Fos Proteins Suppress Dextran Sulfate Sodium-Induced Colitis through Inhibition of NF-κB

Inflammatory bowel disease (IBD), a chronic and relapsing inflammation of the gastrointestinal tract, affects millions of people worldwide. Although the causes of IBD are still unknown, genetic and environmental factors, infectious agents, impairment of local tolerance, and mucosal imbalance leading to activation of the mucosal immune system have been suggested to play roles (1). The resulting imbalance of the mucosal immune system causes the overproduction of inflammatory cytokines, reactive oxygen metabolites, and infiltration of neutrophils into the intestine, thus leading to uncontrolled intestinal inflammation and tissue damage (1). Dextran sulfate sodium (DSS) is commonly used in rodent models to chemically induce acute intestinal inflammation, and the DSS-induced colitis is characterized by weight loss, bloody diarrhea, epithelial cell damage, and immune cell infiltration, as well as an increased production of inflammatory mediators including TNF-α, IL-6, IL-12, and interferons. Two stress-responsive transcription factors—AP-1 and NF-κB—were activated to induce such inflammatory mediators (2, 3). Acute DSS-induced colitis does not require T cells or B cells because it occurs in SCID mice that lack these cells (4). Colitis may result from DSS toxicity to colonic epithelial cells.

AP-1 is a collection of dimeric transcription factors typically composed of one of four Fos family proteins (c-Fos, FosB, Fra-1, and Fra-2) and one of three Jun family proteins (c-Jun, JunB, and JunD). Unlike Fos proteins, Jun proteins can also form dimers within the family. AP-1 is one of the major regulators of bone and immune cells, among other cell types. A wide range of cellular stresses, including serum, UV, and pathogen-associated molecular patterns, such as the cell-wall component LPS of Gram-negative bacteria, sequentially induce transcription of the Fos family genes (2, 5). The activity of both newly synthesized and pre-existing AP-1 components is modulated through their posttranslational modification. Key enzymes are various MAPKs, including JNK, ERK, and p38, which phosphorylate AP-1 components and thereby regulate transcriptional activity.

c-Fos is implicated in the regulation of innate immune responses. Genes encoding proinflammatory mediators, such as TNF-α, IL-6, and IL-12, are transcriptionally activated in response to LPS or TNF-α, mainly through NF-κB activation (6). We previously reported that NF-κB activity is elevated in Fos^−/− mice, resulting in augmented inflammatory responses (7). Injection of LPS into mice lacking c-Fos results in more prominent hypothermia and bradycardia than occur in wild-type (wt) controls, owing to the elevated production of proinflammatory cytokines, especially TNF-α (7). Therefore, in contrast to NF-κB, c-Fos appears to be a negative regulator of proinflammatory responses. Fra-1 is a Fos family protein that is structurally related to, but smaller than, c-Fos and that lacks so-called transactivation domains within the molecule (8). Fra-1 and ΔFosB dramatically enhance osteoblast differentiation when overexpressed in mice (9, 10). However, roles of Fra-1 in inflammatory responses in vivo remain to be determined.

In this study, we hypothesized that Fos proteins may protect cells from excessive stress responses by suppressing NF-κB activity. Our results suggest that these Fos proteins are negative regulators of NF-κB–mediated stress responses.

Fos^−/− mice (11) were bred on a mixed background of 129 and C57BL/6J. A powder diet was provided to Fos^−/− mice. Fra-1 transgenic (Tg) mice (9) were backcrossed onto C57BL/6J mice (Crea, Tokyo, Japan) >10 generations, and both mice groups were maintained under specific pathogen-free conditions. Wt littermates were used as controls. Fra-1–conditional knockout mice on a C57BL/6J × 129 mixed background were produced by crossing Fosl1^flox/flox mice with MORE-Cre mice (12). All experiments were conducted in accordance with the institutional review board-approved protocols of Keio University.

