An NRF2/β3-Adrenoreceptor Axis Drives a Sustained Antioxidant and Metabolic Rewiring Through the Pentose-Phosphate Pathway to Alleviate Cardiac Stress

Abstract

Background: Cardiac β3-adrenergic receptors (ARs) are upregulated in diseased hearts and mediate antithetic effects to those of β1AR and β2AR. β3AR agonists were recently shown to protect against myocardial remodeling in preclinical studies and to improve systolic function in patients with severe heart failure. However, the underlying mechanisms remain elusive.

Methods: To dissect functional, transcriptional, and metabolic effects, hearts and isolated ventricular myocytes from mice harboring a moderate, cardiac-specific expression of a human ADRB3 transgene (β3AR-Tg) and subjected to transverse aortic constriction were assessed with echocardiography, RNA sequencing, positron emission tomography scan, metabolomics, and metabolic flux analysis. Subsequently, signaling and metabolic pathways were further investigated in vivo in β3AR-Tg and ex vivo in neonatal rat ventricular myocytes adenovirally infected to express β3AR and subjected to neurohormonal stress. These results were complemented with an analysis of single-nucleus RNA-sequencing data from human cardiac myocytes from patients with heart failure.

Results: Compared with wild-type littermates, β3AR-Tg mice were protected from hypertrophy after transaortic constriction, and systolic function was preserved. β3AR-expressing hearts displayed enhanced myocardial glucose uptake under stress in the absence of increased lactate levels. Instead, metabolomic and metabolic flux analyses in stressed hearts revealed an increase in intermediates of the pentose-phosphate pathway in β3AR-Tg, an alternative route of glucose utilization, paralleled with increased transcript levels of NADPH-producing and rate-limiting enzymes of the pentose-phosphate pathway, without fueling the hexosamine metabolism. The ensuing increased content of NADPH and of reduced glutathione decreased myocyte oxidant stress, whereas downstream oxidative metabolism assessed by oxygen consumption was preserved with higher glucose oxidation in β3AR-Tg mice after transaortic constriction compared with wild type, together with increased mitochondrial biogenesis. Unbiased transcriptomics and pathway analysis identified NRF2 (NFE2L2) as an upstream transcription factor that was functionally verified in vivo and in β3AR-expressing cardiac myocytes, where its translocation and nuclear activity were dependent on β3AR activation of nitric oxide synthase and nitric oxide production through S-nitrosation of the NRF2-negative regulator Keap1.

Conclusions: Moderate expression of cardiac β3AR, at levels observed in human cardiac myocardium, exerts metabolic and antioxidant effects through activation of the pentose-phosphate pathway and NRF2 pathway through S-nitrosation of Keap1, thereby preserving myocardial oxidative metabolism, function, and integrity under pathophysiological stress.

Keywords: hypertrophy; metabolism; myocytes, cardiac; oxidants; pentose phosphate pathway; receptors, adrenergic, beta-3.

Figure 1.

Cardiac expression of human β3AR attenuates myocardial hypertrophy under hemodynamic stress. A, Total heart weight normalized on tibial length (TL) and (B) total heart weight for all animals subjected to transverse aortic constriction and associated (C) transstenotic gradient severity (by echocardiography) after surgery (A through C; WT, n=35; Tg, n=40). D and E, Echocardiographic follow-up of left ventricular (LV) mass index and LV mass before and 3 and 9 weeks after TAC surgery (WT, n=14 [baseline], n=14 [TAC 9 and 3 weeks]; Tg, n=15 [baseline], n=14 [TAC 9 and 3 weeks]). F, Plasma BNP (brain natriuretic peptide) levels after TAC. G, Representative images and quantification of myocyte area measured histologically (wheat germ agglutinin staining; each dot is mean from >200 myocytes per heart) in LV transverse sections from β3-adrenergic receptor (β3AR)-Tg and WT littermates subjected or not to TAC (scale=20 µm). H, Representative images and quantification of myocyte area in neonatal rat ventricular myocytes infected with Ad-GFP-β3AR (β3AR) or Ad-GFP (GFP) treated with phenylephrine (PE 50µmol/L) or saline (Sal) (scale=20 µm). I through L, Hypertrophic transcriptional program assessed (WT, n=7 [control], n=10 [TAC]; Tg, n=7 [control], n=11 [TAC]) by (I) ANP (atrial natriuretic peptide), (J) BNP, (K) β cardiac myosin heavy chain (βMHC) transcripts abundance, and (L) associated βMHC/α cardiac myosin heavy chain (αMHC) ratio in AVMs isolated from 9-week TAC or control β3AR-Tg or WT mice. Results are expressed as mean±SEM, with *P<0.05 calculated by 2-way ANOVA and corrected for multiple comparisons with the Sidak test, except for A through C and F, which were analyzed by unpaired t test, and D and E, which were analyzed by mixed-effects models corrected with the Sidak test.

