Arrhythmic Risk Assessment of Hypokalaemia Using Human Pluripotent Stem Cell-Derived Cardiac Anisotropic Sheets

Introduction: Hypokalaemia, defined as an extracellular concentration of K+ below 3.5 mM, can cause cardiac arrhythmias by triggered or re-entrant mechanisms. Whilst these effects have been reported in animal and human stem cell-based models, to date there has been no investigation in more complex structures such as the human ventricular cardiac anisotropic sheet (hvCAS). Here, we investigated arrhythmogenicity, electrophysiological, and calcium transient (CaT) changes induced by hypokalaemia using this bioengineered platform. Methods: An optical mapping technique was applied on hvCAS derived from human pluripotent stem cells to visualize electrophysiological and CaT changes under normokalaemic (5 mM KCl) and hypokalaemic (3 mM KCl) conditions. Results: Hypokalaemia significantly increased the proportion of preparations showing spontaneous arrhythmias from 0/14 to 7/14 (Fisher’s exact test, p = 0.003). Hypokalaemia reduced longitudinal conduction velocity (CV) from 7.81 to 7.18 cm⋅s−1 (n = 9, 7; p = 0.036), transverse CV from 5.72 to 4.69 cm⋅s−1 (n = 12, 11; p = 0.030), prolonged action potential at 90% repolarization (APD90) from 83.46 to 97.45 ms (n = 13, 15; p < 0.001), increased action potential amplitude from 0.888 to 1.195 ΔF (n = 12, 14; p < 0.001) and CaT amplitude from 0.76 to 1.37 ΔF (n = 12, 13; p < 0.001), and shortened effective refractory periods from 242 to 165 ms (n = 12, 13; p < 0.001). Conclusion: Hypokalaemia exerts pro-arrhythmic effects on hvCAS, which are associated with alterations in CV, repolarization, refractoriness, and calcium handling. These preparations provide a useful platform for investigating electrophysiological substrates and for conducting arrhythmia screening.


INTRODUCTION
Hypokalaemia is defined as an extracellular concentration of K + below the normal range of 3.5-5.3 mM (Macdonald and Struthers, 2004;Unwin et al., 2011). It is a common biochemical abnormality in cardiac patients that may represent a side effect of diuretic therapy or result from endogenous activation of the renin-angiotensin system and high adrenergic tone. Transient hypokalaemia is also induced by intense exercise during recovery (Atanasovska et al., 2018). It is a risk factor for ventricular arrhythmias that can predispose patients to sudden cardiac death (Osadchii, 2010;Chen et al., 2021). The mortality rate of hospitalized patients with hypokalaemia is tenfold higher than in the generalized hospital population, illustrating its potentially life-threatening consequences (Paltiel et al., 2001). Low levels of K + are replaced with supplements through oral or intravenous routes.
Hypokalaemia can induce several electrophysiological abnormalities, such as delayed ventricular repolarization, which in turn predisposes patients to triggered activity through the development of early afterdepolarizations (EADs) and re-entrant arrhythmias by increasing the dispersion of repolarization (Schulman and Narins, 1990;Macdonald and Struthers, 2004;Osadchii, 2010). Its pro-arrhythmic effects and mechanisms have been demonstrated in previous studies in animal and human stem cell-based models (Tse et al., 2016b;Kuusela et al., 2017). However, to date, there have been no investigations into more complex structures such as the human ventricular cardiac anisotropic sheet (hvCAS). hvCAS is a micro-patterned substrate that is specifically designed to reproduce the anisotropy of native human ventricles and allows for direct visualization of arrhythmic spiral re-entry (Chen et al., 2011;Luna et al., 2011;Wang et al., 2013;Shum et al., 2017;Wong et al., 2019). In this study, we investigated the electrophysiological and calcium transient (CaT) changes induced by hypokalaemia using this bioengineered platform (Wang et al., 2013;Shum et al., 2017). An optical mapping technique was applied on hvCAS derived from human pluripotent stem cells to investigate electrophysiological and CaT changes on normokalaemic (5 mM KCl) and hypokalaemic (3 mM KCl) preparations.

MATERIALS AND METHODS
This study has been approved by the University of Hong Kong/ Hospital Authority Hong Kong West Cluster Institutional Review Board. Informed consent was given by the subject.

