Intracellular Na (Na i ) elevation is a hallmark of the ischemic and failing heart – pathologies in which both acute and chronic metabolic remodeling occurs. Increased Na i is known to decrease mitochondrial calcium (Ca mito ) via enhanced Na/Ca mito exchange, potentially compromising ATP mito production leading to metabolic dysregulation. We aimed to examine whether acute (ouabain+blebbistatin) and chronic myocardial Na i load (PLM 3SA transgenic mouse) is causally linked to metabolic remodeling and if pathological hypertrophy (pressure overload) shares a common Na-mediated metabolic ‘fingerprint’. 23 Na, 31 P and 13 C nuclear magnetic resonance spectroscopy (NMR) was performed in retrogradely perfused mouse hearts followed by 1 H NMR metabolomic profiling, mass spectrometry and in silico modelling. Na i overload (acute, chronic, pathological) resulted in common metabolic perturbations: switch in substrate oxidation from fatty acid to glucose and altered steady-state metabolite concentrations (glycolytic, Krebs cycle, anaplerotic). Inhibition of Na/Ca mito exchange by CGP37157 during both acute and chronic Nai load ameliorated the metabolic changes. In silico modelling predicted increased metabolic fluxes (TCA cycle, OXPHOS, glycolysis, glucose oxidation) at the expense of metabolic coupling (glycolysis/glucose oxidation). Neither acute nor chronic Na i elevation resulted in energetic impairment (PCr/ATP) suggesting an adaptive response. Therefore, elevated Na i leads to complex adaptive metabolic alterations preceding energetic and functional impairment. Early prevention of Na i overload or reversal of the metabolic remodeling (Na/Ca mito inhibition) could potentially ameliorate the origin of metabolic dysregulation in cardiac hypertrophy and failure. These changes constitute a common metabolic ‘fingerprint’ in response to Na i elevation which appear to be independent of its aetiology or duration.
Recently both 3D high resolution and functional studies in muscle cells have revealed a tightly coupled mitochondria reticulum (MR) to rapidly distribute potential energy, in the form of the mitochondrial membrane potential (MMP), throughout skeletal muscle and heart cells. Herein the structural aspects of the MR are described using 3D Focused Ion Beam Scanning Electron Microscopy (FIB-SEM) and presented in muscle cells. A large portion of the MR conductivity is dependent on direct mitochondrial matrix continuity while in some regions of the muscle the connectivity is proposed to occur via poorly characterized electron dense regions (EDR) between adjacent mitochondria. Using a photo-activated mitochondrial uncoupler to regionally perturb the MMP, we have demonstrated that large regions of the MR are electrically coupled via a shared matrix as well as EDR structures between the mitochondria. In murine skeletal muscle cells a large fraction of the mitochondrial volume is located in regions close to capillary indentations in the cell structure. These embedded capillaries are surrounded by large pools of mitochondria near the plasma membrane that have narrow tubes which run along the I-Bands (I-Band Mitochondria Segments (IBMS)) deep into the muscle cell. It has been proposed that these IBMS serve to distribute the MMP from the large sub-sarcolemma mitochondrial pool to the more central ATP-consuming myofibril region of the muscle cell. Consistent with this notion was the observation that there is a 3-fold enhancement of MMP generating oxidative phosphorylation complex in comparison to MMP utilizing ATP synthesis enzymes in the periphery of the muscle cell when compared to central regions near the muscle ATPase activity. In cardiac cells, no IBMS exist and the coupling is exclusively through large mitochondria structures and numerous EDR connections. These data are consistent with a mitochondria reticulum in muscle cells that couples large numbers of mitochondria together providing a rapid and uniform potential energy source throughout the cell to support ATP production.
