Here, we show the generation of human engineered heart tissue from induced pluripotent stem cells (hiPSC)-derived cardiomyocytes. We present a method to analyze contraction force and exemplary alteration of contraction pattern by the hERG channel inhibitor E-4031. This method shows high level of robustness and suitability for cardiac drug screening.
Cardiac tissue engineering describes techniques to constitute three dimensional force-generating engineered tissues. For the implementation of these procedures in basic research and preclinical drug development, it is important to develop protocols for automated generation and analysis under standardized conditions. Here, we present a technique to generate engineered heart tissue (EHT) from cardiomyocytes of different species (rat, mouse, human). The technique relies on the assembly of a fibrin-gel containing dissociated cardiomyocytes between elastic polydimethylsiloxane (PDMS) posts in a 24-well format. Three-dimensional, force-generating EHTs constitute within two weeks after casting. This procedure allows for the generation of several hundred EHTs per week and is technically limited only by the availability of cardiomyocytes (0.4-1.0 x 106/EHT). Evaluation of auxotonic muscle contractions is performed in a modified incubation chamber with a mechanical interlock for 24-well plates and a camera placed on top of this chamber. A software controls a camera moved on an XYZ axis system to each EHT. EHT contractions are detected by an automated figure recognition algorithm, and force is calculated based on shortening of the EHT and the elastic propensity and geometry of the PDMS posts. This procedure allows for automated analysis of high numbers of EHT under standardized and sterile conditions. The reliable detection of drug effects on cardiomyocyte contraction is crucial for cardiac drug development and safety pharmacology. We demonstrate, with the example of the hERG channel inhibitor E-4031, that the human EHT system replicates drug responses on contraction kinetics of the human heart, indicating it to be a promising tool for cardiac drug safety screening.
Cardiac side effects such as the drug-induced long QT syndrome have led to market withdrawals over the past years. Statistics indicate that about 45% of all withdrawals are due to unwanted effects on the cardiovascular system1. This drug failure after the expensive developmental process and approval is the worst-case scenario for pharmaceutical companies. Research and development departments therefore focus on detection of such unwanted cardiovascular effects early on. For economic and ethical concerns, efforts to reduce animal experiments and replace them with new in vitro screening assays are ongoing.
A set of established assays are included in the United States Food and Drug Administration (FDA) and European Medicines Agency (EMA) guidelines for preclinical evaluation of proarrhythmic drug effects2. The technology of reprogramming somatic cells followed by differentiation of human induced pluripotent stem cells (hiPSC) boosted this research field3. It now offers the possibility to screen new drug candidates on human cardiomyocytes in vitro and avoids issues with inter-species differences. Recent cardiac differentiation protocols4,5 provide unlimited supply of cardiomyocytes without ethical concern. However, the measurement of contractile force, the most important and best characterized in vivo parameter of cardiomyocytes, is not well established. This is related to the relative immaturity6 of human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CM) as compared to the adult cardiomyocyte. A possible advancement is to engineer 3-dimensional heart tissue from single cells7 (engineered heart tissue, EHT). The EHT protocol is based on embedding single murine or human cardiomyocytes8,9,10 in fibrin hydrogel between two flexible polydimethylsiloxane (PDMS) posts11 in 24-well format. Within a few days the cardiomyocytes start to contract spontaneously as single cells and start to form cellular networks. After 7-10 days, macroscopic contractions of the entire tissue are visible. During this process the extracellular matrix is remodeled, which leads to a decrease of diameter and length. The shortening of the EHT results in bending of the PDMS post even during rest, subjecting cardiomyocytes in the developing EHT to continuous load. EHTs continue to perform auxotonic muscle contractions over several weeks. Human EHTs show responses to physiological and pharmacological stimulation indicating their suitability for drug screening and disease modeling7.
In this manuscript we present a robust and easy protocol for the generation of human EHT, and the automated contractility analysis of concentration dependent changes of the contraction pattern in the presence of hERG channel inhibitors.
NOTE: The following steps describe a cell culture protocol. Please perform under sterile conditions and use pre-warmed media.
1. Cardiac Differentiation of hiPSC
2. Generation of Engineered Heart Tissue (EHT)
3. Contraction Analysis
NOTE: The contraction analysis is based on video-optical recording in a commercially available EHT analysis instrument (see Table of materials). The central unit of this instrument is an incubation chamber (Figure 3A). The software provided with the instrument calculates contraction force based on deflection of the PDMS posts with known elastic modulus and geometry11 (Supplementary Figure 1).
