Loss of the lupus autoantigen Ro52/Trim21 induces tissue inflammation and systemic autoimmunity by disregulating the IL-23–Th17 pathway

We inserted an IRES-GFP reporter cassette into the Ro52 locus to generate Ro52-deficient mice in which the GFP expression allowed us to track Ro52-expressing cells in vivo (Fig. 1, A–C). Tissue specificity of Ro52 expression has not been investigated systematically, and using GFP as a reporter for Ro52 expression, we observed that Ro52 was predominantly expressed in the lymphoid compartments of spleen, lymph node, and thymus, with little or no expression in nonimmune tissues including liver, pancreas, heart, kidney, and skin (Fig. 1 D and Fig. S1). Some expression was observed in endothelial cells (Fig. S1). Our findings are supported by microarray analysis, showing that Ro52 messenger RNA is predominantly expressed in lymphoid cell populations in both mice and humans (http://biogps.gnf.org). These observations suggest that the primary function of Ro52 may be in the immune system. The vast majority of leukocytes in the bone marrow, thymus, spleen, lymph nodes, and peripheral blood expressed Ro52, as detected by GFP expression, with the highest expression in CD3⁺ and CD11b⁺GR-1⁺ cells (Fig. 1 E and not depicted). Using GFP as the reporter, we noted that, as has previously been reported (Rhodes et al., 2002; Strandberg et al., 2008), the expression of Ro52 was induced by IFN-γ and, to a lesser extent, by IFN-α and IFN-β (Fig. 1 F). Other cytokines (TGF-β or TNF-α) did not affect the expression of Ro52 (Fig. 1 F).

Ro52^−/− mice were born at a slightly reduced Mendelian ratio (Ro52^+/+, 27%; Ro52^+/−, 51%; Ro52^−/−, 22%; P = 0.0001 at n = 280, Chi-Square test) without affecting the male/female ratio. However, the Ro52^−/− mice that were born appeared normal and had no detectable abnormalities in organ morphology or histology. Specifically, we did not observe any differences in the size or organization in spleen, thymus, and lymph nodes or in subsets of immune cells when comparing young unmanipulated Ro52^+/+, Ro52^+/−, and Ro52^−/− mice (Figs. S2 and S3). However, splenic CD3⁺ T cells of Ro52^−/− mice had an increased frequency of T cells with an activated phenotype (CD62L^low; Fig. S3). These data indicate that although there are discreet signs of early activation in splenic T cells in Ro52^−/− mice, there is no apparent immune phenotype or other physical abnormalities present in naive unmanipulated Ro52-deficient mice.

Individual mice are routinely identified either by ear punching or by metallic ear tags. After ear tagging with metal clips, no visible tissue reaction, other than transient erythema and swelling, is normally observed. However, 25 wk after application of metal ear tags, >90% of Ro52^−/− mice had developed a very severe and progressive dermatitis emanating from the tagged ear, which did not occur in Ro52^+/+ WT littermates (Fig. 2, A-C). Dermatitis also developed in Ro52^+/− mice but was not present in Ro52-deficient mice that had lost their ear tag. Histological analysis of the skin lesions showed epidermal hyperplasia with ulcerations and inflammatory infiltrates consisting primarily of neutrophils (Fig. 2 B). Serum immunoglobulin analysis revealed increased IgG levels in Ro52^−/− mice with dermatitis (Fig. 2 D), and the affected mice also developed antinuclear antibodies (ANAs), with specificity for DNA, and autoantibodies to several lupus-associated autoantigens and other autoantigens (Fig. 2 E and Fig. S4). Furthermore, affected mice developed kidney pathology with substantiated mesangium and intraglomerular immunoglobulin depositions, as well as proteinuria, showing impaired renal function (Fig. 2 F). The systemic nature of the inflammation was further supported by the development of splenomegaly and lymphadenopathy in affected mice (Fig. S4 and not depicted). Ro52^−/− mice that were not ear tagged did not develop either dermatitis or systemic autoimmunity when observed up to 40 wk of age, suggesting that initial tissue injury induced by physical ear tagging results in uncontrolled tissue inflammation in Ro52^−/− mice and, subsequently, leads to development of systemic autoimmunity.

