Proarrhythmic toxicity of low dose bisphenol A and its analogs in human iPSC-derived cardiomyocytes and human cardiac organoids through delay of cardiac repolarization.

Acute exposures to 1 nM BPA and its analogs, including BPS, BPF, and BPAF, did not result in any cytotoxicity in hiPSC-CMs as measured by the Cell Counting Kit-8 assay (data not shown). Cellular electrical properties of hiPSC-CMs were examined using perforated patch clamp recording. A mixed population of hiPSC-CMs was observed ( Figure 1A ), exhibiting shorter and triangular APs (atrial-like), APs with prominent phase-4 spontaneous depolarization (nodal-like), and APs with prominent phase-2 plateau (ventricular-like). The ratio of atrial-, nodal-, and ventricular-like cells was 26:22:52 ( Figure 1B ). These electrical properties were similar to previous characterization of hiPSC-CMs and to human cardiac myocytes ( Ma et al., 2011 ; Schmid et al., 2021 ). Acute exposure to 1 nM BPA prolonged AP, particularly phase-2 plateau, of atrial-and ventricular-like hiPSC-CMs, without otherwise changing the AP waveform ( Figure 1C ). APD prolongation occurred within minutes of exposure ( Figure 1D ). On average, BPA prolonged APD 50 and APD 90 by ~11.3% and ~10.9%, respectively ( Figure 1E ). The dose-dependent effect of BPA on APD (measured as APD 90 ) of hiPSC-CMs was nonmonotonic with an inverted U shape ( Figure 1F ). The peak effect of BPA occurred at 10 −9 to 10 −8 M, and the effect diminished at 10 −6 M ( Figure 1F ). To determine the ionic mechanism of BPA-induced APD prolongation in ventricular-like hiPSC-CMs, we assessed the effects of BPA on I Kr and I Ca-L , two ionic currents with major roles in determining human cardiac phase-2 duration ( Roden et al., 2002 ). I Kr was recorded as the tail current at −40 mV following prolonged depolarization steps. Acute exposure to 1 nM BPA significantly reduced the peak amplitude of I Kr tail current in hiPSC-CMs ( Figure 2A , B ). On average, 1 nM BPA inhibited I Kr by 41.8% (P < 0.05; Figure 2C ). Similar to our previously reported results in rat ventricular myocytes ( Liang et al., 2014 ), BPA also produced a moderate suppression of I Ca-L in hiPSC-CMs ( Figure 2D , E ). The peak I Ca-L at +10 mV was reduced by ~16.5% while the current-voltage relationship was not altered ( Figure 2E , F ). 1 nM BPA rapidly increased the automaticity of nodal-like hiPSC-CMs, increasing the average spontaneous firing frequency from 83.3 to 106.9 beats/min (P < 0.05; Figure 3A , B ). The I f current, generated by the HCN pacemaker channel, underlies spontaneous automaticity in nodal cardiomyocytes ( DiFrancesco, 2010 ). A robust I f current was activated upon hyperpolarization in hiPSC-CMs, and the current was markedly stimulated upon acute exposure to 1 nM BPA ( Figure 3C , D ). The average peak I f at −140 mV was increased by 36.2 %, from −5.8 to −7.9 pA/pF ( Figure 3D , E ). Excessive prolongation of cardiac APD can result in ectopic excitation of cardiac myocytes and arrhythmias ( Fozzard, 1992 ). We next examined the impact of BPA on aberrant spontaneous excitation in hiPSC-CMs. In particular, we tested the impact of BPA in myocytes that had a simulated LQT syndrome phenotype and were predisposed to APD-prolongation associated arrhythmias. 10 nM E4031, a selective hERG/I Kr channel blocker ( Zhou et al., 1998 ), was used to partially block the hERG channel, thus pharmacologically mimicking the LQT syndrome type 2 phenotype ( January et al., 2000 ). Excitations of hiPSC-CMs were measured using two independent methods, as either electrical excitation ( Figure 4A ) or Ca 2+ transients ( Figure 4B ). No aberrant excitations were observed in hiPSC-CMs under baseline control condition or in the presence of 1 nM BPA ( Figure 4A – C ). 