Research ArticleHeart Failure

Ablation of the stress protease OMA1 protects against heart failure in mice

See allHide authors and affiliations

Science Translational Medicine  28 Mar 2018:
Vol. 10, Issue 434, eaan4935
DOI: 10.1126/scitranslmed.aan4935

A mitochondrial mark for heart failure

Current therapies for heart failure vary depending on the root cause and stage of the disease. Acin-Perez et al. studied the molecular link between mitochondrial dysfunction and heart failure, focusing on OMA1, a protease involved in mitochondrial inner membrane remodeling and cytochrome c release. Using three mouse models of heart failure, they found that cardiomyocytes were protected from mitochondrial reactive oxygen species–induced cell death when OMA1 was ablated. OMA1 could be a therapeutic target for heart failure.


Heart failure (HF) is a major health and economic burden in developed countries. It has been proposed that the pathogenesis of HF may involve the action of mitochondria. We evaluate three different mouse models of HF: tachycardiomyopathy, HF with preserved left ventricular (LV) ejection fraction (LVEF), and LV myocardial ischemia and hypertrophy. Regardless of whether LVEF is preserved, our results indicate that the three models share common features: an increase in mitochondrial reactive oxygen species followed by ultrastructural alterations in the mitochondrial cristae and loss of mitochondrial integrity that lead to cardiomyocyte death. We show that the ablation of the mitochondrial protease OMA1 averts cardiomyocyte death in all three murine HF models, and thus loss of OMA1 plays a direct role in cardiomyocyte protection. This finding identifies OMA1 as a potential target for preventing the progression of myocardial damage in HF associated with a variety of etiologies.


Heart failure (HF) is a clinical syndrome defined pathophysiologically as the inability of the heart to adequately meet the body’s energy requirements and is the end result of a myriad of cardiac insults. To develop more efficacious treatments to preserve myocardial function, there is a need to define the molecular commonalities and differences between the diverse causes of HF. Current clinical guidelines recommend compliance with lifestyle changes, dietary restrictions, and the use of several therapeutic agents, including β blockers, angiotensin-converting enzyme inhibitors, diuretics, aldosterone receptor antagonists, angiotensin receptor blockers, ivabradine, and the novel introduction of the angiotensin receptor–neprilysin inhibitor (1, 2). Unfortunately, the efficiency of these treatments is limited and varies from patient to patient.

Recent advances have identified that several processes are altered in HF, including cardiomyocyte apoptosis, autophagy, inflammatory responses, mitochondrial remodeling, gene transcription regulation, and energy metabolism. The maintenance of the heart’s pumping action requires functional and morphological integrity of mitochondria to ensure an uninterrupted energy supply. Mitochondria are subcellular energy providers and metabolic integrator organelles whose energy production can generate reactive oxygen species (ROS) as a by-product. ROS act as signaling molecules regulating cell adaptation and, in excess, cause cell damage and, ultimately, cell death. Alteration in mitochondrial physiology and biogenesis plays a central role in the initiation and progression of cardiovascular disease (CVD). In this regard, increased ROS production, energy deficiency, and reduced mitochondrial-dependent metabolism underlie functional changes of HF.

Cardiac muscle contraction strongly depends on calcium homeostasis and calcium redistribution in the cardiomyocyte. Studies on the role of calcium fluxes during contractility induced by acute (3, 4) or chronic stress (5, 6) are consistent with a major role of altered intracellular calcium dynamics in HF (7, 8). Proposals linking HF to oxidative stress and mitochondrial dysfunction (9, 10) are supported by the findings that adrenergic stimulation leads to oxidative stress (11, 12). Several studies have reported increased ROS production in the hearts of patients with congestive HF, and ROS damage biomarkers have been detected in the pericardial fluid (13) and in peripheral blood (14). However, the molecular connection between ROS production and calcium homeostasis in HF remains unclear. Mitochondria are potentially important sources of ROS and also contribute to calcium homeostasis and contractility through the action of the mitochondrial calcium uniporter (MCU) (3, 4, 15, 16).

Mitochondrial inner membrane forms highly organized invaginations known as cristae that enclose the components of the oxidative phosphorylation (OXPHOS) system. Crista structure determines the organization and function of the respiratory chain and its metabolic status, and it is regulated by the action of the inner membrane proteases OMA1 and Yme1l. Silencing OMA1 preserves the cristae, prevents mitochondrial dysfunction, and protects cardiomyocytes against apoptosis (17). OMA1 is activated in response to stress and mediates the proteolytic processing of OPA1, a protein that regulates crista formation and maintenance, mitochondrial fusion, and adenosine 5′-triphosphate (ATP) production (18, 19). The absence of OMA1 rescues the cardiac damage induced by the lack of Yme1l (20) and the neurodegeneration promoted by the lack of prohibitin, a mitochondrial inner membrane protein proposed to have a role as chaperone for respiration chain proteins or as a general structuring scaffold required for optimal mitochondrial morphology and function (21). Lack of OMA1 also protects the kidney against ischemia/reperfusion injury (22). Consistent with these findings, mild overexpression of OPA1, the main OMA1 target identified so far, protects against heart and brain ischemia (23) and ameliorates mitochondrial disease symptoms caused by defects in electron transport chain complexes I and IV (24). Unfortunately, strong OPA1 overexpression causes developmental arrest (23), compromising the use of OPA1 as a therapeutic target. We therefore investigated whether the lack of OMA1 influences the progression of HF induced by a variety of insults.

A more profound understanding of the molecular mechanisms underlying HF would potentially identify novel cellular processes not targeted by current therapies. Here, we induced HF in mice through three experimental approaches: (i) tachycardiomyopathy (25) caused by isoproterenol (ISO)–induced work overload (chronic administration of ISO); (ii) HF with preserved left ventricular (LV) ejection fraction (HFpEF) caused by angiotensin II (AngII)–induced hypertension; and (iii) LV myocardial ischemia and hypertrophy due to pressure overload induced by transaortic constriction (TAC), a pathophysiological model of HFpEF that develops into HFrEF (heart failure with reduced ejection fraction). We evaluated the consequences of modulating mitochondrial ROS (mtROS) and mitochondrial calcium and genetic ablation of OMA1 (OMA1KO) in these models. Our results identify OMA1 as a central player in mediating HF irrespective of the etiology and a potential therapeutic target to protect myocardial integrity.


