Small molecules induce efficient differentiation into insulin-producing cells from human induced pluripotent stem cells

Stem Cell Research (2012) 8, 274–284

Available online at www.sciencedirect.com

www.elsevier.com/locate/scr

Small molecules induce efficient differentiation into insulin-producing cells from human induced pluripotent stem cells Yuya Kunisada, Noriko Tsubooka-Yamazoe, Masanobu Shoji, Masaki Hosoya ⁎ Biology Research Laboratories, Pharmaceutical Research Division, Takeda Pharmaceutical Company Limited, 26-1, Muraokahigashi 2, Fujisawa, Kanagawa, Japan

Received 8 June 2011; received in revised form 28 September 2011; accepted 1 October 2011 Available online 11 October 2011

Abstract Human induced pluripotent stem (hiPS) cells have potential uses for drug discovery and cell therapy, including generation of pancreatic β-cells for diabetes research and treatment. In this study, we developed a simple protocol for generating insulin-producing cells from hiPS cells. Treatment with activin A and a GSK3β inhibitor enhanced efficient endodermal differentiation, and then combined treatment with retinoic acid, a bone morphogenic protein inhibitor, and a transforming growth factor-β (TGF-β) inhibitor induced efficient differentiation of pancreatic progenitor cells from definitive endoderm. Expression of the pancreatic progenitor markers PDX1 and NGN3 was significantly increased at this step and most cells were positive for anti-PDX1 antibody. Moreover, several compounds, including forskolin, dexamethasone, and a TGF-β inhibitor, were found to induce the differentiation of insulin-producing cells from pancreatic progenitor cells. By combined treatment with these compounds, more than 10% of the cells became insulin positive. The differentiated cells secreted human c-peptide in response to various insulin secretagogues. In addition, all five hiPS cell lines that we examined showed efficient differentiation into insulin-producing cells with this protocol. © 2011 Elsevier B.V. All rights reserved.

Introduction Human induced pluripotent stem (hiPS) cells are generated from human somatic cells by the ectopic expression of ⁎ Corresponding author at: Biology Research Laboratories, Pharmaceutical Research Division, Takeda Pharmaceutical Company Limited, 26-1, Muraokahigashi 2, Fujisawa, Kanagawa, 251-8555, Japan. Fax: + 81 466 29 4416. E-mail addresses: [email protected] (Y. Kunisada), [email protected] (N. Tsubooka-Yamazoe), [email protected] (M. Shoji), [email protected] (M. Hosoya). 1873-5061/$ - see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.scr.2011.10.002

defined transcription factors (Takahashi et al., 2007; Yu et al., 2007; Nakagawa et al., 2008). Since hiPS cells are able to proliferate indefinitely and can differentiate into many cell types, they have great potential for use in drug discovery and cell therapy. One possible application of hiPS cells is the generation of pancreatic β-cells for research and treatment of type 1 diabetes, an autoimmune disease that results from destruction of insulin-producing β-cells in the pancreas. Islet transplantation can be an effective therapy for type 1 diabetes, but the shortage of donors limits its clinical application on a large scale. One possible way to overcome the shortage of pancreatic islets is the generation of insulin-producing cells from human pluripotent stem cells,

Insulin-producing cells from hiPS cells such as hiPS cells or human embryonic stem (hES) cells. Insulin-producing cells derived from hiPS or hES cells could also be useful for identifying novel targets for the development of antidiabetic drugs. There have already been several reports on the generation of insulin-producing cells from hiPS cells or hES cells in vitro (D'Amour et al., 2006; Kroon et al., 2008; Jiang et al., 2007; Tateishi et al., 2008; Zhang et al., 2009; Maehr et al., 2009; Chen et al., 2009; Assady et al., 2001; Phillips et al., 2007). For example, stepwise protocols have been developed to achieve differentiation of hiPS or hES cells into insulin-producing cells by mimicking the process of pancreatic development (D'Amour et al., 2006; Kroon et al., 2008; Jiang et al., 2007; Tateishi et al., 2008; Zhang et al., 2009). A protocol utilizing small molecules to induce pancreatic differentiation has also been described (Chen et al., 2009). In addition, it has been demonstrated that hESderived insulin-producing cells can reduce the serum glucose level in diabetic mice (Kroon et al., 2008). These studies have demonstrated that hiPS cells or hES cells can be differentiated into insulin-producing cells that might function in vivo. During the generation of insulin-producing cells from hiPS or hES cells, the process of deriving PDX1-positive pancreatic progenitors from definitive endoderm, which gives rise to the pancreas, seems to be a critical step. Although several methods of differentiating hES or hiPS cells into definitive endoderm have already been reported (D'Amour et al., 2006; Jiang et al., 2007; Maehr et al., 2009; D'Amour et al., 2005; Borowiak et al., 2009), deriving pancreatic progenitors from definitive endoderm is still a challenge. The process by which PDX1-positive cells undergo differentiation into insulin-producing cells seems to be another key step. However, the mechanisms involved are largely unknown. Several methods for inducing efficient differentiation of definitive endoderm into PDX1-positive cells have recently been reported (Chen et al., 2009; Cai et al., 2010; Ameri et al., 2010; Johannesson et al., 2009), but the culture conditions for induction of these cells and their differentiation into insulinproducing cells remain to be optimized. In this study, we developed a simple stepwise protocol for generating insulin-producing cells from hiPS cells. Combined treatment with small molecules induced the efficient differentiation into pancreatic progenitor cells and insulinproducing cells. With this new protocol, multiple hiPS cell lines generated from different donors showed differentiation into insulin-producing cells at a similar level of efficiency.

