PPARα-Sirt1 complex mediates cardiac hypertrophy and failure through suppression of the ERR transcriptional pathway

Abstract

High energy production in mitochondria is essential for maintaining cardiac contraction in the heart. Genes regulating mitochondrial function are commonly downregulated during heart failure. Here we show that both PPARα and Sirt1 are upregulated by pressure overload in the heart. Haploinsufficiency of either PPARα or Sirt1 attenuated pressure overload-induced cardiac hypertrophy and failure, whereas simultaneous upregulation of PPARα and Sirt1 exacerbated the cardiac dysfunction. PPARα and Sirt1 coordinately suppressed genes involved in mitochondrial function that are regulated by estrogen-related receptors (ERRs). PPARα bound and recruited Sirt1 to the ERR response element (ERRE), thereby suppressing ERR target genes in an RXR-independent manner. Downregulation of ERR target genes was also observed during fasting, and this appeared to be an adaptive response of the heart. These results suggest that suppression of the ERR transcriptional pathway by PPARα/Sirt1, a physiological fasting response, is involved in the progression of heart failure by promoting mitochondrial dysfunction.

Figure 1

PPARα and Sirt1 are involved in pressure overload(PO)-induced cardiac hypertrophy. (A) PO induces co-upregulation of PPARα and Sirt1. Nuclear fractions were prepared from wild type mice subjected to transverse aortic constriction (TAC) for 4 weeks. Anti-histone H3 and tubulin antibodies were used to show nuclear and cytosolic fractions, respectively. (B) Sirt1 was co-immunoprecipitated with PPARα from primary cultured cardiac myocytes. Open asterisk represents IgG. (C–E) Haploinsufficiency of PPARα and Sirt1 attenuates PO-induced cardiac hypertrophy (C–D) and systolic dysfunction (E). PPARα^+/− and Sirt1^+/− mice were subjected to 4 weeks of TAC. (C) The heart weight/body weight ratio was measured. Heart lysates were subjected to Western blot analyses with anti-PPARα and anti-Sirt1 antibodies. (D) After 4 weeks of TAC, cell size was measured using WGA staining. Representative images are shown. (E) After 4 weeks of TAC, ejection fraction was measured by echocardiography. (F) Haploinsufficiency of Sirt1 normalizes PPARα-induced cardiac hypertrophy. (G–H) Co-overexpression of PPARα and Sirt1 induces cardiac hypertrophy. Heart weight/body weight ratio and myocytes cell size were measured. (I) Ventricular dysfunction in Tg-PPARα/Tg-Sirt1 bigenic mice was evaluated by echocardiographic measurements. Representative M-mode echocardiographic images of the left ventricle and ejection fraction are shown. The numbers of mice examined in each experimental group were: 4–12 (C), 3–8(D), 6–14 (E), 3–6 (F), 9–21 (G), 3 (H), and 7–19 (I). Error bars represent S. E. M.

Figure 2

The effect of Sirt1 on PPARα activity. (A) Relative mRNA expression of PPARα target genes in transgenic mice. (B) The effect of Sirt1 and PPARα on PPRE-luciferase reporter activity in vivo. Tg-PPRE-luc mice were crossed with Tg-PPARα, Tg-Sirt1, Tg-PPARα/Tg-Sirt1 or control mice (Left). Mice were subjected to 24 hours of fasting (Right). Luciferase assays were performed using heart lysates prepared from these mice. (C) A heat map of known PPARα target genes that were upregulated in Tg-PPARα in microarray analyses. Gray indicates undetectable expression. (D–E) PPARα-induced neutral lipid accumulation was attenuated by Sirt1. Oil red O staining (D) and triglyceride quantification (E) were conducted in each mouse genotype. (F) PPARα-induced increases in H₂O₂ were attenuated by Sirt1. The numbers of mice examined in each experimental group were: 4–8 (A), 4–7 (B), 4–5 (E) and 6–9 (F). Error bars represent S. E. M.

Figure 3

Gene expression profiles of ERR target genes in transgenic mice. (A) A heat map representing the expression profile of ERR target genes in transgenic mice. Known PPAR target genes are indicated by cross mark. Relative gene expression levels of ERR target genes regulating (B) mitochondrial metabolism, (C) contractile work, (D) transcription factors and (E) PGC-1 and ERR isoforms in the indicated transgenic animals are shown (N=8–16). Error bars represent S. E. M.

Figure 4

Mitochondrial function was impaired in Tg-PPARα/Tg-Sirt1 bigenic mice. (A) Electron microscopic analysis of the heart in transgenic mice. (B) Quantitative morphometric analyses of cardiac mitochondrial ultrastructural abnormalities. (C) Mitochondrial volume of the transgenic mice. (D) ATP synthesis in cardiac mitochondria isolated from the transgenic mice (N=7–10). Error bars represent S. E. M.

Figure 5

PPARα binds to the single hexad motif within the ERRE. (A) Schematic representation of topology of PPAR and RXR on the PPRE, and comparison of the PPRE and the ERRE. (B–E) Recombinant PPARα (C) and PPARα in Tg-PPARα heart lysate (D–E) binds to both the PPRE and the ERRE. PPARα was pulled down using biotin-labeled double-stranded DNA with the indicated sequences (B). Non-biotin labeled competitors (1 to 10-fold of biotin-labeled DNA) are shown in blue (C–D). (F) Endogenous PPARα was pulled down with 3×ERRE after 4 weeks of TAC. (G) PPARα binds to the single hexad motif in the heart. ChIP assays were performed with homogenates from Tg-PPARα and Tg-PPARα/Tg-Sirt1 hearts and anti-Flag antibody. Target single hexad motifs are shown in red (AGGTCA and AGGACA) and blue (one nucleotide redundancy in the last 3 letters from AGGTCA or AGGACA) with proximal 7bp sequences. Genomic regions amplified by quantitative PCR are indicated by nucleotide number from first codon under target genes (N=6–12). (H) PPARα recruits Sirt1 to the flanking region of the single hexad motif. ChIP assays were performed in Tg-PPARα hearts using anti-PPARα and anti-Sirt1 antibodies (N=8). (I) Sirt1 recruited by PPARα suppresses gene expression. Quantitative PCR was performed using cDNA prepared from the indicated transgenic mice (N=5−20). Error bars represent S. E. M.

