Introduction

Influenza is an acute respiratory viral disease with approximately 3 to 5 million cases of severe illness and up to 650000 deaths annually, worldwide. Therefore, the underlying mechanisms of the pathogenesis of the influenza virus have attracted extensive attention. Previous reports have demonstrated that influenza virus triggers various inflammatory responses, such as the upregulation of proinflammatory cytokines and chemokines, overproduction of intracellular reactive oxygen species (ROS), and enhanced programmed cell death [1,2,3]. Since these cellular responses can lead to severe lung injury, identification of precise inflammation-associated factors and pathways is important for understanding the pathogenesis of influenza and develo** novel therapeutic strategies.

Acute viral infection is commonly associated with cell death. Since lung epithelial and pulmonary immune cells directly infected with influenza virus undergo apoptotic, necrotic and necroptotic cell death, the underlying molecular mechanisms have been widely studied [4,5,6,7]. Particularly, intracellular interactions between viral and host proteins have attracted the most attention because it has been regarded that viral proteins, especially internal proteins, are expressed and reside in the cytosol. However, it was recently reported that nucleoprotein (NP) can be released from infected cells and increase viral pathogenicity through the activation of Toll-like receptor 4 (TLR4), a membrane receptor, on innate immune cells in the lungs [8]. This suggests a potential role for viral internal proteins in the disease pathogenesis through the interaction with host cells in the extracellular compartment.

Influenza virus matrix 1 (M1) protein is important for viral replication and sha** of virus particles [26,27,28,29]. It is also controversial whether the activation of the NLRP3 inflammasome and production of IL-1β and IL-18 increases or decreases lung pathology [30,31,32]. In this work, we demonstrated that TLR4 recognizes extracellular M1 and aggravates viral pathogenicity in vivo. Interestingly, inactivated H5N1 virus induced acute lung injury in a TLR4-dependent manner, partially supporting our findings [12]. Furthermore, we recently reported the pathogenic role of TLR4 through the interaction with influenza virus NP [8]. Collectively, TLR4 plays a pathogenic role in the pathogenesis of influenza by recognizing a wide range of ligands from oxidized phospholipid to viral proteins.

Excess levels of ROS can lead to mitochondrial damage and consequent apoptotic cell death. In the case of influenza viral infection, oxidative stress by ROS provokes inflammation and acute lung injury [12, 33, 34]. It is well known that oxidized phospholipid can trigger ROS production through the activation of TLR4 [12]. In this study, we additionally showed that M1, the most abundant influenza viral protein, can be another trigger for ROS production in vitro and in vivo. However, it is intriguing that cellular ROS is also required for the optimal induction of the antigen-specific CD8+ T cell response by increasing antigen cross-presentation by dendritic cells [35,36,37]. In our study, we did not investigate the effect of M1 on the fate or behavior of dendritic cells. It is worthwhile to study whether and how M1 differentially modulates the host response to the viral infection depending on physiological conditions such as its micro-concentration and the cell types it encounters.

The activation of TLR4 has different sub-signaling mechanisms depending on the type of ligand [38, 39]. Bacteria-derived LPS activates both MyD88-dependent and TRIF-dependent pathways [38], but monophosphoryl lipid A (MPLA), a low-toxic derivative of lipid A moiety of LPS, predominantly activates the TRIF-dependent pathway [40, 41]. Consequently, TLR4 activation by LPS and MPLA differs in the induction of transcriptome change, chemokine production, inflammation, and memory CD8+ T cell differentiation [42, 43]. Thus, identifying novel ligands of TLR4 and their biological consequence is of continuing interest. Our gene expression profile results showed that M1 and LPS differentially activated downstream target genes, and, interestingly, it did not match those prompted by NP (Figs. 1F and 3F, and Fig. S7A–C). Therefore, further study is required on ligand dependent TLR4 signaling mechanisms in terms of molecular and structural biology.

