All mouse experimental protocols complied with the guidelines of the Beijing Laboratory Animal Welfare and Ethics Committee, and were approved by the Beijing Association for Science and Technology (the approval ID is SYXK-2009-0423). All experiments with live H5/H7 subtype viruses were performed in a biosafety level 3 containment laboratory (the approval number is CNAS BL0017) approved by the Ministry of Agriculture of the People’s Republic of China.

The avian influenza viruses H5N1 (A/Chicken/Henan/1/04) (Hu et al., 2012) and H7N2 (A/Chicken/Hebei/2/02) were isolated from infected chicken flocks, and propagated in the allantoic cavities of 10-day-old embryonated chicken eggs for 24–48 h at 37°C. The working stocks of human influenza virus H1N1 (A/WSN/33) were generated in MDCK cells. Virus titers were determined by standard plaque assays. The 50% lethal dose (LD₅₀) in mice was determined as previously described (Hu et al., 2012). The mouse mastocytoma cell line P815 and the Madin-Darby canine kidney cell line MDCK were cultured as previously described (Hu et al., 2012).

Cell monolayers were formed in tissue culture plates by seeding 6-well (1 × 10⁶ cells/well) or 12-well (5 × 10⁵ cells/well) plates, washed with DMEM and infected with viruses at a multiplicity of infection (MOI) of 0.1 for 1 h at 37°C. After incubation, cells monolayers were washed and DMEM supplemented with 1% bovine serum albumin was added to each well and incubated for the indicated times. Polyinosine-polycytidylic acid (polyI:C), a synthetic mimic of viral double-stranded RNA, was used as a positive control. Liposome-encapsulated PolyI:C (LE-PolyI:C) used in this study was prepared as described previously (Wong et al., 1999), diluted to a final concentration of 10 μg/ml and incubated with cells at 37°C for the indicated times.

Female BALB/c mice (8–10 weeks) were purchased from Vital River Laboratories (Beijing, China), and feed pathogen-free food and water in independent ventilated cages. Mice were first anesthetized with Zotile^® (Virbac, Carros, France), and then infected intra-nasally with PBS-diluted H5N1 virus (5LD50) or PBS alone. The nose and lung tissues were then collected 6 days post-infection.

Tissue samples were fixed in 4% neutral formalin, embedded in paraffin, and serially cut at a thickness of 5 μm. Cultured cells were fixed on a polylysine-coated slide with 4% formaldehyde, and blocked with 3% BSA. To visualize surface receptors, slides containing fixed tissues or cells were directly stained with fluorescein- Sambucus nigra bark lectin (SNA, specific to SAα2,6-Gal) or fluorescein- Maackia amurensis lectin I (MAA-I, specific to SAα2,3-Galβ(1-4)GlcNAc). To confirm the specificity of lectin binding, monolayers were washed and treated with 250 mU/ml of neuraminidase from Clostridium perfringens (New England BioLabs, Beijing, China) for 3 h prior to lectin staining. To detect tryptase expression or IAV nucleoprotein (NP) antigen, cells were permeabilized with 0.5% Triton X-100 before blocking, then tissue sections or cell slides were either incubated with a rabbit anti-mast cell tryptase monoclonal antibody (Abcam, [EPR8476], ab134932) for 2 h at room temperature, or a mouse anti-IAV NP monoclonal antibody (Abcam [AA5H], ab20343) at 4°C overnight. After washing three times with PBS-T, tissue sections were further incubated with a Texas red-conjugated goat anti-mouse or rabbit secondary antibody, and cell slides were incubated with a FITC-conjugated goat anti-mouse secondary antibody (Abcam) for 1 h at room temperature. To visualize the nuclei, all slides were stained with 3 μg/ml 4′,6′-Diamidine-2-phenylindole (DAPI) (Sigma–Aldrich) for 5 min at room temperature and then examined under a laser scanning confocal microscope (Leica TCS SP5 II, Leica Microsystems, Wetzlar, Germany).

Cultured P815 cells (1 × 10⁶) were pelleted, washed twice with DMEM, once with flow buffer (PBS with 2% FBS) and then re-suspended in 200 μl of fluorescein- SAA or fluorescein- MAA I at different dilutions. The cells were incubated for 1 h at 4°C, then washed and re-suspended in flow cytometry buffer for analysis on a BD FACSCalibur using Cell Questpro software (BD Biosciences, California).

