Abstract: Generating stem cell derived pancreatic islets (SC-islets) with functional beta cells has become a reality in recent years. SC-islets show great promise as a renewable cell source for the treatment of insulin-deficient diabetes. However, SC-islets differ metabolically from primary human islets, especially in having low mitochondrial coupling to glucose stimulated insulin secretion (GSIS). The aim of our study was to determine whether there is functional heterogeneity between beta cells within the SC-islets, and if so, what characterizes the more functional beta cells. Mitochondrial function was used to identify populations of interest. As we have previously shown in other systems that mitochondrial age can be used as a proxy for function, we created a reporter hESC line, H1 SNAP-OMP25, which enables temporally controlled labelling of mitochondria. We sorted out four cell populations differing in their mitochondrial age profiles as well as levels of ENTPD3, an ectonucleoside highly expressed in beta cells. Three of the four populations contained over 95% beta cells. We found that beta cells with older mitochondria had higher glucose-stimulated insulin secretion. Additionally, contrary to previous reports on sorting based on ENTPD3 levels, we found that moderate rather than high levels of ENTPD3 were associated with better functionality. The discrepancy is likely due to the old mitochondrial axis allowing us to sort out a pure beta cell population with moderate levels of ENTPD3. We conclude that there is functional heterogeneity within the beta cells of SC-islets identifiable by mitochondrial age as well as ENTPD3 levels. We are currently deciphering the transcriptomic and metabolic basis underlying the identified functional differences. This knowledge could be used to modify the SC-islet differentiation protocol further, allowing the generation of better islets for therapeutic and disease modelling applications.
