Conventional methods to initiate suspension aggregate based cardiac differentiation of human pluripotent stems cells (hPSCs) are plagued with culture heterogeneity with respect to aggregate size and shape. Here, we describe a robust method for cardiac differentiation employing microwells to generate size-controlled hPSC aggregates cultured under cardiac-promoting conditions.
Cardiac differentiation of human pluripotent stems cells (hPSCs) is typically carried out in suspension cell aggregates. Conventional aggregate formation of hPSCs involves dissociating cell colonies into smaller clumps, with size control of the clumps crudely controlled by pipetting the cell suspension until the desired clump size is achieved. One of the main challenges of conventional aggregate-based cardiac differentiation of hPSCs is that culture heterogeneity and spatial disorganization lead to variable and inefficient cardiomyocyte yield. We and others have previously reported that human embryonic stem cell (hESC) aggregate size can be modulated to optimize cardiac induction efficiency. We have addressed this challenge by employing a scalable, microwell-based approach to control physical parameters of aggregate formation, specifically aggregate size and shape. The method we describe here consists of forced aggregation of defined hPSC numbers in microwells, and the subsequent culture of these aggregates in conditions that direct cardiac induction. This protocol can be readily scaled depending on the size and number of wells used. Using this method, we can consistently achieve culture outputs with cardiomyocyte frequencies greater than 70%.
In vitro cell culture can be performed under a number of modalities, but is typically carried out either in two-dimensional adherent conditions or in three-dimensional suspension conditions that more fully recapitulate in vivo systems. Consequently there is a growing trend in many fields of research to develop robust methods for generating three-dimensional tissue constructs. In scenarios where cell types and processes require a supportive extracellular matrix (ECM) surface and adhesion signals, three-dimensional culture may be enabled via scaffolded constructs, where cells are cultured on or in an exogenous supporting matrix1. Cells and processes that do not require adhesion to a supportive matrix can be carried out in suspension as unscaffolded systems composed primarily or exclusively of cells (which may then proceed to generate their own endogenous matrices)2,3. Here, we present a protocol for cardiac differentiation of human pluripotent stem cells (hPSCs — stem cells capable of becoming any cell type in the body, whether from embryonic or other sources) in size-controlled, uniform, unscaffolded aggregates.
Differentiation of hPSCs as suspension aggregates is plagued by large variations in aggregate size both within a run and between runs. This variability is a consequence of the method typically employed to generate these aggregates, which involves mechanical dissociation of cell colonies. To reduce this variability, a number of approaches have been used to control the number of cells per aggregate as well as aggregate diameter and uniformity. Examples include formation of aggregates in microcentrifuge tubes4 or as hanging drops5, micropatterning defined two-dimensional hPSC colonies6 which can then be transferred to suspension, or centrifugation of cells into U- or V-bottom multi-well plates7,8,9. However, all these approaches are limited by their low throughput of aggregate generation. Well-based systems employ a similar approach to V-bottom plate systems, however the smaller size of the microwells (in this protocol each having a width of 400 µm) enables the generation of larger numbers of uniform aggregates from a single culture-plate well (standard diameter of ~ 15.5 mm containing ~ 1,200 microwells) than would be generated from a whole V-bottom plate10. Well-based aggregate formation has been used in a number of settings including differentiation of hPSCs to ectodermal11, endodermal12, mesodermal13 and extraembryonic14 fates; chondrogenesis from mesenchymal stem cells15; generation of uniform substrates for toxicological screening16; and investigations of mechanobiology17.
One major challenge in developing robust manufacturing protocols for the production of hPSC-derived cardiomyocytes has been the lack of reproducibility in cardiac differentiation efficiency between runs. We previously demonstrated that this variability can be attributed to heterogeneity in the starting hPSC population, which comprises both self-renewing hPSCs and differentiating cells that express genes associated with endoderm and neural differentiation6,18. Signals secreted by these differentiating cells impact cardiac induction. Specifically, extraembryonic endoderm promotes cardiac induction, while neural progenitors inhibit cardiac induction. Upon hPSC aggregation, cells within the aggregate differentiate and organize so that undifferentiated hPSCs are surrounded by a layer of extraembryonic endoderm cells that develop on the aggregate surface13. By controlling aggregate size, we can modulate the ratio of cardiac-inducing endoderm cells to undifferentiated hPSCs (surface area to volume ratio) and optimize this ratio for maximum cardiac induction13.
1. Preparation of Medium Components
2. Preparation of the Microwell Plate
NOTE: All steps should be performed in a biological safety cabinet.
3. Formation of hPSC Aggregates in the Microwell Plate
NOTE: All steps should be performed in a biological safety cabinet.
