IκB_NS Regulates Murine Th17 Differentiation during Gut Inflammation and Infection

The identification of the proinflammatory Th17 cell subset marked a significant breakthrough in our understanding of the Th cell compartment (1, 2), expanding the established Th1/Th2 paradigm (3). Previously, only the regulation of cellular immune responses against viruses by Th1 cells and humoral responses by Th2 cells in response to extracellular pathogens were known (3). Today, we know that Th17 cells are essential for host protection against a variety of extracellular bacteria and fungi located predominantly in the gastrointestinal tract, skin, and lungs. Consequently, most Th17 cells are found in these mucosal tissues upon infection; however, some reside in the lamina propria (LP) under steady-state conditions (4, 5). Th17 cells depend on the transcription factor RORγt (6) and primarily produce the cytokine IL-17A, which is a potent inducer of other proinflammatory cytokines, such as IL-6, IL-1β, and TNF-α, in a variety of IL-17R–expressing cell types, including endothelial cells, epithelial cells, and macrophages (7, 8). IL-17A is also involved in the recruitment and activation of neutrophils to sites of infection (7). Moreover, Th17 cells secrete GM-CSF, which controls granulocyte and macrophage maturation. Expression of GM-CSF was reported to be responsible for the pathogenicity of Th17 cells during experimental autoimmune encephalomyelitis (EAE) (9, 10), the murine disease model for multiple sclerosis. Of note, IL-17–producing T cells can be generated in vitro by treating naive CD4⁺ cells with TGF-β and IL-6 (11, 12) because both cytokines induce the expression of RORγt (6).

NF-κB is an inducible transcription factor and is crucial for lymphocyte development and function (13). In T cells, activation of NF-κB signaling is governed by the CARMA1, MALT1, and BCL10 complex that activates the inhibitor of κB kinase (IKK) complex (14). The IKK complex phosphorylates IκBα, which binds to NF-κB transcription factors and causes cytoplasmic retention of NF-κB by blocking the nuclear-localization signal (15). After phosphorylation by IKKβ, IκBα becomes polyubiquitinated and, finally, undergoes proteasomal degradation (13). Consequently, NF-κB is released, enters the nucleus, and regulates gene expression via binding to κB sites. NF-κB transcription factors are dimers composed of two members of the Rel-protein family, comprising p50, p52, p65/RelA, c-Rel, and RelB (13). In the nucleus, these transcription factors are regulated via atypical IκB proteins, which include IκB_NS (encoded by the gene Nfkbid), IκBζ, BCL-3, and IκBη (16).

We reported recently that the atypical NF-κB inhibitor IκB_NS regulates regulatory T cell (Treg) generation (17). Although IκB_NS^−/− mice display a severe reduction in Treg numbers, they do not develop signs of spontaneous autoimmunity (17). Thus, we speculated that IκB_NS might be important for the activation, proliferation, or cytokine production of proinflammatory effector T cells. Indeed, we observed strong expression of IκB_NS in CD4⁺ cells and strongly impaired Th17 responses in vitro and in vivo. Our results demonstrate that IκB_NS is required for Th17 differentiation and function.

B6.129/SV-NFKBID(tm1Clay) mice, later referred to as IκB_NS^−/− mice, were described previously (18). IκB_NS^−/− mice were crossed to Foxp3DTR-eGFP mice, as previously described, later referred as IκB_NS^−/−Foxp3^DTR-eGFP reporter mice (17). IκB_NS^−/−, IκB_NS^−/−Foxp3^DTR-eGFP, and B6.PL-Thy1^a/CyJ mice (Thy1.1 mice) were bred under specific pathogen–free conditions in the animal facility of the Helmholtz Centre for Infection Research. B6.129S7-Rag1^tm1Mom (RAG1^−/−) mice (19) were bred at the animal facility of the Charité – Universitätsmedizin Berlin. All animal experiments and breeding were performed in accordance with the guidelines of national and local authorities.

To induce colitis, mice were fed in four cycles, each comprising 7 d, with drinking water containing 3% (w/v) dextran sodium sulfate (DSS) as an initial dose and 2% (w/v) DSS for all further treatments, followed by 14 d of feeding with DSS-free water. Inflammation of the colon was assessed at the end of the last cycle. Body weight was measured during the progression of colitis. Consistency of the stool and rectal bleeding were scored as described previously (17). At the end of the experiment, the length of the colon was measured from the cecum to the anus. To analyze histological changes in the colon, colon samples were fixed in 4% paraformaldehyde, embedded in paraffin, and cut into 2-μm sections. The colon sections were deparaffinized and stained with H&E. Fixed colon samples were scored in a blinded manner, as described earlier (17). The histological score of DSS colitis is the sum of the individual scores for inflammatory cell infiltrations and tissue damage.

