JunB is crucial for Th17-specific gene expression program

To investigate the effect of JunB deficiency on global gene expression under Th17-polarizing conditions, we performed RNA-seq analyses. Consistent with the results obtained in the qPCR analyses (Fig. 1C), expression of the Th17 signature genes (Il17a, Il17f, Il21, and Il23r) was abrogated in the absence of JunB (Fig. 2A). Pathway enrichment analyses revealed that genes involved in Th17-related functions were significantly enriched in the JunB-regulated gene set (Supplementary Figure 4). A strong correlation was observed between genes up-regulated during Th17 differentiation and those impaired by JunB deficiency (Fig. 2A, red dots): genes highly activated during Th17 differentiation had a tendency to be effectively impaired by the absence of JunB. We also found that expression of another subset of genes was down-regulated during Th17 differentiation in a JunB-dependent manner (Fig. 2B, blue dots): genes severely down-regulated during Th17 differentiation tended to be activated by JunB deficiency (Fig. 2B). The hierarchical clustering analysis revealed that lack of JunB drastically altered gene expression in CD4⁺ T cells after culture under Th17-polarizing conditions (Fig. 2C). By contrast, the absence of JunB did not largely affect expression patterns in naive T helper cells and Th0 cells (Fig. 2C). Importantly, the expression profile of Junb-deficient CD4⁺ T cells under Th17-polarizing conditions was more similar to that of normal Th0 cells than that of Th17 cells (Fig. 2C). Thus JunB is indispensable for generation of a gene expression pattern specific to Th17 cells.

Open in a separate window

Figure 2

JunB governs the Th17-specific gene expression program. (A) Heat map of hierarchical clustering of 234 genes down-regulated in Junb-deficient Th17 cells compared to control Th17 cells. Each column displays scaled FPKM of 234 genes in the indicated sample. Arrows indicate typical Th17 signature genes. (B) Correlation between JunB-regulated genes and those differentially expressed during Th17 differentiation. For each gene, log2FC between control and Junb-deficient Th17 cells are plotted against log2FC during Th17 differentiation from control naive or Th0 cells. Red and blue dots indicate genes significantly up- and down-regulated during Th17 differentiation, respectively (FDR < 0.001, |log2FC| > 1). (C) Dendrogram of unsupervised hierarchical clustering of the transcriptomes of control and Junb-deficient naive, Th0, and Th17 cells.

JunB is dispensable for development of Th1, Th2, and Treg cells

In contrast to the essential role of JunB in Th17 differentiation, neither Th1 nor Th2 differentiation was affected by ablation of Junb: interferon-γ (IFN-γ) and IL-4 were normally synthesized under Th1- and Th2-polarizing conditions, respectively (Fig. 3A–C). In addition, Foxp3 (encoding Foxp3), which specifies differentiation into Treg cells^1,2, was expressed in Junb-deficient cells as much as in control cells under Treg-polarizing conditions (Fig. 3D). Intriguingly, Foxp3 expression under Th17-polarizing conditions was increased in Junb-deficient cells (Fig. 3D); a similar increase has been also observed by ablation of Irf4 and Batf, each being required for development of Th17 cells^7,8. These findings indicate that JunB is selectively required for Th17 differentiation.

Open in a separate window

Figure 3

JunB specifically controls Th17 differentiation. (A) ELISA analysis of cytokine release from CD4⁺ T cells cultured under the indicated conditions. (B) Foxp3 expression in CD4⁺ T cells cultured under the indicated conditions. Junb KO, Meox2 ^Cre; Junb ^f/f mice; Control, littermate mice with Junb ^+/+ (A) or Junb ^f/+ (B). Data are presented as mean ± SD.

JunB-deficient mice are completely resistant to EAE induction

To examine the in vivo role of JunB in Th17 cell differentiation, we evaluated the effects of Junb ablation in EAE, because Th17 cells are the major pathogenic population in this disease^3,4. Junb ^f/+ mice were immunized with myelin oligodendrocyte glycoprotein peptide 35–55 (MOG_35–55) and monitored for clinical signs of EAE. As shown in Fig. 4A, all Junb ^f/+ mice (n = 25) developed severe EAE. In contrast, none of the Junb-deficient mice (n = 22) displayed signs of paralysis during the 48-day period (Fig. 4A). The EAE phenotype of Junb-deficient mice is similar to that of mice lacking Batf ⁷ or Irf4 ⁸. The resistance of Junb-deficient mice to EAE induction was confirmed by histopathological analysis of the spinal cords of Junb ^f/+ and Junb-deficient mice. As shown in Fig. 4B, three to six demyelinated areas were observed in the spinal cord sections of Junb ^f/+ mice; on the other hand, no demyelinated areas in those of Junb-deficient mice (Fig. 4B). In addition, although both CD3ε⁺ T cells and CD11b⁺ myeloid cells were densely infiltrated into the spinal cord of Junb ^f/+ mice, no infiltration of immune cells occurred in Junb-deficient mice (Fig. 4C). Thus, clinical and histopathological analyses indicate that Junb-deficient mice are completely resistant to EAE.

