Inhibition of c-Fos expression attenuates IgE-mediated mast cell activation and allergic inflammation by counteracting an inhibitory AP1/Egr1/IL-4 axis

Background Activator protein-1 (AP1), a c-Fos–JUN transcription factor complex, mediates many cytobiological processes. c-Fos has been implicated in immunoglobulin (Ig)E activation of mast cells (MCs) via high-affinity IgE Fc receptor (FcεRI) binding. This study examined c-Fos involvement in MC activation and tested the effects of the c-Fos/AP1 inhibitor T-5224 on MCs activation and allergic responses. Methods In vitro studies were conducted with two MC model systems: rat basophilic leukemia cells (RBLs) and mouse bone marrow derived mast cells (BMMCs). MC degranulation and effector functions were examined with β-hexosaminidase release and cytokine secretion assays. c-Fos/AP1 was inhibited with T-5224. c-Fos activity was suppressed with short hairpin RNA targeting c-Fos (shFos). In vivo immune responses were evaluated in passive cutaneous anaphylaxis (PCA) and ovalbumin-induced active systemic anaphylaxis (ASA) models, as well as in an oxazolone (OXA)-induced model of atopic dermatitis, a common allergic disease. Results c-Fos expression was elevated transcriptionally and translationally in IgE-stimulated MCs. c-Fos binding of the Egr1 (early growth response 1) promoter upregulated Egr1 transcription, leading to production of interleukin (IL)4. T-5224 reduced FcεRI-mediated MC degranulation (evidenced by β-hexosaminidase activity and histamine levels) and diminished EGR1 and IL4 expression. T-5224 attenuated IgE-mediated allergic responses in PCA and ASA models, and it suppressed MC-mediated atopic dermatitis in mice. Conclusion IgE binding can activate MCs via a c-Fos/Egr1/IL-4 axis. T-5224 suppresses MC activation in vitro and in vivo and thus represents a promising potential strategy for targeting MC activation to treat allergic diseases. Supplementary Information The online version contains supplementary material available at 10.1186/s12967-021-02932-0.


Introduction
Allergic diseases are disruptive to quality of life, can cause serious morbidity, and their incidence has been on the rise [1]. Mast cells (MCs) play a key role in immunoglobulin (Ig)E mediated allergic reactions [2]. Binding of high-affinity IgE Fc receptors (FcεRIs) on the surface of MCs by antigen (Ag)-linked IgE antibodies activates the IgE/FcεRI pathway, which promotes cellular degranulation and the release of  19:261 cytoinflammatory factors [3,4]. FcɛRI aggregation induces phosphorylation of a linker molecule [5] and a signaling cascade involving Bruton's tyrosine kinase, phospholipase γ, and protein kinase C [6], leading to activation of downstream signaling pathways-including calcium signaling [7,8], mitogen-activated protein kinase (MAPK) [9] and nuclear factor κB [10] pathways-which initiate MC degranulation and cytokine production.
Elucidating molecular mechanisms that modulate MC activation could provide crucial insights into the pathophysiology of allergic diseases. Recently, FcεRIstimulated MC activation and anaphylaxis were shown to be dampened by antagonism of an orphan nuclear receptor called nuclear receptor subfamily 4 group A member 1, which acts in opposition to the antiinflammatory liver kinase B1/adenosine monophosphate-activated protein kinase axis [11]. In addition, FcεRI-mediated MC activation has been shown to be negatively regulated by regulator of calcineurin 1 [12]. Activator protein-1 (AP-1)-a transcription factor complex formed by dimerization of c-Fos, a leucine zipper protein, with c-Jun-modulates signaling pathways involved in many cytobiological processes, including proliferation, differentiation, transformation, and cell death [13]. AP-1-dependent regulation of cytokine transcription requires the presence of c-Fos [14], and IgE/Ag treatments upregulate c-Fos transcription in MCs [15,16]. In c-Fos-deficient MCs, FcɛRI-mediated degranulation can be inhibited by down-regulation of degranulation genes [13,17]. Experiments examining anaphylactic processes have shown that released cytokines can induce AP-1-dependent MC secretion of interleukin (IL)6 consequent to activation of the IL6 gene promoter [18]. The molecular mechanism by which FcεRI activation leads to upregulation of c-Fos expression has not been clarified. Moreover, clinical translation studies of the effects of c-Fos/AP-1 inhibition on IgE/Ag-activated MCs and allergic responses are needed.
Here, we used next-generation RNA sequencing (RNA-Seq) and real time quantitative reverse transcriptase (qRT)-polymerase chain reaction (PCR) to analyze the MC transcriptional response to IgE/ Ag activation. We identified MC activation-associated proteins that interact with c-Fos, including early growth response (EGR), IL, and chemokine (C-C motif ) ligand (CCL) proteins [19,20]. We were particularly interested in the potential involvement of EGR1 because it regulates IL-4 secretion in FcεRIactivated MCs [21]. We conducted qRT-PCR and western blotting to measure Egr1 expression levels in stimulated MCs, with and without anti-Fos silencing short hairpin RNA (ShFos). To explore the intracellular mechanism of c-Fos involvement in MC activation, we examined the effects of the c-Fos/AP-1 inhibitor T-5224 on inflammatory cytokine expression and MAPK signaling. Finally, we examined T-5224 effects on MC activation responses in in vivo models.