Cells were trypsinized and fixed using 2.5% (v/v) glutaraldehyde in PBS for 2 h at 4°C. Cells were then washed with PBS, post-fixed in 1% osmium tetroxide, and washed and dehydrated in series of ethanol solutions. The dehydrated pellets were embedded in epoxy resin, and 70-nm sections were cut. Then the sections were placed on copper sieves and stained with uranyl acetate and lead citrate. Images were obtained using a JEM-1230 TEM (JEOL, Japan Electronics Co., Ltd, Tokyo, Japan).

Total RNA was extracted from cells in Trizol reagents (Invitrogen, Carlsbad, CA, USA). DNase I-treated RNA (0.2 μg) was reverse transcribed into cDNA using random or universal primers for IAV (Uni 12) (Hoffmann et al., 2001) with an EasyScript First-Strand cDNA Synthesis Super Mix (TransGen Biotech, China) according to the manufacturer’s instruction. Reactions were performed in triplicate using a Power SYBR^® Green PCR Master Mix (Applied Biosystems, Warrington, UK) and the Applied Biosystems 7500 system. The mRNA expression levels of the genes were normalized to β-actin, compared with mock-infected cells, and quantified by the 2^-ΔΔCT method. The sequences of the primers are listed in Supplementary Table S1. The amplifications were performed as follows: a 10 min hot start at 95°C, followed by 40 cycles of denaturation at 95°C for 15 s, annealing at 55°C for 35 s, and extension at 72°C for 40 s.

Cell-free supernatants were acquired by centrifugation, then the levels of cytokines and chemokines were analyzed by the Mouse Cytokine Array Panel A (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions, which could provide parallel determination of the relative levels of 40 kinds of selected mouse cytokines and chemokines.

The concentration of IL-6, IFN-γ, RANTES, IP-10, IFN-α, IFN-β, and TNF-α in the supernatant of cell cultures was determined using ELISA kits (eBioscience, San Diego, CA, USA) according to the manufacturer’s instructions.

Cells were either mock-treated, LE-PolyI:C-treated, or infected with IAV viruses at a MOI of 0.1. At 6 h post-infection, the cells were washed with cold PBS and lysed for 15 min on ice with RIPA lysis buffer containing 50 mM Tris-HCl (pH7.4), 150 mM NaCl, 1% NP40, 0.25% sodium deoxycholate and a protease inhibitor cocktail (Beyotime Institute of Biotechnology, Beijing, China). Lysed cells were pelleted and the supernatants were incubated with the indicated antibodies (anti-TLR3 antibody or an isotype IgG antibody from Abcam) for 2 h at 4°C with gentle shaking. The samples were then added to a preprocessed EZview^TM Red Protein A/G Affinity Gel (Sigma) and incubated for 6 h. After washing three times with lysis buffer, the beads were boiled in SDS loading buffer, and then analyzed by immunoblot with the indicated antibodies (anti-TLR3 antibody or anti-TRIF antibody from Abcam).

Cell lysates were prepared using lysis buffer as described above, and protein concentrations were determined using a BCA protein assay kit (Beyotime Institute of Biotechnology). Equal amounts of protein were separated by SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore, Beijing, China). The membranes were blocked using 5% non-fat dry milk (BD Biosciences) at room temperature for 2 h, and then incubated overnight at 4°C with antibodies (anti-TLR3 antibody and anti-TRIF antibody from Abcam; Influenza A NS1 antibody from Santa Cruz Biotechnology, Dallas, Texas, USA). After three 10 min washes in Tris-buffered saline containing 0.05% Tween (TBST), the membranes were incubated for 1 h at room temperature with the appropriate horseradish peroxidase-conjugated secondary antibody. Protein bands were visualized using the Western Lightning Plus-ECL (Perkin Elmer, MA, USA). β- Actin was used as a loading control.