Open in a separate window

Figure 4

Junb-deficient mice are resistant to EAE. (A) EAE disease course in Junb ^f/+ (control) (n = 25) and Junb-deficient (KO) (n = 22) mice. Mean clinical scores are presented as mean ± SEM. (B and C) Pathological analysis in the spinal cord sections of mice on day 12 after EAE induction (n = 6). Arrowheads indicate demyelinated areas (Luxol fast blue) in (B). Scale bars in (B), 500 µm. Infiltrated cells were detected by immunostaining for CD3ε (T cells), B220 (B cells), and CD11b (myeloid cells) in (C). Scale bars in (C), 100 µm. (D) IL-17A production in MOG_35–55-restimulated CD4⁺ T cells from mice with EAE. Error bars in (B,C and D) represent SD.

To investigate the presence or absence of Th17 cells in MOG-treated Junb-deficient mice, we prepared CD4⁺ T cells and restimulated them with MOG_35–55 in the presence of APCs. As shown in Fig. 4D, IL-17A was abundantly expressed upon restimulation in CD4⁺ T cells from MOG-immunized Junb ^f/+ mice. By contrast, CD4⁺ T cells from immunized Junb-deficient mice failed to produce IL-17A when stimulated with MOG_35–55 (Fig. 4D), confirming that Th17 cells are absent from Junb-deficient mice. Taken together with the present findings, we conclude that JunB is required for Th17 differentiation both in vitro and in vivo.

Because epidermis-specific deletion of Junb is known to result in skin inflammation¹⁹, we studied the effect of systemic Junb deletion in imiquimod-induced dermatitis, a mouse model for psoriasis-like inflammatory disease²³. Treatment with imiquimod induced ear swelling in Junb-deficient mice to the extent similar to that in control mice (Supplementary Figure 5A). In addition, Junb deletion did not affect the induction of psoriasis-associated genes such as Defb4, Il17f, S100a9, and Il19 in imiquimod-treated skin lesions, although the mRNA level of the two other associated genes IIl23a and Il24 in Junb-deficient mice was slightly higher than that in control mice at day 5 after imiquimod treatment (Supplementary Figure 5B). These findings suggest that JunB plays a marginal, if any, role in imiquimod-induced psoriasis.

c-Jun is much less expressed than JunB and JunD in Th17 cells

The present finding that single gene ablation of Junb is sufficient for effective suppression of Th17 development raised the question why Junb plays such an indispensable role in spite of the presence of other Jun family genes. Indeed the two closely-related proteins c-Jun and JunD as well as JunB were each capable of directly interacting with BATF (Supplementary Figure ^6A,B), an AP-1 protein that is required for Th17 differentiation⁷, and can exist in a complex with BATF on an AP-1 site, as demonstrated by recent analysis using electrophoretic mobility shift assays (EMSAs)^24–26. To know the reason for the dominant role of JunB in Th17 development, we first evaluated the relative amounts of the Jun family proteins expressed in Th17 cells. For this purpose, immunoblot analysis was performed for detection of endogenous JunB, c-Jun, and JunD in Th17 cells using the same amounts of the respective FLAG-tagged Jun proteins to make standard curves (see “Methods”; Fig. 5A; Supplementary Figure 7). As estimated by the analysis, c-Jun was much less expressed than JunB in Th17 cells, whereas the amount of JunD protein was slightly smaller than that of JunB (Fig. 5A). Consistent with this, only a marginal expression of mRNA for c-Jun was observed in Th17 cells compared with Junb mRNA expression (Fig. 5B). The low expression of c-Jun in Th17 cells appears to agree with the previous observation that c-Jun is not involved in the AP-1 complex in Th17 cells, in contrast to JunB and JunD²⁵. In addition, Th17 development was not impaired by knockdown of c-Jun using siRNAs, especially c-Jun siRNA #2, and also c-Jun siRNA #3, but to a lesser extent (Supplementary Figure 8). Thus c-Jun does not appear to play a major role in Th17 development because of its low expression, although c-Jun has an ability to form an AP-1 complex with BATF when overexpressed in HEK293T cells²⁶.

Open in a separate window

Figure 5

JunB but not c-Jun is abundantly expressed in Th17 cells. (A) Immunoblot analysis for evaluation of relative expression levels of endogenous Jun family proteins in Th17 cells. The same amounts of FLAG–JunB, FLAG–c-Jun or FLAG–JunD, which were expressed in HEK293T cells, were estimated by immunoblot with an anti-FLAG antibody (M2) (left panel). Serially diluted proteins and the Th17 cell lysate were subjected to immunoblot analysis with anti-JunB, anti-c-Jun, or anti-JunD antibodies (middle panel), followed by quantification with Odyssey Infrared Imaging System. FLAG-tagged and endogenous proteins were denoted by white and black arrowheads, respectively. Relative protein levels of endogenous JunB, c-Jun, and JunD were shown in the right panel in (A). (B) Real-time PCR analysis for relative mRNA copy numbers of Junb and Jun in naive CD4⁺ T cells and Th17 cells. mRNA copy numbers were estimated from standard curves that were generated using known numbers of a plasmid encoding Junb or Jun. Data are presented as mean ± SD.

