Streptococcus pneumoniae is major cause of otitis media (OM) and life-threatening pneumonia. Overproduction of mucin, the major component of mucus, plays a critical role in the pathogenesis of both OM and pneumonia. However, the molecular mechanisms underlying the tight regulation of mucin upregulation in the mucosal epithelium by S. pneumoniae infection remain largely unknown. In this study, we show that S. pneumoniae pneumolysin (PLY) activates AMP-activated protein kinase α1 (AMPKα1), the master regulator of energy homeostasis, which is required for S. pneumoniae–induced mucin MUC5AC upregulation in vitro and in vivo. Moreover, we found that PLY activates AMPKα1 via cholesterol-dependent membrane binding of PLY and subsequent activation of the Ca2+– Ca2+/calmodulin-dependent kinase kinase β (CaMKKβ) and Cdc42–mixed-lineage protein kinase 3 (MLK3) signaling axis in a TLR2/4-independent manner. AMPKα1 positively regulates PLY-induced MUC5AC expression via negative cross-talk with TLR2/4-dependent activation of MAPK JNK, the negative regulator of MUC5AC expression. Moreover, pharmacological inhibition of AMPKα1 suppressed MUC5AC induction in the S. pneumoniae–induced OM mouse model, thereby demonstrating its therapeutic potential in suppressing mucus overproduction in OM. Taken together, our data unveil a novel mechanism by which negative cross-talk between TLR2/4-independent activation of AMPKα1 and TLR2/4-dependent activation of JNK tightly regulates the S. pneumoniae PLY-induced host mucosal innate immune response.
This article is featured in Top Reads, p.
Streptococcus pneumoniae is one of the important human bacterial pathogens that colonizes the upper respiratory tract and causes the most common childhood middle ear infection, otitis media (OM), and life-threatening pneumonia (1, 2). The antibiotic-resistant strains of S. pneumoniae have increased during the past decade due to inappropriate antibiotics use in the clinic (3). Therefore, development of new nonantibiotic strategies targeting host immune responses is urgently needed for the treatment of S. pneumoniae infections based on full understanding of molecular pathogenesis of S. pneumoniae infections.
The mucosal epithelial cells act as the first line of host mucosal defense against pathogens. Among various components produced by mucosal epithelial cells, mucin, a major protein component in mucus, plays a critical role in mucosal innate defense by providing a physical barrier and trapping microbial pathogens and other inhaled particles for mucociliary clearance (4–8). However, when uncontrolled, the excessive mucin production often leads to impaired mucociliary clearance of mucosal epithelia, resulting in immunopathology in OM and airway obstruction in pulmonary diseases (9, 10). Thus, mucin production must be tightly regulated to maintain an appropriate balance between beneficial and detrimental outcomes in the context of the infectious diseases. Our previous studies showed that S. pneumoniae potently induces mucin MUC5AC, which represents the most important mucin gene in the pathogenesis of OM and pneumonia (11, 12). Upregulation of MUC5AC by S. pneumoniae has been shown in vitro and also in the middle ear of a well-established OM mouse model as well as in human OM patients (7, 13). However, the mechanisms underlying the tight regulation of S. pneumoniae–induced MUC5AC remain largely unknown.
Among numerous virulence factors of S. pneumoniae, pneumolysin (PLY) represents a key virulence factor produced by virtually all clinical S. pneumoniae isolates and released during infections in humans through bacterial spontaneous autolysis (14–16). PLY is defined as a 53-kDa pore-forming toxin belonging to the family of the cholesterol-dependent cytolysin (CDC) (17). Previous reports showed that PLY induces host defense responses in a TLR-dependent and -independent manner (18–21). In the latter aspect, PLY shows lytic activity, which is dependent on the binding activity with cell membrane cholesterol predominantly localized in lipid rafts via its cholesterol binding loop at domain 4, in the target cell membrane (22). S. pneumoniae and PLY have previously been shown to induce mucosal immune response via TLR-dependent activation of MAPK (12, 23, 24). However, how S. pneumoniae and PLY induce mucin in a TLR-independent manner remains largely unknown.
AMP-activated protein kinase (AMPK) is a highly conserved serine/threonine kinase that has emerged as a master regulator of cellular energy homeostasis in mammalian cells (25). AMPK protein exists as a heterotrimer consisting of a catalytic α (α1, α2) subunit and two noncatalytic regulatory subunits, β (β1, β2) and γ (γ1, γ2, γ3) (26). Activation of AMPK is triggered through phosphorylation at threonine 183/172 on the α1/α2 subunits. These phosphorylations are catalyzed by the upstream kinases such as liver kinase B1 (LKB1), Ca2+/calmodulin-dependent kinase kinase β (CaMKKβ), TGF-β–activated kinase 1 (TAK1), and mixed-lineage protein kinase 3 (MLK3) (27–29). Although AMPK has been well known as a master regulator of energy metabolism, recent studies have indicated that AMPK participates in the regulation of many other cellular processes, including cell growth, apoptosis, autophagy, inflammation, and immune responses in multiple cell types. For example, dysfunction of AMPK activity in skeletal muscle, adipocytes, and hepatocytes is involved with metabolic abnormalities such as hyperglycemia, hyperlipidemia, and insulin resistance, as well as chronic inflammation in metabolic syndrome, including obesity and type 2 diabetes (30). In addition, AMPK may coordinate metabolic shifts that support cell growth and proliferation in tumorigenesis (31). Furthermore, AMPK also has potent immunoregulatory effects through modulating the NF-κB signaling pathway in TLR4 or proinflammatory cytokine-stimulated monocytes and macrophages (32, 33). However, despite its known roles in regulating the pathogenic process of certain diseases, the roles of AMPK in infectious diseases, in particular in the host mucosal innate immune responses (e.g., mucus production), in OM pathogenesis caused by S. pneumoniae infection remain largely unknown.