10 nM E4031 resulted in some aberrant excitation events, at an average frequency of 11.8 events/5 min ( Figure 4C ). Also notice the prolongation of APD in the presence of E4031 ( Figure 4A , B ). Remarkably, in the presence of E4031, exposure to BPA resulted in a pronounced increase in aberrant excitations (arrows; Figure 4A , B ). These ectopic excitations, which superimposed on normal excitations, are known as early afterdepolarizations (or EADs), and are the cellular events that underpin APD prolongation-associated cardiac arrhythmias ( Pogwizd and Bers, 2004 ). The frequency of EADs was increased from zero in control and 11.8 events/5 min in E4031 alone, to 92.7 events/5 min in BPA plus E4031 (P < 0.01 vs E4031 alone; Figure 4C ). The impact of BPA on arrhythmias was examined in hiPSC-CM 2D monolayers, including those with mimicked LQT syndrome phenotype produced with E4031 ( Figure 5 ). Similar to the single myocyte level findings, BPA alone did not result in any arrhythmic events, while 10 nM E4031 resulted in some aberrant excitations (i.e., EADs) as well as tachycardia-like events ( Figure 5A – D ). Strikingly, in monolayers treated with E4031, exposure to BPA resulted in a high frequency of arrhythmic events. Frequency of EADs ( Figure 5A , arrows) was increased from zero in control and 32.8 events/5 min in E4031 to 135.8 events/5 min in E4031 + BPA (P < 0.01; Figure 5B ), incidence of EADs was increased from 40% in E4031 alone to 100% (P < 0.05; Figure 5C ), and incidence of tachycardia-like events ( Figure 5A , horizonal lines) was increased from 10% in E4031 alone to 70% (P < 0.05; Figure 5D ). We next compared the effects of BPA and its analogs on the electrical properties of human cardiac organoids. Organoids were fabricated using hiPSC-CMs, human cardiac fibroblasts and human endothelial cells, and stained positive for cardiac myocyte markers troponin-T and sarcomeric α-actinin ( Figure 6A ). Acute exposure to 1 nM BPA, BPS, BPF and BPAF prolonged APD, with BPAF having the largest effect ( Figure 6B , C ). APD 90 was 469.2 ± 13.1 ms, 495.8 ± 7.7 ms, 489.8 ± 12.3 ms, and 508.2 ± 24.1 ms for control, BPA, BPS, and BPAF, respectively ( Figure 6C ). The APD prolongation effects of the bisphenols were moderate, and no EADs were observed when organoids were exposed to any of the chemical alone. By contrast, under E4031-produced LQT2 phenotype, BPA and its analogs triggered the development of numerous EADs ( Figure 6D , E ). EAD frequency under BPA, BPS, and BPF were similar, and was highest under BPAF. Frequency of EADs (events/10 min) was increased from zero in E4031 alone, 15.8 in BPA to 36.5 in BPAF ( Figure 6E ).

All the reagents and solvents used were of the highest purity available. All solutions used were prepared using ultrapure water (18 MΩ; < 6 ppb total oxidizable organics) produced by a Milli-Q IQ7000 water purification system (MilliporeSigma, MA, USA) with an EDS-Pak ® Polisher, Lot F9KA74614, (MilliporeSigma, MA, USA). As per manufacturer’s Certificate of Quality, the setup was capable of delivering ultrapure water with BPA < 0.005 ppb. BPA (CAS 80-05-7; lot 111909; TCI America; ground by Battelle) was provided by the Division of the National Toxicology Program at the National Institute of Environmental Health Sciences. Bisphenol S (BPS, CAS 80-09-1), bisphenol F (BPF, CAS 620-92-8), and bisphenol AF (BPAF, CAS 1478-61-1) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Dimethyl sulfoxide (DMSO) (CAS 67-68-5), nifedipine (CAS 21829-25-4), and gramicidin (CAS 1405-97-6) were also from Sigma-Aldrich. E-4031 dihydrochloride (E4031) (CAS 113559-13-0) was obtained from Tocris Bioscience (Bristol, UK). Other chemicals and drugs were from Sigma-Aldrich unless otherwise stated. All stock solutions of the chemicals were constituted in DMSO. Working solutions for all treatments were prepared by further dilution where the final concentrations of DMSO were ≤ 0.1% as controls were treated with the same vehicle. The concentrations of BPA and BPA analogs tested in the study were decided based on typical human environmental exposure levels. Human exposure level to BPA, based on urine and serum concentrations, is commonly reported to be in the low-nanomolar range ( Calafat et al., 2005 ; vom Saal and Hughes, 2005 ; Vandenberg et al., 2007 ; Calafat et al., 2008 ). Human exposures to BPA analogs are also reported to be in the low-nanomolar range ( Gao et al., 2015 ; Chen et al., 2016 ). The goal of this study was to define the impact of environmentally-relevant concentration of BPA and its analogs on human cardiac