|Home | About | Journals | Submit | Contact Us | Français|
The protocols described here efficiently direct human pluripotent stem cells (hPSCs) to functional cardiomyocytes in a completely defined, serum-free system by temporal modulation of regulators of canonical Wnt signaling. Appropriate temporal application of Gsk3 inhibitor followed by expression of β-catenin shRNA or a chemical Wnt inhibitor is sufficient to produce a high yield (0.8–1.3 million cardiomyocytes/cm2) of virtually pure (80%–98%) functional cardiomyocytes from multiple hPSC lines without cell sorting or selection. Characterization of differentiated cells is performed in qualitative (immunostaining) and quantitative (flow cytometry) manners to assess expression of cardiac transcription factors and myofilament proteins. Flow cytometry of BrdU incorporation or Ki67 expression in conjuction with cardiac sarcomere myosin protein expression can be used to determine the proliferative capacity of hPSC-derived cardiomyocytes. Functional human cardiomyocytes differentiated via these protocols may constitute a potential cell source for heart disease modeling, drug screening, and cell-based therapeutic applications.
Directed differentiation of specific lineages from human pluripotent stem cells (hPSCs), including human embryonic stem cells (hESCs) and induced pluripotent stem cells (iPSCs), is the first critical step toward constructing development or disease models, drug screening tools, or cellular therapies from hPSCs. Because postnatal cardiomyocytes have little or no regenerative capacity, very limited supplies of human cardiomyocytes are available at present. hPSCs could potentially provide an unlimited supply of cardiomyocytes from a single clonal source.
Initial efforts to differentiate hESCs into cardiomyocytes employed embryoid bodies (EBs) in medium containing fetal calf serum, but this method is inefficient, with the culture typically composed of less than 1% cardiomyocytes, and provides variable results in different hPSC lines1. Mouse END-2 (visceral endoderm-like) cell-conditioned medium has been shown to enhance cardiac differentiation in EBs2. The appropriate temporal addition of growth factors important in cardiovascular development, including fibroblast growth factor 2 (FGF2), transforming growth factor β (TGFβ) superfamily growth factors Activin A and BMP4, vascular endothelial growth factor (VEGF), and the Wnt inhibitor DKK-1, can enhance cardiomyocyte differentiation in EBs3. Monitoring the onset of KDR/c-kit3 or Flk1/PDGFRα4 expression during the differentiation protocol enables presentation of these differentiation factors at the appropriate developmental stage, resulting in relatively consistent cardiomyocyte yields in multiple hPSC lines4. In prior work, we reported that undifferentiated hPSC expansion conditions affects cardiomyocyte yield5–8. Pretreatment of hPSCs with a Gsk3 inhibitor before forming EBs greatly enhanced cardiac differentiation using serum-based EB differentiation7.
As an alternative to hPSC differentiation to cardiomyocytes via EBs, a monolayer-based directed differentiation platform was developed. This protocol sequentially exposes the hPSCs to Activin A and BMP4 in defined RPMI/B27 medium, and has been reported to be much more efficient than serum-based EB differentiation, generating greater than 30% cardiomyocytes in the H7 hESC line9, 10. However, the efficiency of the Activin A and BMP4 monolayer directed differentiation protocol is highly variable between cell lines and experimental repeats within the same line11. Here, we modified this protocol in two ways, optimizing Gsk3 inhibitor pretreatment concentration at the undifferentiated hPSC expansion stage and insulin concentration during the first 5 days of differentiation. We found that insulin, present in B27 supplement, greatly inhibits cardiomyocyte yield during the first 5 days of differentiation which is consistent with previous reports that insulin inhibits cardiac differentiation of hPSCs12, 13. We therefore use B27 supplement lacking insulin in the cardiomyocyte differentiation medium. We also found that Gsk3 inhibitor pretreatment of undifferentiated hPSCs is critical for robust cardiac differentiation. We obtained less than 1% cardiomyocytes using the original RPMI/B27 monolayer directed differentiation protocol in several hPSC lines (H9, H13, H14, 19-9-11, 6-9-9 and IMR90C4) that we tested in several experimental repeats (n>5). However, using B27 supplement without insulin and Gsk3 inhibitor pretreatment in the Activin A and BMP4 monolayer directed differentiation protocol generated 30% – 90% cardiomyocytes across these hPSC lines14. Neither B27 lacking insulin nor Gsk3 inhibitor pretreatment alone was sufficient for efficient cardiomyocyte differentiation in this protocol.
Consistent with our findings that hPSC pretreatment with a Gsk3 inhibitor greatly improved cardiac differentiation of hPSCs, Wnt signaling has also been shown to have a biphasic effect on cardiac development in zebrafish, mouse embryos, and mouse embryonic stem cells15, 16, with early Wnt signaling enhancing and later Wnt signaling repressing heart development. Because of the important temporal roles of Wnt/β-catenin on cardiac differentiation, we assessed whether modulation of Wnt/β-catenin signaling, in the absence of exogenous Activin A and BMP4, was sufficient to efficiently produce cardiomyocytes from hPSCs. We found that sequential activation of canonical Wnt signaling by Gsk3 inhibitor treatment and inhibition of Wnt signaling by inducible expression of β-catenin shRNA is sufficient to drive multiple hPSC lines to cardiomyocytes7. Small molecule inhibitors of Wnt ligand production (IWPs) also induced cardiac differentiation as effectively as β-catenin shRNA expression7. Other inhibitors of Wnt signaling, including IWR-1-endo17 and XAV93918, also have been show to promote cardiac differentiation of pluripotent stem cells.
