Implications for iPS-based disease modeling: significant variability is observed in RPE cells differentiated from identical sources

Abstract

Purpose : Many human diseases lack an animal model. Disease modeling using induced pluripotent stem cells (iPSC) derived directly from patients has provided a promising tool for studying otherwise inaccessible human disorders. While this approach allows the direct study of diseased human cells, the procedure involves multiple levels of manipulation with the potential to introduce variability. Quantifying the degree of variability is necessary before making conclusions regarding the differences between diseased and control patients. In this study we derive retinal pigment epithelial (RPE) cells from carefully controlled iPSC lines to determine the inherent variability between what should be identical cell lines.

Methods : To investigate the amount of variability arising from differentiating iPSC into RPE cells we derived multiple clonal iPSC lines designed to control for cell origin, genetic background, reprogramming techniques/factors and cell differentiation protocols. Fibroblasts were infected with lentivirus carrying the inducible reprogramming factors, allowed to divide once, and then induced to generate iPSC colony pairs. Three sets of identical iPSC colony pairs were then differentiated into RPE cells. We performed multiple physiologically-relevant assays to examine cellular morphology, phagocytic capacity, metabolic fitness and stress responses.

Results : Depending on the assay used we saw differing levels of variability. The capacity to differentiate into pigment-expressing RPE cells varied greatly between identical pairs. While the phagocytic capacity showed low variability between identical pairs, measurements of metabolic fitness had a high degree of variability.

Conclusions : The generation and differentiation of iPS-derived cells introduces variability despite being derived from genetically identical cells and being reprogrammed to pluripotency using identical factors and techniques. This inherent variability in the method could make discerning differences between diseased and control patient cells very challenging. Therefore, the level of variability introduced from a differentiation protocol, and the sensitivity of functional assays to this, needs to be considered when determining the number of samples to test, the assays used to analyze cells and confidence assigned to differences detected.

This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.