For adoptive-transfer colitis, spleen and lymph nodes were used to isolate CD4⁺CD25⁻ T cells from wild-type or IκB_NS^−/− mice or CD4⁺GFP⁺ Tregs from wild-type or IκB_NS^−/−Foxp3^DTR-eGFP reporter mice. The cells were sorted via flow cytometry (FACSAria II; BD Biosciences) or by using MACS separation kits (CD4 T Cell Isolation Kit, CD25-PE Kit; Miltenyi Biotec), as described in the supplier’s manual. A total of 200 μl 1× PBS containing 5 × 10⁵ CD4⁺CD25⁻ T cells, alone or together with 5 × 10⁵ CD4⁺GFP⁺ Tregs, were injected i.p. into RAG1^−/− mice. Body weight was monitored during colitis progression. Consistency of the stool and rectal bleeding were scored as described earlier (17). The length of the colon was measured from the cecum to the anus, and colon samples were scored in a blinded manner, as described previously (17).

Mice were infected with C. rodentium strain ICC180 expressing the lux-operon (courtesy of S. Wiles, Auckland, New Zealand). C. rodentium were cultured in Lennox broth medium (Roth) at 37°C overnight. The next day, the bacteria were adjusted to 1 × 10¹¹ CFU in 1 ml 1× PBS. A total of 1 × 10¹⁰ C. rodentium was administered to mice orally. Body weight was monitored during the progression of infection. Bacteria-derived bioluminescence during infection was measured in vivo by bioluminescence imaging using IVIS Spectrum CT. At day 10, the length of the colon was measured from the cecum to the anus. The stool was collected in 1 ml Lennox broth, weighed, homogenized, and diluted serially. To determine the bacterial load in organs, the liver and spleen were homogenized and diluted serially. Serial dilutions were added to MacConkey Agar (Roth) and cultured at 37°C for 24 h. CFU were counted and, for feces, bacterial burden was normalized to stool weight. For histology of C. rodentium–infected mice, whole-colon samples were placed in a shape of a “Swiss roll” and fixed in 4% Roti-Histofix (Roth), embedded in paraffin, cut into 8-μm sections, deparaffinized, stained with H&E, and scored in a blinded manner. The histological sections were scored for epithelial hyperplasia (score based on percentage above the height of the control where 0 = no change, 1 = 1–50%, 2 = 51–100%, and 3 = > 100%). Epithelial integrity was scored as follows: 0, no change; 1, mild epithelial ulceration and cryptic destruction; 2, moderate epithelial ulceration and cryptic destruction; and 3, severe epithelial ulceration and cryptic destruction. Tissue sections were scored for mononuclear cell infiltration as follows: 0, no change; 1, mild; 2, moderate; and 3, severe.

To isolate cells from LP, colons were placed into 15 ml ice-chilled 0.5 mM EDTA for 30 min (on ice), followed by sufficient rinsing with 1× PBS to remove residual epithelium. Then the tissue was cut into fine pieces and digested in DMEM (Life Technologies) supplemented with 1 mg/ml Collagenase D (Roche) and 0.1 mg/ml DNase I (Roche) three times for 30 min at 37°C. After each round of incubation, cells were suspended by passing them through a 100-μm mesh. After centrifugation, the cell pellet was resuspended in a 40% isotonic Percoll solution (GE Healthcare) and underlain with an 80% isotonic Percoll solution. After centrifugation (900 × g, room temperature for 20 min), LP lymphocytes (LPLs) were yielded from the interface cell ring of the 40–80% Percoll gradient. Cells were washed with PBS containing 2% FCS and used for further studies.

CD4⁺CD62L^highCD25⁻ naive T cells were sorted via flow cytometry (FACSAria II; BD Biosciences or MoFlo; Beckman and Coulter) and resuspended in primary T cell medium (IMDM containing 10% FCS, 50 μg/ml penicillin/streptomycin, 25 mM HEPES, 0.05 mM 2-ME, 1% nonessential amino acids, 1 mM sodium pyruvate). Two × 10⁵ cells were seeded per well (96-well plate) and were activated with plate-bound anti-CD3 and anti-CD28 (both from BioLegend) in the presence of priming cytokines and inhibitory Abs, according to the respective Th subset: Th0- 2 μg/ml anti-CD3 (145-2C11; BioLegend), 2 μg/ml anti-CD28 (37.51; BioLegend), 10 μg/ml anti–IL-4 (11B11; made in-house), and 10 μg/ml anti–IFN-γ (XMG1.2; made in-house); Th1- 2 μg/ml anti-CD3 (145-2C11; BioLegend), 2 μg/ml anti-CD28 (37.51; BioLegend), 10 μg/ml anti–IL-4 (11B11; made in-house), and 10 ng/ml IL-12 (R&D Systems); and Th17- 3 μg/ml anti-CD3 (145-2C11; BioLegend), 5 μg/ml anti-CD28 (37.51; BioLegend), 10 μg/ml anti–IL-2 (JES6-1A12; BioLegend), 10 μg/ml anti–IFN-γ (XMG1.2; made in-house), 2 ng/ml porcine TGF-β (R&D Systems), 30 ng/ml IL-6 (R&D Systems), 10 ng/ml IL-1β (R&D Systems), and 20 ng/ml TNF-α (PeproTech). Cells were analyzed on day 4.

