Introduction: Stem cell-derived islets (SC-islets) are unlimited human beta cell sources for transplantation that could provide a curative diabetes treatment. However, few in vitro generated SC-islets achieve an in vitro glucose-stimulated insulin secretion (GSIS) response equivalent to primary human islets in terms of the magnitude of insulin secretion or a biphasic pattern of insulin release1.

Methods: We here optimized the SC-islets differentiation protocol by an additional WNT signaling pathway inhibition at stage 2. The differentiation efficiency was evaluated by flow cytometry analysis. Dynamic and static insulin secretion of the SC-islets derived from the optimized protocol in response to different nutrient sources, including glucose, pyruvate, amino acid, and palmitate, were investigated. The oxygen consumption rate (OCR) of SC-islets in response to different nutrient sources was quantified by the Seahorse analysis. Finally, SC-islets derived from the optimized protocol were transplanted in mice under kidney capsule to examine the in vivo functionality.

Results: Our data showed that the WNT inhibition at stage 2 significantly increased Chromogranin A positive endocrine cell (P<0.001) and C-peptide positive cell (P<0.001) population in the studied induced pluripotency stem cell line than non-WNT inhibition at stage 2 (Figures A and B). SC-islets can maintain GSIS ability for 4 weeks in vitro with an average secretion index (SI) of 1.45. The average SI significantly decreased in week 6. However, this prolonged in vitro culture decreased glucagon-insulin and somatostatin-insulin bi-hormone cell fraction (Figures C and D). SC-islets from 3 batches of differentiation showed an elevated OCR in response to 20 mM glucose stimulation (Figure E). Dynamic GSIS shows that SC-islets respond to pyruvate and glucose stimulation (Figure F). Glutamine cannot induce insulin secretion in human islets and SC-islets; leucine induces elevated insulin secretion in human islets but not in SC-islets. Furthermore, the combination of 20 mM glucose and amino acid amplifies insulin secretion in human islets but not in SC-islets (Figures G and H). Palmitate alone cannot induce insulin secretion in human islets and SC-islets. When glucose presences, palmitate amplifies insulin secretion in human islets, not in SC-islets (Figures I and J). Transplantation of 600 SC-islets under the kidney capsule successfully prevented alloxan-induced diabetes in 3 out of 3 mice.

Conclusions: The SC-islets differentiation efficiency in this study was ensured through an optimized planar differentiation protocol. SC-islets’ insulin secretion in response to different nutrient sources was different from human islets, although they showed elevated OCR in response to glucose stimulation. This study provides a reference for further optimizing SC-islets differentiation protocol by targeting the different nutrient metabolism processes.