High-throughput methods for cardiac cellular electrophysiology studies: the road to personalized medicine

Abstract

Personalized medicine refers to the tailored application of medical treatment at an individual level, considering the specific genotype or phenotype of each patient for targeted therapy. In the context of cardiovascular diseases, implementing personalized medicine is challenging due to the high costs involved and the slow pace of identifying the pathogenicity of genetic variants, deciphering molecular mechanisms of disease, and testing treatment approaches. Scalable cellular models such as human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) serve as useful in vitro tools that reflect individual patient genetics and retain clinical phenotypes. High-throughput functional assessment of these constructs is necessary to rapidly assess cardiac pathogenicity and test new therapeutics if personalized medicine is to become a reality. High-throughput photometry recordings of single cells coupled with potentiometric probes offer cost-effective alternatives to traditional patch-clamp assessments of cardiomyocyte action potential characteristics. Importantly, automated patch-clamp (APC) is rapidly emerging in the pharmaceutical industry and academia as a powerful method to assess individual membrane-bound ionic currents and ion channel biophysics over multiple cells in parallel. Now amenable to primary cell and hiPSC-CM measurement, APC represents an exciting leap forward in the characterization of a multitude of molecular mechanisms that underlie clinical cardiac phenotypes. This review provides a summary of state-of-the-art high-throughput electrophysiological techniques to assess cardiac electrophysiology and an overview of recent works that successfully integrate these methods into basic science research that could potentially facilitate future implementation of personalized medicine at a clinical level.

INTRODUCTION

Precisely coordinated contraction of the heart relies on the unimpeded propagation of electrical signals across all individual cardiomyocytes that make up the myocardium. The action potential (AP), an elegant modus for signal propagation across excitable cells, refers to a transient reversal of membrane potential that is initiated and terminated by the flux of positively charged ions into and out of the cell via various membrane-bound ion channels (Fig. 1). Disease-relevant alterations to these channels can result in aberrant channel function, abnormal AP profiles, and will often translate into severe cardiac diseases, usually with poor clinical outcomes for patients. Knowledge of the mechanics and biophysics of cardiac ion channels and properties of AP propagation are therefore essential in the treatment of cardiac disorders. Ideally, patient treatment would be personalized, whereby specially tailored medical treatment can be provided to individual patients based on their specific disease presentation. Personalized medicine is currently not feasible on a clinical level due to high costs and the slow rate at which molecular mechanisms of disease are determined and treatment modalities can be tested. Physiologically accurate and easily scalable cardiac cellular models are necessary along with an increase in the throughput at which functional measurements can be carried out.

Figure 1.

In silico human induced pluripotent stem cell-derived cardiomyocyte (hiPSC-CM) action potential and contributing currents. A: representative action potential (AP) of hiPSC-CM at 60 days postdifferentiation generated from the iMATURE hiPSC-CM model (2). Key AP phases are marked: phase 0 (depolarization), phase 1 (early repolarization), phase 2 (plateau), phase 3 (repolarization), and phase 4 (resting). B: major currents contributing across the AP duration acquired from the iMATURE model from hiPSC-CM at 60 days postdifferentiation. Currents include Na⁺ current (I_Na), L-type Ca²⁺ current (I_Ca,L), transient outward K⁺ current (I_to), rapid component of the delayed rectifier current (I_Kr), slow component of the delayed rectifier current (I_Ks), and basal inward rectifier current (I_K1). Seibertz et al. (2).

Historically, heterologous expression systems such as human embryonic kidney (HEK293) cells, Chinese hamster ovary (CHO) cells, or Xenopus oocytes have been used to study the biophysical characteristics and pharmacology of cardiac ion channels. Although scalable and inexpensive, these constructs do not faithfully recapitulate the cardiac phenotype in terms of intracellular signaling, Ca²⁺ handling, ion channel trafficking, and protein interactions (1). Their use for detailed study of cardiac electrophysiology in health and disease is therefore somewhat limited. Large animals (pigs, dogs, sheep, rabbits, guinea pigs, and goats) have allowed for the study of global cardiac electrophysiology in disease states, including the transition from paroxysmal atrial fibrillation to sustained atrial fibrillation (3, 4), and the contribution of ionic currents, Ca²⁺ machinery, and electrical propagation heterogeneity to the initiation and maintenance of ventricular arrhythmias (5–8). Smaller models (mice and rats) are cheaper to house and breed and are commonly used for studying genetic modification of cardiac constructs through genetic overexpression, viral gene transfer, or CRISPR-Cas9 knockout/in approaches (9). On a cellular level, ion channel distribution and activity show pronounced species-specific differences, particularly in ion channels involved in repolarization of the AP. This produces severe differences in AP morphology and duration between species that limits the transferability of functional and pharmacological findings to human subjects. Human cardiac samples are available from residual tissue biopsies from various cardiac surgery procedures (10, 11). These samples provide the most accurate modality to investigate cardiac electrophysiology; however, this construct is seriously hampered by limited availability, limited culture applications, and the fact that samples may come from patients with preexisting heart disease (2, 12).