Results Differentiation of hiPS cells into definitive endoderm First, we developed a method for generating definitive endoderm, a germ cell layer that gives rise to the pancreas, from hiPS cells. The 253G1 hiPS cell line (Nakagawa et al., 2008), which was generated from human fibroblasts by overexpression of OCT3/4, SOX2, and KLF4, was used. It has been reported that treatment with activin A and Wnt 3a under low serum conditions induces differentiation of human ES cells into endoderm (D'Amour et al., 2006), so we adapted this

275 method for hiPS cells with slight modifications (Fig. 1A). After incubating hiPS cells with activin A and Wnt 3a for 24 h, the cells were subsequently incubated with activin A alone for an additional two days. Expression of the endodermal markers SOX17 and FOXA2 increased significantly (Fig. 1B), and immunostaining showed that 41.4 ± 22.3% of total cells had differentiated into SOX17 and FOXA2 double-positive cells in seven independent experiments (Fig. 1C and Fig. S1). Fewer SOX17 and FOXA2 doublepositive cells appeared in control cultures that were not treated with any factors or cultures treated with activin A alone (3.0 ± 6.1% and 21.9 ± 20.8% of total cells in seven independent experiments, respectively) (Fig. 1C and Fig. S1). Some of the effects of Wnt/β-catenin signaling are known to be mimicked by inhibition of GSK3β activity (Doble and Woodgett, 2003), and we found that the GSK3β-specific inhibitor CHIR99021 could be used to induce the differentiation of endoderm instead of Wnt 3a. Combined treatment with activin A and CHIR99021 induced expression of SOX17 and FOXA2 (Fig. 1B), and most of the cells were doublepositive for SOX17 and FOXA2 antibodies (74.8 ± 6.7% in seven independent experiments) (Fig. 1C and Fig. S1). These findings suggested that combined treatment with activin A and CHIR99021 could induce efficient differentiation of hiPS cells into definitive endoderm.

Pancreatic differentiation from definitive endoderm Several groups have shown that Noggin, an endogenous antagonist of BMP signaling, and retinoic acid induce pancreatic differentiation of definitive endoderm (Kroon et al., 2008; Jiang et al., 2007; Zhang et al., 2009; Cai et al., 2010; Johannesson et al., 2009; Ostrom et al., 2008; Micallef et al., 2005; Stafford et al., 2006). To examine the effect of these factors, we treated hiPS-derived definitive endoderm with Noggin, retinoic acid, and both agents together (Fig. 2A). When cells were cultured with Noggin or retinoic acid alone for seven days, the yield of PDX1-positive cells was relatively low (Fig. 2B). On the other hand, combined treatment with Noggin and retinoic acid significantly increased the yield of PDX1-positive cells to 42.6 ± 9.3% in three independent experiments (Fig. 2B). Dorsomorphin inhibits BMP type I receptors (ALK2, ALK3, and ALK6) and thus blocks BMP signaling (Yu et al., 2008). We found that dorsomorphin could be used for induction of pancreatic differentiation in this culture system instead of Noggin (Fig. 2A). Treatment with dorsomorphin alone had no influence on the yield of PDX1-positive cells, but the combination of dorsomorphin and retinoic acid increased the yield of PDX1-positive cells to the same extent as treatment with Noggin and retinoic acid (50.5 ± 8.2% in five independent experiments) (Fig. 2B). Since dorsomorphin was expected to be more stable than Noggin, we used it instead of Noggin in the following experiments. Next, we examined the expression of other pancreatic markers under these culture conditions. While the expression of PDX1, HNF6, and HLXB9 was increased by treatment with dorsomorphin and retinoic acid, expression of the pancreatic endocrine progenitor marker NGN3 was not detected (Fig. 2C). Since induction of pancreatic endocrine progenitor cells seemed to be inadequate, we screened for small

276

Y. Kunisada et al.

A

D0

D1

D3

RPMI, 2% FBS

RPMI, 2% FBS

± Activin A

± Activin A

± CHIR99021 ± Wnt 3a

40

SOX17

FOXA2

35

60

30

50

25

40

20 30

15

20

Act+CHIR

Act+Wnt

Act+Wnt

Act

0 Act+CHIR

5

0

Ctrl

10

10 Ctrl

Relative expression

70

Act

B

SOX17 Hoechst

Ctrl

Act

Act+Wnt

Act+CHIR

SOX17 FOXA2

C

Ctrl

Act

Act+Wnt

Act+CHIR

Figure 1 Differentiation of hiPS cells into definitive endoderm. (A) Scheme for inducing differentiation into definitive endoderm. (B) hiPS cells were treated with the indicated factors for three days as shown in A. Expression of SOX17 and FOXA2 was analyzed by quantitative RT-PCR on day 3. The Y-axis indicates the relative change of mRNA expression compared with that by control cells (no factors added). The average of eight independent experiments is shown. Error bars indicate s.d. (C) Cells were immunostained with SOX17 and FOXA2 antibodies on day 3. Scale bar = 100 μm. Ctrl, control (no factors added); Act, activin A; Wnt, Wnt 3a; CHIR, CHIR99021.

molecules that induced the NGN3 expression in the presence of dorsomorphin and retinoic acid from among various compounds known to regulate stem cell differentiation (Table S1). As a result, a TGF-β type I receptor inhibitor (SB431542) was found to induce NGN3 expression in combination with dorsomorphin and retinoic acid (Fig. 2C). Treatment with dorsomorphin, retinoic acid, and SB431542 not only increased the expression of NGN3, but also the expression of other pancreatic markers, such as NEUROD1, PDX1, SOX9, and HLXB9 (Fig. 2C). Furthermore, treatment with these three compounds increased the percentage of PDX1positive cells to 71.9 ± 9.7% in five independent experiments (Fig. 2B). In contrast, incubation with SB431542 alone had no effect on the expression of PDX1 and NGN3 mRNA or the expression of PDX1 protein (Figs. 2B, C). Other TGF-β type I receptor inhibitors, such as Alk5 inhibitor II, A-83-01, and TGFβ RI inhibitor VIII, also induced NGN3 expression in

combination with dorsomorphin and retinoic acid (Fig. S2), indicating that differentiation into pancreatic endocrine progenitor cells was enhanced by blockade of TGF-β signaling. Thus, the combination of dorsomorphin, retinoic acid, and SB431542 efficiently induced pancreatic differentiation from definitive endoderm. To investigate whether PDX1-positive cells could be differentiated into insulin-producing cells, we continued to culture PDX1-positive cells after withdrawal of the above three compounds, and found that insulin (INS) and C-peptide double-positive cells emerged after eight days (Fig. 3A and Fig. S3). In this subpopulation of cells, INS expression increased significantly over time (Fig. 3B). In contrast, cells treated with each compound alone or any other combination of compounds showed little differentiation into insulinproducing cells (Figs. 3A, B, Fig. S3, and data not shown). These results supported the concept that the combination

