Arsenic trioxide, a potent inhibitor of NF-κB, abrogates allergen-induced airway hyperresponsiveness and inflammation

Background Overactivation of nuclear factor κB (NF-κB) orchestrates airway eosinophilia, but does not dampen airway hyperresponsiveness in asthma. NF-κB repression by arsenic trioxide (As2O3) contributes to apoptosis of eosinophils (EOS) in airways. Here we provide evidence that As2O3 abrogates allergen (OVA)-induced airway eosinophilia by modulating the expression of IκBα, an NF-κB inhibitory protein, and decreases the airway hyperresponsiveness. Methods Using a murine model of asthma, the airway hyperresponsiveness was conducted by barometric whole-body plethysmography. Airway eosinophilia, OVA-specific IgE in serum, and chemokine eotaxin and RANTES (regulated upon activation, normal T cell expressed and secreted) in bronchoalveolar lavage fluid were measured by lung histology, Diff-Quick staining, and ELISA. Chemokine-induced EOS chemotactic activity was evaluated using EOS chemotaxis assay. Electrophoretic mobility shift assay and Western blot analysis were performed to assess pulmonary NF-κB activation and IκBα expression, respectively. Results As2O3 attenuated the allergen-induced serum IgE, chemokine expression of eotaxin and RANTES, and the EOS recruitment in bronchoalveolar lavage fluid, which is associated with an increased IκBα expression as well as a decreased NF-κB activation. Also, As2O3 suppressed the chemotaxis of EOS dose-dependently in vitro. Additionally, As2O3 significantly ameliorated the allergen-driven airway hyperresponsiveness, the cardinal feature underlying asthma. Conclusion These findings demonstrate an essential role of NF-κB in airway eosinophilia, and illustrate a potential dissociation between airway inflammation and hyperresponsiveness. As2O3 likely exerts its broad anti-inflammatory effects by suppression of NF-κB activation through augmentation of IκBα expression in asthma.


Background
Asthma is now accepted as a T-helper type 2 (Th2) lymphocyte-mediated chronic inflammatory disorder, characterized by airway eosinophilia and airway hyperresponsiveness (AHR) [1]. Eosinophils (EOS) appear to play a crucial role in the ongoing inflammation due to either an impaired clearance or a delayed apoptosis in the airways, where accumulation of a number of EOS cytotoxic proteins including major basic protein, cationic proteins and peroxidase could occur [2]. Existing data support the notion that morphologic changes in airway tissue to the development and severity of AHR in asthma correlates with the presence of activated airway inflammatory cells, in particular EOS [3].
The molecular regulatory pathways in induction of chronic cytokine expression and recruitment/activation of inflammatory cells in asthma remain elusive. However, there is growing recognition that these processes involve increased transcription of inflammatory genes via transcription factors [4]. One such transcription factor, nuclear factor κB (NF-κB), is abundant of p50 (NF-κB1)/ p65 (RelA) heterodimer. In a latent state, NF-κB is sequestered as an inactive trimer by complexing with IκBα, a 37 kDa inhibitory protein, which promotes cytoplasmic retention and maintains a low basal transcriptional activity. IκBα consists of an N-terminal domain containing specific phosphorylation sites, five ankyrin repeat sequences, and a C-terminal domain of Pro-Glu-Ser-Thr polypeptides [5]. Upon stimulation, IκBα is phosphorylated by the IκB kinase, ubiquitinated and degraded through the 26S proteasome pathway [6]. Subsequently, the nuclear localization sequence of NF-κB is unmasked to allow its translocation into the nucleus, where it binds to DNA and initiates transcription of a wide range of NF-κB-dependent genes in association with immune and inflammatory responses [7].
Arsenic compound has long been considered as a protoplasmic poison that can bind to human sulfydryl-containing proteins with high affinity. Arsenic trioxide (As 2 O 3 ), extracted from arsenic compound, is a powerful ancient medication for a variety of ailments with the principle of "using a toxic against another toxic" in traditional Chinese medicine. Strikingly, As 2 O 3 treatment in a regime of 10 mg/d of intravenous infusion for 28 to 60 days is effective in patients with acute promyelocytic leukemia (APL) without viable toxicity in refractory to the all-trans retinoic acid (ATRA) and the conventional chemotherapy by inducing apoptosis of APL cells [8]. Many studies have demonstrated that NF-κB overactivation underlines the chronicity of airway inflammation characteristic of asthma [9][10][11][12]. Recently, we have reported that As 2 O 3mediated NF-κB repression in airways facilitated EOS apoptosis in a dose-dependent manner, contributing to the resolution of airway eosinophilic inflammation [13]. In this study, we investigated the effects of As 2 O 3 on allergen-induced AHR and NF-κB-mediated airway inflammation in a murine model of asthma. Our data indicate that inhibition of NF-κB activation through induction of IκBα expression may account for the broad anti-inflammatory action of As 2 O 3 .