Increasingly we see an expectation of computational approaches (including AI) to identify previously unforeseen patterns and to inform existing human decision making. The prediction of cardiac safety liability is no exception to this trend, and the Comprehensive in-vitro Proarrhythmia Assay (CiPA) initiative is one such example. However, whilst the strategy for using high-throughput in-vitro data in computational models for calculating proarrhythmic potential is a bold and welcome approach, it is not without risk. One such risk-factor is how we should evaluate the computational models to decide whether they are fit-for-purpose (or at least no worse) in deciding about compound/drug progression. Categorisation of the proarrhythmic potential of a set of well characterised drugs is one such example when used to calibrate and validate the computational model. In this study we show how we have taken a parallel approach to consider the proarrhythmic potential of a series of known drugs in order to offer a more quantitative basis for risk assessment based on real world observation of cardiac adverse events. We demonstrate how this novel use of real world data correlates with previous risk categorisation and discuss the implications for model validation.
Pulmonary arterial hypertension initiates a cascade of pathological events in the heart. The right ventricle (RV) endures pressure overload before it hypertrophies, leading ultimately to right-ventricular failure (RVF). The left ventricle, in contrast, suffers atrophic remodelling, as RV hypertrophy causes the septum to bow leftwards, subjecting the LV to reduced load and prompting it to undergo a reduction in mass and cavity size. Does atrophic remodeling affect the energetics, in particular the energy efficiency, of the LV? To answer this question, we studied RVF rats 6 weeks following a single injection of monocrotaline which progressively damages the endothelial cells lining the vessels of the pulmonary circulation. Experiments were performed at body temperature and were designed to measure LV mechanoenergetic performance at two physiological levels. At the ex vivo working-heart level, the LV generated a lower pressure-volume work output in the RVF group compared with th! e Control group, over a wide range of afterloads. The lower work output was commensurate with a lower oxygen consumption, and, hence, energy efficiency was unchanged. At the in vitro tissue level, LV trabeculae of the RVF group liberated lower heat output. Their force-length work output was not significantly lower and, hence, efficiency was unchanged. Our collective data show that in hypertrophied right-ventricular failure secondary to pulmonary arterial hypertension, the left ventricle, despite undergoing atrophy, has normal energy efficiency, over a wide range of afterloads, at both the organ level and the tissue level.
Current methods of echocardiography provide detailed information about heart wall motion during the cardiac cycle. However, the ability to interpret these data in terms of cardiac muscle properties is limited. The eventual goal of the present work is to develop theoretical models for left ventricle (LV) dynamics that allow estimation of myocardial stiffness and contractility from echocardiogram data. To reduce the computational complexity of this inverse problem, we represent LV deformation in terms of a limited number of modes. We have previously presented an axisymmetric model that describes the essential motions of the LV myocardium with three volume-conserving deformations. Here, we extend this framework to a more general set of volume-conserving deformations. The reference and deformed configurations of the LV are described using prolate spheroidal coordinates. The initial boundaries of the LV are defined by splines. We derive a family of mappings using a relatively small number (< 20) of time-dependent parameters to represent the main modes of LV deformation. These deformations are chosen to closely represent the important characteristics of LV motion, such as contraction, torsion, and elongation. Non-axisymmetric modes of deformation, such as those resulting from differences between the septal and lateral walls, are included. We construct a mechanical model that incorporates muscle fiber orientations, active and passive stresses, and surface tractions. We compare this model to the axisymmetric model to evaluate the relative importance of these additional deformation modes in determining overall cardiac function, as a step toward better quantification of heart disease.
Dialysis is a repetitive sub-lethal ischemia1. Cardiac CT imaging has shown dialysis to alter ventricular perfusion heterogeneity, which may be related to ischemia induced loss of mechanical function. Using a model of human coronary vasculature, we investigated cause-effect relationships between blood viscosity, artery occlusion, and blood vessel radius to perfusion heterogeneity.
CT images from three patients pre- and post-dialysis were analysed. Perfusion heterogeneity in each heart was quantified using fractal dimension2. Anonymised clinical data for each patient were obtained. A bi-ventricular anatomy was generated as truncated ellipsoids3. Morphometry data in conjunction with the space filling algorithm of Beard and Bassingthwaighte4 were used to generate multiple instances of coronary vasculature networks up to a resolution of 0.07 mm (imaging resolution). Conservation and scaling laws were applied to compute biophysical variables including perfusion. The fractal dimension of small sections of the model was computed to establish the baseline perfusion heterogeneity. The model was implemented using in house C language codes. The simulations are being performed on Compute Canada HPC resources.
