Patent Application
20180237745
The present invention provides methods to control the differentiation of pluripotent stem cells. In particular, the present invention provides methods for controlling the differentiation of pluripotent stem cell toward endoderm originated cells.
Stem cells, having the ability to differentiate into many different cell types that make up an organism, are able to differentiate into derivatives of all three embryonic germ layers (endoderm, mesoderm and ectoderm).
A considerable amount of interest has been generated in stem cell committed differentiation. The first intermediate stage of differentiation is the formation of definitive layer type cells. For the purpose of obtaining pure targeted cells, it is advantageous that undesired types of cells are prohibited from differentiation. However, very limited study results have been revealed for the inhibition of differentiation toward definite type of cells.
The present invention is based on the discovery that endogenous TGFβ activated by Activin A drives epithelial-to-mesenchymal transition (EMT) and definitive endoderm (DE) formation in a SNAI1 dependent fashion. Blocking of SNAI1 or the introducing of TGFβ inhibitor will block the EMT of the stem cells, thereby inhibit the stem cells from the acquisition of endoderm character during differentiation. Therefore, it is controllable that no DE originated cell will be formed. As the DE layer type cells are prohibited at the beginning of differentiation, it is more efficient to produce pure cells of other layer type by differentiation.
Cell fate decision during somatic cell reprogramming has been shown to involve transitions between mesenchymal (M) and epithelial (E) states. It is unclear if similar decisions occur during the differentiation from pluripotent to somatic states.
As a basic study for the present invention, we tested this by mapping cell fate changes during the hepatic differentiation of hESCs with bulk RNA-seq and showed that hESCs start differentiation with a synchronous EMT followed by a less synchronous MET before acquiring liver-specific characteristics. We confirmed these changes at single cell resolution and further showed that the EMT is mediated by endogenously produced TGFβ through SNAI1 to initiate differentiation towards hepatocytes. Our study establishes epithelial-mesenchymal-epithelial transitions as a mechanism that drives cell fate decisions during hESC differentiation towards the haptic lineage, and further suggests that differentiation and reprogramming share similar cell fate decisions.
Reprogramming of somatic cells into pluripotent ones with defined factors not only provides a new way to generate functional cells for regenerative medicine, but also establishes a new paradigm for cell fate decisions. For the latter, a cell at a terminally differentiated state can be restored back to pluripotency under well-defined conditions fully observable through molecular and cellular tools. Indeed, the reprogramming process has been analyzed in great detail to reveal novel insights into the mechanism of cell fate changes1-3. Of particular interest is the acquisition of epithelial characteristics from mesenchymal mouse embryonic fibroblasts (MEFs) commonly employed as starting cells in reprogramming experiments4. Termed the MET or mesenchymal to epithelial transition, we and others have described the MET as marking the earliest cellular change upon the simultaneous transduction of reprogramming factors POU5F1 (OCT4), SOX2, KLF4 and MYC into MEFs5-6. However, when delivered sequentially as OK+M+S, they initiate a sequential EMT-MET process that drives reprogramming more efficiently than the simultaneous approach7, suggesting that the switching between mesenchymal and epithelial fates underlies the reprogramming process, i.e., the acquisition of pluripotency. We then speculated that such a sequential EMT-MET process may underlie cell fate decisions in other situations such as differentiation, generally viewed as the reversal of reprogramming with the loss of pluripotency. Here we report that a sequential EMT-MET also drives the differentiation of hESCs towards hepatocytes.
In the present invention, it is reveled that differentiation of human embryonic stem cells towards hepatocyte-like cells undergoes a sequential EMT-MET process: A TGF□ dependent mesenchymal fate transition during the differentiation of human embryonic stem cells towards DE originated cell, e.g., hepatocyte-like cells.
Hereby a method of controlling differentiation of stem cells is provided, comprising
Said method can be used for restraining said stem cells from forming endoderm originated cells, e.g., hepatic cells, biliary ductal cells, exocrine pancreatic cells, endocrine pancreatic cells, lung cells, thyroid cells, as well as cell types of the gastrointestinal tract, during differentiation.
Said stem cells are mammalian stem cells; preferably, human embryonic stem cells or iPS cells.
