Electrophysiological characterization of cardiomyocytes derived from human Pluripotent Stem Cells (hPSC-CMs) is crucial for cardiac disease modeling and for determining drug responses. This protocol provides the necessary information to dissociate and plate hPSC-CMs on multi-electrode arrays, measure their field potential, and a method for analyzing QT and RR intervals.
Cardiomyocytes can now be derived with high efficiency from both human embryonic and human induced-Pluripotent Stem Cells (hPSC). hPSC-derived cardiomyocytes (hPSC-CMs) are increasingly recognized as having great value for modeling cardiovascular diseases in humans, especially arrhythmia syndromes. They have also demonstrated relevance as in vitro systems for predicting drug responses, which makes them potentially useful for drug-screening and discovery, safety pharmacology and perhaps eventually for personalized medicine. This would be facilitated by deriving hPSC-CMs from patients or susceptible individuals as hiPSCs. For all applications, however, precise measurement and analysis of hPSC-CM electrical properties are essential for identifying changes due to cardiac ion channel mutations and/or drugs that target ion channels and can cause sudden cardiac death. Compared with manual patch-clamp, multi-electrode array (MEA) devices offer the advantage of allowing medium- to high-throughput recordings. This protocol describes how to dissociate 2D cell cultures of hPSC-CMs to small aggregates and single cells and plate them on MEAs to record their spontaneous electrical activity as field potential. Methods for analyzing the recorded data to extract specific parameters, such as the QT and the RR intervals, are also described here. Changes in these parameters would be expected in hPSC-CMs carrying mutations responsible for cardiac arrhythmias and following addition of specific drugs, allowing detection of those that carry a cardiotoxic risk.
Human Pluripotent Stem Cells (hPSCs) have the capacity to self-renew and generate virtually any cell type of the human body through differentiation1,2. Detailed protocols on how to direct differentiation of hPSCs into several cardiac lineages (ventricular, atrial, pacemaker-like cardiomyocytes) have been described3,4,5,6,7. Cardiomyocytes are electrically active cells and detailed knowledge of their electrophysiological activity can be extremely informative for understanding heart development and disease8. Patient-specific hiPSC-derived cardiomyocytes (hiPSC-CMs) have been successfully used to model and study the cellular, molecular, and electrical features of several cardiac arrhythmias, including Long QT Syndrome (LQTS)9,10,11,12,13, Brugada syndrome14, and cathecolaminergic polymorphic ventricular tachycardia15,16. Furthermore, multiple drugs have been added to diseased hiPSC-CMs to recapitulate therapeutic intervention and to rescue the cellular pathological phenotypes10,15,20,21,22. More recently, screening platforms based on WT hiPSC-CMs have been developed, in response to the need for human systems to the early phases of drug discovery23,24,25 as rodent cardiomyocytes differ profoundly from humans in ion channel expression and biophysics26.
For this purpose, technologies suitable for medium- to high-throughput application are being developed and implemented. These include optical recordings of membrane potential, Ca2+ transients and strain, impedance measurements (as an indirect measure of cell contractility), and extracellular field potential (FP) measurements (for review see reference24). Multi-electrode Arrays (MEA) devices allow recording of the electrical waveform signals (or FPs) generated and shaped by monolayers or small clusters of cardiomyocytes. FP contour correlates with the cardiac action potential and, to some extent, with the electrocardiogram (ECG) recordings27; they typically show an initial rapid upstroke corresponding to the Na+ influx and membrane depolarization (R/Q peak), a slow wave/plateau phase likely corresponding to the Ca2+ influx, and a repolarization phase corresponding to a predominant K+ efflux (T peak). Perturbation of the FP waveform can be correlated with changes in specific action potential phases28.
Although patch clamp recordings of action potentials could be more informative, especially for parameters as upstroke velocity and resting membrane potential, manual measurements are not feasible for experiments at medium- and high-throughput scale, while automated patch clamp has only recently been applied to hPSC-CMs29. However, since prolonged recordings on MEAs allow both acute and chronic exposure to compounds to be studied, it is now possible to use hPSC-CM platforms for drug screening, discovery24,30 and for safety pharmacology31,32. This holds the promise of future precision or personalized medicine33.
The purpose of this protocol is to provide the necessary information for dissociating and plating hPSC-CMs on MEA chips and measuring their FP. In this procedure, each step has been optimized, ensuring optimal cell survival and recovery after dissociation, optimal cell attachment to the MEA plate and standardized analysis and quantification of parameters. In particular, the procedure for extracellular FP recording, analysis of QT and RR intervals, and evaluation of drug effects are explained and exemplified.
