Trine Krogh-Madsen

Cardiac arrhythmias remain a leading cause of mortality and morbidity. Cardiac arrhythmogenesis is complex with arrhythmia susceptibility depending on a large range of factors, including age, sex, race, genetic mutations, hormonal modulation, electrolyte imbalance, and autonomic imbalance. Therefore, it can be challenging to predict an individual's risk of developing arrhythmias. Similarly, the diversity within a population can influence treatment outcomes for a specific patient, including the efficacy of antiarrhythmic medications.  

As a means to develop a quantitative understanding of the impact of population heterogeneity on arrhythmogenesis, one of our focus areas is development of cell-specific mathematical models of cardiomyocytes. Our current approach includes combining automated parameter identification methods with complex data objectives. The modeling work is closely coupled to development of strategies for best recording electrical activity in cardiac myocytes, with experimental protocols optimized for model calibration and validation. We then use the models to provide new insights into physiological and pathological differences among individual cells.  

Another line of research focuses on determining pro-arrhythmia risk of new drugs. Before advancing to clinical trials, drugs must undergo pre-clinical testing of pro-arrhythmic risk. New strategies and regulations for pre-clinical screening are under development to improve specificity and provide mechanistic insights into drug effects. Our work in this area has included optimizing mathematical models to clinical data from patients with and without a congenital pro-arrhythmic syndrome to more accurately detect arrhythmic risk of ion-channel blocking drugs. In a related project, we applied optimization methods to develop experimental protocols that allow identification of pro-arrhythmia risk and ionic current block, thus developing a candidate drug screening approach that simultaneously gives risk and mechanism of action for each drug. 

A significant proportion of my research has involved developing new computational models or modeling methods. This work has been carried out in close collaboration with experimentalists. The current standard process of developing cardiomyocyte ionic models suffers from several limitations, including post hoc adjustment of model parameters and use of dynamically simple date to create models used to simulate dynamically rich behaviors. We have started to develop tools to overcome these limitations by combining automated parameter identification methods with complex data objectives (Refs. 1-3). As this methodology allows generation of cell-specific models, it increases both the overall fidelity of models and their utility, and we therefore believe it will carry significant impact. In recent work (Ref. 4), we have applied optimization methods to develop experimental protocols that allows simultaneously identification of pro-arrhythmia risk and ionic mechanism of channel-blocking drugs. This approach offers a new strategy for evaluating cardiotoxicity during preclinical drug screening. 

Figure 1

• Mathematical and Computational Modeling
• Systems Biology