Insulin-producing cells from hiPS cells

A

D0

277

D1

RPMI, 2% FBS Activin A CHIR99021

D3

RPMI, 2% FBS Activin A

D10

Improved MEM, 1% B27 ± Noggin ± Dorsomorphin ± Retinoic acid ± SB431542

PDX1 Hoechst

B

C

Ctrl

Noggin

RA

Noggin+RA

17.0±9.1%

14.6±8.0%

27.9±4.2%

42.6±9.3%

DM

DM+RA

SB

DM+RA+SB

9.5±7.6%

50.5±8.2%

20.8±4.4%

71.9±9.7%

6000

1800 1600 1400 1200 1000 800 600 400 200 0

NGN3

5000 4000 3000

Relative expression

2000 1000 0 D5 600 500

D7

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

HNF6

400 300 200 100 0

D5

D7

D 10

3000

PDX1

2500

Ctrl

2000

SB

1500

DM+RA

1000

DM+RA+SB

500 0 D5

D 10

3500

NEUROD1

D7

D5

D 10 60

HLXB9

50

D7

D 10

SOX9

40

Ctrl

30

SB

20

DM+RA

10

DM+RA+SB

0 D5

D7

D 10

D5

D7

D 10

Figure 2 Differentiation of hiPS cells into pancreatic progenitor cells. (A) Scheme for inducing differentiation into pancreatic progenitor cells. (B) Cells were treated with the indicated factors on day 3 and immunostained with PDX1antibody on day 10. Average percentage of PDX1-positive cells in 3–5 independent experiments is shown. Scale bar = 100 μm. (C) Cells were treated with the indicated factors on day 3. Expression of NGN3, NEUROD1, PDX1, HNF6, HLXB9, and SOX9 was analyzed by quantitative RT-PCR on days 5–10. The Y-axis indicates the relative change of mRNA expression compared to that by cells on day 3. Error bars indicate s.d. The results are representative of three independent experiments. Ctrl, control (no factors added); SB, SB431542; DM, dorsomorphin; RA, retinoic acid.

of dorsomorphin, retinoic acid, and SB431542 achieved efficient commitment to differentiation into pancreatic cells.

Enhancement of pancreatic progenitor cell differentiation into insulin-producing cells Although insulin-producing cells were detected in the cultures treated with dorsomorphin, retinoic acid, and SB431542, the percentage of such cells was very low. This suggested that additional signals were necessary for

efficient induction of insulin-producing cells from PDX1positive cells. To enhance the differentiation of insulinproducing cells, we screened for small molecules that increased INS expression by PDX1-positive pancreatic progenitor cells after day 10 from among various compounds known to regulate stem cells differentiation (Fig. 4A and Table S1). As a result, forskolin, dexamethasone, and Alk5 inhibitor II were found to increase INS expression (Fig. 4B). Immunostaining and FACS analysis showed that insulin-producing cells were increased in cultures treated with forskolin, dexamethasone, or Alk5 inhibitor II (Figs. 4C, D).

278

Y. Kunisada et al.

A SB

DM+RA

DM+RA+SB

Cpep INS Hoe

Ctrl

B Relative expression

3000

INS

2500 Ctrl

2000

SB

1500

DM+RA

1000

DM+RA+SB

500 0 D 10

D 14

D 18

Figure 3 Differentiation into insulin-producing cells. (A) Cells were treated with the indicated factors from day 3 to day 10. After removal of the factors on day 10, the cells were cultured further and immunostained with insulin and C-peptide antibodies on day 18. Scale bar = 100 μm. A typical example of C-peptide and insulin double-positive cells is highlighted. (B) Expression of INS was analyzed by quantitative RT-PCR on days 10–18. The Y-axis indicates the relative change of mRNA expression compared with that by cells on day 3. Error bars indicate s.d. The results are representative of three independent experiments. Ctrl, control (no factors added); SB, SB431542; DM, dorsomorphin; RA, retinoic acid; Cpep, C-peptide; INS, insulin; Hoe, Hoechst.

Forskolin is an adenylate cyclase activator that increases the intracellular cyclic AMP (cAMP) level, suggesting that an increment of intracellular cAMP might enhance differentiation of pancreatic endocrine progenitor cells into insulinproducing cells. To verify this hypothesis, we treated PDX1positive cells with the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX) and the cAMP analog dibutyryl cyclic AMP (dbcAMP). We found that IBMX and dbcAMP increased INS expression and the population of insulin-positive cells, as did forskolin (Fig. S4A, B), suggesting that a high intracellular cAMP level contributed to the differentiation of PDX1-positive cells into insulin-producing cells. On the other hand, Alk5 inhibitor II and other TGF-β type I receptor inhibitors (including A-83-01, SB431542, and TGF-β RI inhibitor VIII) also increased INS expression and the number of insulin-positive cells (Fig. S4A, B), suggesting that differentiation into insulin-producing cells could be enhanced by blocking endogenous TGF-β. To examine the additive effect of forskolin, dexamethasone, and Alk5 inhibitor II on differentiation into insulinproducing cells, we cultured PDX1-positive cells with a combination of these compounds. We also added nicotinamide, which has been reported to induce differentiation of pancreatic β-cells (Sjoholm et al., 1994; Otonkoski et al., 1993). While incubation with nicotinamide alone had no effect on INS expression and the number of insulin-positive cells under these culture conditions, it slightly increased the intensity of staining with insulin antibody for the cells treated with three compounds (forskolin, dexamethasone, and Alk5 inhibitor II) (Figs. 4B–D and data not shown). The level of INS expression and the yield of C-peptide-positive cells among

cells treated with all four agents (forskolin, dexamethasone, Alk5 inhibitor II, and nicotinamide) were similar to those for the cells treated with forskolin alone (Figs. 4B, D). However, these cells were more tightly attached to the plates and grew at a higher density than cells treated with forskolin alone (Fig. 4C). Therefore, we used this combination in the following experiment. FACS analysis showed that 7.8% of the cultured cells were C-peptide-positive glucagon-negative cells and 2.0% of the cells were C-peptide-positive glucagonnegative cells (Fig. 4D). The average percentage of insulinpositive cells obtained in seven independent experiments was 11.8% (range: 8.0−16.9%, data not shown). Immunostaining demonstrated that most of the insulinpositive or C-peptide positive cells were also positive for antibodies to various islet-enriched transcription factors, including PDX1, NEUROD1, ISLET-1, PAX6, and NKX2.2 (Fig. 5A and Fig. S5), a finding that matched the process of pancreatic development (Oliver-Krasinski and Stoffers, 2008). Under this culture condition, other types of pancreatic endocrine cells expressing glucagon (GCG), ghrelin, or somatostatin were also identified (Fig. 5A), but pancreatic polypeptide-positive cells were not detected (data not shown). Some of the glucagon- or somatostatin-positive cells were also stained by insulin antibody, whereas most of the ghrelin-positive cells were negative for it.