Cardiac Differentiation and Preparation of EHT
HiPSC were expanded on reduced growth factor basement membrane matrix, dissociated with EDTA and embryoid bodies (EBs) formed in spinner flasks overnight. After mesodermal induction for three days, cardiac differentiation was initiated with the Wnt inhibitor. After ~17 days of differentiation protocol, beating EBs were dissociated into single cells with collagenase type II (Figure 1). EHTs were prepared from freshly isolated cardiomyocytes with 1.0 x 106 cells per 100 µL tissue immediately after EB dissociation (Figure 2). For preparation from frozen cells (e.g. commercial suppliers), cells were thawed by drop-wise resuspension in medium and resuspended in ECM after centrifugation. Within 48 h, spontaneous beating of single cells was observed microscopically in the EHT. Over the next days, cardiomyocytes spread out longitudinally along the force lines of the tissue, coupled and formed a coherently beating EHT. The macroscopically visible deflections of the PDMS posts were measured with the EHT analysis instrument.
Contraction Analysis
Analysis of EHT contractility was based on video-optical recording (Figure 3). Within the system, the 24-well plate with EHTs was positioned in the gas- and temperature-controlled incubator chamber with a glass roof. A camera connected to a motorized XYZ-axis was positioned above this unit, and XYZ-coordinates were defined for each EHT with a software program. Analysis of each EHT was based on figure recognition at the top and bottom ends of the tissue. An automated software algorithm analyzed the filmed EHT contractions and calculated force development based on post deflection and elasticity/geometry of the PDMS posts (Supplementary Figure 1)11. After determination of peaks according to predefined criteria, a pdf report summarized parameters of contractility for all EHTs (Supplementary Figure 2).
Stability in Long Term Measurement
Long term measurement of EHTs was performed in the continuous mode with automated repeated measurement every 20 min. Contraction analysis showed no time-dependent changes (Post test for linear trend was not significant) in contraction force, beating frequency, contraction time T1 or relaxation time T2 for up to 20 h, indicating a high level of stability (Figure 4). As with every biological sample though, the EHTs show some level of variability, as expected for biological replicates. The variation coefficients at beginning and end of long term measurement were 7/13% for force, 12/6% for frequency, 4/3% for contraction time and 4/3% for relaxation time, respectively (n = 4). This variability between biological replicates is not an artifact of figure recognition or analysis as the results for each single EHT are reproducible and have low variability (Supplementary Figure 2).
Drug Screening
Drug screening was performed in serum-free Tyrode's solution. Cumulative concentration response curves were generated with a maximum DMSO vehicle concentration of 0.1%. Baseline contraction analysis showed very little irregularity in beating pattern and very low variability between the replicates. Exposure to the hERG channel blocker E4031 resulted in a significant prolongation of the relaxation time (T2) at 10 nM and irregular beating pattern at higher concentrations (Figure 5). For frequency controlled contraction analysis, measurements were optionally performed with electrical stimulation using carbon electrodes.
Figure 1: Cardiac differentiation. (A) Human induced pluripotent stem cells. Scale bar = 100 µm. (B) Embryoid bodies at the beginning of mesodermal differentiation. Scale bar = 100 µm. (C) Contracting embryoid bodies at the end of cardiac differentiation. Scale bar = 100 µm. (D) Human engineered heart tissue. Scale bar = 1 mm. Please click here to view a larger version of this figure.
Figure 2: Generation of Engineered Heart Tissue (EHT). (A) PTFE spacer. (B) PDMS rack. A, (B) Scale bar = 1 cm (C) Top view on a well of a 24-well-plate with casting mold in agarose after removal of the PTFE spacer. (D) Pair of posts from the PDMS rack positioned in the casting mold. (E) Reconstitution mix pipetted into the casting mold and around the PDMS posts. (F) Freshly generated EHT at day 0, transferred to a new culture dish with medium. (G) Remodeled EHT in medium at day 15. Please click here to view a larger version of this figure.