In the current study, we show that AMPKα1 acts as a positive regulator for S. pneumoniae–induced MUC5AC upregulation in human middle ear epithelia in vitro and in the middle ear epithelial mucosa of mice in vivo. The cholesterol binding activity of PLY is required for S. pneumoniae–induced AMPKα1 activation in a CaMKKβ and Cdc42-MLK3 pathway-dependent but TLR2/4-independent manner. Activation of AMPKα1 leads to upregulation of MUC5AC expression via inhibiting JNK, the negative regulator for S. pneumoniae–induced MUC5AC expression. Pharmacological inhibition of AMPKα1 suppressed MUC5AC expression in the well-established mouse model of OM induced by S. pneumoniae infection. Taken together, our study provides insight into the mechanisms underlying the tight regulation of the S. pneumoniae–induced host mucosal innate response via upregulating MUC5AC expression by PLY-mediated and TLR2/4-independent AMPKα1 activation, and may help the development of new therapeutic strategies for treating OM.
Materials and Methods
Reagents and Abs
Cholesterol and methyl-β-cyclodextrin were purchased from Sigma-Aldrich. Compound C, STO-609 acetate, and CEP-1347 were purchased from Tocris Bioscience. Cytochalasin D, jasplakinolide, and latrunculin B were purchased from Cayman Chemical. PD98059 was purchased from Enzo Life Sciences. SP600125 and U0126 were purchased from Calbiochem. Abs for p-AMPKα (no. 2535), MyD88 (no. 4283), p-ERK1/2 (no. 9101), total ERK1/2 (no. 9102), p-JNK1/2 (no. 9251), total JNK1/2 (no. 9252), anti-rabbit HRP-linked Ab (no. 7074), and anti-mouse HRP-linked Ab (no. 7076) were purchased from Cell Signaling Technology. Abs for AMPKα1 (no. sc-19128), MUC5AC (no. sc-21701), TAK1 (no. sc-7967), MLK3 (no. sc-166592), c-myc (no. sc-40), β-actin (no. sc-8432), and donkey anti-goat HRP-linked Ab (no. sc-2020) were purchased from Santa Cruz Biotechnology. Donkey anti-rabbit IgG Alexa Fluor 488 (no. A-21206) and goat anti-mouse IgG Alexa Fluor 546 (A-11030) were purchased from Thermo Fisher Scientific. The human validated small interfering RNA (siRNA) oligonucleotides for AMPKα1 (no. M-005027-02-0005), TAK1 (no. M-003790-06-0005), and control siRNA (no. D-001206-14-05) were purchased from Dharmacon.
Bacteria strains, culture conditions, and purification of PLY
Clinical isolates of S. pneumoniae strains D39, 6B, 19F, 23F, and D39 isogenic PLY-deficient mutant were used in this study as previously described (23, 34). S. pneumoniae was grown on chocolate agar plates at 37°C in a humidified 5% CO2 overnight and inoculated in Todd-Hewitt broth supplemented with 0.5% yeast extract. After overnight, S. pneumoniae growth was monitored by measurement of the OD600 value and then washed with PBS and resuspended in PBS. For in vitro experiments, S. pneumoniae was treated at a multiplicity of infection of 5 or otherwise as indicated. The purification of PLY was performed as described previously (23).
All media described below were supplemented with 10% FBS (Sigma-Aldrich). Cell cultures included the following: human middle ear epithelial cell (HMEEC)-1 cells (provided by Dr. David J. Lim ) maintained in DMEM (Corning Cellgro) and supplemented with bronchial epithelial cell growth medium SingleQuots (Lonza), HEK293-TLR4/MD2 stably-expressing cells (provided by Dr. Douglas T. Golenbock ) in DMEM supplemented with G418 (InvivoGen) (19), HeLa cells (American Type Culture Collection, no. CCL-2) in DMEM, BEAS-2B cells (American Type Culture Collection, no. CRL-9609) in RPMI 1640 medium (Life Technologies), and A549 cells (American Type Culture Collection, no. CCL-185) in F-12K medium (Life Technologies). Mouse embryonic fibroblasts (MEFs) were isolated from wild-type (WT) or TLR4−/− mice (provided by Dr. Shizuo Akira) and maintained in DMEM. All cells were maintained in a humidified atmosphere of 5% CO2 at 37°C.
RNA isolation and real-time quantitative RT-PCR
Quantitative PCR (qPCR) analysis was performed as described previously (37) and following the manufacturer’s instructions. Total RNA was isolated with TRIzol reagent (Life Technologies). Reverse transcription was performed using MultiScribe reverse transcriptase reagents (Applied Biosystems). For quantitative RT-PCR analysis, PCR amplifications were performed with Fast SYBR Green master mix (Applied Biosystems). Briefly, the reactions were performed in triplicate containing 2× Universal Master Mix, 1 μl of template cDNA, and 400 nM primers in final volume of 12.5 μl, and they were analyzed in 96-well optical reaction plates (Applied Biosystems). Reactions were amplified and quantified by using a StepOnePlus real-time PCR system and StepOne software (v2.3) (Applied Biosystems). Relative quantities of mRNAs were obtained by using the comparative Ct method and were normalized using human cyclophilin for in vitro or mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) for in vivo as an endogenous control. The primers used for RT-PCR are described in Supplemental Table I.
Plasmids, luciferase assay, siRNA-mediated knockdown, and transfections
The expression plasmids of AMPKα1, the truncated form of AMPKα11–312-constitutive active (CA), TLR2ΔTIR-dominant negative (DN), TLR4ΔTIR-DN, Rac1, RhoA, and Cdc42 were cloned from cDNA and the inserts were transferred into the pcDNA3.1/myc-His(−) vector (Invitrogen) with proper restriction enzyme sites after we amplified the insert with primers shown in Supplemental Table I. The expression plasmid of MLK3 was purchased from Sino Biological. The mutant expression plasmids of AMPKα1T183D-CA, AMPKα1D159A-DN, Rac1T17N-DN, RhoAT19N-DN, Cdc42T17N-DN, Cdc42Y40C, and MLK3K144R were generated by a QuikChange II XL site-directed mutagenesis kit (Agilent Technologies) using primers shown in Supplemental Table I. Luciferase assays were performed using the Dual-Luciferase reporter assay system (Promega) following the manufacturer’s instructions. The firefly luciferase reporter of pGL3-MUC5AC-300TK was used as previously described (12). The Renilla luciferase activities expressed by pGL4.74[hRluc/TK] (Promega) were used as an internal control. The sequences of each plasmid were verified form at least three clones. Empty vector was used as a control and was also added where necessary to ensure an equivalent amount of input DNA. Transient transfections were carried out using TransIT-2020 (Mirus) or TransIT-LT1 (Mirus) for plasmid DNA, and DharmaFECT4 (Dharmacon) for siRNA following the manufacturers’ instructions.
