Svetlana Reilly1, Xing Liu2, Ricardo Carnicer2, Timothy Rajakumar2, Rana Sayeed3, George Krasopoulos3, Sander Verheule4, Tudor Fulga5, Ulrich Schotten4, Barbara Casadei2. 1. Radcliffe Department of Medicine, Division of Cardiovascular Medicine, Oxford University, Oxford, UK. Electronic address: svetlana.reilly@cardiov.ox.ac.uk. 2. Radcliffe Department of Medicine, Division of Cardiovascular Medicine, Oxford University, Oxford, UK. 3. Cardiothoracic Surgery, John Radcliffe Hospital, Oxford, UK. 4. Department of Physiology, University of Maastricht, Netherlands. 5. Weatherall Institute of Molecular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, UK.
Abstract
BACKGROUND: The management of atrial fibrillation remains a challenge. This condition remodels atrial electrical properties, which promote resistance to treatment. Although remodelling has long been a therapeutic target in atrial fibrillation, its causes remain incompletely understood. We aimed to evaluate the role of miR-31-dependent reduction in dystrophin and neuronal nitric oxide synthase (nNOS, also known as NOS1) on atrial electrical properties and atrial fibrillation inducibility. METHODS: We recruited 258 patients (209 patients in sinus rhythm and 49 with permanent atrial fibrillation) from the John Radcliffe Hospital, Oxford, UK; written informed consent was obtained from each participant. We also used a goat model of pacing-induced atrial fibrillation (24 with atrial fibrillation vs 20 controls in normal sinus rythm) and nNos-knock-out mice (n=28 compared with 27 wild-type littermates). Gene expression of miR-31, dystrophin, and nNOS was assessed by quantitative RT-PCR; protein content was measured by immunoblotting; NOS activity was evaluated with high-performance liquid chromatography; action potential duration (APD) and rate dependent adaptation were assessed by single-cell patch-clamping, and atrial fibrillation inducibility was evaluated by transoesophageal atrial burst stimulation. FINDINGS: We found that atrial-specific upregulation of miR-31 in human atrial fibrillation caused dystrophin (DYS) translational repression and accelerated mRNA degradation of nNOS leading to a profound reduction in atrial DYS and nNOS protein content and in nitric oxide availability. In human atrial myocytes obtained from patients in sinus rhythm, nNOS inhibition was sufficient to recapitulate hallmark features of remodelling induced by atrial fibrillation, such as shortening of APD and loss of APD rate-dependency, but had no effect in patients with atrial fibrillation. In mice, nNos gene deletion or inhibition shortened atrial APD and increased atrial fibrillation inducibility in vivo. Inhibition of miR-31 in human atrial fibrillation recovered DYS and nNOS, and normalised APD and APD rate-dependency. Prevention of miR-31 binding to nNOS 3'UTR recovered both nNOS protein and gene expression but had no effect on the DYS protein or mRNA level (consistent with the mRNA degradation of nNOS by miR-31). Prevention of miR-31 binding to DYS 3'UTR increased DYS protein but not mRNA is consistent with translation repression of DYS by miR-31; recovery of DYS protein increased nNOS protein but not mRNA in keeping with a stabilising effect of DYS on nNOS protein. In goats, a reduction in dystrophin and nNOS protein content was associated with upregulation of miR-31 in the atria but not in the ventricles. INTERPRETATION: The findings suggest that atrial-specific upregulation of miR-31 in human atrial fibrillation is a key mechanism causing atrial loss of dystrophin and nNOS; this loss leads to the electrical phenotype induced by atrial fibrillation. FUNDING: British Heart Foundation (BHF) Programme grant (for BC and XL), BHF Centre of Excellence in Oxford (SR), Leducq Foundation (in part for BC and SR), the European Union's seventh Framework Programme Grant Agree.
BACKGROUND: The management of atrial fibrillation remains a challenge. This condition remodels atrial electrical properties, which promote resistance to treatment. Although remodelling has long been a therapeutic target in atrial fibrillation, its causes remain incompletely understood. We aimed to evaluate the role of miR-31-dependent reduction in dystrophin and neuronal nitric oxide synthase (nNOS, also known as NOS1) on atrial electrical properties and atrial fibrillation inducibility. METHODS: We recruited 258 patients (209 patients in sinus rhythm and 49 with permanent atrial fibrillation) from the John Radcliffe Hospital, Oxford, UK; written informed consent was obtained from each participant. We also used a goat model of pacing-induced atrial fibrillation (24 with atrial fibrillation vs 20 controls in normal sinus rythm) and nNos-knock-out mice (n=28 compared with 27 wild-type littermates). Gene expression of miR-31, dystrophin, and nNOS was assessed by quantitative RT-PCR; protein content was measured by immunoblotting; NOS activity was evaluated with high-performance liquid chromatography; action potential duration (APD) and rate dependent adaptation were assessed by single-cell patch-clamping, and atrial fibrillation inducibility was evaluated by transoesophageal atrial burst stimulation. FINDINGS: We found that atrial-specific upregulation of miR-31 in human atrial fibrillation caused dystrophin (DYS) translational repression and accelerated mRNA degradation of nNOS leading to a profound reduction in atrial DYS and nNOS protein content and in nitric oxide availability. In human atrial myocytes obtained from patients in sinus rhythm, nNOS inhibition was sufficient to recapitulate hallmark features of remodelling induced by atrial fibrillation, such as shortening of APD and loss of APD rate-dependency, but had no effect in patients with atrial fibrillation. In mice, nNos gene deletion or inhibition shortened atrial APD and increased atrial fibrillation inducibility in vivo. Inhibition of miR-31 in human atrial fibrillation recovered DYS and nNOS, and normalised APD and APD rate-dependency. Prevention of miR-31 binding to nNOS 3'UTR recovered both nNOS protein and gene expression but had no effect on the DYS protein or mRNA level (consistent with the mRNA degradation of nNOS by miR-31). Prevention of miR-31 binding to DYS 3'UTR increased DYS protein but not mRNA is consistent with translation repression of DYS by miR-31; recovery of DYS protein increased nNOS protein but not mRNA in keeping with a stabilising effect of DYS on nNOS protein. In goats, a reduction in dystrophin and nNOS protein content was associated with upregulation of miR-31 in the atria but not in the ventricles. INTERPRETATION: The findings suggest that atrial-specific upregulation of miR-31 in human atrial fibrillation is a key mechanism causing atrial loss of dystrophin and nNOS; this loss leads to the electrical phenotype induced by atrial fibrillation. FUNDING: British Heart Foundation (BHF) Programme grant (for BC and XL), BHF Centre of Excellence in Oxford (SR), Leducq Foundation (in part for BC and SR), the European Union's seventh Framework Programme Grant Agree.