In contrast, human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) are acquired from patients via noninvasive biopsies that contain somatic material (13). Following reprogramming and cardiac differentiation, this construct promises an unlimited source of patient-specific cardiomyocyte material, immensely useful for drug testing and modeling states of cardiac health and disease (13). Revolutionary in the field of medical biology, regenerative medicine, and disease modeling, these constructs also avoid the practical and ethical pitfalls associated with using tissue derived from human embryos or animal models (14). Personalized medicine is consistently hailed as the pinnacle application of hiPSC-CM platforms because of their ability to perfectly reflect the genotype of the donor patient. Although true in theory, the notion of successful and timely generation, differentiation, validation, and pharmacological screening of hiPSC-CMs from a single patient for tailor-made therapy has yet to be achieved on a large and clinically accessible scale. To alleviate this problem, high-throughput electrophysiological methods are emerging as critical techniques to quickly study cardiac cellular activity with direct clinical relevance for patient-centered treatment and care. This review will outline how modern high-throughput methods to investigate cardiac cellular electrophysiology can be successfully implemented into basic science research pipelines.

HIGH-THROUGHPUT METHODS TO MEASURE CELLULAR ELECTROPHYSIOLOGY

Traditional acquisition of cellular ion currents and analysis of ion channel properties is done with the patch-clamp technique. Established in 1976 by Erwin Neher and Bert Sakmann and acknowledged with the Nobel Prize in Physiology or Medicine in 1991, this method involves the direct examination of membrane currents mediated by ion channels in cell membranes through the use of a thin glass micropipette (electrical resistance, 2–10 MΩ) that is carefully, manually micromanipulated and sealed against a small patch of cell membrane (15). Once a small window is broken into the cell, the ionic activity of interest can be isolated and studied (voltage-clamp method). This technique can also be used to measure APs in excitable cells where electrical access to the cell can facilitate recording of the intracellular voltage (current-clamp method). The patch-clamp method necessitates an intricate and laborious procedure that must be undertaken by a skilled experimenter. Although it relays crucial information about electrophysiological cellular function, it is, by definition, low throughput, meaning that the data output is limited by the slow speed at which individual cells can be measured (Fig. 2).

Figure 2.

Methods for measuring cellular electrophysiology. Manual patch-clamp (left) involves manual micromanipulation of a pipette onto the cell of interest and establishing an electrical connection to the cytoplasm. Highly sensitive and fast-voltage sensitive-dyes (middle) function by embedding into the cellular membrane and emitting fluoresce at an intensity relative to the membrane potential. Automated patch-clamp (right) operates using planar substrate onto which the cells passively attach to a precision-drilled micropore that functions as the pipette. Through the application of negative pressure, electrical connection to the cytoplasm can be established. These techniques all represent end-point measurements. After investigation, cells were disposed. hiPSC-CMs, human induced pluripotent stem cell-derived cardiomyocytes.

Automated Patch-Clamp

The revolutionary technique of automated patch-clamp (APC) relies on the use of a planar substrate of quartz, silicon, or glass perforated by fine openings (1∼2 µm diameter) as a substitute for the typical glass patch-clamp pipette (16). Cells are deposited in suspension onto the substrate surface, and through gravity or negative suction pressure, a single cell is positioned onto the opening where a high-resistance seal is able to be established and held for 30 min or more (16–19). The pairing of such a technique with high-throughput platforms containing parallel amplifiers and automated liquid handling capabilities for pipette and bath solutions allows for the simultaneous patch-clamp of multiple cells at once. Since 2004, APC platforms have been commercially available, with market leaders focused on optimizing high-throughput systems for higher cell catch, streamlined liquid handling, GΩ-seal establishment, and current-clamp capabilities (17, 20). Currently, multiple vendors offer flagship APC models that can record ionic currents or action potentials from up to 384 cells in parallel during a single experiment (Fig. 2).

APC offers rapid solution exchange in the internal, or “pipette,” compartments during an experiment that uniquely allows for acquisition of multiple electrophysiological parameters from the same cell. This feature is valuable for increasing data quantity and reducing cell waste. Changing the pipette solution while in whole cell configuration is almost impossible to achieve using manual patch-clamp techniques (21).