Pharmacological stimulation of insulin secretion To examine the functionality of hiPS-derived insulin-positive cells, we measured the C-peptide level in the culture

Insulin-producing cells from hiPS cells

A

D3

D1

D0

RPMI, 2% FBS Activin A CHIR99021

B 0.016

RPMI, 2% FBS Activin A

INS

0.6

0.014

D 10

D 21

Improved MEM, 1% B27 Dorsomorphin Retinoic acid SB431542

Improved MEM, 1% B27 ± Forskolin ± Dexamethasone ± Alk5 inhibitor II ± Nicotinamide

D

GCG

0.5

0.012

Control

0.4

0.01 0.008

0.0

Fsk 0.6

0.0

1.0

0.3

0.006

0.2

0.004 0.1

Ctrl

F+D+A+N

Alki

Alki

Nico

97.4 Dex

F+D+A+N

Alki

Nico

Fsk

0

1.9

Alki 0.5

0.0

2.8

0.8

97.4

1.8

Fsk/Dex/Alki/Nico

Dex 0.1

Insulin Hoechst

6.0

Nico 2.7

93.9

Fsk

93.0

Glucagon

C

Dex

Ctrl

0

Ctrl

0.002

Fsk

Relative expression

279

1.2

0.3

2.0

Nico

94.1

4.5

89.8

7.8

C-peptide

Dex

Fsk/Dex/Alki/Nico

Figure 4 Small molecules inducing differentiation into insulin-producing cells. (A) Scheme for inducing differentiation into insulinproducing cells with small molecules. (B) Cells were treated with the indicated factor on day 10. Expression of INS and GCG was analyzed by quantitative RT-PCR on days 22. Cells cultured in basal medium containing 1% B27 were used as a control (Ctrl). Y-axis is shown as relative expression to that in human islet. Error bars indicate s.d. The results are representative of three independent experiments. (C) Immunostaining with insulin antibody on day 23. Scale bar= 100 μm. (D) FACS analysis of cells treated with the indicated compounds on day 22. Cells were stained with C-peptide and glucagon antibodies. Ctrl, control (no factors added); Fsk, forskolin; Dex, dexamethasone; Alki, Alk5 inhibitor II; Nico, nicotinamide; F + D + A+ N, forskolin + dexamethasone + Alk5 inhibitor II + nicotinamide.

medium after incubation with glucose or insulin secretagogues. Insulin-producing cells that had been induced by treatment with the four compounds (forskolin, dexamethasone, Alk5 inhibitor II, and nicotinamide) were utilized for these experiments. While incubation with glucose did not alter insulin secretion, treatment with tolbutamide (a KATP channel blocker), (−)-Bay K8644 (a L-type Ca 2 + channel

agonist), carbachol (a muscarinic agonist), IBMX (a phosphodiesterase inhibitor), and potassium chloride (a depolarizing agent) increased c-peptide release (Fig. 5B). These results suggested that the cells were able to secrete c-peptide in response to various stimuli, but had not differentiated into fully mature β-cells with the ability to secrete insulin in response to glucose stimulation.

280

Y. Kunisada et al.

A

PDX1 Cpep Hoe

NEUROD Cpep Hoe

ISL1 Cpep Hoe

PAX6 INS Hoe

NKX2.2 INS Hoe

GCG Cpep Hoe

GHRL Cpep Hoe

SST Cpep Hoe

B

C *

1400

1503-iPS

1392-iPS

NHDF-iPS

201B7

1200 1000

*

800

*

600

*

400

*

200 KCl

IBMX

Carbachol

Tolbutamide

(-)-Bay K8644

Glucose 22.5 mM

Glucose 2.5 mM

0

Insulin Hoechst

c-peptide concentration (pM)

1600

+ Glucose 2.5 mM

Figure 5 Characterization of insulin-producing cells derived from hiPS cells. (A) Cells were treated with the four compounds (forskolin, Alk5 inhibitor II, dexamethasone, and nicotinamide) on day 10. Then cells were immunostained with antibodies against C-peptide, insulin, PDX1, NEUROD1, ISLET-1, PAX6, NKX2.2, glucagon, ghrelin, and somatostatin on day 21. (B) C-peptide secretion by differentiated cells on day 22 in response to the indicated concentrations of D-glucose, 2 μM (−)-Bay K8644, 100 μM tolbutamide, 250 μM carbachol, 0.5 mM IBMX, or 30 mM potassium chloride (KCl). Error bars indicate s.d. *P b 0.01 by Student's t-test. The results are representative of three independent experiments. (C) Differentiated cells obtained from the 1503-iPS, 1392-iPS, NHDF-iPS, and 201B7 hiPS cell lines were immunostained with insulin antibody on day 21. Cpep, C-peptide; Hoe, Hoechst; NEUROD, NEUROD1; ISL1, ISLET-1;INS, insulin; GCG, glucagon; CHRL, ghrelin; SST, somatostatin. Scale bar = 100 μm.

Differentiation of multiple hiPS cell lines into insulin-producing cells We used the 253G1 hiPS cell line to develop the present protocol. To examine whether this protocol was also effective for promoting the differentiation of other hiPS cell lines, we tested four more hiPS cell lines, which were 1503-iPS (297A1), 1392-iPS (297F1), NHDF-iPS (297L1), and 201B7. While 201B7 was generated from dermal fibroblasts of the same donor as 253G1 by overexpression of OCT4, KLF4, SOX2, and C-MYC (Takahashi et al., 2007), the other cell lines (1503-iPS (297A1), 1392-iPS (297F1), and NHDFiPS (297L1)) were generated from the fibroblasts of

different donors by overexpression of the same genes (Takahashi et al., 2009). When the cells were treated with activin A and CHIR99021 as described above, a high yield of SOX17and FOXA2 double-positive cells was obtained from all four of the hiPS cell lines tested (Fig. S6). After treatment with the various compounds described above, these cells underwent differentiation into PDX1-positive cells (Fig. S7) and then finally into insulin-positive cells at a similar level of efficiency (Fig. 5C). FACS analysis showed that insulinpositive cells accounted for 15.8% (range: 14.7–17.7%), 11.6% (range: 8.3–14.0%), 12.3% (range: 3.9–17.7%), and 9.2% (range: 7.1–10.6%) of the total cells in cultures of 1503-iPS, (297A1), 1392-iPS (297F1), NHDF-iPS (297L1),

Insulin-producing cells from hiPS cells and 201B7, respectively, in four independent experiments. These findings suggested that our method can be applied to induce the efficient differentiation of pancreatic cells from various hiPS cell lines.