After sorting of CD4⁺CD25⁻ T cells, 4 × 10⁶ cells were expanded in RPMI 1640 supplemented with 10% FCS, 1 mM sodium pyruvate, 1× nonessential amino acids, 50 ng/ml penicillin/streptomycin, and 50 μM 2-ME. For expansion, cells were stimulated in a six-well plate with 1 μg/ml plate-bound anti-CD3 and 2 μg/ml soluble anti-CD28 in the presence of 10 ng/ml murine IL-2. On day 3, cells were transferred into a 10-cm dish in a total volume of 15 ml fresh RPMI 1640 with supplements. On day 6, 5 × 10⁶ to 10 × 10⁶ cells were seeded on a 12-well plate and centrifuged for 2 min at 800 rpm at room temperature. Cells were stimulated with coated anti-CD3 (10 μg/ml) plus anti-CD28 (5 μg/ml) for 15 or 30 min or for 1, 2, or 4 h; stimulated with PMA (10 ng/ml) plus ionomycin (1 μM) for 2 h; or left untreated. Subsequently, cells were analyzed by Western blotting.

For proliferation analysis, sorted naive T cells were stained with CFSE (eBioscience). Briefly, T cells were washed two times with 1× PBS and resuspended in 1× PBS to a concentration of 2 × 10⁶/ml. CFSE was added to a final concentration of 5 μM. Cells were incubated for 10 min (CFSE) at 37°C in the dark and cultured under Th0-, Th1-, or Th17-polarizing conditions. On day 4, cells were stained for surface markers and intracellular proteins and analyzed by flow cytometry.

Naive T cells were isolated and cultured under Th17-polarizing conditions, as described above. On day 4, cells were stimulated with PMA (10 ng/ml) and ionomycin (1 μM; both from Sigma-Aldrich) for 3 h at 37°C. A mouse IL-17 Secretion Assay (Miltenyi Biotec) was used for enrichment of IL-17⁺ cells. The cells were incubated and enriched according to the supplier’s instruction manual, and isolated cells were used to analyze cytokine secretion.

To analyze the cytokine secretion of enriched IL-17–secreting cells by Proteome Profiler Array, 0.8 × 10⁶ cells were stimulated with PMA (10 ng/ml) and ionomycin (1 μM; both from Sigma-Aldrich). After 4 h, the supernatant was taken and used undiluted or diluted 1:10 or 1:50 for cytokine analysis using the Proteome Profiler Array (R&D Systems). The array membrane was activated and thereafter incubated with the Th17 supernatant, as described in the user manual. Finally, the chemiluminescence of the membrane was measured using the Fusion FX-7 camera (Vilber Lourmat), and the volume of bands was quantified with the program BIO-1D (Peqlab).

To analyze the cytokine secretion of Th17 cells by flow cytometry, 0.8 × 10⁶ enriched IL-17–secreting cells were stimulated with PMA (10 ng/ml) and ionomycin (1 μM; both from Sigma-Aldrich). After 4 h, the supernatant was used for cytokine analysis with a FlowCytomix kit (eBioscience). The supernatant was used undiluted or was diluted 1:10 or 1:50. The standard curves and samples were prepared according to the manufacturer’s instructions and measured by flow cytometry in the BD LSR II Flow Cytometer System or the BD LSRFortessa (BD Biosciences).

For intracellular cytokine staining, cells were stimulated for 4 h with PMA (10 ng/ml) and ionomycin (1 μM; both from Sigma-Aldrich). Brefeldin A (10 μg/ml; Sigma-Aldrich) was added during the last 2 h of stimulation. Live/dead staining was performed in 100 μl 1× PBS (Invitrogen) containing one of the LIVE/DEAD dyes from Life Technologies (30 min, 4°C, in the dark). Surface markers were incubated in FACS buffer (1× PBS, 2% w/v BSA, 0.01% v/v NaN₃) at 4°C, for 15 to 20 min in the dark, using the following fluorochrome-conjugated Abs: CD3–allophycocyanin–eFluor 780 (17A2; eBioscience), CD3-FITC (145-2C11; BD), CD3-PECy7 (145-2C11; eBioscience), CD4–Alexa Fluor 488 (RM4-5; BD), CD4-allophycocyanin (RM4-5; eBioscience), CD4-FITC (RM4-5; eBioscience), CD4–Pacific Blue (RM4-5; BioLegend), CD4-PE (RM4-5; BD), CD4–PerCP–Cy5.5 (RM4-5; eBioscience), CD8-allophycocyanin (53-6.7; BD), CD8-FITC (53-6.7; BD), CD25-allophycocyanin (PC61; BD), and CD62L-PerCPCy5.5 (MEL-14; eBioscience). Subsequently, the cells were fixed and permeabilized using the FoxP3 Staining Buffer Set (Miltenyi Biotec) for 30 min (4°C, in the dark). Intracellular cytokine staining was performed in 100 μl 1× permeabilization buffer using IFN-γ–FITC (XMG1.2), IFN-γ–PE (XMG1.2), IFN-γ–PECy7 (XMG1.2), and IL-17A–allophycocyanin (eBio17B7; all from eBioscience) Abs (4°C, for 30 min in the dark). Labeled cells were analyzed with a BD LSR II Flow Cytometer System or BD LSRFortessa (BD Biosciences).

For flow cytometric analysis upon C. rodentium infection, cells from the spleen and mesenteric lymph nodes, as well as LPLs, were restimulated for 6 h in medium containing IL-23 (20 ng/ml) with PMA (10 ng/ml) and ionomycin (1 μg/ml; both from Sigma-Aldrich). Brefeldin A (10 μg/ml; Sigma-Aldrich) was added for the last 3 h. Cells were processed for flow cytometry, as described above, and an Fc-block step (10 min, on ice) was included before cell surface staining and the fixation/permeabilization buffers (eBioscience) were used.