4. Cardiac Induction Stage 1
5. Cardiac Induction Stage 2
6. Cardiac Induction Stage 3
7. Flow Cytometry Analysis of Cardiac Troponin T (cTnT) Expression Frequency of Microwell Culture Output
Size-controlled aggregates of hPSCs can effectively be formed using the microwell system, dependent only on the concentration of cells and the microwell surface area. Following a short centrifugation, the appropriate numbers of cells (1,000 in this protocol) are brought together in each microwell (Figure 1A). Importantly, these cells reestablish intracellular connections within 24 hr, and should no longer fill the well, but appear as compact aggregates with smooth edges (Figure 1B). These aggregates provide the starting materials for further differentiation towards a cardiac fate. If the cells fail to form tight clusters, this suggests possible cell death following dissociation and reaggregation, and the suitability of single cell passaging and ROCK inhibitor concentrations for a particular cell line should be examined. The following three days in microwells show little change in aggregate morphology, although some growth is evident. When removed from the microwells, aggregates should maintain their round, tightly packed morphology and be of a similar size to one another (Figure 1C). Culture in ULA 24-well plates will permit further cell expansion and growth.
By day 8, after exposure first to Activin signaling, and Wnt inhibition, the aggregates will begin to appear as larger and brighter aggregates (Figure 1D). During this period, considerable cellular debris will be evident at the bottom of each well and must be removed by allowing aggregates to settle before removing media. Occasionally, many aggregates will fuse together. This will not inhibit the differentiation of other aggregates in the well, although these "super aggregates" tend not to exhibit the morphological changes seen with smaller aggregates and are less likely to undergo complete differentiation.
Continued differentiation results in notable morphological changes to the aggregates with an increased size and the appearance of organized fibrous regions. By day 12, contracting aggregates can be observed. These will always be made up of large, transparent cells and often include extensive extracellular matrix outside of the aggregate (Figure 2). While aggregate-wide contractions indicate a successful differentiation, cardiac marker expression may also be observed in aggregates that do not appear to contract. Following dissociation of aggregates and immunolabeling, a majority of cells will be positive for the cardiomyocyte marker cTnT by flow cytometry (Figure 3). The expression of this marker is stable in the cells and may be observed in aggregates as late as day 19 of differentiation.
Figure 1: Timeline of hPSC differentiated towards a cardiac fate. Immediately after aggregation, cells nearly fill each microwell (A). One day later, the aggregates appear condensed and smooth (B). This morphology persists even when the aggregates are removed from the microwells and plated into well plates (C). By day 8, aggregates begin to expand and appear lighter in color (D). Scale bar: 250 µm. Please click here to view a larger version of this figure.
Figure 2: Aggregates Begin Contracting by Day 12 of Differentiation. After six days in Cardiac Induction Stage 3 Medium, forceful aggregate-wide contractions were observed (top panel: relaxation, middle panel: contraction). The lower panel is derived from subtracting the upper and middle panels, with most significant differences appearing as black or white (arrowheads). Scale bar: 250 µm. Please click here to view a larger version of this figure.
Figure 3: Immunolabeling for Cardiac Troponin T in Differentiated hPSCs. At day 17, most of the cells are positive for cardiac troponin T by flow cytometry (filled histogram). Also shown are cells stained with secondary antibody alone (unfilled histogram). Please click here to view a larger version of this figure.
It has been observed that efficient cardiac differentiation of pluripotent stem cells is a highly variable process. While it is not surprising that different cell lines exhibit varying propensities for differentiation capacity to specific cell types, it has been observed that cardiac differentiation efficiency fluctuates dramatically between replicate runs using the same cell line6. The protocol described here addresses one major source of this variability by directly controlling the input cell number per aggregate. To further reduce variability between runs, it is recommended that hPSC lines adapted for single cell passaging are used, as this form of hPSC expansion and maintenance results in more consistent pluripotent populations with respect to expression frequencies of pluripotency markers (e.g., Oct4, Nanog, Tra-1-60, etc.).
The protocol as written here specifies an aggregate size of 1,000 cells for optimal cardiac induction from the HES-2 embryonic stem cell line. To apply this protocol to different cell lines, it is critical that an initial aggregate size screen be performed to determine the cell line-specific optimal aggregate size. While it does not directly impact the procedures to be followed here, we remind the reader that changes in aggregate size and overall cell density are expected to affect oxygen delivery. This may become a relevant consideration in downstream applications. Additionally, apoptotic cell death is a concern during dissociation of hPSCs to single cells. Therefore, it is critical to ensure that ROCK inhibitor is present during forced cell aggregation in the microwells. Finally, it is critical that on day 4 of differentiation the aggregates are well washed to remove trace Activin A, present in the Induction 1 Medium, prior to resuspension in Induction 2 Medium. After day 4 of differentiation, Activin A promotes endoderm differentiation at the expense of mesoderm induction20.
