secreted by 3T3L1-adipocytes, is described in Supplementary material online. Briefly, 400 µL of conditioned cell medium was hydrolyzed with sodium hydroxide at 80°C in the presence of methanol and butylated hydroxytoluene. Subsequently, the mixture was neutralized with acetic acid and diluted with methanol (1:10 v/v). Next, the samples were analysed using an LC/MS system Agilent 1290/6470 in electrospray negative ionization mode. Stationary phase was a Kinetex C18 150 mm × 2.1 mm, 2.6 µm, mobile phase a gradient of water and acetonitrile with 0.05% acetic acid from 70% to 98% acetonitrile. Twenty-six individual FAs were detected in single ion monitoring and quantified using deuterated standards. The statistical analysis of the data is described in Supplementary material online.

2.10 Gene expression analysis

Total RNA (n = 7) from frozen cardiac tissue samples was extracted using RNeasy Micro kit according to the manufacture’s protocol (Qiagen, Germany). Total RNA from HL-1 cells was isolated using the NucleoSpin^® RNA II Kit (Machery-Nagel, Germany). cDNA was synthesized by reverse transcription using reverse transcriptase, RNAsin, and dNTPs (all Promega, USA). For gene expression, quantification real-time quantitative polymerase chain reaction (RT-qPCR) was performed on a CFX96 and CFX384 Real-Time PCR System (BioRad, USA) using the SYBR-Green technology. Relative gene expression was calculated by 2^−ΔΔCT method with Heat shock protein 90 α family (HSP90ab1, short-term treatment effect subgroup) or β-actin (long-term treatment effect subgroup, HL-1 cells) as housekeeping gene. The sequences of primers used are listed in Supplementary material online.

2.11 Pathway-focused RT-qPCR based screening of genes involved in cell death

Pathway-focused RT-qPCR analysis of cardiac samples obtained from short-term treatment subgroup (n = 3 per group) was performed (RT² Profiler™ Mouse Necrosis Array; Qiagen, Germany), according the manufacturer’s instructions. For each gene same threshold was used, CT values >35 were considered as no expression. Array contained a profile of 84 cell death-related genes, three reverse transcription controls, three PCR controls, one genomic DNA quality control, and five different housekeeping genes. Analyses were performed using the provided software ‘RT2 Profiler PCR Array Data Analysis’ (version 3.5, Qiagen). Relative gene expression was calculated by 2^−ΔΔCt method with HSP90ab1 as housekeeping gene. Clustergramm, heatmap, and principal component analysis of expression values (2^−ΔΔCt) were created using web-based software ClustVis.

2.12 Western blot

Western blot (WB) analysis was performed as previously described.¹² Briefly, tissue samples and HL-1 cardiomyocytes were lysed in RIPA buffer and lysates were separated by 10%, 12%, or 15% SDS-PAGE gels and blotted onto a PVDF membrane. Proteins were detected using antibodies (ABs) directed against ATGL, BAX, BAK, cleaved-caspase-3, p-HSL (Ser 563), p-HSL (Ser 660), HSL, MGL, Pan-Phospho-PKC, PKCα, PKCδ, GAPDH, RAN, and respective horseradish peroxidase-coupled secondary ABs, all described in the Supplementary material online. For detection, enhanced chemiluminescent reagents (ECL Western Blotting Reagents, GE Healthcare, USA) were used. Signal densities were analysed using Image Lab software (Bio-Rad, version 6.0.1).

2.13 Lactate dehydrogenase release assay

Lactate dehydrogenase (LDH) levels in medium were quantified using CyQUANT™ LDH Cytotoxicity Assay (Invitrogen, USA). The analysis was performed according to the manufacturer’s instructions.

2.14 Ms-based lipidome analysis

Lipid extraction and analysis of murine serum and heart tissue samples were performed at Lipotype GmbH as described previously.¹² Statistical analysis of the lipidome data is described in the Supplementary material online.

2.15 Statistical analysis

Data are presented as mean ± SEM. Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software, USA). Data were analysed for normal distribution using the Shapiro–Wilk test. Differences between two groups were analysed by two-tailed unpaired Student’s t-test. One-way or two-way analysis of variance (ANOVA) followed by Bonferroni post hoc test or Kruskal–Wallis test followed by Dunn post hoc test was used for comparison of more than two groups, as appropriate. A P-value of <0.05 was considered statically significant, and the following levels of statistical significance were used: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Grubbs’ test was performed to identify outliers. Group-size predetermination for in vivo experiments was based on our experience with the ISO model.^16,17

3. Results

3.1 Atglistatin mediates its cardioprotective effects by modulating FA-liberation from WAT

Persistent systemic catecholamine overstimulation induces a pronounced cardiac damage, often linked with fibrosis, inflammation, and apoptosis. On the other hand, catecholamines are known to induce pronounced adipose tissue lipolysis, resulting in the enhanced release of FAs to the circulation. Those FA/lipid species could potentially modulate the pathological effects of catecholamines on cardiomyocytes, participating in cardiac impairment.