Asthma modeling
Specified pathogen-free female BALB/c mice, aged 6 to 8 weeks, were provided by the Chinese Academy of Medical Sciences (Beijing, China). The animal experiment was approved by Nanjing Medical University according to the guidelines of the Institutional Animal Care and Use Committee. A murine asthma model was established as described previously [14] with minor modifications.
On days 0 and 7, mice received intraperitoneal injection of 20 µg of chicken ovalbumin (OVA, Grade V, Sigma-Aldrich, St. Louis, MO) adsorbed to 20 mg of aluminum hydroperoxide gel (Pierce, Rockford, IL). On days 14, mice were randomized to receive aerosol challenge with either 6% OVA in phosphate-buffered saline (PBS) or PBS alone via a nebula (1-5 µM particles, Bohringer Ingelheim, Germany) for 40 min per day up to 7 days. During the treatment period, As 2 O 3 (Yida Pharmaceutics, Harbin, China) at dose of 0.5-4.5 mg/kg, dexamethasone (Dex, Phoenix Pharmaceutics, Belmont, CA) at dose of 2.5 mg/ kg or PBS alone was injected into the peritoneum 30 min before each airway challenge. After the last aerosol exposure, mice were sacrificed at designated timepoints.

Airway physiology
Baseline resistance and AHR induced by nebulized methacholine (Sigma-Aldrich, St. Louis, MO) at dose of 12.5-100 mg/ml in conscious unrestrained-mice were assessed using barometric whole-body plethysmography (Buxco Electronics Inc., Troy, NY) as described previously [15]. Airway resistance is expressed as: where P enh = enhanced pause, T e = expiratory time (sec), T r = relaxation time (sec), P ef = peak expiratory flow (ml/sec), and P if = peak inspiratory flow (ml/sec).

Bronchoalveolar lavage
Four hours after the last airway challenge, mice underwent euthanasia and were cannulated in the trachea. The lungs were washed twice with 1 ml aliquots of PBS to collect the bronchoalveolar lavage fluid (BALF). Subsequently, the lungs were removed, quickly frozen in liquid nitrogen, and stored at -70°C. Additionally, the lungs were collected at 1, 12, and 24 hrs post the last airway challenge to study the kinetics of pulmonary NF-κB activation.

Diff-Quick staining
Diff-Quick staining is a modified Wright's staining [16]. Centrifuged at 300 × g for 10 min, the pelleted cells of BALF were suspended in a serum-free RPMI 1640 medium. The cell viability, evaluated by the trypan blue exclusion method, was over 95%. Total and differential cell counts were enumerated on cytospins (Thermo Shandon, Pittsburgh, PA) in compliance with the Diff-Quick staining profile (Merck, Germany) by counting at least 200 to 500 cells in cross-section.

Enzyme-linked immunosorbant assay (ELISA)
Serum levels of OVA-specific immunoglobulin E (IgE) were analyzed by ELISA using samples collected 24 hrs after the last OVA challenge. Briefly, 96-well plates were coated with either purified anti-mouse IgE (5 µg/ml, BD PharMingen, San Diego, CA) or OVA (100 µg/ml). After addition of serum samples, OVA-specific IgE was detected using horseradish peroxidase (HRP)-conjugated sheep anti-IgG (Calbiochem, La Jolla, CA). Arbitrary units (AU) were calculated according to OD 50 of the standard curve.
Murine chemokines, eotaxin and RANTES (regulated upon activation, normal T cell expressed and secreted), in the BALF samples were measured by utilizing paired antibodies following the manufacturer's recommendations. The ELISA kits were purchased from R&D Systems (Minneapolis, MN) with a minimum detectable levels of 3 and 5 pg/ml for eotaxin and RANTES, respectively.