Figure 2.

Cardiac expression of human β3AR increases glucose entry and insulin sensitivity under hemodynamic stress. A and B, Intraperitoneal glucose (A, GTT; 2 g/kg) and insulin tolerance test (B, ITT; 0.5 U/kg) before and over time up to 9 weeks after aortic constriction in β3-adrenergic receptor (β3AR) transgenic (Tg) and wild-type (WT) mice (A and B, WT; n=17 [baseline], n=9 [TAC]; Tg: n=18 [baseline], n=9 [TAC]). C, Insulinemia measured in serum from β3AR-Tg and WT animals 9 weeks after TAC and measured 30 minutes after intraperitoneal glucose injection. D, Volcano plot displaying differential unbiased serum metabolomics from 9-week TAC β3AR-Tg compared with 9-week TAC WT mice analyzed by reverse-phase liquid chromatography–mass spectrometry in negative (left) and positive (right) ionization mode (1886 and 3881 detected and plotted metabolites, respectively; see Methods for details). E, Myocardial ¹⁸F-fluorodeoxyglucose (¹⁸FDG) uptake measured by positron emission tomography in vivo (expressed as standardized uptake value [SUV]; WT, n=11 [baseline], n=11 [TAC]; Tg: n=12 [baseline], n=10 [TAC]) and (F) increase in tritiated glucose uptake after insulin treatment in adult ventricular myocytes (AVMs) from β3AR-Tg or WT littermates before surgery or at different time points after TAC. G and H, Glucose transporters Glut 1 (G) and Glut 4 (H) transcripts abundance in AVMs isolated from 9-week TAC or control β3AR-Tg or WT mice (WT: n=7 [control], n=10 [TAC]; Tg: n=7 [control], n=11 [TAC]). Results are expressed as mean±SEM, with *P<0.05 calculated by 2-way ANOVA and corrected for multiple comparisons with the Sidak test, except for A and B, which were analyzed by 2-way ANOVA for repeated measures and C, which was analyzed by unpaired t test.

Figure 3.

Myocardial glucose is rerouted away from classic glycolysis by β3AR. A, Schematic representation of the glucose metabolic pathways and their interactions. B through I, Targeted metabolomics measurements of intermediates of glycolysis and accessory pathways by Liquid chromatography–mass spectrometry in cardiac extracts from β3-adrenergic receptor (β3AR) transgenic (Tg) and wild-type (WT) mice at baseline and 9 weeks after transverse aortic constriction (TAC) surgery. B, Heat map of z score of the glycolysis intermediates; (C) UDP-GlcNAc; (D) UDP-glucose; (H) ribose-5-phosphate; and (I) sedoheptulose-7-phosphate. E, Quantification of glycogen content in cardiac extracts. F and G, glycogenolysis enzymes phosphorylase kinase regulatory subunit-γ (PHKG; F) and -α (PHKA; G) transcripts abundance in adult ventricular myocytes isolated from 9-week TAC or control β3AR-Tg or WT mice (WT: n=7 [control], n=10 [TAC]; Tg: n=7 [control], n=11 [TAC]). Results are expressed as mean±SEM, with *P<0.05 calculated by 2-way ANOVA corrected for multiple comparisons with the Sidak test.

Figure 4.

Myocardial glucose is rerouted to fuel the PPP. A through E, Molar percent enrichment of [U-¹³C₆] in upper glycolysis and pentose-phosphate pathway (PPP) intermediates as a function of perfusion duration with [U-¹³C₆]-glucose in Langendorff-perfused hearts from β3-adrenergic receptor (β3AR) transgenic (Tg) and wild-type (WT) mice 9 weeks after transverse aortic constriction (TAC). A, Glucose-6-phosphate M+6; B, fructose-6-phosphate M+6; C, ribose/ribulose-5-phosphate M+5; D, sedoheptulose-7-phosphate M+7; and E, UDP-Glc-NAc M+6 (C). F through H, PPP enzymes: phosphogluconate dehydrogenase (PGD; F), glucose-6-phosphate dehydrogenase (G6PD; G), and transaldolase (Taldo; H) transcripts abundance in adult ventricular myocytes isolated from 9-week TAC β3AR-Tg or wild-type mice (WT: n=7 [control], n=10 [TAC]; Tg: n=7 [control], n=11 [TAC]). Results are expressed as mean±SEM, with *P<0.05 calculated by 2-way ANOVA corrected for multiple comparisons with the Sidak test. For A through E, at 5 minutes, n=3 to 4 per group; at 10 minutes, n=3 to 4 per group; and at 25 minutes, n=6 to 7 per group.

