Mint3/Apba3 depletion ameliorates severe murine influenza pneumonia and macrophage cytokine production in response to the influenza virus

Influenza virus (IFV) infection is a common cause of severe pneumonia. Studies have suggested that excessive activation of the host immune system including macrophages is responsible for the severe pathologies mediated by IFV infection. Here, we focused on the X11 protein family member Mint3/Apba3, known to promote ATP production via glycolysis by activating hypoxia inducible factor-1 (HIF-1) in macrophages, and examined its roles in lung pathogenesis and anti-viral defence upon IFV infection. Mint3-deficient mice exhibited improved influenza pneumonia with reduced inflammatory cytokines/chemokine levels and neutrophil infiltration in the IFV-infected lungs without alteration in viral burden, type-I interferon production, or acquired immunity. In macrophages, Mint3 depletion attenuated NF-κB signalling and the resultant cytokine/chemokine production in response to IFV infection by increasing IκBα and activating the cellular energy sensor AMPK, respectively. Thus, Mint3 might represent one of the likely therapeutic targets for the treatment of severe influenza pneumonia without affecting host anti-viral defence through suppressing macrophage cytokine/chemokine production.

Munc18-1-interacting protein 3 (Mint3), also known as amyloid-β A4 precursor protein-binding family A member 3 (Apba3), is a member of the X11 protein family 20,21 . Mint3 is expressed in most cell types and localizes to the trans-Golgi network by interacting with membrane proteins such as amyloid precursor protein and furin 22,23 . Previously, we have revealed that Mint3 activates hypoxia-inducible factor 1 (HIF-1) even in the presence of oxygen by suppressing its inhibitor, factor-inhibiting HIF-1 (FIH-1) 24 . Mint3 requires a transmembrane protease MT1-MMP/MMP14 to bind and suppress FIH-1 25,26 . Thus, Mint3-mediated HIF-1 activation is limited to MT1-MMP expressing cells such as macrophages and cancer cells [25][26][27] . Mint3-deficient mice are viable and show no apparent defect. However, Mint3-deficent macrophages exhibit a defect in ATP production via glycolysis owing to reduced HIF-1 activity 25,27,28 . Mint3-deficient macrophages also showed impaired cytokine production in response to lipopolysaccharide (LPS) 27 . Since FIH-1 also hydroxylates molecules other than HIF-1α including Iκ Bα 29 which inhibits the NF-κ B signalling, FIH-1 suppression by Mint3 might affect inflammatory cytokine/ chemokine production via not only HIF-1 but also the NFκ B signalling in macrophages. Because macrophages are one of the main sources of inflammatory cytokines/chemokines, here we hypothesized that Mint3 likely played roles in influenza pneumonia and in host protection against IFV.
In this study, we demonstrate a role of Mint3 in influenza pathogenesis and immunity. Our results showed that Mint3 depletion in mice attenuated fatal influenza pneumonia through the reduction of inflammatory cytokine/ chemokine production and neutrophil infiltration without affecting type-I interferon production and anti-viral acquired immunity. Further analyses revealed that Mint3 depletion attenuated NF-κ B signalling in response to IFV infection in macrophages. Therefore, Mint3 inhibition might represent one of the likely targets for the treatment of severe influenza pneumonia.

Mint3 depletion attenuates lethal influenza pneumonia in mice.
To determine whether Mint3 contributes to immune responses during lung IFV infection, we first intranasally infected wild-type (WT: C57BL/6) and Mint3-deficient (Mint3 −/− ) mice with a lethal dose of IFV (10 4 plaque forming units (PFU)/mouse). IFVinfected WT mice appeared visibly ill with ruffled fur and reduced oral intake 6 to 10 days after infection, whereas Mint3 −/− mice appeared more active than the WT mice. Consistent with their apparent activity, the final survival rate on day 21 after infection was significantly improved in Mint3 −/− mice (~80%) as compared with that in WT mice (~50%; Fig. 1a). Histopathological analysis of the IFV-infected lungs on days 4 and 8 revealed that lung inflammation, which was the most obvious on day 8 in WT mice, was much less severe in Mint3 −/− mice (Fig. 1b).
