Computational modelling for improved translation of cardiac inotropic and lusitropic drug effects from rats to humans

Fig. 1.. Transients and definition of characteristic biomarkers. (a) sarcomere length, (b) experiments: fluorescence intensity ratio (F340/F380), simulations: intracellular [Ca2+]. Sarcomere length biomarkers include the resting value (SLRes), the time from stimulus to peak (SLTP), percentage sarcomere shortening (Short) defined as amplitude relative to the resting value expressed as percentage, the times from peak to 50 % (SLTR50) and 90 % return of the transient (SLTR90), the maximum rate during contraction from stimulus to peak (dSLMaxC), and the maximum rate during relaxation from peak onwards (dSLMaxR). Intracellular [Ca2+] biomarkers include the resting value (CaRes), the time from stimulus to peak (CaTP), the amplitude (CaAmp), the time from peak to 50 % (CaTR50) and 90 % return of the transient (CaTR90), the maximum rate from stimulus to peak (dCaMax), and the minimum rate from peak onwards (dCaMin).

Fig. 2.. Concept of the computational drug effect translation approach that predicts human drug effects on sarcomere length and intracellular [Ca2+] biomarkers at multiple concentrations (C) based on measured rat drug effects.

Fig. 3.. Simulated transients before and after recalibration of the rat and human cardiomyocyte models. (a) Sarcomere length and (b) intracellular [Ca2+] transients are shown, with transients from the rat and human cardiomyocyte modelsindicated in red and blue colour, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4.. Evaluation of the rat and human cardiomyocyte models for their representation of physiological behaviour. (a) Sarcomere length and (b) intracellular [Ca2+] transients simulated at the pacing frequencies 0.5, 1, and 2 Hz are shown, and simulated and experimental pacing frequency relationships of biomarkers related to inotropy and lusitropy are compared. Red and blue colour indicate results of the rat and human cardiomyocyte model, respectively. Upwards and downwards arrows indicate positive and negative relationships, respectively. Note that the extracellular ion concentrations in the simulations with the human cardiomyocyte model were not adapted to match the experiments due to incomplete data availability. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5.. Global sensitivity analyses to identify different inotropic and lusitropic responses to inhibition of ion channels and transporters. Normalised total-effect sensitivity indices are shown and higher sensitivity is indicated by darker colour. The chosen biomarkers of inotropy and lusitropy are indicated in bold with framing, with inotropy markers in purple and lusitropy markers in green. The proportion of excluded transients was 2 % and 0 % in the analyses on the rat and human cardiomyocyte model, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 6.. Identifiability analysis to inform the selection of proteins considered in the computational drug effect translation. Scatter plots show inferred versus actual blocking potencies and associated mean absolute errors (MAE). Bold frames indicate proteins considered in the computational drug effect translation (MAE ≤ 0.1).

Fig. 7.. Evaluation of computational drug effect translation using synthetic data. (a) Scatter plots of inferred versus actual blocking potencies for RyR2, SERCA2, and NCX1, with mean absolute errors (MAE) provided. (b) - (o) Scatter plots of synthetic rat drug effects (upper panels) and predicted human drug effects (lower panels) versus synthetic human drug effects on the characteristic biomarkers. Results for increasing drug concentrations (0.01, 0.1, 1, and 10 μM) are shown from left to right, with coefficients of determination (R2) provided.

Fig. 8.. Evaluation of computational drug effect translation using experimental data collected for thapsigargin at four concentrations. (a) Experimental rat drug effects, corresponding fitting results, and predicted human drug effects. (b) Qualitative comparison of experimental rat drug effects, predicted human drug effects, and experimental human drug effects. Green, red, and black colours indicate positive (≥ 10 %), negative (≤ −25 %), and absent (> −25 % & < 10 %) drug effects, in line with the classification used in Nguyen et al. (2017) and Abi-Gerges et al. (2020). (c) Quantitative comparison of experimental rat drug effects, predicted human drug effects, and experimental human drug effects on the percentage sarcomere shortening. Experimental results are presented as means ± standard error of means. Details and sources of the experimental data can be found in Table S7. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)