This protocol describes the procedure for sectioning and culturing human cardiac slices for preclinical drug testing and details the use of optical mapping for recording transmembrane voltage and intracellular calcium signals simultaneously from these slices.
Human cardiac slice preparations have recently been developed as a platform for human physiology studies and therapy testing to bridge the gap between animal and clinical trials. Numerous animal and cell models have been used to examine the effects of drugs, yet these responses often differ in humans. Human cardiac slices offer an advantage for drug testing in that they are directly derived from viable human hearts. In addition to having preserved multicellular structures, cell-cell coupling, and extracellular matrix environments, human cardiac tissue slices can be used to directly test the effect of innumerable drugs on adult human cardiac physiology. What distinguishes this model from other heart preparations, such as whole hearts or wedges, is that slices can be subjected to longer-term culture. As such, cardiac slices allow for studying the acute as well as chronic effects of drugs. Furthermore, the ability to collect several hundred to a thousand slices from a single heart makes this a high-throughput model to test several drugs at varying concentrations and combinations with other drugs at the same time. Slices can be prepared from any given region of the heart. In this protocol, we describe the preparation of left ventricular slices by isolating tissue cubes from the left ventricular free wall and sectioning them into slices using a high precision vibrating microtome. These slices can then either be subjected to acute experiments to measure baseline cardiac electrophysiological function or cultured for chronic drug studies. This protocol also describes dual optical mapping of cardiac slices for simultaneous recordings of transmembrane potentials and intracellular calcium dynamics to determine the effects of the drugs being investigated.
Animal models have been a valuable tool used for understanding the underlying mechanisms of human physiology and pathophysiology, as well as a platform for preliminary testing of therapies to treat various diseases1. Great strides have been taken in the field of biomedical research based on these animal studies2. However, significant interspecies differences exist between human and animal physiologies, including mice, rats, guinea pigs, rabbits, sheep, pigs, and dogs3,4. As a result, there have been numerous drug, gene, and cell therapies that showed promise during the animal testing stage but failed to live up to the results in clinical trials5. To bridge this gap, isolated cardiac myocytes and human induced pluripotent stem cells (iPSCs) were developed as models to test the response of human physiology to various drugs and diseases6. Stem cell-derived cardiomyocytes have been widely used in organ-on-a-chip systems as a surrogate of the heart6,7,8. However, the usefulness of iPSC-derived cardiomyocytes (iPSC-CMs) is impeded by their relatively immature phenotype and the lack of representation of the cardiomyocyte subpopulation; the mature myocardium is a complex structure comprised of several coexisting cell types such as fibroblasts, neurons, macrophages, and endothelial cells. On the other hand, isolated human cardiomyocytes are electrically mature, and different cardiomyocyte subpopulations can be obtained by altering culturing parameters9. Still, these myocytes generally exhibit altered action potential morphologies due to the lack of cell-cell coupling, rapid de-differentiation, and occurrence of proarrhythmic behavior in vitro10,11. Some of the limitations were addressed by 3D cell culture models of iPSC-CMs and cardiac myocytes. These models, which include spheroids, hydrogel scaffold encapsulated 3D cultures, engineered heart tissues (EHTs), and heart-on-a-chip systems, use multiple cardiac cell populations such as cardiomyocytes, fibroblasts, and endothelial cells. They either self-assemble or assemble along a scaffold to form 3D structures, and some even reproduce the complex anisotropic nature of the myocardium. These models have been reported to have cells of mature phenotypes, contractile properties, and molecular profiles similar to cardiac tissue. The heart-on-a-chip system also allows the study of systemic effects in drug testing and disease models. However, in vitro cell-based models lack the native extracellular matrix and therefore cannot accurately mimic organ level electrophysiology. Human cardiac slices, by contrast, have an intact extracellular matrix and native cell-to-cell contacts, making them useful for more accurately examining arrhythmogenic properties of the human myocardium.
Researchers have developed human cardiac organotypic slices as a physiological preclinical platform for acute and chronic drug testing and to study cardiac electrophysiology and cardiac disease progression12,13,14,15,16,17,18,19. When compared with iPSC-derived cardiomyocytes, human cardiac slices more faithfully replicate adult human cardiac electrophysiology with a mature cardiomyocyte phenotype. When compared with isolated human cardiomyocytes, cardiac slices exhibit physiological action potential durations because of the well-preserved cell-cell coupling and the intrinsic existence of their native intra- and extracellular environments.