Western blot and dot blot analysis
Western blot analysis was performed as described previously (38) and following the manufacturer’s instructions. Whole-cell extracts were recovered with protein lysis buffer (20 mM Tris-HCl [pH 7.4], 50 mM NaCl, 50 mM Na4P2O7, 30 mM NaF, 5 μM ZnCl2, 2 mM iodoacetic acid, 1% Triton-X) with freshly added 1 mM sodium orthovanadate and protease inhibitor mixture (Sigma-Aldrich). For Western blot analysis, whole-cell lysate was separated in 8–12% SDS-PAGE gel and transferred to a polyvinylidene difluoride membrane. For dot-blot analysis, cell culture supernatants were loaded into a Bio-Dot SF microfiltration apparatus (Bio-Rad) and transferred to a polyvinylidene difluoride membrane. The membrane was blocked with 5% non-fat dry milk in TBS containing 0.1% Tween 20 (TBST). The membrane was then incubated in a 1:1000 to 1:4000 dilution of a primary Ab in 5% BSA-TBST at room temperature for 1 h, followed by overnight incubation at 4°C. After washing three times with TBST, the membrane was incubated in a 1:5000 dilution of corresponding secondary HRP-conjugated IgG Ab in 2.5% non-fat dry milk–TBST at room temperature for 1 h. Respective proteins were visualized by using Amersham ECL Prime reagent (GE Healthcare Life Sciences) and images were obtained by ChemiDoc XRS+ system (Bio-Rad). ImageJ software (National Institutes of Health) was used for densitometric analysis to determine the fold differences in protein expressions. The intensity of the MUC5AC protein signal obtained by dot-blot analysis was measured by densitometry using ImageJ to represent the protein secretion levels and normalized with protein concentration measured by the bicinchoninic acid (BCA) method. Contrast and brightness adjustments were made equally to all images. These adjustments do not obscure or eliminate any information.
The cytotoxicity of S. pneumoniae and PLY was quantified by a lactate dehydrogenase (LDH) release assay using the CytoTox 96 nonradioactive cytotoxicity assay kit (Promega) following the manufacturer’s instruction.
Targeting strategies for AMPKα1 knockout by the CRISPR-Cas9 system
A 20-bp guide sequence (5′-ATTCGGAGCCTTGATGTGGT-3′) targeting the sense strand of genomic DNA within exon 2 of AMPKα1 was determined by the CRISPR design tool (http://crispr.mit.edu) (39) (Supplemental Fig. 1A). Two complementary oligonucleotides (Supplemental Table I) containing the AMPKα1 guide sequence were purchased from Integrated DNA Technologies. These oligonucleotides were annealed in annealing buffer containing 40 mM Tris-HCl (pH 8.0), 20 mM MgCl2, 50 mM NaCl, and cloned into the BbsI site of the PX458 expression vector (Addgene, no. 48138), which bicistronically expresses single guide RNA (sgRNA) and SpCas9-2A-GFP. HMEEC-1 cells were transfected with PX458-AMPKα1-sgRNA by TransIT-2020. At 48 h posttransfection, cells were suspended in PBS containing 3% FBS, 1 mM EDTA (pH 8.0), and 25 mM HEPES (pH 7.0) and sorted using FACS with an SH800 cell sorter (Sony). The 1 × 103 cells from the population of PX458-derived GFP-expressing cells were subcultured using a limited dilution method in 10-cm dishes. The growing colonies were picked up and then expanded for verification of gene disruption of AMPKα1. The genomic DNA was subjected to PCR amplification with primers shown in Supplemental Table I and cloned into the pUC19 vector. Then, cloned amplicons were analyzed by Sanger sequences to ensure PX458-AMPKα1-sgRNA–derived mutation into the targeting site (Supplemental Fig. 1B). Whole-cell lysates were subjected to Western blotting to confirm the protein expression of AMPKα1.
Cholesterol inactivation experiments
Cholesterol inactivation was performed as previously described with slight modification (40, 41). For masking PLY with cholesterol, ethanol-dissolved cholesterol was mixed in PBS, because the final ethanol concentration in PBS did not exceed 0.5%. The PLY stock was mixed with this cholesterol/PBS suspension and incubated for 30 min at 37°C. After the incubation, cholesterol/PBS suspensions were added to the cells. For depletion of cellular membrane cholesterol, methyl-β-cyclodextrin (MβCD) freshly dissolved in culture medium was added to the cells for 15 min immediately before PLY stimulation. For making inclusion of cholesterol-saturated MβCD, 200 mg of cholesterol was dissolved in 10 mM MβCD at 37°C and then rotated at 700 rpm overnight. The solution was incubated until the stationary phase and centrifuged at 2000 rpm for 5 min. The supernatant was filtered by a 0.45-μm filter and then added to the cells.
The cells were washed with PBS and fixed with 3.7% paraformaldehyde diluted in PBS for 10 min. Then, the cells were permeabilized with 0.1% Triton X-100 for 5 min and incubated with 100 nM Acti-stain 488 fluorescent phalloidin (Cytoskeleton, no. PHDG1) for 30 min in the dark at room temperature. Nuclei were counterstained with mounting medium containing DAPI. Images of stained cells were recorded with a BZ-X710 fluorescence microscope (Keyence).