Mice were given distilled drinking water containing 2% (wt/vol) DSS (m.w. = 36,000–50,000; MP Biomedicals, Irvine, CA) ad libitum. DSS was administered to Fos^−/− mice for 7 d and to Fra-1 Tg mice for 13 d. Bay 11-7085 solution (Biomol International, Plymouth Meeting, PA; final 4% DMSO in PBS) or solution control (mock) was injected 1 h before and each day during DSS administration by i.p. injection at a concentration of 20 mg/kg mouse body weight. The clinical scores of colitis, such as weight change, diarrhea, and colorectal bleeding, and survival were observed. The whole colon was extracted to measure its length and to examine its histological status.

The extracted colon was fixed in 4% paraformaldehyde, embedded in paraffin, and cut into 4-μm sections on slide glasses. The slides were stained with H&E (Wako, Tokyo, Japan) and analyzed under a microscope (Axiovert 135; Carl Ziess, Jena, Germany). Frozen 5-μm sections were also prepared. The histological examination of colonic tissue was performed as described (13). The tissues were scored in a blinded fashion on a scale from 0 to 4, according to intensity of lesions: 0, normal; 1, mild; 2, moderate; 3, severe; and 4, very severe. Intensity measures included the number of infiltrating cells, depth of inflammation, crypt damage, and percent involvement. The average disease activity and the total histological inflammation score are shown.

Frozen sections were stained with rabbit anti–c-Fos (Abcam, Cambridge, UK), mouse anti–Fra-1 (Santa Cruz Biotechnology, Santa Cruz, CA), or Armenian hamster anti-CD11c (Abcam) Abs, followed by incubation with biotin-labeled anti-rabbit, anti-mouse (Histofine, Nichirei Biosciences, Tokyo, Japan), or biotin-labeled anti-Armenian hamster (Abcam) Abs, respectively. The slide was developed by diaminobenzidine solution (EnVision Kit, Dako, Carpinteria, CA), counterstained by hematoxylin (Wako), and mounted with Softmount (Wako). For isotype control staining, DSS-treated colon slides were stained with the specific isotype IgG as a replacement for the first Ab, and followed general staining procedures.

Mice were sacrificed, and total RNAs were extracted from colon fragments, using ISOGEN (Nippon Gene, Toyama, Japan). cDNA was synthesized from 1–2 μg total RNA, using random hexamers (Qiagen, Valencia, CA) with PrimeScript (Takara Bio, Tokyo, Japan). Quantitative RT-PCR (qRT-PCR) was performed on an ABI Prism 7000 machine (Applied Biosystems, Foster City, CA). mRNA levels were quantified using a standard curve generated with serially diluted plasmids containing the PCR amplicon and normalized to Gapdh expression. Experiments were repeated three times with three independent cDNA syntheses.

Mouse serum and culture supernatants were collected, and protein levels of TNF-α, IL-6, IL-10, and IL-12 p40 were measured using ELISA sets (BD Biosciences, San Diego, CA). The production levels of TGF-β and IL-4 also were measured using ELISA sets (eBioscience, San Diego, CA).

NF-κB or AP-1 DNA-binding activity was assessed with EMSA, as described previously (14). Briefly, 3 μg nuclear extracts prepared from M-CSF–dependent macrophages (MDMs) and colon fragments was incubated for 30 min at 37°C with [³²P]-end-labeled 27-mer double-stranded oligonucleotide containing the κB3 binding site of mouse TNF-α promoter (5′-AGCTGAGGAGGGAAAGCCCCCTGTTTG-3′) (15) or 28-mer double-stranded oligonucleotide containing the consensus AP-1 binding site of the human collagenase promoter (16). The DNA-protein complex formed was then separated from free oligonucleotide on 6.6% native polyacrylamide gels. Dried gels were visualized using a BAS5000 imaging analyzer (Fuji Film, Tokyo, Japan), and bands were quantified using MultiGauge software (Fuji Film).