Animals
Female Balb/c mouse pups were used in accordance with ethical protocols. Female Balb/c mice (6-8 weeks, 18-20 g) from Guangdong Medical Laboratory Animal Center (Benxi, Liaoning, China) were bred in an aseptic-specified pathogen-free environment with controlled humidity (55 ± 10%) and temperature (24 ± 1 °C). Animal experiments were carried out according to protocols approved by our University's Animal Care and Use Committee and in compliance with the Guidelines on Animal Welfare of the School of Medicine of Shenzhen University. Mouse health was monitored twice daily, and early humane euthanasia was conducted in cases of weight loss > 15% body weight, lethargy, inability to stand, or anorexia.

Cytotoxicity
Cell viability was evaluated with a Cell Counting Kit 8 assay kit (MedChem Express, Monmouth Junction, NJ). RBLs (2 × 10 3 /well) and BMMCs (1 × 10 4 /well) were each cultured in 96-well plates and incubated with T-5224 for 24 h. Kit-provided solution was added wells and incubated at 37 °C for 1 h. Absorbance at 450 nm was detected by a multi-well plate reader (Bio-Rad, Hercules, CA). Viability was calculated as relative absorbance expressed as a percentage of control values.

Real-time qRT-PCR
Total cellular RNA was isolated with an RNeasy Mini Kit (Qiagen, Duesseldorf, Germany) and reverse transcribed into cDNAs with a TIAN Script II RT Kit (Tiangen Biotech, Beijing, China) and primers (Additional file 1: Table S1) according to the manufacturer's instructions. Real-time qRT-PCRs were performed with TB Green ® Premix ExTaq ™ (Takara, Tokyo, Japan) in a qTOWER 2.2 system (Analytik Jena, Upland, CA). Expression levels were computed by normalizing relative to GAPDH (glyceraldehyde 3-phosphate dehydrogenase) mRNA levels with the 2 − ΔΔCt technique. The primers were shown in Additional file 1: Table S1.

Western blotting
Cell lysate preparation and immunoblotting were performed as previously described [29]. Anti-DNP IgE (50 ng/mL)-sensitized RBLs (5 × 10 5 /well in 6-well plates) were pretreated with T-5224 for 1 h and then stimulated with DNP-HSA for 4 h (p-c-Fos FRA1, and EGR1) or 30 min (others). The cells were washed with PBS twice and lysed with 200 µL RIPA buffer containing protease inhibitor cocktail (MedChem Express, Monmouth Junction, NJ). Lysed samples were cooled on ice for 15 min and centrifuged at 12,000 rpm for 10 min at 4 °C. Protein concentrations were measured with a BCA kit (Beyotime, Beijing, China). Equal amounts of lysate were separated by 10% sodium dodecyl sulfate-polyacrylamide-gel electrophoresis and transferred to polyvinylidene difluoride membranes.
Proteins were crosslinked to DNA by adding 270 µL of 37% formaldehyde to 10 mL culture medium (final concentration, 1%

Dual luciferase promoter-reporter assay
Transient transfection was conducted with liposome transfection reagent. RBLs were seeded in 24-well plates and transfected with the AP-1 luciferasereporter plasmid pAP1-Luc (Promega, USA), and compared to control plasmid pGL3-basic or pGL3control. Activities of Firefly luciferase in pGL3-reporters and Renilla luciferase in pRL-TK were determined 48 h later by dual luciferase reporter assays following the manufacturer's protocol (Promega, Madison, WI). Each transfection was performed in triplicate and each assay was repeated at least three times.