EOS chemotaxis assay (ECA)
Interleukin (IL)-5 transgenic mice (CBA/CaH-TnN) were provided by the Institute of Chemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, China). EOS (~98% purity) were derived from spleen of IL-5 transgenic mice with depletion of B, T, and antigen-presenting cells using anti-B220, anti-CD4, anti-CD8 and anti-class II, as well as rat anti-mouse Ig-conjugated magnetic beads (Miltenyi Biotec, Auburn, CA) as described previously [17]. EOS were seeded at 5 × 10 4 density in triplicate and preincubated for 15 min at room temperature with 0.25-2 µM of As 2 O 3 prior to chemotaxis measurement.
Chemotaxis was assessed in 48-well micro-Boyden chambers using polyvinylpyrrolidone-free polycarbonate membranes (NeuroProbe, Bethesda, MD). Cell suspension and diluted chemokines of eotaxin or RANTES (PeproTech, London, UK) were added into the chamber with RPMI 1640 containing 25 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acids (HEPES, pH 7.4) and 0.05% bovine serum albumin. The plates were incubated for 60 min at 37°C under 5% CO 2 . The migrated cells were counted in five randomly selected high-power fields (magnification was × 1,000). Spontaneous migration was evaluated in the absence of chemoattractant.
The total proteins were prepared by addition of Buffer A to the lung powder and subjected to two freeze/thaw cycles to fracture the nuclear membranes. After centrifugation, the supernatant was collected. The nuclear and total proteins were quantitated using the Bradford assay (Bio-Rad, Hercules, CA), aliquoted and stored at -70°C until use.
For competition assay, a 100-fold excess of unlabelled NF-κB or activator protein 1 (AP-1) oligonucleotide probe was added to the reaction mixture 10 min before addition of the labelled probe. For supershift assay, a 0.5 µg of anti-p50 or anti-p65 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was added to the reaction mixture prior to the labelled probe for 30 min.

Data analysis
Statistical analysis was performed by one-way analysis of variance (ANOVA) and q test with SPSS 11.0 software package (SPSS Inc., Chicago, IL). The negative relationship was evaluated by Pearson correlation analysis. Data were expressed as mean ± SEM, and p < 0.05 was considered statistically significant.

Attenuation of airway EOS recruitment by As 2 O 3
OVA-challenged mice in response to 0.5-4.5 mg/kg of As 2 O 3 reduced the number of EOS in BALF in a dosedependent manner (Fig. 1). Since the anti-inflammatory effects of As 2 O 3 were similar at the doses of 4 and 4.5 mg/ kg, and it was comparable to the effect of 2.5 mg/kg of Dex (p > 0.05), the 4 mg/kg of As 2 O 3 was herein chosen as the effective dosage in the rest of experiments. This dosage was also proved to be relatively safe based on our previous experiments [13,14]. Histological analysis of the OVAchallenged mice lung revealed an enhanced airway eosinophilia as compared to the naïve control mice that were treated with PBS ( Fig. 2A). Conversely, pretreatment of As 2 O 3 protected mice from developing the allergeninduced peribronchial inflammation ( Fig. 2A). Examination of BALF collected from mice at 24 hrs after OVA challenge showed a marked influx of inflammatory cells into the airways, including EOS, lymphocytes, macrophages and neutrophils (Fig. 2B-C). The increased EOS in the BALF was correlated with an increase of EOS recruitment by the Diff-Quick analysis in OVA-challenged mice (Fig.  2B). The number of EOS in BALF from naïve mice was less than 1%, whereas that of OVA-challenged mice was about 49% (p < 0.01). Pretreatment of As 2 O 3 dramatically attenuated the airway eosinophilia in the OVA-challenged mice (p < 0.01; Fig. 2A-C; Table 1).

Amelioration of AHR by As 2 O 3
P enh , relative to the measured airway resistance, was obtained as an index and was normalized to the postsaline -P enh . This readout was used as a measure of AHR.
Mice previously sensitized and challenged with OVA developed a dose-dependent methacholine-induced bronchospasm as compared to the naïve mice that were treated with PBS. As 2 O 3 treatment significantly reduced the effect (p < 0.01; Fig. 3).

