SUMMARY Despite ocular immune privilege, circulating retina-specific T cells can trigger autoimmune uveitis, yet intraocular bleeding—a relatively common event—rarely leads to disease. Using an in vivo immune privilege model, we previously reported that all naive retina-specific T cells entering the eye become primed in situ; about 30% become Foxp3+ regulatory T cells (Tregs), while the rest fail to induce pathology. Here, single-cell transcriptomics and functional validation revealed distinct phenotypes in both populations: ocular Tregs were highly suppressive, whereas non-Tregs expressed suppression- and anergy-associated genes and lacked regulatory function. Trajectory analyses suggested that Tregs and anergic cells arise from a common proliferative precursor in parallel, rather than sequentially. Our data indicate a key checkpoint governing the divergence of anergic and regulatory fates. These findings provide molecular-level insights into ocular immune privilege and may inform strategies to silence autoimmune effector cells or reverse T cell unresponsiveness in cancer, vaccination, or chronic infection. pmc Cell Rep Cell Rep 319 nihpa Cell reports 2211-1247 pmc-is-collection-domain yes pmc-collection-title NIHPA Author Manuscripts PMC12307573 PMC12307573.1 12307573 12307573 NIHMS2092651 40450686 10.1016/j.celrep.2025.115780 NIHMS2092651 NIHPA2092651 1 Article Ocular immune privilege in action: The living eye imposes unique regulatory and anergic gene signatures on uveitogenic T cells Peng Zixuan 1 2 Nagarajan Vijayaraj 1 Horai Reiko 1 Jittayasothorn Yingyos 1 Mattapallil Mary J. 1 * Caspi Rachel R. 1 , 2 In addition to the physical blood-tissue barriers that separate the eye from the immune system, multiple studies have described that the intraocular environment, composed of ocular fluids and ocular resident cells, is immunosuppressive and can inhibit the activity of immunocompetent cells. 3 – 5 Aqueous humor has been shown to reduce proinflammatory cytokine production by T cells in culture and to promote the induction of regulatory T cells (Tregs). 6 , 7 Soluble factors involved in these processes include transforming growth factor β (TGF-β), α-melanocyte-stimulating hormone, vasoactive intestinal peptide, retinoic acid (RA), and others. 18 To address the question of how the eye maintains immune homeostasis, we established a mouse model in which naive retina-specific T cells, 19 , 20 capable of inducing autoimmune uveitis, are injected into the eyes of healthy wild-type (WT) mice. 21 The injected cells express a transgenic T cell receptor (TCR) specific for the retinal Ag, interphotoreceptor retinoid-binding protein (IRBP). This model thus exposes the eye to naive but non-tolerant T cells, as would occur in the case of intraocular bleeding, whether major, as a result of trauma, or minor, as a result of vascular abnormalities (e.g., macular degeneration, diabetic retinopathy, neovascular glaucoma). Tregs and naive T cells from peripheral LTs of Tcra −/− R161H Foxp3 GFP transgenic mice were used for comparison. Proliferation was measured by [ 3 H]-thymidine incorporation in a co-culture system with Ag-presenting cells (APCs) and IRBP peptide as the cognate Ag ( Figure 4A ). Eye-induced nfc cells Limitations of the study While the in vivo model of immune privilege is a powerful tool to dissect eye-specific control of immune cell differentiation, the system also has limitations, both objective and subjective. The level of complexity of an in vivo system precludes analysis of the individual contributions of the array of signals from multiple components, which integrate to produce the final phenotypic and molecular events. In part, this could be addressed by including various ocular resident cells in the analysis. Additionally, RNA-seq performed at multiple time points could provide further insight into the kinetics of the differentiation process that could have strengthened our conclusions from the trajectory analysis. However, technical and logistic aspects inherent to this experimental model precluded addressing this more comprehensively in the present study. Conclusion Our findings shed new light on the concepts of ocular immune privilege and the molecular mechanisms that actively maintain immunological homeostasis. All animals were maintained under specific-pathogen-free conditions at National Eye Institute (NEI) animal facility on standard chow and water ad libitum . All animal procedures were approved by the Animal Care and Use Committee (ACUC) of NEI/NIH. METHOD DETAILS Ocular immune privilege model The ocular privilege model was established and described in our previous study, 21 except that we used R161H TCR Tg Foxp3 GFP reporter mice on Tcra −/− background as donor mice, rather than on the Rag2 −/− background. WT CD90.1-congenic recipient mice were injected intravitreally with 500,000 of these naive T cells in 1.5 μL PBS into each eye, using a 33G needle and Hamilton syringe. The cells were retrieved from donor eyes 7–8 days later and prepared for analysis, as described ahead. Flow cytometry and cell sorting Single-cell suspensions from spleens and lymph nodes (submandibular, axillary, inguinal, and mesenteric lymph nodes) collected from Tcra −/− R161H Foxp3 GFP CD90.2 donor mice were used for isolation of retina-specific T cells. 46 Antigen-specific proliferation assay Ocular immune privilege model was conducted as described above, after one week, Foxp3 GFP+ or Foxp3 GFP− CD4 + CD90.2 + CD90.1 − cells were sorted out from the recipients’ eyes (50,000 cells/well) and co-cultured with human IRBP 161–180 peptide (50 ng/mL) and CD11c + dendritic cells (at a 1:5 ratio to T cells), with or without 100 IU/mL recombinant human IL-2. Dendritic cells were obtained by digesting spleens from WT CD90.1 mice in spleen dissociation medium (Stem Cell) for 30 min, followed by ammonium-chloride-potassium (ACK) lysis and CD11c + enrichment using Micro Beads (Miltenyi Biotec).