Non-invasive marker-independent high content analysis of a microphysiological human pancreas-on-a-chip model.

The increasing prevalence of diabetes, its heterogeneity, and the limited number of treatment options drive the need for physiologically relevant assay platforms with human genetic background that have the potential to improve mechanistic understanding and e\xpedite diabetes-related research and treatment. In this study, we developed an endocrine pancreas-on-a-chip model based on a tailored microfluidic platform, which enables self-guided trapping of single human pseudo-islets. Continuous, low-shear perfusion provides a physiologically relevant microenvironment especially important for modeling and monitoring of the endocrine function as well as sufficient supply with nutrients and oxygen. Human pseudo-islets, generated from the conditionally immortalized EndoC-βH3 cell line, were successfully injected by hydrostatic pressure-driven flow without altered viability. To track insulin secretion kinetics in response to glucose stimulation in a time-resolved manner, dynamic sampling of the supernatant as well as non-invasive real-time monitoring using Raman microspectroscopy was established on-chip. Dynamic sampling indicated a biphasic glucose-stimulated insulin response. Raman microspectroscopy allowed to trace glucose responsiveness in situ and to visualize different molecular structures such as lipids, mitochondria and nuclei. In-depth spectral analyses demonstrated a glucose stimulation-dependent, increased mitochondrial activity, and a switch in lipid composition of insulin secreting vesicles, supporting the high performance of our pancreas-on-a-chip model.