To elucidate the cellular mechanism causing tissue inflammation and systemic autoimmunity in Ro52-deficient mice, we analyzed draining lymph node cells and splenocytes from the affected animals for phenotypic and functional changes. There was a marked expansion of CD11b⁺GR-1⁺ cells in both spleen and lymph nodes, there was an expansion of CD19⁺ B cells and CD3⁺ T cells in the lymph nodes, and the T cells had attained an activated phenotype in the affected mice (Fig. S4). Furthermore, lymph node and spleen cells from affected mice hyperproliferated spontaneously (not depicted) and produced IL-6, IL-12/IL-23p40, IL-21, IL-22, and IL-17 (Fig. 2, G and H), which was further increased upon T cell activation (Fig. 2, G and H). IL-4 and IFN-γ production was not increased in the Ro52^−/− mice compared with the Ro52^+/+ mice (Fig. 2 H). IL-17 in affected mice was produced by CD4⁺ cells (Fig. 2 I), indicating that Th17 cells are induced in the Ro52^−/− mice after skin inflammation. The increased production of IL-17 also begins to explain why dermal lesions are predominantly composed of neutrophils in the Ro52^−/− mice, as IL-17 recruits and activates these cells.

To further study the hyperresponsive immune system of Ro52^−/− mice in a more controlled manner, we tested the response in Ro52^−/− mice to low doses of the contact-sensitizing agent Oxazolone. Mice lacking Ro52 had a significantly higher contact hypersensitivity to Oxazolone than Ro52^+/+ mice, as measured by ear thickness (Fig. 3 A), with affected ears showing epidermal hyperplasia and inflammatory infiltrates consisting predominantly of neutrophils (Fig. 3 B). Analysis of lymphoid cell populations in the draining lymph nodes of the Ro52^−/− mice treated with Oxazolone showed accumulation of T cells with an activated phenotype (Fig. 3 C) and production of high amounts of the proinflammatory cytokines IL-6, IL-12/IL-23p40, TNF-α, and IL-17 and IL-23R expression after application of Oxazolone (Fig. 3 D). Production of IFN-γ was not different between WT and Ro52^−/− mice, and the frequency of FoxP3⁺ cells was not affected (Fig. 3 D). These data demonstrate that a lack of Ro52 leads to dysregulated tissue inflammation with enhanced generation of effector T cells that produce proinflammatory cytokines.

To understand the mechanism by which a deficiency in the Ro52 E3 ligase leads to increased cytokine production and tissue inflammation, we aimed to identify substrates ubiquitinated by Ro52. IRF8 and IRF3 have been suggested as targets for Ro52-mediated ubiquitination (Kong et al., 2007; Higgs et al., 2008). However, as there is a genetic linkage with IRF5 polymorphisms to SLE and Sjögren's syndrome, and proinflammatory cytokines IL-12/IL-23p40, IL-6, and TNF-α, which are increased in Ro52^−/− mice, are also regulated by IRF5, we hypothesized that IRF5 may be a substrate for Ro52. Using in vivo ubiquitination assays, we confirmed that IRF3 and IRF8 are polyubiquitinated by Ro52 and, in addition, we observed polyubiquitination of IRF5 by Ro52 (Fig. 4 A). SMAD4, which has structural homology to the IRF proteins, was used as a negative control in our experiments and was not ubiquitinated by Ro52 (Fig. 4 A).

To directly investigate the effects of Ro52 on the transcriptional activity of IRFs, we used a GAL4 one-hybrid luciferase reporter approach. IRF3 and IRF5 fused to GAL4 were tested for their ability to drive luciferase expression from a 4xGAL4-luciferase reporter plasmid upon transfection into 293T cells. Cotransfection with Ro52 substantially decreased IRF3 and IRF5 transcriptional activity in response to toll-like receptor (TLR) 3 or TLR9 signaling induced by polyI:C or CpG stimulation (Fig. 4 B). These findings indicate that a deficiency in Ro52 would lead to increased IRF activity in vivo and persistent production of cytokines induced by IRF3 and 5, such as IL-6, TNF-α, IFN-α, IFN-β, and IL-12/IL-23p40. To investigate the cytokine production in Ro52^−/− cells after TLR stimulation, the knockout cells were cultured with TLR agonists including CpG, LPS, polyI:C, and Imiquimod. The stimulation of Ro52^−/− splenocytes or bone marrow–derived macrophages resulted in increased expression of cytokines (IL-6, TNF-α, IFN-α, IFN-β, IL-12/IL-23p40, and IL-23p19) compared with Ro52^+/+ cells (Fig. 4 C and not depicted), confirming that the loss of Ro52 indeed leads to increased expression of many proinflammatory cytokines. Together, our data show that IFNs will induce expression of Ro52 (Fig.1F) and that Ro52, in turn, will suppress expression of the proinflammatory cytokines IL-6, TNF-α, IFN-α, IFN-β, and IL-12/IL-23p40 through ubiquitination of IRF3, 5, and 8. Thus, the cellular role of the IFN-induced Ro52 is to act as a negative-feedback regulator of immune responses and inflammation by suppressing the expression of IRF-induced proinflammatory cytokines.