Figure 3: Contraction Analysis of Engineered Heart Tissue (EHT). (A) EHT analysis instrument with computer-controlled camera above the gas- and temperature-controlled incubation chamber with EHTs in 24-well-culture dish on top of a LED panel. (B) Live view of an EHT during analysis with the automated contraction analysis software. (C) Exemplary contraction pattern displaying contraction force over time measured with 100 frames per second and enlarged schematic contraction peak, displaying the analysis parameter force, contraction time (T1), relaxation time (T2), contraction velocity (CV), relaxation velocity (RV; this figure has been reprinted from7 and27). D-H: Parameters for contraction analysis. (D) Examples of different filter levels (blue filtered line) affecting the figure recognition. Left: Filter level = 0. Note that no blue filtered line is visible, but only the red original trace. Middle: Filter level = 3, indicating ideal recognition/analysis of the contraction peaks. Right: Filter level = 20. Note that the blue filtered lines are much smaller than the original traces indicating poor contraction peak analysis. (E) Example of arrhythmic EHT contractions analyzed with a too high force threshold (left) and lower force threshold (right) with all peaks recognized. (F) Examples traces of EHTs exposed to the sodium channel agonist ATX-II resulting in a prolonged relaxation time with early after contractions. Note that the fourth contraction peak on the left panel is not recognized (no green square) due to a too low minimum factor. (G) Examples of contraction peaks with (right panel) and without (right panel) pink curve (first derivative of contraction curve, contraction and relaxation velocity). (H) Summary of parameter recommendations for murine and human EHTs. Note that these parameters might need adjustment as contraction pattern changes during drug exposure (see E, F). Please click here to view a larger version of this figure.
Figure 4: Time Control During Long Term Measurement of Spontaneously Beating Human Engineered Heart Tissue (EHT). Time control during long term measurement of spontaneously beating human engineered heart tissue (EHT). EHTs were positioned in the EHT analysis instrument and incubated for a period of 20 h. Automated contraction analyses were performed every 20 min with not significant post-test for linear trend (n = 4; data represent mean + standard deviation). Please click here to view a larger version of this figure.
Figure 5: Effect of hERG Channel Blocker E4031 on Human Engineered Heart Tissue (EHT).
(A) Exemplary contraction pattern at baseline. (B) Exemplary contraction pattern after incubation in 300 nM E4031. (C) Average contraction peak (n = 4) of electrically stimulated EHTs at baseline (black) and after 30 min incubation in 100 nM E4031 (red). Analysis with 100 frames per second. (D) Concentration response curves for E4031 on spontaneously beating human EHTs. Note the smaller effect size in T2 prolongation in the electrically stimulated EHTs (C) vs. the spontaneously beating EHTs (D). One-way ANOVA, repeated measures with Dunnett's post-test vs. baseline; * p <0.05; n= 4. Z' factor25 for T2 prolongation at 300 nM E4031 is 0.75. Please click here to view a larger version of this figure.
Human | Rat | Mouse | |
Cells | 26.4×106 | 10.8×106 | 18.0×106 |
2xDMEM [µL] | 147 | 147 | 147 |
Basement membrane matrix [µL] | 264 | – | 264 |
Y-27632 [µL] | 2.6 | – | – |
Fibrinogen [µL] | 66.8 | 66.8 | 66.8 |
ECM [µL] | ad 2640 | ad 2640 | ad 2640 |
Table 1: Pipetting Scheme for Preparation of Reconstitution Mix for 24 Human, Rat and Mouse EHTs. Fibrinogen concentration is 5 mg/mL for all EHTs. Note that cell concentration is different for each species. For human EHTs, add basement membrane matrix (e.g. matrigel, see Materials Table) and Rho-kinase inhibitor Y-27632 to the mixture. For mouse EHTs, add basement membrane matrix. Depending on the initial concentration of the cell suspension, add additional ECM to reach a total volume of 2,640 µL reconstitution mix.
Supplementary Figure 1: Calculation Formula for Contraction Force Based on Post Deflection26. Please click here to view a larger version of this figure.
Supplementary Figure 2: Exemplary Data Excerpt from the Contraction Analysis Report of the Software.
Top: Contraction peaks over 20 s of time recorded with 100 frames/s (red: contraction force; pink: contraction velocity). Bottom: Analyzed contraction parameter showing minimum, mean, maximum value and standard deviation of the 18 analyzed contraction peaks marked with a green square (top). Please click here to view a larger version of this figure.