Mice and animal experiments
C57BL/6 mice (10–12 wk old) were purchased from the National Cancer Institute (National Institutes of Health). The mice were anesthetized and transtympanically inoculated with 1 × 107 CFU of S. pneumoniae, and saline was inoculated as control. The inoculated mice were then sacrificed at 6 or 9 h postinoculation. Dissected mouse middle ears were subjected to total RNA extraction and histologic analysis. For inhibitor experiments, mice were pretreated with compound C i.p. for 2 h before S. pneumoniae inoculation. This solvent saline was used as a vehicle control.
Histology and immunofluorescence staining
For histological analysis, formalin-fixed, paraffin-embedded mouse middle ear tissues were serial-sectioned at 4-μm thickness and mounted into slides. Sections were then stained with H&E to visualize inflammatory responses and pathological changes in the middle ear mucosa epithelia. For immunofluorescence (IF) staining, the detection of p-AMPKα and MUC5AC protein was performed using rabbit anti–p-AMPKα or mouse anti-MUC5AC Ab and followed by donkey anti-rabbit IgG Alexa Fluor 488 or goat anti-mouse IgG Alexa Fluor 546 Ab, respectively, in the paraffin section of mouse middle ear. Images of stained tissue sections were recorded with the Axio microscope system (Axiovert 40 CFL, AxioCam MRC, and AxioVision LE image system, Carl Zeiss) or BZ-X710 fluorescence microscope (Keyence).
All animal experiments were carried out in accordance with guidelines of, and were approved by, the Institutional Animal Care and Use Committee at Georgia State University, under approved protocol no. A19057. The study was carried out in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.
Data are shown as the mean ± SD. Statistical analysis was assessed by a two-tailed unpaired Student t test for data with two conditions (k = 2) and by ANOVA (with Tukey’s post hoc test) for data with more than two conditions (k > 2) using GraphPad Prism (GraphPad Software). A p value <0.05 was considered statistically significant.
S. pneumoniae activates AMPKα1 in human middle ear epithelial cells in vitro and in the middle ear of mice in vivo
We first sought to determine whether S. pneumoniae activates AMPK in middle ear epithelial cells in OM pathogenesis. We focused on investigating the role of AMPKα1, which contains a phosphorylation site at threonine 183 (T183) that plays a critical role in mediating activation of the AMPK heterotrimeric complex (25) and is an isoform expressed more highly in middle ear epithelial cells than another isoform, that is, AMPKα2 (Supplemental Fig. 2A, 2B). Interestingly, S. pneumoniae markedly induced AMPKα1 phosphorylation at T183 in a time- and dose-dependent manner in human middle ear epithelial HMEEC-1 cells in vitro (Fig. 1A, 1B). Moreover, phosphorylation of AMPKα1 was also observed in cells stimulated with several OM-causing S. pneumoniae strains (Fig. 1C), thereby suggesting that the activation of AMPKα1 by S. pneumoniae is generalizable. Consistent with these in vitro findings, phosphorylated AMPKα was also observed in middle ear mucosa epithelial cells of mice infected with S. pneumoniae as assessed by IF staining using phosphorylated AMPKα (T183)-specific Ab (Fig. 1D). Taken together, these data provide direct evidence, to our knowledge, for the first time, for induction of AMPKα1 activation by S. pneumoniae in middle ear epithelial cells in vitro and in vivo.
PLY plays a critical role in mediating activation of AMPKα1 by S. pneumoniae
We next determined whether PLY, a key virulence factor in the pathogenesis of S. pneumoniae infection (16), plays an important role in S. pneumoniae–induced AMPKα1 activation. As shown in (Fig. 2A, the S. pneumoniae WT D39 strain, but not its isogenic PLY-deficient mutant, potently induced AMPKα1 activation. Moreover, purified recombinant PLY activated AMPKα1 in a time- and dose-dependent manner in vitro (Fig. 2B, 2C). To determine whether S. pneumoniae– and PLY-induced AMPKα1 activation is influenced by cell lysis, we examined the cytotoxicity of S. pneumoniae and PLY using an LDH release assay in HMEEC-1 cells. As shown in (Fig. 2D, no significant cytotoxicity was observed in HMEEC-1 cells stimulated with S. pneumoniae and PLY for 5–15 min. These data suggest that S. pneumoniae–and PLY-induced AMPKα1 activation is not due to cell lysis. In addition to in vitro findings, PLY also activated AMPKα in mouse middle ear epithelial mucosa in vivo (Fig. 2E). Taken together, these results suggest that S. pneumoniae PLY plays a critical role in mediating activation of AMPKα1 in middle ear epithelial cells both in vitro and in vivo.
PLY-induced activation of AMPKα1 is independent of TLR2/4-MyD88 signaling but dependent on cholesterol-dependent membrane binding of PLY
To further determine the mechanism underlying S. pneumoniae–induced AMPKα1 activation, we investigated how PLY activates AMPKα1. Because TLR2 and TLR4, as well as MyD88 signaling downstream of TLR2 and TLR4, have been shown to play an important role in mediating host response to S. pneumoniae and PLY (20, 21, 23, 42), we first examined whether TLR2, TLR4, or MyD88 is involved in mediating activation of AMPKα1 by PLY. As shown in (Fig. 3A, PLY was still able to induce AMPKα1 activation in HMEEC-1 cells transfected with a DN mutant form of TLR2 or TLR4, the deletion mutant of Toll/IL-1R homologous region (TIR) domain (TLR2ΔTIR-DN or TLR4ΔTIR-DN). Moreover, PLY still activated AMPKα1 in TLR4-deficient MEFs (Fig. 3B). In addition, PLY still induced AMPKα1 activation in HMEEC-1 cells transfected with MyD88 siRNA (Fig. 3C), suggesting that AMPKα1 activation by PLY is independent of TLR2/4-MyD88 signaling. This finding thus led us to further examine the involvement of other mechanisms.
