Abstract
Purpose :
Mobile, ramified macrophage-like cells (MLCs) are detectable at the vitreoretinal interface (VRI) by optical coherence tomography (OCT) imaging in humans. Macrophages (Macs) are heterogeneous cells and MLC identity is unknown. MLCs could include microglia (yolk sac-derived retinal Macs), perivascular Macs (long-lived Macs supporting healthy vasculature), or monocyte-derived Macs (inflammatory Macs derived from blood). The goal of this study was to characterize VRI Macs.
Methods :
OCT and OCT-angiography were performed on C57BL6/J mice. Confocal immunofluorescence (IF) for CD31 (endothelial cells, EC) and IBA1 (pan-Mac marker) were performed at the retinal surface on wholemounts. OCT-angiography and CD31 staining were used to match IF images and en face OCT slabs. Cx3cr1CreER x Rosa26zsGreen (MacGFP) mice were used to fate map Macs 1 week and 1 month post-tamoxifen treatment. IF staining of frozen sections was performed using IBA1, CD31, and/or Tmem119 to delineate microglia. Tmem119GFP mice were used to quantitate GFP+ microglia and Mrc1+ perivascular Macs by multi-parameter flow cytometry.
Results :
En face OCT slabs above the VRI revealed cellular structures. These cells were IBA1+ and showed either a ramified morphology or were colocalized to EC. Fate mapping in MacGFP mice showed that VRI Macs were 84% GFP+ and plexiform Macs were 98% GFP+ at Week 1 (p<0.05). At Month 1, VRI and plexiform layer Macs were equally GFP+ compared to Week 1 (82% vs 96%, p<0.05 between groups). VRI Macs were 72% Tmem119+ or 21% Tmem119negECadjacent (p<0.001). Macs from Tmem119GFP retinas were 90% GFP+Mrc1neg microglia or 10% GFPnegMrc1+ perivascular Macs (p<0.05) by flow cytometry.
Conclusions :
VRI Macs are an emerging disease biomarker. At steady-state, VRI Macs are ~70% microglia and ~20% perivascular Macs without any monocyte-derived Mac component. These findings are potentially important for diseases like multiple sclerosis and Alzheimer’s, where microglia and perivascular Macs have been implicated respectively.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.