Here, we describe a robust protocol for human cardiomyocyte derivation that combines small molecule-modulated cardiac differentiation and glucose deprivation-mediated cardiomyocyte purification, enabling production of purified cardiomyocytes for the purposes of cardiovascular disease modeling and drug screening.
Human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) have become an important cell source to address the lack of primary cardiomyocytes available for basic research and translational applications. To differentiate hiPSCs into cardiomyocytes, various protocols including embryoid body (EB)-based differentiation and growth factor induction have been developed. However, these protocols are inefficient and highly variable in their ability to generate purified cardiomyocytes. Recently, a small molecule-based protocol utilizing modulation of Wnt/β-Catenin signaling was shown to promote cardiac differentiation with high efficiency. With this protocol, greater than 50%-60% of differentiated cells were cardiac troponin-positive cardiomyocytes were consistently observed. To further increase cardiomyocyte purity, the differentiated cells were subjected to glucose starvation to specifically eliminate non-cardiomyocytes based on the metabolic differences between cardiomyocytes and non-cardiomyocytes. Using this selection strategy, we consistently obtained a greater than 30% increase in the ratio of cardiomyocytes to non-cardiomyocytes in a population of differentiated cells. These highly purified cardiomyocytes should enhance the reliability of results from human iPSC-based in vitro disease modeling studies and drug screening assays.
Primary human cardiomyocytes are difficult to obtain because of the requirement for invasive cardiac biopsies, difficulty in dissociating to single cells, and because of poor long-term cell survival in culture. Given this lack of primary human cardiomyocytes, patient-specific human induced pluripotent stem cell-derived cardiomyocyte (hiPSC-CM) technology has been regarded as a powerful alternative cardiomyocyte source for basic research as well as clinical and translational applications such as disease modeling and drug discovery1. Early efforts in differentiating pluripotent stem cells into cardiomyocytes employed differentiation protocols using embryoid bodies (EBs), but this method is inefficient in producing cardiomyocytes because often less than 25% of cells in an EB are beating cardiomyocytes2,3. Comparatively, a monolayer-based differentiation protocol using the cytokines activin A and BMP4 displayed a higher efficiency than EBs, but this protocol is still relatively inefficient, requires expensive growth factors, and only functions in a limited number of human pluripotent stem cell lines4. Recently, a highly efficient, hiPSC monolayer-based cardiomyocyte differentiation protocol was developed by modulating Wnt/β-Catenin signaling5. These hiPSC-CMs express cardiac troponin T and alpha actinin, two sarcomeric proteins that are standard markers of cardiomyocytes6. The protocol describe here is an adaptation of this small molecule-based, feeder cell-free, monolayer differentiation method5,7. We are able to obtain beating cardiomyocytes from hiPSCs after 7-10 days (Figure 1). However, following a cardiomyocyte differentiation resulting in 50% beating cells, immunostaining consistently shows the existence of a population of non-cardiomyocytes that are negative for cardiomyocyte-specific markers such as cardiac-specific troponin T and alpha-actinin. To further purify cardiomyocytes and eliminate non-cardiomyocytes, heterogeneous differentiated cell populations were subjected to glucose starvation by treating them with an extremely low glucose culture medium for multiple days (Figure 2). This treatment selectively eliminates non-cardiomyocytes due to the ability of cardiomyocytes, but not non-cardiomyocytes, to metabolize lactate as the primary energy source in order to survive in a low glucose environment8. After this purification step, a 40% increase in the ratio of cardiomyocytes to non-cardiomyocytes is observed, (Figure 3, Figure 4) and these cells can be used for downstream gene expression analysis, disease modeling, and drug screening assays.
NOTE: Vendor information for all reagents used in this protocol has been listed in Table 1 and Materials List. All solutions and equipment coming into contact with cells must be sterile, and aseptic technique should be used accordingly. Perform all culture incubations in a humidified 37 °C, 5% CO2 incubator unless otherwise specified. In this protocol, all differentiations are performed in 6-well plates, in which the hiPSCs are seeded. Following differentiation and purification, cells can be dissociated and replated for downstream use.