Reprogramming and Cell Culture
A peripheral blood sample was obtained from a normal healthy adult person. CD34 + cells were isolated from peripheral blood mononuclear cells (>95% purity) by using human Miltenyi Biotec). They were expanded for 3 days in StemSpan ™ H3000 (Stemcell Technologies) with the cytokine cocktail CC100 (Stemcell Technologies).
Cells were then seeded to Geltrex (A1413302; Gibco)-coated plates in E8 medium (Life Technologies). After a few days, morphological changes were observed and human induced pluripotent stem (iPS) cell colonies typically started to appear on day 7. Tra-1-60 enrichment was performed on day 13 after nucleofection.

Human iPS Cells Validation
The human iPS cell colonies were treated with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 15 min for fixation and then permeabilized by treating with 0.1% Triton X-100 for 15 min. They were then washed with PBS three times. Staining with anti-OCT4, SSEA-4, and Tra-1-81 for 2 h at room temperature were performed on the fixed samples. They were then treated with flurochrome-conjugated goat antirabbit or anti-mouse secondary antibodies for 1 h. For the pluripotency test, human iPS cells were differentiated to form Embroid Bodies (EB). The EB medium used was DMEM/F12 (Life Technologies) with 20% Knockout Serum Replacement (Life Technologies), 2 mM nonessential amino acids, 2 mM L-glutamine, and 0.1 mM β-mercaptoethanol. EBs were stained at day 15 for three germ layer markers (ie., Tuji, SMA, and AFP). Karyotyping was performed as described by Bates (Bates, 2011).
Flow cytometry analysis was performed on differentiated cells collected at D16-18 to quantify cTnT + cells resembling cardiomyocyte yield.

Fabrication of hvCAS
hvCAS substrates were fabricated according to a published protocol (Shum et al., 2017). The mold was obtained from Ming Wai Lau Centre for Reparative Medicine, Karolinska Institutet of groove dimensions 15 µM (R) × 5 µM (D) × 5 µM (W). Each substrate was 15 mm in diameter. Polystyrene shrink films (Clear Shrink Dinks) were hot-embossed at 180°C for 30 min. They were then treated with ultraviolet ozone (UVO) for 8 min (Jetlight UVO). Finally, they were submerged in 70% ethanol and kept under UV light for at least 15 min for sterilization. The substrates were coated with Matrigel and suctioning was performed for at least 15 min to remove air bubbles. The substrates were kept in a four-well plate (6900A07; Thermo Scientific Nunc) using vaseline.
hvCASs were washed once with PBS and changed to normal Tyrode's solution (for normokalaemia hvCAS) and hypokalaemic Tyrode's solution (for hypokalaemia hvCAS) and maintained at 35-37°C using a culture dish incubator (Warners Instruments) for optical mapping. High-resolution optical mapping was performed using a MiCAM05 (SciMedia, CA, United States) for Action Potential (AP) and CaT parameters. The field of view was 17.6 mm by 17.6 mm obtained by using 1× objective and a 1× condensing lens. The input light source was a highpowered LED illumination system (LEX2, SciMedia, CA, United States) filtered by a 515 ± 35 nm band-pass excitation filter, and the light was split into two using a splitter. One was passed through a >700 nm long-pass emission filter for RH237 (voltage sensitive dye) and the other through a 590 nm band-pass emission filter for Rhod2-AM (calcium sensitive dye). Data acquisition was done at a sampling rate of 5 ms/frame and BVAna software (SciMedia) was used for data analysis followed by clampfit (Axon Instruments). A programmable electrical stimulator (Master9; AMPI, Israel) was used for pacing (10 V, 10 ms pulse duration) through a unipolar pointstimulation electrode (Harvard Apparatus, MA, United States). The point-stimulation electrode was positioned perpendicular to the hvCAS surface at its center. Optical mapping was performed once per sample and unpaired.
Pacing Protocol for hvCAS 1) Steady state pacing was eight trains from 0.5 to 3 Hz (i.e., 30-180 bpm) at an increment of 0.5 Hz unless there was a loss of 1:1 capture. If they captured all the frequencies without arrhythmia, they were subjected to Programmed Electrical Stimulation (PES). 2) PES used eight trains of 1.5 Hz (S1) followed by a premature trigger (S2), where S1-S2 intervals were gradually shortened to find the Effective Refractory Period (ERP). The starting S1-S2 interval was 550 ms, which was then consecutively shortened in steps by 50 ms until capture had failed. S1-S2 interval was then increased by 20 ms, and then by 2 ms decrements until the ERP was achieved, which is the maximum S1-S2 interval that failed to lead to AP propagation ( Figure 5C).