Figure 5.

PPP protects against oxidative stress after TAC through an improved NADPH-mediated cellular detoxification. A, Reaction steps of the pentose-phosphate pathway highlighting the β3-adrenergic receptor (β3AR)–mediated significant increases (red), and no change (gray) in metabolic intermediates. B through D, Liquid chromatography–mass spectrometry (LC/MS) determination of NADPH levels (B), NADP⁺ levels (C), and corresponding NADP⁺/NADPH ratio (D) in cardiac extracts from β3AR-Tg and wild-type (WT) mice 9 weeks after transverse aortic constriction (TAC) determined by hydrophilic interaction chromatography LC/MS. E, Quantification of reduced glutathione (GSH) and ratio of reduced over oxidized glutathione (GSH/GSSG; F) in cardiac extracts from 9-week TAC or control β3AR-Tg or WT mice. G, Glutathione peroxidase 1 (Gpx1) transcript abundance in adult ventricular myocytes (AVMs) isolated from 9-week TAC or control β3AR-Tg or WT mice (WT: n=7 [control], n=10 [TAC]; Tg: n=7 [control], n=11 [TAC]). H, LC/MS quantification of methionine sulfoxide (MetO) in cardiac extracts from 9-week TAC or control β3AR-Tg or wild-type mice. I and J, Comparative increase in intracellular reactive oxygen species (ROS) measured by H₂DCFDA fluorescence in isolated AVMs from β3AR-Tg and WT hearts 9 weeks after TAC in basal conditions (I) and during exposure to extracellular H₂O₂ (20 µmol/L; J; see Methods for details). Results are expressed as mean±SEM and analyzed by 2-way ANOVA corrected for multiple comparisons with the Sidak test (E through H) and (B through D, I, and J) analyzed by unpaired t test. *P<0.05.

Figure 6.

Cardiac β3AR promotes glucose oxidation and mitochondrial biogenesis. A and B, Liquid chromatography–mass spectrometry (MS) determination of adenine-based (A) and guanine-based (B) purines abundance in cardiac extracts from 9-week transverse aortic constriction (TAC) or control β3-adrenergic receptor (β3AR) transgenic (Tg) or wild-type mice. Gas chromatography–MS determination of cardiac tissue pyruvate M+3 perfused for 5 minutes (C) and cardiac effluents lactate M+3 in (D), and lactate efflux rate in effluent of Langendorff-perfused hearts from β3AR-Tg and wild-type (WT) mice 9 weeks after TAC perfused for 25 minutes with [U-¹³C₆]-glucose (E). F, Oxygen consumed by Langendorff-perfused hearts from 9-week TAC β3AR-Tg (n=6) and WT mice (n=6), G and H, Assessment of mitochondrial respiration of adult ventricular myocytes (AVMs) from 9-week TAC β3AR-Tg and WT mice in medium containing 11 mmol/L of glucose and 1 ng/L of insulin by Seahorse analyzer: oxygen consumption rate (OCR; G) and maximal respiration (H). I, Myocyte area enlargement in AVMs treated with phenylephrine or mitochondrial pyruvate carrier inhibitor UK-5099. J, Transcripts abundance of markers of mitochondrial biogenesis (PGC1α and TFAM) and fusion (mitofusin 1, mitofusin 2, and Opa 1) in AVMs isolated from β3AR-Tg (n=11) and WT mice (n=10) 9 weeks after TAC; normalized levels (to HPRT) are reported as relative to levels in AVMs from control WT nonoperated hearts (set as 1). Results are expressed as mean±SEM, with *P<0.05 calculated by 2-way ANOVA corrected for multiple comparisons with the Sidak test, except for C through F, which were analyzed by unpaired t test, and J, which was analyzed by multiple unpaired t test. AUC indicates area under the curve.

Figure 7.