PLY is known as a pore-forming toxin that belongs to the family of CDCs (17). In this aspect, PLY shows lytic activity that is dependent on the presence of target cell membrane cholesterol that acts as the initial binding ligand to anchor PLY into the host cell membrane (22). Cell membrane–linked PLY monomer will oligomerize to form a complex that causes the formation of ring-like pores (19, 43). Thus, we examined whether cell membrane cholesterol is required for PLY-induced AMPKα1 activation by performing cholesterol inactivation experiments in HMEEC-1 cells. To test this hypothesis, we first stimulated the cells with PLY that had been already preincubated with cholesterol at several mass ratios (1:0.5, 1:1, 1:5, 1:10) prior to being added to the cells, leading to a masking PLY cholesterol-binding loop. As shown in (Fig. 3D, cholesterol preincubation suppressed PLY-induced AMPKα1 activation in a dose-dependent manner. Similar results were also observed in the cells in which cholesterol was depleted from the cell membrane by pretreatment with MβCD (Fig. 3E). Moreover, treatment with an inclusion complex of MβCD and cholesterol restored the ability of PLY to activate AMPKα1 (Fig. 3E), thereby suggesting that the host cell membrane cholesterol is required for AMPKα1 activation by PLY.
Because pore formation by PLY leads to actin cytoskeletal remodeling, such as actin polymerization and subsequent stress fiber, lamellipodia or filopodia formation (40), we next sought to determine whether actin cytoskeletal remodeling is required for AMPKα1 activation by PLY. The cells stimulated with S. pneumoniae D39 WT or recombinant PLY protein, but not the PLY-deficient mutant, indeed exhibited strong actin polymerization at 5 min after stimulation as assessed by phalloidin staining (Fig. 3F). We next sought to determine the role of actin polymerization in PLY-induced AMPKα1 activation using actin-filament formation depolymerizing reagents (latrunculin B and cytochalasin D) or stabilizing reagent (jasplakinolide). As shown in (Fig. 3G, latrunculin B and cytochalasin D inhibited, whereas jasplakinolide stabilized, actin-filament formation in HMEEC-1 cells. However, these reagents did not affect PLY-induced AMPKα1 activation (Fig. 3H), suggesting that actin cytoskeletal remodeling by PLY is not involved in AMPKα1 activation. Collectively, our data demonstrate that cholesterol binding activity of PLY is required for PLY-induced AMPKα1 activation in a TLR2/4-MyD88 signaling-independent manner.
Activation of AMPKα1 is required for S. pneumoniae–induced MUC5AC expression
We next sought to determine the role of AMPKα1 activation in S. pneumoniae–induced upregulation of MUC5AC expression. Interestingly, knockdown of AMPKα1 by AMPKα1-specific siRNA (Fig. 4A, Supplemental Fig. 2C, 2E) inhibited S. pneumoniae–induced upregulation of MUC5AC at the mRNA and protein levels (Fig. 4B–D). Notably, these findings are generalizable to multiple OM-causing S. pneumoniae strains (Fig. 4E). Moreover, PLY-induced MUC5AC expression was also inhibited by AMPKα1 siRNA (Fig. 4F).
To further determine the requirement of AMPKα1 in S. pneumoniae–induced MUC5AC expression, we examined in AMPKα1-knockout (KO) HMEEC-1 cells generated by using the CRISPR-Cas9 system (Fig. 4G, Supplemental Fig. 1) (44). In line with AMPKα1 knockdown by siRNA, we observed that S. pneumoniae no longer induced upregulation of MUC5AC expression in AMPKα1-KO cells (Fig. 4H), whereas the reconstitution of AMPKα1-KO cells with the CA form of AMPKα1 (AMPKα11–312-CA) (32) restored their responsiveness to S. pneumoniae (Fig. 4I, 4J). Collectively, these results demonstrate that activation of AMPKα1 is required for mediating S. pneumoniae–induced MUC5AC upregulation in middle ear epithelial cells in vitro.
Pharmacological inhibition of AMPK suppresses S. pneumoniae–induced upregulation of MUC5AC and the development of OM pathology in the mouse model of S. pneumoniae–induced OM
We next determine the in vivo role of AMPKα1 in upregulation of MUC5AC and immunopathogenesis in the well-established S. pneumoniae–induced OM mouse model. As shown in (Fig. 5A, AMPK-specific inhibitor compound C inhibited MUC5AC upregulation at the mRNA level in the middle ear of mice inoculated with S. pneumoniae as assessed by performing qPCR analysis. In addition, compound C also suppressed upregulation of proinflammatory cytokines/chemokines, including IL-6, MIP-2, IL-1β, CCL2, and GM-CSF, in the middle ear of mice inoculated with S. pneumoniae (Fig. 5B–F), thereby suggesting that compound C may suppress S. pneumoniae–induced host mucosal responses. Compound C also inhibited S. pneumoniae– or PLY-induced MUC5AC upregulation at the protein level in middle ear mucosa epithelia of mice as assessed by IF staining (Fig. 5G). Moreover, compound C inhibited S. pneumoniae–induced mucosal thickening and polymorphonuclear neutrophil infiltration (Fig. 5H). Taken together, our findings suggest that AMPKα1 acts as a key positive regulator for mediating S. pneumoniae–induced upregulation of mucin MUC5AC and proinflammatory mediators, thereby leading to development of OM pathology in the mouse model of OM induced by S. pneumoniae.
AMPKα1 mediates S. pneumoniae–induced MUC5AC expression via inhibition of JNK, the negative regulator of MUC5AC induction
Having shown that activation of AMPKα1 mediates S. pneumoniae–induced MUC5AC upregulation, it still remains largely unknown how AMPKα1 regulates MUC5AC expression. Thus, we examined whether activation of AMPKα1 alone is sufficient to induce MUC5AC expression in HMEEC-1 cells in vitro. Overexpression of the AMPKα1 CA form (AMPKα11–312-CA and AMPKα1T183D-CA) (32) alone failed to induce MUC5AC expression (Fig. 6A). However, AMPKα1-CA enhanced S. pneumoniae–induced MUC5AC expression, whereas the DN mutant form of AMPKα1 (AMPKα1D159N-DN) (32) suppressed S. pneumoniae–induced MUC5AC expression (Fig. 6A). These data suggest that activation of AMPKα1 alone is necessary but insufficient for the induction of MUC5AC expression.