Discussion In this study, we developed a simple protocol for inducing the differentiation of human iPS cells into insulin-producing cells. With this method, most iPS cells underwent differentiation into SOX17and FOXA2 double-positive definitive endoderm, and then into PDX1-positive pancreatic progenitor cells. Finally, after treatment with a combination of low molecular weight compounds, differentiation into insulinproducing cells was achieved at a yield of more than 10%. Our protocol does not include complicated processes, such as embryoid body formation, cell sorting by FACS, and reseeding into other plates. While we used activin A and fetal bovine serum (FBS) during early differentiation, each step of differentiation was mainly induced by small molecules, making the differentiation process efficient and stable. As the first step of our method, hiPS cells underwent differentiation into definitive endoderm by treatment with activin A and the GSK3β inhibitor CHIR99021. In other methods, activin A and Wnt 3a have mainly been used to induce definitive endoderm (D'Amour et al., 2006; Kroon et al., 2008; Maehr et al., 2009; Chen et al., 2009; Ameri et al.; Johannesson et al., 2009). However, we found that treatment with activin A plus CHIR99021 induced SOX17 and FOXA2 double-positive definitive endoderm more efficiently compared with activin A plus Wnt 3a. We confirmed that CHIR99021 was a highly selective inhibitor of GSK3β and GSK3α by the kinase profiling assay (KinaseAdvisor™, Caliper LifeSciences) (data not shown), suggesting that its influence on other signaling processes would be very small. In consideration of efficiency, stability, and cost, using CHIR99021 instead of Wnt 3a is preferable for inducing differentiation into definitive endoderm. Next, we induced differentiation of definitive endoderm into pancreatic progenitor cells. A large number of factors, including Cyclopamine-KAAD, (−)-indolactam V, Noggin, retinoic acid, FGF2, FGF4, FGF10, and EGF, have been reported to induce the differentiation of pancreatic progenitor cells (D'Amour et al., 2006; Kroon et al., 2008; Jiang et al., 2007; Chen et al., 2009; Cai et al.; Ameri et al.; Johannesson et al., 2009). Among these factors, we found combined treatment with Noggin and retinoic acid was effective for inducing PDX1-positive cells in our culture condition. We also showed that treatment with the BMP receptor inhibitor dorsomorphin and retinoic acid could induce the differentiation of PDX1-positive cells. This finding is consistent with several recent reports that Noggin can be replaced by dorsomorphin in the induction of ES cell differentiation (Hao et al., 2008; Wada et al., 2009; Nostro et al., 2011). Because of its superior stability and lower cost, dorsomorphin seemed to be a better choice than Noggin for induction of PDX1-positive cells. Furthermore, we found that inhibitors of the TGF-β type I receptor were effective for inducing pancreatic differentiation in combination with retinoic acid and dorsomorphin. This is consistent with a recent report that PDX1 expression was induced by treatment

281 of mouse embryos with SB431542 at the 5–6 somite stage (Wandzioch and Zaret, 2009). Moreover, while we were preparing this manuscript, Rezania et al. and Nostro et al. reported that treatment with TGF-β type I receptor inhibitors following pancreatic specification induced the differentiation into pancreatic endocrine cells from human ES cells (Nostro et al., 2011; Rezania et al.). Our results were also consistent with their reports. Thus, TGF-β signaling might have a negative influence on differentiation into pancreatic progenitor cells. As the next step, we investigated conditions that allowed the efficient generation of insulin-producing cells from PDX1-positive cells. Various compounds, including forskolin, dbcAMP, IBMX, TGF-β type I receptor inhibitors, and dexamethasone, have been found to increase the number of insulin-producing cells. Since forskolin and IBMX are known to increase the intracellular cAMP level and dbcAMP is a cell-permeable cAMP analog, it is suggested that the cellular cAMP level is one of the key factors regulating the differentiation of insulin-producing cells. However, we found that long-term exposure (N 20 days) to forskolin alone decreased the number of insulin-producing cells (data not shown). Thus, a high level of cAMP might be needed in priming for differentiation, but excess cAMP might have a negative effect on the survival of insulin-producing cells. Interestingly, when cells were treated with dexamethasone in addition to forskolin, the number of insulin-producing cells remained stable for a long period. The precise mechanism remains to be analyzed, but this finding indicates that glucocorticoid signaling promotes the survival of insulin-producing cells. TGF-β type I receptor inhibitors were also found to induce differentiation of insulin-producing cells. Among the TGF-β type I receptor inhibitors that we tested, Alk5 inhibitor II showed unique effects on the differentiation of pancreatic endocrine cells. Only Alk5 inhibitor II significantly increased the expression of both INS and GCG, unlike the other TGF-β type I receptor inhibitors (Fig. 4B and data not shown). In the kinase profiling assay, Alk5 inhibitor II inhibited various kinases at the concentration used in this study (data not shown), suggesting that it might regulate the differentiation of pancreatic endocrine cells by inhibiting various other signals in addition to those of TGF-β. The mechanisms involved in differentiation of pancreatic progenitor cells into each type of hormone-producing cell are largely unknown, so elucidating the mechanism by which Alk5 inhibitor II acts would be helpful for understanding the process of pancreatic endocrine cell differentiation. Rezania et al. reported that insulin and glucagon double-positive cells were induced by treatment with Alk5 inhibitor II and that most of these cells finally differentiated into glucagon-producing alpha cells (Rezania et al., 2011). When pancreatic progenitor cells were treated with Alk5 inhibitor II alone in our culture system, the percentage of both glucagon-positive cells and glucagon and insulin double-positive cells was relatively high (Fig. 4D), a finding that was consistent with their results. However, when the cells were treated with four compounds (forskolin, dexamethasone, Alk5 inhibitor II, and nicotinamide), induction of these cells was much lower and most of the insulin-positive cells that we obtained were also positive for PDX1(Figs. 4D, 5A, and Fig. S5), which is not expressed in glucagon-producing cells in the human fetal pancreas (Lyttle et al., 2008). Moreover, the number of insulin-positive (glucagon-negative)

