Abstract
IFN regulatory factor 8 (IRF8) is constitutively expressed in monocytes and B cells and plays a critical role in the functional maturation of microglia cells. It is induced in T cells following Ag stimulation, but its functions are less well understood. However, recent studies in mice with T cell–specific Irf8 disruption under direction of the Lck promoter (LCK-IRF8KO) suggest that IRF8 directs a silencing program for Th17 differentiation, and IL-17 production is markedly increased in IRF8-deficient T cells. Paradoxically, loss of IRF8 in T cells has no effect on the development or severity of experimental autoimmune encephalomyelitis (EAE), although exacerbating colitis in a mouse colitis model. In contrast, mice with a macrophage/microglia-specific Irf8 disruption are resistant to EAE, further confounding our understanding of the roles of IRF8 in host immunity and autoimmunity. To clarify the role of IRF8 in autoimmune diseases, we have generated two mouse strains with targeted deletion of Irf8 in retinal cells, including microglial cells and a third mouse strain with targeted Irf8 deletion in T cells under direction of the nonpromiscuous, CD4 promoter (CD4-IRF8KO). In contrast to the report that IRF8 deletion in T cells has no effect on EAE, experimental autoimmune uveitis is exacerbated in CD4-IRF8KO mice and disease enhancement correlates with significant expansion of Th17 cells and a reduction in T regulatory cells. In contrast to CD4-IRF8KO mice, Irf8 deletion in retinal cells confers protection from uveitis, underscoring divergent and tissue-specific roles of IRF8 in host immunity. These results raise a cautionary note in the context of therapeutic targeting of IRF8.
Introduction
The nine-member IFN regulatory factor (IRF) family of transcription factors is characterized by an N-terminal DNA-binding domain that mediates binding to the core DNA sequence (GAAA) in the IFN-stimulated response element and a C-terminal IRF-association domain, which facilitates heterodimerization with other members of the IRF family as well as ETS family members (1, 2). IRF8, also known as IFN consensus sequence-binding protein, is expressed almost exclusively in hematopoietic cells of both the myeloid and lymphoid lineages (3). It functions to a large extent by forming complexes with IRF1, IRF2, IRF4, or PU.1, and, depending on its interaction partners, IRF8 can act as a transcriptional repressor or activator. In the lymphocyte lineage, it is constitutively expressed by B cells and regulates B cell lineage specification, commitment, and differentiation (4). In contrast, IRF8 is not expressed in naive T cells, but is rapidly induced in response to antigenic stimulation or TCR activation (5), suggesting that it may be required for regulating genes involved in T cell differentiation and/or effector functions. In fact, IRF8 inhibits the Th17 cell differentiation program through its interaction with the Th17 master transcription factor, ROR-γt (6), and integrates TCR and cytokine signaling pathways that drive differentiation of effector CD8+ T cells (7).
IRF8 is also expressed by APCs, including microglia in the CNS, macrophages, and dendritic cells, and plays crucial roles in myeloid cell differentiation, macrophage activation, and functional maturation of CNS microglia (8–10). Furthermore, IRF8 regulates innate immune responses and influences the differentiation of Th cells by modulating transcription of cytokine genes in APCs, including Il12a, which codes the cytokine IL-12p35, a common subunit shared by IL-12 and IL-35. In collaboration with IRF1, IRF8 also activates transcription of Il27p28 and contributes to mechanisms of ocular immune privilege by inducing retinal microglial cells and neurons to express IL-27 and complement factor H (11–13). It is of note that increased expression of the immunosuppressive cytokines, IL-27 and IL-35, in the retina or brain mitigates experimental autoimmune uveitis (EAU) and experimental autoimmune encephalomyelitis (EAE), animal models of uveitis, and multiple sclerosis, respectively (13–16).
Two recent studies have examined the contributions of IRF8 to colitis and encephalitis. Mice with a global Irf8 knockout or T cell–specific deletion of the Irf8 gene (LCK-IRF8KO) developed a more severe inflammation of the colon resulting from enhanced expansion of Th17 cells (6). In the other report, EAE clinical scores were found to be similar between WT and LCK-IRF8KO mice, suggesting that the expression of IRF8 by T cells does not have a consequential role in EAE (17). In this study, we used CD4-Cre mice to generate mice with targeted deletion of Irf8 in T cells to rule out the possibility that different outcomes observed in the colitis and EAE models did not derive in part from use of the relatively “leaky” Lck-Cre mice for generating mice with Irf8 deletion in the T cell compartment. We also generated two mouse strains with targeted deletion of Irf8 in retinal neurons and microglia. We have used these strains to clarify the involvement of IRF8 in autoimmune disease and to investigate whether IRF8 is a potential therapeutic target in uveitis and other CNS autoimmune diseases.
Materials and Methods
Mice
We derived mice with conditional deletion of Irf8 in CD4+ T cells (CD4-IRF8KO) or neurons (αCre-IRF8KO or RX-IRF8KO) by breeding Irf8fl/fl mice with CD4-Cre (Taconic, Hudson, NY) mice or mice expressing the Cre-recombinase under the direction of a retina-specific promoter. For targeted deletion of Irf8 in the neuroretina, we bred the Irf8fl/fl mouse strain with either α-Cre transgenic mice (provided by Dr. P. Gruss, Max-Planck-Institute of Biophysical Chemistry, Gottingen, Germany), which expresses Cre-recombinase only in the retina (αCre-IRF8KO), or RX-Cre transgenic mice (provided by A. Swaroop, National Eye Institute, National Institutes of Health, Bethesda, MD), which expresses Cre-recombinase in the retina as well as the retinal pigmented epithelium (RX-IRF8KO). Littermate Irf8fl/fl mice on the C57BL/6J background were used as wild-type (WT) controls. Mice were maintained and used in accordance with National Eye Institute/National Institutes of Health Animal Care and Use Committee guidelines (ASP Protocol EY000262-19 and EY000372-14).