Reduction of serum IgE and BALF chemokines by As 2 O 3
IgE can augment allergic airway responses in a high affinity receptor-dependent manner. Serum levels of OVA-specific IgE were elevated in OVA-challenged mice compared with the naïve control mice (p < 0.01), whereas pretreatment with As 2 O 3 resulted in a 4.8-fold decrease to the levels of the OVA mice (p < 0.01; Fig. 4A). Eotaxin and RANTES play a critical role in inducing chemotaxis of EOS [19]. ELISA analysis showed that levels of eotaxin and RANTES in BALF were markedly increased in OVA-challenged mice in comparison with the control mice (p < 0.01). However, these chemokine levels were largely reduced by pretreatment with As 2 O 3 (p < 0.05 or 0.01; Fig.  4B).

Ablation of EOS chemotaxis by As 2 O 3
Eotaxin and RANTES with respective concentrations of 1 (10 0 ) and 10 3 nM reached a maximal chemotaxis response indicating that eotaxin is a more active chemotaxin to EOS than RANTES (Fig. 5A). As 2 O 3 significantly inhibited the EOS chemotaxis mediated by eotaxin or RANTES in a dose-dependent manner (p < 0.05 or 0.01; Fig. 5B).

Inhibition of pulmonary NF-κB activation by As 2 O 3
The OVA challenged mice showed a sharp increase in the pulmonary DNA binding activity of NF-κB at various timepoints as compared to the unchallenged mice lung. Indeed, NF-κB activity was increased within 1 hr (p < 0.01), peaked at 4 hrs (p < 0.01), and decreased by 12 (p < 0.01) to 24 hrs (p < 0.05). This effect of OVA challenge was clearly ameliorated by pretreatment with As 2 O 3 (p < 0.01; Fig. 6, lane 6 as compared to lane 3; Table 1). In the competition assay, addition of 100-fold excess of unlabelled NF-κB, but not AP-1, oligonucleotide probe competed away the NF-κB-DNA complexes, verifying the specificity of NF-κB binding. In the supershift assay, addition of antibodies against p50 and p65 resulted in retardation of supershifted bands, with reciprocal decreases in the intensity of the NF-κB bands, confirming the classic subunits of NF-κB heterodimer (Fig. 6).

Augmentation of pulmonary IκBα expression by As 2 O 3
The pulmonary IκBα expression in the lung lysate was relatively decreased in OVA-challenged mice (p < 0.01; Fig.  7; Table 1) compared to the control lung. In contrast, pretreatment of As 2 O 3 accumulated the pulmonary IκBα (p < 0.01). Furthermore, there was a tight negative correlation between EOS recruitment in the BALF or the pulmonary NF-κB activation and IκBα expression (r = -0.82 and -0.94, respectively; p < 0.01).