myocytes. For these reasons, bisphenol chemicals at 1 nM were used in most of our analysis except the dose response experiment shown in Figure 1F . The hiPSC-CMs (iCell Cardiomyocytes 2 ), derived from cells from a healthy donor without arrhythmias-related genetic mutations, were obtained from Cellular Dynamics International Inc. (Madison, WI, USA). hiPSC-CMs were provided as a cryopreserved single cell suspension of ~ 99 % pure population consisting of a mixture of electrically active atrial, nodal and ventricular-like myocytes ( Ma et al., 2011 ). Cryopreserved hiPSC-CMs were thawed in iCell cardiomyocytes plating medium (Cellular Dynamics, Madison, WI, USA) according to the manufacturer’s instructions. Thawed cells were plated on 35 mm cell culture dishes precoated with 0.1% gelatin (Stem Cell Technologies, Canada) in a humidified incubator at 37°C for 2 h at a seeding density of 5 x 10 5 cells/dish. Starting at 4 hours post plating, hiPSC-CMs were maintained in iCell cardiomyocyte maintenance medium (Cellular Dynamics, Madison, WI, USA) at 37°C, 5 % CO 2 with the medium renewed every other day. On day 4, synchronously contracting hiPSC-CM monolayers were established. For electrophysiological experiments, the monolayers were further enzymatically disassociated with TrypLE ™ Express (1X) (Thermo Fisher Scientific, MA, USA). Single hiPSC-CMs were harvested and plated onto glass coverslips coated with fibronectin in PBS supplemented without Ca 2+ and Mg 2+ (5 μg/cm 2 , 2 h at 37 °C; Life Technologies) at a 1 x 10 5 per dish cell density and incubated for 24 h before use. Cells on days 5–11 were used for experiments. Cryopreserved hiPSC-CMs were thawed and plated on the glass-bottom of 35 mm MatTek dishes (MatTek Life Sciences, MA, USA) at a cell density of 1 x 10 5 per dish and cultured for 4 days in the iCell maintenance medium until a monolayered cluster was formed throughout the glass bottom. Monolayers on days 5-10 were used for experiments. 3D cardiac microtissue was fabricated using human cardiomyocytes, endothelial cells, and cardiac fibroblasts. Cryopreserved human primary cardiac fibroblasts and human iCell endothelial cells were obtained from Promocell (Heidelberg, Germany) and FUJIFILM Cellular Dynamics, Inc, respectively. Non-adhesive agarose hydrogel molds containing 35 microwells were prepared using commercial master-micro-molds from Microtissues, Inc (Providence, RI). HiPSC-CMs, cardiac fibroblasts, and iCell endothelial cells were mixed as a cell suspension at a ratio of 5:4:1 in a co-culture medium for a final concentration of ~1.5 × 10 6 cells/ml. The co-culture medium was prepared by mixing the endothelial cell culture medium (Promocell; Heidelberg, Germany) with the iCell cardiomyocyte maintenance medium at a ratio of 1:4. Agarose molds were placed individually in the wells of a 12-well plate and 75 μl of the cell suspension (~150,000 cells) was added to each mold. After the cells settled down into the features of the mold (10 min), additional co-culture medium (2.5ml/well) was added to immerse the molds. The molds were then kept uninterruptedly in a cell culture incubator at 37 °C and 5% CO 2 . After 2 days of organoid self-assembly, spheroid-like organoids were formed and began to contract spontaneously. The organoids were then dislodged carefully from the molds, harvested, and maintained in culture. Cardiac action potential (AP) and ionic currents were recorded from both quiescent or spontaneously contracting hiPSC-CMs using perforated and ruptured whole cell patch-clamp techniques. The rapid effects of BPA on AP and ionic currents in hiPSC-CMs were examined following BPA exposure at 1 nM for 1 to 4 min. For AP recordings in quiescent hiPSC-CMs, AP was evoked by brief stimulating current (4 ms) at a frequency of 1 Hz and recorded in current clamp mode until steady-state AP waveforms were reached. Spontaneous AP from beating hiPSC-CMs were recorded without electrical stimulation. Myocytes were held at −70 mV for L-type Ca 2+ current (I ca-L ) and at −50 mV for hyperpolarization-activated “funny” current (I f ) and rapid delayed rectifier K + current (I kr ) measurements. Measured currents were normalized