Different APC devices possess alternative hardware through which the cells are guided onto the patch-clamp aperture and subsequent external solutions are added. Options include laminar-flow microfluidic chambers through which cells are perfused through small >400-µm channels that contain the patch aperture(s). The caudal end of each channel feeds into a waste reservoir. This organization allows for rapid, low-volume external solution exchange that ensures elegant perfusion and washout of ligands. The 4–8 channel Patchliner (Nanion Technologies), the 16–256 channel IonFlux Mercury line (Fluxion Biosciences), the 8–48 channel QPatch line, and 384-channel Qube (both Sophion Bioscience) use microfluidic channels in their recording plates. In contrast, fixed-well chambers highlighted by the single-channel Port-a-Patch or the 384-channel SyncroPatch 384 (both Nanion Technologies) do not contain microfluidics and instead operate using fixed-volume wells into which all cells and external solutions are applied and removed. Both well formats are also available with multihole configurations to increase the chances of cell catch. Both formats also possess both distinct advantages and disadvantages, for example external fluid wash out with fixed-well hardware requires careful experiment preplanning, particularly when devising multiparameter protocols (i.e., multiple currents from the same cell). Conversely, fixed-well installations may allow for easier access of larger cells such as primary cardiomyocytes to the patch-clamp aperture (22) and also are well suited for the application of overhead LED optical stimulation. This unlocks APC for a wide range of optogenetic applications in the measurement of cells that express channel rhodopsins (23).

Three-hundred eighty-four-channel APC is high throughput by definition; therefore, it follows that it can demand high running costs. Currently available 384-well recording plates are single use and are disposed after each experiment. In addition, it follows that large volumes of solutions are required compared to manual patch-clamp. Bulk solutions for APC applications are therefore often prepared in advance, either in-house or in a central facility that ensures identical batch quality. Overall, the cost per data point is still substantially lower using high-throughput APC compared with manual patch-clamp techniques, where experimenter time introduces a costly variable (22).

APC has been developed over the years to predominately serve the needs of the pharmaceutical and biotech industries for drug discovery (17). Large libraries of new chemical entities (NCE) must be rigorously tested on heterologous systems expressing channels of interest to ensure that the compounds impart limited proarrhythmic risk (24). K_V11.1, conducting the rapid component of the delayed rectifier K⁺ current (I_Kr), is a popular candidate due to the QT prolongation and fatal Torsades de Pointes that can manifest in patients upon channel block (25). In 2013, the FDA implemented the comprehensive in vitro proarrhythmia assay (CiPA), which stresses the need for testing prospective drugs on a broader panel of ionic currents that mediate cardiac repolarization, not just I_Kr (26). Due to its high-throughput nature, APC has been invaluable for drug discovery over the past decades. Historically, the high cost of APC platform manufacture and design has ensured these devices have largely remained outside the realm of academia where funding often remains an obstacle. As APC technology becomes more cost effective and amenable for academic use, this situation is changing. The more diverse needs of academia are being used to push the technology to its highest potential, for example through the assay of primary human cellular material such as red blood cells, cortical neurons, T lymphocytes (18, 27–29), hiPSC-CMs, and recently primary cardiomyocytes (22).

Single-Cell Photometry

Photometry methods with voltage-sensitive dyes (VSDs) provide an alternative method for AP measurement from single cells that is noninvasive, delivers a quality and temporal resolution equivalent to that of traditional patch-clamp methods (30, 31), and keeps the high-throughput moniker for fast AP characterization of single cells or monolayers (Fig. 2). VSDs are chemically synthesized voltage detectors that have been used since the 1970s (30). They interact with the membranes of excitable cells including primary neurons (32), primary cardiomyocytes (33), and hiPSC-CMs (34), and respond to changes in absolute membrane potential with corresponding changes in fluorescence. Modern VSDs should possess fast kinetics to detect and respond to membrane potential changes, and a high sensitivity, meaning that they respond with a high dynamic range during small changes of membrane potential. Ideally, it would also display minimal toxicity and resistance to photobleaching (35, 36). Electrochromic dyes such as the ANEP family bind to the intracellular side of the membrane and change emission spectra based on the cellular electric field. They are fast, but insensitive, yielding 10% ΔF/F per 100 mV. FRET sensors oscillate between leaflets of the plasma membrane and interact with a fixed intracellular fluorophore. The process is slow, but very sensitive, yielding high-intensity fluorescence emission during oscillations in membrane potential (∼56% ΔF/F per 100 mV) (34). Finally, probes that use photoinduced electron transfer (PeT) represent a good middle ground, presenting with high speeds and good sensitivity (∼27% ΔF/F per 100 mV). Here, a modular electron-rich probe passively inserts into the outer leaflet of the plasma membrane and uses the cellular electric field to modulate fluorescence from the rostral fluorophore (37). FluoVolt is a widely available PeT VSD with a FITC fluorophore, operating under standard green fluorescent protein (GFP) spectra. With infrequent low-intensity excitation (to avoid bleaching), FluoVolt experiments on single cells can remain stable for up to 1 h (34, 36). Similar PeT probes such as BeRST1 (38) and RhoVR1 (39) are also available operating under different spectral conditions, but with enhanced sensitivity. PeT probes are not electrochromic, meaning that they do not reversibly change emission wavelengths upon cellular depolarization. This ensures only a single detector is required; however, the resulting lack of signal ratiometry options prevents estimations of resting membrane potential.