Optionally, said TGF inhibitor is TGFβ inhibitor selected from the group consisting of TGFβ signaling inhibitor, TGFβ expression inhibitor, TGFβ interfering agent, TGFβ blocking antibody. As exemplified, the TGFβ inhibitor is selected from the group consisting of RepSox, SB525334.
Optionally, said TGFβ inhibitor is added with a final concentration of no less than 1 μM in the culture medium; preferably 1-20 μM, more preferably 1-10 μM. In one preferred example, it is added with a final concentration of 2 μM.
Optionally, said SNAI1 inhibitor is siRNAs to SNAI1.
In the meanwhile, a method of inhibiting the differentiation of stem cells toward endoderm originated cells is provided, comprising inhibiting the epithelial mesenchymal transition within 3 days from the beginning of differentiation, thereby inhibiting the stem cells from the acquisition of endoderm character.
On the other hand, a method of promoting the differentiation of stem cells toward endoderm originated cells is provided, comprising a stimulation of epithelial-mesenchymal transition followed by a stimulation of mesenchymal-epithelial transition.
Related therapeutic medicine can be developed using the inhibitor disclosed in the present invention.
We presented evidence here that a sequential EMT-MET process occurs during the differentiation of hESCs towards hepatic lineage, mirroring a similar mechanism for somatic cell reprogramming7. These findings may help provide a unified understanding of cell fate decisions in both reprogramming and differentiation. Since reprogramming is generally viewed as the reversal of differentiation, this unified mechanism appears to be quite reasonable as cells must pass through similar phases, albeit in opposite directions, hence involve similar processes and regulators. In agreement, lineage specifiers have been reported to serve as potent reprogramming factors13, 14. Further studies should be directed at elucidating the underlying mechanisms that can orchestrate these cellular and molecular processes during both reprogramming and differentiation in terms of cell fate directions. On the other hand, the transitions between mesenchymal and epithelial states, i.e., EMT and MET, have been observed both in vivo and in vitro15-17. Their coupled role in mediating cell fate decision thus warrant further analysis at multiple levels both in vivo and in vitro and may provide insights into not only normal development, but also disease processes as well.
(A) Schematic of the differentiation protocol and the stages of cell differentiation. Representative live photos as well as images of immunofluorescence staining of marker molecules at the indicated stages were shown. Scale bar is 20 μm.
(B) Principal component analysis indicates the transitions that occur during differentiation. PC2 roughly corresponds to hESC to liver acquisition and PC3 corresponds to an epithelial/mesenchymal phenotype.
(C) Expression of selected marker genes from the RNA-seq data. Gene expression is normalized across the mean of expression over the time course. From top to bottom, the marker genes are: pluripotency, DE, hepatoblast, hepatocyte and EMT/MET.
(D) Broad waves of EMT, differentiation and acquisition, based on the marker genes defined in (C).
(A) Dynamic expressions of CDH1 and CDH2 at the indicated stages of hepatic differentiation of hESCs.
(B) An E-M-E conversion of cells during the hepatic differentiation. The ratio of CDH1 to CDH2 was used as an indicator of the epithelial/mesenchymal states of cells.
(C) Relative expression levels of mesenchymal genes at D0 and D3. The expression level of GAPDH was arbitrary set as 1.
(D) Immunofluorescence staining of CDH1 and CDH2 at the indicated differentiation stages.
(E) Rhodamine-conjugated phalloidin staining of F-actin in day 0 and day 3 cells.
(F-G) Migration assay for hESC and DE. Scale bars represent 20 μm in (D and E) and 100 μm in (F).
(A) Heatmap of the expressions of selected genes.
(B) Relational network plots. Different colors indicate specific days of treatment, node sizes are 2[relative expression]. Open arrows indicate a population of day 5/7 hESC-like cells expressing POU5F1, SOX2, NANOG. Closed arrows indicate a population of cells simultaneously expressing the pluripotent marker genes POU5F1, SOX2, NANOG and the DE markers SOX17, GATA4 and GATA6.
(C) Scatter plot of CDH1 versus CDH2 expression, colors are the same as in panel (B).
(D) Correlation of gene expression for the indicated genes for days 0 through 3.