This protocol describes the process of generating human cardiac slices from whole donor hearts, performing acute (i.e., hours-long) and chronic (i.e., days-long) studies to test cardiac electrophysiology parameters via optical mapping. While this protocol describes only the use of the left ventricular (LV) tissue, it has been successfully applied to other regions of the heart as well as other species such as mice, rats, guinea pigs, and pigs14,20,21,22. Our laboratory uses whole human donor hearts that have been rejected for transplantation for the last 5 years, but it is feasible for these same procedures to be carried out on any donor heart sample tissues obtained by alternative means (e.g., left ventricular assist device [LVAD] implantations, biopsies, myectomies) as long as the tissues have the ability to be sectioned into cubes. Optical mapping is employed for analysis in this study due to its capacity to simultaneously map optical action potentials and calcium transients with high spatial (100 x 100 pixels) and temporal (>1,000 frames/s) resolution. Alternative methods can also be used, such as multielectrode arrays (MEAs) or microelectrodes, but these techniques are limited by their relatively low spatial resolutions. Additionally, MEAs were designed for use with cell cultures, and sharp microelectrodes are more easily managed for use with whole hearts or large tissue wedges.
The goal of the article is to enable more researchers to use human cardiac tissues for cardiac electrophysiology studies. It should be noted that the technology described in this article is relatively simple and beneficial for short-term studies (on the order of several hours to days). More physiological biomimetic culture for longer-term studies (on the order of weeks) has been discussed and described by a number of other studies12,18,23. Electrical stimulation, mechanical loading, and tissue stretching are advantageous conditioning mechanisms that can help limit the onset of in vitro tissue remodelling12,18,23.
All methods described have been performed in compliance with all institutional, national, and international guidelines for human welfare. Research was approved by the Institution Review Board (IRB) at The George Washington University.
NOTE: Donor human hearts were acquired from Washington Regional Transplant Community as deidentified discarded tissue with approval from the George Washington University IRB. Explanted hearts are cardioplegically arrested by flushing the heart with a solution of ice-cold cardioplegia (the blood was cleared from the heart in this process) and transferred to the lab under standard organ transplant conditions.
1. Preparation of solutions
2. Equipment setup
3. Slicing protocol
4. Slice culturing under static conditions
NOTE: To minimize the chance of contamination, sterilize the forceps using a bead sterilizer before each transfer step.
5. Functional characterization―optical mapping
6. Data processing with RHYTHM1.2
NOTE: RHYTHM1.2 is a MATLAB based user interface that is used to display, condition, and analyze optical mapping data acquired by single or dual camera optical mapping systems (Figure 3). It is used in conjunction with the imaging system.
Human organotypic slices were collected from the left ventricle of a donor human heart according to the protocol detailed above and illustrated in Figure 1. A dual camera optical mapping system like that in Figure 2 was used in the upright imaging configuration to perform simultaneous optical mapping of voltage and calcium about 1 h after the completion of the slicing protocol. Data were analyzed using RHYTHM1.2 (Figure 3), an open source optical mapping data analysis tool previously published by our laboratory and freely available on Github (https://github.com/optocardiography/Rhythm-1.2). The electrophysiological parameters measured are illustrated in Figure 4. The action potential and calcium transient traces were signal conditioned and representative traces used in further analysis are illustrated in Figure 4A. Activation times were determined for each pixel and isochronal maps of activation times determined from voltage and calcium traces are illustrated in Figure 4B,C. Note that activation in the calcium isochrone lags behind that of the voltage as expected. CV vectors plotted in Figure 4D were calculated using the activation times and the known interpixel resolution. The average CV in the transverse direction was determined to be 21.2 cm/s in this slice. This CV value is comparable to previously reported ventricular transverse CV measured from explanted whole human hearts (24 ± 4 and 28 ± 7 cm/s in hearts with diffuse and patchy fibrosis)28. The calcium transient decay constant was measured by fitting a polynomial to the decaying portion of the calcium traces and average decay constant was determined to be 105.3 ms in this slice (map in Figure 4E). Next, APD and CaTD were measured as the time duration between activation time and a specified percent of repolarization/calcium removal form the cytoplasm. Average APD80 and CaTD80 were determined to be 343.1 ms and 442.6 ms, respectively (Figure 4F,G). Previous in vivo human studies report the activation-recovery interval (ARI) measured from unipolar electrograms recorded from in vivo human hearts during steady state pacing at 1 Hz as a surrogate for APD. ARI values in these studies range from 250–450 ms29,30. The APD values from human cardiac slices reported here are comparable to the previous ARI values. Regions with motion artifacts were removed from APD and CaTD calculations. Finally, the rise times (i.e., the duration of the upstroke) of the voltage and calcium traces were measured and mapped. These are shown in Figure 4H and Figure 4I, respectively. The average values were determined to be 10.2 ms and 13.3 ms, respectively. The artifacts in the maps to the right of the point of pacing are due to the pacing wire within the field of view. Finally, the use of this slice model in acute drug testing was demonstrated when slices were cultured for 24 h with doxorubicin (DOX), a chemotherapeutic agent known to have cardiotoxic effects. Treatment of slices with 50 μM DOX resulted in a reduction of transverse conduction velocity from 19.4 ± 3.4 cm/s to 9.6 ± 3.2 cm/s, as illustrated by the activation maps in Figure 5A, CV vector maps in Figure 5B, and summary data in Figure 5C. The smaller sample size in the DOX group was due to a higher number of nonviable slices after DOX treatment. Increased motion artifacts in DOX-treated slices prevented the calculation of accurate APD values. Another important parameter that can be measured by dual voltage and calcium imaging is Vm-Ca delay, to determine robust excitation-contraction coupling in the tissue31. Delays or negative values of this parameter could be detrimental.