VSD use is well suited for photodetection systems involving total-signal photometry measurements from single isolated cells. Photomultiplier (PMT) detection systems [available separately (Cairn Research) or integrated (CellOPTIQ, Clyde Biosciences)], coupled with fast-switching illumination components, represent cost-effective hardware that allows for extremely high temporal resolution with a digital signal of 10 kHz and analog sampling of up to 1 kHz. This ensures more experimenter control over signal processing and is important for the accurate capture of AP, signals that display very fast kinetics (in the order of ms). PMT capture involves little to no spatial discrimination, registering all photons from a defined area that hit the detector. When used appropriately with the correct settings, i.e., adequate input voltage, narrow emission bandwidth filters, and with low ambient light/low external noise conditions, PMTs are designed to consistently offer signal-to-noise ratios higher than one with exceptionally fast time responses. This type of speed and control is not possible using customary cameras at the equivalent monetary cost. Common CCD or CMOS cameras usually operate with good signal-to-noise capabilities, i.e., reduced dark current noise, but often focus on the development of spatial capabilities (full well capacity and required bit depth) at the expense of speed. As a result, most typical cameras acquire at <100 Hz (40). This is gradually changing as new technologies emerge for efficient, large-area sensors with fast interface transfers (PCIe cable instead of standard USB connectors), significantly increasing the available acquisition frequency at low bit-depth (Kinetix, Teledyne Photometrics). High-content plate readers increase measurement throughput phenomenally but should be paired with segmentation recognition algorithms to define the boundaries of each cell under investigation (IC-200 KIC, Vala Sciences) to avoid whole well or region-of-interest (ROI) averaging. Combining this technology with low-cellular plating densities is important to ensure that measurements are taken from single, isolated cells to reduce the confounding effects of electrical, mechanical, or paracrine influences of adjacent cell bodies. Because of technical limitations, currently available automated plate readers usually acquire at <100 Hz, which must also be taken into consideration. The level of detail and control offered by total signal photometry via PMT capture is only matched using very expensive, high-resolution, large-area cameras for quantitative imaging (MiCAM line, SciMedia). Here, large full-well capacities and low-read noise values provide exceptional dynamic range and signal-to-noise ratios, respectively. The high spacial resolution provided by these cameras is useful for optical mapping, which involves the detection of spatially sensitive electrophysiological information such as conduction velocity or excitation patterns across intact cardiac tissue.

This review includes single-cell photometry using VSDs as high-throughput alternatives to current-clamp for multiple reasons: 1) the speed at which full data sets can be acquired with this method translates into days rather than months as manual patch-clamp would require, 2) the complexity of acquisition and analysis does not require intensive user training, and 3) noninvasive cell measurement both preserves cellular structure, removes the need for artificial current injection, and avoids systematic error that plagues the “gold standard” current-clamp technique. The limitations of single-cell photometry measurements include the temporal bottleneck that requires a researcher to still investigate individual cells one by one using conventional systems and the inability of the technique to resolve absolute membrane potential in the absence of ratiometric probes and complex acquisition setups (41).

hiPSC-CM IN A HIGH-THROUGHPUT WORLD

The following sections will discuss the current state and relevance of research regarding high-throughput characterization and modeling of hiPSC-CM electrophysiology in health and disease. Readily applicable to native cardiomyocytes (33, 34), VSDs have also been extensively paired with hiPSC-CM technology over the past decade for detection of AP morphology and duration in health and disease (42–45). In contrast, APC techniques are only recently gaining momentum in the robust characterization of multiple hiPSC-CM currents and electrophysiological properties.

hiPSC-CM Physiological Suitability

Typical cardiac differentiation of hiPSC yields cells of a spectrum of regional cardiac subtypes (46). With advances in differentiation protocols, hiPSC can be targeted to produce predominantly ventricular or atrial cardiac subtypes that show subtype-specific structure and function, such as upregulation of the atrial-specific isoform of myosin light chain or ion channels K_v1.5 and K_ir3.1/3.4, which subsequently manifest on atrial AP (47, 48). Repolarization fraction [RF; (APD₉₀ – APD₅₀)/APD₉₀] can be used as an index to highlight the relationship between early and late repolarization, and therefore can act as a mathematical differentiator between AP from atrial-like or ventricular-like hiPSC-CMs (49). This method has been used in great effect to differentiate ventricular and atrial hiPSC-CM AP using VSD analysis (47, 48, 50). APC has also been recently used to differentiate between atrial and ventricular cardiomyocytes regarding AP duration (22) and sensitivity to atrial-specific acetylcholine-activated inwardly rectifying potassium currents mediated by K_ir3.1/3.4 (48).

