Seventy-five male, 6–8-week-old C57BL/6 mice weighing 20–22 g, were purchased from Shanghai Laboratory. Animal Research Center (Shanghai, China). The mice were housed in a constant temperature of 22±2°C and relative humidity of 60±10% environment under 12 h light/dark cycles. The mice were then inoculated with NTHi trans-tympanically at a dose of 5×10⁷ CFU in each mouse. In the control group, saline was inoculated. For quercetin studies in vivo, experimental mice were then intraperitoneally (i.p) injected with quercetin (20, 40 and 80 mg/kg) purchased from Shaanxi Huike Botanical Development Co., Ltd. (Shaanxi, China) 2 h after NTHi inoculation for 6 h. For mRNA analysis, mice were sacrificed 6 h post-NTHi inoculation. Total RNA was extracted from the dissected mouse middle ear. Finally, all mice were sacrificed and the eyeball blood was collected and centrifuged at 15,000 × g for 20 min prepared for following research. Middle ear were then isolated immediately on ice and stored at −80°C for further research. All animal studies here were performed in accordance with the guidelines of, and were approved by, the First Affiliated Hospital of Jinan University.

Human middle ear epithelial cells (HMEECs), renal epithelial cells (HK2), and mouse lung epithelial cells (MLE-12) were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). The human liver epithelial cells (L02) and human lung epithelial cells (BEAS-2B) were purchased from the KeyGen Biotech (Nanjing, China). HMEECs, L02 and BEAS-2B cells were routinely cultured in RPMI-1640 medium, containing 10% fetal bovine serum (FBS) (both from Gibco, Grand Island, NY, USA), 1% penicillin/streptomycin. The cell lines HK2 and MLE-12 were cultured in DMEM (Gibco) supplemented with 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin. The cells were kept in a humidified atmosphere with 5% CO₂ and 95% humidity at 37°C in an incubator.

Clinical isolates of NTHi strains 2019, 12 and 2866 were included in the present study (15,16). NTHi was cultured on chocolate agar plate under 5% CO₂ atmosphere for 16 h, which is followed by culture overnight in the brain heart infusion (BHI) broth supplemented with 10 µg/ml hemoglobin and 3.5 µg/ml NAD (BD Biosciences, Franklin Lakes, NJ, USA). Next, bacteria were then subcultured in fresh BHI broth (5 ml) and the growth situation was monitored by assessment of optical density (OD). Bacteria in log phase were collected, washed and suspended in medium for experiments in vitro and isotonic saline for experiments in vivo. For in vitro experiments across our study the cells were cultured with NTHi at a multiplicity of infection (MOI) of 50. Cells were induced with NTHi for 6 h, or as indicated in our study. For suppression study, cells were pretreated with the specific inhibitor for 2 h ahead of NTHi induction. For post-treatment studies cells were treated with quercetin at different concentrations (40, 80 and 120 µM) 2 h after NTHi stimulation. NF-κB inhibitor, PDTC, and IKKα inhibitor, MRT67307 as well as p38 inhibitor SB203580 were purchased from Biovision (Milpitas, CA, USA), MedChem Express (Monmouth Junction. NJ, USA) and Beyotime (Shanghai, China) respectively.

Concentrations of CXCR4 in the middle ear effusions of mice were determined by the ELISA with the mouse enzyme immunoassay sets (R&D Systems, Inc., Minneapolis, MN, USA) following the manufacturer's instructions. The samples were performed in duplicate.

The expression plasmid investigation, for mutations (Mut) of TLR3, TLR4, MyD88, IRAK1, TAK1, TRAF6, p38, and constitutively active form of NF-κB have been described previously (17). NF-κB luciferase reporter vector was purchased from Promega (Madison, WI, USA). All transient transfections were carried out with Lipofectamine™ 2000 Transfection Reagent (Lipo 2000; Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. HMEECs were analyzed after transfection for 48 h. The empty vector was transfected as a control. The pSV-β-galactosidase vector was performed as a control for luciferase activity assay. The luciferase activity as well as pSV-β-galactosidase activity was determined with luciferase assay system and pSV-β-galactosidase (both from Promega). NF-κB luciferase activity was normalized regarding pSV-β-galactosidase.

