AI-Driven Analysis Unveils Functional Dynamics of Müller Cells in Retinal Autoimmune Inflammation

Müller cell is the most common type of glial cell in the human and mouse retina, playing a crucial role in maintaining retinal homeostasis. In addition to providing structural support to the retina, Müller cells can also supply trophic substances to retinal neurons, remove metabolic waste, mitigate oxidative stress, and promote synaptic activities. However, many roles of Müller cells remain largely unknown, particularly for those in the inflamed retina. In this article, we reanalyzed a single cell RNA-seq (scRNA-seq) dataset from Aire-/- mice, which exhibits autoimmune retinal inflammation, specifically focusing on Müller cells and T cells, identifying nine distinct Müller cell subgroups along with five T cell subgroups. Among them, three subgroups of Müller cells are activated Müller cells, representing over 60% Müller cells in the inflamed retina. Using SCassist - an Artificial Intelligence (AI) based workflow assistant for single-cell analysis, we constructed a comparison matrix to quantify the involvement of pathways characterizing the functions of each Müller cell subpopulation. The activated Müller cells primarily present a macrophage-like phenotype with or without augmentation of the known Müller cell functions. Trajectory analysis further identified two paths, validating the presence of these two phenotypes, governed by Neurod1 and Irf family transcription factors (TFs). We further inferred the interactions between Müller cells and T cells and observed that activated Müller cells do not exhibit extra chemoattraction to Th1 cells compared to other Müller cells but display nearly exclusive expression of immune checkpoint molecules, primarily targeting Th1 cells. Our findings open new avenues for understanding the specialized mechanisms of retinal pathogenic autoimmunity and identifying candidates to explore potential inhibitory pathways in the inflamed retina.