The ease with which functional screening methods can be applied is important because although hiPSC-CMs represent highly scalable constructs, persistent phenotypical variability and immaturity are often reported between batches, between lines, and between laboratories. This somewhat limits their use in disease modeling or therapy development (51, 52). Fast and functional screening is useful to lessen this disadvantage, particularly when combined with modern techniques such as dynamic patch-clamp, a feature that uses an in silico hiPSC-CM model to inject precise amounts of artificial basal inward rectifier current (I_K1) during AP measurement (53). It is common sentiment that many aspects of hiPSC-CM differentiation and measurement conditions need to be standardized in order to reduce heterogeneity between laboratories. APC technology could also be useful in the development of standard recording protocols that generate large volumes of electrophysiological data and help with data generalization. This is a major goal of CiPA, with participating institutions gradually adopting widespread “best practice” guidelines (54). Large blinded, multisite comparisons of drug responses on expression systems appear to show promising homogeneity and applicability for CiPA applications across multiple users (55).

Apart from limitations in heterogeneity and differentiation standardization, it remains difficult for hiPSC-CMs to completely replace heterologous expression systems in the streamlining of CiPA processes using APC. This is because hiPSC-CMs contain very small delayed rectifier activity that is often difficult to measure (56, 57). Current reports on hiPSC-CM I_Kr measurements through APC are limited (58) and mostly measure inward tail current in the presence of high external K⁺ to facilitate a positive shift in E_Rev (59, 60) based on previously established manual patch-clamp protocols (61). I_Kr tail current in hiPSC-CMs has also been assessed through equimolar Cs⁺ conductance (56) using APC (2). To date, APC has not been used to sensitively resolve I_Kr step current in hiPSC-CM, which currently limits deeper biophysical assessment of these channels. This may be due to the enzymatic cellular dissociation protocols before measurement that may impart proteolytic damage to relevant channel subunits. I_Kr in particular appears to be sensitive to enzymatic digestion before voltage-clamp measurement using conventional patch-clamp (62, 63).

Electrophysiological variability in hiPSC-CMs also arises from the age at which hiPSC-CM properties are measured. Significant differences in multiple currents have been reported in the same batches of hiPSC-CMs measured only 2 wk apart (2, 64–66). This presents a clear advantage of higher throughput methods such as APC paired with hiPSC-CM technology, which can acquire multiple data points from many cells at the same point in their development, effectively limiting the aspect of age-dependent heterogeneity that often goes unreported. APC also lends itself well to the construction of data-based in silico models, which are normally limited by the variability and uncertain origin of the underlying data (51, 67).

Increasing studies highlight multiple aspects of robust and cardiac subtype-sensitive hiPSC-CM measurement using APC with recordings of L-type Ca²⁺ current (I_Ca,L), Na⁺ current (I_Na), I_f, transient outward K⁺ current (I_to), I_K1, and I_K,ACh (22, 31, 48, 58, 68–71), which can be routinely acquired in a high-throughput format. Indeed I_Ca,L from the same hiPSC-CM cell line were comparable in measurements from APC (22) and using manual patch clamp (2).

Ciinical Relevance

Measurement of hiPSC-CMs using high-throughput electrophysiological techniques is often carried out with a focus on pharmacological screening (17, 31, 72). High-throughput hiPSC-CM disease models that propose clear roads for therapy development are less established. Among these, single-cell photometry using VSDs has been applied for novel monitoring of cellular arrhythmogenic potential via AP restitution in hiPSC-CMs. AP restitution refers to the relationship between AP duration and the previous diastolic interval. The physiological phenomena of rate dependence describe the shortening of the AP at high frequencies to allow for adequate ventricular filling and coronary perfusion between beats (73, 74). Aberrant (steep) restitution can result in membrane instability and a bifurcation of possible AP durations leading to AP alternans (75, 76) whereby AP properties alternate between two contrasting conditions (e.g., long-short) (77–79). Cells or tissues prone to alternans development represent an arrhythmogenic substrate because under such conditions, some areas of the tissue may remain refractory whereas others can freely propagate (dispersion of refractoriness). A premature ventricular contraction (PVC) can therefore encounter functional block in some areas but not in others that could lead to a reentrant circuit. AP alternans secondary to increased myofilament calcium buffering were recently identified in single hiPSC-CMs derived from a patient with an R173W mutation in troponin T, presenting with dilated cardiomyopathy (52) (Fig. 3D).

Figure 3.