hvCAS Subsets
Acute hypokalaemia: Optical mapping was performed on hvCAS for simultaneously measuring AP and CaT in normokalaemic and hypokalaemic conditions. All the hvCAS were maintained in the normokalaemic medium until optical mapping day (D7-D11 post hvCAS fabrication). On optical mapping day, normal Tyrode's solution was used for one of the subsets (normokalaemia) and hypokalaemic Tyrode's solution was used for another subset (hypokalaemia). The hvCAS were in respective Tyrode's solution from dye loading to data acquisition (30 min to <60 min). The unpaired test was chosen as repeated measurement, which led to a significant decrease in the signal to noise ratio.
Chronic hypokalaemia: The fabricated hvCAS were divided into two subsets (Subset I: control, Subset II: experimental). Both groups were kept at normokalaemic condition until D7 post hvCAS fabrication. From D7-D10, optical mapping was performed for AP and CaT in which subset I was in normakalaemic condition throughout the experiment, whereas subset II was changed to hypokalaemic condition on D7-D10.

AP and CaT Parameters
The following parameters were obtained from the experimental records: 1) Activation Latency or time-to-peak: The time interval from the start of the stimulus to the peak of the AP or CaT; 2) Amplitude: The height of the peak; 3) AP Duration (APD x ) and CaT Duration (CaTD x ): The time interval from the peak of the AP or CaT at x 30, 50, 70, and 90% repolarization or decay respectively; 4) ERP: The maximum S1-S2 interval that failed to elicit an AP; 5) Upstroke Velocity: The velocity at which the fluorescence reaches the peak from baseline; 6) Conduction Velocity (CV): The velocity at which AP propagates between two points; 7) Anisotropic Ratio (AR): The ratio of longitudinal CV to transverse CV.

Statistical Analysis
Prism (  Amplitude of AP and CaT increased in hypokalaemic condition (asterisks, p < 0.001) and also for the APD and CaTD at 30, 50, 70 and 90% decay (asterisks, p < 0.001 for APD and p < 0.005, p < 0.05 for CaTD).
Frontiers in Cell and Developmental Biology | www.frontiersin.org December 2021 | Volume 9 | Article 681665 statistically significant. Data analysis was performed for 1.5 Hz electrical stimulation recordings.

Acute Hypokalaemia Leads to Higher AP and CaT Amplitudes, and Longer APD and CaTD
Arrhythmogenicity and its relationship to AP and CaT activation and recovery properties were examined under normokalaemic and hypokalaemic conditions in hvCAS using optical mapping. During regular steady-state pacing, reproducible AP waveforms could be observed under normokalaemic conditions. By contrast, acute hypokalaemia led to the generation of premature APs that likely reflected triggered activity. Moreover, an S1S2 protocol, which delivered a successively premature S2 extra-stimulus after a regular train of S1 stimuli, was used to provoke arrhythmogenesis. With this protocol, an increase in the proportion of preparations exhibiting spiral wave re-entry was observed during hypokalaemia (hypokalaemia 7/14, 50%; normokalaemia 0/14, 0%; asterisks, p 0.003 (Fisher's exact test); Figure 2).
The AP and CaT values are summarized in Tables 1, 2 respectively.
Together all of the findings associate increased re-entrant arrhythmias observed in the form of spiral waves with depolarization, repolarization, and calcium handling abnormalities in the hypokalaemia model using our stem cellderived bioengineered platform.

