Continuous Inhibition of Sonic Hedgehog Signaling Leads to Differentiation of Human-Induced Pluripotent Stem Cells into Functional Insulin-Producing β Cells

Asan Institute for Life Science, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Republic of Korea Department of Medicine, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Republic of Korea Asan Medical Institute of Convergence Science and Technology (AMIST), Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Republic of Korea Division of Hepato-Biliary and Pancreatic Surgery, Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Republic of Korea


Introduction
Insulin-producing cells (IPCs) derived from human embryonic stem cells (ESCs) or human-induced pluripotent stem cell (iPSCs) can be used not only for transplanting islet cells, which are destroyed by autoimmunity in patients with type 1 diabetes mellitus, but also for identifying novel targets for the development of antidiabetic drugs in vitro. Several studies have differentiated IPCs from ESCs or iPSCs [1][2][3][4][5] by mimicking pancreatic development, and a stepwise protocol has been established [3,4,[6][7][8]. This protocol requires adding various factors to culture media at each stage of pancreatic differentiation, to induce transcription factors that are specific to each stage. Small molecules are popular inducers of differentiation, and the function of IPCs generated in this manner is robust in vivo and in vitro [9][10][11][12].
Although ESC-or iPSC-derived IPCs with functions similar to those of natural islet cells have been generated, they do not secrete insulin in response to glucose [5,20,24]. This is because differentiation and culture conditions in vitro do not accurately mimic the natural stages of human pancreatic development.
Therefore, although combining differentiation-inducing factors is important, the method of producing these cells in culture must also be improved.
We developed a simple and efficient stepwise protocol for generating IPCs from iPSCs using SANT-1 and FR180204. We induced the differentiation of iPSCs into a definitive endoderm, pancreatic progenitor cells, and finally, IPCs. In addition, the functional IPCs were continuously cultured and matured to attain a spheroid morphology similar to that of natural islet cells.

Small
Molecules. Several small molecular compounds that inhibit cellular signaling pathways have been applied to differentiate iPSCs into IPCs. We divided the differentiation process into three steps to mimic the process of embryonic pancreatic development, namely, the formation of a defini-tive endoderm, its differentiation into pancreatic endoderm and pancreatic progenitor cells, and the specific induction of cell differentiation into IPCs. The small molecules were processed at each step, and the results were confirmed by comparing and analyzing the effects of combining these agents (Table 1).

2.
3. Teratoma Analysis. The iPSCs were harvested and dissociated into single-cell suspensions using TrypLE™ Express Table 1: Function and application of small molecules in the differentiation of cells into insulin-producing cells.
Step Name Function Application   3 Stem Cells International (Gibco), and then 2 × 10 6 cells were subcutaneously injected dorsally into 8-week-old mice with severe combined immunodeficiency. The mice were maintained under non-specific pathogen-free (SFP) conditions at an experimental animal facility in Asan Institute, and euthanized after the tumors reached >1 cm 3 , or after an observation period of 40 d. Tissues containing tumors were fixed with 4% paraformaldehyde (PFA), embedded in paraffin, and cut into serial 4 μm sections. The sections were then stained with hematoxylin and eosin and histologically analyzed by microscopy.

Flow
Cytometry. Differentiated cells were dissociated into single cells using Accutase, fixed with 4% PFA, and permeabilized with Perm buffer III (BD Biosciences, San Jose, CA, USA). The cells were then incubated with normal horse serum for 10 min, followed by mouse anti-insulin, rabbit anti-glucagon, and rabbit anti-Ngn3 antibody for 30 min at RT. The cells were stained with Alexa Fluor 488-conjugated donkey antibodies directed against mouse or rabbit IgG for 30 min at RT, and then fluorescence emission was measured by flow cytometry using a FACSAria II (Becton Dickinson).

Statistical
Analyses. Standard deviation was determined using two-tailed unpaired Student's t-tests. Multiple comparisons were assessed, and significance was calculated using two-way ANOVA including P values. All data were statistically analyzed using Prism version 8 (GraphPad Software Inc., San Diego, CA, USA). Values with P < 0:05 were considered statistically significant.