Cardiac β3AR activates nuclear translocation of the transcription factor NRF2. A through C, RNA sequencing of adult ventricular myocytes (AVMs) isolated from β3-adrenergic receptor (β3AR) transgenic (Tg) and wild-type (WT) mice 9 weeks after transverse aortic constriction (TAC). A, Principal component (PC) analysis in which biological replicates are marked in different colors according to subgroups: β3AR-Tg, orange; and WT, black. B, Volcano plot of differentially expressed genes between β3AR-Tg and WT. C, Transcription factors enrichment ratio (β3AR-Tg vs WT) computed with an overrepresentation analysis of modulated genes by reference to the TRRUSTv2 database of transcriptional regulatory interactions. D, Metabolic and intracellular pathways regulated by the transcription factor NRF2 (NFE2L2). E, NRF2 (NFE2L2) transcripts abundance in AVMs isolated from β3AR-Tg and WT mice 9 weeks after TAC relative to levels in AVMs from control WT nonoperated hearts (set as 1; WT: n=7 [control], n=10 [TAC]; Tg: n=7 [control], n=11 [TAC]). F, Binding capacity to antioxidant response element (ARE) measured in nuclear extracts from left ventricles of β3AR-Tg and WT as a proxy of NRF2 nuclear activity. G through I, NRF2 immunostaining in neonatal rat ventricular myocytes (NRVMs) with adenoviral expression of human β3AR (Ad-β3AR) or control (Ad-GFP). G, Representative stainings of NRF2 (orange; scale=20 µm). H, Percentage of NRF2 nuclear localization. I, Percentage of NRF2 cytosolic localization. J, NRF2 (NFE2L2) transcript levels after treatment of NRVMs with siRNA targeting NRF2 or scramble siRNA. K and L, associated myocyte area on treatment with phenylephrine (PE; 50 µmol/L) or saline of neonatal myocytes infected with Ad-GFP (K) or Ad-β3AR (L). M through O, β3AR-Tg and WT mice subjected after TAC to 5-week treatment with Brusatol (2 mg/kg) injected starting 1 month after aortic constriction. M, Total heart weight normalized on tibial length (TL). N, Total heart weight and associated transstenotic gradient severity (O). Results are expressed as mean±SEM, with *P<0.05 calculated by 2-way ANOVA corrected for multiple comparisons with the Sidak test, except for M through O, which were analyzed by unpaired t test.

Figure 8.

Cardiac β3AR regulates NRF2 translocation via NO-mediated S-nitrosation of Keap1. A, DAF-FM diacetate fluorescence in β3-adrenergic receptor (β3AR) transgenic (Tg) and wild-type (WT) murine neonatal ventricular myocytes subjected to prohypertrophic treatment by isoproterenol (1 µmol/L; 4 hours). B through E, Effect of L-nitro-arginine-methyl-ester (L-NAME; nitric oxide [NO] synthase [NOS] inhibitor) administration (100 µmol/L) to neonatal rat ventricular myocytes (NRVMs) adenovirally infected with Ad-β3AR or Ad-GFP and stimulated or not with phenylephrine (PE). B, Representative staining of NRF2 (orange) and DAPI (blue) in PE-treated Ad-β3AR without (top) or with L-NAME treatment (bottom; scale=20 µm); percentage of NRF2 nuclear localization (C); associated myocyte area (D); and Nppb transcript abundance (BNP [brain natriuretic peptide] as marker of hypertrophy) (E). F and H, β3AR-Tg and WT mice subjected to transverse aortic constriction (TAC) combined with L-NAME treatment in drinking water (1 g/L) for 9 weeks. F, Total heart weight normalized on tibial length (TL). G, Total heart weight. H and I, Assessment of mitochondrial respiration in adult ventricular myocytes (AVMs) of β3AR-Tg and WT mice subjected to 9-week aortic constriction and L-NAME (β3AR-Tg, n=5; WT, n=6). Measurements are performed in medium containing 11 mmol/L of glucose, 1 ng/L of insulin, and 100 µmol/L of L-NAME by Seahorse analyzer. H, Oxygen consumption rate (OCR). I, Maximal respiration compared with AVMs without exposure to L-NAME for 9 weeks. J, Keap1 immunoblotting after resin capture of S-nitrosated proteins from NRVMs adenovirally infected with Ad-β3AR or Ad-GFP and stimulated or not with either PE or NO-donor S-nitroso-N-acetylpenicillamine (SNAP), with negative control in the absence of ascorbate (Asc; used to reduce S-nitrosated residues back into free thiols to allow resin capture (see Methods for details). K and L, S-nitrosated Keap1 levels normalized on input as measured by resin capture (K) in NRVMs extracts adenovirally infected with Ad-β3AR or Ad-GFP and stimulated or not with phenylephrine (PE) or (L) in NRVM extracts adenovirally infected with Ad-β3AR and treated with PE in the presence (or absence) of the NOS inhibitor L-NAME. M, Working model of NRF2/β3-adrenoreceptor axis driving a sustained antioxidant and metabolic rewiring through the pentose-phosphate pathway to alleviate cardiac stress. Results are expressed as mean±SEM, with *P<0.05 calculated by 2-way ANOVA corrected for multiple comparisons with the Sidak test, except for A, which was analyzed by the Mann-Whitney test and F, G, and L, which were analyzed by unpaired t test.