Large‐scale generation of megakaryocytes from human embryonic stem cells using transgene‐free and stepwise defined suspension culture conditions

Abstract Objectives Ex vivo engineered production of megakaryocytes (MKs) and platelets (PLTs) from human pluripotent stem cells is an alternative approach to solve shortage of donor‐donated PLTs in clinics and to provide induced PLTs for transfusion. However, low production yields are observed and the generation of clinically applicable MKs and PLTs from human pluripotent stem cells without genetic modifications still needs to be improved. Materials and Methods We defined an optimal, stepwise and completely xeno‐free culture protocol for the generation of MKs from human embryonic stem cells (hESCs). To generate MKs from hESCs on a large scale, we improved the monolayer induction manner to define three‐dimensional (3D) and sphere‐like differentiation systems for MKs by using a special polystyrene CellSTACK culture chamber. Results The 3D manufacturing system could efficiently generate large numbers of MKs from hESCs within 16‐18 days of continuous culturing. Each CellSTACK culture chamber could collect on an average 3.4 × 108 CD41+ MKs after a three‐stage orderly induction process. MKs obtained from hESCs via 3D induction showed significant secretion of IL‐8, thrombospondin‐1 and MMP9. The induced cells derived from hESCs in our culture system were shown to have the characteristics of MKs as well as the function to form proPLTs and release PLTs. Furthermore, we generated clinically applicable MKs from clinical‐grade hESC lines and confirmed the biosafety of these cells. Conclusions We developed a simple, stepwise, 3D and completely xeno‐free/feeder‐free/transgene‐free induction system for the generation of MKs from hESCs. hESC‐derived MKs were shown to have typical MK characteristics and PLT formation ability. This study further enhances the clinical applications of MKs or PLTs derived from pluripotent stem cells.


| INTRODUC TI ON
Platelet (PLT) transfusion is an important therapeutic approach for patients with life-threatening thrombocytopenia. However, there is a severe supply-demand imbalance for PLT transfusion due to the shortage of donations from human volunteers and the increasing demand for the transfusion process in clinics. 1,2 PLTs are primarily generated from megakaryocytes (MKs) and their progenitors in the bone marrow (BM) after birth. [3][4][5] It is difficult, rather impossible, to obtain MKs from normal BM for in vitro PLT production due to their rarity and the painful surgery manipulation. To solve the PLT shortage problem, an alternative approach involving ex vivo engineered production of MKs from human stem cells can be applied. Along with the subsequent generation of PLTs from MKs for transfusion therapy, the induced MKs or their progenitors directly infused into human patients have been accepted as a therapeutic strategy for prophylaxis or treatment of thrombocytopenia. [6][7][8] Thus, stem cellderived MKs and PLTs serve as advanced therapeutic and medicinal products.
The generation of functional MKs and PLTs from hematopoietic stem and progenitor cells (HSPCs) derived from cord blood and BM has been successfully employed in laboratories. 9,10 The limited number and proliferation capacity of HSPCs in these samples have impeded the development of large-scale and standard production of MKs and PLTs. It is known that human pluripotent stem cells, including human embryonic stem cells (hESCs) and induced pluripotent stem cells, have an indefinite expansion capacity. These cells might act as ideal seed cells for in vitro production of large-scale and donorless MKs and PLTs. 11,12 Several studies have shown that hESCs can produce MKs and PLTs using stromal cell co-cultures or embryonic body induction process. [13][14][15] Recently, expandable MK cell lines by overexpression of several genes in pluripotent stem cells have been accepted as the most promising cellular resource for clinically applicable and large-scale generation of PLTs. 16,17 However, low yields of such clinically applicable MKs and PLTs are obtained from hESCs without undergoing genetic modification and the process needs to be further improved.
Here, we defined an optimal, stepwise, and completely defined xeno-free culture method for the generation of MKs from hESCs.
For large-scale generation of MKs from hESCs, we improved the monolayer induction process to define three-dimensional (3D) and sphere-like differentiation models for MKs during the whole process by using a polystyrene CellSTACK culture chamber. The induced cells derived from hESCs in our culture system were shown to exhibit the characteristics of MKs, as well as, proPLT formation and PLT release. The 3D induction protocol could generate

| Human embryonic stem cell culture
The H9 hESC line (from WiCell Research Institute) was cultured in a chemically defined mTeSR1 TM medium on matrigel-coated wells as previously described. 18 Medium changes were performed daily, and confluent cultures were passaged every 4 -6 days using ReLeSR TM .
All cultures were maintained at 37°C in a 5% CO 2 incubator (Thermo Fisher Scientific). Mycoplasma contamination was tested every 2 weeks using the MycoAlert TM Mycoplasma Detection Kit (Lonza).