ISO induces HF

Chronic tachycardia and cardiac work overload were induced in wild-type control C57BL/6JOlaHsd mice by minipump administration of the β-adrenergic agonist ISO over 28 days. Echocardiographic monitoring of LV systolic and diastolic function revealed that ISO increased contractility (% EF) and heart rate (HR), leading to cardiac dysfunction characterized by altered cardiac output, diastolic dysfunction [increased early (E) to late (A) ventricular filling velocity ratio: E/A], and QRS duration without LV hypertrophy. ISO also increased EF and HR in mice lacking OMA1 (OMA1KO); however, ISO did not cause cardiac dysfunction in these animals (Fig. 1A). Moreover, in control mice, chronic ISO therapy induced increased circulating markers of heart damage, including serum creatine kinase (CK) and creatine kinase MB (CK-MB), whereas these parameters remained within normal values in OMA1KO animals (Fig. 1B). In agreement with these findings, fibrotic lesions indicative of cardiomyocyte death were found in the hearts of wild-type mice, but not of OMA1KO mice (Fig. 1, C and D). No alterations were observed in brain, liver, or kidney of ISO-treated wild-type mice (fig. S1).

Fig. 1 Lack of OMA1 prevents tissue damage induced by chronic cardiac work overload in mice.

(A) Echocardiography analysis of ejection fraction (% EF), heart rate (HR), cardiac output (CO), E/A ratio, QRS amplitude, and left ventricular (LV) mass in control mice and isoproterenol (ISO)–treated mice. WT, wild-type. (B) Blood analysis of cardiac damage indicated by increased serum CK (creatine kinase, upper panel) and CK-MB (creatine kinase MB isoform, lower panel); one-way analysis of variance (ANOVA) and Tukey’s multiple comparison test. (C and D) Masson’s trichrome staining and quantification of transverse heart sections. ns, not significant. (E) Representative heart transmission electron microscopy (TEM) images from WT and OMA1KO mice, with and without ISO treatment. Scale bars, 1 μM. (F) Quantification of mitochondrial area, shape (aspect ratio), organization (number of damaged mitochondria), and lipid droplet content between conditions. Data are given as scatter dot plots, and lines are means ± SD. Differences assessed by two-way ANOVA and Sidak’s multiple comparison test (A) or one-way ANOVA and Tukey’s multiple comparison test (B to F). N as described in table S1; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. ISO treatment during 28 days.

Transmission electron microscopy (TEM) revealed that ISO increased the number of damaged mitochondria in the hearts of wild-type mice, but not in OMA1KO mice (Fig. 1, E and F), suggesting that maintenance of the structural integrity of the mitochondrial crista protects OMA1KO mice against work overload–induced heart damage. In addition, we observed a significant decrease in lipid droplets within mitochondria upon ISO treatment in the OMA1KO animals (P < 0.0001; Fig. 1F), suggesting more efficient mobilization of lipids.

To further characterize the molecular mechanisms involved in ISO-induced heart damage and cardiomyocyte cell death, we analyzed OXPHOS performance in heart mitochondria in the early stages (7 days) of ISO administration. ATP synthesis driven by glutamate plus malate or succinate was significantly decreased in ISO-treated heart mitochondria from wild-type mice (P < 0.05, P < 0.0001; fig. S2A) but was unaffected in mitochondria from OMA1KO animals. In control mice, ISO promoted OMA1 activation and concomitant OPA1 processing (fig. S2, B and C), whereas no changes were observed in OMA1KO heart mitochondria.

We next addressed the impact of ISO in mtROS in the heart. First, we found increased Amplex Red signal in isolated heart mitochondria from both genotypes when animals were previously treated with ISO (fig. S2D). Only wild-type ISO-treated mice showed an associated increase in dihydroethidium (DHE) signal in heart slices (fig. S2, E and F). Amplex Red is regularly used as an indicator of H2O2 production (26), whereas DHE is used as an indicator of superoxide production (27). This suggests that ISO induced the production of superoxide and subsequent H2O2 in wild-type hearts, but only H2O2 in OMA1KO cells. However, the major source of H2O2 in mitochondria is superoxide dismutase 2 (SOD2) and this is generated from superoxide. The activity of SOD2 in tissues is very high, and superoxide is quickly dismutated to H2O2, making the detection of hydrogen peroxide easier than superoxide. We found a reduction in the mitochondrial SOD2 and the ROS-sensitive aconitase activities in wild-type ISO-treated heart mitochondria but not OMA1KO (fig. S2, G and H). This is in agreement with the idea that more superoxide is present in the ISO-treated samples. Overall, these results suggest that the maintenance of crista integrity in OMA1KO mice prevents oxidative damage and energetic collapse associated with ISO-mediated ROS increase.

Increased mtROS is required for ISO-induced damage

To determine the role of mtROS in ISO-induced cardiac injury, we administered ISO together with the specific mtROS scavenger MitoQ (MQ). In both genotypes, MQ prevented the ISO-induced increase in heart contractility but not increased HR (Fig. 2A). By doing so, MQ effectively reverted the marker of LV diastolic dysfunction (E/A) to normal values and prevented the increase in cardiac output in wild-type animals (Fig. 2A). At the doses used, ISO + MQ or MQ alone had no deleterious effects on heart structure (fig. S3A). Moreover, MQ prevented ISO-induced increase in mitochondrial H2O2 (Fig. 2B). Blocking ISO-induced mitochondrial H2O2 increase with MQ prevented cardiomyocyte death in wild-type hearts (Fig. 2C) by preserving mitochondrial ATP production (Fig. 2D). As expected, the ISO-induced decrease in OXPHOS in wild-type animals resulted in increased reduced form of nicotinamide adenine dinucleotide (NAD+) (NADH) and NADH/ATP ratio in heart homogenates, which was blocked by MQ; no effects were observed in OMA1KO animals (Fig. 2E). In wild-type mice, MQ also blocked the ISO-induced decrease in aconitase activity (Fig. 2F) and OPA processing (Fig. 2G). TEM revealed that MQ also rescued crista integrity in ISO-treated wild-type animals in both subsarcolemmal and intermyofibrillar mitochondria (Fig. 2, H and I). The number of intermitochondrial junctions (IMJs), an index of highly active mitochondria (28), was decreased in ISO-treated wild-type mitochondria and restored by MQ (Fig. 2, H and I). Notably, integrity of mitochondria in OMA1KO mice was preserved under all conditions.

Fig. 2 The mitochondrial ROS scavenger MitoQ prevents the pathogenic phenotype induced by ISO.