1. Preparation of Solutions and Reagents
2. Sterilization of MEA Chips (Figure 1A)
NOTE: Several different configurations of the MEAs are available, with single- or multi-well formats. The protocol described here uses the single chamber MEA containing 60 recording electrodes in an 8 x 8 grid arrangement (see Table of Materials). The electrode diameter is 30 µm and the distance between electrodes is 200 µm. One reference electrode is also present.
3. Coating of MEA Chips (Figure 1B and 1C)
4. hPSC-CMs Dissociation and Plating (Figure 2)
NOTE: The protocol here described makes use of hPSC-CMs that were differentiated in a monolayer culture using cytokines34 at ~18 days after starting the differentiation. However, it has been proven to be suitable for any 2D and 3D hPSC-CM culture. When using differentiated cultures at earlier or later time points, adjusting the incubation time of the dissociation enzyme (see Table of Materials) might be necessary. The following volumes are meant for a single well of a 12-well plate format (3.8 cm2).
5. Ring Removal and Medium Refreshment (Figure 3)
6. Check Signal Quality (Figure 4)
7. Start Experiment and Recording
8. MEA Cleaning for Reuse
9. Data Export
10. Data Analysis
One day after dissociation and plating, the layer of hPSC-CMs will be visible as a dense and white film covering the center of the MEA chamber (Figure 3A). After removal of the ring (Figure 3B), the layer should remain in place and inspection on a light microscope will show the MEA electrodes covered by the (contracting) hPSC-CM layer (Figure 3C). Because of physical and electrical coupling of the cells, only one electrode (golden electrode) will be used for analysis.
Alternatively, when working with 3D structures, such as embryoid bodies or microtissues, these can be plated so that they are physically and electrically uncoupled. Visual inspection at the microscope could confirm no physical connection between the clusters and nonsynchronized R waves at MEAs confirm no electrical coupling. In this case, multiple independent electrodes can be analyzed.
Typical recordings of FP traces are shown in Figure 4. In particular, a good quality trace can be defined by the presence of a clear peak corresponding to the Na+ influx and membrane depolarization (R/Q peak), a clear repolarization phase corresponding to K+ efflux (T peak), and a high signal to noise ratio (Figure 4A, left: note y-axis scale and Figure 4B). Bad quality traces (Figure 4A, middle) may be the result of failure of hPSC-CMs to attach to the MEA plate or of weak hPSC-CM electrical activity. Waiting 1-3 days for better attachment may improve the signal; however, if no signal improvement is visible, excluding this MEA from experiments is recommended. Noisy traces (Figure 4A, right) may be analyzed after filtering.
Successful analysis of RR interval can be identified by visual inspection of the screen showing peak detection (Figure 5A, 5B). Within the time interval defined by the vertical cursors, blue marks corresponding to each peak should be present. In case the program does not identify one or more peaks, try to move the horizontal cursor and run the analysis again or adjust the detection settings. Similarly, successful analysis of QT interval can be identified by visual inspection of the screen showing FP detection (Figure 7). Within the time interval defined by the vertical cursors, blue marks corresponding to each FP detection should be present. In case the program does not identify one or more FPs, try to redefine the FP template (Figure 6) or adjust the detection settings and run the analysis again.
Extracted individual FP traces or their average (Figure 8) can be used for obtaining QT interval values with specific settings as in Figure 9. Analysis of QT-RR relationship is advisable and is meaningful in diseased and WT hPSC lines (Figure 10A) and to evaluate the need and/or the effect of QT-interval corrections (Figure 10B-10D). hPSC-CMs carrying LQTS-causing mutations have prolonged QT intervals compared with WT controls (Figure 11A). Treatment of hPSC-CMs with a hERG blocker results in QT interval prolongation (Figure 11B); conversely, treatment with a hERG activator results in shortening of the QT interval (Figure 11C). Finally, treatment with drugs affecting the beating frequency of hPSC-CMs should be visible as a change in the RR interval (Figure 11D, RR interval shortening).
Figure 1: Sterilization and Coating of MEA Chips. (A) Scheme representing the sterilization process including placing the MEA chip in an autoclavable glass Petri dish, wrapping up in aluminum foil, steaming for 6 min, and exposing to UV for 30 min. (B) Top (left panel) and side (middle panel) views of custom PTFE ring; the ring has an outer diameter of 1.2 cm, including 4 flaps that allow its positioning in the center of the MEA chip and the inner diameter is 0.4 cm (right panel). (C) Schematic representing preparation of the MEA chip including placing the chip in a standard plastic Petri dish, adding 8 mL of deionized water outside the MEA chamber, placing the PTFE ring in the middle of the chamber, and coating the electrode array with fibronectin. After incubation time, fibronectin is removed and replaced with culture medium. Please click here to view a larger version of this figure.
