Abstract
Purpose :
Here we report on the use of a novel, AI-based cytometry platform to characterize and isolate, without labels, Retinal Cells (RCs) for cell and genetic applications. The differentiation of induced pluripotent stem cells iPSCs to become difficult to obtain cell types such as Retinal Cells is an accepted approach to create material for cell therapy research and ultimately for clinical use. To obtain cells that are both pure and “untouched”, meaning that there are no labels or other artificial markers that could confound downstream uses of the cells. The need for highly pure cell stock with no labels is a key need for the field.
Methods :
Here we used Ghost Cytometry, a new label-free cellular analysis and sorting platform built with proprietary optics and artificial intelligence (AI), to characterize and isolate truly untouched RCs during pluripotent cell differentiation. Ghost Cytometry was used to identify RCs generated from iPSC cell cultures and isolate therapeutically relevant phenotypes, an ongoing need in cell therapy R&D.
Results :
By capturing single-cell morphometric data, we characterized human NPCs from the parent iPSC culture. A set of machine-learning derived classifiers was generated to identify these phenotypic classes in unlabeled iPSC cultures. The classifiers showed area under the curve (AUC) performance ranges for detecting specific, phenotypically defined as NPC and iPSC populations with an SVM score that could clearly separate the two cell types in label free mode.
Conclusions :
We were able to obtain retinal cells that are both pure and “untouched”, free of labels or other artificial markers that could confound downstream uses of the cells. These cells are now available to create material for cell therapy research and ultimately for clinical use.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.