Recently, we have demonstrated that adipocyte-specific deletion of ATGL, a rate-limiting enzyme, regulating lipolysis, had complex effects on the cardiac lipidome and prevented the development of pressure-mediated HF in mice.¹² In this study, we investigated whether transient pharmacological inhibition of ATGL in WAT using Atglistatin has any cardioprotective effects during the development of ISO-induced cardiac damage. Isoproterenol (ISO) is a synthetic catecholamine that specifically induces βAR-derived signalling effects, such as lipolysis in WAT and cardio-inotropic effects in the heart, when applied systemically to mice.^6,16,17 We first asked, whether FA release from WAT/adipocytes under ISO application could participate in the development of cardiac damage. To address this issue, we performed experiments with ISO and ATGL inhibitor (Atglistatin)-treated 3T3-L1 adipocytes (Figure 1A). Adipocytes were terminally differentiated (Figure 1B) using the protocol adopted from Witte et al.²¹ ISO stimulation of adipocytes increased lipolytic activity, as shown by the increased FFA and glycerol release into the culture medium (Figure 1Cand D). Atglistatin treatment significantly reduced this lipolytic activity in our model (Figure 1C and D). Next, we used an LC/MS-based lipid profiling approach to identify FAs liberated form adipocytes by ISO treatment (Figure 1E–G). Using bioinformatic analysis, we identified the three most abundant and highly up-regulated FAs, which were released into the medium upon ISO stimulation: palmitic acid (C16:0), palmitoleic acid (C16:1), and oleic acid (C18:1) (Figure 1E–G). Importantly, those ISO-regulated FAs were significantly reduced in the analysed medium of adipocytes concomitantly treated with ISO and Atglistatin (Figure 1E–G). Subsequently, we used the equimolar mix of palmitic acid, palmitoleic acid, and oleic acid (FA-Mix) for the stimulation of HL-1 cardiomyocytes. Application of FA-Mix resulted in cardiomyocytes death, as indicated by increased LDH levels in conditioned medium using a fluorometric LDH-assay (Figure 1H). TNFα, a potent inductor of apoptosis in cardiomyocytes, was used as a positive control (Figure 1H). Importantly, FA-Mix stimulation led to a significant induction of apoptosis, as demonstrated using WB-analysis of cleaved Caspase 3 (Figure 1I). Next, we measured the expression of apoptosis-related genes in cardiomyocytes treated with the FA-Mix (Figure 1J and K). Treatment of cardiomyocytes with the FA-Mix led to a significant increase of pro-apoptotic gene expression, such as BCL2-associated X protein (BAX), BH3 Interacting Domain Death Agonist (BID) (Figure 1J and K). Importantly, the FA-Mix-derived apoptotic response in cardiomyocytes was significantly attenuated by a pre-incubation with SSO (Sulfo-N-succinimidyl oleate),²² which is known as a potent pharmacological inhibitor of the main FA-transporter CD36²³ (Figure 1J). On the other hand, apoptosis in cardiomyocytes induced by FA-Mix was significantly enhanced by the parallel application of FA-Mix and ETO (etomoxir), a potent inhibitor of carnitine–acetyltransferase enzymes (CPT1) (Figure 1K). CPT1β is known as key enzymes responsible for the initiation of mitochondrial FA-oxidation. Both CPT1 enzymes are crucial regulators of β-oxidation in cardiomyocytes, and CPT1β-deficiency in mice has been recently shown to exaggerate lipotoxicity and cardiac hypertrophy in the TAC-model in mice.²⁴

In summary, these experiments indicate that ISO-stimulated adipocytes liberate a specific set of FAs consisting of palmitic acid, palmitoleic acid, and oleic acid, which was potently inhibited by Atglistatin. When applied to cardiomyocytes, these FAs induce a strong pro-apoptotic response likely involving CD36-mediated cellular FA uptake. Moreover, FA oxidation seems to be beneficial as FA-induced apoptosis was aggravated by simultaneous inhibition of CPT1.

3.2 Atglistatin inhibits lipolytic activity of adipose tissue without affecting cardiac lipolysis

Our study was focused on putative cardioprotective effects of transient pharmacological inhibition of ATGL in WAT using Atglistatin in the model of ISO-induced cardiac damage. In our experiments, we next focused on the inhibitory effects of Atglistatin on the lipolytic activity in adipose tissue (Figure 2A and B). Murine WAT-explants were treated ex vivo with ISO or vehicle in the presence of Atglistatin (Figure 2A and B). Ex vivo treatment of WAT-explants with Atglistatin completely abolished the ISO-dependent lipolytic activity, measured as a liberation of NEFA and glycerol into the explant medium. In addition, adipose tissue lipolysis was directly analysed after explantation of WAT isolated from ATGLi/ISO and VEH/ISO treated mice (Figure 2A). The lipolytic activity of ATGLi/ISO samples, measured as NEFA-release, was significantly reduced, when compared to VEH/ISO samples (Figure 2A).

We next measured basal expression of all three main lipolytic enzymes: ATGL, HSL, and MGL on mRNA (Figure 2C–E) and protein levels (Figure 2F-I) in murine adipose tissue, isolated from mice used in the main experiments (Figures 3 and 4). The WAT-specific expression level of all analysed lipases was not affected by Atglistatin- or ISO-treatment. Deficiency of cardiac ATGL was shown to result in lipid accumulation and severe cardiac dysfunction²⁵ and the expression level of ATGL seems to correlate with its enzymatic activity. To exclude the alterations of cardiac ATGL abundance and function induced by Atglistatin, we next analysed the expression and enzymatic activity of ATGL and HSL in heart tissue sections of the mice (Figure 2J–M). Importantly, ATGL expression and ATGL activity, measured in cardiac tissue were not reduced in the samples isolated from Atglistatin-treated mice, when compared with vehicle-treated controls. As shown in Figure 2K, also phosphorylation status of HSL (Ser 563 and 660) was not affected by systemic Atglistatin application. That implicated no feedback effects of Atglistatin-treatment on cardiac HSL-activity. We also observed a similar cardiac ATGL-activity, measured in cardiac tissues isolated from ATGLi/ISO or VEH/ISO-treated animals (basal activity) (Figure 2N). Importantly, the external addition of recombinant CGI-58, an ATGL co-activator, led to a significant but similar increase of TAG-hydrolase activity in both groups (ATGLi/ISO or VEH/ISO). This indicates clearly that ATGL-activity measured in heart samples of our animals is not affected by systemic in vivo Atglistatin-application. Taken together those experiments strongly imply that systemic Atglistatin application to mice does not inhibit cardiac ATGL activity. Accordingly, myocardial TAG accumulation was not significantly increased under Atglistatin treatment (Figure 2O). As control, TAG accumulation in mice fed an HFD over 15 weeks is shown. These experiments point towards an adipose tissue-targeted action of Atglistatin and make direct cardiac activities of this inhibitor unlikely. In summary, our results indicate that the oral Atglistatin treatment of mice offers an opportunity to investigate the role of adipose tissue lipolysis on myocardial dysfunction in the context of β-adrenergic overactivation.