| Megakaryocyte differentiation
For monolayer induction, single-cell suspensions of hESCs were plated at a density of 1 × 10 4 cells/cm 2 onto 6-well plates.
For induction of the 3D suspension culture, hESC colonies were dissociated into single cells by Accutase TM on day −1. After harvesting the cells and counting them, the cells were resuspended in culture medium supplemented with the Rho kinase (ROCK) inhibitor Y27632 (10 μmol/L) and plated at a density of 10 5 cells/mL in an ultra-low CellSTACK TM culture chamber. On day 0, hESC aggregates were collected and resuspended in differentiation medium 1. On day 2, the EBs were collected and resuspended in differentiation medium 2. On day 6, the EBs were collected and resuspended in differentiation medium 3. The medium was changed daily during the induction process. On day 16-18, the suspended single cells generated from adherent cells or EBs were collected by passing through a 40 μm cell strainer for hematopoietic cell analysis.  Flow cytometry data were analysed using the FlowJo software.

| Flow cytometry analysis
The information of antibodies is described in detail in Table S2.
PLT microaggregates in 50 µL buffer were spread onto glass slides and visualized under a fluorescence microscope (GE Healthcare).

| CFU-MK assay
Colony-forming unit (CFU)-MK assay was performed by plating 2 × 10 4 cells/well from differentiated MKs following manufacturer's protocol (STEMCELL). Cells were incubated at 37°C in a 5% CO 2 incubator, and MK colonies were fixed and stained with CD41 antibody at day 14. CD41 + colonies were counted and classified following manufacturer's criteria.

| Statistical analysis
Data are presented as mean ± SEM. Statistical analyses were performed using the GraphPad Prism software. The statistical significance of the differences was determined using unpaired two-tailed Student's t-tests. Values with P < .05 were considered statistically significant.

| Generation of high purity megakaryocytes in an optimized stepwise 2D induction process
To efficiently generate MKs from hESCs in a xeno-free system, we developed a monolayer, adherent and stepwise induction protocol without feeder cell corporation ( Figure 1A). The hESCs (Q-CTS-hESC-2 and H9 cells) were digested into single cells and plated on wells at a low seeding density (1 × 10 4 cells/cm 2 ). These hESCs presented typical well-defined colony morphology, normal karyotype, positive alkaline phosphatase staining, and expressed the pluripotency-related proteins OCT4, SOX2, and NANOG ( Figure 1B and Figure S1A-C). After induction for 2 days in differentiation medium 1, the cells exhibited high expression levels of key mesoderm marker proteins, such as BRACHYURY (BRA) and ROR2 ( Figure S1D). To refine the hemogenic differentiation potential of these mesoderm progenitor cells, the cells were subsequently cultured using optimized cytokines and small molecule supplements (differentiation medium 2 supplemented with bFGF, VEGF and SB431542) for 3-4 days.
After this induction period, the differentiated cells were digested and sorted for CD34 + cells to enrich for hemogenic cells. The isolated CD34 + cells were subsequently cultured in differentiation medium 3 to attain hematopoietic and MK specification. After 10-12 days of induction, more suspension cell clusters were grown from the adherent cells  Figure 1G). However, the monolayer differentiation protocol led to a low MK yield at approximately 5 MKs/hESCs.

| Large-scale generation of megakaryocytes by a 3D suspension induction method
The manipulation process for two-dimensional (2D) MK generation was complex and produced a low yield. To obtain large numbers of MKs for clinical application, we further developed a 3D suspension differentiation protocol using a CellSTACK culture chamber. The CellSTACK culture chamber has a large culture bottom area of 636 cm 2 and an ultra-low attachment surface, which is recommended for the large-scale expansion of suspension cells (Figure 2A). To initiate a suspension culture process, the hESC colonies were digested into single cells and transferred onto a polystyrene CellSTACK chamber supplemented with E8 medium and Y27632 for 1 day (Figure 2A Figure 2D,E). The 3D differentiation protocol led to a cellular yield of (11.4 ± 2.5) × 10 6 CD41a + MKs/1 × 10 6 hESCs and (7.3 ± 3.0) × 10 6 CD41a + CD42b + MKs/1 × 10 6 hESCs,0.6fold higher than the CD41a + CD42b + cellular yield of monolayer induction ( Figure 2E), which also presented high cellular viability and maintained the cell numbers both before and after preservation for 6 months in liquid nitrogen ( Figure 2F). Within 16-18 days after stepwise induction, (3.4 ± 2.5) × 10 8 CD41a + MKs could be obtained from each CellSTACK chamber. Approximately 1 × 10 9 MKs could be generated using 4 CellSTACK chambers. The induction procedure was easy to manipulate and relatively simple compared to the monolayer adherent induction method. More importantly, large numbers of human-induced MKs could be produced within a 16-18-day culture period from hESCs by using more CellSTACK chambers.

