Abstract
Background Chronic airway inflammation is the main driver of pathogenesis in respiratory diseases such as severe asthma, chronic obstructive pulmonary disease, cystic fibrosis (CF) and bronchiectasis. While the role of common pathogens in airway inflammation is widely recognised, the influence of other microbiota members is still poorly understood.
Methods We hypothesised that the lung microbiota contains bacteria with immunomodulatory activity which modulate net levels of immune activation by key respiratory pathogens. Therefore, we assessed the immunomodulatory effect of several members of the lung microbiota frequently reported as present in CF lower respiratory tract samples.
Results We show that Rothia mucilaginosa, a common resident of the oral cavity that is also often detectable in the lower airways in chronic disease, has an inhibitory effect on pathogen- or lipopolysaccharide-induced pro-inflammatory responses, in vitro (three-dimensional cell culture model) and in vivo (mouse model). Furthermore, in a cohort of adults with bronchiectasis, the abundance of Rothia species was negatively correlated with pro-inflammatory markers (interleukin (IL)-8 and IL-1β) and matrix metalloproteinase (MMP)-1, MMP-8 and MMP-9 in sputum. Mechanistic studies revealed that R. mucilaginosa inhibits NF-κB pathway activation by reducing the phosphorylation of IκBα and consequently the expression of NF-κB target genes.
Conclusions These findings indicate that the presence of R. mucilaginosa in the lower airways potentially mitigates inflammation, which could in turn influence the severity and progression of chronic respiratory disorders.
Abstract
A commensal bacterium of the lower airways, Rothia mucilaginosa, inhibits inflammation by NF-κB pathway inactivation. R. mucilaginosa abundance inversely correlates with sputum pro-inflammatory markers in chronic lung disease, indicating a beneficial role. https://bit.ly/3lNT9th
Introduction
Persistent or dysregulated lung inflammation is a hallmark of chronic respiratory diseases such as severe asthma, chronic obstructive pulmonary disease (COPD), cystic fibrosis (CF) and bronchiectasis [1–9]. In patients with chronic airway disease, the release of pro-inflammatory cytokines by cells of the lung mucosa in response to external triggers (pathogenic microorganisms, cigarette smoke, pollutants or allergens) typically leads to infiltration of innate immune cells such as neutrophils or eosinophils that further exacerbate lung inflammation [5–7, 10, 11]. Persistent inflammation and associated tissue damage impairs lung clearance and facilitates the accumulation of microbes in the lower airways. These microbes represent both recognised respiratory pathogens and oropharyngeal commensal taxa that originate in the upper airways [12].
Colonisation of the lower airways by specific bacterial and fungal species contributes directly to inflammation, tissue damage and lung disease progression [13–15]. These species include Pseudomonas aeruginosa and Staphylococcus aureus in patients with CF [16–18], Haemophilus influenzae and P. aeruginosa in patients with bronchiectasis [19], H. influenzae in patients with neutrophilic asthma [20], and Moraxella catarrhalis in patients with COPD [21]. The presence and abundance of these pathogens is largely viewed as the sole microbial determinant of airway disease. However, in addition to these pathogens, the lower airways of patients with chronic lung disease are also colonised by a much wider microbiota that includes many bacterial taxa not considered to contribute directly to pathogenesis [12, 14, 22–24].
The composition of this wider airway microbiota differs with disease characteristics, severity and patient age [25–27]. However, the potential of microbiota members to influence disease by mediating or moderating the impact of pro-inflammatory external triggers, including airway pathogens, is little understood [28–30].