3.3 Selective inhibition of ATGL in adipose tissue using Atglistatin ameliorates the development of ISO-induced cardiac damage

To assess the role of Atglistatin in the context of β-adrenergic overstimulation, 129/Sv mice were repetitively injected with ISO and analysed after 12 days. Five days prior to ISO application, oral administration of Atglistatin was started (Figure 3A). As shown in Figure 3B, 12 days after the last ISO injection, mice developed cardiac fibrosis, as assessed using cardiac tissue cross-sections, stained with picrosirius red, whereby the fibrotic lesions were mainly found in the subendocardial layer (Figure 3B–D). Development of ISO-induced cardiac fibrosis (Figure 3B and C) was accompanied by a strong up-regulation of the gene expression levels of both cardiac Collagen 1a1 and 3a1 (Figure 3Eand F). Atglistatin treatment of ISO-mice led to a significant reduction of the subendocardial fibrosis (Figure 3Band C). Consistently, cardiac expression of both Col1a1 and Col3a1 genes was significantly attenuated under Atglistatin treatment (Figure 3Eand F). The development of subendocardial fibrosis has been demonstrated to correlate strongly with the predominant disturbance of cardiac function.^16,17 We used echocardiographic analysis to identify the cardiac damage grade in our model. In consonance with our previous studies,^16,17 no differences in the global systolic pump function among all groups, assessed by ejection fraction (EF) and fraction shortening (FS) were observed (Figure 3G–I). Specific analysis of the myocardial deformation was evaluated using speckle-tracking-echocardiography (Figure 3J–N). We observed a markedly impaired global longitudinal strain (GLS) of the ISO-treated mice, when compared to VEH-treated animals (Figure 3K). More importantly, GLS was significantly improved by Atglistatin treatment. Similar to our previous data,¹⁶ GRS (global radial peak strain) and GCS (global circumferential peak strain), were not regulated under ISO treatment (Figure 3M and N), which indicates rather specific sub-endocardial damage under catecholamine application. We also investigated cardiac diastolic parameters using echocardiography (Figure 3O–Q). ISO-treatment resulted in increased E/e’ ratios as a sign of diastolic dysfunction and elevated left-ventricular filling pressures, which was also ameliorated by Atglistatin. Together, these data demonstrate that specific inhibition of ATGL’s lipolytic activity in WAT protects against ISO-induced cardiac damage and dysfunction.

To corroborate that the inhibition of adipocyte ATGL indeed results in an improvement of cardiac function under ISO application, we next applied ISO to adipocyte-specific ATGL-KO (atATGL-KO) mice and their wild-type littermate (control mice, CTRL), both of C57BL/6 background^12,18 (Supplementary material online, Figure S1). The animals were repetitively injected with ISO (5 mg/kg BW) or Vehicle (VEH) for 2 weeks and underwent echocardiography (Supplementary material online, Figure S1A). Similar to the mice treated with Atglistatin (Figure 3), atATGL-KO mice were protected from ISO-induced cardiac damage. The atATGL-KO mice exhibited reduced ISO-induced cardiac fibrosis, improved GLS and diastolic function (E/e’) (Supplementary material online, Figure S1B, E, and I). Systolic parameters (EF and FS), as well as GRS and GCS, were not regulated under ISO-application in all groups of mice (Supplementary material online, Figure S1C, D, and F–H). Together these data confirm that targeting of adipose tissue ATGL, either by pharmacological inhibition or genetic deletion, protects against ISO-mediated cardiac damage.

To explore a putative therapeutic action of Atglistatin treatment, we modified the in vivo study and applied Atglistatin after the induction of ISO-mediated cardiac dysfunction, as depicted in Supplementary material online, Figure S2A. Mice were first injected with VEH/ISO for 4 constitutive days followed by Atglistatin application until Day 17, according to the experiment demonstrated in Figure 3. We observed a significant increase in the development of subendocardial fibrosis in mice treated with ISO (Supplementary material online, Figure S2B). Atglistatin application significantly reduced this pro-fibrotic effect (Supplementary material online, Figure S2B). GLS was significantly elevated in the ISO/VEH group, as expected, but only mildly decreased in the ISO/ATGLi group (Supplementary material online, Figure S2F). Other echocardiographic parameters, such as EF, FS, GRS, and GCS, were not regulated (Supplementary material online, Figure S2D, E, G, and H). Taken together, those experiments indicated a moderate protective action of Atglistatin on ISO-induced cardiac damage, when Atglistatin is given after ISO application compared to the pre-treatment regimen.

3.4 Atglistatin improves early stages of ISO-induced cardiac fibrosis and inflammation in mice

To identify targets of Atglistatin being responsible for ameliorated late stage macroscopic fibrosis, we analysed mice 2 days after final ISO injection (Figure 4A). The histological analysis of cardiac fibrosis indicated a significant increase of collagen content under ISO stimulation. Nevertheless, we observed no significant differences between both the CTRL/ISO and ATGLi/ISO-treated groups (Figure 4B–D). At this early time point, we detected a strong induction of cardiac expression of both Col1a1 and Col3a1 under ISO treatment (Figure 4E and F). Importantly, Atglistatin application, similar to data presented in Figure 2E and F, led to a significant reduction of Col1a1 and Col3a1 expression (Figure 4E and F). Echocardiographic analysis indicated, analogous to the data obtained from the first cohort of mice (Figure 3H and I), no differences in EF and FS among the groups (Figure 4G and H), and a significant decline of GLS under ISO treatment. This was not the case in ATGLi/ISO-treated animals (Figure 4I). Taken together, these results indicate beneficial early effects of Atglistatin on the development of pro-fibrotic gene expression, likely attenuating later macroscopic collagen accumulation (Figure 3).