JunB but not JunD cooperates with BATF to activate Th17 signature genes

To further know the reason why JunB plays the dominant role in Th17 development, we next investigated the role for JunB in BATF-dependent activation of Th17 signature genes and compared it with that for JunD, which is present in Th17 cells at the level comparable to JunB in contrast to the low expression of c-Jun. As described above, JunB as well as JunD is capable of forming a heterodimer with BATF. Although JunB by itself failed to induce transcription via the Il17a promoter, JunB activated the promoter in cooperation with BATF (Fig. 6A,B), indicating that JunB forms a productive dimer with BATF in Il17a activation. On the other hand, JunD did not elicit Il17a transcription even in the presence of BATF. Furthermore, JunB but not JunD cooperated with BATF to activate the Th17 signature genes Il17f and Il23r (Fig. 6C,D). These findings indicate that JunB functions as an indispensable partner of BATF in Th17 development, whereas JunD does not play a major role in regulation of Th17 cells.

Open in a separate window

Figure 6

JunB but not JunD regulates expression of Th17 signature genes. (A,B,C,D and F) Luciferase reporter assay for transcriptional activation of the promoter of Il17a (A,B and F), Il17f (C), or Il23r (D) in HEK293T (A and F), NIH3T3 (B and C), or BW5147α⁻β⁻ (D) cells by exogenous expression of Jun family proteins and/or BATF, and immunoblot (IB) analysis of the exogenously expressed proteins. Data are presented as mean ± SD. (E and G) IL-17A production in Junb-deficient (KO) OT-II CD4⁺ T cells by retroviral overexpression of Jun family proteins (E) or a N-terminally truncated JunB (JunB-∆N) (G) together with EGFP. Production of IL-17A was analyzed by flow cytometry; numbers indicate percent EGFP⁺ cells producing IL-17A.

We also used Junb-deficient CD4⁺ T cells to know the role of Jun family proteins in Th17 differentiation. Compared with JunB, retrovirally expressed JunD only marginally restored Th17 development of the Junb-deficient cells (Fig. 6E). The finding supports the idea that JunD is not a major regulator in Th17 differentiation. The activity of JunB likely depends on the N-terminal region, which is not involved in binding to DNA or dimerization with BATF¹⁵, because truncation of this region resulted in a loss of both activation of Il17a transcription (Fig. 6F) and induction of Th17 differentiation (Fig. 6G). On the other hand, overexpression of c-Jun in the Junb-deficient CD4⁺ T cells partially restored IL-17A production, but less effective than that of JunB (Fig. 6E). Although c-Jun thus may have an ability to replace JunB in Th17 differentiation at least partially, JunB but not c-Jun plays the indispensable role, because c-Jun is expressed in Th17 cells to a much lesser extent than JunB (Fig. 5).

The Fosl2-encoded protein Fra2, another member of the AP-1 family, also has been reported to be important for Th17 cell specification²⁷. Although JunB did not regulate expression of Fosl2 during Th17 development (Supplementary Figure 9A), JunB was capable of forming a complex with Fra2 and activating Il17a transcription (Supplementary Figure ^9B,C), suggesting that Fra2 may control Th17 differentiation by interacting with JunB.

JunB is required for expression of Rorc and Rora

The present findings indicate that JunB regulates Th17 development by forming a heterodimer with BATF. To further know the molecular mechanism whereby JunB functions, we tested its role in expression of the Th17 lineage-specifying factor RORγt (encoded by Rorc)^5,6. It is known that Rorc expression under Th17-polarizing conditions is impaired in CD4⁺ T cells deficient in the Th17-polarizing transcription factors BATF⁷, IRF4⁸, and STAT3⁹, whereas the expression is not affected in CD4⁺ T cells lacking IκBζ (encoded by Nfkbiz), which also participates in Th17 development¹³. As shown in Fig. 7A, expression of Rorc mRNA was prevented in Junb-deficient CD4⁺ T cells at 24 and 72 h after stimulation. The RORγt-related protein RORα is also known to regulate development of Th17 cells, e.g., RORα and RORγt molecularly cross-compensate for Th17 differentiation⁶. Although expression of Rora was enhanced during Th17 differentiation in Junb ^f/+ CD4⁺ T cells, the enhancement did not occur in Junb-deficient CD4⁺ T cells (Fig. 7A). The impaired expression of Rorc and Rora in Junb-deficient cells was confirmed by RNA-seq analysis (Fig. 2A). Furthermore, as shown in a chromatin immunoprecipitation analysis (Supplementary Figure 10), at around the transcription start site of Rorc and Rora as well as that of other Th17 signature genes, Junb deficiency resulted in deacetylation of histones H3 and H4, characteristic of transcriptionally inactive chromatin states. It has been reported that Rora expression is prevented also in Batf-deficient CD4⁺ T cells⁷ but not in Nfkbiz-deficient cells¹³. The reduced transcription of Rorc and Rora in Junb-deficient T cells does not seem to be due to perturbation in IRF4 and BATF, because the amounts of these transcription factors were not affected by Junb deficiency in Th17 cells (Fig. 7B). Of note, compensatory elevation in expression of c-Jun and JunD did not occur in the absence of Junb (Fig. 7B). Thus, JunB as well as BATF stimulates the expression of the ROR nuclear receptors RORγt and RORα, which likely contributes to differentiation of Th17 cells.