1. Medium Preparation
2. Pre-coating 6-well Plates with ECMS
3. Thawing the Frozen hiPSCs
NOTE: Closely follow this procedure as it may lead to optimal culturing and downstream differentiation of hiPSCs into hiPSC-CMs following the freeze/thaw cycle.
4. Passaging of hiPSCs
5. Freezing hiPSCs
NOTE: Closely follow this procedure as it may lead to optimal culturing and downstream differentiation of hiPSCs into hiPSC-CMs following the freeze/thaw cycle.
6. Cardiac Differentiation of Human iPSC
NOTE: All media should be at least at RT when added.
7. Purification of Human Cardiomyocytes through Glucose Starvation
The morphological changes during hiPSC differentiation.
The hiPSC cultured in feeder-free plates grew as flat, two-dimensional colonies. Upon reaching around 85% confluency, hiPSCs were treated with 6 µM CHIR for differentiation (Figure 1A). Substantial amounts of cell death, a normal and common phenomenon, were observed after 24 hr of CHIR treatment. After two days of CHIR treatment, the hiPSCs continued to differentiate towards a mesodermal fate. In comparison to cells in hiPSC colonies, cell size for these day 2 cells increased. Cell number also increased through cell division (Figure 1B). At day 5, the mesodermal cells were directed towards the cardiac lineage (cardiac mesoderm). At this time, cells begin coalescing and forming characteristic, branch-like structures (Figure 1C). At day 10, branch-like structures were highly pronounced, and cardiomyocytes started spontaneously beating (Figure 1D). Terminally-differentiated cardiomyocytes will not proliferate beyond this point. A timeline of the differentiation protocol is shown in Figure 2.
Immunostaining and flow cytometry results showed cardiomyocytes were purified with glucose starvation.
After 10 days of differentiation, more than 50% of cell areas are beating. However, beating cell sheets sometimes are also intermingled with non-cardiomyocytes. To examine cardiomyocyte purity, cells with immunostaining against cardiomyocyte marker cardiac Troponin-T (cTnT) were characterized to find that most cells were cTnT positive. However, in an unpurified population of differentiated cells, a number of cells also lacked expression of cTnT, suggesting the presence of non-cardiomyocytes in this differentiated population at day 13 (Figure 3A). However, with glucose starvation, almost all remaining cells were cTnT positive at day 13 (Figure 3B), indicating successful purification of the cardiomyocytes following glucose starvation. These data were further corroborated using flow cytometry analysis. A population of differentiated cells at day 13 post-differentiation contained approximately 50% TNNT2+ cells when not glucose starved (Figure 4A). A parallel differentiation was also conducted but with a 3 day glucose deprivation beginning at day 10 after differentiation. In contrast to the unstarved differentiation, glucose deprivation led to a purified population containing 90% TNNT2+ cells at day 13 (Figure 4B).
Figure 1. The sequential morphological changes during hiPSC differentiation towards the cardiac lineage. (A) Undifferentiated hiPSCs at day 0 exhibit typical colony morphology and reached around 85% confluency. (B) At day 2, cells after treatment with CHIR for two days reached 100% confluency and entered the mesodermal lineage. (C) At day 5, the cells after treatment with IWR1 for 2 days entered the cardiac mesoderm stage. Some cells began to fuse and form characteristic branch-like morphologies (indicated by arrows). (D) At day 7-10, branch-like structures are highly evident, and cardiomyocytes start spontaneously beating. Bar = 200 μm. Please click here to view a larger version of this figure.
Figure 2. Timeline of hiPSC-CM differentiation protocol and subsequent glucose starvation process. Beating cardiomyocytes are typically first observed at approximately days 7-10. Two rounds of glucose starvation will result in a purified population of cardiomyocytes by day 17. Please click here to view a larger version of this figure.
Figure 3. Immunofluorescence reveals purification of cardiomyocytes following glucose starvation. (A) Unpurified cells after 13 days of differentiation were stained with Cardiac Troponin T (cTnT). Most cells have differentiated into cardiomyocytes and were cTnT positive, but a number of non-cardiac, cTnT-negative cells are also present. (B) In contrast, after a 3 day glucose starvation beginning at day 10, almost all of the surviving cells were cTnT positive by day 13. Scale bar = 200 μm. Please click here to view a larger version of this figure.