β-adrenergic overactivation has been shown to promote cardiac inflammation.² As shown in Figure 4L–N, cardiac inflammatory processes were strongly up-regulated in ISO-treated mice, as demonstrated by the increased gene expression of CD68 and enhanced MAC3-specific staining of infiltrating macrophages in cardiac tissue sections. Atglistatin treatment of ISO-mice led to a significant reduction of cardiac CD68 expression and mild, but not significant, reduction of MAC-positive cells in myocardial tissue samples (Figure 4L–N). To further investigate the putative anti-inflammatory effects of Atglistatin treatment in our model, we performed CD3-specific staining of cardiac tissue section and cardiac qRT-PCR analysis of several genes linked with inflammation and fibrosis (Supplementary material online, Figure S3). We observed a prominent enrichment of CD3 positive cells in cardiac tissue of ISO-treated mice (Supplementary material online, Figure S3A). Atglistatin treatment, however, did not affect the CD3-mediated inflammatory response. In addition, other markers of cardiac inflammation, such as TNFα, Monocyte chemoattractant protein-1 (MCP-1), or Toll-like receptor 2 (TLR2), were significantly induced by ISO, but Atglistatin did not affect this induction (Supplementary material online, Figure S3B, E, and I). The expression of Interleukin 1 receptor-like 1 (ST2/IL1RL1) and Galectin3, which correlates with inflammation and also with cardiac stress and fibrosis, was significantly elevated in the VEH/ISO group of mice. Atglistatin reduced the expression of both genes which was statistically significant for ST2 (Supplementary material online, Figure S3F and G). In summary, the pharmacological inhibition of ATGL in adipose tissue modulates early stages of cardiac fibrosis and has mild and distinct effects on cardiac inflammation.

3.5 Atglistatin attenuates myocardial pro-apoptotic effects induced by ISO application

Myocardial fibrosis and inflammation are closely linked with cardiac apoptosis (reviewed by Xia et al.²⁶) To investigate if Atglistatin modulates cell death of cardiomyocytes, we next performed an apoptosis/necrosis-directed expression array of cardiac tissue samples from mice shortly after ISO application. Principle component analysis and hierarchical clustering were performed (Figure 5A and B). The analysis revealed a robust up-regulation of a cluster of apoptosis-linked genes in cardiac samples from ISO-treated animals (Figure 5A and B). Among apoptosis-linked genes regulated by ISO application, many different TNFα-related genes were identified, such as TNFα-superfamily (TNFSF), FAS-associated protein with death domain (FADD), TNF-receptor type 1-associated death domain (TRADD), as well as BAX, BID, P53-Induced Death Domain Protein 1 (PIDD1), and NADPH Oxidase 4 (NOX4). Observed transcriptional activation of pro-apoptotic genes was completely abolished by the Atglistatin treatment (CTRL/ISO vs ATGLi/ISO) (Figure 5A and B). Additional RT-qPCR- and WB-analysis (Figure 5C–H), further confirmed these effects. Specifically, we performed RT-qPCR analysis of the expression levels of Calpain 6, BID, BAX, PIDD1, and NOX4. Interestingly, both Calpain 6 and NOX4 are involved in the regulation of oxidative-stress pathways leading to apoptosis. All analysed genes were significantly elevated in cardiac tissue of ISO-treated mice and significantly reduced in ATGLi/ISO group of mice (Figure 5C and G). Both BAX and BCL2-antagonist (BAK) were strongly up-regulated under ISO stimulation on protein level, and their expression was reduced in ATGLi/ISO samples (Figure 5D and E). We conclude from this set of experiments that systemic ISO treatment leads to pronounced pro-apoptotic effects in cardiac tissue. Atglistatin attenuated ISO-induced apoptosis revealing that inhibition of lipolytic activity in adipose tissue strongly ameliorates pro-apoptotic processes in the myocardium.

FAs are well known to induce downstream signalling effects by regulating cardiomyocyte PKC activity.²⁷ Next, we tested the involvement of the cardiac PKC-signalling pathway in our mice model. As presented in Supplementary material online, Figure S4A neither the pan-phosphorylation status of PKC nor the protein expression levels of PKCα and PKCδ were regulated in cardiac tissue samples. Complementary, we performed in vitro experiments using HL-1 cells stimulated with the FA-mix in the absence and presence of GO6983, a pan-inhibitor of PKC, and analysed the apoptotic response. We observed a mild reduction of the FA-induced cleaved caspase 3 expression by PKC inhibition that, however, this did not reach statistical significance (Supplementary material online, Figure S4B and C). Altogether, those experiments suggest that in the present models FA-mediated downstream signalling in HL-1 cardiomyocytes, and to some extent in our in vivo study, do not primarily involve PKC.