Here we provide three protocols for efficient generation of functional human cardiomyocytes from hPSCs, the first utilizing TGFβ superfamily growth factors (protocol 1, GiAB) and the others employing small molecule activators of canonical Wnt signaling followed by shRNA of β-catenin expression (protocol 2, GiSB) or small molecule inhibitors of Wnt signaling (protocol 3, GiWi) in a growth factor-free system. Protocol 1 relies upon treatment of undifferentiated hPSCs with Gsk3 inhibitor in mTeSR1, followed by Activin A and BMP4 in RPMI/B27-insulin. The small molecule methods, protocols 2 and 3, use sequential treatment of Gsk3 inhibitors and Wnt signaling inhibitors (or inducible expression of β-catenin shRNA) to stimulate cardiogenesis. Compared with growth factor-induced cardiomyocyte differentiation, the small molecule approaches (protocol 2 and protocol 3) provide more robust cardiac differentiation, producing 82–98% cardiomyocytes from six hPSC lines. While inducible expression of β-catenin shRNA provides specific and facile temporal regulation of canonical Wnt signaling, this method (protocol 2) requires genetic modification of the hPSC line. Protocol 3, which uses Wnt signaling inhibitors instead of β-catenin shRNA, does not require genetic modification and is applicable to any existing hPSC line. All these three protocols can be performed under fully defined conditions with defined medium (RPMI/B27 without or with insulin) and defined substrates (Synthemax plates). These protocols will enable efficient production of human cardiomyocytes for development and disease research, drug screening and testing, and advancing cardiac cellular therapies.
The cardiac differentiation protocol critically depends on the quality of hPSCs, which in turn relies on the quality of matrix and medium, and the methods used to passage and maintain the hPSCs. We recommend mTeSR1 medium in conjunction with hPSC-qualified lots of Matrigel or the chemically defined surface Synthemax for hPSC expansion and maintenance in the undifferentiated state. An enzyme-free method of passaging the cells using Versene is recommended. The hPSCs cultivated in this manner should exhibit a uniform undifferentiated morphology (Fig. 1A). Undifferentiated hPSCs should express Oct4, Nanog, SSEA4, and TRA-1-80. The pluripotent marker Oct4 should be expressed in greater than 95% of the cells, as assessed by flow cytometry (Fig. 1B). Flow cytometry of Oct4 expression should be performed every three passages to validate the lack of differentiation of hPSCs. Partially differentiated hPSCs will diminish the cardiomyocyte differentiation efficiency in the protocols presented here.
The hPSCs are initially cultured on Matrigel-coated plates or Synthemax plates in mTeSR1 medium until fully confluent. Differentiation is initiated by removing the mTeSR1 medium and adding RPMI/B27 medium lacking insulin and containing a Gsk3 inhibitor, such as CHIR99021. 24 hours of culture in this medium generates a high percentage of brachyury-expressing cells (>95% by flow cytometry) (Fig. 2A). In order to direct these brachyury-expressing mesendoderm progenitor cells to a cardiac fate, inhibition of canonical Wnt signaling either by β-catenin shRNA expression or Wnt signaling inhibitors, such as Porcupine inhibitors IWP2 or IWP4, is performed. Cardiac mesoderm cells spontaneously develop into functional contracting cardiomyocytes when cultured in RPMI/B27 medium.
Cardiac differentiation using either the growth factor or the small molecule-based differentiation protocols proceeds rapidly. A relatively pure (>95%) population of brachyury-expressing mesendoderm cells can be detected after one day of differentiation. Gene expression of the cardiac transcription factors NKX2.519 and ISL120, 21 begins at day 4, with the protein detectable at day 5 (Fig. 2B). Cardiac troponin T (cTnT) can be readily detected at day 8 of differentiation. The presence of cardiomyocytes can be easily established by visual observation of spontaneously contracting regions. The first beating cluster of cells can be observed between days 8 to day 10, depending on individual cell line used. Robust spontaneous contraction occurs by day 12. Cardiac marker protein expression after onset of contractions can also be assessed by immunostaining or quantified with flow cytometry. Here we provide procedures for performing these characterizations, including optimized immunostaining and flow cytometry methods, antibody sources, antibody dilutions, and combination of antibodies that facilitate dual staining.
CRITICAL: 293TN cells adhere weakly and minimal mechanical disturbance or contact with cold medium or buffers can cause their detachment, especially when they are confluent. Therefore, transfer of plates and medium changes should be performed very gently. Medium used should be warmed at 37 °C prior to addition to cells.
