Directed pancreatic acinar differentiation of mouse embryonic stem cells via embryonic signalling molecules and exocrine transcription factors

Abstract

Pluripotent embryonic stem cells (ESC) are a promising cellular system for generating an unlimited source of tissue for the treatment of chronic diseases and valuable in vitro differentiation models for drug testing. Our aim was to direct differentiation of mouse ESC into pancreatic acinar cells, which play key roles in pancreatitis and pancreatic cancer. To that end, ESC were first differentiated as embryoid bodies and sequentially incubated with activin A, inhibitors of Sonic hedgehog (Shh) and bone morphogenetic protein (BMP) pathways, fibroblast growth factors (FGF) and retinoic acid (RA) in order to achieve a stepwise increase in the expression of mRNA transcripts encoding for endodermal and pancreatic progenitor markers. Subsequent plating in Matrigel® and concomitant modulation of FGF, glucocorticoid, and folllistatin signalling pathways involved in exocrine differentiation resulted in a significant increase of mRNAs encoding secretory enzymes and in the number of cells co-expressing their protein products. Also, pancreatic endocrine marker expression was down-regulated and accompanied by a significant reduction in the number of hormone-expressing cells with a limited presence of hepatic marker expressing-cells. These findings suggest a selective activation of the acinar differentiation program. The newly differentiated cells were able to release α-amylase and this feature was greatly improved by lentiviral-mediated expression of Rbpjl and Ptf1a, two transcription factors involved in the maximal production of digestive enzymes. This study provides a novel method to produce functional pancreatic exocrine cells from ESC.

Figure 1. Protocol for pancreatic acinar differentiation.

ESC were induced to differentiate during 19 days though 4 sequential stages including the generation of definitive endoderm and pancreatic progenitors as well as the formation and expansion of acinar progenitors. d (days); ActA (activin A); Cyc (cyclopamine); DM (dorsomorphin); Fol (follistatin); Dex (dexamethasone). In some experiments, ESC lines stably expressing GFP or Rbpjl were differentiated throughout the protocol. At the end of stage 2, cells were infected with a lentivirus expressing an ER-fused Ptf1a construct or GFP as control, and treated daily with Tamox until the end.

Figure 2. Gene expression of early germ-layer specific markers at stage 1 by qRT-PCR.

Cells were induced to differentiate as EB in the presence of 100 ng/ml activin A as indicated in Fig. 1. After 3 or 5 days (stage 1), cultures were harvested and subjected to qRT-PCR analysis for the indicated early germ-layer (A–C) and foregut/pancreatic (D) markers. Histograms show the relative expression levels normalized to the loading control Hprt. Error bars indicate the standard deviations of 4 experiments. (d), days; T, treated cells; NT, non-treated cells. In A–C, p is calculated as compared to D1 and in D, as compared to T3. In D, the expression was analyzed from day 3 onwards as some of the markers were undetectable at day 1.

Figure 3. Expression of pancreatic progenitor markers at stage 2.

A) Cells were induced to differentiate through progression of stages 1 and 2 with or without DM as indicated in Fig. 1. After 5 (stage 1) or 7 (stage 2) days, cultures were harvested and subjected to qRT-PCR analysis for the indicated markers. Histograms show the relative expression levels normalized to the loading control Hprt. Error bars indicate the standard deviations of 3 experiments. p values correspond to comparisons between T5 and T7 and between T7 and T7+ DM. NS, not significant. The effect of the pro-pancreatic agents was analyzed at the beginning and the end of the corresponding stage. B) Expression of non-pancreatic markers by qRT-PCR as indicated in A. C) Immunofluorescent staining for HNF1β (a) and Pdx1 (b) in 7-day cultures (stages 1 and 2), 12 hours after plating. Nuclei were stained in blue. The corresponding negative controls are shown in (a’) and (b’). Scale bars, 10 µm.

Figure 4. Expression of pancreatic differentiation markers by qRT-PCR at stage 4.

Cells were induced to differentiate through-out the whole protocol as indicated in Fig. 1 and analyzed by qRT-PCR for the expression of exocrine (A), endocrine (B) or hepatic (C) markers. Histograms show the relative expression levels normalized to the loading control Hprt. Error bars indicate the SEM of at least 3 experiments. Marker expression at day 19 (T19) was compared to non-treated cells after stage 2 induction and cultured during the same period of time (NT19) to specifically study the effect of the pro-exocrine molecules. p, as compared to NT19.

Figure 5. Immunofluorescent analysis of differentiated cell cultures.

Staining was performed for Chymo (a–g, j), Amyl (a–c, h–i), Cpa1 (d–e), Rbpjl (f), Pdx1 (g), Afp (h–i), Gys2 (j), Ins (k–l) and Gluc (k–l) in NT19 (a, d, h, k) and T19 cultures (b–c, e–g, i–j, l) as indicated. Nuclei were stained in blue. Negative controls (m–o) were performed with irrelevant antibodies against rabbit (r), mouse (m), goat (g) or guinea pig (gp) as indicated. Scale bars: a–b, 50 µm; c–o, 10 µm.

Figure 6. Characterization of transgene expression in undifferentiated and differentiated ESC lines.

A) Analysis of transgene expression in RBPL-ES. Undifferentiated RBPL-ES were stained by immunofluorescence with an anti-Rbpjl antibody or an irrelevant one (green) and with Tropo3 (red) to label nuclei. GFP expression in GFP-ES cells was analyzed by confocal microscopy. The engineered ESC lines displayed a normal karyotype and retained their self-renewal capacity (not shown). Scale bars, 50 µm. B) Rbpjl mRNA levels of clone # 50 were comparable to those of mouse adult pancreas by qRT-PCR. C) Immunofluorescence analysis of Ptf1a expression and relocalization in differentiating ESC infected with Lv-Ptf1a-ER and treated with DMSO (−) or with Tamox (+), two days after. Ptf1a expression is shown in green while the nuclei are stained in red. Asterisks (*) show nuclear Ptf1a staining in cells non-exposed to Tamox. Scale bars, 10 µm. D) qRT-PCR analysis of Cpa1 expression in GFP-ES cells differentiated through the protocol until the end of stage 3. Cells were infected with a control LvGFP or the Lv-Ptf1a-ER and incubated with or without Tamox. Ptf1a mRNA expression is also shown as an indicator of LvPtf1a-ER gene transduction. E) qRT-PCR analysis of ectopic Ptf1a mRNA expression at the end of the protocol in GFP-ES and RBPL-ES cultures infected with LvPtf1a-ER.

Figure 7. Digestive enzyme gene expression in transgenic GFP-ES and RBPL-ES differentiated throughout the whole protocol.

A) Analysis of digestive enzyme gene expression by qRT-PCR at the end of the protocol at the indicated culture conditions. T19 cultures of ESC and GFP-ES infected with LvGFP showed no significant differences in gene expression levels (not shown). Histograms show the relative expression levels normalized to the loading control Hprt. Error bars indicate the standard deviation of 2 experiments performed in triplicates. p, as compared to GFP-ES infected with LvGFP. LvPtf1a indicates in this figure LvPtf1a-ER treated with Tamox. B) Secretagogue-mediated exocytosis in differentiated cells. Control GFP-ES or RBPL-ES cells infected with LvGFP or LvPtf1a-ER, respectively, were differentiated through the whole protocol and stimulated for 30 minutes with CCK and carbachol. Amylase activity was measured in both the supernatant and cell lysates.