*3.4. Computer Simulations Corroborate the E*ff*ect of the Non-Cardiac BKCa to Human AP*

To evaluate the impact of BKCa on repolarization quantitatively, we integrated a BKCa model into a human ventricular CM model. Simulation results in Figure 5 demonstrated that a reasonably sized BKCa current can cause a deep notch in the AP. The dynamics of the notch (Figure 5A) match the kinetics of the simulated BKCa current (Figure 5B). Although inclusion of BKCa did not cause oscillations of the membrane voltage similar to in vitro observations, the deep notch enhanced drastically L-type Ca2<sup>+</sup> current (ICa,L, Figure 5C).

**Figure 5.** Impact of including BKCa in a mathematical model of human ventricular cardiomyocytes. BKCa current (**B**) causes a deep notch in the action potential (**A**), and a substantial increase in the amplitude of L-type Ca2<sup>+</sup> current (**C**).

#### **4. Discussion**

The main findings of this study were:

(1) HiPSC-CMs could express ion channels otherwise not found in adult cardiomyocytes.


BKCa channels are widely expressed in various cell types, including electrically excitable and non-excitable cells [31]. Due to the Ca2<sup>+</sup>-sensitivity, they provide relevant negative feedback mechanism in the regulation of intracellular Ca2<sup>+</sup> elevation and membrane potential [32]. Almost every cell type expresses BKCa in the inner mitochondrial membrane (BKmito) [33]. In CMs, pharmacological opening of BKmito reduces ischemia reperfusion injury [34]. Sarcolemmal expression of BKCa is typically found in vascular smooth muscle cells, regulating myogenic tone and thereby blood flow. In CMs from rodents and cardiac tissue from humans, sarcolemmal expression of BKCa is very low or almost non-existent [35–37]. Here we confirm low expression levels of BKCa in human LV myocardium. Among CMs from other species, IBTX sensitive currents were found only in cultured embryonal chicken CMs [38]. Contribution to repolarization in this species is unclear. In rat ventricular myocardium, IBTX does not affect AP shape [39]. Here, we demonstrated for the first time that human ventricular APs were insensitive to IBTX. In consequence, the detection of IBK,Ca in C25-hiPSC-CMs during patch clamp recordings was unexpected and surprising, since BKCa is not known to be expressed in the sarcolemma. The effectivity of the selective blocker of IBK,Ca [26,40], IBTX, which only binds from the external side of the channel [41], confirmed the hypothesis that a non-cardiac channel is present and active at transmembrane level in the hiPSC-CMs. These findings were also confirmed by expression analysis in which the existence of the notch/oscillation in the AP correlated with the high expression levels of *KCNMA1*, encoding for the alpha-subunit (KCa1.1) of the BKCa. In contrast, *KCNMA1* expression was low in human LV tissue and correspondingly IBTX did not show any effect on APs from human LV, which to our knowledge has not been described before.

Recently, overexpression of non-cardiac BKCa in CMs was proposed as a treatment for LQTS [13], since BKCa is a hyperpolarizing channel, which might shorten human AP. Support for this idea came from a study describing the electrophysiological function of BKCa in HL-1 cells by viral overexpression, a cell line derived from a murine atrial tumor. BKCa overexpression reduced the very short AP of this murine model by 50% (APD90 from 30 to 14 ms) and was proposed as a potential genetic therapy to reduce AP duration (APD) of the LQT syndrome [13]. In contrast to the experiments in HL-1 cells, we observed that in hiPSC-CMs, the presence of BKCa induces oscillations in the early plateau phase and no speed-up in the final repolarization [13].

The pronounced initial repolarization could be confirmed by the in-silico integration of the IBK,Ca in modeling ventricular myocyte AP (Figure 5). However, the inclusion of BKCa could not resemble oscillations of the membrane voltage. Nevertheless, we would expect that oscillation might be induced by an alternating feedback mechanism [32] of the IBK,Ca and the L-type Ca2<sup>+</sup> current (ICa,L), since the deep notch drastically enhanced L-type Ca2<sup>+</sup> current (Figure 5C). The afterdepolarizations following the initial repolarization might be also due to activation of ICa,T [11] or the sodium calcium exchanger, however, the exact mechanism remains unclear. More accurate and detailed mathematical modeling would require in vitro data on spatial distribution and localization of the BKCa and Ca2<sup>+</sup> channels, which is beyond the scope of this study.