Previous studies have demonstrated that anti-M1 antibody cannot directly neutralize influenza virus [44]. However, our data showed that anti-M1 antibody alleviates disease severity and improves survival rate against the viral challenge. Anti-M1 plasma or anti-M1 antibody reduced the level of inflammatory cytokines and cell death in infected mice. Antibodies to NP have shown an antiviral effect against influenza virus infection in murine models [45,46,47]. Although it has been previously suggested that NK cells contribute to the antiviral effect via antibody-dependent cellular cytotoxicity, the exact mechanism has not been experimentally proved. In this study, we did not investigate the possibility of involvement of NK cells in the anti-M1 antibody-mediated therapeutic effect. Nonetheless, antibodies targeting M1 could be an auxiliary therapeutic approach for influenza.

In conclusion, we showed that M1 is an important pathogenic factor contributing to influenza viral pathogenicity by increasing ROS-mediated cell death in the lungs. Intriguingly, M1 is released from infected cells to the extracellular compartment and interacts with TLR4, a membrane receptor protein, on the lung epithelial and pulmonary immune cells. This finding not only expands our understanding of the molecular mechanism of influenza virus-induced cell death and viral pathogenesis but could also be useful information for the development of a novel antiviral drug.

Materials and methods

Mice

Seven- to eight-week-old C57BL/6 and BALB/c female mice were purchased from KOATECH (Pyeongtaek, Korea) and maintained in a specific pathogen-free biosafety level-2 facility at the Korea Research Institute of Bioscience and Biotechnology (KRIBB). TLR4 knockout mice were purchased from The Jackson Laboratory (Bar Harbor, Maine, USA). Only 8- to 12-week-old female mice were used in this study. All animal experiments were approved by the Institutional Animal Use and Care Committee of KRIBB (KRIBB-AEC-19173) and performed in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health.

Cell culture

HEK-Blue™ TLR4 (InvivoGen, San Diego, CA, USA) and L929 (ATCC, Manassas, Virginia, USA) cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM, Corning, NY, USA) supplemented with 10% fetal bovine serum (FBS, HyClone, Logan, Utah, USA) and 1× antibiotics (Gibco™, Massachusetts, USA). MLE-12 cells (ATCC) were maintained in DMEM supplemented with 2% FBS and 1× antibiotics. No mycoplasma contamination was detected. To prepare BMDMs, bone marrow cells were harvested from the femur and tibia of C57BL/6 N and TLR4 knockout mice and differentiated in DMEM containing 25% L929-conditioned medium, 10% FBS, and 1× antibiotics. L929-conditioned medium was obtained by culturing L929 cells in DMEM for 7 days until a confluence of >90% was obtained. Afterwards, the medium underwent centrifugation at 1500 rpm for 10 min and the supernatants were collected and filtered through a 0.45 μm filter system.

Recombinant M1 protein

The DNA fragment encoding M1 protein (GenBank: CY147535.1) was cloned into the pEXPR-IBA 103 vector (IBA Lifesciences, Göttingen, Germany), and the plasmid DNA was transfected into ExpiCHO cells. Five days later, the cells were harvested and lysed by sonication (30% amplitude, 10 s on/10 s off, total 10 min). After filtration with a 0.45 μm filter (Corning), the lysate supernatant was bound to Strep-Tactin XT resin (IBA Lifesciences) using a gravity flow column. After washing, the bound protein was collected by treating the column with an elution buffer (IBA Lifesciences). The endotoxin level in the purified M1 was measured using the endotoxin test kit (Thermo Fisher Scientific, Waltham, Massachusetts, USA). The protein was stored at –80 °C until further use.

Virus

The influenza A/Puerto Rico/8/1934 (PR8) virus was cultivated in the allantoic cavity of embryonated chicken eggs. Viruses were titrated by calculating the 50% egg infectious dose (EID50) and 50% tissue culture infectious dose (TCID50) and stored at –80 °C until use.

Anti-M1 polyserum and anti-M1 antibody

The M1-coding DNA fragment was cloned into the pGX-10 plasmid [48], which was used as the DNA vaccine vector in this study. Mice (BALB/c) were intramuscularly immunized twice at 3-week intervals with the plasmid DNA (10 μg) using electroporation. Serum was obtained 2 weeks after the final vaccination. Anti-M1 IgG was purified from the serum of mock- and M1-vaccinated mice using an IgG antibody purification kit (Abcam, Cambridge, UK) according to the manufacturer’s protocol. Serum and purified antibodies were stored at –20 °C.