A telemetric radio transmitter (TA10ETA-F20; Data Sciences, Saint Paul, MN) was used to measure locomotor activity. A transmitter was implanted subcutaneously on the back of each mouse, as previously described (7). The locomotor activity was monitored continuously by receivers (CTR-86; Data Sciences) below the cages and recorded every 5 min by a Data Quest analysis system (Data Sciences). A stock solution of 10 mg/ml LPS from Salmonella enterica serovar Typhimurium (S. Minnesota Re595; Sigma-Aldrich, St. Louis, MO) was prepared in water at 70°C. Before the experiment, LPS was further diluted. In consideration of circadian rhythm, LPS was administered i.p. at 13:00 h for each experiment.

Bone marrow from tibiae and femora of wt and Fra-1 Tg mice (basically at 6 wk old) was harvested by flushing with α-minimal essential medium solution containing 10% FBS and penicillin-streptomycin. After passage through a 70-μm–cell strainer, cells were cultured overnight in 15-cm tissue culture dishes (Greiner, Frickenhausen, Germany). Floating cells were then collected and expanded in the presence of 10 ng/ml M-CSF (R&D Systems, Minneapolis, MN) for 3 d. These cells were used as MDMs, and were seeded at a concentration of 1 × 10⁶ cells in a 24-well plate and treated with 0.1 μg/ml LPS for 24 h. The supernatants were subjected to ELISA to analyze the expression of indicated cytokines and the production of NO. Mouse embryonic fibroblasts (MEFs) were prepared from wt and Fra-1 Tg mice, seeded at a concentration of 1 × 10⁴ cells in a 24-well plate, and treated with 0.1 μg/ml LPS for 24 h. The supernatants were then subjected to ELISA.

The concentration of stable nitrite was determined by Griess reaction, as previously described (14). The absorbance at 530 nm was determined on a plate reader (680XR; BioRad Laboratories, Hercules, CA). A sodium nitrite standard curve was generated in each experiment.

To determine protein expression levels, we prepared whole-cell extracts and fractionated them on 4–12% SDS-polyacrylamide gels (Novex; Invitrogen, Carlsbad, CA) and transferred them onto Hybond nitrocellulose membranes (Amersham, Little Chalfont, UK). The following primary Abs were used: polyclonal anti-TLR4 goat Ab (Santa Cruz Biotechnology), polyclonal anti-CD14 rabbit Ab (Santa Cruz Biotechnology), monoclonal phospho-IκBα mouse Ab (Cell Signaling Technology, Beverly, MA), polyclonal anti-IκBα rabbit Ab (Santa Cruz Biotechnology), monoclonal anti–phospho-p65 mouse Ab (Cell Signaling Technology), polyclonal anti-p65 rabbit Ab (Santa Cruz Biotechnology), monoclonal anti–phospho-JNK mouse Ab (Cell Signaling Technology), polyclonal anti-JNK rabbit Ab (Santa Cruz Biotechnology), monoclonal anti–phospho-ERK1/2 rabbit Ab (Cell Signaling Technology), polyclonal anti-ERK1/2 rabbit Ab (Santa Cruz Biotechnology), monoclonal anti–phospho-p38 mouse Ab (Cell Signaling Technology), and polyclonal anti-p38 rabbit Ab (Santa Cruz Biotechnology). Where indicated, blots were stripped and reprobed with a polyclonal anti-Actin goat Ab (Santa Cruz Biotechnology) to monitor protein loading.

All data were expressed as the mean ± SEM. The statistical significance of differences between values was evaluated by the Student t test.

We first examined the expression of stress-responsive transcription factor AP-1 in DSS-induced colitis. We administered 2% DSS in drinking water to wt mice for 7 d to induce acute intestinal inflammation. DSS induced infiltration of inflammatory cells and crypt damage in the colon, and c-Fos and Fra-1 proteins were detected by immunohistochemistry in various cell types, including infiltrating macrophages (Fig. 1A). We also found that DSS induced expression of CD11c-positive APCs or dendritic cells (Fig. 1A). Furthermore, we analyzed mRNA levels of AP-1 genes by qRT-PCR and noted that expression of all seven AP-1 family members was elevated (Fig. 1B). These data suggest that c-Fos and Fra-1 are expressed mainly in infiltrating cells, including APCs, during the course of DSS injury.