(E) Scatter plots of single cell gene expression for SOX17 and GATA6 versus the epithelial marker gene CDH1 and the mesenchymal marker gene CDH2. Colors are the same as in panel (B).
(F) Immunofluorescence staining of CDH1, CDH2 and SOX17 at day 3.
(A) Secreted protein levels of TGFβ in culture medium.
(B) Single-cell qPCR of selected marker genes on cells treated with RepSox.
(C) Relational maps, constructed as in
(D) Bulk qPCR analysis for a selection of marker genes.
(E) Immunofluorescence staining of day 3 cells with or without RepSox. Scale bar indicates 20
(F) Migration ability of day 3 cells. Assays were performed as in
(G) qRT-PCR analysis for the efficiencies of siRNAs to the indicated genes.
(H) Treatment of cells with siRNAs to SNAI1 effectively blocked the induction of SOX17 and FOXA2.
(A) Immunofluorescence staining for Cytochrome P450 (Red) and Alpha-1-antitrypsin (AAT, Green).
(B) PAS staining for glycogen.
(C) Uptake of Alexa Fluor® 488 labeled LDL (Green).
(D, E) Uptake and release of indocyanine green (ICG).
The expression level of GAPDH is arbitrary set as 1. Data represent mean±SD from three independent biological repeats.
The expression level of these genes in the absence of siRNA treatment is defined as 1. Data represent mean±SD from three independent biological repeats.
During the differentiation of hESCs toward hepatocyte-like cells, they undergo the fate of EMT-MET. Specifically, hESCs express stem cell marker genes POU5F1 and NANOG, and at the same time, they express epithelial cell marker gene CDH1. SOX17 and FOXA2 are markers of definitive endoderm (DE) cell, which in the meantime expresses mesenchymal cell marker gene CDH2. HNF4A and AFP are markers of hepatocytes (HB) cell, which in the meantime express CDH1 and CDH2, which are marker gene of epithelial cell and mesenchymal cell. ALB and TTR are markers of mature hepatocytes (HC), which in the meantime express epithelial Cell marker gene CDH1. During the early stage of the hepatic differentiation phase, endogenous TGFβ that is activated by Activin A promotes the epithelial-mesenchymal transition (EMT), and relies on Snail1 to form definitive endoderm (DE).
English or English abbreviations involved in the drawings:
Day: Day Activin A: Activin A Definitive endoderm: Definitive endoderm
Hepatoblast: Hepatoblast Hepatocyte: Hepatocyte
Change around mean of expression: Change around mean of expression
Average expression of selected markers: Average expression of selected markers
Pluripotency: Pluripotency Epithelial genes: Epithelial genes
Mesenchymal genes: Mesenchymal genes
Relative gene expression: Relative gene expression
Migration distance: Migration distance 24 hr: 24 hours 48 hr: 48 hours
Relative expression: Relative expression Correlation: Correlation
TGFβ1 protein: TGFβ1 protein Blank:Blank Activin A: Activin A
Relative expression: Relative expression Day: Day Fold change: Fold change Migration (um): Migration (um) Relative mRNA level: Relative mRNA level
ICG Uptake: ICG Uptake ICG Release: ICG Release
Activin A: Activin A Day 3: Day 3 Relative mRNA level: Relative mRNA level
For clarity of disclosure, and not by way of limitation, the detailed description of the invention is divided into the following subsections that describe certain features, methods, embodiments or applications of the present invention.
Undifferentiated human H1 ES cells were maintained in monolayer culture on Matrigel (BD Biosciences, 354277) in mTeSR1 medium (Stemcell Technologies, 05850) at 37° C. with 5% CO2. Cells were manually passaged at 1:4 to 1:6 split ratios every 3 to 5 days. For hepatic differentiation, we established a serum-free protocol based on previously described protocols with minor modifications8, 9. Briefly, H1 cells were cultured for 3 days in RPMI/B27 medium (Insulin minus, Gibco, A18956-01) supplemented with 100 ng/ml Activin A (Peprotech, A120-14E), followed by 4 days with 20 ng/ml BMP2 (Peprotech, 120-02) and 30 ng/ml FGF-4 (Peprotech, 100-31) in RPMI/B27 (complete with Insulin, Gibco, 17504-044) medium, then 6 days with 20 ng/ml HGF (Peprotech, 100-39) and KGF (Peprotech, 100-19) in RPMI/B27 (complete with Insulin), then 8 days with 20 ng/ml Oncostatin-M (R&D Systems, 295-OM/CF) in Hepatocyte Culture Media (Lonza, cc-3198) supplemented with SingleQuots (without EGF). Immunofluorescence staining.