3.6 Atglistatin mediates its cardioprotective effects by regulating serum lipid content

Our cell culture experiments, presented in Figure 1, indicated that ISO-stimulated lipolysis in adipocytes regulates the release of specific FAs, and that Atglistatin treatment can potently inhibit that process. To follow that idea, we next asked, whether these processes are also reflected in changes of the serum lipid composition in VEH/ISO- and ATGLi/ISO-treated animals. Therefore, we performed MS-based serum lipidome analysis of mice, sacrificed shortly after the last ISO injection (Figure 6). A total of 303 lipid species from 13 different lipid classes were analysed in serum samples, as shown in Figure 6A and Supplementary material online, Table S2. The most abundant lipid classes in serum were cholesteryl esters (CE), cholesterol (Chol), lyso-phosphatidylcholine (LPC), phosphatidylcholine (PC), and triacylglycerol (TAG) (Figure 6A). Significant differences between the groups were detected in four lipid classes: Ceramide (Cer), Diacylglycerol (DAG), LPC, and lyso-phosphatidylethanolamine (LPE) (Figure 6A). In our further analysis, we focused on those lipid classes, which are known for their cardio-toxic effects,²⁸ such as Cer and DAG (Figure 6B and C). Both classes, Cer and DAG were elevated in the serum of mice treated with ISO whereby only the regulation of Cer reached statistical significance (Figure 6B and C). Atglistatin treatment significantly reduced the abundance of both cardio-toxic lipid intermediates, Cer and DAG, in the serum of ISO-treated mice (Figure 6B and C). Detailed evaluation of individual lipid species within the Cer and DAG class revealed mostly higher lipid species levels in the ISO group and lower levels in the Atglistatin-treated groups (Figure 6D). We identified three Cer species (40:1:2; 42:1:2 and 42:2:2), which were significantly down-regulated by the Atglistatin treatment (Figure 6E). Precise analysis of DAG species indicated an overall reduction of DAG-species abundance with predominant palmitic acid, palmitoleic acid, oleic acid, and C18:2 content (Figure 6D). In contrast to the significant down-regulation of the DAG class by Atglistatin, DAG-species were not significantly decreased by Atglistatin, suggesting that the accumulation of moderate reductions on the DAG species level accumulate in a significant reduction of the DAG class by Atglistatin.

In order to translate the regulation of toxic serum lipid intermediates to an alteration in cardiac tissue, MS-based analysis of cardiac Cer content was performed (Supplementary material online, Figure S5A and B). We focused on cardiac Cer, since Cer-species were most prominently regulated under Atglistatin in serum. At the class level, cardiac ceramides were only slightly reduced in ATGL/ISO group of mice. However, we identified three cardiac ceramide species (Cer 38:2:2, Cer 40:2:2, and 42:2:2) as significantly reduced in the ATGLi/ISO group of animals.

Taken together, we could detect a significant reduction of circulating Cer and DAG, as well as cardiac Cer levels under Atglistatin treatment. These data support a protective action of Atglistatin by regulating serum/cardiac lipid profiles, since both of those lipid classes were shown to induce cardiac lipotoxicity.

4. Discussion

In this study, we demonstrated that catecholamine-derived cardiac damage (apoptosis, inflammation, and fibrosis) depends on adipose tissue lipolysis, and that pharmacological inhibition of ATGL in WAT using Atglistatin, attenuated ISO-mediated deleterious processes in murine hearts.

4.1 Adipose tissue ATGL regulates cardiac lipid metabolism

This study confirmed the important role of ATGL-activity in adipose tissue for the regulation of cardiac function. In the context of heart (dys-)function, ATGL-activity in adipose tissue needs to be strictly separated from its non-adipose activity, e.g. in the heart. Global ATGL-deficient mice showed a reduced lifespan predominantly due to severe cardiac dysfunction, characterized by dramatically reduced systolic function.²⁵ Further studies on those mice showed also reduced mitochondrial substrate oxidation and respiration, leading to excessive cardiac lipid accumulation.²⁹ In addition, cardiomyocytes-specific ATGL-KO mice displayed robust cardiomyopathy, characterized by increased lipid accumulation in cardiomyocytes, cardiac hypertrophy, inflammation, and fibrosis.³⁰ Evident cardiolipotoxicity is also common for patients with neutral lipid storage disease, carrying mutation of the ATGL gene.^31,32 Although adipocyte ATGL-KO mice were reported to display mild obesity, low plasma NEFA levels, and increased insulin sensitivity,³³ the basal cardiac function of those animals was not affected by atATGL-deletion.^12,18 The cardiac phenotype of atATGL-KO mice was initially analysed in a model of exercise-induced physiological cardiac hypertrophy.¹⁸ Our previous study indicated that upon chronic training, atATGL-KO animals showed reduced adipose tissue lipolysis, which was linked to diminished cardiac FA-uptake/utilization, and attenuated development of adaptive physiological cardiac hypertrophy.¹⁸ Preventing that form of cardiac adaptation by adipocyte-specific ATGL deletion might be disadvantageous in terms of the physiological adaptation to exercise. The effect of atATGL on cardiac dysfunction was subsequently studied using a murine TAC-model.^12,15 WT mice, challenged with TAC-induced pressure overload, showed systemic up-regulation of circulating lipids as a response to catecholamine overload and subsequent increase of lipolytic activity in adipose tissue.¹² Moreover, those animals displayed pronounced cardiac fibrosis and significant reduction of systolic EF and FS. Interestingly, atATGL-KO mice, challenged with TAC, showed reduced lipolytic activity in WAT and attenuated cardiac damage with a completely recovered EF and FS¹². Our results are in accordance with the work published recently by others,^15,34 who demonstrated that atATGL-KO mice, as well as Atglistatin-treated animals were protected from TAC-induced systolic dysfunction and catecholamine-induced cardiac damage. The recent study proposed that atATGL inhibition results in an anti-inflammatory response, which mediates the cardio-protective action.³⁴

4.2 Atglistatin treatment improves catecholamine-mediated cardiac damage

In this study, we demonstrated that when used in an ex vivo lipolysis assay, Atglistatin was able to inhibit lipolytic activity of WAT-explants induced by catecholamines. Moreover, lipolysis was significantly reduced in WAT-explants from ISO/ATGLi-treated mice, when compared with ISO-treated animals. A recently published study demonstrated that oral application of Atglistatin to mice significantly inhibited atATGL-activity and protected the mice from HFD-induced insulin resistance, glucose intolerance, and hepatic lipid accumulation.¹⁴ Moreover, this study demonstrated that oral application of Atglistatin led to transient inhibition of ATGL predominantly in adipose tissue and liver. The authors also demonstrated that the starvation-induced plasma NEFA levels were significantly lower in the starved HFD-fed mice, refed for 2 h with HFD-ATGLi, when compared with HFD-VEH-fed controls. In our model, ISO was applied to mice for 4 days, and the application was finalized respectively 2 or 12 days before the animals were sacrificed. That is why we were not able to observe any direct ISO effect on plasma NEFAs level in our mice. In accordance, the results from our present study showed that Atglistatin application to mice predominantly affects adipose tissue. Interestingly, the expression profile of all three lipolytic enzymes (ATGL, HSL, and MGL) was not significantly regulated in WAT under Atglistatin treatment. We also investigated ATGL expression in cardiac tissue of mice at the mRNA and protein level. In our hands, Atglistatin neither induced ATGL-cardiac expression, nor regulated the cardiac TAG content. Moreover, TAG-hydrolase activity measured in cardiac tissue samples was not inhibited by Atglistatin application. These data are clearly in consonance with previous studies performed with HFD-fed mice treated with Atglistatin,¹⁴ which also indicated no significant accumulation and action of Atglistatin in cardiac tissue.