Total-cell lysates were obtained by lysing in TPNE lysis buffer (1× PBS, 300 mM NaCl, 2 mM EDTA, 1% v/v Triton X-100) supplemented with 1 mM PMSF, 1 μg/ml protease inhibitor mix (aprotinin, leupeptin, Pepstatin A, chymostatin), and 0.4 mM sodium orthovanadate. Incubation for 20 min on ice was followed by centrifugation (14,000 rpm, 4°C, 15 min). The protein concentration of the lysate was determined by a BCA Protein Assay (Thermo Scientific), following the manufacturer’s instructions.

For protein separation via SDS-PAGE, 20–40 μg the protein lysate was loaded onto a gel. Proteins were separated on a 12% polyacrylamide gel. For further analysis, the proteins were blotted onto a polyvinylidene difluoride membrane (GE Healthcare) and detected with the following Abs: β-actin (AC-74; Sigma-Aldrich), c-Rel (290512; R&D Systems), ERK (9102; Cell Signaling Technology); IκB_NS (rabbit; made in-house); IκBα (C-21; Santa Cruz), p38 (9212; Cell Signaling Technology), p50/105 (E381; Epitomics), p52/100 (4882; Cell Signaling Technology), p65 (C-20; Santa Cruz Biotechnology), p-Akt (4060; Cell Signaling Technology), p-ERK (4370; Cell Signaling Technology), p-IκBα (2859; Cell Signaling Technology), p-p38 (9211; Cell Signaling Technology), p–SAPK-JNK (9251; Cell Signaling Technology); p-Tyrosine (4G10; kind gift of Prof. B. Schraven, Otto-von-Guericke University), and RelB (C-19; Santa Cruz Biotechnology).

For chromatin immunoprecipitation (ChIP) analyses, Th17 cells were generated and stimulated with PMA/ionomycin for 4 h, as described above. ChIP was performed with a ChIP-IT Express Enzymatic kit (Active Motif), using a modified procedure to achieve ChIP without cross-linking. Hence, procedural steps to remove paraformaldehyde and to reverse cross-linking were not performed. Rabbit polyclonal IκB_NS Ab was used; after elution of the precipitated DNA, IL-10 was amplified by real-time PCR.

cDNA was used as a template for real-time PCR using SYBR Green (Roche). Ubiquitin C (UBC) was used as a housekeeping gene for normalization. Measurements were run in duplicates in the LightCycler 480 system using the following primers: GM-CSF forward (fwd), 5′-GCATGTAGAGGCCATCAAAGA-3′ and GM-CSF reverse (rev), 5′-CGGGTCTGCACACATGTTA-3′; IFN-γ fwd, 5′-ATCTGGAGGAACTGGCAAAA-3′ and IFN-γ rev, 5′-TTCAAGACTTCAAAGAGTCTGAGGTA-3′; IκB_NS fwd, 5′-GGGCTCTTTTCCCATTCTCT-3′ and IκB_NS rev, 5′-GGACACAATCCAGCCTGTCT-3′; IL-2 fwd, 5′-CCTGAGCAGGATGGAGAATTACA-3′ and IL-2 rev, 5′-TCCAGAACATGCCGCAGAG-3′; IL-10 fwd, 5′-TGCCAAGCCTTATCGGAAATG-3′ and IL-10 rev, 5′-CCCAGGGAATTCAAATGCTCC-3′; IL-17a fwd, 5′-CAGGGAGAGCTTCATCTGTGT-3′ and IL-17a rev, 5′-GCTGAGCTTTGAGGGATGAT-3′; IL-17f fwd, 5′-CAAAACCAGGGCATTTCTGT-3′ and IL-17f rev, 5′-ATGGTGCTGTCTTCCTGACC-3′; IL-22 fwd, 5′-TCATCGGGGAGAAACTGTTC-3′ and IL-22 rev, 5′-CATGTAGGGCTGGAACCTGT-3′; MIP-1α fwd, 5′-ATGAAGGTCTCCACCACTG-3′ and MIP-1α rev, 5′-GCATTCAGTTCCAGGTCA-3′; RORγt fwd, 5′-TGCAAGACTCATCGACAAGG-3′ and ROR γt rev, 5′-AGGGGATTCAACATCAGTGC-3′; and UBC fwd, 5′-AAGAGAATCCACAAGGAATTGAATG-3′ and UBC rev, 5′-CAACAGGACCTGCTGAACACTG-3′.

GraphPad Prism software (GraphPad) was used for all statistical analyses. To determine statistical significance, the two-tailed Mann-Whitney U test or Student t test was used; error bars represent the SEM.