(A) Echocardiography analysis of % EF, HR, CO, and E/A ratio in control and treated mice. MQ, MitoQ. (B) Reactive oxygen species (ROS) production measured as mitochondrial H2O2 release using Amplex Red. AU, arbitrary unit. (C) Blood analysis of cardiac damage indicated by increased serum CK-MB. (D) Adenosine 5′-triphosphate (ATP) synthesis rate in heart mitochondria driven by glutamate + malate (Glu + Mal, left) or succinate (Succ, right) in WT and OMA1KO mice. (E) Mitochondrial NADH (reduced form of nicotinamide adenine dinucleotide)/ATP ratio. (F) Spectrophotometrically measured ROS-sensitive aconitase activity in heart mitochondria. (G) Western blot analysis and quantification of OPA1 processing in mitochondrial lysate. (H) Representative TEM images of intermyofibrillar (IM) and subsarcolemmal (SC) mitochondria in cardiac fibers from WT and OMA1KO mice. Scale bars, 1 μM. (I) Quantification of intermitochondrial junctions (IMJs). Data are given as scatter dot plots, and lines are means ± SD. Differences assessed by two-way ANOVA and Sidak’s multiple comparison test (A) or one-way ANOVA and Tukey’s multiple comparison test (B to H). N as described in table S1; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

ISO induces abnormal mitochondrial calcium transport in heart

RNA sequencing (RNA-seq) analysis was performed in heart tissue from wild-type and OMA1KO mice (untreated and treated with ISO). Principal component analysis of the overall heart RNA profile revealed differences between wild-type and OMA1KO transcriptomes (Fig. 3A). After 7 days of ISO administration, the heart transcriptomes in both genotypes underwent modifications, and different clusters were identified in each condition (Fig. 3A). Pathway analysis identified genes involved in calcium and Gq signaling among those increased in ISO-stimulated wild-type mice, whereas genes increased in ISO-stimulated OMA1KO mice included those involved in vasodilatation, endothelial nitric oxide synthase (eNOS), and Gs signaling (Fig. 3, B and C).

Fig. 3 Chronic work overload leads to mitochondrial calcium overload.

(A) Principal component (PC) analysis of RNA sequencing data from the hearts of control and ISO-treated mice (7 days). CPM, continuous phase modulation. (B) Most relevant Ingenuity Canonical Pathways (ICPs) enriched using the differentially expressed genes in OMA1KO ISO versus control. The inner bar circle represents the fold enrichment and the overall changes as a z score. The outer circle displays the logFC of the differentially expressed genes belonging to each category. eNOS, endothelial nitric oxide synthase; mTOR, mammalian target of rapamycin; NO, nitric oxide; iNOS, inducible nitric oxide synthase; ATM, ataxia-telangiectasia mutated. (C) Detailed relationship between changed genes [abs(logFC) ≥ 1 in any contrast] and the pathways selected by Chord Plots [left: logFC of genes in the contrasts ISO versus control of WT (inner) and OMA1KO (outer) mice; right: ICPs connected to the genes]. For (B) and (C), red represents up-regulated genes, and blue represents down-regulated genes. (D) Western blot analysis of mitochondrial lysated and quantification of mitochondrial calcium uniporter (MCU) components (MCU and M1CU1) after short chronic (7 days) ISO administration. (E) Western blot analysis and quantification of MCU expression modulation by mitochondrial ROS (mtROS) after chronic (28 days) ISO administration. (F to H) Effects of mitochondrial fission blockade with Mdivi1. (I to K) Effects of ryanodine receptor 2 (RyR2) activation with caffeine. (L to N) Effects of RyR2 blockade with dantrolene. (F, I, and L) Cardiac contractility (% EF and HR); (G, J, and M) Mitochondrial ATP synthesis; (H, K, and N) OPA1 processing and MCU expression. Loading controls were SDHA, core1, core2, MnSOD, and Tom20. Data are given as scatter dot plots, and lines are means ± SD. Differences assessed by one-way ANOVA and Tukey’s multiple comparison test. N as described in table S1; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

ISO infusion induces an acute increase in HR and mitochondrial calcium mobilization through the MCU to rapidly increase ATP production (3, 4, 29). RNA-seq analysis suggested that calcium mobilization was involved in the induction of HF. To better characterize this finding, we analyzed the expression of two MCU proteins, MCU and M1CU1, in mitochondria from hearts of control and ISO-treated mice. ISO significantly increased MCU (P < 0.0001) and M1CU1 in mitochondria from wild-type but not OMA1KO mice (Fig. 3D), consistent with a detrimental effect of excessive and continuated mobilization of calcium to mitochondria (3). To confirm that mitochondrial calcium overloading was involved in the phenotype of ISO-stimulated damage observed in wild-type mice, we analyzed MCU protein after ISO infusion for 28 days in the presence or absence of MQ. MQ reverted the ISO-induced increase in mitochondrial MCU (Fig. 3E).

To evaluate the causal role of mitochondrial crista remodeling in a work overload model of HF, we pharmacologically blocked mitochondrial fission using the Drp1 inhibitor Mdivi1 (30). Mdivi1 acts on the outer mitochondrial membrane, preventing mitochondrial fission and, indirectly, diminishing crista remodeling. Mdivi1 did not block the ISO-induced increases in EF or HR in wild-type mice (Fig. 3F) but did prevent the ISO-induced drop in ATP synthesis (Fig. 3G), the processing of OPA1, and the increase in MCU (Fig. 3H and fig. S3, B and C). These results confirm that prevention of crista remodeling is necessary and sufficient to preserve cardiac function upon stress, acting downstream of ISO-induced hypercontractility. However, although Mdivi1 prevented ISO-induced energetic decrease and crista remodeling, TEM (fig. S3D) still revealed cardiac fiber damage [as illustrated by the reduction of IMJs per organelle between control (1.5 ± 0.168) and ISO + Mdivi1 (0.651 ± 0.063); P < 0.001], suggesting that full protection may require direct action on the mitochondrial inner membrane.

The results presented so far suggest that calcium mobilization and crista remodeling underlie cardiac dysfunction in an animal model of HF, but the connection between them remains to be determined. To assess the role of calcium in vivo, we modulated the cardiac ryanodine receptor 2 (RyR2). RyR2 was activated with caffeine (xanthine activator) or inhibited with the clinically used antagonist dantrolene. In the hearts of wild-type animals, caffeine mimicked the effect of ISO, increasing the EF and HR while elevating MCU protein expression and decreasing mitochondrial ATP synthesis (Fig. 3, I to K, and fig. S3E). In contrast, dantrolene prevented the ISO-induced changes ATP synthesis (Fig. 3M), crista remodeling, and MCU expression (Fig. 3N and fig. S3F). Dantrolene did not, however, prevent the ISO-induced increase in EF nor HR (Fig. 3L). OMA1KO mice were insensitive to RyR2 activation or inhibition (Fig. 3, I to N, and fig. S3, E and F).