We hypothesised that the lung microbiota contains bacteria with immunomodulatory activity which modulate net levels of immune activation by key respiratory pathogens, such as P. aeruginosa. Therefore, we assessed the immunomodulatory effect of several members of the lung microbiota frequently reported as present in CF lower respiratory tract samples [21, 31, 32]. These species represented recognised pathogens (S. aureus), less frequently recovered pathogens (Achromobacter xylosoxidans and Streptococcus anginosus) and upper respiratory tract commensals thought to be broadly nonpathogenic in this context (Rothia mucilaginosa and Gemella haemolysans). The latter includes the Gram-positive coccus R. mucilaginosa, a common commensal of the oral cavity that is frequently detected in the lungs of patients with chronic lung disease [23, 33–36], and has been shown to be metabolically active and adapted to the CF lung environment [37]. We found that R. mucilaginosa exhibits anti-inflammatory activity in vitro. We subsequently validated our findings in an animal model of lung inflammation and in a cohort of patients with neutrophilic airways disease. Finally, we identified the host pathway that R. mucilaginosa interferes with to exert its anti-inflammatory effect.
Methods
Bacterial species and culturing conditions
Six bacterial species (table 1) commonly isolated from the CF lung were selected for this study, including two pathogens (P. aeruginosa and S. aureus), two less frequently recovered pathogens (S. anginosus and A. xylosoxidans) and two bacteria that are not typically considered as pathogens (R. mucilaginosa and G. haemolysans). Selective media for these species were previously developed [38, 39]. For select experiments we also included additional Rothia species, i.e. Rothia dentocariosa, Rothia terrae, Rothia amarae and Rothia aeria. All isolates were cultured at 37°C under constant shaking conditions (250 rpm) in Brain Heart Infusion broth (Lab M, Bury, UK) until stationary phase. Most isolates were cultured in aerobic conditions except for S. anginosus and G. haemolysans, which were cultured in microaerobic conditions (<1% O2) (Oxoid CampyGen Compact; Thermo Fisher Scientific, Waltham, MA, USA).
Overview of strains used in this study
Three-dimensional lung epithelial cell culture models
A549 cell line
A previously developed organotypic three-dimensional (3D) cell culture model was used to study the immune response to pro-inflammatory stimuli [40–43]. 3D in vivo-like lung epithelial cell culture model systems reflect key aspects of the parental tissue, including 3D architecture, barrier function, apical–basolateral polarity and multicellular complexity [40, 42]. It has been demonstrated that P. aeruginosa adhesion, as well as host-secreted cytokine profiles, in 3D cell culture models of lung epithelial cells are more similar to the in vivo situation than cells grown as a monolayer on plastic [41]. A 3D model of the human adenocarcinoma alveolar lung epithelial cell line A549 (ATCC CCL-185) was generated in GTSF-2 medium (HyClone, Logan, UT, USA) supplemented with 1.5 g·L−1 sodium bicarbonate (Sigma-Aldrich, St Louis, MO, USA), 10% fetal bovine serum (FBS) (Life Technologies, Carlsbad, CA, USA), 2.5 mg·L−1 insulin transferrin sodium selenite (Lonza, Basel, Switzerland) and 1% penicillin/streptomycin (Life Technologies) at 37°C under 5% CO2. 3D A549 cells were cultured for 11–14 days and were transferred to 48-well plates at a concentration of 2.5×105 cells per well containing fresh serum-free GTSF-2 medium on the day of the infection.
IB-3 and S9 cell lines
The aforementioned 3D cell culture model was optimised for use with IB-3 and S9 cell lines [44]. The IB-3 cell line is a bronchial epithelial cell line heterogeneous for F508del (F508del/W1282X). The S9 cell line originates from the IB-3 cell line and is stably transduced with wild-type CF transmembrane conductance regulator (CFTR). LHC-8 medium without gentamicin (Life Technologies) supplemented with 5% FBS and 1% penicillin/streptomycin was used to culture both cell lines. The 3D models of IB-3 and S9 cells were generated as described for the 3D A549 model, except that a cell:bead ratio of 4:1 was used (compared with 2:1 for the 3D A549 model).