Human siRNA (TLR3, TLR4, MyD88, IRAK1, TRAF6 and TAK1; IKKα; NF-κB) were obtained from Generay Biotech Co., Ltd. (Shanghai, China). HMEECs were transfected with 20 nM siRNA with Lipofectamine™ 2000 transfection reagent (Lipo 2000; Invitrogen) following the manufacturer's protocol.

The HEMMCs and the middle ear tissues were harvested. Proteins were extracted from the middle ear tissue samples using T-PER tissue protein extraction reagent kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. Protein concentrations were determined by BCA protein assay kit, and equal amounts of protein were loaded per well on a 10% sodium dodecyl sulphate-polyacrylamide (SDS) gel. Subsequently, proteins were transferred onto polyvinylidene difluoride (PVDF) membrane. The resulting membrane was blocked with Tris-buffered saline containing 0.05% Tween-20 (TBS-T), supplemented with 5% skim milk (Sigma) at room temperature for 2 h on a rotary shaker, and followed by TBS-T washing 5 times. The specific primary antibody, diluted in TBST, was incubated with the membrane at 4°C overnight. Subsequently, the membrane was washed with TBS-T followed by incubation with the peroxidase-conjugated secondary antibody at room temperature for 2 h. The immunoactive proteins were detected by using an enhanced chemiluminescence western blot detection kit. Western blot bands were observed using GE Healthcare ECL western blotting analysis system and exposed to Kodak X-ray film. The primary antibodies used are shown in Table I.

Total RNA from the middle ear tissue samples and cells under different conditions were isolated using TRIzol (Invitrogen) following the manufacturer's instructions. The cDNA was synthesized using SuperScript II reverse transcriptase (Thermo Fisher Scientific). Quantitative PCR was performed with SYBR-Green Real-Time PCR Master mix (Thermo Fisher Scientific). The quantitative expression data were collected and analyzed by a 7900 Real-time PCR system (Applied Biosystems, Foster City, CA, USA). Primers were designed to determine endogenous genes and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) using as the endogenous control. Human (h) and mouse (m) primer sequences were: forward hCXCR4, (5′-3′) TTA CCT ATA TTC TCG GCG TGG ACA G and reverse primers, (5′-3′) CTC GAT GGT CAT GAC TAA GTG TTC; forward mCXCR4, (5′-3′) GGA TAC TTG TAC GAG CAA GC and reverse primers, (5′-3′) GGA GTG GTG CTT GAG TTG ATT; forward GAPDH, (5′-3′) GAC GGA GCT GAG AAC ATG T and reverse primers, (5′-3′) TGT CCG CGT ATT ATG AGA TG.

For fluorescent analysis, the mouse middle ear tissue samples were carefully isolated and fixed in 4% paraformaldehyde for 16 h after cold 4% paraformaldehyde perfusion. Then, optimum cutting temperature (OCT) package tissues were cut to 20–30 µm sections. The tissues were incubated with primary antibodies (p-IKKα and p38) at 4°C overnight after deparaffinized and rehydrated. Fluorophore-conjugated secondary antibodies were treated 1 h at 25°C. The Alexa Fluor 488 labeled anti-rabbit or anti-mouse secondary antibodies (Invitrogen) were used. Sections were subjected to immunofluorescence staining via epifluorescence microscopy (Sunny Co., Beijing, China). Leica TCS SP5 confocal microscope (Leica, Richmond Hill, ON, Canada) was used to obtain images in a blinded manner with respect to treatment groups.

Histopathologic evaluation was performed on mice. Mouse middle ear tissue samples were fixed with 10% buffered formalin, imbedded in paraffin, and sliced. After hematoxylin and eosin (H&E) staining, pathological changes of the tissues were observed under a light microscope.

Data are expressed as means ± SEM Treated tissues and the corresponding controls were compared using GraphPad PRISM (version 6.0; GraphPad Software, La Jolla, CA, USA) by a one-way ANOVA with Dunn's least significant difference tests. Differences between groups were considered significant at p<0.05.