Induction of experimental autoimmune uveoretinitis
Mice were immunized with 150 μg bovine interphotoreceptor retinoid-binding protein (IRBP) and 300 μg human IRBP peptide (1–20) in 0.2 ml emulsion 1:1 v/v with CFA containing Mycobacterium tuberculosis strain H37Ra (2.5 mg/ml), as previously described (14). The human IRBP peptide (1–20) is a highly purified peptide purchased from Sigma-Aldrich (St. Louis, MO), and R. Caspi (National Institutes of Health) provided the bovine IRBP. Mice also received Bordetella pertussis toxin (0.2 μg/mouse) concurrent with immunization, and clinical disease was established by fundoscopy or histological analysis, as described previously (18). For each EAU experiment, 10 mice were used for the WT or IRF8KO strain. For FACS or intracellular cytokine staining, lymph node (LN) and spleen cells from ∼3 mice were pooled and analyzed. Fundoscopy and/or histology were used to monitor progression of pathology in the eyes at 14 or 21 d postimmunization. Briefly, following i.p. injection of ketamine (1.4 mg/mouse) and xylazine (0.12 mg/mouse), pupils were dilated by topical administration of 1% tropicamide ophthalmic solution (Alcon, Fort Worth, TX). To avoid a subjective bias, evaluation of the fundus photographs was conducted without knowledge of the mouse identity by a masked observer. At least six images (two posterior central retinal view; four peripheral retinal views) were taken from each eye by positioning the endoscope and viewing from superior, inferior, lateral, and medial fields. Each individual lesion was identified, mapped, and recorded. The clinical grading system for retinal inflammation by fundoscopy was based on changes at the retina, optic nerve disc, and choroid, as previously established (19). For histology, eyes were harvested and fixed in 10% buffered formalin, and specimens were dehydrated through graded alcohols and embedded in paraffin. Serial vertical sections through the papillary-optic nerve plane were cut and stained with H&E, as described (18). Clinical scores and assessments of disease severity were based on pathological changes at the optic nerve disc and retinal tissues.
Electroretinogram recordings
Before the electroretinogram (ERG) recordings, mice were dark-adapted overnight and experiments were performed under dim red illumination. Mice were anesthetized with a single i.p. injection of ketamine (1.4 mg/mouse) and xylazine (0.12 mg/mouse), and pupils were dilated with Midrin P containing 0.5% tropicamide and 0.5% phenylephrine hydrochloride (Santen Pharmaceutical, Osaka, Japan). ERGs were recorded using an electroretinography console (Espion E2; Diagnosys, Lowell, MA) that generated and controlled the light stimulus. Dark-adapted ERG was recorded with a single flash delivered in a Ganzfeld dome with intensity of 24–1 log cd s/m2 delivered in six steps. Light-adapted ERG was obtained with a 20 cd/m2 background, and light stimuli started at 0.3–30 cd s/m2 in five steps. Gonioscopic prism solution (Alcon Laboratories, Fort Worth, TX) was used to provide good electrical contact and to maintain corneal moisture. A reference electrode (gold wire) was placed in the mouth, and a ground electrode (s.c. stainless steel needle) was positioned at the base of the tail. Signals were differentially amplified and digitized at a rate of 1 kHz. Amplitudes of the major ERG components (a- and b-wave) were measured (Espion software; Diagnosys) using automated and manual methods. Immediately after ERG recording, imaging of the fundus was performed, as described above. For each study, ERG data are based on analyses of six mice (12 eyes) per group.
Analysis of CD4+ Th cells
Freshly isolated T cells or IRBP-stimulated T cells from the spleen, LN (cervical, axillary, inguinal), or retina were analyzed for expression of inflammatory proteins by FACS or an intracellular cytokine expression assay, as described (17). In some experiments, naive T cells were activated by plate-bound anti-CD3 (10 μg/ml) and soluble anti-CD28 Abs (3 μg/ml) in plates without exogenous cytokines. For intracellular cytokine detection, cells were restimulated for 5 h with PMA (20 ng/ml)/ionomycin (1 μM). Golgi-stop was added in the last hour, and intracellular cytokine staining was performed using BD Biosciences Cytofix/Cytoperm kit (BD Pharmingen, San Diego, CA). FACS analysis was performed on a BD Biosciences FACSCalibur using labeled mAbs and corresponding isotype control Abs (BD Pharmingen). For lymphocyte proliferation assays, cells were cultured for 4 d in quintuplet cultures containing IRBP and CFSE. CFSE dilution assays were performed using a commercially available CFSE Cell Proliferation Kit (Molecular Probes, Eugene, OR). Briefly, purified CD4 T cells were washed with labeling buffer (PBS/0.1% BSA) and stained with CFSE (5 μM) for 10 min at 37°C. The cells were then incubated in RPMI 1640 medium for 10 min and washed with CFSE labeling buffer (2×) to remove excess CFSE, and the labeled cells were counted and used for the CFSE dilution assay, as recommended by the manufacturer.
Quantitative and semiquantitative RT-PCR analyses
Total RNA was extracted using the TRIzol reagent, according to the procedures recommended by the manufacturer (Life Technologies, Gaithersburg, MD). All RNA samples were digested with RNase-free DNase I (Life Technologies) and purified by phenol/chloroform extractions. RNA integrity was verified by analysis of 18S and 28S rRNA expression on RNA gels. RNA (10 μg), SuperScript III Reverse Transcriptase (Life Technologies), and oligo(deoxythymidine)12–16 were used for first-strand synthesis, as previously described (20). First-strand synthesis containing each mRNA sample, but no reverse transcriptase, was performed to control for possible DNA contamination; failure to obtain RT-PCR products with any of the PCR amplimers confirmed the absence of DNA templates. All cDNA preparations used were suitable substrates for PCR amplification on the basis of efficient amplification of a β-actin sequence. Real-time PCR was performed on an ABI 7500 (Applied Biosystems), and PCR parameters were as recommended for the TaqMan Universal PCR kit (Applied Biosystems). Primers and probes were purchased from Applied Biosystems.
Statistical analysis
Statistical analyses were performed by independent two-tailed Student t test. Data are presented as mean + SD. Asterisks denote p value (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).