In vitro infection studies
Prior to infection, bacterial cultures were centrifuged and resuspended in serum-free host cell culture medium. Cells were infected with single bacterial cultures for 4 h at a multiplicity of infection (MOI) of 10:1, unless otherwise indicated. Additionally, cells were infected with P. aeruginosa together with each one of the CF lung microbiota members for 4 h at a 10:1 ratio, resulting in an MOI of 20:1. When indicated, 3D cell aggregates were exposed to other pro-inflammatory stimuli, i.e. S. aureus (MOI 10:1), lipopolysaccharide (LPS) (100 µg·mL−1) (Sigma-Aldrich), rhamnolipid (100 µg·mL−1) (Sigma-Aldrich) and H2O2 (1 mM), together with R. mucilaginosa for 4 h. For longer term experiments, 3D cell aggregates were exposed to 100 µg·mL−1 LPS alone or in combination with R. mucilaginosa for 24 h at an MOI of 1:1. For experiments with cell-free supernatant of R. mucilaginosa, cultures were centrifuged and filtered through a 0.22 μm PET filter membrane (GE Healthcare, Chicago, IL, USA). To obtain lysed cell suspension, R. mucilaginosa cultures were subjected to a cold/heat cycle twice (i.e. 5 min ice followed by 5 min at 85°C). Viability of lysed cells was checked by plating on nutrient agar, confirming no culturable bacteria. 3D A549 cells were exposed for 4 h to P. aeruginosa PAO1 (MOI 10:1) with or without R. mucilaginosa supernatant (100 µL cell-free supernatant and 150 µL GTSF-2 cell culture medium without FBS), lysed cells (100 µL lysed cell suspension initially containing 2.5×106 cells, which is equivalent to an MOI of 10:1 and 150 µL cell culture medium) or live cells as described earlier (MOI 10:1).
Cytotoxicity assays
After infection of the 3D cells with various bacteria or exposure to other pro-inflammatory stimuli (LPS, rhamnolipid and H2O2), cell viability was assessed using the annexin V/propidium iodide (Life Technologies) and lactate dehydrogenase (Sigma-Aldrich) assays, according to the manufacturer's instructions.
Bacterial cell adherence analysis
Bacterial adherence to the host cells was determined as described previously [42]. Previously developed selective media were used when host cells were co-exposed to P. aeruginosa and R. mucilaginosa, i.e. Luria–Bertani agar supplemented with 1.25 mg·mL−1 triclosan to select for P. aeruginosa, and nutrient agar supplemented with 5 mg·mL−1 mupirocin and 10 mg·mL−1 colistin sulfate to select for R. mucilaginosa [38].
Evaluation of interleukin-8 degradation
Recombinant interleukin (IL)-8 standard (BioLegend, San Diego, CA, USA) at a concentration of 125 pg·mL−1 was added to R. mucilaginosa culture in GTSF-2 medium without serum (5×107 CFU·mL−1 corresponding to an MOI of 10:1) or to control medium. The IL-8 concentration was measured after 0, 15, 30, 60 and 90 min using ELISA as described in the section on inflammatory marker quantification.
