A novel chemically directed route for the generation of definitive endoderm from human embryonic stem cells based on inhibition of GSK-3

Abstract

The use of small molecules to 'chemically direct' differentiation represents a powerful approach to promote specification of embryonic stem cells (ESCs) towards particular functional cell types for use in regenerative medicine and pharmaceutical applications. Here, we demonstrate a novel route for chemically directed differentiation of human ESCs (hESCs) into definitive endoderm (DE) exploiting a selective small-molecule inhibitor of glycogen synthase kinase 3 (GSK-3). This GSK-3 inhibitor, termed 1m, when used as the only supplement to a chemically defined feeder-free culture system, effectively promoted differentiation of ESC lines towards primitive streak (PS), mesoderm and DE. This contrasts with the role of GSK-3 in murine ESCs, where GSK-3 inhibition promotes pluripotency. Interestingly, 1m-mediated induction of differentiation involved transient NODAL expression and Nodal signalling. Prolonged treatment of hESCs with 1m resulted in the generation of a population of cells displaying hepatoblast characteristics, that is expressing α-fetoprotein and HNF4α. Furthermore, 1m-induced DE had the capacity to mature and generate hepatocyte-like cells capable of producing albumin. These findings describe, for the first time, the utility of GSK-3 inhibition, in a chemically directed approach, to a method of DE generation that is robust, potentially scalable and applicable to different hESC lines.

Fig. 1.

Treatment of hESCs with GSK-3 inhibitors induces differentiation. Shef-3 hESCs were treated with BIO, 1m or vehicle (DMSO) or left untreated (UT) and cultured for 7 days on either MEFs or Matrigel in mTeSR1 medium. (A) Images show the typical colonies that are formed. Scale bar: 1 mm. (B) hESCs were analysed by flow cytometry following immunostaining with antibodies against the pluripotency markers Tra-1-60 and SSEA4. Data show the mean percentage of positive cells (±s.e.m.) from at least three independent experiments. Statistical analysis was conducted using ANOVA and Dunnett's post hoc test to compare each treatment with the untreated control cells. *P<0.05; **P<0.01. An example of a histogram plot from a representative experiment is shown in supplementary material Fig. S2A. (C) RNA was extracted from the cells and RT-PCR analyses were performed using primers specific to the pluripotency genes OCT4 and NANOG and to the house-keeping β-actin-encoding gene. (D) Cell lysates (20 μg) were separated by SDS-PAGE and immunoblotting was performed using an antibody against OCT4. Blots were stripped and re-probed with anti-GAPDH antibodies to assess equal loading.

Fig. 2.

Treatment of hESCs with 1m inhibits GSK-3. (A,B) Shef-1 hESCs, cultured on MEFs or Matrigel in mTeSR1 medium, were treated with BIO or 1m for 30 minutes. Immunoblotting was performed to detect phosphorylated forms of β-catenin and ERK1/2. The same immunoblot in each case was re-probed for total β-catenin and ERK1 to assess loading. The bar graph shows the mean relative β-catenin phosphorylation levels (+s.e.m.; n=3). Data were analysed for statistical significance using two-tailed paired t-tests. *P<0.05; **P<0.01. (C) TOPFlash luciferase reporter assay following 24 hours of treatment of Shef-1 hESCs with BIO, 1m or vehicle (DMSO). Luciferase activity is expressed relative to the normalised TOPFlash luciferase activity in untreated (UT) control cells. A control reporter (FOPFlash), containing mutant TCF-binding sites, was also run in parallel. Data are means±s.e.m. for three independent experiments.

Fig. 3.

1m treatment induces differentiation towards definitive endoderm. Shef-3 hESCs were cultured for a total of 7 days, feeder-free on Matrigel, in chemically defined mTeSR1 medium with 2 μM 1m included in the medium for the number of days indicated (0d, hESCs cultured for 7 days without 1m). (A) RT-PCR analyses were performed using primers selective for the genes specified. (B) Shef-3 hESCs cultured under the conditions indicated for the duration of the experiment in Lumox 24-well trays were analysed by immunofluorescence for expression of brachyury, PDGFRβ, FOXA2, SOX17 and HNF4α. Cells were counterstained with DAPI for nuclear localization and the lower panels show the superimposed images. Scale bars: 50 μm. DMSO controls were cultured for the entire 7-day period.

Fig. 4.

Nodal signalling is involved in 1m-induced differentiation of hESCs. (A) Shef-3 hESCs were cultured for a total of 7 days with 2 μM 1m included in the medium for the number of days indicated and RT-PCR analyses were performed using primers selective for the genes specified. 0d, hESCs cultured for 7d without 1m. (B) Shef-1 hESCs were cultured for a total of 7 days and either treated with 2 μM 1m alone or in combination with 10 μM SB43125 (SB) for the number of days indicated and RT-PCR performed. Con, cultures to which inhibitors were not added. (C) Shef-3 hESCs were treated with 1m alone or in combination with 100 ng/ml activin A (Act) for the times indicated and RT-PCR performed. Con, cultures to which factors were not added. (D,E) Shef-3 hESCs were treated with 2 μM 1m or 100 ng/ml activin A (Act) alone or in combination with 1m for 5 days. Control cultures were grown for the same 5-day period without addition of activin A or 1m. (D) Expression of the endodermal markers FOXA2 and HNF4α were analysed by immunofluorescence. Scale bar: 50 μm. The mean percentage of positive cells (±s.e.m.) over three independent experiments is indicated. (E) Expression of CXCR4 was analysed by flow cytometry. The mean percentage of CXCR4-positive cells (±s.e.m.) for eight individual experiments is indicated.

Fig. 5.

1m-induced DE has hepatic potential. Shef-3 hESCs were treated with 2 μM 1m for 7 days to induce differentiation into DE. Hepatic induction and maturation conditions were then applied (see Materials and Methods). (A) Schematic representation of the differentiation protocol. (B) Images of differentiated cells at stage II (day 14) and stage III (day 25). Scale bars: 50 μm. The bottom panel shows a higher magnification of the boxed area of stage III cultures. (C) RT-PCR analyses of RNA prepared from cells at the indicated stages of differentiation. RNA extracted from human liver (hLiver) was used as a positive control. The β-actin-encoding gene was used as a housekeeping gene. (D) Expression of the early hepatic markers HNF4α, AFP and TTR in cells at stage II and stage III analysed by immunofluorescence. Scale bar: 50 μm. (E) Medium (1 μl) from stage III (day 25) cells, conditioned for 48 hours, and from a medium alone (MA) control was analysed by immunoblotting with an antibody against AFP. Purified AFP and medium conditioned from the Huh7 human hepatocarcinoma cell line were run as controls. (F) ELISA assay for human albumin from stage III (day 25) cell medium. The limit of detection is 12.5 ng/ml. Results are mean+s.d. (n=2 assay repeats). Medium conditioned for 24 hours by primary human hepatocytes is included as a positive control.