Given that inflammatory cytokines/chemokines are linked to lung damage in severe influenza pneumonia [12][13][14][17][18][19]30 , we analysed the infiltrated cells and cytokine/chemokine levels in bronchoalveolar lavage fluid (BALF) collected from the IFV-infected lungs of WT and Mint3 −/− mice. Consistent with the pathological data, the total number of infiltrated cells in BALF from Mint3 −/− mice was decreased compared to that from WT mice on days 4 and 8 after IFV infection ( Fig. 1c and Supplementary Fig. S1a-c). This was attributed to the significant decrease in neutrophils that comprised the major subpopulations of infiltrated cells (Fig. 1c). In contrast, the numbers of T, B, and natural killer cells as well as macrophages were not significantly altered in BALF from the Mint3 −/− lungs (Fig. 1c). The levels of inflammatory cytokines and chemokines, i.e., IL-6, TNF-α , CCL3/MIP-1α , CXCL1/KC, and CXCL10/IP-10, which have been reported to contribute to lung pathology 11,12,16 , also decreased in BALF from Mint3 −/− mice compared to those from WT mice on day 4 after infection (Fig. 1d). In addition, the levels of CXCL1/KC and CXCL10/IP-10 in BALF from Mint3 −/− mice remained lower than those from WT mice on day 8 after infection. These data indicate that host Mint3 contributes to cytokine/chemokine production including the neutrophil chemoattractant CXCL1/KC at an early time point after IFV infection and influences subsequent inflammatory cell recruitment and lung pathology at a later time point. Collectively, Mint3 depletion attenuated the severe influenza pneumonia and improved host mortality in mice.
Mint3 depletion does not compromise anti-viral protective immunity in mice. Activation of innate immunity in response to IFV regulates anti-viral protective immunity 31 . To evaluate the impact of Mint3 deficiency on anti-IFV protection, we first analysed the viral copy number in the IFV-infected lungs of WT and Mint3 −/− mice on day 0, 4, and 8 following the IFV infection. Mint3 depletion did not largely affect the viral copy number in the IFV-infected lungs (Fig. 2a). Type-I interferons (IFN-α /β ) are the primary factors that induce host resistance to IFV. In particular, IFN-α induction in the late stage of IFV infection is essential for protection against IFV 31,32 . In contrast to cytokine/chemokine production, no significant difference was observed in IFN-α production in the infected lungs between WT and Mint3 −/− mice (Fig. 2b). Notably, the level of the type-II interferon IFN-γ was significantly lower in the lungs of Mint3 −/− mice than in those of WT mice on day 8 but not on day 4 after infection ( Fig. 2b) for unknown reasons. Next, we assessed the induction of humoral acquired immunity against IFV in Mint3 −/− mice. WT and Mint3 −/− mice were infected again with IFV on day 21 after the first infection (representing the second infection) and the production of virus-specific antibodies was analysed on day 7 after the second infection. Mint3 depletion in mice affected neither the production of virus-specific IgG in the serum nor virus-specific IgA in the lung mucosa (Fig. 2c). Collectively, these data suggest that Mint3 is dispensable for the induction of protective immunity against IFV infection in mice.
Mint3 depletion impairs inflammatory cytokine production in response to IFV in macrophages but not dendritic cells. Macrophages and dendritic cells are known to comprise an early source of inflammatory cytokines and type-I interferons in pulmonary IFV infection 11,16 . Mint3 depletion affects macrophage function owing to reduced HIF-1 activity and ATP production via glycolysis 25,27,28 . Thus, we focused on the function of Mint3 in macrophages and dendritic cells, and examined whether Mint3 in macrophages and dendritic cells is required for cytokine/chemokine production in response to IFV. Alveolar macrophages (AMFs), thioglycolate-induced peritoneal macrophages (TG-MFs), bone marrow-derived macrophages (BMMFs), bone marrow-derived conventional dendritic cells (cDCs), or Flt3 ligand-induced plasmacytoid dendritic cells (FLT3L-DCs) prepared from WT or Mint3 −/− mice were brought into contact with IFV in vitro and the production of IL-6, TNF-α , and CCL2/MCP-1 was measured. The cytokine/chemokine production in response to IFV  Supplementary Fig. S2c). However, Mint3 depletion did not affect IL-6, TNF-α , or CCL2/ MCP-1 production in cDCs (Fig. 3d) and FLT3-DCs ( Fig. 3e) in response to IFV. Thus, the reduced cytokine/ chemokine production by macrophages in response to IFV, but not by dendritic cells, likely accounts at least in part for the attenuation of influenza pneumonia in Mint3 −/− mice.