In vivo mouse model
12-week-old female BALB/c ByJRj mice (Janvier Labs, Le Genest-Saint-Isle, France) were managed in accordance with the guidelines provided by the European Directive for Laboratory Animal Care (Directive 2010/63/EU of the European Parliament). The laboratory Animal Ethics Committee of the University of Antwerp (Antwerp, Belgium) authorised and approved all animal experiments in this study (file 2019-90). Intratracheal administration of R. mucilaginosa embedded in alginate beads in the presence or absence of LPS was done using a protocol described by Moser et al. [45], with adaptations described in the supplementary material. Embedding bacteria in alginate is a technique developed to retain a sufficient microbial load in the lungs over an extended period of time without immunosuppression [46]. In a preliminary experiment using R. mucilaginosa cell suspension for intratracheal administration, consistent colonisation could not be obtained (supplementary material). Moreover, for R. mucilaginosa, the method used to prepare alginate beads removed bacterial clumps >0.25 mm, thereby limiting the risk of blocking the airways of the mice. Hence, embedding of R. mucilaginosa in alginate beads was used for all experiments. For preparation of the R. mucilaginosa-containing alginate beads, three independent cultures were pooled. Mice were instilled with 10 µg per 50 µL LPS with or without R. mucilaginosa (either R. mucilaginosa DSM20746 or the CF isolate R. mucilaginosa B03V1S1C at the maximal dose that could be obtained in the alginate beads, i.e. ∼1×105 CFU per mouse) embedded in alginate beads for 48 h. Mice were observed during this 48 h for fur quality, posture, state of activity and respiratory symptoms, and were weighed every 24 h. Mice were euthanised by cervical dislocation 48 h post-infection. The left lung was homogenised and used for determination of the microbial load (by plate counting) as well as for cytokine quantification. The spleen and medial lobe of the liver were collected and homogenised to check for dissemination of the instilled bacteria (by plate counting). The right lung was collected in 1 mL of 4% PFA for histological analysis. Sections for histological analysis stained by haematoxylin/eosin were examined blindly and scored as previously described using an EVOS FL Auto microscope (Life Technologies) [47].
Quantification of Rothia species in respiratory samples
Archived, induced sputum samples were obtained from 85 adults with confirmed bronchiectasis and with a history of two or more infective exacerbations (associated with bacterial infection based on the requirement for supplemental systemic antibiotic therapy) in the preceding year, recruited as part of the BLESS randomised controlled trial [48]. Information on the patient cohort (including inclusion/exclusion criteria and sputum induction procedure) is provided in the supplementary methods. The study was approved by the Mater Health Service (South Brisbane, Australia) Human Research Ethics Committee and informed consent was obtained for all participants. Induced sputum was collected at baseline, prior to any intervention, and all patients had a sputum neutrophil abundance of ≥70% (as a percentage of total sputum cells). All patients had a chronic lower airway infection. The relative abundance of Rothia species was measured in 79 out of 85 samples using 16S rRNA gene amplicon sequencing as described previously [49] and as detailed in the supplementary methods. Sequence data are available at the National Center for Biotechnology Information Sequence Read Archive with accession number SRA128000. R. mucilaginosa absolute abundance was measured by quantitative PCR (qPCR) in 82 out of 85 samples using the Qiagen Microbial DNA qPCR Assay for R. mucilaginosa (BPID00297A; Qiagen, Hilden, Germany). Gene copy numbers were determined by comparing with a standard curve as described previously [50].
Inflammatory marker quantification
For in vitro experiments, IL-8 secretion was measured in the cell culture supernatant by a Human IL-8 ELISA MAX Standard assay (BioLegend). For in vivo animal experiments, macrophage inflammatory protein (MIP)-2 secretion was measured in lung homogenate by a MIP-2 Mouse ELISA kit (LabNed, Amstelveen, The Netherlands). All other cytokines for in vitro cell culture and in vivo animal experiments were quantified by Bioplex Multiplex assays (Bio-Rad, Hercules, CA, USA). Inflammatory markers were measured in the sputum of bronchiectasis patients using ELISA (BD Biosciences, San Jose, CA, USA) for IL-8 and IL-1β, and using Magnetic Luminex Performance Assay multiplex kits (R&D Systems, Minneapolis, MN, USA) for matrix metalloproteinase (MMP)-1, MMP-8 and MMP-9 as described previously [51].
Quantitative reverse transcriptase-PCR
3D A549 cells were infected with P. aeruginosa PAO1 or a combination of P. aeruginosa PAO1 and R. mucilaginosa DSM20746 as described earlier. After 4 h of infection, 3D A549 cells were washed three times with Hank's Balanced Salt Solution and RNAprotect Reagent (Qiagen) was added. Total cellular RNA was extracted using the Aurum Total RNA Mini Kit (Bio-Rad) and reverse transcribed into cDNA using the iScript advanced cDNA synthesis kit (Bio-Rad). The expression of various genes involved in inflammation was quantified using the “Bacterial infections in normal airways” PrimePCR panel (Bio-Rad) and SsoAdvanced SYBR Green supermix (Bio-Rad). qPCR experiments were performed on a Bio-Rad CFX96 Real-Time System C1000 Thermal Cycler. TBP, GAPDH and HPRT1 were used as reference genes. Genes with an insufficient expression level were excluded from further analysis.
