Abstract
Purpose :
Macular fibrosis, observed at the end stage of neovascular AMD, is the leading cause of irreversible blindness in AMD and has been identified as a cause of poor anti-VEGF therapeutic outcomes. Although the pathogenesis of macular fibrosis is still unclear, excessive extracellular matrix deposition by myofibroblasts appears to be a key event. Here, using Collagen-1 (Col1)-YFP reporter mice, we compare the dynamics of collagen production in vivo over time in mice receiving the traditional laser-induced choroidal neovascularization (LiCNV) model to mice receiving the recently described two-stage LiCNV model of subretinal fibrosis, and ex vivo, we examine the CNV tissue for fibrovascular markers in both models.
Methods :
LiCNV was carried out using the Micron IV platform (532 nm, 300mW, 100 ms, 50 um spot size, 3 spots per eye) in 8-12 week old Col1-YFP reporter mice. After 7 days, a cohort of the mice received a second laser burn directly over the initial lesion. Mice were fundus imaged every 2-3 days using the YFP filter on the Micron IV to monitor the dynamics of collagen-producing cells infiltrating the CNV lesions. Retinal pigment epithelium flatmounts were harvested 2- and 3-weeks post LiCNV and stained for fibrotic markers, while retinas were harvested for ELISA, RT-PCR and flow cytometry analysis.
Results :
Mice that received the two-stage LiCNV showed significantly larger Col1-YFP lesions from 7 days on and YFP was maintained for the 3 weeks post LiCNV induction. In contrast, much smaller Col1-YFP lesions were observed in mice receiving the one-stage LiCNV and YFP was diminished by 3 weeks post induction. Isolectin IB4 staining revealed a more complex networked vasculature in the two-stage lesions, accompanied with enhanced alpha smooth muscle actin and collagen-hybridizing peptide staining, indicative of increased EMT and collagen remodelling, respectively. Interestingly, flow cytometry showed negligible YFP positive cells in the retina in both LiCNV models, indicating that cells involved in collagen production were not in the retina, which was supported by RT-PCR data showing limited changes in EMT and fibrosis markers.
Conclusions :
Overall, our in vivo and ex vivo data support the two-stage LiCNV model as a more suitable model for studying subretinal fibrosis over the traditional one-stage LiCNV model, which is better representative of a wound healing model.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.