Figure 1: Human cardiac organotypic slice preparation. (A) Human hearts were stored in an ice-cold cardioplegic bath (mixture of cardioplegia solution and cardioplegia ice) upon retrieval. (B) The 1 cm3 cubes of left ventricular tissue were cut and mounted onto a metal tissue holder with 4% agarose gel glued to the back wall of the holder and transferred to an ice-cold bath of oxygenated slicing solution. (C) Once cut, slices were transferred to oxygenated recovery solution at RT in individual 100 µm nylon mesh cell strainers. Meshed washers covered the tissues to keep the slices from curling. (D) For long term culture, slices were washed in PBS and cultured in 6 well plates with 3 mL of tissue medium at 37 °C. (E) Mason’s trichrome stained slice section showing fiber orientation. (F) Representative optical action potential recorded from a slice. (G) Representative activation map of a slice paced at the center (blue). Please click here to view a larger version of this figure.
Figure 2: Dual camera optical mapping system. Dual camera optical mapping system in the upright imaging configuration. System parts include: 1) master and slave cameras for dual mapping; 2) tissue bath with PDMS gel; 3) filter cubes that house the excitation and emission filters and dichroic mirrors; 4) lens holders and lenses; and 5) excitation light source (520 nm green LED). Inset on right details filter and dichroic mirror combinations for dual optical mapping of Vm and calcium using RH237 and Rhod-2-AM, respectively. Please click here to view a larger version of this figure.
Figure 3: RHYTHM1.2 graphical user interface (GUI) of the open source optical mapping data analysis tool for single parameter and multiparameter analysis. The data file loading options and file list are displayed in the red box. Data analysis options are listed in the data analysis dropdown menu indicated in green. Display windows for displaying data maps are indicated in dark blue. Checkboxes to link dual mapping data files for simultaneous analysis of dual camera data are indicated in purple. The waveform window to plot action potential and calcium transient traces is displayed in yellow. Please click here to view a larger version of this figure.
Figure 4: Transmembrane potential and calcium transients mapping from a human organotypic cardiac slice preparation. (A) Representative optical action potential (black) and calcium transient (burgundy). (B,C) Activation maps obtained from Vm and calcium recordings, respectively. (D) Conduction velocity (CV) vector map. (E) Calcium transient decay constant (Tau) map. (F,G) Action potential duration (APD80) and calcium transient duration (CaTD80) map, respectively. (H,I) Maps of the rise time of the upstrokes of action potential and the calcium transient, respectively. Please click here to view a larger version of this figure.
Figure 5: Doxorubicin treatment slows transverse conduction velocity. (A) Activation maps from slices cultured for 24 hours without (control) and with doxorubicin at 50 μM (DOX). (B) Conduction velocity vector maps from control and DOX slices. (C) Average transverse conduction velocity calculated from control and DOX-treated slices. *P < 0.05. Please click here to view a larger version of this figure.
Here, we present step-by-step methods to obtain viable cardiac slices from cardioplegically arrested human hearts and to functionally characterize the slices using dual optical mapping of transmembrane potential and intracellular calcium. With preserved extracellular environment and native cell-cell coupling, human cardiac slices can be used as an accurate model of the human heart for fundamental scientific discovery and for efficacy and cardiotoxicity testing of pharmacological agents and gene therapies. The technology also allows for structural-functional mapping of specific regions of the heart, such as the sinoatrial node, atrioventricular node, and Purkinje fibers. The protocol described here lists basic steps for cardiac slice generation, imaging, and analyses that are best for short-term studies. Static culture of tissue slices have been shown to induce changes in cardiac tissue over prolonged periods of time, so for longer-term studies, we encourage the reader to incorporate more physiological culture conditions, such as electromechanical stimulation12,13,14.