To further investigate the role of MCU in cardiac damage, we treated isolated neonatal cardiomyocytes with the MCU inhibitor Ru360 (because it cannot be used in animals due to toxicity). Cardiomyocytes were obtained from wild-type and OMA1KO mice neonates from day 0 to day 3 after birth and exposed to different stimuli in culture. Mitochondrial performance analysis revealed that ISO increased the maximum respiration rate in wild-type cardiomyocytes, leading to increased glucose utilization; this response was similar to that observed when cardiomyocytes were grown in the presence of serum to mimic hypertrophic conditions (Fig. 4A). ISO did not induce respiration changes in OMA1KO-derived cardiomyocytes. In wild-type–derived cardiomyocytes, the ISO-induced increase in ROS was prevented by supplementation with N-acetyl-l-cysteine (NAC) (Fig. 4B), and mitochondrial calcium overload was prevented by treatment with Ru360 (Fig. 4C). Both drugs blocked the ISO-induced increase in respiration (Fig. 4A); however, the ISO + NAC profile cannot maintain sustained maximum respiration likely because of a nutrient-limited condition. ISO caused a drop in membrane potential (Fig. 4D), resulting in OPA1 processing by OMA1 (Fig. 4E) that preceded MCU-mediated calcium overload (Fig. 4F). Blocking the increase in mtROS preserved membrane potential, and this prevented crista remodeling and calcium overload (Fig. 4, B to F). In contrast, MCU inhibition blocked only calcium overload (Fig. 4, B to F). These results suggest that mitochondrial calcium overload is a downstream element of the pathological response. Analysis of calcium mobilization and mitochondrial distribution by live microscopy imaging showed that the calcium redistribution, indicative of mitochondrial calcium overload, occurs only in ISO-stimulated wild-type cardiomyocytes (Fig. 4, G and H). Both MCU and RyR2 inhibitors abolished mitochondrial calcium redistribution, but only RyR2 inhibition prevented mitochondrial fragmentation (Fig. 4H). In summary, in a model of cardiac work overload, mtROS-induced mitochondrial crista disruption triggers MCU increase, mitochondrial calcium overload, loss of mitochondrial function, and cardiomyocyte death.

Fig. 4 Analysis of ISO response in vitro in neonatal cardiomyocytes.

(A) Mitochondrial respirometry analysis of neonatal cardiomyocytes upon different treatments measured by Seahorse and maximum respiration rate (MRR) quantification. FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; FBS, fetal bovine serum; NAC, N-acetyl-l-cysteine; OCR, oxygen consumption rate. (B) 2′,7′-Dichlorofluorescein diacetate (H2-DCFDA)–mediated ROS quantification in neonatal cardiomyocytes upon different treatments. (C) Fura-2-acetoxymethyl ester (Fura-2 AM)–mediated intramitochondrial calcium quantification in neonatal cardiomyocytes upon different treatments. (D) Tetramethylrhodamine methyl ester perchlorate (TMRM)–mediated mitochondrial membrane potential determination in neonatal cardiomyocytes upon different treatments. (E) Analysis of OPA1 processing by Western blot. (F) Analysis of MCU expression by Western blot of mitochondrial lysate. SDHA and core2 are used as loading controls. (G) Kurtosis analysis as indicator of calcium distribution in the cardiomyocytes at the indicated treatments. Basal and final treatments correspond to time 0 and time 9 min, respectively. (H) Time-lapse images (40×) at the indicated time showing calcium (Fura-2 AM) and mitochondrial (MitoTracker Deep Red) distribution under different treatments. Scale bars, 20 μM. Data in upper and middle panels of (A) are given as means ± SEM and in the rest of graphics as scatter dot plots, and lines are means ± SD. Differences assessed by two-way ANOVA and Sidak’s multiple comparison test (A) or one-way ANOVA and Tukey’s multiple comparison test (B to F). N as described in table S1; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

AngII induces HF

To investigate the role of mitochondrial integrity in other models of HF, we evaluated a model of HFpEF mediated by AngII-induced vasoconstriction and arterial blood pressure increase. AngII acts through the angiotensin receptor (AT1-2). The existence of an organelle-autonomous AngII signaling pathway in mitochondria has been proposed (31). Chronic AngII infusion did not alter EF or HR in either genotype (fig. S4A). However, wild-type mice developed cardiac hypertrophy, systolic arterial hypertension, and alterations in heart conductance, which were not observed in OMA1KO hearts (Fig. 5A and fig. S4B). Consequently, only wild-type hearts developed fibrosis (Fig. 5B). Despite the increase in ROS in both genotypes, only wild-type hearts showed decreased mitochondrial ATP synthesis and aconitase activity and increased NADH/ATP ratio (Fig. 5, C to F). Similarly, wild-type, but not OMA1KO, hearts showed disrupted mitochondrial ultrastructure (Fig. 5, G and H) and elevated MCU content (Fig. 5I). Although ISO and AngII each have specific and distinct receptors that activate different signaling pathways, they both converged in the promotion of downstream mtROS (Figs. 2B and 5C).

Fig. 5 Lack of OMA1 protects against AngII-induced hypertension.

(A) Effect of chronic (28 days) angiotensin II (AngII) or AngII + MQ administration on hypertrophy indicators [intraventricular septum (IVS) and left ventricular posterior wall (LVPW) thickness], systolic blood pressure (SBP), and QRS amplitude in WT and OMA1KO mice. d, diameter. (B) Masson’s trichrome staining and quantification on transverse heart sections. The quantification of the relative collagen observed with respect to AngII untreated animals in the indicated mouse strains is shown in chart; **P < 0.01. (C) mtROS production measured as Amplex Red–detected H2O2 release. (D) Spectrophotometric determination of ROS-sensitive aconitase activity in heart. (E) Cardiac mitochondrial ATP synthesis rate driven by malate Glu + Mal (left) or Succ (right). (F) Mitochondrial NADH/ATP ratio. (G) Representative TEM images of intermyofibrillar (IM) and subsarcolemmal (SC) mitochondria from WT and OMA1KO mouse hearts. Scale bars, 0.5 μM. The chart shows quantification of IMJs; control values are those used in Fig. 3H. (H) Western blot and quantification of OPA1 processing. (I) Western blot and quantification of MCU expression. SDHA, core1, and Tom20 were used as loading controls. Data are given as scatter dot plots, and lines are means ± SD. Differences assessed by one-way ANOVA and Tukey’s multiple comparison test. N as described in table S1; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Similarly to ISO-induced HF, liver and brain showed no evidence of AngII-induced damage (fig. S4C). However, kidney function was affected in AngII-treated wild-type mice but was preserved in OMA1KO mice (fig. S4, D and E), confirming that the ablation of OMA1 protects kidney function. In addition, analysis of aorta, assessed by estimation of media-to-lumen ratio and collagen area, showed that OMA1 ablation also protected the vasculature against the AngII challenge (fig. S4, F to H). MQ treatment also protected against AngII-induced cardiac damage (Fig. 5, A to I, and fig. S4B). Although a role for mtROS in AngII-mediated heart damage has been suggested previously (3234), our results further confirm this relationship and demonstrate that OMA1 activation and crista remodeling are downstream processes.