In response to intestinal infections and inflammation, Th17 cells accumulate in the gut and surrounding lymphoid tissues (20, 21). To unveil whether IκB_NS deficiency affects Th cell differentiation, we induced chronic colitis that is promoted by Th1 and Th17, as well as B cells (22), in IκB_NS^−/− and wild-type mice by repeated DSS feeding. Although DSS-fed IκB_NS^−/− and wild-type mice did not display marked histological differences, there was inherent colon damage in the nontreated IκB_NS^−/− mice that was not present in wild-type mice (Supplemental Fig. 1A–D). Most importantly, the frequency of IL-17A⁺ T cells was decreased in LPLs of IκB_NS^−/− mice compared with wild-type mice (Supplemental Fig. 1E, 1F). Interestingly, Th17 frequencies already were reduced in nontreated, IκB_NS^−/− animals compared with respective wild-type mice (Supplemental Fig. 1E, 1F).

Next, we investigated whether Th17 differentiation also is impaired in T cell–based transfer colitis. Rag1^−/− mice receiving IκB_NS^−/− T cells exhibited significantly more colon damage than did wild-type controls (Fig. 1A). Importantly, the induction of both IL-17A⁺IFN-γ⁻ (Fig. 1B, 1C) and IL-17A⁺IFN-γ⁺ (Fig. 1D) T cells was completely blunted in IκB_NS^−/− mice. In contrast, the frequency of IL-17A⁻IFN-γ⁺ T cells (Fig. 1E) was slightly, but significantly, increased in IκB_NS^−/− mice, demonstrating the importance of IκB_NS for the formation of Th17 cells in the inflamed gut. Previously, we reported a reduced number of Tregs in IκB_NS^−/− mice (17). To analyze whether the increased frequency of Th1 cells in IκB_NS-deficient animals arose from reduced numbers of Tregs, Tregs from IκB_NS^−/−Foxp3^DTR-eGFP or IκB_NS^+/+Foxp3^DTR-eGFP reporter mice were cotransferred (Fig. 1F). The cotransfer of Tregs had no effect on the increased Th1 cell frequency in IκB_NS^−/− mice, regardless of whether the Tregs were IκB_NS deficient or proficient. In summary, both colitis models revealed a requirement for IκB_NS in Th17 cell differentiation during gut inflammation.

To determine whether IκB_NS is differentially expressed in Th cell subsets, we analyzed IκB_NS (Nfkbid) expression at the mRNA level by real-time PCR (Fig. 2). We detected comparable Nfkbid mRNA expression in unprimed T cells (Th0) and in Th1- and Th17-polarized cells upon restimulation with PMA and ionomycin (Fig. 2A). In addition, high IκB_NS expression was observed at the protein level in reactivated T cells (Fig. 2B). Similar findings were obtained when T cells were restimulated with anti-CD3 and anti-CD28 Abs (Fig. 2C, 2D). Th1 cells exhibited a higher IκB_NS protein expression than did Th17 cells. Notably, the nonactivated Th17 cells displayed higher IκB_NS expression compared with the other nonactivated cells (Fig. 2B, 2D), substantiating a pivotal role for IκB_NS specifically in this T cell subset. Indeed, during in vitro priming we monitored severely reduced frequencies of both IL-17A⁺ Th17 cells and IFN-γ⁺ Th1 cells in the absence of IκB_NS (Fig. 2E, 2F). We conclude that IκB_NS expression is not restricted to a single Th cell subset but forces the differentiation of Th17 and Th1 cells in vitro.

To identify the reason for the reduced frequency of Th17 and Th1 cells induced by IκB_NS deficiency, we first analyzed Th cell proliferation in vitro. Although all T cells proliferated equally during the first 2 d of culture, IκB_NS^−/− T cells primed for both Th17 and Th1 differentiation proliferated much less compared with the wild-type cells in the following days (Supplemental Fig. 2). To determine whether wild-type cells could compensate for the proliferation defect of IκB_NS^−/− cells when cocultured, we differentiated IκB_NS^−/− (CD90.2⁺) T cells in the presence of congenic wild-type (CD90.1⁺) cells, both stained with CFSE. Again, the frequencies of IκB_NS^−/− Th0, Th1, and Th17 cells showed a 50% reduction compared with wild-type cells, suggesting a cell-intrinsic effect of IκB_NS on T cell differentiation (Fig. 3A). The CFSE profile and mean number of divisions/cell (division index) revealed enhanced proliferation of wild-type Th0, Th1, and Th17 cells compared with IκB_NS^−/− cells (Fig. 3B, 3C). Remarkably, the proliferation defect caused by IκB_NS deficiency was most pronounced in Th17-differentiated cells. Of note, the proliferation defect was not due to impaired proximal TCR signaling because tyrosine phosphorylation; the activation of the kinases Akt, ERK, and p38; expression of the NF-κB family members c-Rel, NF-κB1, NF-κB2, p65, and RelB; and the expression and phosphorylation status of IκBα were unaltered (Supplemental Fig. 3).