Results
Generation and characterization of CD4-IRF8KO mice
We generated mice with targeted deletion of Irf8 in the CD4+ T cell compartment (CD4-IRF8KO) to investigate the function of IRF8 in T cells during the organ-specific autoimmune disease, uveitis. PCR analysis of tail DNA of mice from the cross between CD4-Cre and Irf8fl/fl mouse strains on C57BL/6J background confirmed the generation of CD4-Cre/IRF8fl/fl double-positive mice (Fig. 1A). Use of the mouse CD4-Cre strain expressing the Cre recombinase under the CD4 promoter element (CD4-Cre strain) resulted in the targeted deletion of Irf8 in CD4+ T cells, as previously reported, but also in CD8+ T cells due to intrathymic deletion at the double-positive stage (21, 22). Naive LN and splenic CD4+ T cells, purified by sorting from CD4-IRF8KO or WT mice, were stimulated with anti-CD3/CD28 Abs for 4 d. Western blot analysis of the total cellular protein extract confirmed that the IRF8 protein was indeed absent, specifically in the CD4 T cell compartment of the CD4-IRF8KO mice (Fig. 1B). This result is consistent with previous studies showing that, whereas IRF8 is not constitutively expressed in naive T cells, expression is rapidly induced in WT CD4+ T cells following TCR activation (5). To further characterize the phenotype of the CD4-IRF8KO T cells, we examined whether the loss of IRF8 affected their proliferative capacity. Naive CD4+ T cells from WT or CD4-IRF8KO mice were stimulated with anti-CD3/CD28 Abs for 4 d under nonpolarizing condition. Analysis of thymidine incorporation indicated a significant increase in the proliferative response of T cells from CD4-IRF8KO mice as compared with cells from WT mice (Fig. 1C). We further found that a higher proportion of IRF8-deficient CD4+ T cells produced IL-2 than their WT counterparts (Fig. 1D), providing suggestive evidence that IRF8 may function in vivo to restrain T lymphocyte proliferation in response to Ag stimulation. Consistent with the observed effects of IRF8 on the proliferation of CD4+ T cells, we also found that CD4-IRF8KO T cells exhibited an enhanced activation phenotype, as indicated by increased proportion of cells expressing CD44 and CD25, the high-affinity IL-2R (Fig. 1E). The increased proliferation and activation phenotype exhibited by the CD4-IRF8KO T cells compared with the WT T cells were accompanied by an increase in the percentage of the activated CD4+ T cells undergoing apoptosis (Fig. 1F) and a corresponding elevation in the expression of proapoptotic genes (Fig. 1G). Consistent with findings of a recent study showing increased production of IL-17 in activated CD4+ T cells derived from LCK-IRF8KO mice (6), we observed a marked increase in cells expressing high intracellular levels of IL-17 in cultures of CD4-IRF8KO cells (Fig. 1H). We also stimulated CD4+ T cells from WT or CD4-IRF8KO mice with anti-CD3/CD28 Abs for 4 d under the Th17-polarizing conditions. The cells were then analyzed by an intracellular cytokine-staining assay. Interestingly, the increased proportion of IL-17+ cells was accompanied by a reduction in the frequency of Th17 cells expressing ROR-γt at high levels and an increased proportion of the IL-17–expressing CD4+ T cells that were ROR-γtlow (Fig. 1I). Taken together, these findings suggest that functions of IRF8 in T cells may include the regulation of proliferation and protection of T cells from activation-induced cell death.
Generation and characterization of IRF8 conditional KO mice. Irf8fl/fl mice were crossed with CD4-Cre mice to generate mice with deletion of IRF8 in CD4+ T cells (CD4-IRF8KO). The CD4-IRF8KO mice were identified by PCR analysis of mouse tail genomic DNA (A). Naive CD4+ T cells from LN and spleens of WT or F8 generation CD4-IRF8KO mice (after eight cycles of brother–sister mating) were stimulated with anti-CD3/CD28 Abs and analyzed by Western blotting (B) and thymidine incorporation assays (C). Data shown in (C) are mean ± SEM from five replicate cultures. *p < 0.05. The immunophenotype of the stimulated T cells was characterized by FACS and intracellular cytokine-staining assays. Numbers in quadrants indicate percentages of IL-2–expressing CD4+ T cells (D) or percentages of activated CD4+ T cells (E). Histogram presented in (D) and (E) represents percentages of IL-2–expressing cells or CD44highCD25+ cells, respectively (mean and SEM, n = 3). CD4+ T cells from the WT or CD4-IRF8KO mice were stimulated with anti-CD3/CD28 Abs for 48 h and analyzed for apoptosis using the annexin V assay (F) and RT-PCR (G). Numbers in quadrants and histogram in (F) indicate percentages of CD4+ T cells undergoing apoptosis or necrosis (mean and SEM, n = 3). (H and I) Naive CD4+ T cells from WT or CD4-IRF8KO mice were stimulated with anti-CD3/CD28 Abs for 4 d under Th17-polarizing condition. The cells were then analyzed by FACS for intracellular cytokine staining. Numbers in quadrants and histograms indicate percentages of CD4+ T cells expressing IL-17 and/or ROR-γt (mean and SEM, n = 3). Results are representative of at least three independent experiments. *p < 0.05, **p < 0.01 (Student two-tailed t test).