Single-cell photometry using voltage sensitive dyes (VSDs) on intact cardiomyocytes. A: representative photomicrographs of a native cardiomyocyte loaded with FluoVolt VSD under brightfield conditions (left), and during excitation with a 470-nm LED, emission captured at 353 nm (right). B: absorption and emission spectra of FluoVolt. C: simplified render of molecular FluoVolt integration (loading) into the cellular membrane. D: examples of continuous recordings of membrane voltage from various studies using human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CM). hiPSC-CM derived from healthy patients (blue) can be used as wild-type controls (WT, top), or gene edited to represent a particular cardiac phenotype to investigate mechanisms of disease and propose novel, feasible treatment options (middle). Disease modeling using hiPSC-CM directly derived from patients with known genetic variants (red) also represent useful tools to investigate therapeutic solutions (bottom). Figures and data adapted from Refs. 2, 31, 34, 44, and 52 with permission where required.

Comprehensive single-cell photometry analysis has also recently allowed for the identification of the arrhythmogenic potential of CAV3 variants seen in patients with long-QT syndrome type 9. Here, significant early afterdepolarization (EAD) events were observed in CAV3 knockout (KO) hiPSC-CMs compared with wild type (WT) (44). EADs are also relevant markers of pathophysiological function and increase the likelihood of reentrant circuit initiation in myocardial tissue. In this study, arrhythmogenic behavior was elegantly attributed to upregulation of the late component of the sodium current. In a different study, VSDs were also used to successfully validate novel and highly selective late sodium current blockers in hiPSC-CMs derived from patients with long-QT syndrome type-3 with limited off-target effects (31) (Fig. 3D).

Other advances in the realm of direct clinical translation using high-throughput methods involve patients with Brugada syndrome who are at severe risk for the development of ventricular arrhythmias. APC has recently been used to extensively characterize hiPSC-CMs derived from patients with a p.S1812X mutation in the gene encoding for SCN5A (70). Robust measurements revealed decreased I_Na and increased I_to in line with previous manual patch-clamp data from the same laboratory (80). Importantly, APC was used to elegantly screen phosphodiesterase (PDE) blockers commonly used in clinical settings. Cilostazol and milrinone both decreased I_to in patient-derived Brugada syndrome hiPSC-CMs, but not in WT controls indicating that PDE inhibition could represent a novel therapy for patients with Brugada syndrome (80). Monitoring I_Na in hiPSC-CM using APC has also recently uncovered potential roles for SGLT2 inhibitors such as dapagliflozin in the atrial-specific regulation of cardiac excitability (81). This is particularly important because effective atrial-specific drugs for rhythm control that do not induce proarrhythmic ventricular side effects are severely needed on a clinical level to treat patients with atrial fibrillation (AF).

AF manifests in patients as a vicious cycle of tachycardia-induced ion channel remodeling in the atria that, in turn, increases the propensity for reentry. Mechanistic insights and pharmacological discovery should therefore be undertaken in samples that exhibit endogenous AF-associated remodeling of ion channels. Animal models have historically fit this purpose; however, issues regarding mechanistic translatability from animals to humans limit their conclusions (2). Recently, APC and single-cell photometry methods have been used to establish an atrial-specific hiPSC-CM model of AF-associated remodeling (48). The study used optogenetic tools to produce atrial-specific hiPSC-CM which, after 7 days of continuous high-frequency pacing, exhibited all major hallmarks of AF-associated remodeling including decreased I_Ca,L, increased I_K1, and the establishment of a constitutively active I_K,ACh (I_K,ACh,c) (48). APC will likely prove incredibly useful in future screening for atrial-specific therapeutics to counteract these remodeled characteristics.

Importantly, high-throughput electrophysiological analysis in these aforementioned studies offered rapid deep assessment of electrophysiological function and proposed clear options for future therapy development and application. As these techniques advance further and a broader panel of hiPSC-CM currents and parameters can be used, APC paired with hiPSC-CM technology will be immensely valuable for future industrial paradigms, academic projects, and clinical applications.

Technical Considerations

A number of technical aspects also need to be considered when assessing the usefulness of hiPSC-CM technology paired with APC. High-throughput APC requires substantial amounts of biomaterial, particularly for 384-channel systems. Recent work suggests that ∼400,000 cells/mL are necessary to ensure 50% cell attachment to single-patch aperture plates in the absence of suction pressure. When using a minimal amount of 100,000 cells/mL, a 20% success rate for whole cell configuration was observed (22). This is in line with a previous report focusing on minimizing cell use for APC applications (18). With hiPSC-CM culture becoming more cost effective and with optimization of differentiation and subsequent digestion procedures, this success rate will be readily improved.