Ro52^−/− mice developing dermatitis and systemic autoimmunity produced IL-6 and IL-12/IL-23p40 and had increased levels of Th17 cells in the draining lymph nodes. IL-6 and IL-23 are needed for differentiation and maturation of Th17 cells (Cua et al., 2003; Bettelli et al., 2006), and the IL-23–IL-17 axis appears to be hyperactive in lupus and Sjögren's syndrome (Crispin et al., 2008; Nguyen et al., 2008). We therefore hypothesized that genetic disruption of the IL-23–Th17 pathway rather than the IL-12–IFN-γ pathway would result in inhibition of the tissue inflammation and systemic autoimmunity in Ro52^−/− mice. To address this hypothesis, we crossed Ro52^−/− mice with IL-23p19^−/− mice to disrupt the IL-23–Th17 pathway. The increased hypersensitivity with dermatitis and inflammatory infiltrates observed in Ro52^−/− mice after Oxazolone application was completely reversed in the Ro52^−/− mice lacking IL-23p19 chains (Fig. 5, A and B). The enhanced T cell activation observed in Ro52^−/− mice was also reversed in the double knockout mice (Fig. 5 C). In addition, the increased expression of IL-23R, IL-12/IL-23p40, IL-17, and RoRγt in the draining lymph nodes of Ro52^−/− mice was reversed to baseline levels in the Ro52^−/−p19^−/− mice (Fig. 5 D). Furthermore, the skin inflammation after ear tagging in the Ro52^−/− mice did not develop in ear-tagged Ro52^−/−p19^−/− in up to 6 mo (30 wk) of observation (Fig. 5, E and F). The Ro52^−/−p19^−/− mice did not develop ANA or anti-DNA antibodies and did not develop kidney pathology as did the ear-tagged Ro52^−/− mice (Fig. 5, G and H). Together, these data demonstrate that abrogating the IL-23–Th17 pathway abolishes tissue inflammation and systemic autoimmunity observed in the Ro52-deficient mice after ear tagging and Oxazolone treatment.

In summary, our studies demonstrate that Ro52 is a key negative regulator of inflammation. Although expressed at a steady-state level, Ro52 is further induced by IFNs (Rhodes et al., 2002; Strandberg et al., 2008), and our data show that Ro52 then acts to control and down-regulate the proinflammatory cytokines IL-6, TNF-α, IFN-α, IFN-β, and IL-12/IL-23p40, thus providing a negative-feedback loop by suppressing IFN-mediated immune activation. Recent studies indicate dual roles for Ro52 (Kim and Ozato, 2009; Yang et al., 2009) and indicate that this E3 ligase may act to sustain IRF3 and IRF8 stability under some conditions but may promote polyubiquitination and degradation of the same transcription factors after TLR- and IFN-mediated activation of cells (Higgs et al., 2008; Kim and Ozato, 2009). Different functions for Ro52 depending on the degree of cellular activation is consistent with our previous observations that Ro52 resides in the cytoplasm in unstimulated cells but translocates into the nucleus upon IFN stimulation (Strandberg et al., 2008) and that Ro52-mediated ubiquitination is supported both by UbcH5 and UbcH6 ubiquitin-conjugating enzymes/E2 ligases that are present in the cytoplasm and nucleus, respectively (Espinosa et al., 2008). In Ro52^−/− mice, we observed no obvious phenotype in naive unmanipulated mice, but we did observe a severe unrelenting tissue inflammation after tissue injury, progressing into systemic autoimmunity with several lupus-like features, which was dependent on the IL-23–IL-17 pathway, stressing the importance of Ro52 as a negative regulator of inflammatory responses. Our data begin to build an understanding of how loss or genetic deficiency of Ro52 would result in dysregulation of IFN pathways, a feature which is common to many systemic autoimmune diseases as they present clinically with an “IFN signature” (Baechler et al., 2004; Ronnblom et al., 2006). The Ro52 haploinsufficient mice developed disease at the same frequency and with similar phenotypical expression as the mice that completely lacked Ro52, suggesting an important role of Ro52 in regulating immune homeostasis and that even a partial loss of function for this key regulatory protein would result in exaggerated tissue inflammation and systemic autoimmunity. From our data, one would predict that the human genetic linkage to polymorphisms in the Ro52 gene most likely relate to reduction in the level, binding, or function of Ro52 in systemic autoimmune patients.

