DPhil the Future: Using computers to get to the heart of the matter

DPhil student Leto L Riebel is using computational tools to progress the challenge of regenerating the injured human heart 

In the Computational Cardiovascular Science group, at the Department of Computer Science, our goal is to utilise computational power to investigate the response of the human heart to disease and therapy and to augment experimental and clinical investigations. In my project, I am simulating the response of the injured human heart to regenerative therapy, specifically, stem cell therapy.  

‘Despite improvements in therapies and clinical interventions, cardiovascular disease remains the leading cause of death worldwide.’ DPhil student Leto L Riebel

Myocardial infarction, commonly known as a heart attack, is often caused by the build-up and rupture of plaques in the coronary arteries, resulting in a lack of oxygen and nutrients supplied to the heart’s tissue. The affected tissue becomes necrotic and unable to contract properly, with consequent reduction of cardiac function, which can pave the way for heart failure.  

After myocardial infarction, reperfusion therapies can restore blood supply and limit infarct size. However, once damage has occurred, treatment options to reverse it are very limited. Stem cell therapy is being explored to restore cardiac function by replacing damaged tissue with new lab-grown heart muscle cells. Experimental studies and early clinical trials have been investigating optimal stem cell culture, maturation, and delivery conditions. However, despite continuous progress in the field, many challenges remain. Lab-grown heart muscle cells remain largely immature, expressing electrophysiological, structural, and functional properties more similar to foetal than adult human heart muscle cells. Slowed conductivity and spontaneous beating activity are some of the most concerning immature characteristics, as once the cells are introduced into the adult human heart, they could lead to desynchronised propagation of the heart’s electrical signal and ultimately hamper coordinated contraction.  

Using computational methods, we can build detailed bio-physically accurate and multiscale models, from a single heart muscle cell to the whole human heart. Such models have shown great capabilities, for example in simulating the heart’s function under disease conditions and predicting drug effects.  

Computational modelling and simulation provide a unique tool to control individual parameters and investigate mechanisms on a multiscale level, unfeasible in experimental or clinical settings. They can therefore add to insights from and even guide experimental and clinical studies, whilst being fast, cost-effective, and reducing the need to use animal testing.  

The goal of my DPhil project, which is funded by a BBSRC Industrial CASE scholarship in collaboration with AstraZeneca, is to develop a computational modelling and simulation framework to investigate the safety and efficacy of stem cell therapy in the human heart after myocardial infarction. Offering a pivotal perspective on the development and application of disease therapies, the collaboration with AstraZeneca has been crucial in defining and driving my project forward.  

During my first year, I focused on investigations at the cellular level, to identify the key model parameters characterising the differences between human stem cell-derived and healthy adult heart muscle cells. From these single cell insights, I then progressed on to introducing the stem cell-derived heart muscle cells into a three-dimensional computational model of the human heart. The 3D model, which consists of two heart chambers (left and right ventricle), also implements information acquired experimentally, such as observed regional differences in cellular properties and patient-specific features. Our 3D model also includes an infarcted region with specific properties, eg, slowed conductivity and ionic changes, based on experimental and clinical measurements. In total, the 3D model contains up to a few million elements, each represented by a single cell model, linked together through terms of conductivity. Solving the underlying mathematical equations in space and time is time and resource consuming. We are therefore using the fast state-of-the-art parallelised GPU model solver MonoAlg3D (Sachetto Oliveira et al., 2018). Close collaboration with the developer team in Brazil has allowed us to utilise the software’s full capabilities and efficiently adapt it to our projects’ needs and has sparked new joint research questions.  

In the coming months, I will use this computational framework to explore the safety of different stem cell delivery conditions (such as delivery location relative to the infarcted region), patient-specific characteristics (such as infarct size and progression), and ultimately suggest therapeutic targets to increase synchronicity of the heart’s electrical signal and reduce the risk of abnormal activity. My project joins many others in the Computational Cardiovascular Science group in continuing to display the promising capabilities of computational tools to investigate disease mechanisms and inform clinical decisions, aimed at optimising personalised treatment of cardiovascular disease.