Flow cytometric assay
To study the effects of T-5224 on the differentiation of progenitor cells to BMMCs, cells were treated with 20 µM T-5224. Culture media were changed every 3 days and differentiation was assessed every 6 days. Single cell suspensions from the BMMCs were stained with the antibody mix including CD177-PE and FcεRIα-APC (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). To set the gates, flow cytometric dot plots were based on comparison with isotype controls, fluorescence minus one, and unstained cells. The percentage of CD177 and FcεRIα double-positive cells were measured by flow cytometry (CytoFLEX, Beckman Coulter, Miami, FL).

IgE-mediated passive cutaneous anaphylaxis (PCA) mouse model
A PCA mouse model was used to assess T-5224 effects on allergic reactions [31]. In this model, extravasation reflects vessel permeability. Twenty-four mice were randomized into four groups (6/group): PBS; DNP-HSA; DNP-HSA plus T-5224; and DNP-HSA plus ketotifen. Left ears were injected subcutaneously with anti-DNP IgE (500 ng in 0.9% saline). Twentyfour hours later, IgE-sensitized mice were given an intraperitoneal (i.p.) injection of T-5224 (60 mg/ kg) or ketotifen (50 mg/kg), and 1 h later challenged with a tail vein injection of 200 µg of DNP-HSA in 1× PBS with 0.5% Evans blue dye. Mice were euthanized 1 h after the challenge; their ears were removed and dissolved in 700 μL formamide overnight at 62 °C. Extravasated dye was quantitated with 620-nm spectrophotometry (Bio-Rad, Hercules, CA).

Ovalbumin (OVA)-induced active systemic anaphylaxis (ASA) mouse model
The OVA-induced ASA model is widely used to examine immediate-type hypersensitivity, which is strongly associated with MCs [32]. A cohort of 24 mice was randomized into four groups (6/group): negative control; OVA sensitization/challenge; OVA sensitization/ challenge plus T-5224; and OVA sensitization/challenge plus ketotifen (50 mg/kg). The mice were sensitized on days 0 and 7 with OVA (100 µg i.p. with 2 mg alum adjuvant). From experimental day 9 to 13, mice received T-5224 (60 mg/kg) and ketotifen (50 mg/ kg), i.p. on alternating days. On day 14, the mice were challenged with 200 μg OVA (i.p.) and then monitored for 90 min, during which their rectal temperatures were measured every 10 min. Immediately thereafter, orbital venous plexus blood was sampled for serum histamine and IL-4 level measurement by enzymelinked immunosorbent assay (ELISA).

Oxazolone (OXA)-induced atopic dermatitis model
In the OXA-induced atopic dermatitis model, massive inflammatory cell infiltration leads to skin thickening [33]. To produce the model [34], the right ears of Balb/c mice were sensitized with 20 μL of 1% OXA (TargetMol, Shanghai, China) in 4:1 (v/v) acetoneolive oil mixture (experimental day 0). Plain acetone/ sesame seed oil was injected in the left ears to obtain within-animal controls. On days 7, 9, and 11, the ears were challenged with 20 μL of 1% OXA three times a week for 1 week. At the same time, T-5224 (60 mg/kg, i.p.) was administered. On day 13, the mice were sacrificed, ear samples were collected for further analysis. Ear swelling change was used as a measure of allergic response.

Statistical analyses
All clinical and histological evaluations were performed in a blinded manner. Data were analyzed in Prism 8 (GraphPad, La Jolla, CA). For normally distributed quantitative variables, means of triplicate values are reported with standard deviations (SDs). Kruskal-Wallis tests were performed to detect significant differences between treatment groups. Oneway analyses of variance (ANOVAs) and Dunnett's post-hoc tests were used to find differences between treatment groups in the in vivo experiments. Differences was considered significant at p < 0.05.

