Key role for IL-21 in experimental autoimmune uveitis

Identification of IL-21– and IL-2–Expressing Cells in Vivo Using Il21-mCherry/Il2-emGFP Reporter Mice.

To evaluate IL-21 and IL-2 gene expression, we generated Il21-mCherry and Il2-emGFP BAC dual reporter Tg mice (Fig. 1A). For most experiments, we used a founder line containing one integrated copy of the reporter construct, but behavior of a second independent BAC reporter mouse line was similar (Fig. S1 A and B). To validate the reporter mice, we stimulated CD4⁺ T cells enriched from the Tg reporter and wild-type (WT) littermate mice with anti-CD3 and anti-CD28 and analyzed the expression of emGFP and IL-2 by flow cytometry (Fig. 1B). The frequency of IL-2⁺ cells determined by intracellular staining was ∼35% in both Tg reporter mice and WT controls. In the reporter mice, >95% of emGFP⁺ cells also produced IL-2, and ∼70% of IL-2–producing cells were emGFP⁺, even though our fixation procedure included paraformaldehyde, which can quench the emGFP fluorecence. We also stimulated CD4⁺ T cells with anti-CD3 + anti-CD28 and sorted emGFP⁻ and emGFP⁺ cells. By intracellular staining, 60% of the emGFP⁺ but only 11% of emGFP⁻ cells were IL-2⁺ (Fig. S2A). Moreover, we evaluated IL-2 and IL-21 mRNA expression of sorted emGFP⁻mCherry⁻ double-negative (DN), emGFP⁺ and mCherry⁺ single-positive, and emGFP⁺mCherry⁺ double-positive (DP) cells by real-time PCR. mCherry⁺ and DP cells had ∼30-fold higher IL-21 mRNA expression than DN cells. As expected, IL-21 mRNA expression in GFP⁺ cells was weaker. emGFP⁺ cells had higher IL-2 mRNA expression than DN and mCherry⁺ cells, and IL-2 mRNA expression of DP cells was less than in GFP⁺ cells but still fivefold higher than in DN cells (Fig. 1C). We also stimulated CD4⁺ T cells with anti-CD3 + anti-CD28 with or without IL-6 and TGF-β for 18 h and evaluated mRNA expression of mCherry, IL-21, emGFP, and IL-2 by RT-PCR. Levels of mCherry and IL-21 mRNA were highly correlated (R² = 0.99) as were emGFP and IL-2 mRNA (R² = 0.89) (Fig. 1D), further validating the Il2-emGFP/Il21-mCherry reporter mice.

We next stimulated CD4⁺ T cells with anti-CD3 + anti-CD28 under neutral conditions (presence of anti–IFN-γ + anti–IL-4) versus Th1 or Th17 conditions and analyzed cells at days 1, 2, and 3 (Fig. 1E). emGFP as a marker for IL-2 increased under all conditions but least under Th17 conditions, whereas mCherry as a marker of IL-21 production was highest in Th17 cells. As expected, Th1 cells expressed both emGFP and mCherry, consistent with their known production of IL-21 (1) as well as of IL-2. Sorting of cells after Th1 stimulation revealed that ∼28.5% of emGFP⁺ cells produced IFN-γ versus <2% of mGFP⁻ cells (Fig. S2B). Similarly, sorting of cells after Th17 differentiation revealed that ∼11% of the mCherry⁺ cells produced IL-17A versus <3% of mCherry⁻ cells (Fig. S2C).

To identify IL-21– and IL-2–producing cells in vivo, we next examined Il21-mCherry/Il2-emGFP expression in liver, spleen, and other lymphoid tissues. CD4⁺ cells that expressed mCherry, emGFP, or both were present in all tissues examined, but the frequencies of the cell subsets differed in different tissues (Fig. 1F). Approximately 1–2% of CD4⁺ T cells in spleen, as well as axillary, cervical, and mesenteric LNs, were Il2-emGFP⁺, in contrast to ∼0.5% in the liver, whereas there were more of these cells (3–4% of CD4⁺ T cells) in the small intestine lamina propria lymphocytes (SI LPLs) and in the colon LPLs, consistent with higher basal stimulation in the gut (Fig. 1F, Top). For Il21-mCherry, expression tended to be lowest in liver, axillary, and cervical LNs, slightly higher in spleen, mesenteric LNs, and colon, and highest (10% of cells) in SI LPLs (Middle). Cells expressing both Il2-emGFP⁺ and Il21-mCherry were detected in all tissues, but were highest in the gut (Bottom). Thus, Il2-emGFP⁺ and Il21-mCherry⁺ cells were present in all immune compartments examined, with relative enrichment in the intestine. Evaluation of splenic CD4⁺ T cells revealed that ∼80% of these Il2-emGFP⁺ and Il21-mCherry⁺ cells were CD44^hiCD4⁺ effector/memory type T cells (Fig. 1G). When we gated on CD4⁺CD44⁺mCherry⁺ and CD4⁺CD44⁺emGFP⁺ cells, 39% of Il21-mCherry⁺ cells and 38.6% of Il2-emGFP⁺ cells were both mCherry⁺ and emGFP⁺ (Fig. 1H).