Among other results, an increase of fractal dimension between pre- and post dialysis CT scan data. The model indicated that a major factor is reduction of small arteriole diameter.
1C. W. McIntyre,Seminars in dialysis 23, 449 (2010). 2J. B. Bassingthwaighte, et al.,Circ Res 65, 578 (1989). 3J. C. Mercier, et al.,Circulation 65, 4962 (1982). 4D. A. Beard, et al.,Journal of vascular research 37, 282 (2000).
Cardiac muscle converts the chemical free energy obtained by oxidation of metabolic substrates into mechanical work. Overall thermodynamic efficiency is the fraction of substrate free energy that appears as work. The overall energy cost of a cardiac twitch comprises two conceptually distinct contiguous components: "Initial metabolism (I)" and "Recovery metabolism (R)". Initial metabolism funds electrical excitation, Ca2+ triggering and work by the contractile proteins, events fuelled by the free energy of ATP hydrolysis. There is sufficient ATP to fuel only a brief period of activity. Buffering of ATP by PCr and regeneration of PCr via mitochondrial ATP production underwrites sustained activity. Regeneration of PCr by ATP produced in the mitochondria constitutes Recovery Metabolism. The initial and recovery processes are coupled in series. Therefore, overall thermodynamic efficiency is the product of the efficiencies of the initial and recovery processes.
Thermodynamic efficiency cannot be measured directly, so must be inferred from mechanical efficiency, the ratio of work to overall enthalpy output. Since the difference between enthalpy and free energy arising from the oxidation of either carbohydrates or fatty acids is negligible, overall thermodynamic efficiency can be safely approximated by overall enthalpy production. Rescaling the free energy of ATP hydrolysis (60 kJ/mol) by the enthalpy of PCr hydrolysis (34 kJ/mol), and adopting our measured value of activation enthalpy (20% of total energy expenditure), allows us to conclude that the thermodynamic efficiency of mitochondrial oxidative phosphorylation of either fats or carbohydrates is approximately 0.8 while that of cross-bridge cycling is approximately 0.24.
Approximately 6.5 million Americans are affected by heart failure, a condition where the heart is unable to meet the blood and oxygen demands of the body. Pathophysiological changes that occur during heart failure include remodeling of the myocardium and alterations in the metabolic state that are associated with the depletion of key metabolic pools in the heart. Based on multi-scale computer simulation of mechanical-metabolic interactions in the myocardium, we predict that depletion of the myocardial adenine nucleotide pool in heart failure impedes the ability of oxidative phosphorylation to supply ATP at the hydrolysis potential necessary for proper physiological function of the heart. Furthermore, we hypothesize that the resulting changes to the phosphate metabolite levels impair mechanical function of the heart, and contribute to the heart failure phenotype. To test these hypotheses, transverse aortic constriction (TAC) surgery was performed on 3-week-old Sprague Dawley rats to induce heart failure. Mechanical function of the heart was measured by echocardiography at several time points post-surgery and, after 15 weeks, hearts were excised to measure mitochondrial oxidative capacity and myocardial cytosolic metabolites. We observe reductions in fatty acid and carbohydrate oxidative capacities and adenine nucleotide pool levels that correlate with impaired mechanical function. Furthermore, with decreased ejection fraction there is a decrease in the magnitude of the ATP hydrolysis potential. Taken together these results suggest that depletion of the adenine nucleotide pool and reduction in mitochondria oxidative capacity both contribute to metabolic and mechanical dysfunction in this model. Future studies include investigating whether cardiac function can be improved in heart failure by inhibiting or reversing adenine nucleotide degradation and efflux. Preliminary studies on an inducible ventricular 5' nucleotidase knockout suggest that by inhibiting adenine nucleotide pool depletion, the energetic status and mechanical function of the myocardium in heart failure may be improved.