TAC induces HF

The protective effect of OMA1 deletion from cardiac work overload and systolic arterial hypertension-induced HF prompted us to evaluate its potential protective role in myocardial ischemia due to pressure overload induced by TAC. The efficacy of the TAC was monitored by echocardiographic determination of the descending aortic velocity (Fig. 6A). Wild-type animals subjected to TAC had elevated systolic blood pressure (Fig. 6B) and QRS (Fig. 6C) and developed hypertrophy with cardiac fibrosis (Fig. 6, D and E). OMA1KO hearts were protected against TAC-induced injury, showing signs of mild hypertrophy but no tissue damage (Fig. 6, D and E). TAC rendered wild-type cardiomyocyte mitochondria dysfunctional (Fig. 6F) and induced mitochondrial crista remodeling (Fig. 6G), whereas mitochondria from OMA1KO hearts were insensitive to TAC. TAC triggered a decrease in EF, contrasting with the preserved EF with AngII and the increase in response to ISO, and we did not observe major alterations in factors related to calcium homeostasis (Fig. 6G). Minipump administration of MQ to wild-type mice before TAC was cardioprotective (Fig. 6, A to E) and preserved cardiac mitochondrial performance (Fig. 6, F and G).

Fig. 6 Lack of OMA1 protects from hypertrophy induced by TAC.

(A) Descending aortic velocity (Des Ao Vel) in control and MQ-treated WT and OMA1KO mice 3 weeks after transaortic constriction (TAC) surgery [3 weeks post surgery (ps)]. (B) SBP at 3 weeks ps. (C) QRS analyzed by echocardiogram at 3 weeks ps. (D) Hypertrophy parameters at 3 weeks ps (left, IVS; center, LVPW; and right, LV mass). (E) Masson’s trichrome staining and quantification in transverse heart sections at 3 weeks ps. (F) Cardiac mitochondrial ATP synthesis rate driven by glutamate plus malate (Glu + Mal, left) or succinate (Succ, right) at 3 weeks ps. (G) Western blot analysis of crista remodeling (OPA1 processing) and mitochondrial calcium overload (MCU and M1CU1) at 3 weeks ps. (H) Descending aortic velocity (left), LV mass (middle), and ejection fraction (right) in WT and OMA1KO mice at the indicated times after TAC surgery. (I) Treadmill endurance performance of mice at 12 weeks ps. (J) Cardiac mitochondrial ATP synthesis rate driven by glutamate plus malate (Glu + Mal, left) or succinate (Succ, right) at 12 weeks ps. (K) Western blot analysis of crista remodeling (OPA1 processing) in mitochondrial lysate at 12 weeks ps and quantification. SDHA and core2 were used as loading controls. Data, except in (B), are given as scatter dot plots, and lines are means ± SD. Data in (B) are given as a box plot, whiskers being minimum to maximum. Differences assessed by one-way ANOVA and Tukey’s multiple comparison test. N as described in table S1; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

A cohort of wild-type and OMA1KO mice were maintained for up to 3 months after TAC surgery to evaluate the long-term consequences of the cardiac injury and duration of protection. In wild-type mice, TAC induced a persistent increase in descending aortic velocity (Fig. 6H), which led to cardiac hypertrophy and LV systolic dysfunction (EF decrease) (Fig. 6H). TAC also compromised endurance exercise performance at 12 weeks after surgery in wild-type animals (Fig. 6I) due to decreased ATP synthesis (Fig. 6J) and OPA1 processing (Fig. 6K). These negative effects were not observed in OMA1KO mice (Fig. 6, H to K).

OMA1 plays a pathophysiological role in HF

We have previously reported that deletion of OMA1 promotes the development of obesity and insulin resistance in mice placed on a high-fat diet (HFD) (19). Because these conditions are associated with cardiovascular impairment, the protective effect of OMA1 deletion reported here could seem counterintuitive. To evaluate the effect of HFD on cardiac function, we fed wild-type and OMA1KO mice an HFD for 8 weeks. Contrary to our previous findings with mice on a mixed 129/C57BL/6 background, purebred C57BL/6 OMA1KO animals fed with the HFD did not gain more weight than controls (fig. S5A). A glucose tolerance test (GTT) showed that purebred OMA1KO mice still developed insulin resistance (fig. S5, B and C). However, the HFD-fed OMA1KO animals had a normal cardiac phenotype (fig. S5D) and normal mitochondrial function (fig. S5E). These results resemble the paradoxical phenotype observed in metabolically healthy but obese people (MHO phenotype) of relevant clinical interest (35). Together, our results suggest that inhibition of OMA1 protease could be used as a possible therapeutic target to prevent HF, because its ablation has minor physiological alterations under normal conditions.

To consider the translational potential of the inhibition of OMA1 function in HF, we used an adenovirus with heart tropism expressing short hairpin RNA (shRNA) targeting OMA1 under the cardiac troponin and miR30 promoter. This assay also allowed to determine whether cardiac protection required the loss of OMA1 function in organs besides the heart. Adenovirus treatment prevented OPA1 processing by OMA1 in the heart without affecting liver processing (Fig. 7, A and B). Next, we assessed the impact of viral particle injection before TAC in wild-type C57BL/6 mice. Cardiac-specific silencing of OMA1 was sufficient to prevent cardiac and mitochondrial dysfunction induced by subsequent TAC (Fig. 7, C and D). Moreover, the viral treatment also induced protection against TAC-induced HF in an outbred mouse strain (CD1; Fig. 7, Cand D).

Fig. 7 Heart-specific OMA1 down-regulation protects from hypertrophy induced by TAC.

OPA1 processing analysis by Western blot (A) and quantification (B) in isolated mitochondria from heart and liver of mice injected with sh scrambled (shScr) or shOMA1 viral particles. Mitochondria were incubated in the presence of carbonyl cyanide m-chlorophenyl hydrazone for the indicated time points. MW, molecular weight; nt, not treated. (C) Des Ao Vel, heart weight/body weight (HW/BW), % EF, and rate of ATP synthesis in WT inbred (C57BL/6JOlaHsd) and outbred (CD1) mice injected with sh shScr or shOMA1 virus and then subjected to TAC. Analyses were performed 3 weeks after TAC surgery. (D) OPA1 processing analysis by Western blot (left) and quantification OPA1 (right) in mitochondrial heart fractions form WT inbred (C57BL/6JOlaHsd) and outbred (CD1) mice injected with sh scr or sh OMA1 for the indicated conditions. Data are given as scatter dot plots, and lines are means ± SD. Differences assessed by one-way ANOVA and Tukey’s multiple comparison test. N as described in table S1; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.