Generation and characterization of IRF8 conditional KO mice. Irf8fl/fl mice were crossed with CD4-Cre mice to generate mice with deletion of IRF8 in CD4+ T cells (CD4-IRF8KO). The CD4-IRF8KO mice were identified by PCR analysis of mouse tail genomic DNA (A). Naive CD4+ T cells from LN and spleens of WT or F8 generation CD4-IRF8KO mice (after eight cycles of brother–sister mating) were stimulated with anti-CD3/CD28 Abs and analyzed by Western blotting (B) and thymidine incorporation assays (C). Data shown in (C) are mean ± SEM from five replicate cultures. *p < 0.05. The immunophenotype of the stimulated T cells was characterized by FACS and intracellular cytokine-staining assays. Numbers in quadrants indicate percentages of IL-2–expressing CD4+ T cells (D) or percentages of activated CD4+ T cells (E). Histogram presented in (D) and (E) represents percentages of IL-2–expressing cells or CD44highCD25+ cells, respectively (mean and SEM, n = 3). CD4+ T cells from the WT or CD4-IRF8KO mice were stimulated with anti-CD3/CD28 Abs for 48 h and analyzed for apoptosis using the annexin V assay (F) and RT-PCR (G). Numbers in quadrants and histogram in (F) indicate percentages of CD4+ T cells undergoing apoptosis or necrosis (mean and SEM, n = 3). (H and I) Naive CD4+ T cells from WT or CD4-IRF8KO mice were stimulated with anti-CD3/CD28 Abs for 4 d under Th17-polarizing condition. The cells were then analyzed by FACS for intracellular cytokine staining. Numbers in quadrants and histograms indicate percentages of CD4+ T cells expressing IL-17 and/or ROR-γt (mean and SEM, n = 3). Results are representative of at least three independent experiments. *p < 0.05, **p < 0.01 (Student two-tailed t test).
CD4-IRF8KO mice develop more severe EAU
To investigate the impact of deleting IRF8 in CD4+ T cells on development of uveitis, we induced EAU in WT (Irf8fl/fl) and CD4-IRF8KO mice by active immunization with IRBP, a 140-kDa retinal glycoprotein produced by photoreceptor cells (18). Initial signs of uveitis in the C57BL/6J EAU model are generally observed by day 12 postimmunization (p.i.) with full-blown uveitis occurring between days 16 and 22 p.i. with a disease incidence of 100% (18). Disease progression was monitored by fundoscopy and confirmed histologically by assessing the extent of lymphocyte infiltration of the neuroretina and ocular pathology. Fundoscopic images obtained on days 14 and 21 p.i. showed the development of ocular inflammation in the WT mouse eyes characterized by blurred optic disc margins and enlarged juxtapapillary areas, retinal vasculitis with moderate cuffing, vitreitis, choroiditis, and yellow-whitish retinal and choroidal infiltrates (Fig. 2A). In contrast, the fundus images reveal more severe disease in the eyes of CD4-IRF8KO mice with significantly higher clinical scores compared with the eyes of WT mice (Fig. 2A). Histological analysis of retinas from eyes harvested 21 d p.i. underscored the severity of EAU in the eyes of CD4-IRF8KO mice. Compared with EAU in the immunized WT mice, the disease in the CD4-IRF8KO mice was characterized by the infiltration of massive numbers of inflammatory cells into the retina, resulting in substantial destruction of retinal cells, development of more retinal folding, serous retinal detachment, vasculitis, retinitis, choroiditis, and vitritis. These pathological features are hallmarks of severe acute uveitis (18, 23, 24), further reflected by the clinical scores shown in Fig. 2B.
IRF8KO mice develop more severe EAU. EAU was induced in WT or CD4-IRF8KO mice by immunization with IRBP in CFA, and disease progression was analyzed by fundoscopy or histology. (A) Fundus images of eyes harvested 14 or 21 d postimmunization were taken using an otoendoscopic imaging system. Assessment of the severity of the inflammatory disease (EAU scores) was based on changes at the optic nerve disc and retinal vessels or tissues. Black arrow indicates inflammation with blurred optic disc margins and enlarged juxtapapillary areas; blue arrows indicate retinal vasculitis with moderate cuffing; white arrow shows yellow-whitish retinal and choroidal infiltrates. (B) For histological analysis, eyes were harvested 21 d postimmunization. Histologic sections through the retina were stained with H&E, and EAU scores were determined, as described (18). White arrowheads indicate the presence of inflammatory cells in vitreous (V); blue asterisks indicate retinal folds. Results represent at least three independent experiments. **p < 0.01, ***p < 0.001 (Student two-tailed t test).
IRF8KO mice develop more severe EAU. EAU was induced in WT or CD4-IRF8KO mice by immunization with IRBP in CFA, and disease progression was analyzed by fundoscopy or histology. (A) Fundus images of eyes harvested 14 or 21 d postimmunization were taken using an otoendoscopic imaging system. Assessment of the severity of the inflammatory disease (EAU scores) was based on changes at the optic nerve disc and retinal vessels or tissues. Black arrow indicates inflammation with blurred optic disc margins and enlarged juxtapapillary areas; blue arrows indicate retinal vasculitis with moderate cuffing; white arrow shows yellow-whitish retinal and choroidal infiltrates. (B) For histological analysis, eyes were harvested 21 d postimmunization. Histologic sections through the retina were stained with H&E, and EAU scores were determined, as described (18). White arrowheads indicate the presence of inflammatory cells in vitreous (V); blue asterisks indicate retinal folds. Results represent at least three independent experiments. **p < 0.01, ***p < 0.001 (Student two-tailed t test).
IRF8 deficiency in CD4+ T cells promotes the expansion of Th17 cells during EAU
Previous studies of the LCK-IRF8KO mouse strain suggested that T cell–specific Irf8 disruption promotes colitis by increasing Th17 expansion. By comparison, no significant differences in the development or severity of EAE were observed between WT and LCK-IRF8KO mice (6, 17). To investigate the mechanistic basis for the development of a more severe uveitis in CD4-IRF8KO compared with WT mice, we analyzed the immune phenotype and cytokine expression profile of CD4+ T cells in the LN of both mouse strains. We isolated LN cells of unimmunized mice or IRBP-immunized WT or CD4-IRF8KO mice, stimulated the cells with IRBP for 4 d in medium containing CFSE, and performed intracellular cytokine-staining analysis. Consistent with previous studies, development of EAU in the WT mice was associated with increase in Th1 and Th17 cells (13). In keeping with results of the colitis study (6), our data revealed a significant increase in Th17 cells in the LN of CD4-IRF8KO mice as compared with IRF8f/f mice during EAU (Fig. 3). This result suggests that enhanced severity of EAU in mice with IRF8-deficient CD4+ T cells derived, in part, from the expansion of Th17 cells. T regulatory (Treg) cells have recently been shown to bring about resolution of autoimmune uveitis (25). We therefore examined whether severe EAU in the CD4-IRF8KO mice could have resulted from perturbation of the levels of Treg cells. Compared with EAU in the WT mice, the percentage of Foxp3+ Treg cells was significantly reduced in the LN of CD4-IRF8KO compared with WT mice. However, the difference in the levels of the Foxp3+ T cells between the WT and CD4-IRF8KO mice is relatively small, and thus, the extent to which the modest reduction in Treg cells contributes to disease exacerbation in mice with IRF8-deficient CD4+ T cells is not clear (Fig. 3).