Because RORγt is an essential transcription factor for Th17 differentiation and function (6), we analyzed the expression of Rorc (encoding RORγt) in IκB_NS^−/− and wild-type Th1 and Th17 cells by real-time PCR. Despite defective Th17 differentiation of IκB_NS^−/− cells, Rorc expression was comparable in wild-type and IκB_NS^−/− Th17 cells (Fig. 4A). Next, we analyzed whether IκB_NS^−/− Th17 cells display an altered cytokine expression profile. Coculturing wild-type and IκB_NS^−/− T cells under Th17-polarizing conditions, we observed, in addition to the proliferation defect (Fig. 3), a significantly decreased mean fluorescence intensity (MFI) of IL-17A in IκB_NS^−/− cells (Fig. 4B, 4C), indicating a role for IκB_NS in Th17 cell function. To determine whether IκB_NS also impairs the expression of other cytokines, a cytokine array membrane was performed. We enriched in vitro–generated wild-type and IκB_NS^−/− Th17 cells to obtain comparable numbers of IL-17A–secreting cells. The results revealed reduced secretion of IL-2, IL-10, GM-CSF (Csf2), and MIP-1α (Ccl3) by IκB_NS^−/− Th17 cells (Fig. 4D, Supplemental Fig. 4). Using a bead-based assay and flow cytometry, the reduced secretion of the cytokines IL-2, IL-10, and MIP-1α in IκB_NS^−/− Th17 cells was verified (Fig. 4E). Nevertheless, secretion of the Th17 cytokines IL-17A, IL-17F, and IL-22 was not affected by IκB_NS deficiency. Furthermore, we analyzed the effect of IκB_NS on mRNA expression of the respective cytokines in stimulated Th17 cells by real-time PCR. In accordance with the protein data, we detected reduced mRNA expression of Il10 and Csf2 in IκB_NS^−/− Th17 cells, whereas Il17a and Il17f expression was not affected (Fig. 4F). Interestingly, in contrast to IL-22 cytokine secretion, IκB_NS^−/− Th17 cells showed an increased expression of IL-22 mRNA. We conclude that IκB_NS regulates the expression of multiple key Th17 cytokines.

C. rodentium is a noninvasive pathogen that establishes acute infections in the murine large intestine (23). The adaptive immune response plays an important role for the host response to C. rodentium because mice lacking B and T cells are unable to clear the infection (24). C. rodentium induces a strong Th17 response (25, 26). Therefore, we examined whether IκB_NS is required for an appropriate host response against C. rodentium.

The bacterial burden in feces of IκB_NS^−/− mice at day 10 postinfection was significantly increased compared with wild-type mice (Fig. 5A). A closer investigation of colonization using live in vivo imaging indicated a similar infection kinetic in mice of both genotypes until day 5 (Fig. 5B, 5C). However, although the bioluminescent signal from the gut of wild-type mice started to decrease between days 7 and 10, the signal intensity increased in IκB_NS^−/− mice, confirming impaired C. rodentium clearance in these animals (Fig. 5B, 5C). In line with this observation, IκB_NS^−/− mice also suffered from an elevated systemic infection, because dissemination of the pathogen to the liver was significantly enhanced in IκB_NS^−/− mice compared with wild-type mice (Fig. 5D). These findings indicate that IκB_NS is required for restricting C. rodentium infections to the gut.

We next studied the impact of IκB_NS deficiency on the host T cell response. We found that the frequencies of Th17 cells were reduced by ∼50% in the colonic LP and in the spleen (Fig. 6A, 6B). Interestingly, the frequency of IL-17A⁺IFN-γ⁺ double-positive cells showed a similar reduction in both organs (Fig. 6A, 6C), whereas the frequency of IFN-γ⁺ Th1 cells was slightly reduced in the spleen of IκB_NS^−/− mice but not in the other organs investigated (Fig. 6A, 6D). In addition, we examined LP cell–derived mRNA levels of the cytokines IFN-γ, IL-10, IL-17A, IL-17F, and IL-22 by real-time PCR. Consistent with the previous data, IκB_NS deficiency resulted in decreased mRNA expression of Ifng, Il17a, Il17f, and Il10 by LPLs of the colon (Fig. 6E). In contrast with the in vitro data, but consistent with the other cytokines in vivo, the expression of Il22 mRNA was slightly downregulated in IκB_NS^−/− mice (Fig. 6E).

To assess whether IκB_NS deficiency also affects pathology of the colon, we investigated cellular infiltration and the level of tissue damage in the colon of infected mice. Despite the increased bacterial burden, IκB_NS^−/− mice exhibited significantly less immunopathology in the large intestine compared with wild-type mice (Fig. 6F), suggesting that the defective Th17 immune response in these mice results in less inflammation-induced tissue destruction. Altogether, our data demonstrate that IκB_NS promotes Th17 differentiation and cytokine expression to fight intestinal infections.

The expression of several cytokines is regulated by the proteins of the IκB family (16). For instance, it was described that IκBζ and IκB_NS regulate the expression of IL-6 by binding to its promoter (27, 28). In this study, we consistently found that expression of IL-10 is affected by IκB_NS deficiency in various in vitro and in vivo experiments. To investigate whether IκB_NS regulates the expression of IL-10 directly by binding to its gene locus, we performed native ChIP experiments. We detected increased binding of IκB_NS to the Il10 gene locus in stimulated Th17 cells compared with unstimulated Th17 cells and IκB_NS-deficient controls (Fig. 7). This suggests that IL-10 is a direct IκB_NS target gene and that IκB_NS regulates Th17 differentiation, at least in part, by direct control of cytokine expression.