to the membrane capacitance. The bath solution for AP, I f and I kr measurements contained (in mM): 148 NaCl, 5.4 KCl, 1.8 CaCl 2 , 1 MgCl 2 , 0.4 NaH 2 PO 4 , 5.5 glucose, 15 HEPES (pH 7.4 with NaOH). For measuring I kr 5 μM nifedipine was added in the bath solution to block I ca-L . The pipette solution for AP, I f and I kr measurements contained (in mM): 150 KCl, 5 EGTA, 5 MgATP, 10 HEPES, 5 NaCl, 2 CaCl 2 (pH 7.2 with KOH). For all AP recordings, gramicidin was added into the pipette solution at a concentration of 50 μg/ml. The extracellular solution for I ca-L recordings contained (in mM): 144 TEA-Cl, 5 CaCl 2 , 1 MgCl 2 , 10 HEPES, 10 glucose, and 2 4-aminopyridine (pH 7.4 with CsOH), whereas the pipette solution contained (in mM): 125 Cs-aspartate, 20 CsCl, 5 EGTA, 10 HEPES, 2 CaCl 2 , 5 NaCl and 5 Mg-ATP (pH 7.2 with CsOH). All patch-clamp recordings were performed using an Axopatch-200B amplifier. Data were acquired with an Axon Digidata 1322A data acquisition system (Axon instruments, CA, USA) and analyzed using the pCLAMP9 software (Axon instruments, CA, USA). All experiments were performed at 37°C with the exception of I ca-L recordings, which were performed at room temperature (24 °C). Ca 2+ transients were recorded from spontaneously beating dispersed single cells or either small monolayered clusters or individual organoids using the fluorescent Ca 2+ indicator dye Fluo-4 AM (Thermo Fisher Scientific, MA, USA). On the day of the experiment, single hiPSC-CMs or hiPSC-CM monolayers and organoids were incubated with the loading medium containing 3 μM Fluo-4 AM and 0.1 % Pluronic F-127 (Thermo Fisher Scientific, MA, USA) at 37 °C for 30 min and 60 min respectively, and followed by 20 min de-esterification in fresh medium at 37 °C. Before imaging, the dishes containing hiPSC-CM monolayers/organoids or glass coverslips with dispersed single hiPSC-CMs were placed into a chamber mounted onto a Leica Stellaris 8 confocal microscope (Leica Microsystems, Inc., Exton, PA) providing a 37 °C, 5% CO 2 and humidified atmosphere. The dye was excited at 488 nm and the fluorescence signals were measured at > 500 nm. Ca 2+ transients in single hiPSC-CMs were recorded with line-scan imaging at 1.53 ms intervals, with each line comprising 512 pixels spaced at 0.078 μm. Ca 2+ signals in hiPSC-CM monolayers were acquired using frame mode of scanning of a 512 × 20 pixel field at a rate of 85 frames/s. Ca 2+ transients in organoids were imaged using the resonant scanner mode with a frame size of 512 × 64 pixels at a rate of 125 frames/s. Image processing and data analysis were performed using Leica LAS X imaging and Clampfit 10.3 software. The fluorescence signals were continuously recorded and quantified in relation to baseline fluorescence (F/F 0 ). Treatments (i.e., bisphenol chemical, E4031, E4031 plus bisphenol chemical, or vehicle control, as indicated in the Results section and the figures) were added to the media after at least 5 min of stabilization under control, unexposed conditions. The AP in organoids were measured using FluoVolt ™ membrane potential kit (Thermo Fisher Scientific, MA, USA) according to the manufacturer’s instructions. Briefly, organoids were staining with a FluoVlot ™ working solution of 1: 100 dilution component B and 1:1000 dilution component A in media and incubated for 45 min at 37 °C, 5% CO 2 . Following two washes with fresh media and an additional 30 min incubation for de-esterification, organoids were imaged using the Leica Stellaris 8 confocal microscope with excitation wavelength of 522 nm. The fluorescence signal at > 535 nm was measured continuously for 5 min after a stabilization was visualized under control conditions. Image processing and data analysis were performed using the Leica LAS X imaging and Clampfit 10.3 software. For each organoid, the normalized AP trace profile was generated and the APD 90 was expressed as average APD 90 from three consecutive AP traces. Organoids were fixed in 4% paraformaldehyde (Thermo Fisher Scientific, USA) for 15 min, washed with phosphate-buffered saline (PBS), and incubated with blocking buffer (PBS containing 1% BSA and 0.3% triton X-100) for overnight