An Abbreviated Protocol for In Vitro Generation of Functional Human Embryonic Stem Cell-Derived Beta-Like Cells

Abstract

The ability to yield glucose-responsive pancreatic beta-cells from human pluripotent stem cells in vitro will facilitate the development of the cell replacement therapies for the treatment of Type 1 Diabetes. Here, through the sequential in vitro targeting of selected signaling pathways, we have developed an abbreviated five-stage protocol (25-30 days) to generate human Embryonic Stem Cell-Derived Beta-like Cells (ES-DBCs). We showed that Geltrex, as an extracellular matrix, could support the generation of ES-DBCs more efficiently than that of the previously described culture systems. The activation of FGF and Retinoic Acid along with the inhibition of BMP, SHH and TGF-beta led to the generation of 75% NKX6.1+/NGN3+ Endocrine Progenitors. The inhibition of Notch and tyrosine kinase receptor AXL, and the treatment with Exendin-4 and T3 in the final stage resulted in 35% mono-hormonal insulin positive cells, 1% insulin and glucagon positive cells and 30% insulin and NKX6.1 co-expressing cells. Functionally, ES-DBCs were responsive to high glucose in static incubation and perifusion studies, and could secrete insulin in response to successive glucose stimulations. Mitochondrial metabolic flux analyses using Seahorse demonstrated that the ES-DBCs could efficiently metabolize glucose and generate intracellular signals to trigger insulin secretion. In conclusion, targeting selected signaling pathways for 25-30 days was sufficient to generate ES-DBCs in vitro. The ability of ES-DBCs to secrete insulin in response to glucose renders them a promising model for the in vitro screening of drugs, small molecules or genes that may have potential to influence beta-cell function.

Fig 1. Short protocol outline.

(A) Schematic overview of the 25 to 30-day protocol to generate human H1 ES-derived beta-like cells (DBCs). Below, images of the differentiated H1 cells and the control cells (Non-Treated ES cell) at each stage are shown. The arrow symbol identifies tube-like structure in the differentiated cells in the stage 2. The star symbol identifies detached dead cells as spheres in the Non-Treated cells in stage 4. Scale bar = 100μm for all cell images. The red font indicates modifications to molecules or timing in comparison to the protocol described by Rezania et al [9]. (B) Expression analyses of SOX17, FOXA2 and Gooscoid as Definitive Endoderm (DE), Sox1 as ectoderm, and Brachyury as mesoderm-specific markers in the H1 ES cells differentiated on MEF, Mouse Embryonic Fibroblast; as EB (Embryoid Bodies) or on Geltrex, analyzed by quantitative RT-PCR. (* p< 0.05, **p< 0.01, p***<0.001, significant differences between the treated and control cells in each condition, unpaired two-tailed t-test, n = 3).

Fig 2. The efficiency of Definitive Endoderm (DE) and Gut Tube Endoderm formation at the stage 1 and 2 of the differentiation protocol.

(A) Flow cytometry, and immunofluorescence staining for DE-specific markers in the differentiated H1 ES cells. (B) Quantitative RT-PCR results for Gut Tube Endoderm-specific markers are shown in (B), showing genes up-regulated in the stage 1, and (C) maintained highly expressed genes in the stage 2. Scale bar = 40μm. (*p< 0.05, **p< 0.01, p***<0.001, unpaired two-tailed t-test, n = 3).

Fig 3. Characterization of the differentiated H1 ES cells at the Pancreatic Progenitor (PP) and the Endocrine Progenitor (EN) stages.

(A) From left to right, flow cytometry for PDX1/FOXA2, immunofluorescence staining for PDX1, and qRT-PCR analysis for the PP-specific genes in the differentiated cells at stage 3. (B) Flow cytometry for NGN3/NKX6.1, immunofluorescence staining for NGN3, qRT-PCR analysis for the EP-specific genes and below, immunofluorescence staining for PDX1/NKX6.1 in the differentiated cells at stage 4. (C) Immunofluorescence staining for ARX/PAX4, and qRT-PCR analysis for ARX and PAX4 in differentiated cells at the stage 4. Scale bar = 40μm. (*p< 0.05, **p< 0.01, p***<0.001, unpaired two-tailed t-test, n = 3).

Fig 4. Study of insulin and beta-cell marker expression in the human H1 ES-DBCs at stage 5.

(A) Flow cytometry and immunofluorescence staining for C-peptide/Glucagon. From left to right, gating of flow cytometry for detection of C-peptide and glucagon, flow cytometry for C-peptide and glucagon and immunofluorescence staining for C-peptide/glucagon in the ES-DBCs. (B) Flow cytometry and immunofluorescence staining for C-peptide/Somatostatin and (C) Insulin/NKX6.1, in the ES-DBCs. (D) Immunofluorescence staining for C-peptide/MAFA, (E) C-peptide/NeuroD1,(F) C-peptide/Syntaxin-1A, and (G) C-peptide/Synaptophysin in the ES-DBCs at the stage 5. Scale bar = 20μm. GCG: Glucagon, SS: Somatostatin.

Fig 5. The mRNA expression analysis of pancreatic islet, beta-cell and related genes in the differentiated human H1 ES-DBCs.

(A) Exact copy number of insulin mRNA molecules in the ES-DBCs and human islets by digital droplet RT-PCR (GAPDH was used for normalization). Quantitative real time RT-PCR analysis for (B) endocrine hormones, (C) Chromogranin A, (D) pancreatic transcription factors, Ca⁺² and K⁺ channels genes, (E) Glucose transporters (GLUT1 and 2) and PCSK2 as the enzyme required for pro-insulin processing and in the ES-DBCs compared to human islets. (*p< 0.05, **p< 0.01, p***<0.001, unpaired two-tailed t-test, n = 3).

Fig 6. Comparison of gene expression in human H1 ES-DBCs and mature beta-cells.

Expression of the top-ten most significantly enriched mRNAs in either adult mature or fetal beta-cells as described by Hrvatin et al. [26] were examined in ES-DBCs vs. the human adult islets via real time RT-PCR assay. (*p< 0.05, **p< 0.01, p***<0.001, unpaired two-tailed t-test, n = 3).

Fig 7. Examination of beta-cell stimulus-secretion coupling in human ES-DBCs vs. human islets.

(A) Measurement of C-peptide in the supernatant, and (B) lysates of H1 ES-DBCs and the human islets after stimulation by glucose. (C) Normalized secretion compared to intracellular C-peptide. (D) Temporal insulin secretion by perifusion in ES-DBCs and human islets. Correlation between (E) MAFA expression analyzed by qRT-PCR and (F) insulin secretion, in response to glucose stimulation in EN and ES-DBCs at stage 5. EN: ENdocrine cells as referred in Fig 1A. (*p< 0.05, **p< 0.01, p***<0.001, paired two-tailed t-test, n = 5).

Fig 8. Analyses of Ca⁺² flux, and respiration capacities of the human H1 ES-DBCs.

(A) Measurement of glucose-stimulated cytosolic Ca⁺² flux in the ES-DBCs, Non-Treated cell and MIN-6 beta-cell population. (B) Mitochondrial respiration (the potential of mitochondria to reserve energy) in ES-DBCs, Non-Treated and MIN-6 cells using the seahorse technique. (n = 4)-two technical replicates per batch, data are presented as Mean±SD. (*p< 0.05, **p< 0.01, p***<0.001, paired two-tailed t-test, n = 4).