Autoreactive memory CD4+ T lymphocytes that mediate chronic uveitis reside in the bone marrow through STAT3-dependent mechanisms

Abstract

Organ-specific autoimmune diseases are usually characterized by repeated cycles of remission and recurrent inflammation. However, where the autoreactive memory T cells reside in between episodes of recurrent inflammation is largely unknown. In this study, we have established a mouse model of chronic uveitis characterized by progressive photoreceptor cell loss, retinal degeneration, focal retinitis, retinal vasculitis, multifocal choroiditis, and choroidal neovascularization, providing for the first time to our knowledge a useful model for studying long-term pathological consequences of chronic inflammation of the neuroretina. We show that several months after inception of acute uveitis, autoreactive memory T cells specific to retinal autoantigen, interphotoreceptor retinoid-binding protein (IRBP), relocated to bone marrow (BM). The IRBP-specific memory T cells (IL-7Rα(High)Ly6C(High)CD4(+)) resided in BM in resting state but upon restimulation converted to IL-17/IFN-γ-expressing effectors (IL-7Rα(Low)Ly6C(Low)CD4(+)) that mediated uveitis. We further show that T cells from STAT3-deficient (CD4-STAT3KO) mice are defective in α4β1 and osteopontin expression, defects that correlated with inability of IRBP-specific memory CD4-STAT3KO T cells to traffic into BM. We adoptively transferred uveitis to naive mice using BM cells from wild-type mice with chronic uveitis but not BM cells from CD4-STAT3KO, providing direct evidence that memory T cells that mediate uveitis reside in BM and that STAT3-dependent mechanism may be required for migration into and retention of memory T cells in BM. Identifying BM as a survival niche for T cells that cause uveitis suggests that BM stromal cells that provide survival signals to autoreactive memory T cells and STAT3-dependent mechanisms that mediate their relocation into BM are attractive therapeutic targets that can be exploited to selectively deplete memory T cells that drive chronic inflammation.

Figure 1

Chronic intraocular inflammation induces retinal degeneration and neovascularization. (A) Fundus images were taken from normal or EAU mice using an otoendoscopic imaging system and assessment of severity of the inflammatory disease was based on changes at the optic nerve disc and retinal vessels or tissues (top panels). Left panels: show control retina with well-circumscribed optic disc and normal retinal vessels and histology revealed intact retinal architecture (bottom panels). Middle panels: Fundus image of retina 21 days after EAU induction (d-21 EAU) showing severe inflammation with blurred optic disc margins and enlarged juxtapapillary area (black arrowhead), retinal vasculitis with moderate cuffing (black arrows) and yellow-whitish retinal and choroidal infiltrates (white arrows). Histological analysis reveals substantial numbers of inflammatory cells in the vitreous, photoreceptor cell damage (white asterisk), granuloma (white arrowhead), choroiditis (black arrowhead), and retinal edema (the thickened retina) (bottom panels). Right panels: 92 days after EAU induction (d-92 EAU) fundus image reveal severe vasculitis with cuffing (black arrows) with part of the vessel segment not visible. Retinal structural damage was observed, including evidence of atrophic retina (thinning) and sclerotic vessel (red arrow) with multiple whitish infiltrates (white arrow) and brownish chorioretinal scars (blue arrows). Histology revealed photoreceptor cell loss (red asterisk), retinal vasculitis (black arrows), retinal sclerotic vessel (white arrow), choroiditis (black arrowhead), and retinal degeneration (bottom panels). OpN, optic nerve; V, vitreous; R, retina; GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; RPE, retinal pigment epithelial layer; CH, choroid. (B) Layered structure of the retina was visualized by Spectral Domain Optical Coherence Tomography (SD-OCT) and used to follow the evolution of pathology induced by chronic uveitis. Change in the thickness of retina (white bar) or RPE-Choroid (RPE-CH) layer (black bar) was used to quantify severity of the retinal degeneration. Oval circle highlights hypo-dense area indicative of retinal edema while speckles in the vitreous represent clusters of residual inflammatory cells. (C) Quantitative analysis of inflammatory cells in the day-21- or day-92-mouse retina. Numbers in circle indicate percent of CD4⁺ T cells and result is representative of 3 independent experiments.

Figure 2

Prolonged intraocular inflammation in neuroretina caused vision loss and blindness. B10.A mice were immunized with CFA alone or IRBP/CFA emulsion and electroretinogram (ERG) recordings were obtained and analyzed 185 or 225 days post-immunization. ERGs of control unimmunized mice or control mice immunized with CFA alone, reveal the characteristic normal a-, b- and cone waves (A). In contrast, light-induced response was barely elicited from retina of IRBP-induced EAU mice (A). Immediately after ERG recordings, imaging of the fundus was performed as previously described (B).