Mint3 depletion affects NF-κB activation in macrophages. IFV is recognized by pattern recognition
receptors such as Toll-like receptor 7 and retinoic acid inducible gene 1, which then activate the NF-κ B signalling pathway in cells 31,33 . Although Mint3 −/− macrophages exhibit reduced HIF-1 activity 25,27,28 , it remains unclear whether Mint3 also influences the activation of NF-κ B, a transcription factor that induces the expression of inflammatory cytokines/chemokines 34 . Thus, the expression levels of NF-κ B target genes such as Il6, Tnf, and Ccl2 in IFV-stimulated WT and Mint3 −/− BMMFs were analysed by a quantitative polymerase chain reaction (qPCR) method. The expression levels of Il6, Tnf, and Ccl2 mRNA decreased in Mint3 −/− BMMFs following IFV stimulation (Fig. 4a). Next, nuclear translocation of NF-κ B in IFV-stimulated BMMFs was examined by immunoblotting. Notably, NF-κ B protein levels in the nuclear extracts of Mint3 −/− BMMFs were lower than those of WT cells at the starting point (0 h) and remained lower during the observed time after IFV infection (Fig. 4b).
Translocation of cytoplasmic NF-κ B to the nucleus is usually suppressed by inhibitor of κ B (Iκ B); however, Iκ B kinase (IKK) phosphorylates Iκ B and thereby promotes proteasomal degradation of Iκ B when the NF-κ B signalling pathway is activated 34 . Thus, we analysed Iκ B and IKK in the cytoplasmic extracts from IFV-infected WT and Mint3 −/− BMMFs. The expression and phosphorylation levels of IKKα and IKKβ showed no apparent differences between WT and Mint3 −/− BMMFs (Fig. 4c). However, the expression levels of Iκ Bα markedly increased in Mint3 −/− BMMFs at the starting point (0 h) and remained higher than those of WT cells during the observed time after IFV stimulation (Fig. 4c). In turn, a lower proportion of Iκ Bα was phosphorylated in Mint3 −/− BMMFs after IFV stimulation (Fig. 4c). Taken together, these findings indicate that Mint3 depletion resulted in the accumulation of Iκ Bα in the cytoplasm and suppressed the activation of NF-κ B in response to IFV stimulation in macrophages.   . 5a). When cells are starved, adenosine monophosphate-activated protein (AMP)-activated kinase (AMPK) is activated by phosphorylation 35 . Accordingly, the phosphorylation levels of AMPKα increased in WT BMMFs after IFV infection (Fig. 5b). Mint3 −/− BMMFs showed increased phosphorylation levels of AMPKα at the starting point reflecting their decreased ATP levels (Fig. 5a) and sustained their AMPKα phosphorylation after IFV infection (Fig. 5b). Notably, the amount of AMPKα protein increased after IFV stimulation in Mint3 −/− BMMFs (Fig. 5b).

Mint3 deficiency affects NF-κB activation in macrophages via both AMPK activation and
AMPK is known to suppress NF-κ B signalling via several pathways such as SIRT1, PGCα 1, and p53/FoxO 36 . We hypothesized that the defect of ATP production via glycolysis and the resulting AMPK activation by Mint3 depletion also affects the IFV-induced NF-κ B signalling in macrophages. To address this, WT and Mint3 −/− BMMFs were infected with IFV in the presence of a glycolysis inhibitor, 2-deoxyglucose (2-DG), a mitochondrial oxidative phosphorylation inhibitor, oligomycin, or vehicle, and the inflammatory cytokine/chemokine production was analysed. 2-DG but not oligomycin treatment decreased the inflammatory cytokine/chemokine production in WT BMMFs to the levels of Mint3 −/− cells. The AMPK activator, 5-aminoimidazole-4 -carbox-amide-1-β -D-ribofuranoside (AICAR), also decreased the inflammatory cytokine/chemokine production in WT BMMFs to the levels of Mint3 −/− cells (Fig. 5c). Next, we examined whether AMPK activation via the glycolysis defect relates to the Iκ B accumulation observed in Mint3 −/− BMMFs (Fig. 4c). To address this issue, Iκ Bα and AMPKα in IFV-infected WT BMMFs treated with or without 2-DG were analysed. 2-DG treatment did not impact either total or phosphorylated Iκ Bα levels whereas both total and phosphorylated AMPKα levels increased in IFV-infected WT BMMFs (Fig. 5d) and nuclear NF-κ B levels decreased (Fig. 5e). These results indicate that AMPK activation via the glycolysis defect attenuates NF-κ B signalling independently from Iκ B accumulation in IFV-infected macrophages. In turn, we next examined whether Iκ B accumulation alone could attenuate NF-κ B signalling in IFV-infected macrophages. To address this possibility, IFV-infected WT BMMFs were treated with or without an Iκ Bα phosphorylation inhibitor, BAY 11-7082 37 . BAY 11-7082 treatment decreased phosphorylated Iκ Bα levels and increased total Iκ Bα levels without affecting total and phosphorylated AMPKα levels in IFV-infected WT BMMFs (Fig. 5f). Under this condition, BAY 11-7082 treatment decreased nuclear NF-κ B levels in WT BMMFs 3-6 h after IFV infection (Fig. 5g). BAY 11-7082 treatment also decreased the inflammatory cytokine/chemokine production in WT BMMFs to the levels of Mint3 −/− cells (Fig. 5h). Thus, Mint3 depletion independently caused both AMPK activation and Iκ B accumulation, each of which could separately attenuate NF-κ B signalling independently in macrophages.