Characterization of iPSCs.
We cultured iPSCs in vitronectin-coated culture dishes without feeder cells. We found that colonies of iPSCs could be maintained stably in vitro for up to 2 weeks (Figure 1(a)). Specific nuclear markers of iPSCs, namely, the homeodomain transcription factor of the POU family, Oct4 (Figure 1(b)), Nanog (Figure 1(c)), and Sox2 (Figure 1(d)), were identified by the immunostaining of undifferentiated cells. We also detected SSEA4, a representative cell surface marker of iPSCs (Figure 1(e)). We examined the pluripotency of the iPSCs using a teratoma analysis. The iPSCs that were transplanted into the mice grew into a solid mass of tissue for 40 d, and then clusters of various tissues appeared on the tumor surface (Figure 1(f)). Sections of the solid masses histologically varied, and representative endoderm, ectoderm, and mesoderm tissues were identified. These characteristics confirmed that the iPSCs were pluripotent and had the potential to differentiate into IPCs.

Induction of iPSC Differentiation into a Definitive
Endoderm Using Small Molecules. We investigated small molecules that could differentiate iPSCs into a definitive endoderm (Figure 2(a)). We compared the differentiation efficiency of CHIR99021 and LY294002 based on activin A concentrations (Figure 2(b)) on day 1 of a total induction period of 3 d. Immunostaining revealed that cells expressed more abundant FoxA2 (Figure 2(c)), and more cells expressed CXCR4 when incubated with CHIR99021 and LY294002 together than separately (Figure 2(d)). The findings of qPCR revealed slightly less FoxA2 expression in cells incubated with both compounds than with CHIR99021 alone. More Sox17 and CXCR4 were expressed by cells incubated with both compounds (Figure 2(e)). These findings showed that CHIR99021 and LY294002 together induced the differentiation of iPSCs into a definitive endoderm more effectively than the use of either alone.

Induction of Differentiation of iPSCs into IPCs Using
Combinations of Small Molecules. Although the iPSCs differentiated into a definitive endoderm, they retained the internal ability to differentiate into mesodermal and ectodermal cells. Therefore, we used small molecules to inhibit all signaling by cell differentiation pathways, except that of pancreatic progenitor cells (Figure 2(a)). We compared the effects of      Stem Cells International small molecules that can inhibit the differentiation of cells into hepatic progenitor cells. We incubated cells in differentiation-inducing media containing Dor and SB431542 that inhibit the differentiation into ectodermal and mesodermal cells, respectively, and compared cells incubated with fibroblast growth factor 2 (FGF2), the ERK signaling inhibitor FR180204, and the hedgehog signaling inhibitor SANT-1 to cause the differentiation of cells into mature pancreatic progenitor cells. The morphology of 6-day cultured cells differed according to the conditions under which differentiation was induced. In general, all conditions induced vigorous cell growth, resulting in high cell density. However, the morphology and growth rates of the cells that did not significantly differ (Figure 3(a)). Therefore, we analyzed the expression of the transcription factor characteristic of the pancreatic progenitor cells, Pdx1, Ngn3, Nkx6.1, Sox9, and NeuroD, under all conditions. The overall expression of the transcription factors was higher than that of the basic differentiation factors in group 1. However, the expression of each transcription factor was significantly increased in groups 4 and 6 incubated with SANT-1 compared with each transcription factor in the other groups (Figure 3(b)). These results showed that inhibiting hedgehog signaling by a specific pathway resulted in a more effective differentiation into pancreatic progenitor cells. After differentiation into pancreatic progenitor cells induced under step 2 conditions, we similarly induced differentiation into IPCs over 8 days in culture medium containing Dex, Nic, and For (Figure 4(a)). We then used qPCR to analyze the gene expression of insulin and glucagon, which are representative hormones of mature pancreatic islet cells and the transcription factors, Pdx1, and Nkx6.1. The mean expression of each marker in all groups tended to increase compared with group 1 but did not reach statistical significance. However, the gene expression was markedly increased in groups 4 and 6 incubated with SANT-1 compared with group 1, thus confirming the statistical significance of all groups (Figure 4(b)). Group 6 that had relatively more complete differentiation than the other groups was analyzed by immunostaining and flow cytometry. Cells that expressed insulin and glucagon were scattered among the cells in group 6. In addition, differentiation occurred in small groups rather than in large clusters (Figure 4(c)). Cells expressing Pdx1 were identified in most groups, but rarely among cells expressing insulin (Figure 4(d)). The flow cytometry findings indicated that approximately 38% of all differentiated cells expressed insulin, and 23% expressed glucagon. The most important transcription factor of pancreatic islet cells, Ngn3, was expressed in approximately 66% of all differentiated cells (Figure 4(e)). These results indicated that a few cells could directly differentiate into islet cells even in the presence of a large number of islet progenitor cells expressing Pdx1 and Ngn3. Thus, differentiation conditions could be improved to increase differentiation efficiency.