Afterdepolarizations might complicate the evaluation of how the BKCa affects APD. In vitro, there was a slightly longer APD90 when BKCa was present, which could be confirmed in silico. However, the inhibition of IBK,Ca by iberiotoxin in AP with notch/oscillations did not significantly alter APD90, averaged values tended even to longer APD90. The apparent differential contribution of BKCa to the APD revealed at baseline level or by pharmacological intervention might be due to remodeling of other ion channels downstream to BK expression or due to potential off-target effects of iberiotoxin. Taken together, our results do not support the idea that BKCa overexpression can cure LQTS in humans, since BKCa might lead to afterdepolarization and arrhythmia.

Previously, it was shown that the IBK,Ca current contribute to outward currents in the murine sinoatrial node and the selective blocker paxilline decreased beating rate by more than 50% [42]. However, the exact role of IBK,Ca in pacemaking is widely unclear, since substantial decrease in diastolic depolarization by paxilline does not fit to very small contribution of BKCa to total potassium outward currents activated at positive membrane potentials. Nevertheless, spontaneous beating is a peculiarity of hiPSC-CMs and the autonomic activity of hiPSC-CMs is not fully understood. Therefore, it seems reasonable to speculate that BKCa may be involved in pacemaking of hiPSC-CMs. However, we would not expect a large impact of IBK,Ca to pacemaking in EHT, since beating rate in EHT with and without expressing BKCa did not differ.

The reason for the unexpected expression of BKCa in individual differentiation batches of a single hiPSC-CM cell line (C25) is very difficult to evaluate retrospectively. Cell line C25 did not show chromosomal anomalies at passage number 40 and 92 (Supplementary Figure S3). Since 6 out of 7 individual differentiation batches were done from lower passage numbers than 92, it seems that BK expression was not the consequence of karyotype abnormalities. As notch/oscillation were observed also in preparations with very high differentiation efficiencies (90% and 93%) and hiPSC-CM fraction in EHT was even enriched in comparison to 2D culture (Supplementary Materials Figure S2B), we are confident that BK expression is not due to an extraordinary high fraction of non-cardiac cells within the EHT. Strong batch effects, as either all or none of the EHTs from one preparation depicted signs of BKCa expression, argue for an upregulation of BKCa due to events occurring during stem cell culture and cardiac differentiation. There are two factors that might have raised the likelihood for spontaneous mutations in the stem-cell culture. The C25 cell line was reprogrammed by lentivirus and it was passaged to very high number (up to 107 passages). To avoid these factors, we changed to Sendai virus and restricted passage number for future experiments. Since *KCNMA1* related genes did not show major alterations in contrast to *KCNMA1* itself, we would account a genetic alteration more likely than an upregulation due to regulatory pathways. In addition, various reports show that iPS-cells frequently acquire genetic alterations in cell culture. Kilpinen et al. [43] showed that chromosome 10, harboring BKCa/*KCNMA1*, was among the most susceptible loci to copy number alterations. In addition, in a study searching for variants that provide mutated cells with a growth advantage in culture, *KCNMA1* candidate mosaic variants were identified in two independent hES cell lines [44]. Thus, a genetic alteration leading to a reoccurring overgrowth of BK misexpressing hiPSC is the most likely explanation for our finding. Regulatory expression profiling might reveal this change of BKCa activity in advance.

#### **5. Conclusions**

Our results clearly demonstrated that hiPSC-CMs could possess even non-cardiac ion channels affecting AP waveform causing afterdepolarizations and oscillations in the AP. This might serve as an example that iPS cell culture could lead to genetic alterations with functional consequences. Therefore, we felt that cell culture and differentiation protocols should be standardized as much as possible and that expression of non-cardiac sarcolemmal ion channels should be considered before hiPSC-CMs are used for pharmacological studies. Screening for the expression of the *KCNMA1* gene might be one potential quality parameter.