The main application of this technique is to screen aggregate sizes that promote efficient cardiac differentiation. However, one of the limitations of the current technique is that it is challenging to scale cardiac production to clinically relevant levels using microwell plates. Scale up of cardiac differentiation is typically carried out in bulk culture conditions in stirred suspension bioreactors21. Therefore, once the microwell system has been used to determine acceptable ranges of aggregate size for efficient cardiac induction, the next step to scale up is to determine bioreactor impeller speeds that can generate the desired cell aggregate size.
One of the significant differences of this technique with respect to other methods for aggregate-based cardiac differentiation is that it enables direct investigations into modulating the effects of endogenous signaling in aggregates as well as the co-culture of inductive/inhibitory tissue types with the hPSCs in the aggregate13. These types of investigations can inform process development of large scale cardiac production.
The authors have nothing to disclose.
We thank Dr. Peter Zandstra, in whose laboratory this protocol was developed, and Drs. Mark Gagliardi and Gordon Keller who provided assistance in establishing the initial methods on which this process was based. Protocol development was supported by an Ontario Graduate Scholarship in Science and Technology to C.B. and a grant from the Heart and Stroke Foundation of Ontario to Peter Zandstra.
Biological safety cabinet | |||
Pipette aid | |||
Serological pipettes (5 to 25 mL) | |||
Aspirator | |||
Aspirator or Pasteur pipettes | |||
15 and 50 mL conical tubes | |||
Fume hood | |||
0.22 µm syringe filter | |||
5% CO2, 5% O2, and humidity controlled cell culture incubator | Hypoxic (low oxygen) incubator | ||
5% CO2, 20% O2, and humidity controlled cell culture incubator | |||
Low speed centrifuge with a swinging bucket rotor fitted with a plate holder | |||
P2, P20, P200, and P1000 micropipettors and associated tips | |||
Inverted microscope with 4X, 10X and 20X phase objectives | |||
Ultra-Low Attachment (ULA) 24 well plates | Corning/Costar | 3473 | |
1.5 mL microcentrifuge tubes | |||
Bench-top microcentrifuge | |||
L-Ascorbic Acid | Sigma-Aldrich | A4403 | |
Sterile Ultrapure distilled water | Sigma-Aldrich | W3500 | |
Vortex | |||
Ice | |||
-20C freezer | |||
Monothioglycerol | Sigma-Aldrich | M6145 | Toxic; Aliquoting of MTG is strongly recommended to minimize oxidation due to repeated opening. Aliquots can be stored at 4 °C for up to 3 months, -20 °C is recommended for long-term storage. |
StemPro-34 Medium | Thermo Fisher Scientific | 10639-011 | Basal Cardiac Induction Medium; The supplement is stored at -20 °C and the basal medium at 4 °C. |
Transferrin | Roche | 10652202001 | |
BMP-4 | R&D Technologies | 314-BP | |
bFGF | R&D Technologies | 233-FB | |
VEGF | R&D Technologies | 293-VE | |
Activin A | R&D Technologies | 338-AC | |
IWP-2 | Reagents Direct | 57-G89 | |
Phosphate buffered saline (PBS) | Thermo Fisher Scientific | 14190 | |
Bovine Serum Albumin (BSA) | Thermo Fisher Scientific | 15561 | |
Hydrochloric acid | Sigma-Aldrich | 258148 | Corrosive |
Dimethylsulfoxide (DMSO) | Sigma-Aldrich | D2650 | |
DMEM/F-12 | Thermo Fisher Scientific | 12660 | |
100X Penicillin/Streptomycin | Thermo Fisher Scientific | 15140 | |
100X L-glutamine | Thermo Fisher Scientific | 25030 | |
Knockout Serum Replacement | Thermo Fisher Scientific | 10828010 | |
TrypLE Select | Thermo Fisher Scientific | 12563 | Dissociation enzyme |
Hemocytometer | |||
Trypan Blue | |||
Aggrewell 400 plates | StemCell Technologies | 27845 | Microwell Plates |
Aggrewell Rinsing Solution | StemCell Technologies | 7010 | Microwell Rinsing Solution |
Y-27632 ROCK Inhibitor | Tocris | 1254 | |
Collagenase Type II | Sigma-Aldrich | C6885 | |
Hank's Balanced Salt Solution | Thermo Fisher Scientific | 14025092 | |
Fetal Bovine Serum | Thermo Fisher Scientific | 12483 | |
96 well plate (for FACS staining) | |||
Intraprep Permeabilization Reagent | Beckman Coulter | IM2389 | Kit with 2 parts: Fixation Solution and Permeabilization Solution; Toxic |
cTnT antibody | Neomarkers | MS-295 | |
goat anti-mouse-IgG APC antibodyThermoFisher | Molecular Probes | A865 | |
5 mL round bottom flow cytometry tubes | FACS machine dependent |