Augmented Expression of IL-21 and IL-2 Expression in EAU.

We next immunized the reporter Tg mice with IRBP and examined mCherry and emGFP expression at day 21, at which point mice had developed severe uveitis. After immunization, we detected Il2-emGFP⁺ CD4⁺ T cells in draining LNs (Fig. 2 A and Left Panel of B). mCherry also increased from 2% of CD4⁺ T cells before immunization to ∼9% afterward (Fig. 2B, Center). After immunization, ∼4% of CD4⁺ cells were Il21-mCherry⁺/Il2-emGFP⁺ DP cells (Fig. 2B, Right). From 50,000 LN cells for each sample, mCherry⁺ cells increased from ∼220 to ∼590 cells after immunization. emGFP⁺ cells and DP cells increased from ∼200 to ∼500 and from ∼90 to ∼210 cells, respectively (Fig. 2C). Similar results were obtained with a second reporter BAC Tg mouse line containing six copies of the BAC clone rather than one (Fig. S1). We also sorted DN, emGFP⁺, mCherry⁺, and DP cells from CD4⁺-draining lymphocytes 21 d after immunization and performed both RT-PCR and intracellular staining. As expected, mCherry⁺ cells had high Il17a mRNA expression (Fig. 2D), consistent with IL-17A cytokine production (Fig. 2E). Ifng mRNA was expressed in emGFP⁺, mCherry⁺, and DP cells (Fig. 2D), and correspondingly, IFN-γ protein was produced by these three populations (Fig. 2E). We also examined IL-21– and IL-2–producing cells in draining LNs and retina using confocal microscopy and reporter Tg mice not immunized or immunized with IRBP. Il21-mCherry and Il2-emGFP were expressed in draining LNs at day 7 (Fig. 2F) and in the eye at day 21 after immunization (Fig. 2G). Of the emGFP⁺ and/or mCherry⁺ cells infiltrating the retina, ∼50% were Il2-emGFP⁺, ∼25% were Il21-mCherry⁺, and ∼25% were DP cells (Fig. S3). Thus, during EAU development, expression of both IL-2 and IL-21 was induced in autoreactive CD4⁺ T cells, and cells expressing each cytokine as well as cells expressing both cytokines were present in the inflammatory cells infiltrating the retina.

Less Severe EAU in Il21r^−/− Mice.

Given the role of IL-21 in the development of autoimmune diseases and its expression within the retina, we investigated the development of EAU in WT versus Il21r^−/− mice, monitoring by fundoscopy from day 10 after immunization to the end of the study period (approximately day 25). Compared with the normal control (Fig. 3A, Left), WT retinas at day 21 showed severe inflammation with blurring of the optic disk margins, retinal vasculitis with cuffing (red asterisks), and inflammatory infiltrates (black arrows) (Fig. 3A, Center). Histological analysis revealed massive inflammatory cell infiltration in the vitreous (black arrows), photoreceptor cell damage and retinal folds (Fig. 3B, blue arrows), and choroiditis (Fig. 3B, Center panels). In contrast, Il21r^−/− mice had modest fundoscopic changes, with less inflammation, only slight retinal vasculitis with cuffing (red asterisk) (Fig. 3A, Right), and less severe histological changes (Fig. 3B, Right panels). EAU scores based on fundoscopy (31) (Fig. 3C) or histology (32) (Fig. 3D) showed much less disease in the Il21r^−/− mice. Correspondingly, at day 21, the proliferation of draining LN cells from EAU WT mice was greater than in unimmunized animals in response to IRBP (Fig. 3E). Proliferation was induced by immunization with IRBP in Il21r^−/− mice but was lower than in WT mice (Fig. 3E). IL-17–expressing and IFN-γ–expressing T cells have been implicated in EAU (26). We therefore examined the production of IL-2, IL-17, and IFN-γ from CD4⁺ T cells in peripheral blood mononuclear cells (PBMCs) after EAU induction. There was significantly less IL-17 and IFN-γ production in Il21r^−/− mice, but IL-2 production was similar in WT and Il21r^−/− mice (Fig. 3 F and G). Moreover, when draining LN cells were harvested at day 21 and then stimulated with IRBP for 3 d, there was less IL-17A and IL-1β but more IL-10 production in Il21r^−/− mice (Fig. 3H). IL-6 production tended to be lower, but the difference was not significant (Fig. 3H). Overall, these data indicate a possible shift from pro- to anti-inflammatory cytokines in the absence of IL-21 signaling.

Adoptive Transfer of IRBP-Specific Uveitogenic Lymphocytes from Il21r^−/− Mice Induces Less Severe EAU than WT Cells.