In a sterile hood, add 23 ml of cold (4 °C) DMEM/F12 to a 50 ml conical tube and keep it cold by placing it on ice. Remove one Matrigel aliquot (2 mg) from the freezer, and add 1 ml of cold DMEM/F12 to it. Gently pipette the Matrigel solution with a P1000 tip to thaw and dissolve the Matrigel. Immediately transfer the Matrigel solution to the 50 ml conical tube that contains 23 ml cold DMEM/F12. Immediately add 1 ml/well Matrigel in DMEM/F12 for 6-well plates, 0.5 ml/well for 12-well plates, 250 μl/well for 24-well plates, or 100 μl/well for 96-well plates. Allow the Matrigel to set for 30 minutes at room temperature before use. The Matrigel-coated plates can be stored at 4 °C for up to 3 weeks. CRITICAL: We recommend dissolving 0.5 mg Matrigel into 6 ml cold DMEM/F12. Use lots of Matrigel qualified by BD Biosciences for hESC/iPSC culture. Some lots of Matrigel do not support hPSC self-renewal. We also recommend the use of Synthemax plates as an alternative matrix for consistent hPSC maintenance.
Autoclave the coverslips at 121°C, 15 psi for 30 minutes and place one sterile coverslip in each well of a 12-well plate. Add 1 ml of 0.1% gelatin per well and incubate at 37 °C overnight. Store the gelatin coated coverslips at 4 °C for up to 2 months.
|Primers for quantitative RT-PCR|
|GAPDH||F: 5′ GTGGACCTGACCTGCCGTCT 3′|
R: 5′ GGAGGAGTGGGTGTCGCTGT 3′
|Size 152 bp|
|CTNNB1||F: 5′ CCCACTAATGTCCAGCGTTT 3′|
R: 5′ AACGCATGATAGCGTGTCTG 3′
|Size 217 bp|
Robust spontaneous contraction should occur by day 12. Two typical spontaneously contracting results from 19-9-11 inducible β-catenin shRNA line are shown in Movie S1 (day 15 cardiomyocytes) and Movie S2 (day 180 cardiomyocytes).
Because we use small molecule inhibitors of Wnt signaling instead of β-catenin shRNA in the GiWi protocol, transgenic modification of hPSC to inducibly express β-catenin shRNA is not required for this protocol. GiWi protocol is applicable to any existing hPSC line. A summary of this protocol is shown in Fig. 3.
Greater than 80% cardiomyocytes were obtained in the six hESC and iPSC lines that we tested using GiSB or GiWi protocols (Table 1).
Scientists may be interested in comparing the small molecule-derived cardiomyocytes with cardiomyocytes differentiated from hPSCs using growth factors. A brief outline for generating cardiomyocytes via our activin/BMP method is provided in Box 1.
We recommend flow cytometry for quantitative analysis of the purity of hPSC-derived cardiomyocytes. Antibody combinations of cTnT/SMA, cTnT/MLC2a, and MF20/Ki67 are recommended for double staining.
The protocol for analysis of BrdU and MF20 is different from other combinations of antibodies.
This protocol presents a rapid and efficient (80%–98% cTnT+ cells after two weeks) method for the generation of functional cardiomyocytes from multiple hPSC lines. Before starting the differentiation protocol, well-maintained hPSCs should have a high ratio of nucleus to cytoplasm, prominent nucleoli morphology (Fig. 1A) and be uniformly positive for pluripotency markers, including Oct4, Nanog, TRA-1-80 and SSEA4 (Fig. 1B–C). 24 hours after initiation of differentiation with CHIR99021, at least 90% of the total differentiated cells should express brachyury, the mesendoderm marker (Fig. 2A). Mesendoderm differentiation below 90% could be caused by poor quality of the starting hPSCs. After 5–6 days of differentiation, cells will express the cardiac progenitor marker protein Isl1 (Fig. 2B). The first beating cluster of cells can be observed between day 8 to day 10, depending on the individual cell line used. Robust spontaneous contraction occurs by day 12 in all hPSCs that we tested. Cardiac sarcomere proteins, such as α-actinin, MLC2a, cTnT, will be expressed in more than 80% of the differentiated cells (Fig. 3–5) by day 15. About 20% of the day 20 cardiomyocytes should show proliferative capacity via Ki67 or BrdU incorporation analysis (Fig. 5B). We typically generate approximately 3–5 million cardiomyocytes (80%–98% cTnT+) per well of a 12-well plate (surface area = 3.8 cm2) resulting a total density of day 15 cardiomyocytes equal to 0.8–1.3 million cardiomyocytes/cm2.
This study was supported by NIH grants R01 EB007534 and U01 HL099773 and NSF grant EFRI 0735903.
AUTHOR CONTRIBUTIONSX.L. designed and performed experiments, analyzed data and wrote the paper; J.Z., S.M.A., K.Z., L.H., X.B. and C.H. contributed to the development of this protocol. T.J.K. and S.P.P. supervised the project, wrote and approved the final paper.
COMPETING FINANCIAL INTERESTS
T.J.K. is a founder and consultant for Cellular Dynamics International, a company that uses human stem cells for drug testing. All the other authors declare no competing financial interests.