at 4 °C. Afterwards, incubation with primary antibodies was performed in a staining buffer (PBS containing 1% BSA) for 2 days at 4°C. Primary antibodies were used at the following dilutions: anti- troponin T (1:200 dilution; Thermo Fisher, catalog# MA5-12960) and anti-sarcomeric alpha actinin (1:100 dilution; Thermo Fisher, catalog# MA1-22863). The organoids were then washed two times for 1h each with PBS and incubated with the secondary antibodies (Alexa Fluor ™ plus 488, 1:500 dilution; Thermo Fisher, catalog# A32723) in the staining buffer containing 0.5ug/ml DAPI (Thermo Fisher, catalog# 62248) in the dark room for 4 hours at room temperature. Following incubation, organoids were washed and mounted on coverslips using Prolong Gold Antifade Mountant (Thermo Fisher, catalog# P10144 ). The bright field and fluorescent images of organoids were acquired using a Leica Stellaris 8 confocal microscope (Leica Microsystems, Inc., Exton, PA). For 3D fluorescent imaging, z-stacking was used for imaging of individual organoid where z-stacks of 2-3 μm thickness with 0.5-1 μm step sizes were used. Image processing was performed using Imaris imaging software (Andor Technology, Belfast, UK). Data are presented as mean ± standard deviation (SD). Statistical significance was assessed using either a paired t -test or one-way analysis of variance (ANOVA) and a multiple comparison posttest. Difference in frequency of events (e.g., incidences of ectopic excitations and arrhythmic events) was assessed by a chi-square (χ 2 ) test. Data was analyzed with Clampfit 10.3 and Leica LAS X imaging software, and graphed with GraphPad Prism version 8.0. P < 0.05 was considered statistically significant.

The potential cardiovascular toxicity of BPA has been demonstrated in epidemiological studies ( Gao and Wang, 2014 ; Dias et al., 2022 ) and in experimental studies ( Yan et al., 2011 ; Gao et al., 2013 ; Zhang et al., 2020 ) mostly in rodents. However, little is known about the impact of low dose BPA on human cardiac electrophysiology. Using hiPSC-CM-based models, which have strong relevance to human heart health, we show that rapid exposure to environmentally-relevant low dose BPA delayed cardiac repolarization through I kr inhibition, a marker for arrhythmogenic risk ( Witchel, 2011 ). The pro-arrhythmic impact of BPA was particularly pronounced in hiPSC-derived myocytes and tissues under simulated LQT syndrome setting. In addition, BPA increased cardiac automaticity in hiPSC-CMs through stimulation of the pacemaker channel I f . The repolarization delay and pro-arrhythmic effects of BPA were shared by its analog chemicals in human cardiac organoids. Our findings suggest a potential contributing role of BPA and its analogs in arrhythmogenesis in human heart cells and tissues, particularly in hearts with preexisting pathophysiological conditions such as LQT syndrome. Cardiac AP in the hearts of large mammals is characterized by its long duration and distinctive “spike-and-dome” morphology ( Rosati et al., 2008 ). Excessive APD prolongation can result in ectopic excitations of cardiomyocytes (i,e,. EADs), which in turn can trigger malignant ventricular tachycardia and sudden cardiac death ( Yan et al., 2001 ; Lankipalli et al., 2005 ; Schwartz et al., 2012 ). Our results suggest that BPA exposure may be a risk factor for repolarization delay-associated aberrant excitation in human cardiomyocytes, particularly in LQT syndrome patients. The APD prolongation effect of BPA in hiPSC-CMs is mediated by hERG/I Kr inhibition, partially countered by inhibition of I Ca-L . Interestingly, the rapid effects of BPA on I Kr and I Ca-L are similar to those of natural endogenous estrogen, suggesting an estrogenic mechanism. Acute exposure to nM 17β-estradiol has been shown to inhibit I Kr and I Ca-L and promote APD prolongation in female guinea pig cardiomyocytes ( Kurokawa et al., 2008 ; Yang et al., 2010 ). Clinically-defined cardiac arrhythmias occur at the myocardium and organ level. Ectopic excitation at the myocyte level does not necessarily propagate and produce arrhythmias at the myocardium level. 