Analysis of NF-κB activation
A 3D epithelial cell model of NF-κB–luciferase-transfected A549 cells (BPS Bioscience, San Diego, CA, USA) was developed as described for the 3D A549 model. Cells were exposed for 4 h to single or co-cultures of P. aeruginosa and R. mucilaginosa (or its cell-free supernatant) as described earlier. Likewise, screening of NF-κB activation by LPS (100 µg·mL−1) with or without 36 isolates representing five different Rothia species (table 1) was performed. When indicated, the effect of Rothia species (R. mucilaginosa DSM20146, R. dentocariosa HVOC18-02, R. aeria HVOC13-01 and R. terrae HVOC29-01) on NF-κB pathway activation by LPS (100 µg·mL−1) was evaluated under microaerobic conditions (3% O2, 5% CO2 and 92% N2) in a BACTROX-2 Hypoxia Chamber (SHEL LAB, Cornelius, OR, USA). The One-Step Luciferase Assay System (BPS Bioscience) was used to quantify NF-κB pathway activation and luminescence was measured by an EnVision luminometer (PerkinElmer, Waltham, MA, USA).
Western blot analysis
The Western blot experiments were done as described previously [52]. Mouse monoclonal antibodies, at concentrations recommended by the manufacturer, were used to detect A20 (Santa Cruz Biotechnology, Dallas, TX, USA), NF-κB p65 (Santa Cruz Biotechnology) and phosphorylated (p)-IκBα (Cell Signaling Technology, Danvers, MA, USA) and a goat polyclonal antibody was used to detect IκBα (Santa Cruz Biotechnology). As secondary antibody, an anti-mouse IgG horseradish peroxidase (HRP)-linked antibody (Santa Cruz Biotechnology) or anti-goat IgG HRP-linked antibody (Santa Cruz Biotechnology) was used.
Statistical analysis
All in vitro experiments were carried out at least in biological triplicates (n≥3). In vitro infection experiments contained two technical replicates for assessment of the viability, quantification of CFU by plating and quantification of cytokines. For animal experiments, MIP-2, CFU, disease symptoms in the animals, colonisation and dissemination were determined for two independent experiments with R. mucilaginosa DSM20746. Validation of the in vivo data with another R. mucilaginosa strain (B03V1S1C) and quantification of additional cytokines as well as histological analysis were done for one experiment for both strains.
Statistical analysis was performed using SPSS version 24.0 (IBM, Armonk, NY, USA). The Shapiro–Wilk test was applied to evaluate normality of the data. For normally distributed data, differences between the means were assessed by an independent samples t-test or ANOVA followed by Dunnett's post hoc analysis. Nonnormally distributed data were further analysed by a Kruskal–Wallis nonparametric test or Mann–Whitney test. Statistical significance is assumed at p<0.05. For correlation analysis, between Rothia species abundance (relative or absolute) and cytokine/MMP concentrations, correlation coefficients and p-values were determined by either Pearson's test (r) for parametric distributions or Spearman's test (rs) for nonparametric distributions.