Figure 4. Flow cytometry reveals purification of cardiomyocytes following glucose starvation. (A) After 13 days of cardiac differentiation without glucose starvation, a population of cells exhibited roughly 50% TNNT2+ cardiomyocytes. Red indicates the differentiated cell population and blue indicates undifferentiated hiPSCs as a negative control. (B) After 10 days of cardiac differentiation and followed by 3 days of glucose starvation, a population of cells from the same batch of differentiations was purified to contain 90% TNNT2+ cardiomyocytes. Please click here to view a larger version of this figure.
The compositions for E8 medium | |||
E8 | Volume: 1 L | Company | Catalog number |
DMEM/F12 with Glutamine and HEPES | 1,000 ml | Invitrogen | 11330-032 |
NaHCO3 (7.5%, 75 mg/ml) | 7.24 ml | Invitrogen | 25080-094 |
L-Ascorbic acid 2-phosphate (64 mg/ml) | 1 ml | Sigma | A8960 |
sodium selenite (70 µg/ml) | 200 µl | Sigma | S5261 |
transferrin (50 mg/ml) | 214 µl | Sigma | T3705 |
insulin (4 mg/ml) | 5 ml | Invitrogen | 12585-014 |
FGF2 (200 ng/µl) | 500 µl | Peprotech | 100-18B |
TGFB1 (100 ng/µl) | 20 µl | Peprotech | 100-21 |
The stock solutions for E8 compositions | |
L-Ascorbic acid-2-phosphate (64 mg/ml) | 3.2 g in 50 ml Ultrapure water, store 500 µl aliquots at -20 °C |
Transferrin (50 mg/ml) | 500 mg in 10 ml of Ultrapure water, store 107 µl aliquots at -20 °C |
Sodium selenite (70 µg/ml) | 35 mg sodium selenite into 500 ml ultrapure water, store 100 µl aliquots at -20 °C |
FGF2 (200 ng/µl) | 1 mg in 5 ml cold D-PBS, store 250 µl aliquots at -20 °C |
TGFB1 (100 ng/µl) | 100 µg in 1 ml cold 10 mM citric acid, pH 3, store 10 µl aliquots at -20 °C |
Table 1. Composition of E8 Medium.
Obtaining a large amount of highly purified hiPSC-derived cardiomyocytes is critical for basic cardiac research as well as clinical and translational applications. Cardiac differentiation protocols have undergone tremendous improvements in recent years, transitioning from embryoid body-based methods utilizing cardiogenic growth factors2, to matrix sandwich methods12, and finally to small molecule-modulated and monolayer-based methods5. Of the aforementioned protocols, the protocol described here displayed the highest and the most reproducible cardiac differentiation efficiency for different hiPSC cell lines by modulating Wnt/β-catenin signaling using the small molecules CHIR and IWR5,7. Specifically, the undifferentiated hiPSCs were induced to undergo mesodermal differentiation with the GSK3-beta inhibitor CHIR and subsequent cardiac differentiation with the Wnt signaling inhibitor IWR1. These sequential steps of Wnt signaling modulation, aided by insulin depletion during the cardiac lineage induction phase, led to highly efficient cardiac differentiation. CHIR is a common small molecule GSK3-beta inhibitor used for mesodermal induction, but other small molecules for GSK3-beta inhibition have been tested in cardiomyocyte differentiation protocols11. Multiple options are also used for the subsequent small molecule Wnt inhibitor. For example, a recent publication focusing on chemically-defined differentiation of pluripotent stem cells to cardiomyocytes uses 2 µM Wnt-C59 for effective Wnt inhibition and cardiac mesoderm induction11.