3D human cardiac organoids recapitulate the multicellular properties and some physiological functions of the heart, and are as close a model as we can obtain to human cardiac tissue ( Lewis-Israeli et al., 2021 ; Zhao et al., 2021 ; Cho et al., 2022 ). We showed that BPA exposure triggered numerous EAD-like arrhythmias in human cardiac organoids under LQT2 setting (but no tachycardia-like activities). These results at the cardiac tissue level further demonstrate the potential pro-arrhythmic toxicity of BPA. We show that BPA and several BPA analogs have similar APD prolongation effects in human cardiac organoids. These BPA analog chemicals are structurally similar to BPA, and are used as BPA alternatives in various BPA-free products ( Rochester and Bolden, 2015 ; Chen et al., 2016 ). Despite their implied or perceived better safety ( Scherer et al., 2014 ), BPA analogs have been shown to share generally similar endocrine disrupting effects and toxicities of BPA in various cell/tissue types ( Kinch et al., 2015 ; Mao et al., 2020 ; Harnett et al., 2021 ; Winkler et al., 2022 ). We have previously shown that BPS and BPA have nearly identical effects on Ca 2+ handling disruption and arrhythmogenesis in rodent heart ( Gao et al., 2015 ). BPA and several of its analogs have been shown to have similar effects on cardiac ion channels ( Prudencio et al., 2021 ). Among the commonly used BPA substitutes, we show that BPAF had a worse toxicity profile as measured by APD prolongation and EAD development. This is consistent with reported comparisons of the toxicities of bisphenol-type chemicals. In a mouse model of endometriosis, BPAF has been shown to cause greater endometriosis lesion growth than BPA ( Jones et al., 2018 ). In comparison with other bisphenols, BPAF induced the highest oxidative stress and damage in human red blood cells ( Maćczak et al., 2017 ). BPAF at nanomolar concentrations exhibited stronger estrogenic effects than BPA via GPER pathway ( Cao et al., 2017 ). Moreover, BPAF was found to induce stronger toxicity and estrogenic activity than BPA in zebrafish ( Moreman et al., 2017 ). Combined, these results argue that the biological and toxicological effects of these BPA analogs need to be evaluated before their safety is assumed. APD prolongation increases the risk of arrhythmias but does not necessary trigger arrhythmias in all hearts, under all conditions. Indeed, we show that the response to BPA and its analogs differed markedly in normal human heart cells and tissues vs those with drug-simulated LQT syndrome phenotype. BPA and its analogs resulted in modest APD prolongation but no ectopic excitation or arrhythmia events in baseline condition, whereas in models under simulated LQT setting, these chemicals rapidly promoted arrhythmogenic EADs in cells and tissues, and even tachycardia-like events in 2D monolayers. These results resemble our previous findings in rodent hearts, where rapid exposure to BPA resulted in pronounced arrhythmia events only under pathophysiological conditions, but not in hearts under normal conditions ( Yan et al., 2011 ; Yan et al., 2013 ). These results highlight the role of preexisting arrhythmia susceptibility in determining the response of the heart to BPA and other environmental EDCs with pro-arrhythmic toxicity. It is reasonable to assume that in healthy individuals, BPA, at typical environmental exposure levels, does not by itself lead to clinically significant cardiac events. What may play a key role in determining the response of hearts to BPA and related EDCs is the existence of pathophysiological conditions such as heart disease or genetic mutations; such pathophysiological conditions provide the necessary condition for arrhythmogenesis, making the heart vulnerable to BPA-induced arrhythmias. In hiPSC-CMs, the repolarization capacity of myocytes likely determines their response to BPA-induced repolarization delay. Normal cardiac myocytes have redundant repolarization ionic mechanisms involving multiple K + channels ( Roden et al., 2002 ). Partial blockade of the hERG channel with E4031, mimicking LQT syndrome type 2 ( January et al., 2000 ), reduced repolarization capacity of myocytes and allowed the APD prolongation effect of BPA to manifest