The mechanism, by which Atglistatin is applied systemically to mice, seems to exert WAT-specific inhibitory properties remains elusive. Tissue selectivity could rely on putative differences in ATGL activity among diverse tissue/cells types, or differences in the expression and/or activity of ATGL-regulating co-activators/co-repressors, such as CGI-58, perilipins, or G0/G1 switch gene 2 (GOS). In addition, Atglistatin was reported to inhibit ATGL activity up to 78%.¹³ That could explain, at least in part, selective and transient inhibitory effects Atglistatin in different organs/tissues.

A predominant part of the cardiac tissue consists of cardiomyocytes. We still cannot exclude that systemic Atglistatin application may inhibit ATGL-activity of other cardiac cell types, such as endothelial cells, immune cells, and/or fibroblasts. Diminished ATGL-activity in endothelial cells was recently linked to micro- and macrovascular endothelial dysfunction³⁵ and disturbed lipid droplet biogenesis.³⁶ Importantly, as demonstrated in the work performed by the group of Kratky et al.,³⁷ ATGL-activity plays also a crucial role in the regulation of pro-inflammatory effects in myeloid cells. Myeloid-specific ATGL-KO animals showed increased TAG accumulation in neutrophil granulocytes, when compared with wt controls. Experiments on neutrophils treated with FA/ATGLi indicated a strong enhancing impact of Atglistatin-activity on immune responses ex vivo. Nevertheless, the present experiments indicated only a mild effect of Atglistatin-treatment on cardiac inflammatory response in our model.

The mouse model of ISO-induced cardiac damage, applied in the present work, has been carefully characterized in our previous studies.^16,17 The ISO-induced cardiac injury seemed to be restricted to the subendocardium and linked with disturbed myocardial microstructure and function.¹⁷ We reported that under catecholamine application subendocardial cell loss, inflammation, and subsequent collagen deposition and fibrosis occur.¹⁶ The underlying mechanism of these processes is not completely understood. Others reported that catecholamine-induced cardiac damage was linked to its direct inotropic effects on cardiac function mediated by β1-AR. That led to an imbalance between increased transmural oxygen demand and reduced oxygen supply.^38,39 A recently published study demonstrated that a 2-week application of ISO results in pro-inflammatory effects in adipose tissue and heart that could be inhibited by inhibition of ATGL in adipose tissue.³⁴ Importantly, the characteristic of ISO-induced cardiac damage differs, depending on dose, duration, route of application, and mouse strain.^40–42 Wallner et al.⁴³ demonstrated alterations of cardiomyocyte membrane permeability by ISO application to mice (200 mg/kg, 6 days, C57Bl/6 mice) without apparent apoptosis. Other studies on rodents indicated pro-apoptotic effects of systemic ISO application, presumably mediated by catecholamine-induced cardiac oxidative stress.^44–46 Understanding the pathophysiological mechanism underlying catecholamine-derived cardiac damage is important, as chronic activation of the sympathetic nervous system strongly contributes to the aetiology of HF.¹¹ In this study, as a proof of concept, we also applied a modified ISO-model to atATGL-KO animals. As expected, atATGL-KO mice showed improved cardiac function and reduced cardiac fibrosis, similar to the pharmacological inhibition of ATGL by Atglistatin. To further determine the therapeutic potential of Atglistatin after ISO-induced cardiac damage has been established, we applied Atglistatin to mice one day after the last ISO application. Atglistatin also induced beneficial cardiac actions when given after the initial ISO-mediated damage. However, these effects were more moderate when compared with the pre-treatment regimen, which is possibly caused by the fact that the ISO-induced damage is manifested at an early stage in this model and can only be partially stopped by later treatment.

Systemic and prolonged loss of ATGL function may lead to severe deficits in ATP production due to a shortage of lipids/FA as an important myocardial source of energy. Importantly, several studies pointed towards a metabolic switch in myocardial energy supply due to the development of pathological cardiac hypertrophy and HF.⁴⁷ In those studies, myocardial energy production was demonstrated to be strongly dependent on glucose, and not lipids/FA. It is still an open question, whether that adaptation could be considered as a result of the pathology, or is rather an adaptive and necessary process to manage the hypertrophic/HF-dependent lack of efficient FA-utilization/FA-oxidation. It has been recently described that Atglistatin’s pharmacological activity, when given orally, is likely limited to the phase, in which the mice displayed increased food consumption during night phase.¹⁴ That may indicate a periodical and not constant ATGL inhibition in WAT, and may explain, at least in part, why the animals showed no obvious cardiac ATP-deprivation, but improved cardiac function in ISO-model of cardiac damage. Analysis of cardiac ATP production during Atglistatin treatment would finally answer this question, however, because of the technical limitations, we were not able to measure the total ATP production in cardiac tissue of our mice.