Preparation of lung samples (whole lung cells, BAL fluid, and BAL cells)

Total lung cells and BAL fluid samples were obtained as described previously [8]. Briefly, lung tissue was minced and incubated in 1.5 ml RPMI 1640 media containing collagenase D (150 unit/ml, Gibco), DNase I (50 μg/ml, Merck), 10% FBS, and 1× antibiotics. After enzymatic digestion, cells were isolated from the tissue using a strainer (SPL, Pyeongtaek, Korea). For BAL fluid preparation, a catheter was inserted in the trachea of anesthetized mice and 1 ml of cold PBS containing protease inhibitors (Merck) was instilled into the lungs. After gentle aspiration, BAL fluid and cells were separated by centrifugation (4000 rpm, 5 min, 4 °C).

Experimental schedule

Mice (C57BL/6) were treated intranasally with recombinant M1 protein (rM1), after which, BAL fluid, BAL cells and lung cells were obtained at each indicated time point for enzyme-linked immunosorbent assay (ELISA), western blotting, and flow cytometry analysis. Mice (C57BL/6) infected with PR8 influenza virus (32 PFU) were intranasally treated with PBS or M1 immediately, 3 days pre- or post-infection. Influenza virus-infected mice were intraperitoneally administered with 200 μl of naïve or anti-M1 serum or 15–200 μg of anti-M1 purified IgG antibody daily for 3 days. The change in bodyweight and survival rate of the mice were monitored for 14 days post infection.

Western blot

Cells were lysed in CETi lysis buffer (TransLab, Daejeon, Korea) and the lysate was separated by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE). After transferring the proteins onto polyvinylidene fluoride membranes (Merck), the membranes were blocked and incubated with primary antibodies overnight at 4 °C. After washing, horseradish peroxidase (HRP)-conjugated secondary antibodies were added. Bound antibodies were visualized using Immobilon Forte Western HRP substrate (Merck) and imaged using EZ-Capture II (ATTO Corporation, Tokyo, Japan). The following antibodies were used: anti-phospho-SAPK/JNK (Cat# 9255), anti-SAPK/JNK (Cat# 9252), anti-phospho-p44/42 MAPK (Cat# 4370), anti-p44/42 (Cat# 4695), anti-phospho-p38 MAPK (Cat# 4511), anti-p38 MAPK (Cat# 54470), anti-phospho-IκB (Cat# 2859), anti-IκB-alpha (Cat# 4812), anti-phospho-NF-κB p65 (Cat# 3033), anti-COX1 (Cat# 9896), anti-COX2 (Cat# 12282), anti-iNOS (Cat# 13120), anti-Bim (Cat# 2933), anti-phospho-Bad (Cat# 5284), anti-Bak (Cat# 12105), anti-caspase-3 (Cat# 9665), anti-caspase-6 (Cat# 9762), anti-caspase-7 (Cat# 8438), anti-caspase-9 (Cat# 9508), HRP-linked anti-mouse IgG (Cat# 7076), HRP-linked anti-rabbit IgG (Cat# 7074),(all from CST, Danvers, Massachusetts, USA), and anti-Strep-tag II (Abcam, Cat# ab76949).

IP

For IP, BMDMs were lysed in CETi lysis buffer for 10 min on ice. Lysates were centrifuged and the supernatants were transferred to fresh tubes containing antibody coated Dynabeads as per the manufacturer’s instructions (Invitrogen), followed by western blot analysis. The following antibodies were used: anti-TLR4 (Santa Cruz Biotechnology, Dallas, Texas, USA; and Thermo Fisher Scientific, Cat# sc-293072) and anti-influenza M1 (Sino biological, Bei**g, China, Cat# 40010-RP01).

ELISA, LDH, and ROS assay

BMDM, MLE-12, and BAL cell culture supernatants were collected after treatment with M1 or LPS for 20 h for ELISA and after 0–48 h for LDH and ROS assays. BAL fluid samples were obtained from the mice, and cytokines were measured using IL-6, CXCL1, CXCL-10, CCL2, CCL5 ELISA kits (all from R&D Systems, Minneapolis, Minnesota, USA). LDH and ROS were detected using the CyQUANT™ LDH cytotoxicity assay (Invitrogen) and in vitro ROS/RNS assay kits (Cell Biolabs Inc. San Diego, CA, US), respectively, according to the manufacturers’ protocols.