Engineered heart tissue offers a valuable option to the tool box of cardiovascular research. EHTs in the 24-well format have proven valuable for disease modeling8,14, drug safety screening7,8,10,11,15, or basic cardiovascular research16,17.
Potential modifications of the basic EHT protocols include studies on cellular interaction with defined mixtures of cardiomyocytes and non-cardiomyocytes18 and afterload enhancement14. Important aspects for troubleshooting include the quality of the input cell population and the quality of fibrinogen. If the initial gel formation is irregular or not compact enough to stick to PDMS posts, or if the gel is rapidly degenerated within a few days after casting due to low aprotinin concentration, EHTs generation will not be successful.
The following limitations have to be considered. (i) The throughput in 24-well format is not well suited for a primary screening setup, but is sufficient for a secondary confirmation screen. (ii) The integrated readout of force, contraction kinetics and rhythm have low specificity, which makes it difficult to link changes in force to specific molecular mechanisms. (iii) While the 3D culture format and the stretch imposed by bent PDMS posts improve morphological and functional aspects of maturation, the lack of true intercalated discs, spontaneous beating, and lower than normal adrenergic responses indicate that full cardiac maturation is not yet achieved with this protocol. (iv) PDMS has been shown to absorb different drugs to various extend19 and the actual effective concentration on the EHTs might vary. This bias might be more prominent in chronic toxicity studies15 than those investigating acute drug effects7,8 and clearly depends on the drug at hand. As a consequence, any PDMS rack exposed to drugs should be discarded after use and the potential absorption of different drugs will need to be taken into consideration when analyzing the data. (v) Given the fact that 1×106 hiPSCs are required per EHT the costs per data point are high.
The significance of this protocol compared to ex vivo heart contractility analysis is the potential to work in a human system, the lack of experimental run-down, and the standardized production of EHTs which allows setting up large studies on specific genetic backgrounds (e.g. disease modeling). In comparison to other hiPSC test systems (multi electrode array, impedance), this system analyzes contractile force as the most important physiological parameter of cardiomyocytes in a three-dimensional structure and allows cultivation under defined load. EHTs contract regularly and at stable force, frequency and rhythm over many hours and weeks, allowing long-term measurements and evaluation of chronic toxicity. EHT analysis can be performed under electrical stimulation which is important with regards to the force-frequency relationship of muscle contraction and its potential influence on drug effects. With regard to the immature status of the hiPSC-CM, electrical stimulation has the potential to push the cells towards a more mature phenotype20,21 and could be included in EHT maintenance. In addition, the EHT test system provides high content readout including all standard histological parameters as well as biochemistry (RNA, DNA, protein isolation)11. Potential future applications include aspects of drug development (target validation, safety pharmacology) and disease modeling in particular in combination with CRISPR/Cas9 mediated generation of isogenic controls.
During maintenance of EHTs and performance of concentration-response curves, it is important to carefully follow instructions for sterile cell culture procedures to avoid fungal contaminations during transfer of PDMS racks between 24-well plates. The most critical step with EHT though, is the cardiomyocyte differentiation protocol. For the EHTs to form coherently beating muscle strips, the input population has to consist of at least 60% cardiomyocytes. Fortunately, the differentiation protocols became more robust and efficient over the years and cardiomyocytes are readily commercially available by now. The role of non-myocytes in the development of hiPSC-derived engineered heart tissue is not fully understood, yet. Experiments with hiPSC-CM of high purity led to surprisingly good EHT7, but integration of other cell types into the tissue has been shown to improve tissue function and might push the phenotype towards a mature physiology22,23,24. As the tissues are generated in a 24-well-format with one million cells per tissue, this platform is still rather expensive. Scaling down to a 96-well format is aspired, but so far robustness outbalances this shortcoming. The protocol presented in this manuscript is technically very robust with little variability between the replicates. The intra-batch variability is very low for the kinetics (mean coefficient of variation = 5%), and somewhat large for the spontaneous beating frequency (mean coefficient of variation 10%) and contraction force (mean coefficient of variation 17%). The technique is very robust. Across different experimenters, the efficiency to generate beating EHTs is >90% of casted hydrogels.
The authors have nothing to disclose.