Th17 cells are expanded in CD4-IRF8KO mice during EAU. (A) The immunophenotype of freshly isolated LN cells of unimmunized mice or IRBP-immunized WT or CD4-IRF8KO mice was characterized by FACS and intracellular cytokine-staining assays. Numbers in quadrants indicate percentages of IL-17–, ROR-γt–, and/or IFN-γ–expressing CD4+ or CD8+ T cells. (B) LN cells of unimmunized mice or IRBP-immunized WT or CD4-IRF8KO mice were labeled with CFSE and stimulated with IRBP for 4 d before analysis by FACS for intracellular cytokine staining. The cells were gated on CD3/CD4 cells, and numbers in quadrants indicate percentages of CD4+ T cells expressing IL-17, IFN-γ, or Foxp3. Statistical analysis of the percentage of cytokine-expressing T cells was based on analysis of six mice per group. Results represent at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Student two-tailed t test).
Th17 cells are expanded in CD4-IRF8KO mice during EAU. (A) The immunophenotype of freshly isolated LN cells of unimmunized mice or IRBP-immunized WT or CD4-IRF8KO mice was characterized by FACS and intracellular cytokine-staining assays. Numbers in quadrants indicate percentages of IL-17–, ROR-γt–, and/or IFN-γ–expressing CD4+ or CD8+ T cells. (B) LN cells of unimmunized mice or IRBP-immunized WT or CD4-IRF8KO mice were labeled with CFSE and stimulated with IRBP for 4 d before analysis by FACS for intracellular cytokine staining. The cells were gated on CD3/CD4 cells, and numbers in quadrants indicate percentages of CD4+ T cells expressing IL-17, IFN-γ, or Foxp3. Statistical analysis of the percentage of cytokine-expressing T cells was based on analysis of six mice per group. Results represent at least three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Student two-tailed t test).
Loss of Irf8 in retinal microglial cells and neurons confers protection from uveitis
In previous reports, we showed that retinal cells express IRF8 during EAU and that IRF8 expression is localized to the photoreceptors and microglial cells in the ganglion cell layer of the retina (12, 13, 26). In view of recent reports that microglial cells of the brain express IRF8 and that activation of microglia by IRF8 exacerbates neuroinflammation (9, 10, 17), we sought to determine whether the loss of IRF8 in the retina would confer protection from EAU similar to EAE (17). In the EAE model, mice with global deletion of Irf8 or loss of IRF8 in monocyte and macrophage are resistant to EAE (6). In this study, it was most expedient to address the role of IRF8 in mechanisms that regulate ocular inflammation in the retina by use of a Cre-lox strategy to generate mice that lack IRF8 in the retina. We generated mice with a targeted deletion of Irf8 in the retina using two mouse strains with overlapping patterns of Cre recombinase expression in the retina (27, 28). The Pax6/αCre mouse strain specifically expresses the Cre-recombinase in photoreceptors, retinal neurons, microglia cells, and all retinal cell types with the exception of amacrine cells (27), whereas the Rx-Cre mouse that expresses Cre recombinase under direction of the medaka fish promoter element mediates inactivation of gene expression in the developing retina (28).
PCR analysis of tail DNA of mice from the cross between Pax6/αCre-Cre or Pax6-Cre and Irf8fl/fl mouse strains on the C57BL/6 background confirmed the generation of Rx-IRF8KO (Fig. 4A) and αCre-IRF8KO (Fig. 4B) mice with targeted deletion of IRF8 in retinal cells. To directly examine the effects of loss of IRF8 in retinal cells, we induced EAU in WT (Irf8fl/fl), αCre-IRF8KO, and RX-IRF8KO mice by active immunization with IRBP in CFA and monitored the disease by fundoscopy or histology. We first established that IRF8 transcripts were indeed absent in retinal cells by isolating RNA from retinal cells of WT, RX-Irf8KO, or αCre-Irf8KO mice, and then analyzing for expression of IRF8 mRNA by RT-PCR. We show that, although IRF8 transcripts were detected in the retina of WT mice, they were not detectable in retinal cells of RX-Cre-IRF8KO or αCre-IRF8KO mice (Fig. 4C). Fundus images and histological analyses of the day 21 retina show a significant reduction in EAU pathology in the αCre-IRF8KO mice (Fig. 4D). Compared with EAU in the immunized WT mice, we detected fewer numbers of inflammatory cells in the vitreous, less retinal infolding, as well as reduction in other key features of severe uveitis—vasculitis, retinitis, choroiditis, and vitritis. Similar reductions in pathological features of EAU are also observed in day 21 retina of Rx-IRF8KO mice immunized with IRBP (Fig. 4E). The reduced disease observed in both Irf8-deficient mouse strains, as reflected by significantly lower clinical scores, provided strong confirmatory evidence that expression of IRF8 by retinal cells, including retinal microglial cells, may promote neuroinflammation and contribute to the immunopathogenic mechanisms that underlie the development of uveitis.