We recently reported that IκB_NS regulates the induction of Foxp3 expression in Tregs (17). Although IκB_NS^−/− mice displayed a severe reduction in immunosuppressive Tregs, we found no signs of autoimmune disease. Thus, we speculated that IκB_NS promotes the development or function of proinflammatory T cells. In this study, we show that IκB_NS additionally controls the proliferation and cytokine expression of differentiated CD4⁺ effector T cells. Although impaired in vitro proliferation of IκB_NS^−/− thymic and peripheral CD4⁺ and CD8⁺ cells upon anti-CD3 and anti-CD28 stimulation was reported (18), our data indicate that Th17 cells are particularly affected by IκB_NS deficiency. We further unveiled that the proliferation defect of IκB_NS^−/− T cells primed for Th17 differentiation could not be overcome by the presence of wild-type cells, indicating that IκB_NS’s function is T cell intrinsic.

Remarkably, we also detected expression of IκB_NS in acutely stimulated Th1 and Th0 cells. Overall, IκB_NS expression in Th1 and Th0 cells was comparable to the expression measured in Th17 cells. Differences between mRNA and protein expression might be due to posttranscriptional control of IκB_NS expression by Roquin, which was described previously in cell lines (29). Of note, fewer Th1 cells were generated in vitro from IκB_NS^−/− naive cells as a result of impaired proliferation. Importantly, IκB_NS does not affect proximal TCR signaling. Thus, IκB_NS forces proliferation further downstream in the signaling cascade. Noteworthy, in vivo proliferation of IκB_NS^−/− Tregs is not altered (17). Therefore, IκB_NS acts differentially in Tregs and effector T cells. For instance, IκB_NS is required for development but not function of Tregs (17), whereas in differentiated CD4⁺ Th cells, it regulates proliferation and effector cytokine production. Thus, our findings provide an explanation for our previous observation that Treg reduction does not cause an autoimmune phenotype in IκB_NS^−/− mice (17).

We (17) and other investigators (30) previously described severe pathological changes in the colon in IκB_NS^−/− mice as a result of DSS-induced and transfer colitis, respectively. In the current study, chronic DSS colitis led to a reduced number of IL-17A⁺ Th17 cells in IκB_NS^−/− mice compared with wild-type mice. Furthermore, using a T cell–dependent transfer colitis model, we demonstrated that almost no Th17 cells or IL-17A⁺ IFN-γ⁺ T cells developed from transferred IκB_NS^−/− T cells. Hence, IκB_NS is not only an important regulator of Th17 priming in vitro, it is also essential for Th17 and IL-17A⁺IFN-γ⁺ T cell development in mouse models of gut inflammation. Furthermore, tissue damage in the gut was increased in recipients of IκB_NS^−/− T cells as a result of the reduced numbers of immunosuppressive Tregs, as previously reported (17). In addition to a lower Treg frequency, the elevated frequency of IFN-γ⁺ Th1 cells is likely to be responsible for the enhanced pathological damage. Although both IκB_NS^−/− Th1 and Th17 cells show a proliferation defect in vitro, Th1 development was not affected during colon inflammation in vivo. Because cotransfer of Tregs did not affect the enhanced Th1 differentiation, this might indicate the presence of additional signals that promote Th1 development in vivo.

In addition to the autoimmune models of colitis, we addressed the question of whether IκB_NS is instrumental for Th17 responses during infection. To this end, we used C. rodentium because this murine pathogen induces strong Th17 responses (25, 26, 31). C. rodentium shares virulence factors with enteropathogenic Escherichia coli and enterohemorrhagic E. coli and, thus, is a commonly used murine model to study immune responses against these gut-tropic pathogens (32). Similar to the colitis models, IκB_NS deficiency resulted in impaired Th17 and IL-17A⁺IFN-γ⁺ T cell generation in infected mice. Because IFN-γ, as well as the p40 and the p19 subunit of IL-23, which is known to stabilize the Th17 phenotype (33), are required for C. rodentium clearance (25, 34, 35), it is conceivable that the lack of IL-17A⁺ and IFN-γ⁺ T cells in IκB_NS^−/− mice upon C. rodentium infection is responsible for the increased bacterial burden in these mice. Interestingly, NF-κB1/p50^−/− mice also are defective in C. rodentium clearance (36), which is in line with the observation that IκB_NS preferentially interacts with this NF-κB family member (17, 37). Furthermore, because of the normal IL-17A and IL-17F expression in vitro, we speculate that the reduced mRNA expression in vivo arises from the decreased frequency of Th17 and IL-17A⁺IFN-γ⁺ cells.

In addition to IL-17A and IL-17F, which were involved in late host defense against C. rodentium, the IL-10 cytokine family member IL-22 is essential in early host protection (26, 38). IL-22^−/− mice exhibit increased damage to the intestinal epithelium, systemic bacterial burden, and mortality (39). In this study, we observed an increased mRNA, but unaltered protein, expression of IL-22 by IκB_NS^−/− cells in vitro. Of note, Il22 expression was reduced in vivo upon C. rodentium infection. The reduced IL-22 levels might be sufficient to support protection of the intestinal epithelium; however, the defects in the expression of IL-22 and other cytokines were insufficient to reduce bacterial burden.

