Modelling the human heart: an integrated experimental and computational study

The proposed project will provide new experimental data from human hearts that will be used to make a realistic computer model of the human heart.

Heart disease is the main cause of mortality in the developed world. This disease is characterised by a reduced ability of the heart to pump blood, due to changes to the mechanical and electrical properties of heart cells. However, despite extensive experimental studies, the complicated sequence of events leading from altered function at the cellular level to life-threatening pump failure remains poorly understood.

This situation has motivated rapid advances in the development of computer models of the heart that now provide new and powerful quantitative tools for understanding the triggers and progression of heart disease. These models have delivered an important means for capturing the complex function of the heart by establishing a consistent, quantitative and multi-level framework for integrating measurements and understanding. From this work, important insights into the inter-relationships between cell contraction, heart shape and muscle structure have already been revealed.

However, while the scientific importance and significant clinical potential in this approach is widely acknowledged, the promise of these computer models to increase our understanding of human heart function remains largely unfulfilled. This is because the vast majority of cardiac mathematical models are currently developed and validated using data collected from measurements in animal, rather than human, experiments. Furthermore these experiments are often performed under conditions that are very different from the environment of either a normal or diseased heart in the body. In particular, individual cells are held at constant length and are studied in the cold, whereas in the intact heart the cell length changes during the heartbeat and the cells are at body temperature; these differences alter profoundly the amount of force muscle cells can produce.

This situation means that there are inherent limitations to using animal-based model frameworks and experimental data for understanding human heart function and for answering clinical questions. Thus an important challenge to address is the development of a model for the human heart that can be applied directly in clinical contexts.

Recently we have developed the capacity to collect unique data on isolated cells from human hearts. Importantly these measurements can be performed at body temperature and the cell length can be changed to mimic the full range of conditions the cells experience as the heart beats. This information enables, for the first time, the ability to construct a model that will be able to directly capture human heart function. To achieve this goal, mathematical equations representing these human heart cells will be developed and combined using high-performance computers to construct a computational model - a 'virtual' heart. The heart's pumping capacity will then be computed under different conditions and linked back to the human heart experiments.

Using this virtual heart we will be able to isolate the important mechanisms that govern how the human heart responds to meet the wide range of requirements placed on it. Specific examples include understanding the cellular changes that enable the heart to pump larger volumes of blood, such as in exercise, or produce more force in conditions of high blood pressure.

Finally, by publishing the experimental data and model, and by making all of our computer code freely available, we will enable other heart modellers to use our model to perform their own human-based simulations. Through this work this study will provide a new way to investigate and understand human heart function and ultimately the progression of heart disease, together with ways to improve its diagnosis and prevention.

Rapid advances in computing now provide the potential for the development of powerful quantitative models for understanding the working of the human heart in health and disease. Finite-element based models that accurately represent heart structure and microstructure have incorporated cellular models of electrical activation and force generation to produce cardiac contraction. However, all previous heart models are severely limited in their applicability to the human heart because they used myofibrillar contraction data that were obtained from experiments using animal myofibrils at sub-physiological temperatures (the latter due to preparation degradation at higher temperatures); the contractile properties of human cardiac myofibrils at 37oC are currently unknown.

We have devised a novel technique for studying the contraction of human (and animal) cardiac myofibrils at 37oC. In a comprehensive study, we will establish the calcium-, sarcomere length-, and velocity-dependence of force production during calcium-activated contractions of human cardiac myofibrils. This rich, physiologically-relevant dataset will be shared with the academic community. In addition, we will use this dataset to populate a mathematical model of human myofibrillar contractile dynamics, the predictions of which will be tested with further experimentation, so that we can refine the model until it provides excellent simulations of cross-bridge dynamics under all conditions relevant to heart function. This myofibrillar model will then be integrated into a model of the contracting cardiac myocyte which will, in turn, form the core element in a realistic whole-heart computational model. The heart model will be validated against real data obtained from patient imaging studies. The coupling of the unique human data acquired during this project with our heart model will be used to explore which parameters of myofibrillar function are key determinants of the pump function of the working human heart.

The output of this project will be a novel human heart contraction model and a unique analysis of the coupling mechanisms by which individual cardiac cells regulate their contractile response to produce integrated contraction of the whole heart. This work will provide the ability to quantitatively characterise the highly complex regulation of contraction in the human heart that is currently beyond analysis by clinical observation and intuition alone. For these reasons the project output has potential to provide significant impact for cardiac healthcare workers, the patient cohorts they treat and the cardiac imaging industry.

Specifically the work of this proposal will enhance the capacity of our simulation software to predict human physiological function This addition will in turn enable the application and implementation of our models within the clinical environment and the ability to engage medical imaging companies seeking to integrate simulation software within their current hardware products. Two specific examples of translational pathways to impact are outlined in further detail in the sections below.

Improved patient selection for Cardiac Resynchronisation Therapy (CRT): Despite the increasing prevalence of heart failure, it continues to have a terrible prognosis with 50% mortality in the first 3 years after diagnosis, worse than most malignancies. Randomized, controlled clinical trials have shown that some patients benefit from CRT, in which a pacemaker is embedded into the heart wall. However, there are still major issues associated with patient selection, since up to a third of the patients treated do not show any response to this very expensive therapy. Our preliminary work has shown the fundamental cellular excitation-contraction coupling mechanisms are likely to underlie the clinical response. However, these results are sensitive to the exact parameterization. If this result is confirmed via detailed model development with consistent human-derived cell data, this project will have a significant impact on the treatment and selection of CRT patients.

Coronary Artery Disease (CAD) analysis of images: Despite its significance, the determination of optimal clinical diagnosis and treatment strategies for CAD patients remains problematic. Exacerbated by the high risk of the disease, and the difficulty in excluding it, the clinical problem is demonstrated tangibly by the large number of patients who currently undergo invasive angiography unnecessarily and achieve negative results. Central to both diagnosis and treatment of CAD is the relationship between perfusion and contraction. In particular, understanding if a region of tissue is mechanically viable and thus a candidate for reperfusion is extremely valuable. The application of a human contraction model using the strain- and motion-derived imaging information will provide the ability to estimate tension and thus contractility of tissue. From this information a much-improved prediction of the resulting benefits from revascularisation can be determined, which, in turn, will provide the ability to select patients more accurately.

Finally, in addition to dissemination of this research in high-quality journals and scientific meetings, the investigators will continue their work engaging with media to further the impact of the project with the general public. Recent examples of this engagement include appearance on the BBC Horizon documentary "How to Mend a Broken Heart" and winner of the BHF-sponsored Reflections on Research award.