Cells on glass coverslips were fixed in 4% paraformaldehyde (PFA, Jingxin, GI001, Guangzhou) for 30 min, washed with PBS for 3 times and permeabilized in 0.3% Triton X-100 (Sigma, T9284)/PBS for 30 min. After two brief washes in PBS, cells were blocked in 10% FBS (ExCell Biology, FSP500)/PBS for 1 hr at RT. For actin staining, cells were incubated with 1 unit/ml rhodamine phalloidin (Invitrogen, R415) for 60 min at RT. For Immunofluorescence staining, samples were then incubated with primary antibody (diluted in 1×PBS with 0.3% Triton X-100/10% FBS) overnight at 4° C., washed 3 times with blocking solution and incubated with a secondary antibody for 1 hr at RT. The cells were washed with blocking solution and PBS, counter stained with DAPI (Sigma, D9542) for 5 min, and then imaged with the Zeiss LSM 710 confocal microscope (Carl Zeiss). Antibodies used in this study are listed in supplementary Table 51.
Total RNA was extracted using the Trizol reagent (MRC, TR1187) and 2 μg of total RNA was reverse-transcribed using the ReverTraAce Kit (TOYOBO, 34520B1). The product (cDNA) was properly diluted and used as PCR template. PCR reactions were performed with the SYBR®Premix Ex Taq™ Kit (TAKARA, RR420A) on the CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad). GAPDH was used as the internal control. Three to four biological replicates were performed for each assay and data represents mean±SD.
siRNA Treatment.
Three independent siRNAs for each gene were designed and synthesized by Ribobio (Guangzhou, China) to target human SNAI1, SNAI2, TWIST1, TWIST2 and ZEB1. Transfection of siRNAs was performed using the Lipofectamine RNAiMAX Reagent (Life Technologies, 13778-150) according to the manufacturer's protocol and the final concentration of each siRNA was 30 nM. Three consecutive transfections (every 48 hrs) were performed on 70˜80% confluent H1 cells maintained in mTeSR1. Cells were then treated with Activin A after the third transfection for three days as described above. The knockdown efficiency of siRNA was determined by real-time RT-PCR 96 hrs post the third transfection. Two of the most effective siRNAs to each gene were combined and used in further experiments.
Total RNA was harvested during the differentiation of hESCs to hepatocyte-like cells on days 0, 1, 2, 3, 5, 7, 9, 11, 13, and 21. Approximately 4 μg of RNA was used to generate sequencing-ready cDNA library with the TruSeq RNA Sample Prep Kit (Illumina, RS-122-2001). DNA fragments (250-300 bp) were recovered from the gel slice using the QIAquick gel extraction kit (QIAGEN, 28704). The concentration of cDNA library was determined with the Qubit® dsDNA HS Assay Kit (Invitrogen, Q32851). Samples were sequenced on a MiSeq according to the manufacturer's instructions to an average depth of 2 million sequence tags. Reads were aligned to the ENSEMBL (mm10 v76) transcriptome using Bowtie2 (v2.2.0)18 and RSEM (v1.2.17)19, GC-normalized using EDASeq (v2.0.0)20. Analysis was performed using glbase21. Reads were deposited with GEO under the accession number: GSEXXXXXX Single cell qPCR and analysis.