Epithelial cells are reported to be of great importance against various injuries under different stresses through inflammation response regulation (18,19). Hence, here we attempted to calculate if NTHi could induce CXCR4 activation in epithelial cells from human middle ear (HMEEC). As shown in Fig. 1A and B, CXCR4 expression was highly activated after NTHi stimulation, shown in a dose- and time-dependent manner. In addition, CXCR4 protein levels in HMEEC cells were significantly upregulated due to NTHi exposure through ELISA method (Fig. 1C). Moreover, CXCR4 levels in human renal epithelial HK-2 cells, human liver epithelial L02 cells, human lung epithelial BEAS-2B cells as well as in the mouse lung epithelial MLE-12 cells were detected by RT-qPCR assays. CXCR4 gene levels were highly upregulated in HMEECs after NTHi stimulation in a time-dependent manner (Fig. 1D–G). Another two stains of NTHi, 12 and 2866, reported to be effective in inducing otitis media, were also used to investigate whether CXCR4 could be upregulated. As shown in Fig. 1H, increased CXCR4 mRNA levels were observed in NTHi-induced HMEECs compared to the Con group. Consistently, CXCR4 mRNA levels were also stimulated in the mouse middle ear tissue samples induced by NTHi (Fig. 1I). Collectively, the data above indicated that CXCR4 may be of essential importance in NTHi-induced otitis media.

TLRs are well known as vital cell surface receptors, playing an essential role in regulating inflammation response against various pathogens under different conditions (20). Up until now, ~13 TLRs have been reported. Among these receptors, TLR3 and TLR4 are significant in realizing LPS, which is a Gram-negative bacteria characteristic (21). TLRs are involved in NTHi-induced inflammatory response, which may be also related to CXCR4 activation. Thus, here the HMEECs were transfected with TLR3 and TLR4 mutants. Of note, TLR3 mutant in high expression apparently reduced NTHi-induced CXCR4 expression at the gene levels (Fig. 2A). However, no significant difference related to CXCR4 expression was observed in TLR4-mutation cells after NTHi stimulation compared to the TLR3-Mut⁻/TLR4-Mut⁻ group. The TLRs down-streaming signals, MyD88/IRAK1, and TRAF6/TAK1 were also investigated. As shown in Fig. 2B and C, we found that both MyD88 and IRAK1 mutations downregulated CXCR4 mRNA levels in NTHi-treated cells. Also, TRAF6/TAK1 mutations dramatically decreased CXCR4 mRNA expression through RT-qPCR analysis. Moreover, depletion of TLR3 with TLR3 siRNA also decreased CXCR4 expression from the gene levels. In the results above, TLR4 knockdown showed no significant difference on CXCR4 gene expression (Fig. 2D). Furthermore, MyD88 and IRAK1 silence through specific RNA knockdown reduced CXCR4 expression in HMEECs after NTHi induction (Fig. 2E). Finally, in order to further confirm TRAF6 and TAK1 in CXCR4 regulation, we found that TRAF6 and TAK1 silence considerably reduced CXCR4 gene expression levels (Fig. 2F). The data above indicated that TLR3/MyD88/IRAK1/TRAF6/TAK1 signaling pathway was, at least partly, involved in NTHi-induced otitis media.

According to previous studies, IKKα and p38 MAPK are two important signaling pathways in regulating otitis media progression induced by NTHi (22). IKKα is considered as an essential molecule, activated by TLRs/MyD88 signaling pathway, contributing to inflammation response (23). As shown in Fig. 3A, we found that IKKα phosphorylation was upregulated due to NTHi stimulation. Next, IKKα inhibitor MRT67307, was used to suppress IKKα expression, and with the increasing of MRT67307 concentration, CXCR4 mRNA levels were reduced, which was comparable to the NTHi-treated group in the absence of MRT67307 (Fig. 3B). Moreover, IKKα mutation significantly reduced CXCR4 mRNA levels in HMEECs stimulated by NTHi (Fig. 3C). Similarly, IKKα knockdown indicated that CXCR4 mRNA levels were decreased after NTHi stimulation (Fig. 3D). Fig. 3E suggested that IKKα was successfully silenced for knockdown to inhibit its activation. In contrast, IKKα activator was used to improve its phosphorylation, leading to CXCR4 upregulation in a dose-dependent manner (Fig. 3F). Additionally, p38 phosphorylated levels were evaluated after NTHi stimulation in HMEECs. p38 phosphorylation was also upregulated with the increasing time of NTHi induction (Fig. 3G). Further, p38 inhibitor SB203580 usage reduced CXCR4 mRNA levels in a dose-dependent manner (Fig. 3H). Finally, p38 mutation was included to decrease CXCR4 mRNA levels through RT-qPCR analysis (Fig. 3I). The data above illustrated that IKKα and p38 signaling pathways were included in NTHi-induced otitis media.