282 cells was much larger than that of glucagon-positive cells, even after additional four weeks of culture (data not shown). These results indicate that cells were more likely to undergo differentiation into pancreatic β-cells by treatment with the four compounds (forskolin, dexamethasone, Alk5 inhibitor II, and nicotinamide) in our culture system. It is noteworthy that multiple hiPS cell lines showed differentiation into insulin-producing cells by our protocol. The hiPS cell lines that we utilized in this study were generated from the fibroblasts of various individuals, including a neonatal male and a 73-year-old woman (Takahashi et al., 2007; Nakagawa et al., 2008; Takahashi et al., 2009). Thus, our results suggest that this is a robust protocol and that optimization for individual cell lines might not be required. In another report, the rate of differentiation into insulinproducing cells varied among different hES cell lines (D'Amour et al., 2006). Since stable differentiation from various hiPS cell lines is necessary for clinical application, our protocol could be beneficial for development of cell therapy in the future. Although this method was effective for inducing insulinproducing cells, there are still several issues to be addressed. First, the insulin-producing cells generated by our protocol showed little insulin secretion in response to glucose stimulation, and some of the cells co-expressed glucagon or somatostatin, suggesting that the cells we obtained were not fully functional mature β-cells. To induce the maturation, further optimization of the culture condition seems to be required. Second, the current protocol utilizes mouse embryonic fibroblasts and FBS, but animal-derived substances are undesirable for clinical use. Third, this protocol has not been applied to hES cells. Since hES cells are recognized as the gold standard among pluripotent stem cells and were utilized in many other studies, experiments on hES cells would be useful for comparing our protocol with other protocols and for improving our understanding the mechanism of human pancreatic development. In conclusion, we developed a simple method for promoting the differentiation of hiPS cells into insulin-producing cells. With this method, all of the hiPS cell lines that we tested showed efficient differentiation into insulinproducing cells, suggesting that our protocol achieves efficient and reproducible generation of insulin-producing cells in vitro. This protocol has the potential to contribute to drug discovery and cell therapy for diabetes.

Materials and methods Cell culture Human iPS cell lines (253G1, 201B7, 1503-iPS (297A1), 1392iPS (297F1), and NHDF-iPS (297L1)) were kindly provided by the Center for iPS Cell Research and Application of Kyoto University (Kyoto, Japan). The cells were maintained on mitomycin-C treated-mouse embryonic fibroblasts (MEFs, Oriental Yeast, Tokyo, Japan) in primate ES medium (ReproCell, Tokyo, Japan) containing 4 ng/ml of human fibroblast growth factor 2 (FGF2, ReproCell) and penicillin/streptomycin (Invitrogen, Carlsbad, CA). Cells were passaged at 1:3–1:4 by using dissociation solution (ReproCell). Culture was carried out at 37 °C under 5% CO2 in air.

Y. Kunisada et al. For pancreatic differentiation, hiPS cells were dissociated with trypsin-EDTA, plated on MEFs at a density of 6.3 × 10 4 cells/cm 2, and cultured for 24 h in primate ES medium containing 10 μM Y-276342 (Wako, Osaka, Japan), 4 ng/ml FGF2, and penicillin/streptomycin. Then the cells were cultured in primate ES medium containing 4 ng/ml FGF2 and penicillin/streptomycin for an additional 48 h. Next, the cells were cultured in RPMI 1640 medium (Invitrogen) containing 2% fetal bovine serum (FBS, Invitrogen) with or without 100 ng/ml human activin A (PeproTech, Rocky Hill, NJ), 25 ng/ml mouse Wnt 3a (R&D, Minneapolis, MN), or 3 μM CHIR99021 (Axon Medchem, Groningen, The Netherlands) for 24 h, and then in RPMI 1640 medium containing 2% FBS with or without 100 ng/ml activin A for 48 h. Subsequently, the medium was replaced with Improved MEM Zinc Option medium (Invitrogen) containing 1% B27 (Invitrogen) with or without 100 ng/ml human Noggin (R&D), 1 μM dorsomorphin (Calbiochem, San Diego, CA), 2 μM retinoic acid (Sigma, St. Louis, MO), or 10 μM SB431542 (Sigma). Cells were cultured for seven days and then these media were replaced with Improved MEM Zinc Option medium supplemented with 1% B27. To enhance differentiation into insulin-producing cells, 10 μM forskolin (Wako), 10 μM dexamethasone (Wako), 5 μM Alk5 inhibitor II (Calbiochem), or 10 mM nicotinamide (StemCell Technologies, Vancouver, BC) was added to the medium, and two thirds of the medium was replaced on every third or fourth day of culture. The use of hiPS cells was approved by ethical review committee of Takeda Pharmaceutical Company Limited.

Reverse transcriptase polymerase chain reaction (RT-PCR)

Total RNA was isolated from cells with an RNeasy 96 kit (Qiagen, Valencia, CA), and cDNA was prepared with a PrimeScript RT reagent kit (Takara Bio, Shiga, Japan) according to the manufacturer's instructions. Quantitative RT-PCR analysis was performed with the TaqMan system on a 7900HT Fast Real-Time PCR analyzer (Applied Biosystems, Foster City, CA). Primers and fluorogenic probes for RT-PCR are listed in Supplementary Table 2. The reaction mixture consisted of 1 × Platinum Quantitative PCR SuperMix-UDG with ROX (Invitrogen), 125 nM of the probe, 450 nM of the forward and reverse primers, and template cDNA. To quantify PDX1 mRNA expression, a 0.5× probe and primer set (HS00236830, Applied Biosystems) was utilized. Expression of each target gene was normalized for the amount of total RNA or the level of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression. Human islets were obtained from National Disease Research Interchange (NDRI, PA, USA) through Human and Animal Bridging (HAB) Research Organization (Chiba, Japan) following written consent. The use of human islets was approved by ethical review committee of Takeda Pharmaceutical Company Limited.