Because MAPK ERK is a critical positive regulator for S. pneumoniae–induced MUC5AC expression whereas MAPK JNK acts as a negative regulator (12), we next sought to determine whether activation of AMPKα1 actually regulates MUC5AC upregulation by modulating these MAPK signaling pathways. We first confirmed the activation of AMPKα1, ERK1/2, and JNK1/2 in S. pneumoniae–stimulated HMEEC-1 cells (Fig. 6B). Interestingly, AMPKα1 knockdown by siRNA enhanced JNK1/2 activation but did not affect ERK1/2 activation (Fig. 6C), suggesting that AMPKα1 acts upstream of JNK1/2. We next sought to determine the effects of inhibition of ERK or JNK on AMPKα1 activation by specific ERK inhibitors PD98059 and U0126 and specific JNK inhibitor SP600125, respectively. As shown in (Fig. 6D and 6E, inhibition of ERK or JNK did not affect AMPKα1 activation by S. pneumoniae, confirming that AMPKα1 does not act downstream of ERK1/2 and JNK1/2. We further evaluated the effect of blocking the JNK pathway by SP600125 on S. pneumoniae–induced MUC5AC expression in HMEEC-1 cells transfected with AMPKα1 siRNA. As shown in (Fig. 6F, JNK inhibition by SP600125 enhanced S. pneumoniae–induced MUC5AC expression in mock cells, which can no longer be suppressed by AMPKα1 knockdown. Moreover, we also assessed the effect of AMPKα1 depletion on S. pneumoniae and PLY-induced JNK activation in middle ear epithelial cells. S. pneumoniae– and PLY-induced JNK activation was significantly enhanced in AMPKα1-KO HMEEC-1 cells (Fig. 6G, 6H). Thus, these results suggest that AMPKα1 mediates upregulation of MUC5AC by S. pneumoniae via inhibition of JNK.
AMPK kinases CaMKKβ and MLK3 mediate S. pneumoniae–induced AMPKα1 activation for regulating MUC5AC expression in human middle ear epithelial cells in vitro
We next sought to investigate how S. pneumoniae PLY activates AMPKα1 by determining the upstream kinases critical for activation of AMPKα1. Because LKB1, CaMKKβ, TAK1, and MLK3 have been reported to induce AMPKα1 phosphorylation at T183 (25), we examined the role of these kinases in S. pneumoniae–induced AMPKα1 activation. As shown in (Fig. 7A, S. pneumoniae–induced phosphorylation of AMPKα1 was observed in the cell lines lacking LKB1 including human cervical carcinoma HeLa cells and lung carcinoma A549 cells (45) as well as in the cell lines harboring WT LKB1 such as HMEEC-1 cells and human bronchial epithelia BEAS-2B cells. Moreover, we observed S. pneumoniae–induced phosphorylation of AMPKα1 in TAK1-knockdown cells by siRNA (Fig. 7B). Because CaMKKβ and MLK3 induce AMPKα1 phosphorylation at T183 to modulate AMPKα1-mediated downstream signaling (25), we next examined the role of CaMKKβ and MLK3 in S. pneumoniae–induced phosphorylation of AMPKα1 and MUC5AC upregulation using specific inhibitors for CaMKKβ and MLK3. Interestingly, treatment with MLK3 inhibitor CEP-1347 (46) and CaMKKβ inhibitor STO-609 (47) inhibited S. pneumoniae–induced phosphorylation of AMPKα1 (Fig. 7C, 7D). Moreover, CEP-1347 and STO-609 also inhibited S. pneumoniae–induced MUC5AC upregulation similarly to the knockdown of AMPKα1 (Fig. 7E). We further confirmed that CEP-1347 and STO-609 also inhibited PLY-induced AMPKα1 activation and MUC5AC upregulation (Fig. 7F–H). Taken together, these data thus suggest that CaMKKβ and MLK3 but not LKB1 and TAK1 are required for S. pneumoniae PLY-induced AMPKα1 activation and upregulation of MUC5AC.
Because activation of small GTPase Rac1, RhoA, and Cdc42 by PLY penetration in the cell membrane regulates their interaction with many effector molecules that control the downstream signaling pathway (40, 48), we evaluated the involvement of small GTPase in PLY-induced activation of AMPKα1. Interestingly, the overexpression of WT Cdc42 but not Rac1 and RhoA enhanced PLY-induced AMPKα1 activation (Fig. 7I). In contrast, overexpression of a DN mutant of Cdc42T17N (Cdc42T17N-DN) but not Rac1T17N (Rac1T17N-DN) and RhoAT19N (RhoAT19N-DN) attenuated PLY-induced AMPKα1 activation (Fig. 7J), suggesting that Cdc42 is involved in PLY-induced AMPKα1 activation in HMEEC-1 cells. These results led us to further postulate that small GTPase Cdc42 may regulate AMPK kinase, leading to the activation of AMPKα1 by PLY. It has been shown that Cdc42 interacts with and activates MLK3 (49, 50), and our data in (Fig. 7C and 7F indicated that MLK3 acts as an AMPK kinase during PLY stimulation. Thus, we further determined the involvement of Cdc42 in MLK3-induced activation of AMPKα1 by using a kinase-dead mutant MLK3K144R and Cdc42Y40C mutant in which its binding with Cdc42/Rac interactive binding motif (CRIB)-containing protein is defective (51, 52). As shown in (Fig. 7K, overexpression of MLK3K144R attenuated enhancement of AMPKα1 activation in the cells cotransfected with WT Cdc42. Moreover, overexpression of Cdc42Y40C attenuated PLY-induced AMPKα1 activation. These results suggest that small GTPase Cdc42 also plays an important role in MLK3-induced AMPKα1 activation. Collectively, our data demonstrate that S. pneumoniae PLY induces AMPKα1 activation for regulating MUC5AC expression in a Ca2+-CaMKKβ and Cdc42-MLK3 pathway-dependent manner.