Second, a historical feature of planar patch-clamp involves the addition of fluoride ions (F⁻) into the internal solution (82). When used in conjunction with a solution containing divalent ions, this fosters GΩ-quality seals ensuring high-quality recordings (83–85). It is possible that F⁻ in the internal solution can have a negligible, but detectable impact on the biophysical properties of various Na⁺ (activation curve shift), Ca²⁺ (increased open probability), and K⁺ (increased amplitude) channels and should be considered when reporting APC-acquired data. In recent years, F⁻ free applications have been heavily developed and are now possible across a range of APC devices (85).

Current-clamp methods in both manual patch-clamp and APC applications are somewhat limited by the nature of their technical function. Typical patch-clamp amplifiers are designed for measurements of ionic currents in voltage-clamp mode. When in current-clamp mode, patch-clamp amplifiers are characterized by relatively low intrinsic input resistance resulting in aberrant current absorption from the cell during experiments. This causes a voltage drop across the cell-to-amplifier series resistance and alters the amount of current actually charging the membrane capacitance. Taken together, this may produce distortions in AP amplitude and depolarizing/repolarizing characteristics. Consequently, extra holding currents are manually delivered to keep resting membrane potential (RMP) consistently low (86, 87). This is an inherent disadvantage of the current-clamp technique in general, as the majority of papers report artificial RMP values due to the presence of these holding currents, which also influence global ionic flux and therefore AP shape. For measurement of true RMP values, sharp microelectrode techniques using “voltage follower” or bridge amplifiers are recommended, along with pipettes of very high resistances (>50 MΩ) (48, 87).

In traditional current-clamp measurements, seal quality can also impart large impacts on the reported RMP because a “leaky” seal will result in further current loss and will alter (increase) the measured membrane voltage (88). RMP is largely determined by the inward rectifier I_K1. In small cells such as hiPSC-CM, absolute inward rectifier conductivity is lower, which results in larger membrane resistance [driving force of I_K1 (V_M − E_Rev) divided by absolute Ba²⁺-sensitive I_K1]. This acts in series with seal resistance, causing a disparity in the measured RMP. A very clear and detailed description of this phenomenon is outlined by Horváth et al. (88). Furthermore, it is suggested that seal resistance must be at least fivefold higher than membrane resistance to negate these effects (89). This systematic error must be considered when conducting current-clamp experiments, particularly in small cells such as hiPSC-CM (88). This is poignant when discussing APC installations, as planar patch-clamp has historically operated under a different paradigm of acceptable seal resistances for hiPSC-CM, sometimes below 1 GΩ (17, 22, 90) compared with manual patch-clamp in which seal resistances of 1–10 GΩ are common (88). Indeed, a recent APC study using native cardiomyocytes reports a membrane resistance of up to 3 GΩ with seal resistances of <1 GΩ, not meeting the fivefold criteria specified above, which perhaps offers an explanation for the short AP reported (22). This implies that current-clamp measurements should be interpreted with caution, as both the amplifier circuitry and low seal resistances can necessitate high amounts of injected current and produce artificial readouts of RMP, respectively, therefore possibly misreporting AP morphology. Naturally, experimental conditions such as working temperature and solution composition also need to be considered.

It is important to note that standard practice in cardiac cellular electrophysiology studies requires measurement from isolated cells that do not contain connections to other cell bodies or aggregates. This attempts to standardize experimental conditions and remove confounding environmental factors that may be imparted from neighboring cells. In APC studies, all samples are dissociated to single cells and harvested before measurement. It follows that the cellular harvesting protocol is crucial for the success and quality of the measurement (17, 18, 70, 90). The enzymatic digestion process directly before APC measurement produces single ball-shaped hiPSC-CMs. In addition to enzymatic protein cleavage, it is unclear what affects this membrane and cell body contortion has on ion channel expression and function of hiPSC-CMs. A recent study incorporated EC uncoupler blebbistatin into the dissociation procedure to limit membrane disruption and prevent hypercontractility-related Ca²⁺ damage, which appeared to help hiPSC-CMs maintain rod-like shapes after digestion and increased subsequent I_Na density using APC (70).

Single-cell photometry measurements using VSD do not involve cells in suspension, and instead rely on plated constructs. It is clear that plating hiPSC-CMs in low-density conditions will likely introduce functional differences, potentially impacting their development in culture and ionic current expression (91, 92). hiPSC-CMs grown in isolated conditions do show more AP heterogeneity and a more depolarized RMP than when they are connected in a monolayer or three-dimensional construct such as Engineered Human Myocardium (EHM) (91, 93). This is likely due to a lower membrane resistance and increased pooled I_K1 between electrically connected cells, again highlighting the issue of systematic error when patching small cells to measure membrane voltage. This could also be due to cell-cell adhesion-dependent factors. Indeed, primary rodent cardiomyocytes showed larger I_Na density in the intercalated disk region when cells were paired compared with when they are isolated (94). K_ir2.1 is also known to colocalize with Na_V1.5, which implies that cell adhesion-dependent processes may also regulate K_ir2.1 functional expression (95). Intercellular variability in ion current expression has even been reported in rabbit cardiomyocytes that are isolated from the same region of the same heart (33). It is proposed that global AP characteristics are retained between cells on the tissue level through local relationships of ion channel conductances, which form a codependent system with a high redundancy to compensate for individual differences in ion channel expression. This system is completely lost when cells are isolated, increasing the variability measured between cells. It follows that this is also lost to some extent in isolated hiPSC-CMs, which provides a novel explanation for typical hiPSC-CM-related heterogeneity and highlights a disadvantage of typical single-cell measurements.

