Reproducibility and Robustness of PhysioMimix LC12, a Liver Microphysiological System, Under Varying Culture Conditions and Cell Type Combinations

Abstract

Traditionally, safety evaluation of drugs and chemicals is performed using in vivo animal testing. Although animal models have been thoroughly validated and have decades of accumulated data, issues such as the complexities of translating in vivo data to human data and the biological incompatibilities between animals and humans continue to persist. There has also been a growing push from both the public and government agencies to move away from and reduce animal testing. In response, NAMS, or New Approach Methodologies, have been proposed as alternative methods to animal testing. One type of NAMs is microphysiological systems, or MPS. According to the US Food and Drug Administration, MPS are microscale cell culture platforms for in vitro modeling of functional features of a specific tissue or organ of human or animal origin that expose cells to a microenvironment that mimics the physiological aspects important for their function or pathophysiological condition. While many MPS models have been developed, independent validation and characterization of each of these models to determine their specific utility and context of use is still lacking. As such, this study aimed to determine the utility and reproducibility of a specific liver MPS model, the PhysioMimix™ LC12, using different human hepatocyte sources in conjunction with different combinations of nonparenchymal cells (NPCs).

To accomplish this aim, multiple replicate studies averaging 14 days in length were conducted using the PhysioMimix™ LC12 seeded with either primary human hepatocytes (PHHs) or induced pluripotent stem cell-derived hepatocytes (iPSC-derived hepatocytes), with and without NPCs. Albumin and urea secretion, lactate dehydrogenase leakage, CYP3A4 activity and drug metabolism (using the CYP3A4 substrate, midazolam (MDZ)) were measured to assess basal function and metabolic capacity. Monocultured PHHs and co-cultured PHHs demonstrated moderately high hepatic function at the start of culture that gradually decreased over 14 days of culture, with variability across replicate studies averaging 20-40% (coefficient of variation). Two other liver MPS models tested previously in-house displayed slightly lower levels of albumin and urea secretion that were sustained over 14 days of culture, and CYP3A4 activity levels were approximately five-fold less. Quad-cultured PHHs and iHeps with and without NPCs seeded in the PhysioMimix™ LC12 exhibited low to negligible hepatic function.

Overall, we found that the PhysioMimix™ LC12 seeded with PHHs or PHHs + THP-1s is a fairly robust, reproducible, and functional liver MPS model. With its relatively high CYP3A4 activity that is sustained at levels above the two liver MPS and 2D comparators over 14 days of culture, we determined that this model’s utility lies in the investigation of metabolism of compounds, particularly in the context of drug discovery and development. By applying an in-depth methodology to independent validation of a liver MPS, this study provides an example of how to validate future MPS, emphasizes the importance of defining each MPS’s specific context of use, and increases confidence and trust in MPS, ultimately laying important groundwork for further characterization of the PhysioMimix™ LC12 as well as increased implementation of MPS in regulatory and safety evaluation.