as significant increase in arrhythmic events. Our results argue that the cardiac toxicity of BPA and other related EDCs should be evaluated not only in the general population, but particularly in vulnerable subpopulations and individuals who are predisposed to arrhythmogenesis and other cardiac abnormalities. Such “precision environmental health” approach is necessary to better define the toxicity of EDCs which may be masked in the general population but is revealed in susceptible subpopulations, and to better inform the protection and treatment of individual patients. Future studies using hiPSC-CM models carrying disease-related mutations introduced by gene editing or derived from patients with idiopathic arrhythmias are warranted to directly define the individual-specific cardiac toxicity of EDCs in human heart cells. At such low nM concentration, we show that BPA prolonged APD in hiPSC-CMs through inhibition of I Kr and, to a lesser degree, I Ca-L . Our results are opposite to those on high dose BPA in hiPSC-CMs ( Hyun et al., 2021 ; Kofron et al., 2021 ; Prudencio et al., 2021 ). Acute BPA exposure at high doses of 1-100 μM has been shown to concentration-dependently shorten APD in hiPSC-CMs ( Hyun et al., 2021 ). Similarly, mid to high μM BPA has been shown to shorten APD in hiPSC-CMs and inhibit I Ca-L and hERG expressed in HEK 293 ( Prudencio et al., 2021 ). A recent study using hiPSC-CM cardiac microtissue demonstrated a biphasic effect of BPA on hiPSC-CM repolarization; BPA prolonged APD at 1 nM, and shortened APD at μM concentration ( Kofron et al., 2021 ). We showed that the dose response curve for BPA’s effect on APD in hiPSC-CM was nonmonotonic ( Figure 1 ). A slight nonmonotonic dose response was also noted in the study by Prudencio et al, where BPA slightly prolonged APD at submicromolar concentration and shortened APD at high dose ( Prudencio et al., 2021 ). The nonmonotonic, bidirectional effect of BPA on cardiac repolarization demonstrates the importance of using environmentally-relevant doses in EDC studies. Altered automaticity is a mechanism for cardiac arrhythmia ( Antzelevitch and Burashnikov, 2011 ; Paci et al., 2020 ). We show that 1 nM BPA rapidly increased the beating rate of hiPSC-CMs through stimulation of I f , which acts as the dominant pacemaker current in cardiac pacemaker cells ( DiFrancesco, 2010 ). This mirrors our previous finding that low dose BPS rapidly increased heart rate in female rat heats ( Gao et al., 2015 ), possibly through rapid stimulation of intracellular cAMP level ( Gao et al., 2013 ). By itself, the moderate effect of BPA on automaticity is unlikely to result in any clinically significant cardiac pathophysiological events, but may be detrimental to hearts with existing tachycardiac conditions such as sinus tachycardia. The effect of low dose BPA on cardiac automaticity is opposite to those of high dose BPA. Acute exposure to BPA, at micromolar up to millimolar concentrations, reduced the beating rate of rat atrial tissue, rat hearts and hiPSC-CMs ( Pant et al., 2011 ; Posnack et al., 2014 ; Hyun et al., 2021 ). The contrasting effects of BPA at high and low doses on cardiac automaticity, once again, argues the critical importance of dose in the toxicity assessment of EDCs.

We show that rapid exposure to environmentally-relevant low dose BPA had pro-arrhythmic toxicity through delay of cardiac repolarization and inhibition of the hERG channel in in vitro human cardiomyocyte models. This pro-arrhythmic action was particularly significant in myocytes and tissues under simulated LQT syndrome setting. We show that the adverse effects of BPA on repolarization delay and arrhythmogenesis were shared by its analog chemicals. Our study suggests that exposure to BPA and its related EDCs may be a risk factor for repolarization delay-associated arrhythmias in human hearts, particularly in LQT syndrome patients, and demonstrates that the preexisting pathophysiological conditions of the heart plays an important role in determining the cardiac impact of pro-arrhythmic EDCs. A “precision environmental health” approach is necessary for individualized risk assessment of EDCs toxicity.

