Delivery of shRNA via lentivirus in human pseudoislets provides a model to test dynamic regulation of insulin secretion and gene function in human islets

The loss of functional beta cell mass is the central pathology for both type 1 and type 2 diabetes (Kahn 2001; Atkinson et al. 2014; Chen et al. 2017). As the three‐dimensional structure of pancreatic islets supports viability and function of beta cells through cell‐cell and cell‐matrix communications (Rutter and Hodson 2015; Arous and Wehrle‐Haller 2017; Reissaus and Piston 2017; Briant et al. 2018), it is critical to address beta cell pathophysiology in pancreatic islets. Rodent islets are readily available, cost effective, can be easily genetically manipulated, and can be compared with syngeneic animals to connect in vitro observations to in vivo phenotype. However, human islets differ substantially from their rodent counterparts anatomically and functionally. Humans and mice show distinct islet innervation, cell distribution, and ratio of beta to alpha cells (Arrojo e Drigo et al. 2015). Glucose transporters, ion channels, the ratio of first/second phase of glucose‐stimulated insulin secretion (GSIS), and amyloid deposition also differ between human and mouse islets (Arrojo e Drigo et al. 2015; Dai et al. 2016; Skelin Klemen et al. 2017). Thus, studies of human islets from organ donors are important for understanding the regulation of islet function and beta cell viability in humans. However, human islets are limited in availability, costly, difficult to maintain in culture, and challenging to genetically modify. Islet function, including GSIS, reduces over time in culture (Paraskevas et al. 2000; Arzouni et al. 2018). Genetic manipulation of intact islets by liposomal or viral‐mediated vehicles has low efficiency and typically requires partial dispersion or enzyme digestion that compromises insulin secretion and removes cell‐cell communication. The enormous heterogeneity of islet sizes also introduces high variability in assays. To overcome the variability of isolated human islets, islet spheroids or pseudoislets composed of dissociated reaggregated islet cells has been used. Reaggregated islet cells form uniformly sized pseudoislets that maintain similar spatial distribution of beta and alpha cells with better first phase GSIS compared with dispersed cells (Hopcroft et al. 1985; Halban et al. 1987) and original islets (Zuellig et al. 2017; Yu et al. 2018). Pseudoislets also lend themselves to more efficient gene modification (Caton et al. 2003; Arda et al. 2016; Peiris et al. 2018). Caton et al. reported that lentiviral‐mediated overexpression of connexin cDNA does not interfere with pseudoislet formation and allows for the transduction of a large proportion of cells (Caton et al. 2003). Thus, pseudoislets appear to offer a unique and useful model to assess human islet function after the modulation of gene expression. Using readily available reagents and resources, we first characterized pseudoislets for insulin secretion in response to secretagogues by perifusion and analyzed their expression of markers for islet cell types, beta cell function, cell stress, and extracellular matrix (ECM). We compared transduction efficiency between adenovirus and lentivirus side by side and determine the impact of transduction on GSIS by perifusion. Following characterization of the pseudoislet platform, we tested short hairpin RNA (shRNA) delivered with a lentiviral vector targeting glucokinase (GCK) as a model target to test dynamism of insulin secretion by perifusion. Our data demonstrate that lentiviral‐mediated gene downregulation combined with a simple protocol to form human pseudoislets is a useful tool that enables assessment of the impact of gene function on islet GSIS.