The authors are grateful to Alessandra Moretti and Dennis Schade for their kind contribution of material. We acknowledge the great support of the iPS and EHT working group at the Department of Experimental Pharmacology and Toxicology of the UKE. The work of the authors is supported by grants from the DZHK (German Centre for Cardiovascular Research) and the German Ministry of Education and Research (BMBF), the German Research Foundation (DFG Es 88/12-1, HA 3423/5-1), British National Centre for the Replacement Refinement & Reduction of Animals in Research (NC3Rs CRACK-IT grant 35911-259146), the British Heart Foundation RM/13/30157, the European Research Council (Advanced Grant IndivuHeart), the German Heart Foundation and the Freie und Hansestadt Hamburg.
EHT analysis intrument | EHT Technologies GmbH | A0001 | Software is included |
EHT PDMS rack | EHT Technologies GmbH | C0001 | |
EHT PTFE spacer | EHT Technologies GmbH | C0002 | |
EHT electrode | EHT Technologies GmbH | P0001 | |
EHT pacing adapter/cable | EHT Technologies GmbH | P0002 | |
24-well-plate | Nunc | 144530 | |
6 well-cell culture plate | Nunc | 140675 | |
15 ml falcon tube, graduated | Sarstedt | 62,554,502 | |
Cell scraper | Sarstedt | 831,830 | |
Spinner flask | Integra | 182 101 | |
Stirrer Variomag/ Cimarec Biosystem Direct | Thermo scientific | 70101 | Adjust rotor speed to 40 rpm |
T175 cell culture flask | Sarstedt | 831,812,002 | |
V-shaped sedimentation rack | Custom made at UKE Hamburg | na | |
10× DMEM | Gibco | 52100 | |
1-Thioglycerol | Sigma Aldrich | M6145 | |
2-Phospho-L-ascorbic acid trisodium salt | Sigma Aldrich | 49752 | |
Activin-A | R&D systems | 338-AC | |
Agarose | Invitrogen | 15510-019 | |
Aprotinin | Sigma Aldrich | A1153 | |
Aqua ad injectabilia | Baxter GmbH | 1428 | |
B27 PLUS insulin | Gibco | 17504-044 | |
BMP-4 | R&D systems | 314-BP | |
Collagenase II | Worthington | LS004176 | |
DMEM | Biochrom | F0415 | |
DMSO | Sigma Aldrich | D4540 | |
DNase II, type V (from bovine spleen) | Sigma | D8764 | |
Dorsomorphin | abcam | ab120843 | |
EDTA | Roth | 8043.2 | |
Fetal calf serum | Gibco | 10437028 | |
FGF2 | Miltenyi Biotec | 130-104-921 | |
Fibrinogen (bovine) | Sigma Aldrich | F8630 | |
Geltrex | Gibco | A1413302 | For coating: 1:200 dilution |
HBSS w/o Ca2+/Mg2+ | Gibco | 14175-053 | |
HEPES | Roth | 9105.4 | |
Horse serum | Life technologies | 26050088 | |
Human serum albumin | Biological Industries | 05-720-1B | |
Insulin, human | Sigma Aldrich | I9278 | |
L-Glutamin | Gibco | 25030-024 | |
Lipidmix | Sigma Aldrich | L5146 | |
Matrigel | BD Biosciences | 354234 | For EHT reconsitutionmix. |
N-Benzyl-p-Toluenesulfonamide | TCI | B3082-25G | |
PBS w/o MgCl2/CaCl2 | Biochrom | 14190 | |
Penicillin/Streptomycin | Gibco | 15140 | |
Pluronic F-127 | Sigma Aldrich | P2443 | |
Polyvinyl alcohol | Sigma Aldrich | P8136 | |
RPMI 1640 | Gibco | 21875 | |
Sodium selenite | Sigma Aldrich | S5261 | |
TGFß1 | Peprotech | 100-21 | |
Thrombin | Sigma Aldrich | T7513 | |
Transferrin | Sigma Aldrich | T8158 | |
Y-27632 | Biorbyt | orb6014 | |
hiPSC | Custom made at UKE hamburg | na | |
iCell cardiomyocytes kit | Cellular Dynamics International | CMC-100-010-001 | |
Pluricyte cardiomyocyte kit | Pluriomics | PCK-1.5 | |
Cor.4U – HiPSC cardiomyocytes kit | Axiogenesis AG | Ax-C-HC02-FR3 | |
Cellartis cardiomyocytes | Takara Bio USA, Inc. | Y10075 |