To confirm that the Il21r^−/− mouse defect occurred in the generation of IRBP-specific pathogenic T cells, we adoptively transferred WT and Il21r^−/− IRBP-specific lymphocytes into WT mice. We found similar CD3 and CD4/CD8 profiles (Fig. 4A, Left and Center panels), but the Il21r^−/− cells produced less IL-17A and IFN-γ but similar levels of IL-2 (Fig. 4A). Compared with mice receiving WT cells, animals receiving Il21r^−/− cells exhibited less severe EAU as evaluated by fundoscopy (Fig. 4B), histological analysis (Fig. 4C), and clinical EAU score (Fig. 4D).

Mice.

Il21r^−/− mice (10) were produced by breeding heterozygous mice. C57BL/6 mice were from the Jackson Laboratory. To generate Il2-emGFP/Il21-mCherry Tg reporter mice, we used the RP23-98I15 BAC clone (Invitrogen), which contains the Il21 and Il2 genes (Fig. 1A). Recombineering (37) was used to introduce the emGFP (Invitrogen) and mCherry (38) coding sequences linked to the SV40 poly(A) site into the first exon of the Il2 and Il21 genes, respectively. Recombineering comprises a two-step method to create precise genetic changes: first, amp-sacB is placed in the DNA and then this is replaced with the emGFP or mCherry coding sequences in a second recombineering event. The targeting cassettes with amp-sacB were amplified using primers with ≥50 bp 5′ overhangs that were homologous to the regions of the BAC just outside of the segment being replaced. The primers are provided in SI Materials and Methods. BAC clones that had integrated the targeting construct were selected, and the location of the insert was confirmed by PCR and sequencing. The BAC was prepped using a Large-Construct kit (Qiagen), digested with NotI, DNA purified by phenol/chloroform extraction, and microinjected into fertilized C57BL/6J × CBA/J oocytes. Resulting pups were screened for the emGFP and mCherry reporters by PCR (see SI Materials and Methods for the primers). Founder lines were backcrossed at least six generations to C57BL/6 mice. Experiments were performed under protocols approved by the National Eye Institute and/or National Heart, Lung, and Blood Institute Animal Care and Use Committees.

Analysis of CD4⁺ T-Helper Cells.

For Th1 conditions, CD4⁺ T cells (>98% pure) were treated with plate-bound 2 μg/mL anti-CD3 + soluble 1 μg/mL anti-CD28, 10 μg/mL anti–IL-4, and 10 ng/mL IL-12 for 3 d. For Th17 conditions, cells were treated with anti-CD3/CD28, 20 ng/mL IL-6, 2 ng/mL TGF-β1, and 10 μg/mL each of anti–IFN-γ and anti–IL-4. Cytokines and antibodies were from R & D Systems. For intracellular cytokine detection, cells were stimulated for 4 h with 10 ng/mL phorbol 2-myristate 3-acetate (PMA) + 1 μM ionomycin, Golgi-stop added in the last 2 h, and intracellular cytokine staining performed using a Cytofix/Cytoperm kit (BD Pharmingen). Flow cytometry was performed using a FACSCalibur (BD Biosciences). All mAbs were from BD PharMingen. Flow cytometry was performed on an LSRII (BD Biosciences). mCherry was detected using a 561-nm laser with a 605/40 filter.

Real-Time PCR.

mRNA was isolated from 10⁶ cells (RNeasy mini kit; Qiagen) and cDNA was prepared (Omniscript RT kit, Qiagen). Real-time PCR was performed with an ABI PRISM 7700 sequence detection system with site-specific primers and probes (Applied Biosystems).

Induction and Evaluation of EAU.

Mice were immunized with 150 μg bovine IRBP and 300 μg human IRBP peptide (amino acids 1–20) in 0.2 mL 1:1 vol/vol emulsion with complete Freund's adjuvant (CFA) containing Mycobacterium tuberculosis strain H37RA (2.5 mg/mL). Mice simultaneously received 0.3 μg Bordetella pertussis toxin. Clinical disease was scored by fundoscopy (31) and histological analysis (32), as detailed in SI Materials and Methods. Tissue sections were also examined using a multiphoton laser-scanning confocal microscope with fluorescence emission detected with 515/30 (emerald-GFP, green) and 600/40 (mCherry, red), as described (26, 31). For adoptive transfer experiments, draining LN cells from WT and Il21r^−/− EAU mice (at day 21) were stimulated with IRBP 20 μg/mL for 4 d, and 10⁷ cells in 200 μL of PBS were injected i.v. into sex- and age-matched recipient mice. Fundoscopic examinations were performed 14 d after adoptive transfer.

Cytokine Analysis and Lymphocyte Proliferation Assays.

A total of 5 × 10⁶ draining LN cells were cultured for 3 d in medium containing 10 μg/mL IRBP protein, and IL-6, IL-10, IL-1β, and IL-17 levels were measured using an ELISA kit (BD PharMingen). For proliferation, normal control and EAU draining lymphocytes (2 × 10⁶/mL) were stimulated with 20 μg/mL IRBP. After 48 h, cultures were pulsed with [³H]thymidine (0.5 μCi/10 μl/well) and incorporation was measured 12 h later.

Statistical Analyses.

Nonparametric t tests and Prism software (Graphpad Software) were used.