P815 cell monolayers were pre-incubated with TLR3/dsRNA complex inhibitor (Calbiochem, Darmstadt, Germany) at a concentration of 25 and 50 μM (diluted with DMSO) for 12 h. They were then were infected with virus at a MOI of 0.1, exposed to LE-PolyI:C, or mock treated as described above. The same concentration of inhibitors was immediately added after 1 h of viral incubation. Samples were collected 24 and 36 h after infection.

Statistical analysis was performed by one-way ANOVA using the SPSS software suite (version 12.0), and a P-value of <0.05 was considered statistically significant. Results were expressed as mean ± standard deviation (SD) of at least three independent experiments.

Influenza viral HA proteins initiate infection by interacting with sialic acid residues coating the surface of host cells. In general, human influenza viruses and S. nigra bark lectin (SNA) preferentially bind α-2,6- linked SA receptors, while avian influenza viruses and M. amurensis lectin (MAA) predominantly bind to α-2,3- linked SA receptors (Springer et al., 1969; Suzuki et al., 2000; Ibricevic et al., 2006; Shinya et al., 2006). Here we used the mouse mastocytoma cell line, P815, as a mast cell model. To determine the susceptibility of P815 cells to influenza viruses, we first analyzed the distribution of surface SA receptors by lectin staining as described in the materials and methods. Both α-2,3- and α-2,6- linked SA receptors were expressed on the surface of P815 mouse mastocytoma cells, with the intensities of SNA being visually stronger than one of the two isoforms of MAA (Figure 1A). We treated P815 cells with a broad-spectrum neuraminidase to cleave sialic moieties abolished lectin binding, and confirmed the specificity of SNA and MAA staining (Figure 1A, insets).

To quantitatively analyze the sialic acid residues, P815 cells were stained with different concentrations of FITC-conjugated lectins and assayed by flow cytometry. As shown in Figure 1B, at an lectin concentration of 5 μg/ml, 95.54% of P815 cells were detected positive expression of α-2,3-linked SA receptors (FITC-MAA) and 98.57% of α-2,6-linked receptors (FITC-SNA). At higher concentrations (10 and 20 μg/ml), almost all cells positively expressed both SA receptors (>99%). The mean fluorescence intensity (MFI) depended on the lectin concentration, and the MFI of SNA was significantly higher (>fivefold) than that of MAA at all the concentrations. These data suggested that the expression of α-2,6- linked SA receptors was more abundant on P815 mastocytoma cells than α-2,3- linked SA receptors.

Taken together, these data suggested that both α-2,3- and α-2,6- linked SA receptors were expressed on the surface of P815 mouse mastocytoma cells.

We previously demonstrated that H5N1 infection could activate mast cells (Hu et al., 2012). To determine if IAV could enter and replicate productively in mastocytoma cells, we examined the replication kinetics of human and avian IAVs in P815 cells. As shown in Supplementary Figure S1, all three subtypes of IAVs productively replicated in P815 cells as measured by hemagglutination assay (left), plaque formation (middle), and viral RNA expression (right). In P815 cells, the replications H1N1 and H5N1 viruses were more efficient than H7N2 virus. These data indicated that IAVs replicated well in P815 cells with some degree of tropism selectivity.

To corroborate this finding, P815 cells were co-stained with α-2,3- or α-2,6- linked SA receptors and viral NP (Figure 2A). The wide distribution of α-2,6- linked SA receptors was consistent the effective replication of H1N1. In addition, a large number of NP-positive cells were observed when infected with H5N1, but were less abundant in H7N2 infected cells, which was consistent with the viral replication profiles (Supplementary Figure S1). To further validate the permissiveness of P815 for IAVs replication, we used transmission electron microscopy. As shown in Figure 2B, budding virions were present on the surface of cells infected with each of the three virus subtypes. Moreover, many viral particles were found to be associated with the surface of the cells infected with H1N1 and H5N1, but were less obvious on H7N2 infected cells. Together, the above data suggested that IAVs could bind and enter into P815 mastocytoma cells, where the efficient replication was supported.

To determine if IAVs can infect mast cells in vivo, we used a mouse H5N1 virus infection model. Mice were either infected with H5N1 or PBS and the nasal mucosa and lung tissues were resected and analyzed by immunofluorescence. Cells double positive for the virus-specific antigen NP and the mast cell specific protein tryptase were present in H5N1 infected mice but not in PBS treated mice (Figure 3). These data suggested that IAVs probably could actively infect mast cells in vivo.