4.3 Atglistatin treatment improves ISO-induced cardiac damage by inhibiting lipid-mediated cardiac apoptosis

Our present study strongly supports the hypothesis that catecholamine-induced FA release from adipose tissue into the circulation mediates pro-apoptotic effects in the myocardium. More importantly, we have demonstrated that pharmacological inhibition of ATGL activity in adipose tissue leads to a significant attenuation of pro-apoptotic processes in cardiac tissue of ISO-treated animals. The striking down-regulation of selected apoptosis/necrosis-related genes in cardiac tissue of Atglistatin/ISO-treated animals provides the first hint to understand the observed cardioprotective effects of Atglistatin application in the model of catecholamine overactivation. The present study complements our previously published experiments, performed on atATGL-KO and control mice challenged with TAC-induced pressure overload.¹² Under TAC-induced pressure overload, control mice were characterized by enhanced WAT-lipolysis, linked with global changes of the myocardial lipidome, including phosphatidylethanolamine/phosphatidylcholine ratio (PE/PC ratio), and reduced cardiac mitochondrial cardiolipin abundance. Importantly, PE/PC ratio was previously suggested to play an important role in maintaining cell membrane integrity⁴⁸ and seems to be specifically important under pro-apoptotic/necrotic conditions in the heart.¹² Importantly, cell membrane permeability seems to be affected in ISO-treated animals as well, as discussed above.⁴³ Notably, atATGL-KO mice used in our previous study¹² showed a maintained PE/PC ratio during pressure-mediated HF. That suggests a significant influence of catecholamine-induced release of NEFA from WAT on cardiac lipid content and functionality.

To investigate the ISO-induced pro-apoptotic effects in detail, we next stimulated adipocytes with the combination of ISO/Atglistatin or ISO alone. Dual application of ISO/Atglistatin resulted in a marked reduction of lipolytic activity in adipocytes. LC/MS-based lipid profile analysis allowed us to identify a specific set of palmitic acid, palmitoleic acid, and oleic acid FAs, which were highly up-regulated upon ISO stimulation of adipocytes, and were entirely down-regulated by ISO/Atglistatin treatment. Equimolar application of palmitic acid, palmitoleic acid, and oleic acid (FA-Mix) to cultured cardiomyocytes led to a significant induction of apoptosis in those cells, which pointed out that serum FA-load could be responsible for the pro-apoptotic effects observed in our model in vivo. The general concept of FA-mediated lipotoxicity in cardiomyocytes characterized by apoptosis and necrosis has been investigated previously. Palmitic acid was able to induce mitochondrial dysfunction, oxidative stress, and apoptosis in cardiomyocytes.^49–51 Moreover, when applied to cardiomyocytes, FAs were reported to induce apoptosis mediated by accumulation of ceramides, reduced synthesis of cardiolipins, and consequent loss of mitochondrial membrane integrity and release of cytochrome C, all hallmarks of cellular BAX-BAK mediated apoptosis.^26,28 To the contrary, a certain set of FAs (myristic acid, palmitic acid, and palmitoleic acid) or palmitoleic acid alone were shown to induce physiological cardiac hypertrophy or hypertrophic response in vitro without fibrosis, inflammation, and apoptosis.^18,52 The study published by Adrian et al.⁵³ showed that only myristic acid was able to induce a hypertrophic response in mice and in cell culture experiments using neonatal cardiomyocytes. The authors identified palmitic acid as the FA with pro-apoptotic properties, mainly due to the increased activation of MAPK [stress-activated protein kinase (SAPK)/c-Jun N-terminal kinase (JNK)] pathway. Intriguingly, both myristic acid and palmitoleic acid, as well as the combination of all used FAs (myristic acid, palmitic acid, and palmitoleic acid) did not induce apoptosis in those cells.⁵³ Interestingly, oleic acid-supplementation was also shown to have some putative cardioprotective effects in mice models of HF. The study published by Lahey et al.⁵⁴ demonstrated that oleic acid increases TAG turnover and FA-oxidation rate through activation of PPARα, a key regulator of myocardial lipid metabolism. Taken together, the pro-apoptotic effects in cardiomyocytes vary dramatically, depending on the applied FAs (mono-, poly-, or unsaturated) and the experimental protocol used. Our results pointed to increased expression of pro-apoptotic genes, such as BAX and BAK, in cardiac tissue samples isolated from ISO-treated mice as well as in cardiomyocytes stimulated with FAs in vitro. This suggested that NEFA/FAs are able to activate the intrinsic/mitochondrial apoptotic pathway mostly by oxidative stress and BAX/BAK-dependent mitochondrial outer membrane permeabilization and consecutive cytochrome C release.²⁶ In accordance, our in vitro experiments demonstrated that FA stimulation led to release of cleaved caspase 3 and concomitant liberation of LDH indicating acute cellular damage. In addition, our in vivo data demonstrate that the observed pro-apoptotic effects induced by ISO application could also involve extrinsic apoptosis.²⁶ Multiple different TNFα-related genes, such as TNFSF, FADD, and TRADD, were significantly up-regulated by ISO, and strongly reduced in ATGLi/ISO-derived murine cardiac tissue. Taken together, our data suggest that FA-induced cardiac damage/apoptosis may result from catecholamine-induced adipose tissue lipolysis, a process potentially targeted by Atglistatin treatment.