FUTURE APPLICATIONS OF HIGH-THROUGHPUT ELECTROPHYSIOLOGY

APC has very recently been applied to the characterization of primary cardiomyocyte electrophysiology (22). This is particularly exciting because residual cardiac material from cardiac surgeries can be acquired for the valuable study of human cardiac mechanisms in health and disease. For example, right atrial appendages are routinely removed during open heart surgery for technical reasons, along with left ventricular spindle samples from patients being implanted with left ventricular assist devices (79, 96, 97). Electrophysiological characterization of human material is usually limited by low-throughput manual patch-clamp approaches that limit functional data quantity. APC provides a unique opportunity to scale up these measurements for comprehensive study of patient-specific cellular electrophysiology and compilation of detailed ionic current and AP information.

APC has been shown to be compatible with multiparametric recordings (I_Ca,L, I_K1, I_K,ACh, and AP) from hundreds of primary cardiomyocytes over only 3 days of data acquisition (22). Should this output be paired with currently existing artificial intelligence (AI) neural network assemblies, the availability of patient-specific data could aid in the rapid diagnosis and treatment of patients based on various clinical biomarkers (Fig. 4). Previous studies have used such AI networks to precisely characterize and diagnose patients based on small patterns detected in their clinical ECG readouts (98). Furthermore, AI has recently been used to assess future atrial fibrillation-related risk in patients based on ECG alone (99, 100).

Figure 4.

High-throughput strategies for direct measurement of key cellular electrophysiology parameters aid the formation of predictive models for personalized medicine. Modern measurements allow for complete assessment of membrane voltage and ionic currents from both native and in vitro preparations such as human induced pluripotent stem cell derived cardiomyocytes (hiPSC-CMs) in high-throughput conditions. Miniscule deviations in action potential (AP) waveform can be linked to pathologies of major ionic currents across thousands of individual measurements to train models of machine learning. Deep learning configurations could then be used to categorize patients into recommended treatment regimens based on fast hiPSC-CM investigation of minimal parameters. Figure adapted from Ref. 22 with permission.

APC technology lends itself very well to AI-related data intensive tools where vast amounts of AP and ionic current data could be fed into a pattern-recognizing neural network to generate predictive readouts of ion channel function based on AP morphology or vice versa (101). In conjunction with variant characterization, this would allow for rapid patient-specific disease identification and appropriate personalized treatment regimes. This type of AI technology could also be paired with autologous hiPSC-CM equivalents from the same patients to phase out the requirements for invasive native tissue acquisition. Currently, hiPSC-CM phenotypical variability and the high costs of hiPSC-CM line generation and maintenance limit the widespread accessibility to patient-specific “personalized” medicine at a clinical level. A feasible solution is the utilization of machine learning techniques to translate high-throughput data into predictive models for direct use in systems medicine. Tens of thousands of APC-acquired experimental recordings from the same site and under the same conditions can be used then to build predictive tools regarding the conversion of characteristics between hiPSC-CM phenotype and clinical presentation to negate the classical limitations of hiPSC-CM use. APC could therefore be instrumental in shortening the gap between mechanistic studies and clinical applications in terms of tailoring patient treatment.

In summary, the concept of personalized medicine is still not yet a reality; however, it is clear that the unprecedented power of APC holds many promising future utilities. Widespread symbiosis of novel mechanistic and machine-guided studies coupled with high-throughput quantification of ion channel function will over time generate a vast landscape of new pharmacological paradigms for efficient and patient-tailored clinical cardiovascular care.

GRANTS

N.V. is supported by Deutsche Forschungsgemeinschaft Grants VO 1568/3-1 and VO 1568/4-1, Germany’s Excellence Strategy SFB1002 Project A13 and Grant EXC 2067/1-390729940, and German Center for Cardiovascular Research (DZHK) Grants 81X4300102 and DNAfix.

DISCLOSURES

F.S. is partially employed by Nanion Technologies, GmbH, Munich, Germany. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

F.S. prepared figures; F.S. drafted manuscript; F.S. and N.V. edited and revised manuscript; N.V. approved final version of manuscript.