Understanding and predicting how ion currents contribute to cellular electrophysiology is critically dependent on the characterization of ion channel kinetics. This characterization is in turn critically dependent on mathematical models of the kinetics, but there is little consensus on which recordings to use (and how to use them) to derive the models. As a result, different literature models of the same current often provide very different predictions. We present a method for rapidly exploring and characterizing ion channel kinetics, using the hERG channel, responsible for cardiac IKr current, as an example. We fit a mathematical model to currents evoked by a novel 8 second sinusoidal voltage clamp. The model is then used to predict over 5 minutes of recordings in the same cell in response to further voltage clamp protocols, including a new collection of physiological action potentials. Our technique allows rapid collection of data from single cells, produces more predictive ion current models than traditional approaches, and will be widely applicable to many electrophysiological systems..
Despite extensive studies attempted to identify clinical indices for the selection of patient candidates who benefit from cardiac resynchronization therapy (CRT), there still remains a significant number of non-responders among those treated by this invasive therapy. A multi-scale heart simulation capable of reproducing the electrophysiology and mechanics of a beating heart may help resolve this problem. In a retrospective study including nine patients, we tested whether we can reproduce the response to CRT using the patient-specific multi-scale heart simulator. Based on the clinical data collected before CRT, we created patient-specific failing heart models with conduction block. Each model was tailored to reproduce the surface electrocardiogram and hemodynamics of each patient. To each heart model, we performed bi-ventricular pacing according to the actual pacing protocol and compared the results with the clinical data recorded after the treatment. CRT simulation improved the electromechanical dyssynchrony by varying degree for each patient. The best correlation between simulated and clinical results was obtained for the simulated maximum value of the time derivative of ventricular pressure (dP/dtmax) and the clinically observed improvement in the ejection fraction (EF). By integrating the complex pathophysiology of the failing heart with conduction block, patient-specific, multi-scale heart simulation could successfully reproduce the response to CRT. With further verification and improvement through the prospective study, this technique could be a useful tool in clinical decision making.
Increasingly we see an expectation of computational approaches (including AI) to identify previously unforeseen patterns and to inform existing human decision making. The prediction of cardiac safety liability is no exception to this trend, and the Comprehensive in-vitro Proarrhythmia Assay (CiPA) initiative is one such example. However, whilst the strategy for using high-throughput in-vitro data in computational models for calculating proarrhythmic potential is a bold and welcome approach, it is not without risk. One such risk-factor is how we should evaluate the computational models to decide whether they are fit-for-purpose (or at least no worse) in deciding about compound/drug progression. Categorisation of the proarrhythmic potential of a set of well characterised drugs is one such example when used to calibrate and validate the computational model. In this study we show how we have taken a parallel approach to consider the proarrhythmic potential of a series of known ! drugs in order to offer a more quantitative basis for risk assessment based on real world observation of cardiac adverse events. We demonstrate how this novel use of real world data correlates with previous risk categorisation and discuss the implications for model validation.
Introduction: Atrial fibrillation (AF) is the most common arrhythmia, and current treatment is suboptimal. Accurate representation of 3D atrial anatomy and its underlying structure from medical images provides an effective patient-specific approach for clinical diagnosis of atrial scar and targeted treatment for patients with AF. We propose an algorithm to automatically segment the atria to visualize its geometry in 3D.
Methods: A gadolinium-enhanced (GE)-MRI dataset (N=60), with a spatial resolution of 0.625x0.625x1.25 mm3 for patients with AF(provided by the University of Utah), was used to develop a deep convolutional neural network to recognize features specific to atrial and non-atrial tissue. The network contained a dual pathway for both local and global feature detection. The network was fully convolutional and made pixel-wise binary classifications. The MRI image slices were standardized and fed into the algorithm. The algorithm was trained with 50 patient datasets on one GPU (384 cores) and then made to perform slice-by-slice predictions on the remaining 10 patients datasets to evaluate the performance accuracy.
Results: The overall accuracy for the 10 validation datasets was 76 ± 12.6%. Predictions for each patient took ~1 minute. The outputs showed clear detail of the atrial wall thickness but were heavily influenced by the imaging quality of the initial MRI scans.
Conclusions: We have demonstrated a deep neural network to perform fully automatic 3D atria reconstruction on medical images. Potentially, future computer modeling based on the robust, accurate atrial anatomy will lead to our improved understanding of AF and to guide AF treatment more effectively.