Why do we see decreased histological damage in C. rodentium–infected IκB_NS^−/− mice despite the increase in bacterial burden? Interestingly, IL-23 also was reported to induce GM-CSF expression in Th17 cells, which causes pathogenicity in EAE, the murine model for multiple sclerosis (9, 10). Thus, the reduction in GM-CSF production by IκB_NS^−/− Th17 cells may account for the reduced tissue damage in the gut. Alternatively, impaired bacterial clearance in conjunction with reduced tissue damage was reported recently (40). In this infection model, Th17 differentiation depended on the presence of Tregs and the consumption of IL-2. Thus, it appears that Th17 cells during C. rodentium infection are not only involved in pathogen clearance (i.e., beneficial), they also cause tissue damage and, therefore, can be pathogenic.

In addition to proinflammatory or pathogenic Th17 cells, protective or suppressive Th17 cells have been described (21, 41). For instance, IL-10–expressing Th17 cells that were primed with TGF-β and IL-6 were not pathogenic (41). Therefore, IL-10 coexpression by Th17 cells is presumably a self-regulatory mechanism to reduce the inflammatory damage of Th17 cells. Although it was initially believed that Th17 cells produce IL-10 early on in an infection (41), it was later shown that Th17 cells migrate from the spleen to the gut in a CCR6-dependent manner and acquire suppressive function and IL-10 expression there (21). We found decreased IL-10 expression in IκB_NS^−/− Th17 cells both when generated in vitro and in mice infected with C. rodentium. Therefore, IκB_NS is crucial for the differentiation of proinflammatory, as well as suppressive, Th17 cells.

We showed previously that IκB_NS in Tregs is essential to induce the expression of the transcription factor Foxp3 (17). IκB_NS binds together with p50 and c-Rel to CNS3 and the promoter of Foxp3. Because of the extenuated expression of IL-10 in vivo and in vitro, we analyzed the influence of IκB_NS in IL-10 expression by ChIP, which revealed binding of IκB_NS to the IL-10 gene locus. Previous studies identified the binding of NF-κB to the IL-10 gene locus (42, 43). Thus, we suggest that IκB_NS binding to the IL-10 gene locus is mediated by p50 and/or c-Rel, because IκB_NS itself has no DNA-binding domain. Although IL-10 appears to be directly regulated by IκB_NS, IL-17A, IL-17F, and IL-22 do not contain canonical NF-κB–binding sites. Consequently, we conclude that IκB_NS does not regulate these genes directly; further experiments are needed to clarify how IκB_NS regulates the expression of cytokines in an indirect manner.

Several components of the NF-κB signaling pathway are important for Th17 development. Most upstream in the signaling complex, CARMA1 deficiency results in defective Th17 cell differentiation (44). MALT1 is highly expressed in Th17 cells generated in vitro, and MALT1^−/− Th17 cells produce less IL-17 and GM-CSF (45). Loss of p65/RelA or c-Rel results in impaired Th17 induction in vitro, and c-Rel was shown to interact directly with the Rorc and Il21 promoters (46, 47). Moreover, the atypical nuclear IκB protein IκBζ, another nuclear IκB protein with high sequence similarity to IκB_NS, regulates Th17 proliferation in a cell-intrinsic way (48). IκBζ interacts with RORα and RORγt to regulate Th17 differentiation and Il17 gene expression (48). Notably, CARMA1 and c-Rel are also critical for the generation of Tregs (49, 50), indicating at least a partially overlapping developmental program (51). Thus, we believe that IκB_NS is a component of the common Th17/Treg NF-κB signaling cascade, because mice that are also deficient in c-Rel or CARMA1 fail to develop spontaneous autoimmunity (52, 53). In summary, our data reveal that IκB_NS is necessary for murine Th17 differentiation during inflammation and infection, and it belongs to the group of NF-κB–signaling proteins that commonly regulate Treg and Th17 cell differentiation. Hence, IκB_NS exhibits diverse T cell proliferation and cytokine secretion functions and may represent a specific pharmacological target in future drug development.

While this manuscript was in revision, it was reported that IκB_NS^−/− mice are less susceptible to EAE induction, with lower Th17 frequencies in draining lymph nodes and a lack of lymphocyte infiltration in spinal cord (54). Consistent with our data, the gene expression of IL-17A was suggested to be independent of IκB_NS. Furthermore, Heissmeyer and colleagues (55) demonstrated that Roquin and Regnase-1 inhibit IκB_NS expression and, thus, impair Th17 cell differentiation. Consistent with our results, knockdown of IκB_NS in T cells under Th17-priming conditions resulted in impaired Th17 differentiation. Taken together, IκB_NS appears to be a crucial regulator of Th17 differentiation.

We thank Sabrina Schumann for technical assistance, Alisha Bronietzki and Dr. Yvonne Neumann for critical reading of the manuscript, David Dettbarn and colleagues at the Helmholtz Centre for Infection Research (HZI) Animal Facility for excellent mouse stock breeding, and Dr. Marina Pils (HZI) and Elena Wiebe (Medical School Hannover) for help with the histology. We are grateful to the FACS facilities, especially Dr. Lothar Groebe (HZI) and Dr. Desiree Kunkel (Berlin-Brandenburg Centre for Regenerative Therapies–Flow Cytometry Laboratory, Berlin, Germany), for help with cell sorting. Rag1^−/− mice were kindly provided by Dr. Thomas Blankenstein (Charité – Universitätsmedizin Berlin).

The authors have no financial conflicts of interests.