Loss of Irf8 in neuroretinal cells confers protection from uveitis. (A and B) PCR genotype analysis of tail DNA isolated from WT Irf8f/f or heterozygous Irf8f/− mice (A and B), RX-Cre-Irf8KO (A) or αCre-Irf8KO (B) mice. (C) RNA was isolated from retinal cells of WT, RX-Cre-Irf8KO, αCre-Irf8KO mice and analyzed for IRF8 expression by RT-PCR. The white lines in (A) and (C) indicate where parts of the image were joined. (D and E) EAU was induced in WT, αCre-Irf8KO (D), or RX-Cre-Irf8KO mice (E) by active immunization with IRBP in CFA. Progression and severity of EAU were monitored by funduscopy or histology. Top panels, Fundus images were taken using an otoendoscopic imaging system, and the development of papillitis (black arrows), retinal vasculitis (blue arrows), and inflammatory infiltrates (white arrows) is indicated. Clinical scoring was based on changes at the optic nerve disc, retinal vessels, and surrounding tissues, as described in 2Materials and Methods. Bottom panels, Eyes were enucleated on day 21 postimmunization, fixed in formalin, and embedded in paraffin, and sections were stained with H&E. Histologic sections through the retina were stained with H&E, and EAU scores were determined as described (18). Infiltrated inflammatory cells in the vitreous (V) are denoted by black arrows; blue asterisks indicate retinal folds. Results are representative of at least three independent experiments. ****p < 0.0001 (Student two-tailed t test). Ch, choroid; GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; OpN, optic nerve; RPE, retinal pigment epithelial layer.
Loss of Irf8 in neuroretinal cells confers protection from uveitis. (A and B) PCR genotype analysis of tail DNA isolated from WT Irf8f/f or heterozygous Irf8f/− mice (A and B), RX-Cre-Irf8KO (A) or αCre-Irf8KO (B) mice. (C) RNA was isolated from retinal cells of WT, RX-Cre-Irf8KO, αCre-Irf8KO mice and analyzed for IRF8 expression by RT-PCR. The white lines in (A) and (C) indicate where parts of the image were joined. (D and E) EAU was induced in WT, αCre-Irf8KO (D), or RX-Cre-Irf8KO mice (E) by active immunization with IRBP in CFA. Progression and severity of EAU were monitored by funduscopy or histology. Top panels, Fundus images were taken using an otoendoscopic imaging system, and the development of papillitis (black arrows), retinal vasculitis (blue arrows), and inflammatory infiltrates (white arrows) is indicated. Clinical scoring was based on changes at the optic nerve disc, retinal vessels, and surrounding tissues, as described in 2Materials and Methods. Bottom panels, Eyes were enucleated on day 21 postimmunization, fixed in formalin, and embedded in paraffin, and sections were stained with H&E. Histologic sections through the retina were stained with H&E, and EAU scores were determined as described (18). Infiltrated inflammatory cells in the vitreous (V) are denoted by black arrows; blue asterisks indicate retinal folds. Results are representative of at least three independent experiments. ****p < 0.0001 (Student two-tailed t test). Ch, choroid; GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer; OpN, optic nerve; RPE, retinal pigment epithelial layer.
Upregulation of anti-inflammatory cytokines in the retina correlates with amelioration of EAU in mice with retina-specific disruption of Irf8
Production of IL-12 family cytokines such as IL-12 and IL-23 by APCs has been implicated in pathogenic mechanisms that mediate EAU and EAE (29, 30), whereas IL-27 and IL-35 are thought to contribute to the suppression of both diseases (13–16). We therefore examined whether mitigation of EAU in αCre-IRF8KO and RX-IRF8KO mice derived in part from perturbations in the levels of pro- and/or anti-inflammatory cytkines in the retina during EAU. We obtained cells from the retina of unimmunized C57BL/6J mice or inflamed eyes of IRBP-immunized WT or αCre-IRF8KO mice. RNA isolated from the cells was analyzed by quantitative PCR to assess the transcript levels of cytokines that regulate inflammation. In line with published data, we detected elevated levels of IFN-γ and IL-17 transcripts in the retina of WT mice with EAU (Fig. 5A). In contrast, the levels of these proinflammatory cytokines were significantly lower in the retinas of Cre-IRF8KO compared with those of WT mice during EAU (Fig. 5A). We also detected a significant increase in IL-10 transcripts in the retina of αCre-IRF8KO mice (Fig. 5A). This suggested that the marked reduction of EAU in Irf8 null mouse strains derived in part from enhanced production of the immunosuppressive cytokine, IL-10, and reductions in expression of the inflammatory Th1 and Th17 cytokines in the target tissue. IL-12 family cytokines have profound influences on Ag presentation and lymphocyte lineage commitment (31, 32). In the EAE model, IRF8 produced by activated microglia exacerbated neuroinflammation (17) by upregulating the production of IL-12 and IL-23 while downregulating IL-27, resulting in the induction of a cytokine milieu that favored the expansion of pathogenic Th17 cells. We therefore examined whether loss of IRF8 expression in the retina modulates expression of IL-12 family cytokines in the retina during EAU. We found that expression of transcripts for Ebi3, a cytokine subunit shared by IL-27 (IL-27p28/Ebi3) and IL-35 (IL-12p35/Ebi3), was >3-fold higher in the retinas of αCre-IRF8KO compared with WT mice (Fig. 5B). In contrast, we did not observe any differences between WT and CD4-IRF8KO mice for expression of IL-12p40, a cytokine subunit shared by IL-12 (IL-12p35/IL-12p40) and IL-23 (IL-23p19/IL-12p40). Taken together, these results suggest that mitigation of ocular pathology in αCre-IRF8KO mice during EAU may derive in part from increased expression of IL-27 and IL-35.
Anti-inflammatory cytokines are upregulated in the retina of IRF8-deficient mice during EAU. EAU was induced in WT or αCre-Irf8KO mice by active immunization with IRBP in CFA. Retina isolated from the eyes of the mice on day 21 postimmunization was digested with collagenase, and retina cells were analyzed for the expression of IFN-γ, IL-17, and IL-10 (A) and IL-35 (EBI3, IL-12p35) or IL-12 (IL-12p40) (B) by quantitative RT-PCR. Results represent at least three independent experiments. *p < 0.05, **p < 0.01, ****p < 0.0001 (Student two-tailed t test).