Our results suggest that the development of HF requires an increase in mtROS that activates OMA1, resulting in mitochondrial crista disruption due to enhanced OPA1 processing. In the case of HFpEF, crista remodeling is followed by MCU accumulation and excessive mitocondrial calcium influx. In all three HF models used in this study, disruption of mitochondrial ultrastructure triggered cell death, as revealed by the significant increase in fibrosis. This cascade of events can be arrested at several stages: MQ-induced reduction in mtROS, OMA1 deficiency–mediated prevention of crista remodeling, and Ru360 inhibition of mitochondrial calcium influx. All of these measures result in the preservation of cardiac structure and function and prevent the onset of HF.

ROS generation is a double-edged sword that can have regulatory or pathological consequences. Excessive mtROS production is known to have detrimental effects on cell survival, whereas ROS is required in physiologically relevant amounts for cell signaling purposes (36). A number of stress conditions can cause a spike in mtROS production in the context of CVD. Our results show that this mtROS increase is the trigger that promotes mitochondrial crista disruption and ultimately leads to cell death.

The concomitant increase in superoxide and the decrease in SOD2 activity we saw with ISO treatment could allow the detection of superoxide increase in wild-type samples, whereas in the OMA1KO samples, the induction of superoxide was hidden by the maintenance of full SOD2 activity. Fluorescent assessment of ROS production has to be taken with care. Amplex Red can be artifactually converted to its “H2O2-modified” resorufin form by the mitochondrial carboxylesterase, thus yielding false-positive results (37). Therefore, the production of H2O2 that we observed could be overestimated.

Our results are in line with previous observations showing correlation between increased ROS and mitochondrial dysfunction (9, 10) and also with studies showing that diverse chemical or peptide scavengers ameliorate fatal consequences in various models of HF (3234, 38). MQ is currently being tested as a general clinical mitochondrial antioxidant for a number of conditions. Clinical trials have failed to show benefits in efficacy endpoints thus far, although studies have shown that the treatment has low toxicity. This does not preclude the use of mitochondrial antioxidants as a future treatment strategy; however, concern remains that under certain circumstances, mtROS ablation may interfere with their role as intracellular secondary messenger. Therefore, there is great interest in identifying novel mitochondrial targets for the treatment of HF.

The role of MCU in cardiac pathology has been studied in conditional knockout mouse models. Whereas lack of MCU protects against ischemia-reperfusion injury, its absence is detrimental to exercise performance, revealing a role in acute but not chronic heart stress (39). In our HF models with preserved or increased EF, we observed a pathological imbalance in mitochondrial calcium homeostasis due to increased MCU expression and activity that ultimately led to cardiomyocyte death. Recently, a novel regulatory pathway involving IF1 [the adenosine triphosphatase (ATPase) inhibitor factor] has been proposed to prevent OMA1 activation and thus OPA1 processing and crista remodeling (40). We hypothesize that the cross-talk between OMA1/OPA1/ATPase dimerization defines the prosurvival induction and that the protective action of OPA1 on mitochondrial function requires ATP synthase dimerization. Together, the results of Faccenda et al. (40) and ours suggest that both pathways converge on OMA1 as the downstream target that promotes cell survival.

Given the favorable efficacy and safety profile demonstrated in preclinical animal models in the current study, OMA1 seems to be a potential target for the prevention of myocardial cell damage, regardless of the cardiac insult. The fact that OMA1 inhibition protected both inbred and outbred mouse strains negates the hypothesis that protection might depend on factors restricted to a defined genetic context. In addition, it may be reasonable to anticipate that OMA1 inhibitors will have limited side effects, because OMA1 is only activated upon during stress conditions, and Oma1-deficient mice breed normally and do not present altered phenotype, other than exacerbated HFD-induced obesity and altered thermoregulation upon cold exposure. Further studies are required to elucidate the clinical potential of OMA1 inhibition in the context of HF; however, the preclinical results presented in multiple mouse models in this report support its potential as a pharmacological target.


Study design

We aimed to evaluate whether the lack of the mitochondrial stress protease OMA1, responsible for triggering of crista remodeling and cytochrome c release from mitochondria, could preserve cardiomyocyte survival and prevent HF. Experiments were designed with groups of at least four mice per condition. In some cases, experimental sets were repeated under the same conditions but at different times of the year to exclude seasonal variations. Mice were subjected to pharmacological or interventional (surgery) treatments to mimic HF. Experimental groups were randomly assigned. Blinded analysis was performed in the cardiac assessment in vivo. Biological replicates are noted in table S1.

Animal handling and procedures

C57BL/6JOlaHsd and CD1 wild-type mice were purchased from Harlan Laboratories. OMA1KO mice, on the C57BL/6 background, were generated as described in (19). All animal procedures conformed to European Union (EU) Directive 86/609/EEC and Recommendation 2007/526/EC regarding the protection of animals used for experimental and other scientific purposes, enforced in Spanish law under Real Decreto 1201/2005. Control and treated groups were assessed at the same time during experiments. Groups were randomly established, where control mice were housed with untreated mice. Mice were never segregated by treatment group. ISO (10 mg/kg per day; Sigma-Aldrich), AngII (1 μg/kg per min; Sigma-Aldrich), and MQ (0.15 mg/day, a gift from M. Murphy) were chronically administered with minipumps (Mini-Osmotic Pump Model 2004, Alzet) for 7 or 28 days. Maladaptive cardiac hypertrophy was induced by transverse aortic constriction (29). Mdivi1 (50 mg/kg; Sigma-Aldrich), caffeine (50 mg/kg; Sigma-Aldrich), and dantrolene (40 mg/kg; Sigma-Aldrich) were administered by daily intraperitoneal injection for 3 days (Mdivi1) and 8 days (caffeine and dantrolene).

Neonatal cardiomyocytes

Neonatal cardiomyocytes were isolated from 30 to 35 pups from day 0 to day 3 after birth, using the Worthington Neonatal Cardiomyocyte Isolation System. Cardiomyocytes were maintained in culture for 48 hours before experiments. Treatments were performed overnight with ISO (10 μM; Sigma-Aldrich), NAC (5 mM; Sigma-Aldrich), MQ (2 μM, a gift from M. Murphy), or Ru360 (10 μM; Calbiochem).

Design and injection of viral particles in mice

To produce adenovirus, we used the following vectors (VectorBuilder): for OMA1 silencing, pAAV[Exp]-cTNT>EGFP:5′miR30-mOma1[shRNA#1]-3′miR30; as controls, pAAV[Exp]-cTNT>EGFP:5′miR30-Scramble shRNA-3′miR30. Viral particles (AAV2/9) were generated at Boston Children Hospital, Boston, MA.

