Cardiac-specific ablation of ARNT leads to lipotoxicity and cardiomyopathy

Abstract

Patients with type 2 diabetes often present with cardiovascular complications; however, it is not clear how diabetes promotes cardiac dysfunction. In murine models, deletion of the gene encoding aryl hydrocarbon nuclear translocator (ARNT, also known as HIF1β) in the liver or pancreas leads to a diabetic phenotype; however, the role of ARNT in cardiac metabolism is unknown. Here, we determined that cardiac-specific deletion of Arnt in adult mice results in rapid development of cardiomyopathy (CM) that is characterized by accumulation of lipid droplets. Compared with hearts from ARNT-expressing mice, ex vivo analysis of ARNT-deficient hearts revealed a 2-fold increase in fatty acid (FA) oxidation as well as a substantial increase in the expression of PPARα and its target genes. Furthermore, deletion of both Arnt and Ppara preserved cardiac function, improved survival, and completely reversed the FA accumulation phenotype, indicating that PPARα mediates the detrimental effects of Arnt deletion in the heart. Finally, we determined that ARNT directly regulates Ppara expression by binding to its promoter and forming a complex with HIF2α. Together, these findings suggest that ARNT is a critical regulator of myocardial FA metabolism and that its deletion leads to CM and an increase in triglyceride accumulation through PPARα.

Figure 7. Ppara deletion reverses the cardiac dysfunction of csArnt^–/– mice.

(A) Representative m-mode echocardiographic images of hearts from control, csArnt^–/–, Ppara^–/–, and csArnt^–/– Ppara^–/– double-knockout mice 4 weeks after tamoxifen treatment. The experiment was repeated 7–11 times. (B–D) Echocardiographic data, including cardiac output (CO) (B), FS (C), and EF (D) from control, csArnt^–/–, Ppara^–/–, and csArnt^–/– Ppara^–/– double-knockout mice (n = 7–11 independent experiments). (E) Kaplan–Meier survival curve of csArnt^–/– (n = 35) and csArnt^–/– Ppara^–/– double-knockout mice (n = 44) mice. There is a significant difference between the 2 curves based on log-rank analysis. (F) Schematic model for ARNT regulation of cardiac FA metabolism. Data are presented as the mean ± SEM. *P < 0.01.

Figure 6. PPARα knockdown reverses the metabolic changes associated with Arnt deletion.

(A) mRNA levels of genes involved in FAU and FAO in NRCMs treated with control, ARNT, PPARα, and ARNT plus PPARα siRNA. MCAD, medium-chain acyl-CoA dehydrogenase (n = 6 independent experiments). (B) FAO as determined by oxygen consumption with the exogenous addition of palmitate in NRCMs treated with control, ARNT, PPARα, and ARNT plus PPARα siRNA (n = 6 independent experiments). (C) Myocardial TAG levels in control, csArnt^–/–, and csArnt^–/– Ppara^–/– mice (n = 6–8 independent experiments). (D) Myocardial FAO determined in isolated working hearts of csArnt^–/–, csArnt^–/– Ppara^–/–, or littermate controls using ³H-labeled palmitate (n = 5 hearts). (E) mRNA levels of gene targets of Ppara in control, csArnt^–/–, and csArnt^–/– Ppara^–/– hearts (n = 6 independent experiments). (F) Oil red O, TUNEL, MT, and H&E staining of hearts from control, csArnt^–/–, and csArnt^–/– Ppara^–/– mice. Scale bar: 100 μm. Experiments were repeated in triplicate. (G) Summary of TUNEL⁺ cells normalized to the number of nuclei in hearts from control, csArnt^–/–, and csArnt^–/– Ppara^–/– mice (n = 4 independent experiments). Data are presented as the mean ± SEM. *P < 0.01; ^#P < 0.01 for PPARα siRNA vs. ARNT plus PPARα siRNA.

Figure 5. ARNT regulates PPAR at the transcriptional level.

(A) Schematic presentation of the Ppara promoter depicting its HRE sequences. (B) Relative luciferase activity in cells treated with control or ARNT siRNA and transfected with a Ppara promoter-luciferase reporter plasmid or constructs with a truncated Ppara promoter. Renilla luciferase activity was used as an internal control and for normalization of transfection efficiency (n = 6 independent experiments). Data are presented as the fold change of luciferase activity in ARNT siRNA–treated cells over control siRNA–treated cells and normalized to that of the empty vector reporter. Trunc, truncation. (C) Results of ChIP on the second HRE upstream of the PPARα initiation site and using ARNT antibody (n = 3 independent experiments). (D) Changes in Ppara mRNA levels in response to knockdown of ARNT and its partners in NRCMs (n = 6 independent experiments). (E) Sequential ChIP studies using HIF2α and ARNT antibodies. The first ChIP was performed with HIF2α antibody, and the subsequent ChIP was conducted with the indicated antibodies (n = 3 independent experiments). (F) Luciferase activity of WT and a construct with the deletion of the second HRE in cells treated with either control or ARNT siRNA (n = 6 independent experiments). (G) Luciferase activity of WT and a construct with mutations in the second HRE in cells treated with either control or ARNT siRNA (n = 6 independent experiments). Data in F and G are represented as fold change of luciferase activity in ARNT siRNA–treated cells over control siRNA–treated cells and normalized to that of the empty vector reporter. Data are presented as the mean ± SEM. *P < 0.01.