DISCUSSION
The main findings of this study are that hypokalaemia can induce arrhythmogenesis, which is associated with depolarization, repolarization, and calcium handling abnormalities in stem cell-derived bioengineered platforms. These confirm that (n 12, 13) Conduction Velocity decreased in hypokalaemic condition (asterisk, p < 0.05 for both LCV and TCV). ERP of hvCAS paced at 1.5 Hz (8 trains) followed by an extra premature stimulus which was gradually shortened to find ERP was decreased in hypokalaemic condition (asterisk, p < 0.001). AR is between 1.4 and 1.8, resembling more closer to the native human ventricle (AR 3) in normokalaemia as well as in hypokalaemia.
Frontiers in Cell and Developmental Biology | www.frontiersin.org December 2021 | Volume 9 | Article 681665 8 hvCAS preparations can be used as a model system for studying human arrhythmic syndromes, providing opportunities for efficient drug screening programs for pro-arrhythmic effects.
Low extracellular K + levels lead to a steeper transmembrane voltage gradient and membrane hyperpolarization (Tse et al., 2021). This low level also reduces the repolarization reserve by suppressing the rapid delayed rectifier (I Kr ) channel, producing APD prolongation (Osadchii, 2010;Pezhouman et al., 2015). APD prolongation can in turn lead to the re-activation of L-type Ca 2+ channels, causing EAD phenomena and triggering activity (Killeen et al., 2007). Moreover, it can predispose patients to reentry through reduced CV, reduced tissue refractoriness, increased transmural dispersion of repolarization, and increased steepness of electrical restitution that can generate electrical alternans (Surawicz, 1985;Sabir et al., 2008;Tse et al., 2016a;Tse et al., 2016b). Reduction in CV is more likely due to APD prolongation. The membrane hyperpolarization also causes an increase in the availability of fast Na + channels that leads to an increase in upstroke velocity (Osadchii, 2010). Despite the higher upstroke velocity, time-to-peak is prolonged, which is possibly due to higher amplitude caused by hyperpolarization. In our study, greater AP amplitudes, reduced CV, and longer APDs were observed in hvCAS. Moreover, both longitudinal and transverse conduction velocities were decreased to similar extents, leading to no significant alteration in the AR. This took a mean value between 1.4 and 1.8 in both under normokalaemic and hypokalaemic conditions, which is fairly close to the AR value for the human ventricle of 3 (Camelliti et al., 2011).
The amplitude and time-to-peak of CaTs were higher under hypokalaemic conditions. This can be explained by the suppression of sodium-calcium (Na + -Ca 2+ ) exchanger and sodium-potassium pump during hypokalaemia, which leads to the accumulation of Ca 2+ and Na + in the cytoplasm (Pezhouman et al., 2015). Together with increased Ca 2+ influx owing to L-type Ca 2+ channel reactivation, this can lead to secondary depolarization events if the activation threshold is reached, an extra premature AP. The suppression of Na + -Ca 2+ exchanger leads to the Ca 2+ overload in the cytoplasm which in turn prolongs the time taken for Ca 2+ to be removed from the cytosol via the following routes: Na + -Ca 2+ exchanger to the extracellular space, sarco/endoplasmic reticulum Ca 2+ -ATPase, into the sarcoplasmic reticulum.
The different electrophysiological mechanisms at play induced by hypokalaemia are summarized in Figure 6.

LIMITATIONS OF THE STUDY
Although the cardiomyocytes (CMs) that we produce through directed cardiac differentiation are mostly ventricular CMs, there are still a small proportion of other subtypes of CMs (Weng et al., 2014). Additionally, stem cell-derived CMs show a relatively immature phenotype as reported in previous studies (Karakikes et al., 2015). By performing optical mapping on the hvCAS, relative values of voltage change and CaT can be obtained from the fluorescence intensity of the fast-responding voltage sensitive and Ca 2+ dyes respectively. However, the fluorescence signal is not calibrated and absolute resting membrane potential cannot be determined. Furthermore, hvCAS contains a stack of cells at each pixel and the data presented here are from those FIGURE 6 | Diagram showing the changes caused by hypokalaemia in human pluripotent stem cell-derived human ventricular cardiac anisotropic sheet (hvCAS) during acute hypokalaemic (hypoK) experiment where the treatment was for 15 min before data acquisition. Hypokalaemia refers to 3 mM KCl in Tyrode solution on the samples. Those changes make cardiomyocytes prone to arrhythmia in hypokalaemic conditions. The boxes highlighted in grey are adapted from literature research. Those changes leading to arrhythmia can be visualized on the hvCAS.
Frontiers in Cell and Developmental Biology | www.frontiersin.org December 2021 | Volume 9 | Article 681665 multiple cells. This makes the AP upstroke phase appear slow compared to single-cell actual AP measured by patch-clamp or microelectrode array in other studies. Nevertheless, we can see the differences between normokalaemia and hypokalaemia hvCASs, and most importantly, arrhythmia can be visualized as spiral waves by using optical mapping, which is not possible by other techniques.

CONCLUSION
The pro-arrhythmic effects of hypokalaemia are associated with depolarization, repolarization, and calcium handling abnormalities in a human multicellular monolayer model using a stem cell-derived bioengineered platform. This platform can be used as a model system for cardiac arrhythmic risk assessment where arrhythmia can be visualized as spiral waves.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Materials, further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by The University of Hong Kong/Hospital Authority Hong Kong West Cluster Institutional Review Board. The patients/participants provided their written informed consent to participate in this study.