Immunostaining Cells were fixed with 2% paraformaldehyde (PFA) for 10 min and then with 4% PFA for 20 min at room temperature. After washing with PBS, the cells were permeabilized and blocked in PBS containing 0.1% Triton X-100 and 5% normal donkey serum for 30 min at room temperature. Then the cells

Insulin-producing cells from hiPS cells were incubated overnight at 4 °C with the following primary antibodies: goat anti-SOX17, 1:500 (AF1924, R&D); rabbit anti-FOXA2, 1:100 (07-633, Millipore, Billerica, MA); goat anti-PDX1, 1:200 or 1:400 (AF2419, R&D); guinea pig antiinsulin, 1:500 (A0564, Dako); mouse anti-C-peptide, 1:1000 (MON5021, Sanbio, Uden, The Netherlands); goat antiglucagon, 1:400 (SC-7780, Santa Cruz Biotechnology, Santa Cruz, CA); goat anti-ghrelin, 1:400 (SC-10368, Santa Cruz Biotechnology); rabbit anti-somatostatin, 1:400 (A0566, Dako); rabbit anti-PP, 1:50 (180043, Invitrogen); mouse anti-PAX6, 1:100 (PAX6, Developmental Studies Hybridoma Bank, Iowa City, IA); mouse anti-NKX2.2, 1:100 (74.5A5, Developmental Studies Hybridoma Bank); goat anti-ISLET1, 1:100 (AF1837, R&D); and goat anti-NEUROD1, 1:100 (AF2746, R&D). After washing with PBS, the cells were incubated for one hour at room temperature with Alexa Fluor 488- or 568-conjugated donkey or goat antibodies directed against mouse, rabbit, goat, or guinea pig IgG at 1:500 dilutions. For quantification, 15 images per well were obtained using an IN Cell Analyzer 1000 (GE Healthcare, Buckinghamshire, UK) and quantitative analysis was done with Developer software (GE Healthcare).

Flow cytometry Differentiated cells were dissociated with trypsin/EDTA, fixed with 2% paraformaldehyde, and permeabilized with BD Perm/Wash buffer (Becton Dickinson, Mountain View, CA). The cells were incubated with mouse anti-C-peptide antibody (1:1000, MON5021, Sanbio) and goat anti-glucagon antibody (1:400, SC-7780, Santa Cruz Biotechnology) for 30 min at room temperature and then stained with Alexa Fluor 488-conjugated donkey antibodies directed against mouse IgG (1:500, Invitrogen) and Alexa Fluor 647conjugated donkey antibodies directed against goat IgG (1:500, Invitrogen) for 30 min at room temperature. Flow cytometry was done with a FACScalibur (Becton Dickinson) or a FACSAria II (Becton Dickinson).

C-peptide release assay Differentiated cells were preincubated for 2 h at 37 °C in Krebs-Ringer bicarbonate HEPES buffer (KRBH; 116 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl2 1.2 mM KH2PO4, 1.2 mM MgSO4, 24 mM HEPES, 25 mM NaHCO3, and 0.1% BSA) containing 2.5 mM glucose. Then, the cells were incubated for 1 h at 37 °C in KRBH containing the indicated concentrations of glucose or stimulants, including 2 μM (−)-Bay K8644, 100 μM tolbutamide, 250 μM carbachol, 0.5 mM IBMX, or 30 mM potassium chloride (KCl). C-peptide was measured with a human C-peptide ELISA kit (Mercodia AB, Uppsala, Sweden) according to the manufacturer's instructions. Supplementary materials related to this article can be found online at doi:10.1016/j.scr.2011.10.002.

Acknowledgments We thank Dr. Nobuhiro Suzuki, Atushi Nakanishi, and Nobuyuki Miyajima for helpful suggestions and comments, and we thank our many colleagues in Biology Research

283 Laboratories for technical advice. We also thank Dr. Shinya Yamanaka (center for iPS Cell research and Application of Kyoto University) for providing hiPS cell lines, National Disease Research Interchange (NDRI, PA, USA) and Human and Animal Bridging (HAB) Research Organization (Chiba, Japan) for providing human islets, and Developmental Studies Hybridoma Bank (developed under the auspices of the NICHD and maintained by the University of Iowa, Department of Biology) for providing antibodies.

References Ameri, J., Stahlberg, A., Pedersen, J., Johansson, J.K., Johannesson, M., Artner, I., Semb, H., 2010. FGF2 specifies hESC-derived definitive endoderm into foregut/midgut cell lineages in a concentration-dependent manner. Stem Cells 28, 45–56. Assady, S., Maor, G., Amit, M., Itskovitz-Eldor, J., Skorecki, K.L., Tzukerman, M., 2001. Insulin production by human embryonic stem cells. Diabetes 50, 1691–1697. Borowiak, M., Maehr, R., Chen, S., Chen, A.E., Tang, W., Fox, J.L., Schreiber, S.L., Melton, D.A., 2009. Small molecules efficiently direct endodermal differentiation of mouse and human embryonic stem cells. Cell Stem Cell 4, 348–358. Cai, J., Yu, C., Liu, Y., Chen, S., Guo, Y., Yong, J., Lu, W., Ding, M., Deng, H., 2010. Generation of homogeneous PDX1(+) pancreatic progenitors from human ES cell-derived endoderm cells. J, Mol Cell Biol 2, 50–60. Chen, S., Borowiak, M., Fox, J.L., Maehr, R., Osafune, K., Davidow, L., Lam, K., Peng, L.F., Schreiber, S.L., Rubin, L.L., Melton, D., 2009. A small molecule that directs differentiation of human ESCs into the pancreatic lineage. Nat. Chem. Biol. 5, 258–265. D'Amour, K.A., Agulnick, A.D., Eliazer, S., Kelly, O.G., Kroon, E., Baetge, E.E., 2005. Efficient differentiation of human embryonic stem cells to definitive endoderm. Nat. Biotechnol. 23, 1534–1541. D'Amour, K.A., Bang, A.G., Eliazer, S., Kelly, O.G., Agulnick, A.D., Smart, N.G., Moorman, M.A., Kroon, E., Carpenter, M.K., Baetge, E.E., 2006. Production of pancreatic hormone-expressing endocrine cells from human embryonic stem cells. Nat. Biotechnol. 24, 1392–1401. Doble, B.W., Woodgett, J.R., 2003. GSK-3: tricks of the trade for a multi-tasking kinase. J. Cell Sci. 116, 1175–1186. Hao, J., Daleo, M.A., Murphy, C.K., Yu, P.B., Ho, J.N., Hu, J., Peterson, R.T., Hatzopoulos, A.K., Hong, C.C., 2008. Dorsomorphin, a selective small molecule inhibitor of BMP signaling, promotes cardiomyogenesis in embryonic stem cells. PLoS One 3, e2904. Jiang, J., Au, M., Lu, K., Eshpeter, A., Korbutt, G., Fisk, G., Majumdar, A.S., 2007. Generation of insulin-producing isletlike clusters from human embryonic stem cells. Stem Cells 25, 1940–1953. Johannesson, M., Stahlberg, A., Ameri, J., Sand, F.W., Norrman, K., Semb, H., 2009. FGF4 and retinoic acid direct differentiation of hESCs into PDX1-expressing foregut endoderm in a time- and concentration-dependent manner. PLoS One 4, e4794. Kroon, E., Martinson, L.A., Kadoya, K., Bang, A.G., Kelly, O.G., Eliazer, S., Young, H., Richardson, M., Smart, N.G., Cunningham, J., Agulnick, A.D., D'Amour, K.A., Carpenter, M.K., Baetge, E.E., 2008. Pancreatic endoderm derived from human embryonic stem cells generates glucose-responsive insulin-secreting cells in vivo. Nat. Biotechnol. 26, 443–452. Lyttle, B.M., Li, J., Krishnamurthy, M., Fellows, F., Wheeler, M.B., Goodyer, C.G., Wang, R., 2008. Transcription factor expression in the developing human fetal endocrine pancreas. Diabetologia 51, 1169–1180. Maehr, R., Chen, S., Snitow, M., Ludwig, T., Yagasaki, L., Goland, R., Leibel, R.L., Melton, D.A., 2009. Generation of pluripotent