Discussion
In this study, we have shown that the Mint3-mediated pathway contributes to influenza pneumonia in mice. Mint3 depletion in mice reduced inflammatory cytokine/chemokine production and the infiltration of inflammatory cells in IFV-infected lungs, and improved mortality rates. In macrophages, Mint3 depletion also attenuated cytokine/chemokine production in response to IFV infection by suppressing the NF-κ B signalling pathway. This suppression in Mint3 −/− macrophages appeared to be mediated by two independent mechanisms: Iκ Bα accumulation and AMPK activation (Fig. 6). Although how AMPK activation suppressed the NF-κ B signalling pathway in Mint3 −/− macrophages was unclear, several reported pathways such as SIRT1, PGCα 1, and p53/FoxO might contribute to this suppression of the NF-κ B signalling pathway by AMPK 36 .
Mint3 was considered dispensable for anti-IFV immunity because Mint3 depletion did not alter viral burden, the elevation of IFN-α , or the induction of anti-viral adaptive B cell responses in the IFV-infected mice. Thus, other innate mechanisms independent of the Mint3 pathway may play a more dominant role in protective immunity against IFV. Type II interferon IFN-γ production was reduced in the lungs of Mint3 −/− mice compared to that in the lungs of WT mice on day 8 but not on day 4 after IFV infection (Fig. 2b). As IFN-γ expression is regulated by HIF-1 38 , Mint3 depletion might affect IFN-γ expression after IFV infection in mice. IFN-γ is not essential for viral clearance 39 ; however, its presence during IFV infection ameliorates the severity of inflammation and lung injury 40 . Although how much the reduction of IFN-γ levels by Mint3 depletion affects the IFV-mediated lung injury remains unclear, supplementation with IFN-γ might support therapies of IFV-infected lung diseases based on Mint3 inhibition.
Mint3 promotes ATP production via glycolysis by activating HIF-1 in macrophages 25,27,28 . AMPK is activated when the ATP concentration decreases in cells 35 . Thus, Mint3 depletion activated AMPK and thereby suppressed the NF-κ B signalling in macrophages. Notably, AMPKα protein levels increased after IFV infection in Mint3 −/− and 2-DG treated WT macrophages, whereas the basal protein levels of AMPKα were comparable between WT and Mint3 −/− macrophages (Fig. 5b,d). Infectious stress might boost the AMPK signalling by increasing the amount of AMPKα protein in macrophages in combination with energy stress. Besides AMPK, HIF-1 also influences cytokine/chemokine transcription directly 41,42 . Thus, Mint3-mediated HIF-1 activation might also contribute to the cytokine/chemokine production independently from AMPK in macrophages. In addition to AMPK, Iκ Bα was also accumulated in Mint3 −/− macrophages. Although activated/phosphorylated IKKα /β levels were comparable, less Iκ Bα was phosphorylated in Mint3 −/− macrophages than in WT cells after IFV infection. Mint3 activates HIF-1 by suppressing its inhibitor, FIH-1. FIH-1 is an asparaginyl hydroxylase and hydroxylates the α subunit of HIF-1 and other molecules including Iκ Bα 29 . Although the hydroxylation of Iκ Bα by FIH-1 has not been reported to affect the stability of Iκ Bα in vitro 43 , FIH-1 might influence the phosphorylation state of Iκ Bα in macrophages.