We investigated the susceptibility of Fos^−/− mice to DSS-induced colitis. We administered 2% DSS in drinking water to Fos^−/− mice and their wt littermates, and monitored clinical parameters, cytokine profiles, and NF-κB activity. Some of the Fos^−/− mice died within 6 d after DSS administration (Fig. 2A), and surviving Fos^−/− mice showed more severe clinical colitis profiles than did wt controls (Fig. 2B–D). Seven days after DSS administration, the histological status of each colon was analyzed. In the Fos^−/− mice, the percentage of colon shortening, which reflects inflammation at the site, was four times higher (Fig. 2E), and invasiveness of inflammatory lesions and crypt damage were significantly greater than in wt controls (Fig. 2F). DSS-induced cytokine production in serum was also examined. We found that production of proinflammatory cytokines such as TNF-α, IL-6, and IL-12 was significantly enhanced in DSS-treated Fos^−/− mice. Although TGF-β expression was increased, anti-inflammatory cytokines IL-4 and IL-10 were not elevated in DSS-treated Fos^−/− mice compared with wt controls (Fig. 2G). We also noted that TNF-α and IL-6 production (Supplemental Fig. 1), NF-κB–binding activity (Fig. 2H), and MAPK activity (Fig. 2I) were higher in colon fragments of DSS-treated Fos^−/− mice than in those of DSS-treated wt mice. The enhanced production of TNF-α and IL-6 in the absence of c-Fos was also observed in colon fragments treated with LPS (Supplemental Fig. 2). These results demonstrate that the loss of c-Fos protein enhances susceptibility to DSS-induced colitis.

To clarify the role of NF-κB activation in DSS-induced colitis in Fos^−/− mice, we administered the NF-κB inhibitor Bay 11-7085 along with DSS treatment, using wt and Fos^−/− mice. The Fos^−/− mice that were administered Bay 11-7085 demonstrated lower clinical colitis profiles than did mock-treated Fos^−/− mice (Fig. 3A–C). The histological status of the colon showed that treatment of Bay 11-7085 suppressed colon shortening in Fos^−/− mice (Fig. 3D), and invasiveness of inflammatory lesions and crypt damage were significantly lower than in mock-treated Fos^−/− mice (Fig. 3E). The production levels of inflammatory cytokines TNF-α and IL-6 were significantly suppressed in Fos^−/− mice administered Bay 11-7085 (Fig. 3F). Therefore, the loss of c-Fos enhances DSS-induced colitis mainly through NF-κB. Together with enhanced LPS-induced cytokine production in the absence of c-Fos (7), these results strongly suggest that c-Fos is an anti-inflammatory transcription factor in response to administration of both DSS and LPS.

We next asked whether the loss of Fra-1, another Fos family protein, also enhances inflammatory responses. Because Fra-1 knockout mice are embryonic lethal owing to a placental defect, we examined the LPS response of Fra-1–conditional knockout (Fosl1^Δ/Δ) mice, in which Fra-1 was deleted in the entire body but not in the placenta (12). We found that serum levels of TNF-α and IL-6 in response to LPS injection were comparable between Fosl1^Δ/Δ and control mice, suggesting that, as long as c-Fos is present, the loss of Fra-1 does not result in increased cytokine production (data not shown).

