Simultaneous measurement of excitation-contraction coupling parameters identifies mechanisms underlying contractile responses of hiPSC-derived cardiomyocytes

Abstract

Cardiomyocytes from human induced pluripotent stem cells (hiPSC-CMs) are increasingly recognized as valuable for determining the effects of drugs on ion channels but they do not always accurately predict contractile responses of the human heart. This is in part attributable to their immaturity but the sensitivity of measurement tools may also be limiting. Measuring action potential, calcium flux or contraction individually misses critical information that is captured when interrogating the complete excitation-contraction coupling cascade simultaneously. Here, we develop an hypothesis-based statistical algorithm that identifies mechanisms of action. We design and build a high-speed optical system to measure action potential, cytosolic calcium and contraction simultaneously using fluorescent sensors. These measurements are automatically processed, quantified and then assessed by the algorithm. Multiplexing these three critical physical features of hiPSC-CMs allows identification of all major drug classes affecting contractility with detection sensitivities higher than individual measurement of action potential, cytosolic calcium or contraction.

Fig. 1

Schematic of critical mechanisms affecting cardiomyocyte contraction targeted by drugs. Most therapeutics that target the heart or have off-target effects on the heart fall within the mechanisms shown. 1: Since no beta-adrenergic agonists are present in in vitro cultures with defined medium, this category is not relevant in this study. 2: This group is small and the mechanisms are not specific. For example, while blebbistatin reduces contractility, it also affects cardiomyocyte electrophysiology. Therefore, this category is not included in this study. Asterisk: These compounds exhibit both PDE inhibiting and myofilament Ca²⁺ sensitizing properties, with the dominant mechanism typically depending on the concentration

Fig. 2

Experimental setup and example output. a Schematic layout of the experimental setup; b output example of contraction amplitude in a concentration-response plot in which individual points show averaged values of a well and the red dotted line indicates the known therapeutic concentration; c corresponding probability density function (grey) and PS_↑ and PS_↓ (red lines) describing the three spaces for the three Probability Scores

Fig. 3

Triple Transient Measurement system for determining AP, Ca and Co simultaneously. a Hardware setup: the hiPSC-CM model of interest (2D or 3D) is incubated with the voltage sensitive dye (ANNINE-6plus), calcium sensitive dye (Rhod 3) and cell membrane label (CellMask Deep Red) used for measuring contraction. The hiPSC-CMs are excited using three rapidly alternating LEDs which are synchronized with a high-speed camera. The camera receives the emitted signal from the hiPSC-CMs after photon amplification by the image intensifier. The data are then fed into the data processing pipeline. b The video frames are split into three stacks that contain the information regarding the AP (blue), Ca (yellow) and Co (red). The information of each stack is extracted by either plotting the intensity changes and/or applying MUSCLEMOTION. c The data are then “cleaned” to remove the motion artefacts, invert the signal when necessary, calculate ∆F/F, apply a minimal filter and/or average the events within one recording. d The data are quantified by summarizing the kinetic parameters of interest determined in Table 1. e Finally, the data are plotted and meanwhile fed into the hypothesis algorithm

Fig. 4

Averaged representative transients of hiPSC-CMs in response to vehicle (DMSO) and drugs. All measurements were normalized to their individual baseline measurements. Cells were paced at 1.2 Hz. Simultaneous measurements show action potential (blue), cytosolic calcium (orange) and contraction (red) of Pluricytes under baseline conditions and in response to 0.3 µM epinephrine (a), 100 µM pimobendan (b), 0.1 µM verapamil (c), 30 µM captopril (d), 3 µM forskolin (e), 30 µM aspirin (f), 3 µM doxorubicin (g), 0.3 µM omecamtiv mecarbil (h), 0.3 µM levosimendan (i), 3 µM atenolol (j), 10 µM sunitinib (k) and 1 µM ouabain (l)