Open in a separate window

Figure 7

Expression of RORγt and RORα is impaired in Junb-deficient cells. (A) Real-time PCR analysis of Rorc and Rora expression in Junb ^f/+ (control) and Junb-deficient (KO) CD4⁺ T cells cultured under Th17-polarizing conditions. Data are presented as mean ± SD. (B) Immunoblot analysis of production of IRF4, BATF, and Jun family proteins in CD4⁺ T cells cultured under the indicated conditions. (C) Bi-cistronic retroviral overexpression of RORγt together with EGFP and its effect on IL-17A production in CD4⁺ T cells cultured under Th17-polarizing conditions. IL-17A production was analyzed by flow cytometry; numbers indicate percent EGFP⁺ cells producing IL-17A.

We next tested whether overexpression of RORγt is able to rescue impaired development of Th17 cells in Junb-deficient cells. Retroviral expression of RORγt restored IL-17A production impaired in Junb-deficient CD4⁺ T cells under Th17-polarizing conditions, but to a small extent (less than 10%) (Fig. 7C). Similarly, a loss of Th17 differentiation by ablation of Batf or Irf4 is only slightly rescued by overexpression of RORγt^7,8. The partial restoration indicates that RORγt expression is not sufficient for differentiation of Th17 cells, and thus raises a possibility that RORγt function is also supported by the Th17-polarizing factors JunB, BATF, and IRF4. It has recently been proposed that early binding of BATF and IRF4 to Th17-associated genes governs chromatin accessibility and subsequent recruitment of RORγt²⁷, and thus it seems likely that the BATF-partner JunB also participates in initial chromatin accessibility of RORγt for Th17 cell specification.

RORγt-dependent expression of IL-17A is not only a feature of Th17 cells but also that of other RORγt⁺ cells including γδT cells²⁸ and NKp46⁻CCR6⁺CD4⁺ group 3 innate lymphoid cells (ILC3)²⁹. To know the role of JunB in IL-17A production by these cell types, we analyzed lymphocytes from short intestinal lamina propria of Junb-deficient and control mice. As expected, ablation of Junb led to a severe impairment in IL-17 production by Th17-containing populations such as CD4⁺TCRβ⁺ cells (Supplementary Figure 11A) and CD3ε⁺TCRγδ⁻ cells (Supplementary Figure 11B). By contrast, Junb depletion only modestly reduced IL-17A production by CD4⁺TCRβ⁻ cells, containing IL-17-producing ILC3 cells (Supplementary Figure 11A), and that by CD3ε⁺TCRγδ⁺ cells (γδT cells) (Supplementary Figure 11B). These findings indicate that JunB plays an indispensable role in IL-17A expression by Th17 cells, but not in that by ILC3 and γδT cells.

In vitro differentiation of CD4⁺ T cells

Naive CD4⁺ T cells were prepared from the spleen and lymph nodes of 6–9 week old mice by magnetic sorting using Dynabeads^® Mouse CD4 (L3T4, Invitrogen) and DETACHaBEAD^® Mouse CD4 (Invitrogen), as described previously⁴⁷. The CD4⁺ T cells were activated for 3 days via TCR with plate-bound 5 μg/ml of an anti-CD3ε antibody (145-2C11, BioLegend) and soluble 2.5 μg/ml of an anti-CD28 antibody (35.71, BioLegend) under the following differentiation conditions: 10 μg/ml of an anti-IFN-γ antibody (XMG 1.2, BioLegend) and 10 μg/ml of an anti-IL-4 antibody (11B11, BioLegend) for Th0; 10 μg/ml of the anti-IFN-γ antibody, 10 μg/ml of the anti-IL-4 antibody, 40 ng/ml of mouse IL-6 (Peprotech), and 4 ng/ml of human TGF-β1 (R&D) for Th17; 10 μg/ml of the anti-IL-4 antibody and 40 ng/ml of mouse IL-12 p70 (BioLegend) for Th1; 10 μg/ml of the anti-IFN-γ antibody and 40 ng/ml of mouse IL-4 (Peprotech) for Th2; 10 μg/ml of the anti-IFN-γ antibody, 10 μg/ml of the anti-IL-4 antibody, 4 ng/ml of human TGF-β1 for Treg.

Cell surface staining and intracellular cytokine staining

For cell surface staining, naive CD4⁺ T cells were stained with fluorescence-labeled antibodies: fluorescein isothiocyanate (FITC)-labeled anti-mouse CD8a (53–6.7, BioLegend), phycoerythrin (PE)-labeled anti-CD4 (RM4-5, BioLegend), allophycocyanin (APC)-labeled anti-CD4 (RM4-5, TONBO), APC-labeled anti-CD3ε (145-2C11, TONBO), APC/Cy7-labeled anti-CD45.2 (clone104, BioLegend), PE/Cy7-labeled anti-TCRβ (H57-597, TONBO), PE/Cy7-labeled anti-TCRγ/δ (GL3, BioLegend), FITC-labeled anti-mouse CD62L (MEL-14, BioLegend), PE/Cy5-labeled anti-mouse/human CD44 (1M7, BioLegend), FITC-labeled anti-mouse Ly6G (1A8, TONBO), and PE-labeled anti-mouse CD11b (M1/70 TONBO) antibodies were used. For intracellular cytokine staining, cells were stimulated for 4 h with phorbol 12-myristate 13-acetate (PMA) (50 ng/ml) and ionomycin (500 ng/ml) in the presence of GolgiPlug (BD Bioscience), and then fixed in a buffer containing 4% paraformaldehyde (BioLegend). The fixed cells were permeabilized in a permeabilization buffer (BioLegend) and stained with PE-labeled anti-mouse IL-17A (TC11-18H10, BD Pharmingen) and FITC-labeled anti-IFN-γ (XMG1.2, BioLegend) antibodies. Flow cytometry was carried out using FACS Calibur (BD Bioscience) at the Research Support Center, Research Center for Human Disease modeling, Graduate School of Medical Sciences, Kyushu University or LSR Fortessa X-20 cell analyzer (BD Bioscience).