In order to understand the molecular mechanism of FA-mediated apoptosis observed in our model, we used the pharmacological inhibitor of the CD36 transporter (SSO) in further experiments. SSO inhibited FA uptake in cardiomyocytes and significantly reduced FA-induced apoptosis. That also indicates that pro-apoptotic signals generated in response to the FA-Mix stimulation are not dependent on the extracellular action of FAs on putative membrane receptors, but rather involve the active transport of FAs via CD36 transporter. Similar results were published recently in a study conducted by Hua et al.²² on podocytes. The authors showed that podocyte-specific CD36 expression was strongly elevated in diabetic patients with hyperlipidaemia. When applied to podocytes in vitro, palmitic acid led to lipid accumulation, ROS production, and apoptosis. Those processes were significantly inhibited by SSO treatment, indicating that CD36 is involved in the cellular transport of palmitic acid.²² As described above, CD36 seems to play an important role in the regulation of myocardial lipid uptake, especially in patients with pronounced hyperlipidaemia. Interestingly, recent studies in mice with cardiomyocyte-specific CD36-overexpression in a TAC model indicated improved cardiac mitochondrial function of those mice.⁵⁵ In addition, our experiments indicated that inhibition of the β-oxidation in FA-stimulated cardiomyocytes enhances the pro-apoptotic response of FAs. As discussed in a review from Xia et al.,²⁶ inhibition of β-oxidation by blocking the CPT1-pathway accelerates the lipotoxic effects of FAs,^56,57 which could lead to enhanced apoptosis. Accordingly, impaired β-oxidation has been suggested as a key mechanism involved in the development of HF.⁵⁸ Interestingly, some FAs such as palmitic acid were recently demonstrated to impair catecholamine-induced cardiac signalling via PKC activation.²⁷ Two isoforms of PKC (PKCα and PKCδ) were demonstrated to be significantly induced by the palmitic acid stimulation in vitro.²⁷ In addition, increased PKC activity was described in the model of ISO-derived cardiomyopathy in rats.⁵⁹ Moreover, PKC is involved in the regulation of cardiac apoptosis.⁶⁰ Our experiments performed on HL-1 cells, stimulated with the FA-MIX, in conjunction with the pan-PKC inhibitor GO6983, indicated no significant regulation of apoptosis via PKC-signalling. We observed no regulation of phosphorylated-PKC and PKCα/PKCδ isoforms in protein extracts of cardiac tissue samples from our in vivo studies, which may be a result of the timing of tissue selection. Indeed, the PKC activity (abundance) was shown to be significantly up-regulated in the membrane fractions of cardiomyocytes. We used whole lysates of cardiac tissues, as the isolated cardiac samples were nitrogen-shock frozen directly after the mice were sacrificed. That is why we were not able to perform cell fractionation experiments. To address that issue more precisely, one would need to perform experiments applying our ISO-model to cardiomyocytes-specific PKCα- or PKCδ-deficient mice.

To support the relevance of an Atglistatin-mediated serum lipid regulation for its cardioprotective actions, we next performed MS-based lipidome analysis of serum samples. Both ceramides and DAG species, reported previously to mediate cardio-toxic effects, were elevated in the serum of ISO-treated mice, when compared with the ISO/ATGLi group of animals. More importantly, Atglistatin treatment led to an overall decrease of serum DAG abundance with predominant palmitic acid, palmitoleic acid, oleic acid, and C18:2 content, as well as a reduction of ceramide species (40:1:2; 42:1:2 and 42:2:2). Subsequently, Atglistatin treatment also reduced cardiac ceramide species levels. These data suggest that the anti-lipolytic actions of Atglistatin in adipose tissue result in a reduction of toxic lipid intermediates in serum which then translates into lower abundance of these toxic, pro-apoptotic mediators in the heart.²⁸

In summary, our study indicates that augmented adrenergic activation leads to development of cardiac apoptosis, inflammation, and fibrosis. The pharmacological inhibition of ATGL activity in adipose tissue improves cardiac function in a murine model of catecholamine-induced cardiac damage. The systemic actions of ISO on cardiac tissue are accompanied by a significant induction of adipose tissue lipolysis in WAT. These data implicate an adipose tissue/heart communication in the context of pathological conditions associated with catecholamine overload. Since the ISO-model displays some characteristic of a cardiac HF with preserved EF (HFpEF) phenotype, including cardiac ischemia, fibrosis, diastolic dysfunction, and preserved EF, these results may translate into a new preventive therapeutic approach for this disease. Taking into account all Atglistatin-related studies^14,15 and the present data one could postulate that Atglistatin-based treatment provides a novel pharmacological option for HF prevention including HFrEF and/or HFpEF by targeting non-cardiac tissue.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Author contributions

A.T. performed experiments, analysis of the data, participated in conception and design of experiments and drafting of the manuscript (MS). H.B. participated in design of apoptosis experiments and analysis of the data, as well as in writing of the MS. K.L., D.R., N.B., J.S.S., R.K., M.R., A.K.M., E.S., J.G., R.B., E.E.K., and G.F.G. contributed to design and research data analysis, performed experiments, or provided research material. M.S., N.W., and R.Z. reviewed/edited MS. U.K. and A.F.-L. participated in conception and design of experiments, wrote MS, contributed to discussion, reviewed/edited MS.

Acknowledgements

The authors thank Beata Hoeft for her excellent technical assistance and Annelie Blumrich (Institute of Pharmacology, Charité - Universitätsmedizin Berlin) for the help with mice experiments. The scientific figure illustrations were created using images adapted from Servier Medical Art by Servier (http://www.servier.com). Parts of this work will be used in the MD/PhD thesis of A.T.

Funding

This work was supported by DZHK (German Centre for Cardiovascular Research, BER 5.4PR) and by the DFG (Deutsche Forschungsgemeinschaft, DFG-KI712/10-1). A.T. was supported by research grants of the German Cardiac Society (DGK) and German Heart Foundation (DHS). J.S.S and M.S. are supported by the DFG (DFG–Schu 2546/5-1 to M.S.). N.W. is participant in the Clinician Scientist Program funded by the Berlin Institute of Health (BIH) and is supported by the DFG (DFG, Project number 394046635—SFB 1365), a grant from the Corona-Stiftung (Deutsches Stiftungszentrum, Essen, Germany), and by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (852796). E.E.K. received research support from NIH R01 DK090166. R.Z. is an Einstein BIH Visiting Fellow supported by the Foundation Charité; U.K. is supported by the DZHK; BER 5.4 PR, the DFG (DFG—KI 712/10-1), the BMBF/BfR1328-564m, and the Einstein Foundation/Foundation Charité (EVF-BIH-2018-440).

Data Availability

The lipidomic datasets reported in this article are available in Supplementary material online, Tables S2 and S3.

References

Author notes