Figure 3

Analysis of CD4⁺ T cells in the BM, spleen, lymph nodes, blood and retina of control and d-92 EAU mice. EAU was induced in C57BL6 mice. (a) Cells from the BM, spleen, blood, LN or retina were isolated 92 days post-immunization with IRBP/CFA (EAU) or CFA alone (control), re-stimulated in vitro with IRBP and then subjected to FACS analysis. Numbers in quadrants indicate percent of CD4⁺ T cells in the various tissues. (b) Graphical representation of the absolute numbers of CD4+ T cells in the various tissues. Results are representative of 3 independent experiments.

Figure 4

Autoreactive CD4⁺ T-cells that mediate uveitis are maintained in resting non-proliferative state in the bone marrow. EAU was induced in mice and primary BM, spleen, blood, LN or retina cells from EAU day-92 (a–f) or d-21, d-50 and d-75 (g, h) mice were re-stimulated in vitro with IRBP (20µg/ml). (a) Cells were gated on CD4 and the percentages of cells responding to IRBP stimulation were detected by FACS analysis of CD154 cell-surface expression. (b) Graphical representation of the percentage of the IRBP-responsive CD4⁺ T cells. (c) FACS analysis of the relative percentage of IRBP-specific T cells expressing the memory T cell marker, IL-7Rα. Number in quadrants present the percentage of CD4⁺ T cells. Graphical representation of the absolute numbers of IL-7Rα⁺ cells (d) or IRBP-responsive CD4⁺ memory T cells (e). Cells from BM of day-92 EAU mice were sorted into IL-7Rα^low or IL-7rα^high populations on a cell sorter and were then re-stimulated in vitro with IRBP (f). Conversion of IRBP-specific, IL-7Rα^High memory T cells to IL-7Rα^Low effectors in response to re-stimulation with IRBP was assessed by FACS analysis of IL-7Rα and CD154 expression. (g) Analysis of absolute numbers of CD4⁺ T cells responding to IRBP-stimulation in the BM of mice at various time points post-immunization. (h) BM cells from day-75 EAU mice were analyzed for expression of the memory T-cell markers, IL-7Rα and Ly6C. Numbers in quadrants indicate percent of T-cells expressing CD154, IL-7Ra, Ly6C or CD4. Results are representative of 3 independent experiments.

Figure 5

BM cells from d-92 EAU mice express IL-17 and IFN-γ and transferred EAU to naïve mice through STAT3-dependent mechanisms. (A) BM cells from control or d-92 EAU mice were stimulated with IRBP for 12 h and then assayed for intracellular cytokine expression. Numbers in quadrants indicate percent CD154-, IFN-γ̃ or IL-17-expressing CD4⁺ T-cells. (B) IRBP-stimulated BM cells from control or d-92 EAU mice were transferred (1×10⁷ cells/mouse) into naïve syngeneic C57BL/6 recipient mice. Ten days after adoptive cell transfer, fundoscopy and histology reveal the development of retinal vasculitis (black arrows) and inflammatory infiltrates (white arrows). (C) WT or CD4-STAT3KO mice were immunized with IRBP in CFA and fundus images were taken on day-92 post-immunization. (D) BM cells were re-stimulated with IRBP and percentage of IRBP-specific CD4⁺ T-cells (CD154⁺) was quantified by Flow cytometry. (E) BM cells from day-92 p.i. mice were re-stimulated in vitro with IRBP and IL-7Rα- and CD154-expressing CD4+ T-cells were detected by FACS analysis.

Figure 6

Recruitment and maintenance of IRBP-specific memory T cells in BM require STAT3. (A) BM cells from control or d-92 EAU mice were stimulated with IRBP for 12 h and then assayed for intracellular cytokine expression. Numbers in quadrants indicate percent IL-2, IFN-γ̃ or IL-17-expressing CD4⁺ T-cells. (B) IRBP-stimulated BM cells from IRBP/CFA-immunized WT or CD4-STAT3KO mice were transferred (1×10⁷ cells/mouse) into naïve WT or CD4-STAT3KO mice recipient mice. Ten days after adoptive cell transfer, disease was assessed by fundoscopy. (C) WT and CD4-STAT3KO mice were immunized with IRBP and after 7 days T cells were isolated from peripheral blood and expression of integrins and osteopontin was analyzed by RT-PCR. (D, E) (d, e) EAU was induced in B10A mice (D) or C57BL6 (E) and BM or spleen cells were isolated from control or d-200 p.i. mice. IL-7Rα-, β1 integrin or α4 integrin integrin-expressing CD4⁺ T-cells were detected by FACS analysis. Results are representative of 3 independent experiments.