Induction of EAE and histological analysis

The synthetic peptide MOG_35–55 (MEVGWYRSPFSRVVHLYRNGK) was purchased from Scrum Inc. EAE was induced by subcutaneous immunization with MOG_35–55 in complete Freund’s adjuvant and intraperitoneal injection of pertussis toxin, according to the method of Langrish et al.⁴⁸. On day 0, mice were subcutaneously immunized with MOG_35–55 (200 μg per mouse) emulsified in complete Freund’s adjuvant supplemented with 500 μg per mouse of Mycobacterium tuberculosis H37RA (Difco). Pertussis toxin (500 ng per mouse, Calbiochem) was intraperitoneally injected on day 0 and day 2. Disease severity was scored based on the EAE clinical signs as follows: 0, no clinical sign; 1, tail limpness; 2, hind limb weakness; 3, hind limb paralysis; 4, fore limb weakness; 5, quadriplegia; 6, death. For myelin staining, freshly prepared spinal cords were fixed in 4% paraformaldehyde overnight, and embedded in paraffin. Tissue sections (5 μm) were rehydrated and stained with Luxol fast blue. For immunohistochemical analysis, spinal cords were fixed in 4% paraformaldehyde on ice for 1 h, embedded in Tissue-Tek OCT compound (Sakura Fineteck), and frozen at −80 °C. Cryosections (10 μm) were blocked for 1 h with PBS containing 1% BSA, and stained with FITC-conjugated anti-mouse CD3ε (145-2C11, BD Biosciences) and PE-conjugated anti-mouse CD11b (M1/70, BD Biosciences) antibodies, and 4′,6-diamidino-2-phenylindole (DAPI) (DOJINDO).

RNA-seq and bioinformatic analysis

Two and a half micrograms of total RNA was subjected to ribosomal RNA depletion using Ribo-Zero Magnetic Gold Kit (Human/Mouse/Rat) (Epicentre) followed by RNA-seq library preparation using NEBNext Ultra Directional RNA Library Prep Kit for Illumina (NEB) according to the manufacturer’s instructions. Each library was quantified using Kapa Library Quantitation Kit (Kappa) and sequenced on Illumina HiSeq. 2500 to generate 101-bp paired-end reads. Raw FASTQ reads were trimmed for adapter sequences and low-quality reads using Trimmomatic version 0.32⁴⁹. Trimmed reads were then aligned to the mouse reference genome mm9 and the transcriptome defined by the mm9 genes.gtf table obtained from the UCSC Genome Browser using TopHat version 2.0.13⁵⁰. Mapped reads were assigned to all exons using featureCounts⁵¹. The differential analysis was conducted using the Bioconductor package edgeR⁵², applying trimmed Mean of M-values library normalization. A gene was defined as differentially expressed, if the false discovery rate (FDR) corrected with Benjamini-Hochberg method was less than 0.001 and if the log2 fold change (log2FC) was more than 1 (up-regulated) or less than −1 (down-regulated). Differentially expressed genes were subjected to pathway enrichment analysis using Ingenuity Pathway Analysis (Qiagen). Over-represented pathways were defined as those with Benjamini-Hochberg–corrected P values of Fisher’s exact test less than 0.01. The same gene set was also analyzed using ConsensusPathDB⁵³, which enables an integrative usage of widely used pathway databases, namely KEGG, Pathway Interaction Database, Reactome, and Wikipathways. To exploit the latest human pathway datasets (Release 32, 2017) of ConsensusPathDB, human orthologues of the differentially expressed genes were identified with HomoloGene (https://www.ncbi.nlm.nih.gov/homologene) and Mouse Genome Informatics (www.informatics.jax.org) and used as a query to find enriched pathways. FPKM values were used for unsupervised hierarchical clustering based on the Spearman correlation distance and the Ward’s linkage clustering algorithm with the hclust function implemented in R software.

Cell culture

The RAW264.7 mouse macrophage-like cells, the HEK293T human embryonic kidney cells, and the NIH3T3 mouse fibroblasts were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 µg/ml). The BW5147α⁻β⁻ mouse thymoma cells were cultured in RPMI medium supplemented with 10% fetal bovine serum, penicillin (100 units/ml), and streptomycin (100 µg/ml).

