Comparative Spatial Profiling of Mouse Retinopathy of Prematurity (ROP) Model To Elucidate Cell-Specific Interactions In Retinal Neovascularization

Abstract

Purpose : The retina is organized into >9 different layers where function of cell types in each layer contribute to the overall eye integrity. In ROP, the early stages are characterized by angiogenic inhibition and regression, which can lead to neovascularization. Underlying this progression are cell-cell interactions that exacerbate this disease severity. We therefore conducted a comparative spatial transcriptomics study to profile these granular disease-specific interactions to investigate mechanisms underlying ROP progression.

Methods : MERFISH-based spatial transcriptomics is used in our comparative study. The panel consists of 500 genes to distinguish across 13 major cell types probing for genes underlying neovascularization (E.g VEGF, adhesion, and HIF-related genes). Eye cross sections (10μm) were selected based on RNA quality and morphological preservation. Decoding, segmentation, and cell type identification via cell type markers were then used to obtain spatial maps for each tissue.

Results : Progression of neovascularization can be observed in mice subjected to hyperoxia between vascularized and avascular retina regions. To elucidate cell-specific interactions underlying ROP, we spatially profiled mouse retinas from eight mice with/without undergoing oxygen-induced retinopathy (OIR). Leveraging on public scRNA dataset, our gene panel validated known inter-endothelial interactions from VEGF-related and independent pathways (Fig 1) present during OIR but absent under Normoxia. Our preliminary analysis suggests that these disease-specific interactions drive downstream effectors which contribute to neovascularization.

Conclusions : We demonstrate that spatial profiling can resolve retinal cell-cell interactions with its corresponding spatial proximity between the different major cell types. This profiling enables high-sensitivity detection of cell types to create unique disease-specific spatial maps, thereby tracking mechanisms underlying ROP progression for application in monitoring intervention efficacy.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.

 

Fig 1: Cell type ligand-receptor pairs (p<0.05) in P14 OIR (left) and Normoxia (right) mined from Binet et al (2020). Science. Colors represent “source” ligand connected to the “target” receptors while line thickness represents the expression of each pair. (ACs: Amacrine cells, VE: Vascular endothelium)