Our previous data suggested that H5N1-activated mast cells could intensify lung injury by releasing the pro-inflammatory mediators including histamine, tryptase, and IFN-γ (Hu et al., 2012). To more specifically examine the cytokines and chemokines released by P815 cells upon avian and human IAVs infection, we performed an antibody array analysis. As shown in Figure 4A, the production of sICAM-1, IL-6, IL-13, IP-10, M-CSF, CCL-2, CCL-12, and TNF-α was augmented in the supernatants of P815 cells infected with all three viruses subtypes. The release of G-CSF, GM-CSF, CCL-1, IFN-γ, IL-1α, IL-4, and CXCL-1 was only moderately increased. To further confirm these findings, we used ELISA to conduct kinetic profiles of the selected cytokines and chemokines that were potentially involved in responses to influenza infection. The three virus subtypes, and LE-PolyI:C, induced significantly higher levels of IL-6 and IFN-γ compared to mock treated cells (Figure 4B). The production of CCL-2 did not occur until 12 h post-infection, but then increased from 24 to 48 h post-infection (Figure 4B), while the secretion of IP-10 and TNF-α increased gradually peaking at 36 h post-infection. In contrast, all three IAV subtypes induced a relatively low expression of CCL-5 for 24 h post-infection, but this increased at later time points. The expressions of antiviral cytokines IFN-α and IFN-β did not change from 2 to 48 h post-infection (Figure 4B).

To further analyze the expression kinetics of various cytokines and chemokines released by P815 cells infected with each of the three virus subtypes and LE-PolyI:C, we performed quantitative RT-PCR. The mRNA expression profiles of IL-6, IFN-γ, TNF-α, CCL-2, CCL-5, and IP-10 were similar to the data generated from ELISA (Supplementary Table S2). However, while the expression levels of a large number of pro-inflammatory cytokines and chemokines were up-regulated, the mRNA levels of the antiviral genes IFN-α, IFN-β and the anti-inflammatory cytokine IL-10 were unchanged during viral infection (Supplementary Table S2). Taken together, these data suggested that following IAV infection P815 cells released a range of pro-inflammatory cytokines and chemokines.

Given that IAVs could infect P815 cells and promote the release of inflammatory cytokines and chemokines, we next examined expression level of viral RNA sensor TLR3 involved in the transduction of inflammatory molecule signals. The mRNA expression level of TLR3 in P815 cells infected with IAVs peaked 4 h post-infections, and returned to baseline levels by 12 h post-infection (Figure 5A). In addition, the mRNA expression profiles of the adaptor molecules TRIF increased in a manner consistent with TLR3 levels; however, the fold increase compared to mock infected cells was much lower. Consistent with mRNA levels, the protein expression of TRL3 and TRIF was strongly augmented in IAV infected P815 cells (Figure 5B). To confirm if TLR3 were indeed activated, co-immunoprecipitation experiments were used to test the endogenous interactions of TLR3 with TRIF. Co-precipitation of TLR3/TRIF was evident in P815 cells infected with all three IAVs (Figure 5C). These data indicated that the viral RNA sensors TLR3 were expressed following IAV infection.

To further investigate the role of TLR3 in P815 cells during IAV infection, we utilized a novel TLR3/dsRNA complex inhibitor to disrupt the interaction between IAV and TLR3 in P815 cells. The effects of the inhibitors on the release of pro-inflammatory cytokines and chemokines, including IL-6, IFN-γ, TNF-α, and CCL-2, were examined. The expression of these pro-inflammatory cytokines and chemokines was significantly decreased in TLR3/dsRNA complex inhibitor treatment group 24 and 36 h after infection (Figure 6). In addition, treatment of IAV-infected P815 cells with TLR3/dsRNA complex inhibitor dramatically decreased viral titers 24 and 36 h after infection (Figure 7). Taken together, these data suggested that following IAV infection P815 cells actively participated in promoting inflammation by releasing a range of pro-inflammatory cytokines and chemokines possibly through TLR3 signaling pathways.