Mass spectrometric analysis of IκBζ-binding proteins

Modified RAW264.7 cells expressing FLAG–IκBζ under the control of the zinc-inducible sheep metallothionein Ia promoter¹¹ were stimulated with 50 nM ZnSO₄ for 3 h and subsequently with LPS (100 ng/ml) for 2 h. Nuclei prepared from the stimulated cells were sonicated in a buffer containing 140 mM KCl, 0.5% NP-40, 5% glycerol, and 20 mM HEPES, pH 8.0; the sonicates were subjected to immunoprecipitation using anti-FLAG (M2) antibody-conjugated agarose (Sigma). After washing with PBS (137 mM NaCl, 2.68 mM KCl, 8.1 mM Na₂HPO₄, and 1.47 mM KH₂PO₄, pH 7.4) containing 0.1% Triton X-100, the precipitated proteins were separated by SDS–PAGE (10%), followed by staining with Coomassie Brilliant Blue (CBB). Protein identification using liquid chromatography and tandem mass spectrometry (LC-MS/MS) was performed at the Laboratory for Technical Support, Medical Institute of Bioregulation, Kyushu University. Bands separated by SDS–PAGE were excised from the gel, and subjected to reduction with DTT, S-carbamidomethylation with iodoacetamide, and in-gel digestion with trypsin. Fragmented peptides were analyzed by LC-MS/MS as previously described⁵⁴. The analysis reproducibly identified JunB in addition to the NF-κB p50 subunit, a well-known partner of IκBζ^10–12.

An in vitro pull-down binding assay

Maltose-binding protein (MBP)-fusion proteins were expressed in E. coli BL21 strain and purified using Amylose Resin (New England Biolabs). FLAG-tagged proteins were synthesized in vitro using TNT^® T7 Quick for PCR (Promega). The MBP-fusion protein bound to the resin was mixed with the FLAG-tagged protein in a binding buffer (150 mM NaCl, 5% glycerol, 1 mM DTT, 1 mM EDTA and 25 mM Tris-Cl, pH 8.0) and incubated for 1 h at 4 °C. After washing with the binding buffer containing 0.05% Triton X-100, proteins were eluted with 20 mM maltose. The eluate was subjected to SDS–PAGE, followed by staining with CBB or by immunoblot analysis with the anti-FLAG antibody (M2, Sigma).

Activation of CD4⁺ T cells prepared from OT-II TCR transgenic mice

CD4⁺ T cells prepared from OT-II TCR transgenic mice, which express TCR specific for chicken ovalbumin (OVA)^22,55 were co-cultured for 3 days with the OVA_323–339 peptide (25 μg/ml) in the presence of APCs; APCs were prepared from splenocytes by depletion of T cells using CD90.2 MicroBeads (Miltenyi Biotec), followed by γ-irradiation (30 Gy).

Retroviral transduction

cDNAs for RORγt, IκBζ, JunB, c-Jun, and JunD were cloned into the bicistronic retroviral vector pMXs-IG⁵⁶, which expresses EGFP under the control of an internal ribosomal entry site. The Plat-E packaging cells⁵⁷ were transfected with the retroviral vector and cultured for 2 days. The culture supernatant was filtrated using a surfactant-free cellulose acetate filter (0.45-μm, Sartorius). Activated CD4⁺ T cells were infected with the retroviral supernatant on days 1, 2 and 3 by centrifugation at 650 × g for 2 h at 32 °C in the presence of polybrene (10 μg/ml). On day 4, IL-17 production was measured by intracellular cytokine staining.

Co-immunoprecipitation and immunoblot analysis

HEK293T cells were transfected with pcDNA3 encoding FLAG-tagged Jun family proteins and haemagglutinin (HA)-tagged BATF. After 24 h, the FLAG-tagged protein was immunoprecipitated using anti-FLAG (M2)-conjugated agarose (Sigma) and analyzed by immunoblot with the anti-FLAG (M2, Sigma) and anti-HA (16B12, Covance) antibodies. Endogenous proteins in T cell lysates were analyzed by immunoblot with anti-IRF4 (sc-6059, Santa Cruz), anti-BATF (sc-100974, Santa Cruz), anti-JunB (sc-46, Santa Cruz), anti-c-Jun (sc-1694 and sc-74543, Santa Cruz), anti-JunD (sc-74, Santa Cruz), and anti-β-actin (sc-47778, Santa Cruz) antibodies.

Estimation of relative amounts for endogenous Jun family proteins in Th17 cells

HEK293T cells were transfected with pcDNA3 encoding FLAG-tagged JunB, c-Jun, or JunD to express substantial amounts of these proteins. FLAG–JunB, FLAG–c-Jun or FLAG–JunD in the cell lysates were used as standard proteins in immunoblot analysis for estimation of relative amounts of endogenous JunB, c-Jun, and JunD in Th17 cells. The same amounts of FLAG–JunB, FLAG–c-Jun or FLAG–JunD, estimated by immunoblot using the anti-FLAG antibody (M2, Sigma), were serially diluted and subjected to immunoblot analysis; in the analysis, antibodies specific for JunB (sc-46, Santa Cruz), c-Jun (sc-74543, Santa Cruz), or JunD (sc-74, Santa Cruz) were used for estimation of the corresponding endogenous Jun protein in the Th17 cell lysate (see Fig. 4a). These primary antibodies were probed with the fluorescently labeled secondary antibody, anti-mouse IRDye800CW-conjugated (LI-COR Biosciences). Detection and quantification were performed using the Odyssey Infrared Imaging System (LI-COR Biosciences)^58,59. Standard curves were generated by plotting the band intensities against the amounts of FLAG-tagged Jun-family proteins, and used for calculation of the relative expression level of endogenous JunB, c-Jun, and JunD in Th17 cells (see supplemental Fig. 5).