Figure 2: Dissociation and Plating of hPSC-CMs. Schematic representing the processes of hPSC-CMs enzymatic dissociation, centrifugation, resuspension, and plating on the center of the MEA chamber. Please click here to view a larger version of this figure.
Figure 3: MEA Chips with hPSC-CM Layer. (A) Top views of MEA chip containing the ring and the layer of hPSC-CMs. (B) side view of MEA chip after ring removal. (C) Bright field image of hPSC-CMs layer plated on the micro-electrode array; 4X magnification. Please click here to view a larger version of this figure.
Figure 4: MEA Recording of hPSC-CMs. (A) Representative traces recorded with the MEA showing a good quality trace with R/Q and T peaks clearly visible with high signal to noise ratio (left), a bad quality trace without clearly visible R/Q and T peaks (middle), and a noisy trace with R/Q and T peaks clearly visible but with low signal to noise ratio (right). (B) Representative examples of good quality FP traces with different morphologies that can be recorded during MEA experiments using hPSC-CMs. The shaded area represents the QT interval measured during the analysis. Since the FP at MEA resembles the first derivative of the action potential28, we have calculated the integral of the FP trace, shown as red dotted line, as theoretical demonstration of a T wave choice close to a complete action potential repolarization. Please click here to view a larger version of this figure.
Figure 5: RR Interval Calculation. (A) Example of RR interval analysis with automatic peak detection (top) and data extraction (bottom) using the analysis software (see Table of Materials). Vertical cursors identify the time interval of interest and the horizontal cursor is crossing all the events that are detected and identified by blue marks. (B) The magnified column shows the extracted data used to calculate the RR interval. Please click here to view a larger version of this figure.
Figure 6: Creation of FP Template. Example of template selection using vertical cursors placed before and after a single FP. This template is used for FP automatic identification. Please click here to view a larger version of this figure.
Figure 7: FP Template Automatic Identification. Example of template search throughout an interval defined by two vertical cursors. All the detected events are identified by blue marks and are automatically superimposed in the inset. Please click here to view a larger version of this figure.
Figure 8: Quantification of FP Parameters. Analysis of the saved events, with the R peak manually identified within the first two cursors and the T peak manually identified within the last two cursors. Please click here to view a larger version of this figure.
Figure 9: Analysis Window. Parameters used in the Statistics window to detect R peak. For the detection of the T peak, change cursors and (if necessary) polarity. Please click here to view a larger version of this figure.
Figure 10: QT-RR Relationship. (A) Example of different relationship between QT and RR intervals between WT (LQT1corr) and Long QT syndrome type 1 (LQT1R190Q) hPSC-CMs. Shaded areas show that the same shift in RR interval generates a bigger shift in the QT interval of the LQT1 diseased line, thus likely increasing arrhythmia susceptibility. (B) Relationship between uncorrected QT and RR intervals measured at MEA in CMs from 7 different hPSC lines. The effect of QT correction for Bazett's (C) or Fridericia's (D) formulae is shown and visible as change in the slope of the QT-RR interval relationship. Figures adapted from reference30. Please click here to view a larger version of this figure.
Figure 11: Disease- or Drug-induced QT and RR Interval Variations. (A) Example of QT interval prolongation in hPSC-CM derived from a patient carrying a Long-QT Syndrome (LQTS) mutation compared to its isogenic WT control. (B) Example of QT interval prolongation in hPSC-CM upon pharmacological hERG block. (C) QT interval shortening upon treatment with increasing doses of hERG activator. Arrow indicates the direction of the shortening. (D) Example of drug-induced RR interval shortening. Panel (C) was adapted from reference30. Please click here to view a larger version of this figure.
Low-insulin, BSA, polyvinylalcohol, essential lipids (LI-BPEL) Medium | |||
Component | Quantity for 100 mL | ||
IMDM | 43 mL | ||
F12 | 43 mL | ||
Ascorbic Acid 2-phosphate (5 mg/mL in distilled water) | 1 mL | ||
Cell culture supplement (direct substitute for L-glutamine) | 1 mL | ||
Penicillin/Streptomycin | 0.5 mL | ||
Phenol Red | 1 mg | ||
Protein Free Hybridoma Medium-II (PFHMII) | 5 mL | ||
BSA (10% wt/vol in IMDM) | 2.5 mL | ||
PVA (5% wt/vol in distilled water) | 2.5 mL | ||
Chemically Defined Lipid Concentrate (CDLC) | 1 mL | ||
Insulin-Transferrin-Selenium-Ethanolamine (ITS-X) 100X | 0.1 mL | ||
α-Monothioglycerol (13 µL in 1 mL IMDM) | 0.3 mL | ||
Combine the reagents, filter with a 0.22 µm pore filter and store the medium at 4 °C for up to 2 weeks. |
Table 1: Li-BPEL Medium Composition.