| Characteristics of MKs derived from hESCs after 3D induction
We evaluated the characteristics of MKs derived from hESCs after the 3D induction process. The CFU-MK assay showed that

| Cytokine expression and secretion by MKs derived from hESCs in a 3D-induced manner
It has been reported that MKs and PLTs can secret various cytokines with multiple functions. [22][23][24][25][26] To dissect the secretion of cytokines by these differentiated MKs derived from hESCs, we profiled the supernatants from the 3D or monolayer-induced cells using a cytokine array.
We found that over 30 cytokines were secreted by these differentiated MKs ( Figure 4A

| Evaluation of clinically-applicable MKs derived from hESCs
To generate clinically applicable MKs from hESCs, the use of animal resources should be avoided in the agents used in the differentia-  (Table 1). BSA was undetected, and no endotoxin and cytokine residues were detected in the supernatants of the suspended single cells after washing 3 times. We also evaluated the safety of the differentiated cells in vivo. The mice injected with these cells via the blood or subcutaneously showed no teratoma formation after 8-week inoculation. The organs of these mice, such as bone marrow, liver, heart, brain, and spleen, showed normal morphology and tissue structure (data not shown). However, significant tumours were found in mice inoculated with HeLa cells ( Figure S3).
These results indicate that suspended single cells generated from hESCs after 3D induction are clinically biosafe.

| CellSTACK-cultured MKs induced from hESCs produce proPLTs and PLTs
To investigate whether these MKs derived from hESCs in a 3D induction manner can differentiate into proplatelets (proPLTs) and PLTs, the cells were transferred into maturation medium with SCF, TPO, and GM6001 and rotated on a plate at 10 rpm for 48 hours ( Figure 6A). We found that some MKs began to develop into proPLTforming cells, exhibiting thick cytoplasmic projections ( Figure 6B).
The proPLT-forming cells expressed CD41, CD42 and β1-TUBULIN ( Figure 6C), indicating the existence of a proPLT-forming cell phenotype. The supernatants were collected for PLTs, and their PLT characteristics were evaluated by immunofluorescence staining.
The results showed that the induced PLTs strongly expressed CD42 and were co-expressed with CD41, CD110, PF4, VWF and phalloidin ( Figure 6D). We used flow cytometry to analyse the surface markers of these differentiated PLTs. We found that >80% of the cell debris was co-expressed with CD41a and CD61, and >50% of the cell debris co-expressed with CD41a and CD42b ( Figure 6E), which showed a phenotype similar to that of peripheral blood PLTs. These PLTs presented similar aggregation capacity on fibrinogen compared with blood PLTs ( Figure 6G). After thrombin stimulation, both the induced and natural PLTs showed increased CD62P expression and clot aggregation ( Figure 6F,H), indicating that an active state of PLTs could be induced by thrombin.

| D ISCUSS I ON
The frequency of MK cells in bone marrow is only about 0.01 % of total nucleated cells, and they are derived from hematopoietic stem and progenitor cells. 27 MK is the seed cell that generates PLTs and maintains normal platelet levels in peripheral blood by generating up to 3000 PLTs per MK. 28 In certain disease conditions, such as   In summary, we developed a defined 3D differentiation protocol for the generation of MKs from hESCs using a special polystyrene CellSTACK culture chamber. The 3D manufacturing system could efficiently generate large numbers of MKs from hESCs after three stages of orderly induction within a culture period of 16-18 days. The induced cells derived from hESCs in our culture system were shown to have the characteristics of MKs, as well as, proPLT formation and PLT release.
Importantly, we generated clinically applicable MKs from clinical-grade hESC lines and analysed the biosafety of these cells. The current protocol is a simple, practicable and upscalable method that does not require feeder cells or genetic modification, further enhancing the clinical application of MKs or PLTs derived from pluripotent stem cells.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R CO NTR I B UTI O N S
The contributions of each author made to the study are specified as YHL and BWZ prepared the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analysed during this study are included in the published article and its supplementary information files.