Anti-inflammatory cytokines are upregulated in the retina of IRF8-deficient mice during EAU. EAU was induced in WT or αCre-Irf8KO mice by active immunization with IRBP in CFA. Retina isolated from the eyes of the mice on day 21 postimmunization was digested with collagenase, and retina cells were analyzed for the expression of IFN-γ, IL-17, and IL-10 (A) and IL-35 (EBI3, IL-12p35) or IL-12 (IL-12p40) (B) by quantitative RT-PCR. Results represent at least three independent experiments. *p < 0.05, **p < 0.01, ****p < 0.0001 (Student two-tailed t test).
Loss of Irf8 in the retina prevents the decline of retinal function during uveitis
Electroretinography measures electrical potential change of the retina in response to light stimulation arising predominantly from changes in the neural activity of photoreceptors and second-order neurons in the retina and is a well-established tool for uncovering gross physiologic changes in the intact retina. Thus, it is routinely used to analyze layer-by-layer changes of retinal function in patients and animals. Changes in the ERG occurring during EAU are indicative of alterations in visual function (33). We analyzed scotopic (dark adaptation) and photopic (light adaptation) ERG responses in EAU mice at day 21 postimmunization. We observed similar ERG responses after dark or light adaptation in unimmunized WT and α-CRE-IRF8KO mice (Fig. 6A). EAU development was associated with a marked reduction in photopic b-wave amplitudes (Fig. 6B). Importantly, the lowest amplitudes were observed in the retinas of WT mice with EAU. The light-adapted b-wave amplitudes of the retinas of Irf8-deficient mice with EAU were higher than that of the retinas of WT mice with disease, but lower than the amplitudes for retinas of unimmunized WT mice (Fig. 6B). These functional changes reflect corresponding pathological lesions documented for these mouse strains by histological findings and fundus changes and confirm results of previous studies showing that rod and cone functions, as measured by ERG, are perturbed during the ocular inflammation associated with EAU (33).
Visual function is altered in IRF8KO mice during EAU. (A and B) Visual function of WT or αCre-Irf8KO mice was analyzed by ERG. (A) ERG response after dark or light adaptation was analyzed in unimmunized mice (three mice or six eyes; n = 6). (B) ERG response after light adaptation was analyzed in IRBP/CFA-immunized mice with EAU. Data are presented as the mean ± SEM of five mice from two individual experiments. Gold asterisks: ERG comparison between unimmunized and immunized wild-type mice (10 eyes at each time point, n = 10). Blue asterisks: ERG comparison between unimmunized wild-type mice and immunized α-CRE-IRF8KO mice (10 eyes at each time point, n = 10). Black asterisks: ERG comparison between IRBP/CFA-immunized wild-type mice and IRBP/CFA-immunized α-CRE-IRF8KO mice (10 eyes at each time point, n = 10). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Student two-tailed t test).
Visual function is altered in IRF8KO mice during EAU. (A and B) Visual function of WT or αCre-Irf8KO mice was analyzed by ERG. (A) ERG response after dark or light adaptation was analyzed in unimmunized mice (three mice or six eyes; n = 6). (B) ERG response after light adaptation was analyzed in IRBP/CFA-immunized mice with EAU. Data are presented as the mean ± SEM of five mice from two individual experiments. Gold asterisks: ERG comparison between unimmunized and immunized wild-type mice (10 eyes at each time point, n = 10). Blue asterisks: ERG comparison between unimmunized wild-type mice and immunized α-CRE-IRF8KO mice (10 eyes at each time point, n = 10). Black asterisks: ERG comparison between IRBP/CFA-immunized wild-type mice and IRBP/CFA-immunized α-CRE-IRF8KO mice (10 eyes at each time point, n = 10). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Student two-tailed t test).
Discussion
IRF8 is known primarily as a transcription factor that plays critical roles in regulating myeloid and B cell lineage commitment and maturation (4, 8–10). However, its functional role in T lymphocytes other than CD8+ T cells is not well understood. In this study, we generated mice with a targeted deletion of Irf8 in the CD4+ T cell compartment (CD4-IRF8KO) and used them to investigate the contribution of IRF8 to the regulation of CD4+ T cell effector functions with particular emphasis on their role in regulating immunopathogenic mechanisms that mediate the CNS autoimmune diserase, uveitis (34, 35). CD4-IRF8KO CD4+ T cells proliferated more than T cells from WT mice. The increased proliferation was accompanied by elevated levels of CD4+ T cells undergoing apopotosis and upregulated expression of proapoptotic genes. This suggests that a major function of IRF8 in CD4+ T cells may be to restrain proliferation and protect CD4+ T cells from activation-induced cell death. It is, however, of note that IRF8 is also absent from CD8+ T cells because of Cre recombinase-mediated deletion of Irf8 at the double-positive (CD4+CD8+) stage of T cell development in the thymus. We and others have shown that CD4-Cre mice do indeed mediate targeted gene deletion in both CD4+ and CD8+ T cells (36–38). However, in this study, we have focused on effects of Irf8 deletion in CD4+ T cells because EAU is predominantly mediated by CD4+ T cells and, in particular, Th17 and Th1 cells.