We then wondered whether overexpression of Fos proteins might suppress the inflammatory response. Whereas c-Fos Tg mice develop osteosarcoma (17), Fra-1 Tg mice, which widely overexpress Fra-1 under the control of the H-2^k promoter, show enhanced bone formation (9). We next determined the effects of forced Fra-1 expression on pro- and anti-inflammatory cytokine production, using Fra-1 Tg mice. To prove the principle of acute inflammatory responses, we injected 4 mg/kg of LPS i.p. into wt and Fra-1 Tg mice, and measured cytokine levels in serum. Levels of proinflammatory cytokines, TNF-α, IL-6, and IL-12, were significantly lower in Fra-1 Tg mice than in wt control mice (Fig. 4A). The production levels of anti-inflammatory cytokines TGF-β and IL-4 were also lower in LPS-injected Fra-1 Tg mice, whereas that of IL-10 was not altered (Fig. 4B). LPS injection is known to reduce the locomotor activity of mice through the induction of inflammatory mediators. To determine whether reduced proinflammatory cytokine production in Fra-1 Tg mice could block the LPS-induced decrease in locomotor activity, we implanted a telemetric radio transmitter into wt and Fra-1 Tg mice. After injection of a low dose (1 mg/kg) of LPS, wt control mice often stayed in a crouched position, resulting in reduced locomotor activity, but Fra-1 Tg mice tolerated LPS injection and remained active (Fig. 4C). Therefore, Fra-1 overexpression reduces the production of proinflammatory cytokines such as TNF-α in vivo.

We then determined how Fra-1–overexpressing cells respond to LPS. We prepared MDMs and MEFs from wt and Fra-1 Tg mice. Fra-1–overexpressing MDMs and MEFs produced lower levels of IL-6 than did control wt cells in response to increasing LPS concentration (Fig. 5A). NO production was also reduced (Fig. 5B), suggesting that various inflammatory responses are diminished in Fra-1 Tg mice. To examine the molecular mechanisms underlying reduced inflammatory responses, we compared AP-1 and NF-κB binding activities between wt and Fra-1 Tg MEFs treated with LPS, using EMSA. In Fra-1 Tg MEFs, AP-1–binding activity was higher than in the control mice, as expected. In contrast, LPS-induced NF-κB–binding activity was lower in Fra-1 Tg than in wt MEFs (Fig. 5C). These findings suggest that diminished inflammatory responses to LPS in Fra-1–overexpressing cells result from NF-κB suppression.

To gain insight into the molecular mechanisms underlying the suppression of NF-κB by Fra-1, we used MEFs to analyze protein levels of LPS receptors, NF-κB, and MAPKs. As shown in Fig. 6A, expression levels of two LPS receptors, TLR4 and CD14, were comparable between wt and Fra-1–overexpressing cells. Activation of NF-κB requires the sequential activation of IκBα kinase (IKK), degradation of IκBα, and phosphorylation and nuclear translocation of p65 (18). We therefore treated wt and Fra-1–overexpressing MEFs with 0.1 μg/ml of LPS for different time periods, prepared the whole-cell extracts, and analyzed the activation status of these molecules. In wt MEFs, we observed the LPS-induced phosphorylation and degradation of IκBα, as well as the phosphorylation of p65. Interestingly, the phosphorylation of IκBα was decreased in Fra-1–overexpressing MEFs, and degradation of IκBα and phosphorylation of p65 were also reduced (Fig. 6B). To explore the role of Fra-1 in LPS-induced MAPK activation, we analyzed LPS-induced phosphorylation of JNK, ERK, and p38. In wt MEFs, phosphorylated forms of JNK, ERK, and p38 were observed in a time-dependent manner, with maximum activation at 15 to 30 min. In contrast, in Fra-1–overexpressing MEFs, the phosphorylation of each MAPK was reduced and delayed (Fig. 6C). Thus, our results suggest that Fra-1 overexpression suppresses the activation of NF-κB and MAPKs by LPS.