Discussion
Multiple upstream signal events converge on the NF-κBinducing kinase (NIK) [20]. Activation of NIK results in phosphorylation of IκB kinases, which render the phosphorylation of IκBα at N-terminal serines 32 and 36 (Ser32 and Ser36) residues, leading to a proteolytic degradation of IκBα. Consequently, the activated NF-κB translocates to the nucleus, where it bonds to specific κB sites to facilitate the transcription of target genes. This results in expression of numerous pro-inflammatory cytokines, chemokines and adhesion molecules [21]. These proinflammatory mediators are essential in the recruitment of airway inflammatory cells, including EOS and CD4 + T lymphocytes, which in turn secret Th2 cytokines [22]. Therefore, NF-κB repression in airways via suppression of IκBα degradation or augmentation of IκBα synthesis would decrease the transcription of a myriad of NF-κBdependent genes. This strategy proved to be more effective than that of blocking a single downstream inflammatory or an immune gene among the inflammatory cascade [23,24].
Several lines of evidence suggest a central role of NF-κB in the pathogenesis of asthma. Activated NF-κB has been identified in sputum-induced macrophages and bronchial biopsy specimens of asthmatic patients [25]. Agents such as allergens, ozone and viral infections, which are associated with exacerbation of asthma, stimulate activation of NF-κB [26]. As the major effective treatment for asthma, As 2 O 3 decreases EOS recruitment in BALF in a dose-dependent manner Figure 1 As 2 O 3 decreases EOS recruitment in BALF in a dose-dependent manner. Intraperitoneal administration of OVAchallenged mice with As 2 O 3 (0.5-4.5 mg/kg) reduced the EOS in BALF, in which both 4, 4.5 mg/kg of As 2 O 3 and 2.5 mg/kg of Dex achieved the similar anti-inflammatory effects. BALF EOS, stained with Diff-Quick solution, were counted using a hematocytometer, and expressed as a percentage in total leukocytes. Data represent the mean ± SEM of four separate experiments (n = 6 per group). # p < 0.05, *p < 0.01, vs the control mice; ‡ p < 0.05, † p < 0.01, vs the OVA-challenged mice.   Data represent the mean ± SEM of four independent experiments (n = 6 per group). # p < 0.05, * p < 0.01, vs the control mice; † p < 0.01, vs the OVA-challenged mice at 4 hrs. glucocorticoids are potent blockers of NF-κB activation [27]. Furthermore, mice lacking the NF-κB subunits p50 or c-Rel develop less airway inflammation upon antigen challenge [28]. Nevertheless, NF-κB activation orchestrates allergen-induced inflammation and subsequent adaptive responses, but does not appear to modulate AHR, the cardinal feature that underlies asthma, signifying a potential dissociation between airway inflammation and AHR [29]. Clearly, additional airway signaling pathways activated, residual NF-κB activity or other inflammatory processes may be responsible for the AHR. Alternatively, events localized more distally within the alveolar compartments, such as microvasculature leakage of macromolecules, alveolar injury or surfactant dysfunction might dominate the genesis of AHR [30][31][32].

As 2 O 3 prohibits allergen-induced AHR
induces the apoptosis in t (15;17) APL cell line NB4 in vitro and in APL patients without significant myelosuppression in vivo [8]. We and others have confirmed that inhibition of NF-κB was essential to arsenic-induced apoptosis [13,33]. In this report, despite a decreased serum OVA-specific IgE production, we demonstrated an inhibitory effect of As 2 O 3 on EOS recruitment from OVA-challenged BALF, in agreement with our previous observation that As 2 O 3 promoted EOS apoptosis in the airway eosinophilic inflammation [13]. Additionally, both eotaxin and RANTES, downstream genes of NF-κB, demonstrated potent chemoattractants to EOS and Th2 lymphocytes [34]. Presumably, the ablation of airway eosinophilia by As 2 O 3 results from a collective effects of NF-κB inhibition such as a reduced specific IgE secretion, chemokine expression and Th2 cytokine production as well as an altered eosinophilic cytoskeletal rearrangement [35,36]. Overall, As 2 O 3 might exert its multiple antiinflammatory action through augmentation of IκBα expression and suppression of NF-κB activation in the airways. This is partially in accordance with the therapeutic role of glucocorticoid-mediated NF-κB repression in asthma [37,38]. Interestingly, in this model of asthma, As 2 O 3 abrogated both allergic airway inflammation and AHR in contrast with the previous report [29], suggesting a specific effect of As 2 O 3 besides NF-κB suppression. Taken together, these findings not only prove an essential role of NF-κB-mediated airway inflammation, but also illustrate the importance of alternative signaling pathway and additional cell types in the airways, and the complicated interactions between them in dictating the pathophysiology of asthma.

Conclusion
Our data demonstrate that a broader anti-inflammatory activity of As 2 O 3 lies in the inhibition of NF-κB activation through induction of IκBα expression in the airways. Clinically, low dosage of As 2 O 3 may have a potential benefit in treating patients with asthma, especially in those with steroid-dependent and -resistant asthma [8,13]. It is anticipated that specific inhibitors of NF-κB may be developed by modifying the poisonous group(s) of As 2 O 3 and screen As 2 O 3 analogues in the libraries of chemical compounds. Moreover, novel nondegradable IκBα mutant, namely super-repressor of NF-κB, may be achieved by completely deleting the phosphorylation sites of Ser32 and Ser36 residues [18,37]. This will offer promising strategies for future immunotherapy of asthma as well as the infectious, inflammatory, cancerous and autoimmune diseases associated with aberrant NF-κB activation [1,5,[39][40][41][42].