This protocol shows how to dissociate and prepare hPSC-CMs for measuring their FP using MEAs. hPSC-CMs usually display spontaneous electrical activity, which can be measured as FP and can provide meaningful data with respect to beating frequency, QT interval duration, and arrhythmic events.
Dissociation of 2D cardiac differentiated cultures is needed for recreating a beating layer on the MEA and it represents a critical step. Mechanical stress by repeated pipetting and/or aggressive dissociation enzyme treatments may result in high cell mortality, failure to attach to the MEA plate, and lack of spontaneous electrical activity. This protocol has been optimized for monolayer cultures. However, a similar approach can be used for three-dimensional (3D) cultures (e.g., embryoid bodies or EBs) with minor modifications, such as collection of the EBs followed by PBS wash and longer incubation time with the dissociating enzyme. Importantly, in both 2D and 3D differentiated cultures, the older the differentiated cells, the longer incubation time required might be to detach the cells because of increased extracellular matrix deposition.
The protocol described here for quantifying FP parameters can be used to generate dose-response curves for cardioactive drugs. As recently described by Cavero et al.31, the starting concentration of a drug might profoundly affect the outcome of a MEA measurement. Therefore, to improve accuracy and reliability of the results, we suggest the following: 1) in case of irreversible activators/blockers, use relatively large volumes of medium containing the drug to be tested. More in detail, remove 10-50% of medium volume from the MEA chip and add an equal volume of medium in which the drug was previously dissolved at the appropriate concentration. In this case, for calculating the final drug concentration, it is critical to consider the change in concentration after medium removal. 2) In case of reversible activators/blockers, add 10 µL of each drug dose from a 100X stock solution.
The majority of the cardiac differentiation protocols results in a variable mixed population of nodal-like, atrial-like and ventricular-like cardiomyocytes, with the ventricular type being the most represented. This might constitute a limitation when modeling cardiac diseases affecting a specific cardiomyocyte subtype or drugs acting on cardiac subtype-specific ion channels. Although several studies have optimized conditions to direct more controlled specification during cardiac differentiation3,5,37,38, their broader applicability is still under investigation.
Furthermore, variable efficiency of differentiation (in different experiments and in different hPSC lines) might be observed39,40,41,42,43,44. Cardiomyocyte-enriching strategies based on surface protein expression35,45 (by florescence-assisted cell sorting or by magnetic-bead selection46,47), and metabolic selection44,48 may represent valid strategies that can be applied to any (genetically modified or unmodified) hPSC-line prior plating of the hPSC-CMs, to improve the electrical signal.
Although hPSC-CMs are notoriously immature as compared to human adult cardiomyocytes4,49, they have proven to be valuable in recapitulating and identifying specific disease-related changes (e.g., in channelopathies)19,20,50 and drug-induced responses (e.g., cardiac ion channel blockers)4,51. Furthermore, immature cells are easier to dissociate, and recover better than adult cardiomyocytes after dissociation and plating44 therefore, hPSC-CM immaturity may be rewarded as an advantage in this respect. However, to be able to recapitulate e.g. late onset cardiac diseases and faithfully reproduce drug responses of adult cardiomyocytes, a more mature mechanical, metabolic, and electrical hPSC-CM state should be obtained. Methods to mature these cells include prolonged time in culture52, mechanical strain53, electrical pacing54, addition of small molecules55, 3D-culture56, co-culture with other cell types57, and even a combination of these approaches58; to date, none of these approaches has led to an adult-like phenotype.
As part of the immaturity features, hPSC-CMs show electrical automaticity. Here, details are provided on how to accurately quantify QT and RR intervals. One limitation of measuring spontaneous electrical activity is that comparison of QT intervals may be difficult when hPSC-CMs display different beating frequencies. In this case, Bazett's or Fridericia's formulae can be used to correct the QT interval for the frequency. However, as previously reported30, we strongly recommend performing Major-Axis regression analysis by plotting QT interval versus RR interval for both raw and corrected data, to exclude any possible bias due to the correction method itself.
