Glia Modulates Immune Responses in the Retina Through Distinct MHC Pathways

Abstract

Glia antigen-presenting cells (APCs) are pivotal regulators of immune surveillance within the retina, maintaining tissue homeostasis and promptly responding to insults. However, the intricate mechanisms underlying their local coordination and activation remain unclear. Our study integrates an animal model of retinal injury, retrospective analysis of human retinas, and in vitro experiments to gain insights into the crucial role of antigen presentation in neuroimmunology during retinal degeneration (RD), uncovering the involvement of various glial cells, notably Müller glia and microglia. Glial cells act as sentinels, detecting antigens released during degeneration and interacting with T-cells via MHC molecules, which are essential for immune responses. Microglia function as APCs via the MHC Class II pathway, upregulating key molecules such as Csf1r and cytokines. In contrast, Müller cells act through the MHC Class I pathway, exhibiting upregulated antigen processing genes and promoting a CD8⁺ T-cell response. Distinct cytokine signaling pathways, including TNF-α and IFN Type I, contribute to the immune balance. Human retinal specimens corroborate these findings, demonstrating glial activation and MHC expression correlating with degenerative changes. In vitro assays also confirmed differential T-cell migration responses to activated microglia and Müller cells, highlighting their role in shaping the immune milieu within the retina. In summary, our study emphasizes the involvement of retinal glial cells in modulating the immune response after insults to the retinal parenchyma. Unraveling the intricacies of glia-mediated antigen presentation in RD is essential for developing precise therapeutic interventions for retinal pathologies.

Keywords: Müller cells; T‐cells; antigen‐presenting cells; microglia; retina.

FIGURE 1

Retinal injury triggers microglial MHC II expression and CD4⁺ T‐cell signaling upregulation. (a–d) Analysis of MHC expression by microglia after injury using a microglia marker (Iba1) and MHC Class I (MHC I) or Class II (MHC II) marker. (a) Representative sections stained for Iba1 (green) and MHC I (red). Scale bars equals 100 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (b) Quantification of the MHC I⁺Iba1⁺ area on total Iba1⁺ area per lesion before injury (baseline) and at pre‐defined time points (Days 1, 3, and 7). Significant differences between baseline and the different time points were determined by using a post hoc Bonferroni one‐way ANOVA test (n = 5). (c) Representative sections stained for Iba1 (green) and MHC II (red). Scale bars equals 100 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (d) Quantification of the percentages of the MHC II⁺Iba1⁺ area on total Iba1⁺ area per lesion before injury (baseline) and at pre‐defined time points (Days 1, 3, and 7). Significant differences (*p < 0.1 and ***p < 0.001) between baseline and the different time points were determined by using a post hoc Bonferroni one‐way ANOVA test (n = 5). (e) Heatmaps of differentially expressed APC‐related genes in Csfr1^EGFP cells, represented as z‐scores. Genes were selected from KEGG pathways (mmu04612).

FIGURE 2

Differential gene expression profiles of CSF receptors, interferon receptor–ligand genes, and TNF pathway genes in microglia after laser injury. (a) Heatmaps of differentially expressed genes that encode receptors for CSFs in Csfr1^EGFP (microglia), represented as z‐scores. Data are expressed as fold‐changes between different time points (Days 1, 3, and 7) compared with negative controls (Csfr1^EGFP cells from uninjured retinas). (b) Heatmaps of differentially interferon receptor–ligand genes in Csfr1^EGFP (microglia), represented as z‐scores. Data are expressed as fold‐changes between different time points (Days 1, 3, and 7) compared with negative controls (Csfr1^EGFP cells from uninjured retinas). (c) Heatmaps of differentially expressed genes of the TNF signaling pathway in Csfr1^EGFP (microglia), represented as z‐scores. Genes were selected from KEGG pathways (mmu04668).

FIGURE 3

Retinal injury induces the expression of MHC Class I in Müller glia and upregulates CD8 T‐cell signaling. (a–d) Analysis of MHC proteins expression by Müller cells after injury using a Müller cell marker (GS) and MHC Class I (MHC I) or Class II marker (MHC II). (a) Representative sections stained for GS (green) and MHC I (red). Scale bars equals 100 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (b) Quantification of the percentages of the MHC I + GS+ area on total GS+ area per lesion before injury (baseline) and at pre‐defined time points (Days 1, 3, and 7). Significant differences (****p < 0.0001) between baseline and the different time points were determined by using a post hoc Bonferroni one‐way ANOVA test (n = 5). (c) Representative sections stained for GS (green) and MHC II (red). Scale bars equals 100 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (d) Quantification of the percentages of the MHC II + GS+ area on total GS+ area per lesion before injury (baseline) and at pre‐defined time points (Days 1, 3, and 7). Significant differences between baseline and the different time points were determined by using a post hoc Bonferroni one‐way ANOVA test (n = 5). (e) Heatmaps of differentially expressed APC‐related genes in Rlbp1GFP cells, represented as z‐scores. Genes were selected from KEGG pathways (mmu04612).

FIGURE 4

Differential gene expression profiles of CSF receptors, interferon receptor–ligand genes, and TNF pathway genes in Müller glia after laser injury. (a) Heatmaps of differentially expressed genes that encode receptors for CSFs in Rlbp1^GFP cells (Müller glia), represented as z‐scores. Data are expressed as fold‐changes between different time points (Days 1, 3, and 7) compared with negative controls (Rlbp1^GFP cells from uninjured retinas). (b) Heatmaps of differentially interferon receptor‐ligand genes in Rlbp1^GFP cells (Müller glia), represented as z‐scores. Data are expressed as fold‐changes between different time points (Days 1, 3, and 7) compared with negative controls (Rlbp1^GFP cells from uninjured retinas). (c) Heatmaps of differentially expressed genes of the TNF signaling pathway in Rlbp1^GFP cells (Müller glia), represented as z‐scores. Genes were selected from KEGG pathways (mmu04668).