Single cell qPCR was performed using a Fluidigm C1 and BioMark HD as described by the manufacturer. Briefly, a cell suspension of a concentration of 166˜250 K/mL was loaded into a 10-17 μm C1™ Single-Cell Auto Prep IFC chamber (Fluidigm, PN100-5479), and cell capture was performed on the Fluidigm C1™ System. Both the empty wells and doublet-occupied wells were excluded from further analysis. Upon capture, reverse transcription and cDNA pre-amplification were performed using the Ambion Single Cell-to-CT Kit (Life Technologies, PN 4458237) and C1™ Single-Cell Auto Prep Module 2 Kit (Fluidigm, PN100-5519). The pre-amplified products were diluted 10-fold prior to analysis with TaqMan® Gene Expression Master Mix (Life Technologies, 4369016) and inventoried TaqMan® Gene Expression assays (20×, Applied Biosystems) in 96.96 dynamic Arrays™ on a BioMark System (Fluidigm). Inventoried TaqMan primers were used for single-cell qPCR. Relative expression was calculated as described in Buganim et al., 201222 except that a Ct value of 25 was used for low expressed genes. Relational network plots (mdsquish) were implemented as part of glbase21 and will be described in detail elsewhere. Briefly, the normalized Euclidean distance between all cells was measured for singular value decomposed principal components 2, 3, 4 and 5. A network was then constructed using a threshold of 0.92 for weak links (dotted lines) and 0.99 for strong links (solid lines) with a maximum of the 50 best scoring edges per node, the network was then laid out using graphviz ‘neato’ layout. Node sizes are 2[relative expression].
Scratch assay was used to determine the migration activity of H1-derived cells. Briefly, cells in a confluent monolayer were scratched with a needle to form a cell-free zone into which cells at the edges of the wound can migrate. The denuded area was imaged to measure the boundary of the wound at pre-migration. Images of cell movement were captured at regular intervals within a 24-48 hr period for data analysis.
The protein level of TGFβ in the Activin A stimulated H1 cell culture media was determined with an ELISA Kit (R&D Systems, DB100B) as described by the manufacturer.
PAS staining was performed using the PAS staining kit (Polysciences, 24200-1) according to the manufacturer's instructions.
Cells were washed with PBS and incubated in culture medium containing 4 μg/ml low-density lipoprotein (LDL) (Invitrogen, L23380) for 30 min at 37° C. Cells were then fixed with 4% formaldehyde and stained with DAPI (Sigma, D9542). LDL uptake by cells was examined under a fluorescence microscope.
Indocyanine green (ICG) (Sigma, 1340009) was dissolved in DMSO at 5 mg/ml. When cells were ready, ICG was diluted freshly in culture medium to 1 mg/ml. Diluted ICG was added to cultured cells for 30 min at 37° C. After washing with PBS, the cellular uptake of ICG was examined under a microscope. Then cells were refilled with the culture medium and incubated for 6 hrs and the release of cellular ICG was examined.
A Sequential EMT-MET Connects hESCs to Hepatocytes
With robust expression of E-cadherin (CDH1), hESCs should be considered as epithelial cells in a pluripotent state. Conversely, hepatocytes are also epithelial cells, but are somatic and fully differentiated. Thus, the generation of hepatocytes from hESCs should follow a process from one type of epithelial cells to another with the gradual loss of pluripotency and gain of hepatic characteristics, without the necessity to pass through a mesenchymal state. To map the cell fates along the differentiation pathway between hESCs and hepatocytes, we adopted a serum-free, chemically defined protocol of hepatic differentiation of hESCs based on the stepwise addition of Activin A, FGF4/BMP2, HGF/KGF and then Oncostatin M8, 9. As shown in
A Transitional Mesenchymal State During Differentiation
The mesenchymal state uncovered by bulk RNA-seq in
hESCs Begin Differentiation with a Near Synchronous EMT
The Bulk RNA-seq dataset presented in
Surprisingly, the synchrony appears to break down at day 5 with many cells simultaneously expressing CDH1 and CDH2 and the correlation with the bulk is also considerably reduced (R2=0.45), suggesting substantial heterogeneity. In agreement with the synchronous to asynchronous transition, the starting EMT is exclusive: cells at day 3 express only CDH2, but the following MET is heterogeneous and many cells express both CDH1 and CDH2 at day 5 and only a small number of cells at day 7 no longer express CDH2 (