Quantitative reverse transcription (RT)–PCR and ELISA

Extraction of total RNA, reverse transcription and real-time quantitative PCR were performed as described previously¹¹. The sequences of the primers used for quantitative PCR are as follows: Hprt, 5′-GCAGTACAGCCCCAAAATGG-3′ and 5′-AACAAAGTCTGGCCTGTATCCAA-3′; Il17a, 5′-CTGGAGGATAACACTGTGAGAG-3′ and 5′-TGCTGAATGGCGACGGAGTTC-3′; Il17f, 5′-CAAAACCAGGGCATTTCTGT-3′ and 5′-ATGGTGCTGTCTTCCTGACC-3′; Il21, 5′-AGCCCCCAAGGGCCAGATCGC-3′ and 5′-AGCTGCATGCTCACAGTGCCCCTTT-3′; Il23r, 5′-CCCAGACAGTTTCCCAGGTTACAGC-3′ and 5′-TGGCCAAGAAGACCATTCCCGACA-3′; Rorc, 5′-CCGCTGAGAGGGCTTCAC-3′ and 5′-TCCCCACAGATCTTGCAAGG-3′; Rora, 5′-GAACACCTTGCCCAGAACAT-3′ and 5′-AGCTGCCACATCACCTCTCT-3′; Foxp3, 5′-GGCGAAAGTGGCAGAGAGG-3′ and 5′-AAGGCAGAGTCAGGAGAAGTTG-3′; Batf, 5′-CCAGAAGAGCCGACAGAGAC-3′ and 5′-GAGCTGCGTTCTGTTTCTCC-3′; Jun, 5′-GCCTACGGCTACAGTAACCC-3′ and 5′-GAGAAGGTCCGAGTTCTTGG-3′; Junb, 5′-CTGAAACCCACCTTGGCGC-3′ and 5′-CCGAAAAGTAGCTGCCTGCC-3′; Fosl2, 5′-GACTCCCAGGTGGGAGTGC-3′ and 5′-CAGATAGGAGGAGTCTGAGG-3′; Acp5, 5′-CTGCTTGTCCGCTAACGGG-3′ and 5′-TCTGCCAGAGTACCAGGGC-3′; Mmp9, 5′-CTCTCTACTGGGCGTTAGGG-3′ and 5′-AGAGGAGTCTGGGGTCTGG-3′; Defb4, 5′-AACCAGCAAGATGAATAAATTGC-3′ and 5′-TGTGCATCCCCTAGAACTGG-3′; Il23a, 5′-GGGGAACATTATACTTTCCTGG-3′ and 5′-CTAGATTCTGTTAGAACTGAGG-3′; S100a9, 5′-ATCTTTGCCTGTCATGAGAAGC-3′ and 5′-GTTGGGCAGCTGTCACATGG-3′; Il19, 5′-GCCAACTCTTTCCTCTGCGT-3′ and 5′-GGTGGCTTCCTGACTGCAGT-3′; and Il24, 5′-TTGGATGCTCCGACTGACC-3′ and 5′-TACGGTTTCCAGGGAAGGTG-3′. Expression level of mRNA was normalized to that of Hrpt. For ELISA, CD4⁺ T cells cultured under Th0-, Th1-, Th2-, and Th17-polarizing conditions were stimulated for 24 h with PMA (50 ng/ml) and ionomycin (500 ng/ml). Protein concentrations for IL-17A, IFN-γ and IL-4 in the culture supernatant were measured using ELISA MAXTM Deluxe Set Mouse IL-17A, IFN-γ, and IL-4 (BioLegend), respectively.

Imiquimod-induced model of psoriasis-like skin inflammation

Psoriasis-like dermatitis was induced by treatment of mice with imiquimod, according to the method by Yoshiki et al.⁶⁰. Mice were topically treated with a daily dose of 50 mg of imiquimod cream (5%) (Mochida Pharmaceutical) on one ear for 5 consecutive days. Severity of the dermatitis was quantified by the extent of ear swelling and by the induction level of psoriasis-associated genes in samples obtained by ear biopsy. Ear thickness was measured daily with a Mitutoyo digmatic micrometer (Mitsutoyo). For estimation of psoriasis-associated gene expression, total RNA was extracted with TRIsure (BIOLINE) using Micro Smash MS-100 (TOMY), followed by RT-qPCR analysis. Expression level of mRNA was normalized to that of Hrpt.

Luciferase reporter Assay

A genomic DNA fragment corresponding to the 5′-upstream region of murine Il17a (−6,971/+56), Il17f (−2,107/+71), or Il23r (−1,540/+75) was amplified by PCR and inserted into pGL3-basic vector (Promega). HEK293T, NIH3T3, or BW5147α⁻β⁻ cells were transfected using X-tremeGENE HP DNA transfection Reagent (Roche) with pcDNA3 encoding a FLAG-tagged Jun family proteins (c-Jun, JunB, or JunD), pcDNA3-FLAG–BATF, the reporter plasmid containing the promoter of Il17a, Il17f, or, Il23r, and the internal control plasmid pRL-TK (Promega). After cultured for 18 h (NIH3T3 and BW5147α⁻β⁻ cells) or 24 h (HEK293T cells), luciferase activities were measured using the Dual-Luciferase Reporter Assay System (Promega).