The protocol presented here, together with previously described methods59,60 helps the standardization of the procedures and the analysis of hPSC-CM FPs, improving data reproducibility and allowing a better comparison of inter-laboratory results.
The authors have nothing to disclose.
This work was supported by the following grants: CVON (HUSTCARE): the Netherlands CardioVascular Research Initiative (the Dutch Heart Foundation, Dutch Federation of University Medical Centres, the Netherlands Organisation for Health Research and Development and the Royal Netherlands Academy of Sciences); the European Research Council (ERCAdG 323182 STEMCARDIOVASC). We thank E. Giacomelli (LUMC) for help with hPSC cardiac differentiation.
Biological safety cabinet/laminar flow hood | Cleanair | ||
MEA2100 in vitro recording system | Multi Channel Systems | ||
CO2 cell culture incubator | Sanyo | MCO-15A | |
Centrifuge | Hitachi | himac-CT6EL | |
Leica stereomicroscope | Leica Microsystems | MS5 | |
Handheld pipetman (P-10 (10μL), P-200 (200μL), P-1000 (1000μL)) | Gilson International | ||
Filter tips (10 μL, 200μL, 1000μL) | Corning | 4807 (10 μL), 4810 (200μL), 4809 (1000μL) | |
Disposable bottle top filter (0.22 μm pore size) | Millipore | SCGVU02RE | |
Sterile plastic pipette | Greiner Bio-One | 606180 (5 mL), 607180 (10 mL), 760180 (25 mL) | |
Tweezers | Dumont | ||
Autoclavable Petri dishes | VWR/ Duran Group | 391-0860 | |
MEA chip | Multi Channel Systems | MEA200/30iR-Ti-gr | |
Phosphate-Buffered Saline (PBS) calcium, magnesium | Thermo Fisher Scientific | 14040-091 | |
Phosphate-Buffered Saline (PBS) no calcium, no magnesium | Thermo Fisher Scientific | 14190-169 | |
Human recombinant fibronectin | Tebu-Bio | J64560 | |
Custom made polytetrafluoroethylene (PTFE) MEA rings | LUMC: department of Instrument Development | ||
Dissociation enzyme – TrypLE Select 1X | Thermo Fisher Scientific | 12563-029 | |
Tergazyme enzyme detergent | Sigma-Aldrich | Z273287-1EA | |
Distilled Water | Thermo Fisher Scientific | 15230-089 | |
Name | Company | Catalog Number | Comments |
Reagents for LI-BPEL medium | |||
IMDM | Thermo Fisher Scientific | 21056-023 | |
F12 | Thermo Fisher Scientific | 31765-027 | |
Ascorbic Acid 2-phosphate | Sigma-Aldrich | A8960 | |
Glutamax | Thermo Fisher Scientific | 35050-038 | |
Penicillin/Streptomycin | Thermo Fisher Scientific | 15070-063 | |
Phenol Red | Sigma-Aldrich | P3532 | |
Protein Free Hybridoma Medium-II (PFHMII) | Thermo Fisher Scientific | 12040-077 | |
Bovine Serum Albumin (BSA) | Bovogen Biologicals Australia | BSAS05 | |
Poly(vinyl alcohol) (PVA) | Sigma-Aldrich | P8136 | |
Chemically Defined Lipid Concentrate (CDLC) | Thermo Fisher Scientific | 11905-031 | |
Insulin-Transferrin-Selenium-Ethanolamine (ITS-X) 100X | Thermo Fisher Scientific | 51599-056 | |
α-Monothioglycerol | Sigma-Aldrich | M6145 | |
Name | Company | Software version | Comments |
MC_Rack | Multi Channel Systems | 4.6.2 | Alternatively, data can be recorded using Cardio2D or MC_Experimenter (MultiChannel Systems) |
TCX Control | Multi Channel Systems | 1.3.4 | |
MEA Select | Multi Channel Systems | 1.3.0 | |
MC_Data Tool | Multi Channel Systems | 2.6.15 | Alternatively, Multi Channel Data Manager (MultiChannel Systems) can be used when custom data export is required (HDF5, EDF, etc.) |
Clampfit | Molecular Devices | 7.0.0 | Used in step 10.1 for analyzing, graphing, and formatting of all of data. To use Clampfit, download and install the electrophysiology data acquisition and pClam (latest version, 10.7.0), available on the Molecular Devices Website. Once complete, launch the software Clampfit. Alternatively, data can be analysed using Cardio2D (MultiChannel Systems) or MatLab custom code. |