Figure 4. csArnt^–/– hearts display increased FAO and PPARα levels.

(A) Myocardial FAO in control and csArnt^–/– hearts measured in isolated working hearts using ³H-labeled palmitate (n = 5–6 independent experiments). (B) FAU in NRCMs treated with control or ARNT siRNA (n = 6 independent experiments). (C) FAO in control or ARNT siRNA–treated NRCMs, assessed by oxygen consumption with the addition of exogenous palmitate (n = 5–6 independent experiments). (D) qRT-PCR of Pparg, Pparb/d, and Ppara in NRCMs treated with control or ARNT siRNA (n = 6 independent experiments). (E) Western blot analysis of PPARα expression in the hearts of csArnt^–/– and control mice (n = 5 independent experiments). (F) qRT-PCR of genes involved in FA metabolism in NRCMs treated with control or ARNT siRNA (n = 6 independent experiments). Data are presented as the mean ± SEM. *P < 0.01.

Figure 3. csArnt^–/– hearts display increased lipid accumulation.

(A) Representative cardiac histological sections of control (Arnt^fl/fl littermates) and csArnt^–/– mice 4 weeks after initiation of tamoxifen treatment. MT, Masson’s trichrome. Scale bars: 1 mm (H&E-stained ×4 images) and 100 μm (H&E-stained ×200 and MT-stained images). Experiments were repeated 3 times. (B) TUNEL staining of hearts from control and csArnt^–/– mice. TUNEL⁺ cells were normalized to the total number of nuclei and are represented as a bar graph next to the images (n = 6 independent experiments). Scale bars: 500 μm. (C) Electron microscopic images of control and csArnt^–/– hearts indicated the presence of fat droplets. Red arrows point to lipid vacuoles in the csArnt^–/– hearts. Scale bars: 1 μm (×4,800 images) and 500 nm (×9,300 images). (D) Oil red O staining of control and csArnt^–/– hearts. Scale bars: 500 μm (original magnification, ×100 and ×500 [insets]). Experiments were repeated 3 times. (E) Summary bar graph of images in D (n = 3 independent experiments). (F) TAG content in the heart of control and csArnt^–/– mice (n = 6 independent experiments). Data are presented as the mean ± SEM. *P < 0.01.

Figure 2. Deletion of Arnt in the heart leads to CM.

(A) Gross images of hearts from 3 sets of controls (Cre, Arnt^fl/fl on a normal chow diet, Cre Tg mice on a tamoxifen-treated chow diet, and Arnt^fl/fl mice on a tamoxifen chow diet) and csArnt^–/– mice. (B–E) Echocardiographic data on control (Arnt^fl/fl mice treated with tamoxifen) and csArnt^–/– mice (n = 12 independent experiments). FS (B), EF (C), LVID-d (D), and E/A ratio (E), as assessed by echocardiography (n = 12 independent experiments). FS and EF were measured 2 and 4 weeks after completion of tamoxifen treatment, while LVID-d and the E/A ratio were assessed at 4 weeks. LVID-d and the E/A ratio are summarized in bar graphs below the figures. Positive and negative dp/dt (F) and EDP (G) using invasive hemodynamic measurements in control and csArnt^–/– mice (n = 6 independent experiments). Data are presented as the mean ± SEM. *P < 0.01.

Figure 1. Generation of cardiac-specific knockout of ARNT.

(A) Western blot of ARNT in the hearts of 32-week-old db/db and age-matched control mice (n = 7 independent experiments for the bar graph). (B) Schematic diagram of Arnt deletion in the heart. Mice with loxP sequences flanking exon 8 of Arnt were crossed with Mcm Tg mice under the Myh6 promoter. (C) PCR analysis of genomic DNA of Arnt homozygous and heterozygous floxed mice with or without Cre and in the absence of tamoxifen. (D) PCR analysis of Arnt^fl/fl mice after crossing with Cre and tamoxifen treatment. DNA samples from 2 normal chow–fed and 2 tamoxifen chow–fed Cre Arnt^fl/fl mice are shown. (E) qRT-PCR results of Arnt transcript levels in control and csArnt^–/– hearts (n = 6 independent experiments). (F) Western blot analyses of ARNT in control and csArnt^–/– hearts (n = 6 independent experiments). A summary bar graph is shown below the blot. Data are presented as the mean ± SEM. *P < 0.01.