The Gram-positive bacterium S. pneumoniae is the most common bacterial pathogen causing OM and upregulates mucin MUC5AC expression, leading to the development of conductive hearing loss (9). The molecular mechanisms underlying the tight regulation of S. pneumoniae–induced MUC5AC expression still remain largely unknown. In this study, we showed that S. pneumoniae PLY markedly induced activation of AMPKα1, previously known as a master regulator of energy homeostasis, and AMPKα1 positively regulated S. pneumoniae–induced MUC5AC expression in the middle ear epithelial cells in vitro and in the middle ear mucosal epithelia of mice in vivo. We also showed that cholesterol binding activity of PLY is required for TLR2/4-independent activation of AMPKα1 in a Ca2+-CaMKKβ and Cdc42-MLK3 pathway-dependent manner. We also found that AMPKα1 positively regulates PLY-induced MUC5AC upregulation by inhibiting JNK, the negative regulator for MUC5AC induction. In addition, pharmacological inhibition of AMPKα1 suppressed S. pneumoniae–induced MUC5AC upregulation and attenuated the pathological changes in the well-established OM model of mice. Taken together, our data unveil a novel mechanism by which negative cross-talk between TLR2/4-independent activation of AMPKα1 and TLR2/4-dependent activation of JNK controls the S. pneumoniae PLY-induced host innate mucosal response by regulating MUC5AC expression and may lead to the development of new therapeutic strategies for OM (Fig. 8).
One of the major interesting findings in our study is the identification of a previously unreported role of AMPKα1, a master regulator of energy homeostasis, in tightly regulating mucin MUC5AC induction by S. pneumoniae in the pathogenesis of OM. Importantly, overexpression of the AMPKα1-CA form alone is insufficient for inducing MUC5AC expression, indicating the requirement of other signaling pathways activated by S. pneumoniae (Fig. 6). Previously, we showed that S. pneumoniae PLY upregulates MUC5AC expression via activation of TLR4-MEK1-ERK1/2. In contrast, S. pneumoniae also induces TLR4-MEKK3-JNK1/2 activation and ERK1/2-dependent MAPK phosphatase-1 (MKP-1) upregulation, which in turn limits ERK1/2-mediated induction of MUC5AC (12, 24). In this study, we identified a novel mechanism by which TLR2/4-independent activation of AMPKα1 suppresses TLR4-dependent activation of JNK1/2, thus leading to the upregulation of PLY-induced upregulation of MUC5AC (Figs. 3, 6, 7). Thus, these data unveil a previously unreported positive regulatory mechanism for MUC5AC induction via AMPKα1. Future study is required to further elucidate the precise mechanisms by which activation of AMPKα1 by S. pneumoniae inhibits JNK activation. For example, MKP-1 also interacts with and dephosphorylates JNK1/2, thereby leading to upregulation of MUC5AC during S. pneumoniae infection (24). Nicotine, a major component of cigarette smoke, activates AMPKα2, which in turn phosphorylates MKP-1 at serine 334, initiating its proteasome-dependent degradation and facilitating aberrant JNK1/2 activation in adipocytes (53). Thus, it is likely that activation of AMPKα1 by S. pneumoniae may also regulate JNK1/2 activation through phosphorylation and degradation of MKP-1. Taken together, these data suggest that both TLR2/4-dependent and -independent mechanisms are coordinated to tightly regulate host mucosal responses during S. pneumoniae infection.
The second major finding of this study is that PLY is specifically required for S. pneumoniae–induced AMPKα1 activation via a TLR2/4-independent mechanism to tightly control TLR2/4-dependent upregulation of MUC5AC expression. Among a variety of S. pneumoniae virulence factors that have been identified, including its capsule, PspA/C, LytA, and ClpL (48, 54, 55), PLY is well known as a major virulence factor to induce the host innate immune responses and contributes significantly to the morbidity and mortality associated with S. pneumoniae infection (14, 15). Previous studies reported that PLY exerts its virulence activity through two distinct TLR2/4-dependent and -independent mechanisms: 1) TLR4-dependent activation of downstream MyD88/TRAF6 signaling axis, and 2) membrane pore-forming activity via its cholesterol-dependent binding to the cell membrane (20, 22, 23, 56). As shown in this study, disrupting TLR2/4-MyD88 signaling exhibited no effect on PLY-induced AMPKα1 activation. In contrast, preincubation of PLY with cholesterol or cholesterol depletion by MβCD significantly suppressed PLY-induced AMPKα1 activation, whereas the cholesterol-MβCD inclusion complex restored it, thereby providing supportive evidence for the critical contribution of cholesterol binding activity for AMPKα1 activation by PLY. It is noteworthy that, in addition to S. pneumoniae, a number of Gram-positive bacterial pathogens such as Clostridium perfringens, Streptococcus pyogenes, and Listeria monocytogenes also produce CDC toxin perfringolysin O, streptolysin O, and listeriolysin O, respectively. Importantly, structural analysis revealed that the tryptophan-rich motif at domain 4 and residues at the interface for monomer packing are conserved in members of CDC toxin, suggesting that these CDC toxins share a common mechanism for its assembly to prime cell membrane penetration and pore formation (57, 58). Thus, future studies are needed to investigate whether these CDC toxin-producing pathogens also induce AMPKα1 activation and further establish the generalizability of our finding.