Because mice overexpressing Fra-1 showed reduced LPS-induced inflammatory responses, we studied whether Fra-1 overexpression could protect mice from DSS-induced colitis. We treated Fra-1 Tg mice and their wt littermates with 2% DSS in drinking water and induced acute intestinal inflammation. The Fra-1 Tg mice and wt mice similarly survived the DSS treatment (Fig. 7A). Although wt mice showed progressive clinical colitis profiles, however, these profiles were all suppressed in Fra-1 Tg mice (Fig. 7B–D). Thirteen days after DSS administration, the colon was histologically analyzed. In Fra-1 Tg mice, the percentage of colon shortening was half that of wt controls (Fig. 7E), and both the invasiveness of inflammatory lesions and crypt damage were significantly reduced (Fig. 7F). We found that production levels of inflammatory cytokines such as TNF-α, IL-6, IL-12, TGF-β, and IL-4 were significantly lower in DSS-treated Fra-1 Tg mice, whereas IL-10 expression was comparable between DSS-treated wt and Fra-1 Tg mice (Fig. 7G). We also noted that TNF-α and IL-6 production (Supplemental Fig.1), NF-κB–binding activity (Fig. 7H), and MAPK activity (Fig. 7I) in colon fragments were lower in DSS-treated Fra-1 Tg mice than in DSS-treated wt mice. The suppressed production of TNF-α and IL-6 in Fra-1 overexpression was also observed in colon fragments treated with LPS (Supplemental Fig. 2). Taken together, these results demonstrate that overexpression of Fra-1 reduces susceptibility to DSS-induced colitis.

In this study, we determined the functions of Fos proteins, c-Fos and Fra-1, in inflammatory stress responses using LPS injection and a DSS-induced colitis model. Upon LPS injection, Fra-1 Tg mice showed decreased production of proinflammatory cytokines and attenuated endotoxin shock as illustrated by continuous locomotor activity. Moreover, when treated with LPS, the MDMs and MEFs derived from Fra-1 Tg mice showed reduced activation of NF-κB and MAPKs, as well as reduced production of proinflammatory cytokines and NO, compared with wt control cells. The magnitude of DSS-induced colitis, proinflammatory cytokine production, and NF-κB activation were diminished in Fra-1 Tg mice. Therefore, the Fra-1–containing AP-1 transcription factor acts as a negative regulator of NF-κB and inflammatory responses.

Because bone formation is enhanced in Fra-1 Tg mice (9, 12, 19), we separately analyzed the healing process after tibial transverse fracture and consistently found that production of inflammatory cytokines, including TNF-α and IL-6, around the bone fracture site is impaired in these mice (20). Furthermore, using the murine macrophage cell line RAW264.7 overexpressing Fra-1 by rentivirus, Morishita et al. (21) reported that Fra-1 negatively regulates LPS-induced production of TNF-α, IL-1, IL-6, and NO. In contrast, RAW264.7 cells transfected with Fra-1 small interfering RNA showed increased LPS-induced expression of proinflammatory cytokines (21). These observations indicate that Fra-1 is a negative regulator of proinflammatory cytokine production.

Because c-Fos is also a negative regulator in LPS-induced expression of proinflammatory cytokines (7, 22–28), we examined the effect of c-Fos disruption in DSS-induced colitis. Fos^−/− mice exhibited higher susceptibility to DSS-induced colitis associated with higher levels of proinflammatory cytokines and NF-κB activity than did wt mice. Furthermore, administration of NF-κB inhibitor Bay 11-7085 suppressed the enhancement of DSS-induced colitis in Fos^−/− mice. These findings establish that both c-Fos and Fra-1 are negative regulators in LPS-induced proinflammatory cytokine production and NF-κB activation.

In this study, we observed that the enhanced response to LPS is associated with an increased susceptibility to DSS in mice lacking c-Fos. In contrast, mice lacking TLR4, TLR2, or MyD88 also show increased susceptibility to mucosal inflammation (29–31). How can these two observations be reconciled? First, the enhanced susceptibility of Fos^−/− mice to DSS is apparently mediated not only by TLR-MyD88 signaling but also by other mechanisms such as direct damage to colon epithelial cells after DSS treatment. Second, the mixed genetic background of mice lacking c-Fos (129 and C57BL/6J), which was unavoidable owing to the reduced fertility of Fos^−/− mice on a C57BL/6J background, might affect DSS susceptibility. It will be informative to study the roles of Fos proteins in various chemically induced colitis models such as those using acetic acid and trinitrobenzene sulfuric acid (32).