We further show that the effect of IRF8 on host immunity is contextual, depending on cell type and the target organ system responding to IRF8 signaling. CD4-IRF8KO mice developed more severe disease characterized by the development of hallmark features of severe uveitis and enhanced expansion of uveitogenic Th17 cells. These results are consistent with a previous report of exacerbated colitis in LCK-IRF8KO mice lacking IRF8 in T cells during the later stages of T cell development (6). However, in contrast to the mouse colitis model in which severe colitis in the LCK-IRF8KO mice resulted from increased effects of Th17 cells without perturbations of Th1 cells, exacerbation of ocular inflammation in the CD4-IRF8KO mice was also associated with significant inhibition of Th1 cells. It is therefore noteworthy that the marked decrease of Th1 cells in CD4-IRF8KO mice with exacerbated uveitis is consistent with the notion that Th1 cells may have a protective role in uveitis (13, 39). In contrast to the pathogenic roles played by Th17 cells in human and mouse uveitis (13, 39), Treg cells are immunosuppressive and have recently been shown to be effective in ameliorating EAU (25). We show in this work that the enhanced EAU pathology seen in CD4-IRF8KO also correlated with a small, but significant reduction in the levels of Foxp3+ Treg cells. However, it is doubtful that the modest reduction in Treg cells played a major role in the disease exacerbation observed in the IRF8-deficient mice. It is also not clear how loss of IRF8 results in disparate effects on Th17 and Th1 cells as well as Treg cells, three important CD4+ T subsets that play critical roles in the development or regulation of autoimmune diseases. Clues to the possible involvement of IRF8 in the regulation of Th cell effector functions or lineage specification can be informed by recent reports showing that TCR signaling activates pioneering transcription factors such as BATF and IRF4 to recruit inflammation-associated transcription factor to target gene structures during Th cell differentiation (40, 41). In the context of the Th17 differentiation program, proinflammatory environmental cues, such as hypoxia, have recently been shown to stabilize the Th17 lineage while simultaneously destabilizing the Treg lineage by promoting the interactions of hypoxia-inducible factor 1α with ROR-γt and Foxp3, respectively (41, 42). In contrast, interactions of IRF8 with ROR-γt have the specific effect of silencing Th17 developmental pathways (6). Data presented in this work indicate that IRF8 may antagonize T cell activation and proliferation, raising the possibility that competition between IRF8 and hypoxia-inducible factor 1α for binding to lineage-specific master transcription factors may result in diametrically opposed effects on Th cell development and effector functions. However, because IRF8 activities are dependent on its interactions with other transcription factors (1), its exact role in host immunity is difficult to discern, as it is highly dependent on these partners and their ability to recruit IRF8 to active chromatin loci.
Another surprising observation in this study is that IRF8 expression by T cells may have different effects depending on the organ-specific autoimmune disease. For example, EAU and EAE are two well-characterized models of the human CNS autoimmune diseases, uveitis and multiple sclerosis, respectively, and both diseases share essential immunopathological mechanisms in terms of critical roles of Th17 cells in their etiology (35, 43). It is therefore surprising that loss of IRF8 in CD4+ T cells exacerbates EAU, but does not have a significant role in EAE (17). A possible explanation for the different outcomes may relate to the different strategies used to delete IRF8 in T cells. In this study, we deleted Irf8 in T cells using the CD4-Cre mice with highly restricted expression of the Cre recombinase in the CD4 compartment (21, 22), whereas, in the EAE model, the Lck-Cre mice was used to delete Irf8 in T cells. It is of note that the Lck-Cre mouse strain does not faithfully recapitulate the expression pattern of the corresponding endogenous Lck gene, which can result in off-target expression of the Cre recombinase (21, 22, 44). Nonetheless, if the difference derives from use of CD4-Cre versus the Lck-Cre mice, it is difficult to reconcile the fact that colitis was exacerbated, whereas development or severity of EAE was not affected despite the use of Lck-IRF8KO mice in both studies (6).
Data presented in this work showing that the loss of IRF8 in retinal cells prevents the decline in retinal function during uveitis extend our previous observation that the expression of IRF8 by retinal cells is upregulated during ocular inflammation (12). Recent reports have identified critical roles of IRF8 in the functional maturation of brain microglia (8–10), suggesting potential functional relevance of IRF8 expression by retinal microglial cells and neurons. In this study, we generated mice with loss of IRF8 in retinal cells to examine whether expression of this putative immunosuppressive transcription factor by endogenous retinal cells, including photoreceptors, retinal neurons, and microglia cells, can influence the development or severity of uveitis or contribute to mechanisms of ocular immune privilege. Similar to the EAE model, loss of IRF8 in retinal microglial cells and neurons confers protection against neuroinflammation. In EAU, the protective effect was associated with induction of a cytokine milieu characterized by upregulation of the immunosuppressive cytokines IL-27, IL-35, and IL-10, as well as complement factor H. Interestingly, the loss of IRF8 in the retina had no effects on vision. However, during ocular inflammation, rod and cone functions are compromised, leading to diminished sensitivity to light stimuli and vision loss associated with chronic uveitis (33). Surprisingly, our evaluation of visual function in the αCre-IRF8KO mice by electroretinography revealed that the loss of IRF8 partially rescued the loss of visual function and suggests that the increase in IRF8 expression by microglial cells during ocular inflammation may lead to aberrent activation of microglial cells and compromise visual function.
Uveitis comprises a heterogeneous group of potentially sight-threatening inflammatory diseases that includes sympathetic ophthalmia, birdshot retinochoroidopathy, Behcet’s disease, Vogt-Koyanagi–Harada disease, and ocular sarcoidosis and may account for >10% of severe visual handicaps in the United States (45, 46). Conventional treatments of uveitis such as corticosteroids can cause serious systemic side effects, and there is considerable impetus for seeking alternative therapies such as biologics that can be used to target proinflammatory pathways that mediate autoimmune diseases. Data presented in this study suggest that blockade of the IRF8 signaling pathway may be used to downregulate Th17 cells, a pathogenic T cell subset that mediates uveitis. In contrast, IRF8 activation in microglial cells and APCs has recently been shown to activate TGF-β signaling and induce expansion of pathogenic T cells that mediate neuroinflammation (17). We show in this work that the loss of IRF8 in retinal microglia confers protection from severe uveitis. These observations thus highlight the multiple roles played by IRF8 in host immunity and raise a cautionary note in the context of therapeutic targeting of IRF8.
Acknowledgements
We thank Dr. Haohua Qian and Yichao Li (Visual Function Core, National Eye Institute, National Institutes of Health) for ERG technical assistance and Rafael Villasmil (National Eye Institute/National Institutes of Health FLOW Cytometry Core facility) for cell sorting and FACS analysis. We also thank Rashid M. Mahdi (Molecular Immunology Section, National Eye Institute) for technical assistance.
Footnotes
References
Disclosures
The authors have no financial conflicts of interest.