Given that PLY is a critical inducer for AMPKα1 activation in a cholesterol binding–dependent manner, it is likely that PLY-mediated pore-forming events in the host cell membrane lead to AMPKα1 activation. Previous studies revealed that the pore-forming by PLY penetration into the cell membrane via cholesterol-dependent binding facilitates not only Ca2+ influx into the cells through the pore but also activation of small GTPases Rac1, RhoA, and Cdc42 for promoting actin cytoskeletal remodeling (40, 43). To date, LKB1, CaMKKβ, TAK1, and MLK3 have been recognized as the AMPK kinase in the context of different physiological/pathological conditions (25). Consistent with previous findings demonstrating that CaMKKβ activates AMPKα1 in response to elevated intracellular Ca2+ concentrations, PLY-induced AMPKα1 activation is inhibited by the CaMKKβ inhibitor STO-609 (Fig. 7). Interestingly, the overexpression of WT Cdc42 enhanced, whereas the overexpression of Cdc42T17N-DN inhibited, PLY-induced AMPKα1 activation (Fig. 7). We also observed the overexpression of the Cdc42Y40C mutant that is defective in interacting with the CRIB motif of MLK3, the MLK3 inhibitor CEP-1347, and a kinase-dead mutant MLK3K144R suppressed PLY-induced AMPKα1 activation (Fig. 7). In contrast, depletion of LKB1 or TAK1 exhibited no effect on PLY-induced AMPKα1 activation (Fig. 7). Thus, it is likely that both CaMKKβ and Cdc42-MLK3 are involved in mediating AMPKα1 activation by PLY. This finding may provide novel insight into the regulation of AMPKα1 activity via the activation of small GTPase Cdc42 in a context of certain physiological conditions, although the precise molecular mechanism of AMPKα1 activation by Cdc42-MLK3 remains to be understood. One possible explanation is the possible involvement of osmotic stresses caused by PLY-induced pore formation into the cell membrane in epithelial cells (59). Future studies are also needed for further elucidating the underlying mechanism.
In this study, we found that S. pneumoniae induced AMPKα1 activation at early time points (9 h after S. pneumoniae inoculation in the middle ear of mouse). In contrast, Hoogendijk et al. (60) previously showed that AMPKα activity declines at late time points (24–48 h after S. pneumoniae inoculation in the lungs of mice). This discrepancy may be attributed to several differences. First, we observed S. pneumoniae–induced phosphorylation of AMPKα1 in middle ear mucosa epithelial cells of mice as assessed by IF staining (Fig. 1D), whereas Hoogendijk et al. observed S. pneumoniae–induced phosphorylation of AMPKα1 in the homogenized lysates of mouse lung as assessed by Western blot analysis. Second, we observed AMPKα1 activation at early time points (9 h after S. pneumoniae inoculation), whereas Hoogendijk et al. showed that AMPKα activity increased slightly at 3 h, peaked at 6 h, and declined thereafter at late time points (24–48 h after S. pneumoniae inoculation). Third, we inoculated S. pneumoniae (1 × 107 CFU) into the middle ear of mice whereas Hoogendijk et al. intranasally inoculated S. pneumoniae (5 × 104 CFU) to the mice. Future studies are needed to further determine whether these differences indeed lead to the discrepancy described above.
Initially, AMPK was predominantly characterized as a master regulator of energy homeostasis (25). In addition, the recent accumulating studies suggest that AMPK regulates diverse physiological processes and could be dysregulated in the context of certain chronic diseases such as metabolic syndrome, inflammation, and cancer (26). Thus, AMPK is emerging as one of the promising targets for the prevention and treatment of these diseases (61–63). Of particular interest in the current study is that we demonstrated for the first time, to our knowledge, the new physiological/pathological role of AMPKα1 in mediating the host mucosal responses through upregulation of mucin MUC5AC during bacterial infection in mucosal epithelial cells. For examples, we observed that pharmacological inhibition of AMPKα1 by the AMPK-specific inhibitor compound C suppressed S. pneumoniae–induced MUC5AC upregulation as well as mucosal thickening in the middle ear epithelial mucosa in the OM mouse model (Fig. 5). These findings suggest that upregulated MUC5AC production via activation of AMPKα1 in mucosal epithelial cells may contribute significantly to the pathological development of S. pneumoniae infection in the middle ear. In addition to suppression of mucin, pharmacological inhibition of AMPKα1 by compound C also suppressed S. pneumoniae–induced expression of cytokines/chemokines and polymorphonuclear neutrophil infiltration in middle ear epithelial mucosa in OM model mice (Fig. 5). All contributed significantly to the development of conductive hearing loss (1, 5). In addition to in vivo findings, compound C suppressed S. pneumoniae–induced upregulation of MUC5AC and cytokines/chemokines in HMEEC-1 cells in vitro (Supplemental Fig. 3). These data suggest that inhibition of AMPKα1 may suppress S. pneumoniae–induced host mucosal responses due to less damaging mucus production. Future studies will also focus on exploring the therapeutic potential of targeting AMPKα1 for controlling mucin overproduction in respiratory diseases including OM and chronic obstructive pulmonary disease.
We thank Dr. David Briles, Dr. David J. Lim, Dr. Douglas T. Golenbock, Dr. Shizuo Akira, and Dr. Jingren Zhang for providing various reagents.
This work was supported by National Institutes of Health Grants DC013833, DC015557, and DC019512 (to J.-D.L.), the Japan Society for the Promotion Science (JSPS, https://www.jsps.go.jp/english/) program on Strategic Young Researchers Overseas Visits Program for Accelerating Brain Circulation (Grant S2510 to H.K.), Grants-in-Aid for JSPS Fellows (to S.M.), and by the Program for Leading Graduate Schools “HIGO” in Kumamoto University from the Ministry of Education, Culture, Sports, Science and Technology (http://www.mext.go.jp/en/), Japan. J.-D.L. is a Georgia Research Alliance Eminent Scholar in Inflammation and Immunity.
S.M. designed and performed experiments, analyzed data, and wrote the manuscript; K.K. contributed to experimental design and data analysis and wrote the manuscript; B.-C.L., Y.T., M.M., T.S., and H.K. contributed to experimental design, data analysis, and manuscript preparation; H.X. contributed to data analysis and manuscript preparation; J.-D.L. was responsible for overall study design and data analysis, and wrote the manuscript and supervised the project.
The online version of this article contains supplemental material.
Abbreviations used in this article:
AMP-activated protein kinase
Ca2+/calmodulin-dependent kinase kinase β
human middle ear epithelial cell
liver kinase B1
mouse embryonic fibroblast
mixed-lineage protein kinase 3
single guide RNA
small interfering RNA
TGF-β–activated kinase 1
Toll/IL-1R homologous region
The authors have no financial conflicts of interest.