RNA sequencing and analysis

(1) RNA isolation

BMDM cells were treated with PBS, M1, NP, or LPS for 6 h, after which total RNA was isolated using Trizol reagent (Invitrogen Corp., Carlsbad, CA, USA). RNA purity and integrity were evaluated using a ND-2000 Spectrophotometer (NanoDrop, Wilmington, USA) and Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, USA), respectively.

(2) Affymetrix whole-transcript expression array

The Affymetrix Whole-transcript Expression array was performed using the GeneChip Whole Transcript PLUS reagent Kit (Thermo Fisher Scientific) per the manufacturer’s protocol. cDNA was synthesized using the GeneChip WT (Whole Transcript) Amplification kit as described by the manufacturer. The sense cDNA was then fragmented and biotin-labeled with terminal deoxynucleotidyl transferase (TdT) using the GeneChip WT Terminal labeling kit. Approximately 5.5 μg of labeled DNA target was hybridized to the Affymetrix GeneChip Mouse ST 2.0 Array at 45 °C for 16 h. Hybridized arrays were washed and stained on a GeneChip Fluidics Station 450 and scanned on a GCS3000 Scanner (Affymetrix). Signal values were computed using the Affymetrix® GeneChip™ Command Console software.

(3) Data analysis

Expression data were generated by Transcriptome Analysis console 4.0.1. For the normalization, robust multi-average (RMA) algorithm implemented in Transcriptome Analysis Console software was used. Data mining and graphic visualization were performed using ExDEGA (Ebiogen Inc., Korea). To define differentially expressed genes (DEGs), adjusted |log2fold-change (FC)| ≥ 1.5 and P < 0.05 were selected as cut-off values.

TLR reporter cell line assay

The TLR reporter cell lines (Invivogen) were cultured in DMEM containing 10% FBS and 1× antibiotics. When the HEK-Blue™ TLR4 reporter cells reached 80% confluency, they were harvested using a scraper and resuspended in HEK-Blue™ Detection media (InvivoGen). The resuspended cells (5 × 104 cells/well; 180 μl) were then incubated with 20 μl of each stimulant in a 96-well plate. After 20 h of culture, the optical density at 630 nm (OD630) was measured.

Flow cytometry

Single lung cell suspensions were blocked using anti-mouse CD16/CD32 (mouse BD Fc Block™; BD Biosciences) at room temperature for 15 min before staining. The surface antigens were stained with the indicated conjugated antibodies at 4 °C for 15 min. The following antibodies were used: APC anti-CD4 (Thermo Fisher Scientific, Cat# 17-0042-82), PerCP-eFluor 710 anti-CD3e (Thermo Fisher Scientific, Cat# 46-0033-82), PE/Cyanine7 anti-CD45R (Thermo Fisher Scientific, Cat# 25-0452-82), PE anti-Siglec-F (Thermo Fisher Scientific, Cat# 552126), PerCD-eFluor 710 anti-Ly6G (Thermo Fisher Scientific, Cat# 46-9668-82), Alexa Fluor 700 anti-MHC Class II I-A/I-E (Thermo Fisher Scientific, Cat# 56-5321-80), and eFluor 450 anti-F4/80 (Thermo Fisher Scientific, Cat# 48-4801-82); APC/cyanine7 anti-CD45 (BioLegend, San Diego, CA, USA, Cat# 103116); APC anti-CD11b (BioLegend, Cat# 101212) and PE anti-CD49b (BioLegend, Cat# 103506); and FITC anti-NK1.1 (BD Biosciences, Cat# 553164). Stained cells were analyzed using a Gallios flow cytometer (Beckman Coulter, Brea, CA, US) with FlowJo™ software (Becton, Dickinson and Company, New Jersey, USA).

Statistical analysis

Statistical differences among groups were assessed using a two-tailed Student’s t-test and a log-rank test with Prism software (GraphPad Software, USA). Mice were randomly assigned to experimental groups without blinding method for injections. No animal exclusion criteria were applied, and the variance was comparable among the groups that were statistically compared. Values with p < 0.05 were considered to be statistically significant.