Chromatin immunoprecipitation (ChIP)

ChIP assays were performed as previously described¹¹. Briefly, formaldehyde-fixed chromatin was subjected to sonication to obtain DNA fragments ranging in size from 100 to 400 bp. Chromatin fragments were immunoprecipitated using anti-acetylated histone H3 antibody (Millipore, #06-599), or anti-acetylated histone H4 antisera (Millipore, #06-866), or rabbit control Immunoglobulin (DAKO). After reversal of formaldehyde crosslinks, precipitated DNA was analyzed by quantitative real-time PCR using primers as follows: Il17a promoter, 5′-CACCTCACACGAGGCACAAG-3′ and 5′-ATGTTTGCGCGTCCTGATC-3′; Il17f promoter, 5′-GGCTGCTTCTTCCCTCAGG-3′ and 5′-TAAAACTGACAGGTACTACTGC-3′; Il21 promoter, 5′-CTGCAATGGGAGGGCTTGG-3′ and 5′-CTTCAACCTGACTGTGCACAG-3′; Il23r promoter, 5′-CAAGAGTCCTTAAAACCCACC-3′ and 5′-CATGGGAAGTGGCATTATTAGG-3′; Rorc (RORγt) promoter, 5′-CAGAAACACTGGGGGAGAGC-3′ and 5′-ACACAGCTGGCAGTGGAGG-3′; Rora (RORα isoform 4) promoter, 5′-GCAAGGCAGAGAGCTTCCG-3′ and 5′- CACCAAAGTCCCTCGCCAC-3′; Il5 3′-end, 5′-ATGAGAGGATGAATGAATGAATG-3′ and 5′-AGCTCTTCATCCTTGTACAGC-3′; Actb promoter, 5′-CTGTGGCGTCCTATAAAACCC-3′ and 5′-CGAAGGAGCTGCAAAGAAGC-3′.

Osteoclast differentiation

Differentiation of osteoclasts from bone marrow cells was performed as reported by Matsubara et al.⁶¹. Bone marrow cells were cultured in the presence of M-CSF (10 ng/ml) and sRANKL (50 ng/ml) for 5 days, and analyzed by a tartrate-resistant acid phosphatase (TRAP) assay using Acid Phosphatase, Leukocyte (TRAP) kit (Sigma) or by qPCR for quantifying expression of the osteoclast marker genes Acp5 and Mmp9 ²⁰.

Preparation of lymphocytes from short intestinal lamina propria

Lymphocytes were prepared from short intestinal lamina propria as described by Satoh-Takayama et al.⁶² with minor modifications. Briefly, after removal of epithelia with 1.0 mM EDTA, fragments of the small intestine were digested with collagenase (1 mg/ml, Wako). The tissues suspended in a 40% Percoll solution (GE Healthcare) were underlaid with an 80% Percoll solution. Following centrifugation for 20 min at 880 × g, cells at the 40%–80% interface were harvested.

Knockdown using short interfering RNA (siRNA)

Knowkdown using siRNA in CD4⁺ T cells were performed as described by Ciofani et al.²⁷. CD4⁺ T cells from wild-type mice were cultured for 16 h under Th0 conditions, and 300 pmol of control or c-Jun siRNA (Invitrogen) was introduced using the Mouse T cell Nucleofector Kit (Lonza) according the manufacture’s instruction. After recovery by incubation in fully supplemented culture media (Lonza), the cells were cultured for 3 days under Th17 conditions. Sequences for c-Jun siRNAs are as follows: #1, 5′-UACUGUAGCCGUAGGCACCGCUCUC-3′; #2, 5′-AUGACUUUCUGCUUAAGCUGUGCCA-3′; #3, 5′-AAACGUUUGCAACUGCUGCGUUAGC-3′.

We thank Masako Suzuki, Yoko Kage, and Namiko Kubo for technical assistance, Drs Naoko Satoh and Hiroshi Ohno (RIKEN) for advise on preparation of lamina propria lymphocytes from mouse short intestine, and Dr. Riko Nishimura (Osaka University) for advise on an osteoclast differentiation experiment; and technical supports from Research Support Center, Research Center for Human Disease Modeling, Kyushu University Graduate School of Medical Sciences, from the Laboratory for Technical Support, Medical Institute of Bioregulation, Kyushu University, and from Advanced Medical Research Center, Toho University Graduate School of Medicine. MS12 cells were a generous gift from the Meiji Dairies Corporation. This work was supported in part by grants from MEXT (the Ministry of Education, Culture, Sports, Science and Technology), from JSPS (Japan Society for the Promotion of Science: a Grant-in-Aid for Scientific Research on Innovative Areas “Oxygen Biology: a new criterion for integrated understanding of life” No. 26111009), from a Grant-in Aid for Scientific Research (C) (23590338 to S.Y.), from the Research Promotion Grant from Toho University Graduate School of Medicine (No.17-02 to S.Y.), from Private University Research Branding project (to H.N.) from the MEXT, and from the Naito Foundation; and by the Platform Project for Supporting in Drug Discovery and Life Science Research (Platform for Drug Discovery, Informatics, and Structural Life Science) from AMED (Japan Agency for Medical Research and Development).