The findings above revealed that IKKα and p38 had a close relationship with otitis media (24). NF-κB is known as an essential signal in regulating inflammation response, which is regulated by IKKα activity (25). As shown in Fig. 4A and B, CXCR4 gene and protein levels were significantly reduced for MRT67307 and SB203580 treatment alone, especially in combination. Of note, NF-κB inhibitor, PDTC, was used here to suppress NF-κB activation. CXCR4 mRNA levels were highly reduced, particularly in the highest concentration of PDTC (Fig. 4C). Pretreatment with PDTC markedly abolished NTHi-triggered NF-κB promoter-induced luciferase activity (Fig. 4D). Depletion of NF-κB by the use of NF-κB siRNA, reduced CXCR4 mRNA expression levels (Fig. 4E). Notably, Fig. 4F overexpression of NF-κB further promoted NTHi-triggered CXCR4 mRNA expression. In oder to calculate if p38 triggers CXCR4 expression through NF-κB, HMEECs were transfected with NF-κB combined with SB203580 or not, prior to stimulation by NTHi. SB203580 reduced CXCR4 mRNA expression in NF-κB-transfected HMEECs (Fig. 4G). In addition, pretreatment with SB203580 apparently reduced NTHi-stimulated NF-κB promoter-driven luciferase activity, which as shown in Fig. 4H. p38 mutation decreased NF-κB promoter activity induced by NTHi (Fig. 4I). The data above indicated that p38-regulated CXCR4 expression induced by NTHi relied on NF-κB signaling pathway.

The findings above indicated the possible molecular mechanism by which otitis media was induced for NTHi, which was related to CXCR4 expression levels. Quercetin has been reported before to have an essential role in suppressing inflammation response through various signaling pathways, including TLR4/NF-κB and MAPKs (26). Thus, we attempted to explore if quercetin could improve NTHi-induced otitis media in vivo and in vitro through targeting CXCR4. As shown in Fig. 5A and B, after NTHi stimulation, overexpressed CXCR4 gene and protein levels were reduced for quercetin administration in a dose-dependent manner. Also, in other NTHi strains, 2019, 12 and 2866, CXCR4 high expression was significantly downregulated for quercetin treatment under different concentrations in HMEECs stimulated by NTHi (Fig. 5C–E). Finally, in in vivo study, we found that CXCR4 mRNA levels in the middle ear of mice were upregulated by NTHi induction, which was downregulated for quercetin treatment in a dose-dependent manner (Fig. 5F). The histologic sections and the mucosa thickness in the mouse middle ear were observed through H&E staining (Fig. 5G). By assessment, the mucosa in the roof of NTHi-treated mice with otitis media was much thicker compared to the control ones, which was reduced for quercetin treatment at different concentrations. The data above indicated that quercetin has potential value for ameliorating NTHi-induced inflammation seen in otitis media.

Next, we attempted to investigate the molecular mechanism by which quercetin suppressed CXCR4 expression. IKKα and p38 signaling pathway activation has been revealed to be related with NTHi-stimulated CXCR4 expression, we calculated the role of quercetin in the two signaling pathways. Quercetin abolished NTHi-induced IKKα and p38 phosphorylation through western blot analysis (Fig. 6A). Additionally, quer-cetin suppressed IKKα activator-induced CXCR4 expression in vitro (Fig. 6B). Finally, fluorescent analysis indicated that IKKα and p38 phosphorylated levels induced by NTHi were apparently reduced for quercetin treatment, which was in a dose-dependent manner in vivo (Fig. 6C and D). Collectively, the data above indicated that quercetin suppressed NTHi-induced CXCR4 expression and activation by inhibiting IKKα and p38 MAPK signaling pathways.