Cell-autonomous RPE dysfunction is sufficient to instigate CC atrophy and CNV in an iPSC model of Macular Degeneration

Abstract

Purpose : The purpose of this study was to use a representative hydrogel-based model of the RPE-CC complex to investigate RPE-choriocapillaris (CC) cross-talk in the development of CC atrophy and choroidal neovascularization (CNV), disease hallmarks of age-related macular degeneration (AMD) and related macular dystrophies, like Sorsby’s fundus dystrophy (SFD).

Methods : SFD (TIMP3, S204C) and control induced pluripotent stem cells (iPSCs) were differentiated into RPE, endothelial cells (ECs) and mesenchymal stem cells (MSCs). To recapitulate in-vivo RPE-CC structure, hiPSC-ECs were encapsulated in PEG hydrogels plated on a layer of hiPSC-MSCs and hiPSC-RPE was grown on the vascularized gel surface. Immunostaining and Western blotting analyses were performed to characterize RPE and CC characteristics in the hydrogel-based control vs. SFD RPE-CC tissue mimetic. Furthermore, Alamar blue cell proliferation assay, extracellular matrix (ECM)-cell invasion and migration assay and immunocytochemical analyses of vascular network characteristics (number, size, length) were used to evaluate the effect of control vs. SFD RPE-secreted factors on CC atrophy and CNV-relevant changes. All experiments were performed at least in triplicates and unpaired t-test was utilized to determine statistical significance (p<0.05).

Results : Characterization of hiPSC-RPE-CC model showed an RPE monolayer that expressed tight junction marker ZO-1 overlying a CC-like fenestrated endothelium that showed co-expression of EC-specific protein, CD31 and fenestration marker, PLVAP. Furthermore, consistent with CC atrophy and CNV in SFD, SFD-hiPSC-RPE-CC model showed vascular EC atrophy and invasion of RPE by ECs. Furthermore, SFD-hiPSC-RPE conditioned media alone was sufficient to induce CC atrophy and CNV relevant phenotypes (ECM invasion and migration (p< 0.05), EC proliferation(p< 0.05)) in ECs/CC-like vascular networks. Notably, pharmacological inhibition of FGF2 was sufficient to suppress CNV relevant phenotypes in this SFD hiPSC model (p< 0.01).

Conclusions : Using an iPSC-RPE-CC model, we have shown that alterations in RPE-secreted factors are sufficient to cause vascular changes, CC atrophy and CNV, observed in SFD/AMD. Furthermore, iPSC-RPE-CC model can be used to pharmacologically-target SFD/AMD relevant pathological changes, like CNV

This is a 2020 ARVO Annual Meeting abstract.