Mint3-mediated HIF-1 activation depends on an invasion-promoting membrane protease MT1-MMP/ MMP14 in macrophages and cancer cells 25,26 . We have recently reported that the monooxygenase NECAB3 and mTOR signalling also promote Mint3-mediated HIF-1 activation in cancer cells 44,45 . Although whether NECAB3 relates to inflammation is unclear, mTOR contributes to inflammatory responses 46,47 . Thus, Mint3 may function as a hub among these molecules and HIF-1, AMPK, and NF-κ B signalling pathways to orchestrate inflammation, metabolism, and invasion in macrophages.
In conclusion, Mint3 in macrophages acts to contribute to severe influenza pneumonia, but is dispensable for host protection against IFV. Thus, Mint3 inhibition might represent one of the likely therapeutic targets to control severe IFV infection without affecting viral clearance.

Materials and Methods
Mice. Mint3 −/− mice have been previously described 27 (Riken Center for Developmental Biology (CDB) Accession No. CDB0589K; http://www2.clst.riken.jp/arg/mutant%20mice%20list.html). These mice were backcrossed at least 10 times onto C57BL/6 mice. C57BL/6 mice were purchased from CLEA Japan, Inc. (Tokyo, Japan). The animals were housed in specific pathogen-free conditions. All experiments were approved by the Institutional Animal Care and Use Committee for Kitasato University Medical Center and animals were treated in accordance with the Regulations for Animal Experiments in Kitasato University.  Histology. Mice were euthanized by intraperitoneal administration of sodium pentobarbital at 0, 4, and 8 days after IFV infection and whole lungs were collected. Paraffin embedding and hematoxylin and eosin staining of tissues were performed using standard methodologies.

Bronchoalveolar lavage (BAL)
. BAL was carried out as described previously 15,30,48 . In brief, the tracheas of mice were cannulated with 1.2-mm diameter polyethylene catheters. The lungs were instilled with 1 mL pre-warmed phosphate buffered saline (PBS) containing 5 mM ethylenediaminetetraacetic acid, followed by the retrieval of lavage fluid aliquots. Cells in the BAL fluid (BALF) were counted after red blood cell lysis and subjected to flow cytometric analysis. The supernatants of the BALF were subjected to multi cytokine/chemokine expression analysis and enzyme-linked immunosorbent assay (ELISA). Antigen-specific B cell responses. B cell-mediated humoral responses were measured as virion-specific immunoglobulin production by ELISA, as previously described 51 . Briefly, 96-well ELISA plates (Corning, Corning, NY) were coated with ultrasonicated influenza virion (A/California/07/2009 strain) at 5 × 10 6 PFU/mL in a carbonate buffer (pH 9.6) and incubated overnight at 4 °C. Plates were then washed with PBS containing 0.05% Tween 20 (Wako). Serum and BALF collected from mice at day 7 after the secondary infection were serially diluted with PBS/Tween 20 containing 5% skim milk, applied onto the virion-coated plates, and incubated for 2 h at room temperature. After washing, goat anti-mouse total IgG or IgA conjugated to horseradish peroxidase (Jackson Immunoresearch, Baltimore Pike, PA) was applied and incubated for 2 h at room temperature. After washing, the plates were stained with a TMB Substrate Set (Biolegend). The reaction was terminated with 1 M H 2 SO 4 (Wako) and the absorbance was measured.
Scientific RepoRts | 6:37815 | DOI: 10.1038/srep37815 Nuclear and cytoplasmic extraction of BMMFs. BMMFs were prepared as described above and stimulated with IFV (M.O.I. = 10). After 1, 2, 3, 4, 5, and 6 h of incubation, the BMMFs were harvested with a cell scraper. Nuclear and cytoplasmic extraction of BMMFs was performed using the Nuclear Extract Kit (Active Motif, Carlsbad, CA) according to manufacturer instruction.
Measurement of ATP concentrations. BMMFs were prepared as described above and stimulated with IFV (M.O.I. = 10). After 2, 4, and 6 h of incubation, the BMMFs were harvested with a cell scraper. ATP levels were determined using the ATP bioluminescence assay kit CLS II (Roche Applied Science, Upper Bavaria, Germany) as previously described 26,27 . ATP levels were normalized to the total protein concentration as determined using a Bradford assay kit (Bio-Rad Laboratories, Hercules, CA).
Statistical analysis. Survival curves were generated by the Kaplan-Meier method and statistical analyses were performed using the log-rank test. The statistical significance was assessed by Student's t-tests. A P value < 0.05 was considered significant.