Although the role of IL-17 and the Th17 pathway in lupus and Sjögren's syndrome is only beginning to be explored, several studies implicate a role for Th17 effector cells also in these autoimmune conditions (Crispin et al., 2008; Nguyen et al., 2008; Sakai et al., 2008; Wong et al., 2008). Furthermore, our data suggest that loss of IL-17 also suppresses development of exaggerated skin inflammation in Ro52^−/− animals. Together, the present data link Ro52, an E3 ligase which targets many transcription factors in the IFN pathway, to disregulated proinflammatory cytokine production and to the development of tissue specific inflammation and systemic autoimmunity.

Ro52^−/− mice were generated using standard gene targeting techniques. In brief, a short (1.7 kb) and a long (5 kb) arm were PCR cloned from the C57BL/6J BAC clone RP23-311N6 using the following primers: short arm, forward, 5′-ATCGATTGCAGAGGCAAAGCCATGGGAAAG-3′, and reverse, 5′-ATCGATTCCTCCTCTCCAGCTCTGAGATCA-3′; and long arm, forward, 5′-GTCGACAGGCTACAACAGCTGTGCTCAGA-3′, and reverse, 5′-GCGGCCGCACCCTGGCCTTGATTTCTAACAG-3′. They were then ligated into the gene targeting cassette vector pKS-TK-NeoloxP-IRES-GFP. The C57BL/6-derived embryonic stem cell line V26.2 was used for homologous recombination, and targeted embryonic stem cell clones were identified by Southern blotting and injected into BALB/c embryos. Chimeras carrying the target allele were bred to C57BL/6J mice. Ro52^−/−p19^−/− mice were generated by intercrossing Ro52^−/− and p19^−/− mice obtained from the Mutant mouse Regional Resource Center (www.mmrrc.org). The study was approved by the Ethical Review Committee North, Stockholm County and Standing Committee on Animals at Harvard Medical School.

Organs were fixed in 4% paraformaldehyde for 24 h, embedded in paraffin, and 4–5-µm sections stained with hematoxylin/eosin or Toluidine blue or were used in immunohistochemistry with rabbit polyclonal anti-GFP (Invitrogen), rat monoclonal anti-CD3 (clone SP7; Thermo Fisher Scientific), or rat monoclonal anti-B220 (clone RA3-6B2; BD). Antigen retrieval was performed using Retriever 2100 (PickCell Laboratories). For electron microscopy, small pieces of kidneys were further fixed in 2.5% glutaraldehyde and postfixed in 1% OsO₄ plus 1% potassium ferrocyanide, followed by staining with 0.5% uranyl acetate. After dehydration, specimens were embedded in epoxy resin (agar 100) and cured by heat. 50-nm sections were contrasted with lead citrate and uranyl acetate and examined with an electron microscope (912AB; Carl Zeiss, Inc.) equipped with a MegaView III camera (Olympus) for digital image capture. Immunofluorescence staining of kidneys was performed on 7-µm paraformaldehyde-fixed sections of snap-frozen organs with an Alexa Flour 594–conjugated donkey anti–mouse IgG (Invitrogen).

Single cell suspensions were prepared from lymphoid organs and cells blocked in Fc block (BD) for 30 min before staining by fluorophore-conjugated antibodies (BD). For intracellular cytokine staining, cells stimulated in vitro with soluble anti-CD3 were restimulated for 4 h with 50 ng/ml PMA (Sigma-Aldrich), 1 µg/ml ionomycin (Sigma-Aldrich), and 1 µl/ml GolgiStop (BD), followed by surface marker and intracellular cytokine staining according to the manufacturer's instructions. Cells were analyzed on a flow cytometer (BD).

Immunoglobulin isotype serum leves were analyzed using the SBA Clonotyping System/AP kit (SouthernBiotech), ANA by Hep-2 slides (Immuno Concepts), and anti-DNA antibodies, as previously described (Wermeling et al., 2007), in ELISA using plates coated with methylated bovine serum albumin and activated calf thymus DNA (Sigma-Aldrich). Lupus autoantigen microarrays were generated and used, as previously described, with 41 autoantigens spotted in replicates of six (Wu et al., 2008). Sera were diluted 1:50–1:100 in the assays. Proteinuria was detected using Multistix 5 (Bayer).