c-Fos upregulates EGR1 transcription through EGR1 promoter binding
Network analysis suggested c-Fos protein-protein interactions with a number of MC activation-associated proteins, including early growth response 1 (Egr1), Egr2, Egr3, IL-4, and Ccl2 ( Fig. 2A). Subsequently, qRT-PCR and western blot analyses showed that IgE/Ag-stimulated MCs had increased Egr1 mRNA and EGR1 protein levels, compared to nonstimulated MCs (Fig. 2B, C), consistent with our RNA-Seq data. Treatment with shFos decreased EGR1 mRNA and protein levels, compared to non-treated controls (NCs) (Fig. 2D, E). To verify that EGR1 is activated by c-Fos in MCs, we obtained the first three Fos-binding sequences within the Egr1 promoter from JASPAR (Fig. 2F). A ChIP assay confirmed that Fos binds the Egr1 promoter (Fig. 2G).
Regarding AP1-inhibition effects on MC activation signaling, qRT-PCR and western blot analyses showed that T-5224 pretreatment decreased transcription and translation of EGR1 in IgE-stimulated RBLs, with the effect being stronger with the high dose (20 µM) than with the low dose (10 µM) (Fig. 3F, G). Western blot analysis showed that T-5224 inhibited activation of MAPKs (p38, JNK, and ERK) in activated RBLs, as evidenced by decreased levels of p-p38, p-JNK and p-ERK1/2 (Fig. 3H), with the high dose resulting in more pronounced reductions.

T-5224 inhibits IgE-mediated PCA in vivo
As shown in Fig. 5, IgE/Ag stimulation induced robust in vivo PCA reactions, evidenced by dye diffusion, ear thickening, and dye extrusion from injected ears. T-5224 attenuated these responses, as shown by decreased dye diffusion (Fig. 5A, B), reduced ear thickness (Fig. 5C), and diminished dye extrusion (Fig. 5D) compared to the activated group not given T-5224. The effects of T-5224 resembled those of the anti-allergy drug ketotifen.

T-5224 attenuates MC-induced allergic reaction in OVA-induced ASA model
To establish the OVA-induced allergy model (Fig. 5E), mice were sensitized with repeat administrations of OVA Ag plus alum adjuvant; then an allergic reaction was induced with an OVA challenge (Fig. 5E).
During the 90-min post-OVA challenge period, rectal temperatures of control mice decreased over 40-50 min, and then recovered. This reduction was attenuated by injection of T-5224 or ketotifen (positive control) (Fig. 5F). Notably, 80 min after the challenge induction, the mean rectal temperature for the OVA-sensitized group remained low (29.36 ± 0.93 °C) whereas that of the T5224-treated OVA-sensitized group (34.60 ± 0.83 °C) was similar to that of non-sensitized controls. Concomitantly, serum histamine levels increased after the OVA challenge (non-sensitized control group, 1126 ± 81 ng/mL) and this increase was diminished significantly by T-5224 (211 ± 122 ng/mL) (Fig. 5G). The similar result was found in the serum IL4 levels (Fig. 5H).

T-5224 suppresses OXA-induced atopic dermatitis symptoms in mice
The atopic dermatitis model mice exhibited marked ear skin thickening and these dermatitis symptoms were attenuated by T-5224 (p < 0.01; Fig. 6A, B). Toluidine blue staining showed that the infiltration of mast cells in the dermis was observed in the dermis of the control group. The number of mast cells in the ear dermis was decreased significantly after T-5224 treatment (p < 0.01; Fig. 6C, D). In addition, increased the permeability of blood vessels in allergic skin allowed eosinophils to infiltrate into tissue. H&E staining revealed that eosinophil infiltration of the ear skin was attenuated by treatment in the T-5224 groups compared with the AD group (p < 0.01; Fig. 6E, F).
The results indicated T-5224 had a protective effect against OXA-induced AD in mice.