Interestingly, at day 5 and day 7, we can find small numbers (11 and 12, respectively) of hESC-like cells. These cells uniformly express POU5F1, SOX2, NANOG and CDH1 and do not express HNF4A or SOX17 (
We then wished to resolve the relationship between EMT and the acquisition of DE markers. A previous report indicated that the EMT could be uncoupled from DE markers when EOMES was knocked-down11. Here we took advantage of the single cell-qPCR analysis, which when organized into relational maps suggests the temporal order of DE acquisition and EMT. When we clustered the single-cell qPCR correlated gene expression for days 0-3 (
We were intrigued by the rapid and synchronous EMT at the start of the differentiation process. However, Activin A is not a strong inducer of EMT so it might function through the stimulation of other EMT-inducing signals. We previously reported that TGFβ is a major barrier for reprogramming MEFs into iPSCs: It must be suppressed by reprogramming factors for the MET and somatic reprogramming to occur5. Our RNA-seq data indicated that TGFβ was strongly induced by Activin A treatment so the endogenously produced TGFβ might induce EMT during the formation of DE. To test this hypothesis, we measured TGFB1 gene expression by qRT-PCR and showed a robust activation of TGFB1 by Activin A at day 3 (
We further confirmed the role of TGFβ by analyzing EMT and other lineage markers by qRT-PCR and showed that RepSox blocked the expression of CDH2, SNAI1 SNAI2, ZEB1, KLF8 and VIM while maintaining the expression of CDH1, blocked the expression of mesendoderm/endoderm markers such as GSC, EOMES, MIXL1, SOX17, FOXA2, GATA4, GATA6, HHEX and LGR5 (
We then turned out attention to transcriptional factors up-regulated at day 3 and decided to focus on SNAI1 and TWIST family members based on RNA-seq dataset (
The following examples are provided for the purpose of illustrating the invention, and should not be construed as limiting.
Inhibition of differentiation of hESCs towards definitive endoderm using TGFβ inhibitor.
Culturing condition during differentiation is as follows. Nearly 50% confluent undifferentiated human H1 ES cells obtained from ATCC were passaged by Accutase 24 hr after seeding, cells were almost 90% confluent at 37° C. with 5% CO2. Then H1 cells were cultured for 3 days in RPMI/B27 medium (Insulin minus) supplemented with 100 ng/ml Activin A in the presence of 2 μM Repsox or not.
TGFβ-signaling inhibitor RepSox was added with a final concentration of 2 μM at day 0 from the start of differentiation of hESCs.
Expression of various markers of definite layer type cells, i.e., SOX1/PAX6 (ectoderm markers), GATA2/BMP4 (mesoderm markers), and SOX17/FOXA2 (endoderm markers) were tested during differentiation to evaluate the differentiated cells.
Techniques such as single cell qPCR and bioinformatics, immunofluorescence staining and real-time RT-PCR are used to evaluate the effect of TGFβ inhibitor on hepatic differentiation of hESCs. A detailed description of the methods is presented in the experimental description section.
The results are presented in
As control group, differentiation of hESCs was conducted using the same culturing conditions but without the addition of TGFβ-signaling inhibitor. Most of the hESCs-derived cells robustly expressed endodermal markers, such as SOX17, FOXA2, GATA4, and GATA6, while the markers of other layer type cells were hardly detected (
It can be concluded that the addition of TGFβ-signaling inhibitor can inhibit the differentiation of hESCs from the formation of definitive endoderm originated cells.
In the meanwhile, TGFβ inhibitor enhances the expression of ectoderm and mesoderm markers (
Inhibition of differentiation of hESCs towards definitive endoderm using siRNAs to SNAI1.
Culturing condition during differentiation is as follows. H1 cells after the third transfection were cultured for 3 days in RPMI/B27 medium (Insulin minus) supplemented with 20 ng/ml Activin A.
Expression of various markers of definitive layer type cells, i.e., SOX17, FOXA2, GSC, GATA4, and GATA6, were tested during differentiation to evaluate the differentiated cells.
Methods include siRNA treatment and real-time RT-PCR. A detailed description of the methods is presented in the experimental description section.
The results are presented in
| Number | Date | Country | Kind |
|---|---|---|---|
| 201510491300.9 | Aug 2015 | CN | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/CN2016/094380 | 8/10/2016 | WO | 00 |