Results
R. mucilaginosa inhibits the production of pro-inflammatory cytokines by lung epithelial cells in vitro
Exposure of an in vivo-like 3D alveolar epithelial model to P. aeruginosa PAO1 (MOI 30:1) induced high IL-8 production, while low or moderate IL-8 production was observed following exposure to S. aureus SP123, S. anginosus LMG14696, A. xylosoxidans LMG26680, G. haemolysans LMG18984 and Rothia mucilaginosa DSM20746 (figure 1a). When 3D A549 cells were co-exposed to P. aeruginosa and one of each of the five other bacterial species, R. mucilaginosa completely abolished the P. aeruginosa-induced IL-8 response (figure 1b). R. mucilaginosa exposure also lowered the IL-8 response induced by P. aeruginosa at an MOI of 100:1 and 10:1 (supplementary figure S1a). The effect of R. mucilaginosa DSM20746 on P. aeruginosa-induced IL-8 production was confirmed with two other strains of R. mucilaginosa (ATCC 49042 and a CF sputum isolate B03V1S1C) (supplementary figure S1b). The three R. mucilaginosa strains also reduced the IL-8 response induced by various P. aeruginosa CF sputum isolates (supplementary figure S1b). Reduction in IL-8 response by R. mucilaginosa was confirmed using LPS (100 µg·mL−1) as the pro-inflammatory stimulus, both for short-term (4 h) and long-term (24 h) exposure (figure 1c). Under the experimental conditions tested, a similar number of R. mucilaginosa CFU was observed between the initial inoculum (0 h) and the total CFU that associated with the host cells and was present in the surrounding liquid at 4 and 24 h, indicating that bacteria remained viable throughout the experiment (supplementary figure S1c). After 4 h, a slight increase in total R. mucilaginosa CFU was observed compared with the inoculum (supplementary figure S1c).
Three-dimensional (3D) lung epithelial responses to pro-inflammatory stimuli in the presence and absence of members of the lung microbiota: Pseudomonas aeruginosa PAO1 (Pa), Rothia mucilaginosa DSM20746 (Rm), Staphylococcus aureus SP123 (S), Streptococcus anginosus LMG14696 (St), Achromobacter xylosoxidans LMG26680 (A) and Gemella haemolysans LMG18984 (G). a, b) Interleukin (IL)-8 production by 3D A549 cells after 4 h exposure to a) single bacterial cultures or b) co-cultures of various lung microbiota members with P. aeruginosa PAO1 at an MOI of 30:1. c) IL-8 production by 3D A549 cells after 4 or 24 h exposure to 100 μg·mL−1 lipopolysaccharide (LPS) alone or in co-culture with R. mucilaginosa at an MOI 10:1 (4 h) or 1:1 (24 h). d) IL-6, IL-8, granulocyte–macrophage colony-stimulating factor (GM-CSF) and monocyte chemoattractant protein (MCP)-1 production of 3D A549 cells after 4 h exposure to P. aeruginosa alone or in co-culture with R. mucilaginosa at an MOI of 10:1. NC: negative control (uninfected 3D epithelial cells in serum-free medium). Data represent mean±sem or mean, n≥3. *: p<0.05; **: p<0.001.
R. mucilaginosa also significantly reduced IL-8 levels promoted by the Gram-positive pathogen S. aureus (MOI 10:1) in 3D A549 cells (supplementary figure S1d). The anti-inflammatory effect of R. mucilaginosa was confirmed in a 3D model of CF bronchial epithelial cells (IB-3 cell line) and in the CFTR corrected control (S9 cell line) (supplementary figure S1e and f).
In addition, R. mucilaginosa reduced levels of other P. aeruginosa-induced pro-inflammatory cytokines (IL-6, IL-8, granulocyte–macrophage colony-stimulating factor (GM-CSF) and monocyte chemoattractant protein (MCP)-1) (figure 1d). Exposure to P. aeruginosa, R. mucilaginosa or the combination showed normal morphology based on microscopic analysis and cells retained >80% viability after infection (supplementary figure S2).