To test the efficiency in targeting OMA1 viral particles in the heart, 107 viral particles in 100 μl of phosphate-buffered saline (PBS) were injected via tail vein. Two serial injections (days 0 and 5) were performed, and then mitochondria from heart and liver were isolated as described (41). Mitochondria were incubated in the presence of carbonyl cyanide m-chlorophenyl hydrazone (Sigma-Aldrich) at 100 μM final concentration, for 0, 5, or 20 min at 37°C to activate OMA1 in vitro and thus OPA1 processing. Efficacy was evaluated by Western blot followed by quantification of OPA1 different isoforms.

Adeno-associated virus (AAV) was injected 6 and 2 days before TAC surgery. Mice were analyzed for in vivo cardiac function and mitochondrial performance 21 days after surgery.

Assessment of cardiac function in vivo

Ultrahigh resolution echocardiography was performed by operators blind to the origin of the animals with a linear 30-MHz transducer (Vevo 2100, VisualSonics Inc.) before and 21 days (TAC) or 28 days (minipumps) after the cardiac insult in 4- to 6-month-old mice. Mice were lightly anesthetized with 1 to 1.5% isoflurane in 100% oxygen and placed on a heated table to preserve physiological body temperature. HR was monitored and anesthesia delivery was adjusted to maintain an HR of 500 ± 50 beats/min. Two-dimensional and M-mode (MM) echocardiography was performed in parasternal long- and short-axis views.

LV ejection fraction (EF) by Teichholz formula (42) was assessed using M-mode images displayed over time and obtained by a single line in the middle of LV. LV mass was calculated from the same M-mode images using diastolic LV diameters of LV internal diameter (LVIDd), LV posterior wall (LVPWd), and interventricular septum (IVSd) as follows: LV mass (mg) = 1.05 [(LVIDd + LVPWd + IVSd)3 − LVIDd3]. Corrected LV mass = (LV mass) 0.8.

To assess LV diastolic function, spectral Doppler echocardiography (Vevo 770 system; VisualSonics Inc.), equipped with a 30-MHz linear transducer probe, was recorded for analyzing early (E wave) and late (A wave) velocities of mitral flow as previously described (43). Doppler mitral velocity flow pattern in mild LV diastolic dysfunction consists of a decreased E/A ratio (44), which is the result of elevated LV diastolic pressure and normal left atrium (LA) pressure. In a more severe stage of LV diastolic dysfunction, LA pressure is elevated, and consequently, there is an increase in the ratio of early/late mitral flow velocity (E/A ratio) (45).

Blood pressure

Blood pressure was measured by the noninvasive tail-cuff method using the BP-2000 apparatus from Visitech Systems.

Treadmill analysis

For the treadmill endurance test, mice were trained for 5 min on a treadmill at 10 cm/s and 20° slope the day before [LE8700 (76-0303), Treadmill Panlab, Harvard Apparatus]. On the experimental day, the treadmill was set at the same training conditions, and mice were placed on the treadmill to run. The experiment was stopped when the mouse spent 3 s in the back of a rack.

Serum biochemistry

Blood samples were obtained from cardiac puncture and kept at 4°C during serum preparation. Serum was obtained by centrifugation (12,000 rpm) during 12 min at 4°C. Serum biochemistry was assessed the day of extraction with a Dimension RxL Max automated analyzer.

Glucose metabolism assessment

Mice were fed an HFD (60% kcal fat, 1.5% kcal cholesterol; Research Diets Inc.) for 8 to 10 weeks and weighted every week. For GTTs and insulin tolerance tests (ITTs), mice were fasted for 16 or 2 hours, respectively, with free access to water. For intraperitoneal GTT, the mice received glucose injections of 1 mg/kg. For intraperitoneal ITT, the mice received insulin injections of 0.75 U/kg. Mice were bled from a tail clip, and blood glucose was measured with a handheld glucometer before injection (time 0) and at the indicated times after injection.

Histopathological staining

Tissue samples were fixed in 4% paraformaldehyde (24 hours) and processed and embedded in paraffin. Sections (5 μm) were prepared and mounted on slides for staining with hematoxylin and eosin or Masson’s trichrome. Samples were digitalized using the NDP.view2 software, and collagen fibers were calculated as trichrome-positive area/total area for heart sections or as trichrome-positive area/total perimeter for aortic sections. Media-to-lumen ratios were calculated by measuring the media diameter and lumen area in aortic sections using ImageJ.

Determination of superoxide in cardiac tissue

Cardiac tissue was dissected, mounted in optimal cutting temperature embedding compound and cryostat-sectioned (12 μm). Slides were defrosted and incubated for 45 min at 37°C with 5 mM dihydroethidium (DHE; Thermo Fisher Scientific). Samples were washed twice with PBS and incubated 5 min at room temperature with 4′,6-diamidino-2-phenylindole (DAPI) solution (0.1 mg/ml) for counterstaining of all nuclei. Slides were washed twice with PBS and water and mounted in ProLong Gold Antifade Reagent. Confocal fluorescence microscopy images were obtained on a Leica SPE-Upright confocal microscope using HCX PL APO CS 40× 1.25 oil objective. Images were acquired with LAS-AF version 2.3.6 acquisition software (Leica Microsystems). DHE and DAPI fluorescence of cardiac sections was quantified using ImageJ version 2.0.0. The mean DHE fluorescence was quantified and expressed relative to values obtained for DAPI mean fluorescence.

Isolation of mitochondria and mitochondrial function assessment

Mitochondria were isolated from mice as described (41). ATP synthesis was assessed in isolated mitochondria from heart using glutamate plus malate or succinate as fuels in the presence of adenosine 5′-diphosphate (15 to 25 μg of mitochondrial protein) by a kinetic luminescence assay using the luciferin/luciferase reaction (46). Aconitase activity was measured in isolated mitochondria as reported (36).

ROS measurement in isolated mitochondria

H2O2 release in isolated mitochondria was measured by Amplex Red according to the manufacturer’s instructions (Molecular Probes).

Transmission electron microscopy

Heart samples were collected from untreated and treated mice and fixed in 2.5% glutaraldehyde and 4% formaldehyde in 0.1 M Hepes buffer for 4 to 5 hours. Samples were processed as described (47). Samples were examined on a JEOL 10-10 electron microscope through 5000×, 40,000×, and 80,000× objectives. Mitochondrial morphometry was analyzed using ImageJ version 1.48 (National Institutes of Health).