In addition, the differentiated cardiomyocytes with a glucose starvation method were further purified, which takes advantage of the distinct glucose metabolic flow existing between cardiomyocytes and non-cardiomyocytes8. It is essential to note that the effectiveness of the glucose starvation method is density-dependent (i.e., differentiations that yielded higher percentages of cardiomyocytes are more likely to achieve greater cardiomyocyte survival). However, the transition to the low glucose medium is also a stressful condition for the cardiomyocytes. It is important to change the cells back to the regular RPMI/B27 with insulin medium after 3 days of glucose starvation. In this protocol, a feeder-free growth system was utilized, in which pluripotent stem cells were not grown on feeder mouse embryonic fibroblasts (MEFs). However, prior cardiac differentiation protocols have utilized feeder-based systems to produce cardiomyocytes from pluripotent stem cells5. Likewise, prior protocols have also used mTeSR1 medium for stem cell maintenance, but the E8 media and feeder-free system utilized here is superior due to its simplicity and lack of excess xenogeneic components such as bovine serum albumin and mouse embryonic fibroblasts.
Currently, pluripotent stem cell differentiation towards cardiomyocyte lineages is largely robust, but still suffers from hiPSC line-to-line variability in terms of efficiency. This remains a major issue in the cardiac differentiation field and will require further study. The differentiation efficiency is improved by proper maintenance of hiPSC lines, and in particular, preventing overconfluency during hiPSC maintenance. Additional variability arises from cell seeding during the passaging process, as an evenly seeded monolayer of hiPSCs tends to give rise to the best overall differentiations. Here, a mouse-derived, Matrigel-based ECMS for cell seeding during cell culture was successfully utilized, but an eventual transition to xeno-free substrates is recommended. In regards to cell seeding on ECMS, uneven seeding often results in uneven distribution of cardiomyocytes within a particular well. For example, if hiPSCs are seeded unevenly, the edges of a well in a six-well plate may give rise to higher numbers of cardiomyocytes than the center of the well. These uneven seeding conditions may give rise to low efficiency differentiations and a larger number of non-cardiomyocyte, mesodermal derivatives such as fibroblasts and smooth muscle cells. We have also observed that low efficiency differentiations (below 50%) are much more difficult to purify using the glucose deprivation process mentioned here. Another side effect of long-term glucose starvation is a potential loss in cardiomyocyte viability. Although the cardiomyocytes are able to metabolize lactate in the absence of glucose, we find that this switch to a low-glucose environment is stressful for the cells. Cells may stop spontaneously beating, and some cell loss may be observed if glucose starvation is prolonged beyond the recommended time. This enhanced sensitivity to glucose deprivation may be reflective of the well-established developmental, functional, and electrophysiological immaturity of stem cell-derived cardiomyocytes in comparison to true adult cardiomyocytes13.
In summary, the protocol described here combines the small molecule-based, hiPSC monolayer cardiac differentiation method with a cardiomyocyte-purifying glucose starvation method. This protocol allows for the reproducible generation of highly purified cardiomyocytes and should facilitate various downstream assays relevant to cardiovascular disease modeling and drug screening.
The authors have nothing to disclose.
This work was supported in part by the NIH/NHBI (U01 HL099776-5), the NIH Director’s New Innovator Award (DP2 OD004411-2), the California Institute of Regenerative Medicine (RB3-05129), the American Heart Association (14GRNT18630016) and the Endowed Faculty Scholar Award from the Lucile Packard Foundation for Children and the Child Health Research Institute at Stanford (to SMW). We also acknowledge funding support from the American Heart Association Predoctoral Fellowship 13PRE15770000, and National Science Foundation Graduate Research Fellowship Program DGE-114747 (AS).
Name | Company | Catalog number |
Matrigel (9-12 mg/mL) | BD Biosciences | 354277 |
RPMI media | Invitrogen | 11835055 |
Glucose free RPMI media | Invitrogen | 11879-020 |
B27 Minus Insulin | Invitrogen | A1895601 |
B27 Supplement (w/ insulin) | Invitrogen | 17504-044 |
Pen-strep antibiotic | Invitrogen | 15140122 |
Fetal bovine serum | BenchMark | 100-106 |
DMSO | Sigma | D-2650 |
ROCK inhibitor Y-27632 | EMD Millipore | 688000 |
CHIR99021 | Thermo Fisher | 508306 |
IWR1 | Sigma | I0161 |
EDTA | Invitrogen | 15575-020 |
Accutase | Millipore | SCR005 |
Cell lifter | Fisher | 08-100-240 |
Cryovial | Fisher (NUNC tubes) | 375418 |
TrypLE Select Enzyme | Invitrogen | 12563-011 |