How do Fra-1 and c-Fos negatively regulate NF-κB–mediated proinflammatory cytokine production? In the DSS-induced colitis model in Fos^−/− and Fra-1 Tg mice, the anti-inflammatory mediator IL-10 was not likely involved, because the IL-10 expression levels in those mice were comparable with those in wt mice. In the current study, we also showed that production of anti-inflammatory cytokines such as TGF-β and IL-4 in DSS-treated mice was mainly paralleled by the activation of NF-κB and DSS-induced injury. Therefore, these changes in the production of anti-inflammatory cytokines do not seem to contribute to the negative regulation of NF-κB or to stress responses by Fos proteins.

Because NF-κB–binding activity in the colon was increased in Fos^−/− mice and decreased in Fra-1 Tg mice after DSS treatment, we analyzed LPS-induced IKK activation by detecting phosphorylation and degradation of IκBα, which leads to NF-κB activation. In Fra-1–overexpressing cells, we found that phosphorylation and degradation of IκBα were reduced, which explains the reduced NF-κB activation and impaired production of proinflammatory cytokines. Interestingly, the activation of MAPKs in Fra-1–overexpressing cells was also suppressed. These results raise the possibility that Fos/AP-1 might inhibit the expression or activity of upstream kinase(s), which activate both MAPK and NF-κB pathways. It is worth noting that TGF-β-activated kinase 1 (TAK1) and Tpl-2 (Map3k8, Cot), kinases of the MAPK family, are known to regulate activation of NF-κB, JNK, ERK, and p38 (33, 34). The possible impaired expression and/or activation of TAK1 and Tpl-2 in Fra-1 Tg mice need to be further explored. The direct association between AP-1 and NF-κB might also be a mechanism (35).

One intriguing possibility is that the AP-1 and NF-κB pathways are reciprocally regulated. NF-κB activation appears to downregulate the MAPK–AP-1 signaling pathway, such as NF-κB subunit p65–mediated suppression of MAPK kinase activities (36–38). It was reported that increased reactive oxygen species (ROS) in IBD are implicated as mediators of intestinal inflammation (39). The generation of ROS is increased in response to external stimuli, including LPS and TNF-α. ROS, such as H₂O₂, induces NF-κB activation by tyrosine phosphorylation of Syk, followed by IKK activation (40), and downregulates the activity of AP-1 (41). The antioxidants pyrrolidine dithiocarbamate and N-acetyl-l-cysteine, which inhibit NF-κB activation, enhance DNA binding and transactivation of AP-1 in response to phorbol ester (42). Hirata et al. (43) reported that pyrrolidine dithiocarbamate suppressed DSS-induced TNF-α production and NF-κB activation. These studies raised the possibility that NF-κB and AP-1 activities are regulated by mutual negative feedback mechanisms in stress responses, including those during DSS-induced colitis. Curiously, mutual negative regulation between AP-1 and NF-κB has been observed in Drosophila (44, 45).

Therapeutic applications of decoy oligodeoxynucleotides (ODN) for NF-κB and AP-1 have been studied; the NF-κB decoy ODN exerts potent immunosuppressive effects with inflammatory diseases (46), whereas the AP-1 decoy ODN is also reported to have the therapeutic potential to attenuate intestinal inflammation (47). Our results and others collectively suggest that suppression of either AP-1 or NF-κB might elevate the activity of the other. Elucidation of the reciprocal relationship between AP-1 and NF-κB in various tissues and stresses may clarify the mechanisms underlying the regulation of inflammation magnitude.

We thank Toru Yamaguchi for producing Fra-1 Tg mice, Kenta Maruyama for discussion, and Mayako Asakawa for technical support.

Disclosures The authors have no financial conflicts of interest