Western blot analysis

Protein expression was analyzed in mitochondrial or cell lysates by Western blot after SDS–polyacrylamide gel electrophoresis. All Western blots show a representative gel of two or three gels from independent experiments. The following antibodies were used: SDHA (Novex); ATP5B, OPA1, glyceraldehyde-3-phosphate dehydrogenase, core1, core2, and MnSOD (Abcam); MCU and M1CU1 (Sigma-Aldrich); and Tom20 (Santa Cruz Biotechnology).

NADH measurement

NADH was measured spectrophotometrically in heart homogenates with a Colorimetric NAD/NADH kit (Abcam).

Mitochondrial SOD activity

The activity of mitochondrial SOD (KCN-insensitive) was measured in heart homogenates (50 μg) using the SOD Assay Kit (Fluka/Sigma-Aldrich).

Transcriptome analysis

Transcriptome analysis was conducted with nine 10-week-old male mice per genotype and treatment as described (47). Sequencing adapter contaminations were removed from raw reads using cutadapt software (48), and the resulting reads were mapped and quantified on the transcriptome (Ensembl gene-build 70) using RSEM version 1.2.20 (49). Only genes with at least 1 count per million in at least three samples were considered for statistical analysis. Data were then normalized using trimmed mean of M values and deferential expression tested using the bioconductor package edgeR (50). We considered as differentially expressed those genes with a Benjamini-Hochberg adjusted P ≤ 0.05. Functional analysis on differentially expressed genes was performed on Ingenuity Pathway Analysis (IPA) platform ( for those genes with a logFC over 0.5. The IPA Canonical Pathways represented in the figures, using GOPlot R package (51), have been manually selected by biological relevance among significantly enriched pathways (Padj < 0.05).

Seahorse analysis

Oxygen consumption in neonatal cardiomyocytes (75,000 cells per well) was measured using the XF96 MitoStress Test (Seahorse Bioscience). Oxygen consumption rates were normalized to cell number using CyQuant (Molecular Probes).

ROS, mitochondrial membrane potential, and calcium measurements in neonatal cardiomyocytes

ROS, mitochondrial membrane potential, and intracellular (intramitochondrial) and extracellular calcium were measured in neonatal cardiomyocytes cultured in 96-well plates, using 2′,7′-dichlorofluorescein diacetate (H2-DCFDA, 0.4 μM final concentration; Thermo Fisher Scientific), tetramethylrhodamine methyl ester perchlorate (200 nM final concentration; Sigma-Aldrich), and fura-2-acetoxymethyl ester (Fura-2 AM; 1 μM final concentration; Thermo Fisher Scientific), respectively, according to the manufacturer’s instructions.

In vivo live imaging of neonatal cardiomyocytes

Mitochondria were labeled with MitoTracker Deep Red FM (50 nM final concentration; Molecular Probes) for 30 min at 37°C. Fura-2 AM (1 μM final concentration) was then added to cardiomyocytes for 10 min before commencing time-lapse live imaging (10 min) with a Nikon ECLIPSE Ti-Time Lapse inverted microscope coupled to a Hamamatsu Orca ER camera. Compounds were added at the following concentrations just before monitoring of calcium and mitochondrial distribution: ISO (100 μM), Ru360 (10 μM), and Dant (10 μM). Calcium (Fura-2 AM) distribution was quantified using ImageJ version 2.0.0. Kurtosis analysis was performed as a measurement of signal clustering by comparing basal (time 0) to final (time 9) Fura-2 AM signal in individual cardiomyocytes.

Statistical analysis

Comparisons between groups were made by one-way or two-way analysis of variance (ANOVA). Correction for multiple comparisons was made by Tukey or Sidak tests when appropriate. Pairwise comparisons were made by two-sided Student’s t test. Differences were considered statistically significant at P < 0.05. In the figures, asterisks denote statistical significance (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001). Data were analyzed with GraphPad Prism 7. In the figures, each point represents a biological replicate.


Fig. S1. Histological analysis of brain, liver, and kidney sections upon work overload.

Fig. S2. Short chronic ISO administration (7 days) promotes ROS increase and mitochondrial crista remodeling.

Fig. S3. Modulation of mitochondrial remodeling and calcium homeostasis.

Fig. S4. Analysis of heart performance upon pressure overload.

Fig. S5. Assessment of cardiac function after HFD administration.

Table S1. Number of samples used in the experiments shown in Figs. 1 to 7.


Acknowledgments: We thank A. Cecconi and R. Rizzuto for scientific discussion; M. Murphy for scientific discussion and reagents; J. M. Redondo for reagents; M. De la Cueva for mouse work; A. Dopazo, S. Callejas, and A. Benguría for RNA-seq experiments and design; A. de Molina and R. Doohan for histology and pathology interpretation; M. L. García-Gil and Instalaciones Científico-Tecnológicas Singulares-Centro Nacional de Microscopía Electronica (Universidad Complutense de Madrid) for technical support; A. V. Alonso and L. Flores for echocardiography acquisition and analysis; N. A. Miller for imaging interpretation and analysis; and S. Bartlett [Centro Nacional de Invesigaciones Cardiovasculares (CNIC)] for English editing. Funding: This study was supported by grants from the Ministerio de Economía y Competitividad (SAF2012-32776 and SAF2015-65633-R), the Comunidad de Madrid (S2011/BMD-2402), the EU (ITN GA317433 and MC-CIG GA304217), BBVA Foundation Grant for Researchers and Cultural Creators and the Instituto de Salud Carlos III (Fondo de Investigaciones Sanitarias grants PI09-00946, PI12/01297, and PI11-00078). The CNIC is supported by the Ministerio de Economía y Competitividad and the Pro-CNIC Foundation and is a Severo Ochoa Center of Excellence (Ministerio de Economía y Competitividad award SEV-2015-0505). This study was also funded by FEDER. Author contributions: R.A.-P.: molecular and biochemical experiments; A.V.L.-V.: TEM analysis; M.d.M.M., R.N.-A., A.G.-G., and I.C.: animal handling and animal experimental setup; C.B.: neonatal cardiomyocyte isolation; C.T. and F.S.-C.: transcriptome data analysis; P.M.Q. and C.L.-O.: provision of OMA1KO mice; L.J.J.-B.: echocardiography analysis; J.R.-C.: imaging analysis; C.J. and J.M.C.: scientific discussion; J.A.E., R.A.-P., C.J., J.M.C., and L.J.J.-B.: paper writing; and J.A.E. and R.A.-P.: designing of the project and the experiments. Competing interests: The authors declare that they have no competing interests. Data and materials availability: Raw RNA-seq data were submitted to the Gene Expression Omnibus repository corresponding to GSE96057 accession number.

Stay Connected to Science Translational Medicine

Navigate This Article