BPA is synthetic chemicals that is used extensively in the manufacturing of plastic products and consumer goods. There is well documented human exposure to BPA from diet, inhalation and other exposure routes ( Vandenberg et al., 2007 ; Calafat et al., 2008 ; Geens et al., 2012 ). The CDC “National Report on Human Exposure to Environmental Chemicals” shows that the geometric mean of urinary total BPA in the US population is around 2 μg/L, or 9 nM ( CDC, 2017 ). BPA is an estrogenic endocrine disrupting chemical (EDC) with a range of potentially adverse impacts on human health ( Diamanti-Kandarakis et al., 2009 ; Gore et al., 2015 ; Kahn et al., 2020 ; Catenza et al., 2021 ). Epidemiological studies demonstrate that higher BPA exposure is associated with cardiovascular diseases or cardiovascular disease risk factors in humans ( Lang et al., 2008 ; Melzer et al., 2010 ; Shankar et al., 2012 ; Cai et al., 2020 ; Fu et al., 2020 ; Moon et al., 2021 ). Experimental studies shown that in rodent models, BPA exposure contributes to cardiac remodeling, heart development defects, alteration of cardiac function, oxidative stress, and arrhythmogenesis ( Gao and Wang, 2014 ; Zhang et al., 2020 ; Nagarajan et al., 2021 ; Zhou et al., 2021 ; Cooper and Posnack, 2022 ; Dias et al., 2022 ; Sirasanagandla et al., 2022 ). With increased recognition of the potential adverse health impact of BPA, manufacturers are switching away from BPA-based consumer plastics and have introduced various “BPA-free” products. These plastics are made from BPA analog chemicals, such as bisphenol S (BPS) and bisphenol F (BPF). Contrary to the implied better safety of these alternative chemicals, they have been found to share many of the endocrine disrupting actions and toxicities of BPA, including cardiac toxicity ( Chen et al., 2016 ; Zhang et al., 2020 ; Dias et al., 2022 ). Electrical excitation is fundamental to cardiac physiology. The morphology and duration of the cardiac action potential (AP) is tightly controlled by the sequential opening and closing of a number of cardiac ion channels ( Grant, 2009 ). Perturbation of cardiac electrical properties is a major contributor to cardiac arrhythmias ( Pogwizd and Bers, 2004 ; Schmitt et al., 2014 ). Hearts of large mammals are characterized by broad APs ( Rosati et al., 2008 ) and have distinct arrhythmogenic mechanisms. In particular, excessive prolongation of the ventricular AP, which manifests as prolongation of the QT interval on the electrocardiography (ECG), can lead to aberrant excitations of ventricular myocytes, malignant arrhythmias and sudden cardiac death, and is a key arrhythmogenic mechanism in large mammals ( Pogwizd and Bers, 2004 ). Such delay in ventricular repolarization can occur due to genetic mutations of cardiac ion channels, resulting in idiopathic long QT (LQT) syndrome ( Kass and Moss, 2003 ; Pogwizd and Bers, 2004 ; Sanguinetti and Tristani-Firouzi, 2006 ; Bezzina et al., 2015 ). Delay in cardiac repolarization can also occur due to adverse blockade of cardiac ion channels by drugs or environmental chemicals ( Cavero and Crumb, 2005 ; Sanguinetti and Tristani-Firouzi, 2006 ; Kannankeril et al., 2010 ; Pollard et al., 2010 ). One cardiac ion channel that is particularly susceptible to blockade or inhibition by chemicals is the rapid delayed rectifier K + channel I Kr channel), also known as the hERG channel, encoded by the hERG (or KCNH2 ) gene ( Sanguinetti et al., 1995 ; Sanguinetti and Tristani-Firouzi, 2006 ). The I Kr channel plays an important role in regulating cardiac action potential duration and repolarization in the hearts of large animals, such as humans, and adverse blockade/inhibition of I Kr is responsible for the vast majority of drug/chemicals induced QT prolongation and arrhythmias ( Recanatini et al., 2005 ; Witchel, 2011 ; Vandenberg et al., 2012 ). QT prolongation due to adverse effects of chemical or pharmaceutical agents is one of the central issues in cardiac safety and toxicology ( Cavero and Crumb, 2005 ; Kannankeril et al., 2010 ; Pollard et al., 2010 ). Examples of cardiac toxicity of environmental chemicals mediated by QT prolongation include trivalent arsenic ( Chen et al., 2010 ; Chen et al., 2013 ; Moon et al., 2018 ), phenanthrene ( Brette et al., 2017 ) and organophosphate ( Shiyovich et al., 2018 ; Thandar et al., 2021 ). Although BPA is a near ubiquitous environmental chemical, the impact of environmentally relevant low dose BPA on cardiac electrical properties, particularly in model systems that are directly relevant to human heart, is not known. As a promising human in vitro cardiomyocyte model, human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) have cellular electrical properties and ionic current characteristics that are similar to adult human cardiomyocytes ( Karakikes et al., 2015 ). They can be used for modelling human heart diseases and testing drug cardiotoxicity ( Hoekstra et al., 2012 ; Yamamoto et al., 2016 ). In the present study, we examined the acute effect of low dose BPA, the prototypical bisphenol-type EDC, on the cellular electrical properties of hiPSC-CM-based cardiac models, with a particular focus on action potential prolongation, and the susceptibility to repolarization delay-associated arrhythmia. The actions of three common BPA analogs including BPS, BPF, BPAF on cardiac repolarization and arrhythmogenesis were also examined.