Ability of Differentiated IPCs to Secrete Insulin.
We quantified the differentiation of IPCs based on combinations of inducing factors and determined changes in the glucoseregulated quality of functional IPCs. Concentrations of insulin spontaneously secreted in culture medium were analyzed on day 5 of differentiation into mature IPCs. All groups secreted insulin, but the amount did not significantly differ among the groups (Figure 5(a)). Insulin concentrations were approximately 15-fold higher on day 8 than on day 5 in all groups. In particular, SANT-1-treated groups 4 and 6 showed higher insulin concentrations than the other groups, and the increase in concentration did not significantly differ among the other groups ( Figure 5(b)). Forskolin is an adenylate cyclase activator that is involved in the vitality and growth of cells during IPC differentiation, but it also causes mature pancreatic cells to release more insulin. Therefore, we incubated the cells for another 4 d after removing For and then measured insulin concentrations in the media. The insulin concentrations were decreased by >2-fold and slightly decreased in the cells incubated without and with SANT-1, respectively ( Figure 5(c)).
We also determined whether glucose could stimulate differentiated IPCs to secrete insulin. Group  Comp-FITC-A 9 Stem Cells International showed increased concentration of insulin in the medium when KCL was added during differentiation, but no significant secretion based on glucose concentration was observed ( Figure 5(d)). However, the group with SANT showed an increase in insulin secretion in response to a high concentration of glucose (Figures 5(e) and 5(f)). Overall, high levels of natural insulin and the ability to control insulin secretion using glucose were evident only in the group incubated with SANT-1 that inhibited hedgehog signaling mechanisms. These results suggested that inhibiting hedgehog signaling plays a major role in the differentiation of iPSCs into mature pancreatic islet cells.
3.5. IPC Culture to Create Cells Similar to Islet Cells. We differentiated IPCs using small molecules in 2D adherent cultures and attempted to cultivate spheroid clusters of IPCs similar to natural islet cells. After 10 d of differentiation step 3, the cells were separated, and the cell density was high. Small cell aggregates formed on day 1, and smaller spheroid clusters of a uniform size formed on day 5 (Figure 6(a)). Step 3 5d (a) Medium insulin (uIU/mL) Step 3 8d 10 Stem Cells International The spheroid IPC clusters were coordinated, coherent, and expressed insulin and glucagon ( Figure 6(b)). We confirmed that the spheroid morphology and basic insulin secretion ability of the IPCs could be maintained in long-term culture in vitro before transplantation in vivo. Spheroid IPCs were maintained for a considerable period under any culture conditions. However, their ability to secrete insulin gradually decreased when incubated in a differentiation induction medium without FBS.
Basal c-peptide secretion was low but significantly decreased on day 11 when For was removed from the differentiation induction medium. Similarly, c-peptide secretion by islet cells cultured in CMRL 1066 medium was lower without, than with For, but remained stable for 7 d (Figure 6(c)). We suggest that the optimal conditions for culturing spheroid IPCs comprise the use of CMRL 1066 medium for 7 d, as maintaining stable cultured normal islet cells in vitro beyond 7 d is difficult.