Care should be taken for a number of steps to maintain tissue health. For example, the time between heart harvesting and tissue slicing should be minimized. As shown previously21, extended length of cardioplegic arrest can lead to altered electrophysiology. Also, to preserve the integrity of the extracellular matrix and cell-cell coupling, excessive handling and stretching of the slices should be avoided during slice collection, culturing, and functional studies. Excessive stretching of the slices could result in conduction blocks. Additionally, all solutions described in this protocol should be kept at a pH of 7.4. The Tyrode’s solution described here utilizes HEPES to maintain the proper pH using 100% O2. Sodium bicarbonate may be used to control pH instead, but this solution should be bubbled with 95% O2 and 5% CO2. Finally, during transportation and slicing of the heart, one should ensure that the cardioplegia and slicing solution are kept as close to freezing temperature as possible to preserve tissue viability.
The authors have nothing to disclose.
Funding by NIH (grants R21 EB023106, R44 HL139248, and R01 HL126802), by Leducq foundation (project RHYTHM) and an American Heart Association Postdoctoral Fellowship (19POST34370122) are gratefully acknowledged.
1mL BD Syringe | Thomas Scientific | 309597 | |
2,3-butanedione monoxime | Sigma-Aldrich | B0753 | |
6 well culture plates | Corning | 3516 | |
Biosafety cabinet | ThermoFisher Scientific | 1377 | |
Blebbistatin | Cayman | 13186 | |
Bubble Trap | Radnoti | 130149 | |
Calcium chloride | Sigma-Aldrich | C1016 | |
Corning Cell Strainers | Fisher Scientific | 07-201-432 | |
Di-4-ANEPPS | Biotium | stock solution at 1.25 mg/mL in DMSO | |
DMSO | Sigma-Aldrich | D2650 | |
Dumont #3c Forceps | Fine Science Tools | 11231-20 | |
Emission dichroic mirror | Chroma | T630LPXR-UF1 | |
Emission filter (RH237) | Chroma | ET690/50m | |
Emission Filter (Rhod2AM) | Chroma | ET590/33m | |
Excitation dichroic mirror | Chroma | T550LPXR-UF1 | |
Excitation Filter | Chroma | ET500/40x | |
Falcon 50mL Conical Centrifuge Tubes | Fisher Scientific | 14-959-49A | |
Glucose | Sigma-Aldrich | G8270 | |
Heat exchanger | Radnoti | 158821 | |
HEPES | Sigma-Aldrich | H3375 | |
Incubator | ThermoFisher Scientific | 50145502 | |
Insulin Transferrin Selenium (ITS) | Sigma-Aldrich | I3146 | |
LED excitation light source | Prizmatix | UHP-Mic-LED-520 | |
Magnessium chloride hexahydrate | Sigma-Aldrich | M9272 | |
Medium 199 | ThermoFisher Scientific | 11150059 | |
Micam Ultima L type CMOS camera | Scimedia | N/A | |
Minutien Pins | Fine Science Tools | 26002-10 | |
Pennicillin-Streptomycin | Sigma-Aldrich | P4333 | |
Peristaltic Pump | Cole Parmer | EW-07522-20 | |
Platinum pacing wire | Alfa Aesar | 43275 | |
Pluronic F127 | ThermoFisher Scientific | P6867 | nonionic, surfactant polyol |
Potassium chloride | Sigma-Aldrich | P3911 | |
Powerlab data acquisition and stimulator | AD Instruments | Powerlab 4/26 | |
RH237 | Biotium | 61018 | |
Rhod2AM | ThermoFisher Scientific | R1245MP | |
Rhod-2AM | Invitrogen, Carlsbad, CA | ||
Sodium bicarbonate | Sigma-Aldrich | S6014 | |
Sodium chloride | Sigma-Aldrich | S9625 | |
Sterilizer, dry bead | Sigma-Aldrich | Z378550 | |
Stone Oxygen Diffuser | Waterwood | B00O0NUVM0 | |
TissueSeal – Histoacryl Topical Skin Adhesive | gobiomed | AESCULAP | |
UltraPure Low Melting Point Agarose | Thermo Fisher Scientific | 16520100 | |
Ultrasound sonicator | Branson 1800 | ||
Vibratome | Campden Instruments | 7000 smz |