284 stem cells from patients with type 1 diabetes. Proc. Natl. Acad. Sci. U. S. A. 106, 15768–15773. Micallef, S.J., Janes, M.E., Knezevic, K., Davis, R.P., Elefanty, A.G., Stanley, E.G., 2005. Retinoic acid induces Pdx1-positive endoderm in differentiating mouse embryonic stem cells. Diabetes 54, 301–305. Nakagawa, M., Koyanagi, M., Tanabe, K., Takahashi, K., Ichisaka, T., Aoi, T., Okita, K., Mochiduki, Y., Takizawa, N., Yamanaka, S., 2008. Generation of induced pluripotent stem cells without Myc from mouse and human fibroblasts. Nat. Biotechnol. 26, 101–106. Nostro, M.C., Sarangi, F., Ogawa, S., Holtzinger, A., Corneo, B., Li, X., Micallef, S.J., Park, I.H., Basford, C., Wheeler, M.B., Daley, G.Q., Elefanty, A.G., Stanley, E.G., Keller, G., 2011. Stagespecific signaling through TGFbeta family members and WNT regulates patterning and pancreatic specification of human pluripotent stem cells. Development 138, 861–871. Oliver-Krasinski, J.M., Stoffers, D.A., 2008. On the origin of the beta cell. Genes Dev. 22, 1998–2021. Ostrom, M., Loffler, K.A., Edfalk, S., Selander, L., Dahl, U., Ricordi, C., Jeon, J., Correa-Medina, M., Diez, J., Edlund, H., 2008. Retinoic acid promotes the generation of pancreatic endocrine progenitor cells and their further differentiation into beta-cells. PLoS One 3, e2841. Otonkoski, T., Beattie, G.M., Mally, M.I., Ricordi, C., Hayek, A., 1993. Nicotinamide is a potent inducer of endocrine differentiation in cultured human fetal pancreatic cells. J. Clin. Invest. 92, 1459–1466. Phillips, B.W., Hentze, H., Rust, W.L., Chen, Q.P., Chipperfield, H., Tan, E.K., Abraham, S., Sadasivam, A., Soong, P.L., Wang, S.T., Lim, R., Sun, W., Colman, A., Dunn, N.R., 2007. Directed differentiation of human embryonic stem cells into the pancreatic endocrine lineage. Stem Cells Dev. 16, 561–578. Rezania, A., Riedel, M.J., Wideman, R.D., Karanu, F., Ao, Z., Warnock, G.L., Kieffer, T.J., 2011. Production of Functional GlucagonSecreting Alpha Cells from Human Embryonic Stem Cells. Diabetes 60, 239–247.

Y. Kunisada et al. Sjoholm, A., Korsgren, O., Andersson, A., 1994. Polyamine requirements in nicotinamide-stimulated beta-cell differentiation in fetal porcine islet-like cell clusters. Endocrinology 135, 1559–1565. Stafford, D., White, R.J., Kinkel, M.D., Linville, A., Schilling, T.F., Prince, V.E., 2006. Retinoids signal directly to zebrafish endoderm to specify insulin-expressing beta-cells. Development 133, 949–956. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., Yamanaka, S., 2007. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131, 861–872. Takahashi, K., Narita, M., Yokura, M., Ichisaka, T., Yamanaka, S., 2009. Human induced pluripotent stem cells on autologous feeders. PLoS One 4, e8067. Tateishi, K., He, J., Taranova, O., Liang, G., D'Alessio, A.C., Zhang, Y., 2008. Generation of insulin-secreting islet-like clusters from human skin fibroblasts. J. Biol. Chem. 283, 31601–31607. Wada, T., Honda, M., Minami, I., Tooi, N., Amagai, Y., Nakatsuji, N., Aiba, K., 2009. Highly efficient differentiation and enrichment of spinal motor neurons derived from human and monkey embryonic stem cells. PLoS One 4, e6722. Wandzioch, E., Zaret, K.S., 2009. Dynamic signaling network for the specification of embryonic pancreas and liver progenitors. Science 324, 1707–1710. Yu, J., Vodyanik, M.A., Smuga-Otto, K., Antosiewicz-Bourget, J., Frane, J.L., Tian, S., Nie, J., Jonsdottir, G.A., Ruotti, V., Stewart, R., Slukvin, I.I., Thomson, J.A., 2007. Induced pluripotent stem cell lines derived from human somatic cells. Science 318, 1917–1920. Yu, P.B., Hong, C.C., Sachidanandan, C., Babitt, J.L., Deng, D.Y., Hoyng, S.A., Lin, H.Y., Bloch, K.D., Peterson, R.T., 2008. Dorsomorphin inhibits BMP signals required for embryogenesis and iron metabolism. Nat. Chem. Biol. 4, 33–41. Zhang, D., Jiang, W., Liu, M., Sui, X., Yin, X., Chen, S., Shi, Y., Deng, H., 2009. Highly efficient differentiation of human ES cells and iPS cells into mature pancreatic insulin-producing cells. Cell Res. 19, 429–438.