Discussion
The same method has been applied to differentiate ESCs and iPSCs in most studies of insulin-producing cells. This is because even though they are induced to differentiate by RNA-based viruses, iPSCs have the same differentiation ability as ESCs. However, iPSCs can avoid the ethical problems involved in collecting ESCs and are easily produced from adult cells. However, ESCs and iPSCs that remain undifferen-tiated after the differentiation process pose a risk of tumor formation if transplanted. This problem can be resolved using adult MSCs, but they have limited differentiation ability to differentiate into target cells. Therefore, methods of removing undifferentiated cells to prevent posttransplant tumor formation and of blocking transplanted cell growth by encapsulation are under investigation. Therefore, many other studies aim to safely transplant iPSCs in vivo.
We simply differentiated iPSCs into IPCs using small molecules to inhibit all differentiation pathways except those of IPCs. That is, we combined small molecules to ensure the natural differentiation of iPSCs towards IPCs.
We combined each of the GSK3β-specific inhibitor CHIR99021 [20] and the phosphoinositide 3-kinase inhibitor LY294002 with activin A to induce the differentiation of iPSCs into a definitive endoderm. Slightly more FoxA2 was expressed in the group induced with CHIR99021 and activin A. However, the results of the expression of other specific markers showed that adding LY294002 would be more effective. Pancreatic progenitor cells were induced by incubating the definitive endoderm with the FGF family members, RA and Dor, as described. Therefore, we investigated whether IPC differentiation was more effective after incubation with FR180204, an ATP-competitive inhibitor of ERK1 and ERK2 [20,27] and SANT-1 [23,24], which inhibits differentiation into hepatocyte precursor cells [23][24][25][26]. The findings showed that SANT-1 and FR180204 increased the induction of differentiation of iPSCs into IPCs, increased insulin levels

11
Stem Cells International in culture medium, and produced functional IPCs that responded to glucose.
Many studies have investigated the generation of insulinproducing cells from stem cells. However, insulin secretion, mitochondrial metabolism, and specific gene expression of adult islet cells remain insufficient in insulin-producing cells derived from stem cells. This problem has been overcome by differentiating cells with better insulin-producing ability using PCL/PVA nanofibrous scaffolds and collagen-coated, electrospun polyethersulfone nanofibers [31,32]. More diverse combinations of differentiation-inducing factors and signaling inhibitors have also been applied [4,33]. Therefore, the production of mature insulin-producing cells should be further investigated from a complex perspective that considers the environment and cell signal transduction, rather than focusing on a single differentiation inducing factor. Here, we described a simple and efficient differentiation method and an important mechanism.

Conclusions
We rendered IPCs more similar to natural islet cells by culturing them to acquire a spheroid morphology because the 2D culture resulted in IPCs that did not secrete insulin in response to glucose stimulation. We applied combinations of small molecules to 2D cultures to induce the differentiation of iPSCs into functional IPCs that respond to glucose; however, further maturation and growth allowed the IPCs to form spheroids that were easy to transplant and could survive posttransplantation.
We suggest that inducing the differentiation of iPSCs into IPCs using this method would be suitable for human transplantation. The iPSCs were maintained and differentiated on xeno-free matrix vitronectin that supported the growth and differentiation of human iPSCs under serum-free feeder-free conditions. In addition, we induced differentiation using small molecules, which are stable, safer than protein growth factors, and economical enough for mass production.
This study differentiated iPSCs into insulin-producing cells using a new method. However, these cells did not have the same maturity as adult islet cells, and differentiation rates were low. The amount of insulin secreted by adult islet cells according to the glucose concentration is most important. The maturation stages in the differentiation process will require further detailed investigation before cells with the same ability to produce insulin as adult islet cells can be derived from stem cells. Therefore, we plan to investigate embryonic pancreas development to identify key important points involved in maturation that would be suitable for application in clinical trials.