The Microbiota Determines Susceptibility to Experimental Autoimmune Uveoretinitis

Abstract

The microbiota is a crucial modulator of the immune system. Here, we evaluated how its absence or reduction modifies the inflammatory response in the murine model of experimental autoimmune uveoretinitis (EAU). We induced EAU in germ-free (GF) or conventionally housed (CV) mice and in CV mice treated with a combination of broad-spectrum antibiotics either from the day of EAU induction or from one week prior to induction of disease. The severity of the inflammation was assessed by fundus biomicroscopy or by histology, including immunohistology. The immunophenotyping of T cells in local and distant lymph nodes was performed by flow cytometry. We found that GF mice and mice where the microbiota was reduced one week before EAU induction were protected from severe autoimmune inflammation. GF mice had lower numbers of infiltrating macrophages and significantly less T cell infiltration in the retina than CV mice with EAU. GF mice also had reduced numbers of IFN-γ and IL-17-producing T cells and increased numbers of regulatory T cells in the eye-draining lymph nodes. These data suggest that the presence of microbiota during autoantigen recognition regulates the inflammatory response by influencing the adaptive immune response.

Figure 1

Severity of ocular inflammation in germ-free (GF) and conventional (CV) mice 35 days after EAU induction. (a) Representative photographs of retinal fundus of GF (grade 0) and CV (grade 2) mouse. The figure shows small and linear lesions (arrow), minimal optic disc inflammation, and moderate vascular cuffing (arrowheads). (b) Quantification of the clinical EAU score. By clinical fundoscopy, no inflammation was observed in the GF mice (3 animals) at day 35 after induction, whereas in control CV mice (6 animals), severe inflammation was observed. The red line in the graphs represents mean. ^∗∗∗ p < 0.001 (Mann-Whitney test). (c) Representative microphotographs of hematoxylin and eosin-stained retina of GF and CV mice. The figure shows a large infiltrate (star) located in inner retinal layer, mild vitritis (arrow), and small retinal folds (arrowhead). (d) Quantification of histological EAU score. On histological evaluation, minimal to no signs of uveitis were observed in GF mice (6 animals) compared to severe uveitis in CV mice (6 animals). The red line in the graphs represents mean. ^∗∗∗ p < 0.001 (Mann-Whitney test).

Figure 2

Reduced severity of EAU in mice treated with antibiotics (ATB) from one week before EAU induction (15 animals). Mice treated with metronidazole and ciprofloxacin commencing one week prior to EAU induction and continued for the course of the experiment had significantly lower levels of EAU compared to controls (15 animals) both clinically and histologically. (a) Representative photographs of retinal fundus at days 14, 21, 28 and day 35 after EAU induction show the development of ocular pathology in ATB and control mice. (b) Quantification of clinical EAU score. The red lines in the graphs represent mean. ^∗ p < 0.05 (Mann-Whitney test). (c) Representative microphotographs of hematoxylin and eosin-stained retina of ATB-treated and control mice at day 35 after induction. Fewer signs of inflammation are present in ATB-treated compared to control mice, including cells in the vitreous (arrows) and small retinal folds (arrowheads), retinal neovascularization (black star), and vasculitis (white star). (d) Quantification of histological EAU score is shown. The red lines in the graphs represent mean. ^∗ p < 0.05 (Mann-Whitney test).

Figure 3

Antibiotics (ATB) administered from the day of EAU induction do not reduce the EAU severity. Quantification of (a) clinical and (b) histological EAU score at day 35 after induction is shown. The clinical data are from one of several independent experiments (ATB 11 mice, water 15 mice). The red lines in the graphs represent mean.

Figure 4

T cell and macrophage infiltration of the retina from germ-free (GF) mice (a, b), from mice treated with antibiotics (ATB) administered from one week (c, d), or from the day of EAU induction (e, f) and conventional (CV, water) mice 35 days after EAU induction. (a, c, e) CD3⁺ cells (T-lymphocytes) are shown both distributed as single cells in inner and outer retinal layers (arrows), in the vitreous (arrowheads) and concentrated as clumps in granulomas (stars). (b, d, f) F4/80⁺ cells (macrophages) are either present as single cells, in inner retinal layers (arrows), or accumulated in the periphery of granulomas (stars). Each point in the graphs shows the sum of all positive cells counted in two sections, one from periphery and one from the centre, from one randomly selected eye. The red lines in the graphs represent mean. ^∗ p < 0.05 (Mann-Whitney test).

Figure 5

Flow cytometric analysis of lymphocyte populations in cervical and mesenteric lymph nodes. In the cervical lymph nodes of conventional (CV) mice, (a) the percentage of IFN-γ and IL-17-producing CD4⁺ T cells increased and (c) the percentage of regulatory Foxp3-expressing CD4⁺ T cells decreased compared to germ-free (GF) mice. In the mesenteric lymph nodes of CV mice, the environment was less proinflammatory showing only a small but significant increase of IFN-g-producing CD4⁺ T cells compared to GF mice. In both cervical and mesenteric lymph nodes, (b) the percentage of TNF-α-producing CD4⁺ T cells remained unchanged. The dot plots are representative of two independent experiments. The column graphs summarize the frequency of lymphocyte subpopulations in (d) the cervical and (e) the mesenteric lymph nodes. Each graph represents data from two independent experiments. ^∗ p < 0.05, ^∗∗ p < 0.01 (Mann-Whitney test). n.s. = not significant.