Discussion
We confirmed that DNP-HSA triggered MC degranulation, as evidenced by β-hexosaminidase release and upregulation of allergic reaction-related genes, thereby confirming the validity of our IgE-activated RBL model of activated MCs [36]. Transcriptome analysis revealed that cFos and FosB were upregulated by DNP-HSA stimulation of RBLs. Stimulated RBLs had increased c-Fos phosphorylation. Inhibition of c-Fos/AP1 with T-5224 reduced degranulation, suppressed MC activation-associated morphological changes, and attenuated MC activation effects on IL4 and EGR1 expression. IgE/Ag-stimulated MCs showed upregulation of EGF1, both transcriptionally and translationally, that could be attenuated by suppression of c-Fos activity. We demonstrated in a ChIP assay that Fos bound the Egr1 promoter region, indicating that Egr1 transcription was likely activated by c-Fos binding its promoter during MC activation. Furthermore, T-5224 pretreatment decreased EGF1 transcription and translation and inhibited MAPK activation in activated MCs. Finally, in multiple models, T-5224 had a protective effect against the development of allergic responses including swelling, body temperature reduction, histamine release, and infiltration of MCs and eosinophils. MCs represent a potential therapy target to alleviate allergic reactions [37]. Extending prior findings implicating Fos upregulation in IgE/FcεRI-activated MCs [13], we demonstrated that IgE/Ag treatment enhances c-Fos transcription and translation, and that c-Fos acts on the Egr1 promoter to upregulate EGR1 expression. EGR1 is a zinc-finger transcription factor that regulates the expression of various inflammatory cytokines following FcεRI activation, including TNF, IL6, IL13, and IL4 [38,39].
Previously, EGR1 was shown to induce IL4 secretion in FcεRI-activated MCs by binding the Il4 promoter [21]. IL4 enhances IgE-mediated MC responses and is a necessary mediator of allergy development [20,40]. IL4 is critical for directing Th2 cell differentiation and B cell antibody class switching to allergenic IgEs [41]. FcεRI-mediated MC-derived IL4 production is important for allergy pathogenesis [42]. Our results indicating that c-Fos binding of the Egr1 promoter can drive EGR1 expression are consistent with a new cFos/Egr1/ IL-4 axis of MC activation. AP1 activity is positively associated with the development of IgE-mediated allergic diseases, and AP1 inhibition can suppress allergic asthma symptoms [43][44][45][46]. To examine the allergenic role of AP1, we employed T-5224, a selective c-Fos/AP1 inhibitor produced by Aikawa et al. based on the crystal structure of the AP1-DNA complex [47]. T-5224 has been shown to suppress collagen-induced arthritis [47], lipopolysaccharide-induced liver/kidney injury [48,49], intervertebral disc degeneration and associated pain [50], and breast cancer [51]. However, it was not known whether T-5224 may impede MC-induced allergic responses.
Our results showing that T-5224 reduced FcεRImediated degranulation in MCs were similar to previously reported effects of palbociclib (cyclin-dependent kinase inhibitor) and tozasertib (Aurora kinase inhibitor) [22,52]. Mechanistically, T-5224 decreased IgE/ Ag-induced expression of EGR1 and IL4 and also blocked MAPK activation in activated MCs. A variety of established medicines, including omeprazole (proton pump inhibitor) [7], tozasertib [52], and berberine (adenosine monophosphate-activated protein kinase stimulator) [53] inhibit MC activation-associated MAPK signaling. Meanwhile, T-5224 did not affect maturation of precursor cells into MCs (Additional file 1: Figure S5). Together, these data support exploration of T-5224 for potential repurposing as a MC activation inhibitor. We proceeded to conduct in vivo studies in murine models of allergic disease employing doses appropriate for clinical application (30-300 mg/ kg).
Previously, oral administration of T-5224 (30 mg/kg per day for 3 weeks) was shown to resolve collageninduced arthritis in a preclinical model [47]. Additionally, oral administration of T-5224 (300 mg/kg) has been shown to ameliorate lipopolysaccharide-induced liver/kidney injury in mice through a mechanism that involves reducing production of proinflammatory cytokines [48,49]. T-5224 was developed into a potential therapeutic agent for rheumatoid arthritis and has advanced into phase-II clinical trials [54]. The major metabolites of T-5224, glucuronides, occur predominantly in the liver [55]. The T-5224 treatments employed in our PCA and ASA model mice, which translate to a 60-mg/kg dose, showed robust effectiveness without signs of cytotoxicity. We proceeded to test T-5224's ability to suppress MC activation in a model of atopic dermatitis, which is characterized by MC activation and intra-dermal infiltration of MCs that release vasoactive and proinflammatory mediators [56,57]. T-5224 attenuated MC and eosinophil infiltration in OXA-induced allergic skin, consistent with a protective effect against allergic dermatitis.
In conclusion, the present experiments demonstrated a previously unknown molecular mechanism whereby IgE/Ag promotes MC activation via a c-Fos/Egr1/IL4 axis. The c-Fos/AP1 inhibitor T-5224 reduced MC degranulation and cytokine production and attenuated allergic responses in vivo. The ability of T-5224 to inhibit MC degranulation and allergic disease presentation suggest that it should be considered as a candidate for repurposing into an allergic/ inflammatory disease pharmacotherapy. Clinical trials aimed at determining the anti-allergy efficacy and toxicity of T-5224 are warranted.  Figure S1. Volcano plot of RNA-seq data. Figure S2. Results of KEGG pathway analysis of upregulated DEGs. Figure  S3. Results of GO enrichment analysis of upregulated DEGs. Figure S4.
Non-effect of T-5224 on maturation of precursor cells to MCs. Table S1.
Primers sequences for quantitative real-time PCR.