Figure 6

The role of RXR in PPARα/Sirt1-mediated ERRE suppression. (A) PPARα and Sirt1 cooperatively suppress ERR target gene promoters. (B) The ERRE is an important interface for PPARα/Sirt1-mediated gene suppression. (A–B) Schematic representations of each promoter are shown in these panels. The common hexad motifs of HREs are indicated by a red line (AGGTCA and AGGACA), blue line (one nucleotide redundancy in the last 3 letters from AGGTCA or AGGACA) or black line (ERRE mutant: TCGGAATCA). (C) Sirt1 is a crucial partner for PPARα-induced ERRE suppression. (D) Transcriptional activation of PPARα is not essential for ERRE suppression. Schematic representation of PPARα and PPARαΔAF2 are shown. AF1: Activation function 1, DBD: DNA binding domain, LBD: ligand binding domain, AF2: activation function 2. HA-Sirt1 was co-immunoprecipitated with both Myc-PPARα and Myc-PPARαΔAF2. (E) PGC-1α fails to normalize PPARα/Sirt1-mediated ERRE suppression. (F) RXR is not necessary for interaction between PPARα and Sirt1. RXR was removed from heart lysates by immunodeprivation. PPARα was then immunoprecipitated with anti-Flag-antibody. (G) The effect of an RXRα mutant lacking the DNA binding domain (RXRα(ΔN)) on PPARα-induced PPRE activation and ERRE suppression. (H) The effect of full-length RXRα on PPARα-induced PPRE activation and ERRE suppression. (I) The dosage effect of PPRE-reporter on PPARα-induced reporter activity. The indicated amounts of pPPRE-luc reporter plasmids were transfected together with PPARα and RXRα expression vectors (0.3 μg). (J) The dosage effect of PPRE-reporter on PPARα/Sirt1-mediated transcriptional regulation. The mean value for activity with PPARα expression alone was expressed as 1. (K) Knockdown of Sirt1 enhances PPARα-induced PPRE activation. (L) RXRα counteracts PPARα/Sirt1-mediated suppression of PPRE reporter activity. (A–F and G–L) Luciferase assays were performed after 1 (A–B, D–E, and G–L) and 3 days (C and K) of transduction and transfection with the indicated constructs (N=6–12). Error bars represent S. E. M. (M) Schematic representation of the data shown in this figure. PPARα-induced PPRE activation is enhanced by RXRα but prevented by RXRα(ΔN). In contrast, PPARα/Sirt1-mediated ERRE suppression is not affected by RXRα or RXRα(ΔN) (Left). When RXR does not cover the half core site of the PPRE, the PPARα/Sirt1 complex suppresses transcription through the single hexad unit within the PPRE (Right).

Figure 7

Involvement of PPARα/Sirt1-mediated ERRE suppression in pressure overload(PO)-induced cardiac hypertrophy and fasting. (A–B) After 4 weeks of TAC, ChIP assays were performed with anti-Sirt1, anti-PPARα and anti-histone H3 acetyllysine 9 specific antibodies (A) or anti-ERRα antibody (B) (N=4–8). PPARα and Sirt1 bind to the single hexad motifs of HREs in response to PO in the heart (A), whereas the occupancy of ERRα on the ERRE decreases (B). (B) The control region corresponds with Pgc1a (−137 to 0). The number above the control IgG column represents raw data of the copy number (×1000) (C) Haploinsufficiencies of PPARα and Sirt1 attenuate PO-induced downregulation of ERR target gene expression. The expression levels of the indicated genes were examined by quantitative PCR (N=4–12). (D) PPARα/Sirt1-mediated ERRE suppression is a fasting response. PPARα^+/− and Sirt1^+/− mice were subjected to fasting for 24 hours. The expression levels of the indicated genes were examined (N=8–20). (E) The NAD/NADH ratio is increased by 24 hours of fasting and 4 weeks of TAC (N=4–8). (F) Fasting-induced attenuation of systolic function is mediated by PPARα and Sirt1. Ejection fraction was measured in PPARα^+/− and Sirt1^+/− mice after 24 hours of fasting. (N=18–20). (G) Hearts were isolated from PPARα^+/− and Sirt1^+/− mice and subjected to the Langendorff perfused heart experiment without any nutrients in the perfusate. The duration of consecutive beating was evaluated in each heart (N=7–9). (H) PPARα and Sirt1 prolong survival of cultured cardiac myocytes under glucose-free conditions. The indicated expression or knockdown was introduced with adenovirus vectors. Cell viability was evaluated with Trypan blue dye exclusion following incubation with complete or glucose-free medium for 72 hours (N=3). Error bars represent S. E. M. (I) Schematic representations of the regulatory roles of PPARα, Sirt1 and RXR on the PPRE and the ERRE during pressure overload and fasting. PPARα shifts the energy source to fatty acids through PPRE activation, and minimizes nutritional usage through ERRE suppression, with the functional binding partners RXR and Sirt1, respectively. PPARα-mediated ERRE suppression is triggered by pressure overload, which exacerbates cardiac dysfunction due to insufficient energy production.