R. mucilaginosa lowers the LPS-induced pro-inflammatory response in an in vivo mouse model
Next, we evaluated the in vivo anti-inflammatory effect of R. mucilaginosa by measuring mouse lung inflammatory cytokine levels following LPS instillation or LPS and R. mucilaginosa (either DSM20746 and B03V1S1C strains) co-exposure (figure 2a). After 48 h of co-exposure, LPS-induced MIP-2 (homologue for human IL-8) was reduced by 40% by strain DSM20746 (figure 2b). The results obtained were validated using another R. mucilaginosa strain (B03V1S1C), which demonstrated a reduction of LPS-induced MIP-2 production of 70% (figure 2b). Levels of other cytokines (MCP-1, IL-6, IL-1α, IL-5, IL-10 and GM-CSF) were also significantly lower in mice exposed to LPS and R. mucilaginosa B03V1S1C compared with LPS alone (supplementary figure S3). The lungs of R. mucilaginosa-infected mice contained a similar number of bacteria (CFU·mL−1) in the presence and absence of LPS (figure 2c). No bacteria were detected in the liver and spleen, suggesting no dissemination, and mice showed an overall normal fur quality, posture, body weight and no respiratory symptoms in all test conditions. Haematoxylin/eosin-stained lung sections of LPS-instilled mice showed significantly more pronounced tissue damage compared with lung sections of mice co-exposed to LPS and R. mucilaginosa B03V1S1C, and the same trend for R. mucilaginosa DSM20746 was observed (figure 2d and e).
Influence of Rothia mucilaginosa (Rm) on the in vivo responses to lipopolysaccharide (LPS). a) Timeline of animal infection and necropsy, b) macrophage inflammatory protein (MIP)-2 concentration (measured by ELISA), and c) number of CFU·mL−1 in mice lung homogenates after 48 h exposure to vehicle (n=7), LPS (n=18), R. mucilaginosa suspension (n=9 for R. mucilaginosa DSM20746 and n=3 for R. mucilaginosa B03V1S1C) or a combination of LPS+R. mucilaginosa (n=18 for R. mucilaginosa DSM20746 and n=3 for R. mucilaginosa B03V1S1C). d) Lung histopathological score on haematoxylin/eosin-stained sections of the right lung. e) Lung haematoxylin/eosin-stained sections of vehicle, LPS, R. mucilaginosa B03V1S1C and LPS+R. mucilaginosa B03V1S1C instilled mice. Scale bar: 200 μm. Vehicle: sterile alginate beads; LPS: 10 μg per 50 μL. The data presented is from two independent animal experiments for strain DSM20746 and one experiment for strain B03V1S1C. Data represent mean±sem, n≥3. *: p<0.05; **: p<0.01.
Rothia species inhibit NF-κB activation in epithelial cells
To decipher the mode of action of R. mucilaginosa, the expression of genes that are part of major pro-inflammatory pathways was evaluated in 3D A549 cells exposed to P. aeruginosa or co-exposed to P. aeruginosa and R. mucilaginosa. R. mucilaginosa co-exposure significantly downregulated the expression of genes encoding IL-8 (IL8), IL-6 (IL6) and pro-IL-1β (IL1B) (figure 3) caused by P. aeruginosa. In addition, R. mucilaginosa significantly downregulated NFKB1 expression (figure 3). Regulation of NFKBIA, NFKBIE and REL genes showed a downward trend that further indicates an effect of R. mucilaginosa on the NF-κB pathway.
Differential expression of genes involved in inflammation by three-dimensional (3D) A549 alveolar epithelial cells exposed to Pseudomonas aeruginosa versus P. aeruginosa in combination with Rothia mucilaginosa. Quantitative PCR analysis showing average fold changes in mRNA levels of 3D A549 cells stimulated by P. aeruginosa PAO1 versus a co-culture of P. aeruginosa and R. mucilaginosa. n=3. *: statistically significant (p<0.05) fold change.
