The management of preretinal fibrovascular membranes, a devastating complication of advanced diabetic retinopathy (DR), remains challenging. We characterized the molecular profile of cell populations in these fibrovascular membranes to identify potentially new therapeutic targets. Preretinal fibrovascular membranes were surgically removed from patients and submitted for single-cell RNA-Seq (scRNA-Seq). Differential gene expression was implemented to define the transcriptomics profile of these cells and revealed the presence of endothelial, inflammatory, and stromal cells. Endothelial cell reclustering identified subclusters characterized by noncanonical transcriptomics profile and active angiogenesis. Deeper investigation of the inflammatory cells showed a subcluster of macrophages expressing proangiogenic cytokines, presumably contributing to angiogenesis. The stromal cell cluster included a pericyte-myofibroblast transdifferentiating subcluster, indicating the involvement of pericytes in fibrogenesis. Differentially expressed gene analysis showed that Adipocyte Enhancer-binding Protein 1, AEBP1 , was significantly upregulated in myofibroblast clusters, suggesting that this molecule may have a role in transformation. Cell culture experiments with human retinal pericytes (HRP) in high-glucose condition confirmed the molecular transformation of pericytes toward myofibroblastic lineage. AEBP1 siRNA transfection in HRP reduced the expression of profibrotic markers in high glucose. In conclusion, AEBP1 signaling modulates pericyte-myofibroblast transformation, suggesting that targeting AEBP1 could prevent scar tissue formation in advanced DR.

Background: Diabetic retinopathy and retinal vein occlusion are vision threatening retinal vascular diseases. Current first-line therapy targets the vascular component, but many patients are treatment-resistant due to unchecked inflammation. Non-invasive inflammatory imaging biomarkers are a significant unmet clinical need for patients. Imaging of macrophage-like cells on the surface of the retina using clinical optical coherence tomography (OCT) is an emerging field. These cells are increased in patients with retinal vascular disease, and could be a potential inflammatory biomarker. However, since OCT is limited by an axial resolution of 5–10 microns, the exact location and identity of these retinal cells is currently unknown. Methods: We performed OCT followed by confocal immunofluorescence in wild-type mice to identify macrophages within 5–10 microns of the vitreoretinal interface. Next, we used Cx3cr1^CreER/+; Rosa26^zsGreen/+ mice to fate map retinal surface macrophages. Using confocal immunofluorescence of retinal sections and flatmounts, we quantified IBA1⁺Tmem119⁺CD169^neg microglia, IBA1⁺Tmem119^negCD169^neg perivascular macrophages, and IBA1⁺Tmem119^negCD169⁺ vitreal hyalocytes. Finally, we modeled neuroinflammation with CCL2 treatment and characterized retinal surface macrophages using flow cytometry, OCT, and confocal immunofluorescence. Results: We were able to detect IBA1⁺ macrophages within 5–10 microns of the vitreoretinal interface in wild-type mice using OCT followed by confirmatory confocal immunofluorescence. Retinal surface macrophages were 83.5% GFP⁺ at Week 1 and 82.4% GFP⁺ at Week 4 using fate mapping mice. At steady state, these macrophages included 82% IBA1⁺Tmem119⁺CD169^neg microglia, 9% IBA1⁺Tmem119^negCD169⁺ vitreal hyalocytes, and 9% IBA1⁺Tmem119^negCD169^neg perivascular macrophages. After CCL2-driven neuroinflammation, many Ly6C⁺ cells were detectable on the retinal surface using OCT followed by confocal immunofluorescence. Conclusions: Macrophages within close proximity to the vitreoretinal interface are self-renewing cells, and predominantly microglia with minor populations of perivascular macrophages and vitreal hyalocytes at steady state. In the context of neuroinflammation, monocytes and monocyte-derived macrophages are a significant component of retinal surface macrophages. Human OCT-based imaging of retinal surface macrophages is a potential biomarker for inflammation during retinal vascular disease.

Patients with neovascular AMD (nAMD) suffer vision loss from destructive angiogenesis, termed choroidal neovascularization (CNV). Macrophages are found in CNV lesions from patients with nAMD. Additionally, Ccr2 ^–/– mice, which lack classical monocyte–derived macrophages, show reduced CNV size. However, macrophages are highly diverse cells that can perform multiple functions. We performed single-cell RNA-Seq on immune cells from WT and Ccr2 ^–/– eyes to uncover macrophage heterogeneity during the laser-induced CNV mouse model of nAMD. We identified 12 macrophage clusters, including Spp1 ⁺ macrophages. Spp1 ⁺ macrophages were enriched from WT lasered eyes and expressed a proangiogenic transcriptome via multiple pathways, including vascular endothelial growth factor signaling, endothelial cell sprouting, cytokine signaling, and fibrosis. Additionally, Spp1 ⁺ macrophages expressed the marker CD11c, and CD11c ⁺ macrophages were increased by laser and present in CNV lesions. Finally, CD11c ⁺ macrophage depletion reduced CNV size by 40%. These findings broaden our understanding of ocular macrophage heterogeneity and implicate CD11c ⁺ macrophages as potential therapeutic targets for treatment-resistant patients with nAMD.