Culture supernatants were collected after stimulation, as defined in each experiment, to assess cytokine production by ELISA or cytometric bead array (R&D Systems; BD). IFN-α was quantified by dissociation-enhanced lanthanide fluoroimmunoassay as previously described (Strandberg et al., 2008). RNA was extracted with RNeasy columns (QIAGEN) from cells ex vivo and analyzed by quantitative RT-PCR according to the manufacturer's instructions (Applied Biosystems) and expression related to HPRT.

Oxazolone (4-ethozymethylene-2-phenyl-2-oxazolin-5-one) provocation was performed using the ear swelling test as described in Yokozeki et al. (2000). Mice were sensitized on shaved back skin with 150 µl of 0.3% Oxazolone (Sigma-Aldrich) dissolved in acetone/olive oil (4:1) applied on two consecutive days. 5 d later, the ears were challenged with 50 µl of a 0.1% Oxazolone solution on one ear and with the vehicle on the other. This was repeated on three consecutive days, and ear swelling was measured by a micrometer (Kroeplin). The whole procedure was repeated up to four times. Change in thickness was calculated by subtracting swelling caused by the vehicle.

293T cells were transfected with 500 ng FLAG-IRF or FLAG-SMAD4, 500 ng Ro52-GFP, and 500 ng Myc-ubiquitin. The SMAD4 plasmid was a gift from C.-H. Heldin (Uppsala University, Uppsala, Sweden). 10 µM MG132 was added 28 h after transfection and cells were harvested 6 h later. Cells were lysed in immunoprecipitation buffer and, after clearing of the lysate, FLAG-M2-agarose beads (Sigma-Aldrich) were added, followed by incubation for 2 h. After washing, FLAG-IRFs were eluted with FLAG peptide (Sigma-Aldrich) and separated by SDS-PAGE, and ubiquitinated FLAG-IRFs were detected by immunoblotting against Myc (Santa Cruz Biotechnology, Inc.). FLAG-IRFs and GFP-Ro52 were detected in whole cell lysates by immunoblotting against FLAG (Sigma-Aldrich) and GFP (Sigma-Aldrich).

For measurement of IRF activity, 10⁵ 293T cells were transfected in triplicates in 24-well plates using Fugene HD (Roche). Each well was transfected with 20 ng GAL4-IRF, 20 ng MycRo52 or empty Myc vector, 20 ng pRL-TK-Renilla, 20 ng pcDNA3.1-TLR9 or pFLAG-CMV1-TLR3 (gift from R. Medzhitov, Yale University, New Haven, CT; plasmids 13084 and 13091, respectively; Addgene), and 50 ng GAL4 reporter plasmid MH100-luc. After 16 h, cells were harvested and prepared for the reporter assay (Dual-Luciferase Reporter Assay system; Promega). Firefly luciferase luminescence was normalized to Renilla luminescence and used as a measure of IRF activity. Results are shown relative to transfections in which GAL4 was not fused to any IRF.

Data are expressed as mean ± SEM. Analysis of nonparametric data between groups was performed with a Mann-Whitney U test. Analysis, including repeated measures, was performed with an ANOVA mixed model, and a Fischer's exact test was used for analysis of frequencies between groups in small samples.

Fig. S1 shows expression of GFP as a reporter for Ro52. Fig. S2 shows lymphoid organization in Ro52^+/+, Ro52^+/−, and Ro52^−/− mice. Fig. S3 shows leukocyte cell populations in naive mice. Fig. S4 shows autoantibody reactivity and splenomegaly in Ro52^−/− mice.

We thank Nora Ramirez, Vijole Dzikaite, and Ann Trönnberg for skillful technical assistance, Mohamed Oukka for targeting vector design, Henrik Källberg for statistical analyses, and Professor Carl-Henrik Heldin (Uppsala University, Uppsala, Sweden) for the gift of the SMAD4 plasmid.

This work was supported by grants from the Swedish Research Council, the Swedish Foundation for Strategic Research, the Stockholm County Council, King Gustaf the Vth 80-year foundation, the Göran Gustafsson Foundation, the Torsten and Ragnar Söderberg Foundation, the Heart-Lung Foundation, Science Foundation Ireland, and the Swedish Rheumatism Association.

The authors have no conflicting financial interests.