FIGURE 5

Pathological examination of ocular tissue from human donors' eyes. (a–d) H&E staining of human retinas showing presence of drusen and its integrity. (a) Representative detailed view of healthy cuboidal RPE (CTRL) and retinas presenting drusen (RD). Scale bars equals 50 μm. (b) Quantification of drusen size, categorized as either ≤ 25 μm or > 25 μm in width. Significant differences (****p < 0.0001) between CTRL and RD groups were determined by using two‐tailed Mann–Whitney test analysis (n = 15). (c) Representative detailed view of healthy retinal structure (CTRL) and atrophic retinas (RD). Scale bars equals 50 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (d) Quantification of cell nuclei per retina. Significant differences (****p < 0.0001) between CTRL and RD groups were determined by using two‐tailed Mann–Whitney test analysis (n = 15). (e) Representative sections stained with Picro Sirius red staining for histological visualization of collagen fibers and retinal fibrosis. Scale bars equals 50 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer.

FIGURE 6

Glial responses correlate MHCs in human retina during degeneration. (a, b) Analysis of pro‐inflammatory phenotype in microglia in healthy (CTRL) and degenerated retinas (RD) using a microglia marker (IBA1) and a pro‐inflammatory marker (iNOS). (a) Representative sections stained for Iba1 (green) and iNOS (red). Scale bars equals 100 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (b) Quantification of the percentages of iNOS⁺IBA1⁺ area on total IBA1⁺ area in CTRL and RD retinas. Significant differences (****p < 0.0001) between CTRL and RD were determined by using two‐tailed Mann–Whitney test analysis (n = 15). (c, d) Analysis of Müller cell reactivity in healthy (CTRL) and degenerated retinas (RD) using a Müller cell marker (GS) and a reactivity marker (GFAP). (c) Representative sections stained for GS (green) and GFAP (red). Scale bars equals 100 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (d) Quantification of the percentages of the GFAP⁺GS⁺ area on total GS⁺ area in CTRL and RD retinas. Significant differences (****p < 0.0001) between CTRL and RD were determined by using two‐tailed Mann–Whitney test analysis (n = 15). (e–h) Analysis of MHC proteins by glia in degenerated retinas. (e) Representative sections stained for Iba1 (green) and MHC I or II (red). Scale bars equals 100 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (f) Quantification of the percentages of the MHC I⁺ IBA1⁺ and MHC II⁺ IBA1⁺ area on total IBA1⁺ area in CTRL and RD retinas. Significant differences (****p < 0.0001) between CTRL and RD were determined by using two‐tailed Mann–Whitney test analysis (n = 15). (g) Representative sections stained for GS (green) and MHC I or II (red). Scale bars equals 100 μm. GCL, ganglion cells layer; INL, inner nuclear layer; ONL, outer nuclear layer. (h) Quantification of the percentages of the MHC I⁺GS⁺ and MHC II⁺GS⁺ area on total GS⁺ area in degenerated retinas. Significant differences (****p < 0.0001) between CTRL and RD were determined by using two‐tailed Mann–Whitney test analysis (n = 15). (i) Spearman correlation between MHC II⁺ IBA1⁺ cells with iNOS⁺ IBA1⁺ cells in the degenerated retinas. (j) Spearman correlation between MHC I⁺GS⁺ cells with GFAP⁺GS⁺ cells in the degenerated retinas.

FIGURE 7

Reciprocal interplays between glia and T‐cells in vitro. (a) Quantification of the MHC I⁺ and MHC II⁺ HMC‐3 cells untreated and LPS treated. Significant differences (**p < 0.1) between untreated and LPS‐treated cells were determined using two‐tailed Mann–Whitney test analysis (n = 5). (b) Mean ± SD of the percentages of CD4⁺CD3⁺ and CD8⁺CD3⁺ cells migrated toward microglia. Significant differences (*p < 0.1) between CD4⁺CD3⁺ and CD8⁺CD3⁺ cells migrated were determined using two‐tailed Mann–Whitney test analysis (n = 3). (c) Quantification of the MHC I⁺ and MHC II⁺ MIO‐M1 cells untreated and LPS treated. Significant differences (**p < 0.1) between untreated and LPS‐treated cells were determined using two‐tailed Mann–Whitney test analysis (n = 3). (d) Mean ± SD of the percentages of CD4⁺CD3⁺ and CD8⁺CD3⁺ cells migrated toward Müller cells. Significant differences (*p < 0.1) between CD4⁺CD3⁺ and CD8⁺CD3⁺ cells migrated were determined using two‐tailed Mann–Whitney test analysis (n = 3). (e) Quantification of the iNOS⁺ HMC‐3 cells untreated, treated with LPS, T‐cell media or co‐cultured with T‐cells. Significant differences (**p < 0.01 and ****p < 0.0001) between conditions were determined by using a post hoc Bonferroni one‐way ANOVA test (n = 5). (f) Quantification of the GFAP⁺ MIO‐M1 cells untreated, treated with T‐cell media or co‐cultured with T‐cells. Significant differences (*p < 0.1 and **p < 0.001) between conditions were determined by using a post hoc Bonferroni one‐way ANOVA test (n = 5).