Purpose: To quantitatively characterize macrophage-like cells (MLCs) at the vitreoretinal interface in different severity stages of diabetic retinopathy (DR) using optical coherence tomography angiography (OCTA). Methods: The study included 72 eyes of 72 subjects: 18 healthy controls, 22 diabetes mellitus (DM) without DR, 17 nonproliferative DR (NPDR), and 15 proliferative DR (PDR). We obtained repeated (average, 6.5; range, 3–10) macular OCTA scans for each eye. We registered and averaged the 3-µm OCT slab above the vitreoretinal interface to visualize MLCs. Using a semiautomated method, we binarized and quantified MLCs and compared MLC densities among groups. We also evaluated MLC distribution relative to underlying superficial capillary plexus vasculature and quantified MLCs overlying blood vessels within the perivascular 30-µm watershed region and within ischemic zones (defined as >30 µm from the nearest vessel). Results: MLC density was 2.8- to 3.8-fold higher in PDR compared with all other groups (P < 0.05 for all). MLC density in PDR was most increased in perivascular areas (3.3- to 4.2-fold; P < 0.05 vs. all) and on blood vessels (3.0- to 4.0-fold; P < 0.05 vs. all), and elevated to a lesser extent in ischemic areas (2.3- to 3.4-fold; P < 0.05 vs. all). MLCs were more likely to localize on blood vessels in DM without DR, NPDR, and PDR (P < 0.05 for all), but not healthy eyes. Conclusions: MLC density was significantly increased in PDR. MLCs clustered on blood vessels in diabetic but not in healthy eyes. Further studies are needed to confirm the origin, identity, and function of MLCs during DR.

In the modern era, patients often resort to the internet for answers to their health-related concerns, and clinics face challenges to providing timely response to patient concerns. This has led to a need to investigate the capabilities of AI chatbots for ophthalmic diagnosis and triage. In this in silico study, 80 simulated patient complaints in ophthalmology with varying urgency levels and clinical descriptors were entered into both ChatGPT and Bard in a systematic 3-step submission process asking chatbots to triage, diagnose, and evaluate urgency. Three ophthalmologists graded chatbot responses. Chatbots were significantly better at ophthalmic triage than diagnosis (90.0% appropriate triage vs. 48.8% correct leading diagnosis; p < 0.001), and GPT-4 performed better than Bard for appropriate triage recommendations (96.3% vs. 83.8%; p = 0.008), grader satisfaction for patient use (81.3% vs. 55.0%; p < 0.001), and lower potential harm rates (6.3% vs. 20.0%; p = 0.010). More descriptors improved the accuracy of diagnosis for both GPT-4 and Bard. These results indicate that chatbots may not need to recognize the correct diagnosis to provide appropriate ophthalmic triage, and there is a potential utility of these tools in aiding patients or triage staff; however, they are not a replacement for professional ophthalmic evaluation or advice.

Background: Neovascular age-related macular degeneration (nAMD) commonly causes vision loss from aberrant angiogenesis, termed choroidal neovascularization (CNV). Macrophages are heterogeneous cells that are necessary for experimental CNV, present in human CNV samples, and can display diverse functions, which are dependent upon both their origin and tissue microenvironment. Despite these associations, choroidal macrophage heterogeneity remains unexplored. Methods: We performed multi-parameter flow cytometry on wildtype (WT) and Ccr2^−/− mice after laser injury to identify macrophage subtypes, and determine which subsets originate from classical monocytes. To fate map tissue resident macrophages at steady state and after laser injury, we used the Cx3cr1^CreER/+ ; Rosa26^zsGFP/+ mouse model. We reanalyzed previously published single-cell RNA-seq of human choroid samples from healthy and nAMD patients to investigate human macrophage heterogeneity, disease association, and function. Results: We identified 4 macrophage subsets in mice: microglia, MHCII⁺CD11c⁻, MHCII⁺CD11c⁺, and MHCII⁻. Microglia are tissue resident macrophages at steady state and unaffected by laser injury. At steady state, MHCII⁻ macrophages are long lived, tissue resident macrophages, while MHCII⁺CD11c⁻ and MHCII⁺CD11c⁺ macrophages are partially replenished from blood monocytes. After laser injury, MHCII⁺CD11c⁻ macrophages are entirely derived from classical monocytes, MHCII⁻ macrophages originate from classical monocytes (90%) and an expansion of tissue resident macrophages (10%), and MHCII⁺CD11c⁺ macrophages are derived from classical monocytes (70%), non-classical monocytes (10%), and an expansion of tissue resident macrophages (20%). Single-cell RNA-seq analysis of human choroid found 5 macrophage subsets: two MHCII⁺CD11C⁻ and three MHCII⁺CD11C⁺ populations. One MHCII⁺CD11C⁺ subset was 78% derived from a patient with nAMD. Differential expression analysis identified up-regulation of pro-angiogenic gene expression in one MHCII⁺CD11C⁻ and two MHCII⁺CD11C⁺ subsets, including the disease-associated cluster. The upregulated MHCII⁺CD11C⁻ pro-angiogenic genes were unique compared to the increased MHCII⁺CD11C⁺ angiogenesis genes. Conclusions: Macrophage origin impacts heterogeneity at steady state and after laser injury in mice. Both mice and human patients demonstrate similar macrophage subtypes. Two discrete pro-angiogenic macrophage populations exist in the human choroid. Targeting specific, pro-angiogenic macrophage subsets is a potential novel therapeutic for nAMD.