The influence of R. mucilaginosa on activation of the NF-κB pathway was confirmed in a NF-κB–luciferase reporter 3D A549 cell model, showing a significant reduction of NF-κB activity by co-culturing R. mucilaginosa with P. aeruginosa compared with P. aeruginosa single exposure (figure 4a). The inhibition of the NF-κB pathway by R. mucilaginosa was also shown using LPS as a pro-inflammatory stimulus and was observed for 36 isolates representing five different Rothia species from clinical and environmental sources (supplementary figure S4a and b). The anti-inflammatory effect of different Rothia species on NF-κB pathway activation by LPS was also observed under microaerobic conditions (supplementary figure S4c), which can be encountered by microorganisms in the mucus of patients with chronic lung disease such as CF [53]. The minimal effective dose of R. mucilaginosa to inhibit P. aeruginosa-induced inflammation was an MOI of at least 5:1 for all doses of P. aeruginosa tested (up to MOI 100:1) (supplementary figure S5a). We also tested higher doses of R. mucilaginosa (MOI 100:1 and 1000:1) and the anti-inflammatory effect was confirmed (supplementary figure S5b).
Effect of Rothia mucilaginosa (Rm) on NF-κB pathway activation by Pseudomonas aeruginosa (Pa). a) Activation of the NF-κB pathway measured via luminescence of three-dimensional (3D) NF-κB reporter A549 cells. 3D cells were exposed for 4 h to P. aeruginosa PAO1 alone or in co-culture with R. mucilaginosa DSM20746. b) Semiquantitative determination (by Western blotting) of proteins (i.e. A20, IκBα, p65 and phosphorylated (p)-IκBα) produced by 3D A549 cells stimulated with P. aeruginosa PAO1 with or without R. mucilaginosa DSM20746 for 15 min (15′), 30 min (30′), 1 h and 4 h. c) Band intensity (normalised to β-actin) of Western blot at 4 h of 3D A549 cells stimulated with P. aeruginosa PAO1 with or without R. mucilaginosa DSM20746. RLU: relative light units; NC: negative control (uninfected 3D NF-κB reporter A549 cells in serum-free GTSF-2 medium). Data represent mean±sem, n=3. *: p<0.05; **: p<0.01; ***: p<0.001.
We subsequently performed Western blot analysis to investigate at which stage R. mucilaginosa affects NF-κB signalling. After 4 h co-culture of P. aeruginosa and R. mucilaginosa, R. mucilaginosa prevented P. aeruginosa-induced IκBα phosphorylation (figure 4b and c). This suggests that R. mucilaginosa affects NF-κB signalling upstream or at the level of the IκBα kinase complex (IKK). In line with this, the expression of both IκBα and A20, two NF-κB target genes, was greatly reduced in the co-culture samples (figure 4c).
Finally, we excluded degradation of IL-8 or inhibition of P. aeruginosa adhesion to 3D A549 epithelial cells by R. mucilaginosa as alternative explanations for the observed anti-inflammatory effect (supplementary figure S6).
R. mucilaginosa supernatant inhibits IL-8 production and NF-κB pathway activation in response to pro-inflammatory stimuli in 3D lung epithelial cells
Exposure of 3D A549 cells to P. aeruginosa PAO1 (MOI 10:1) in the presence of R. mucilaginosa cell-free supernatant, but not lysed cells, strongly reduced IL-8 production (figure 5a). We confirmed that R. mucilaginosa DSM20746 cell-free supernatant fully abolished LPS-stimulated NF-κB pathway activation (figure 5b).
The anti-inflammatory effect of Rothia mucilaginosa (Rm) is mediated by its cell-free supernatant. a) Interleukin (IL)-8 production by three-dimensional (3D) A549 epithelial cells after exposure to P. aeruginosa PAO1 (Pa) for 4 h, with or without R. mucilaginosa DSM20746 live cells, lysed cells or cell-free supernatant. b) Quantification of NF-κB pathway activation after 4 h exposure to lipopolysaccharide (LPS) (100 μg·mL−1) with or without R. mucilaginosa DSM20746 cell-free supernatant. The multiplicity of infection was 10:1 in all experiments. RLU: relative light units; NC: negative control (uninfected 3D epithelial cells). Data represent mean±sem, n≥3. ***: p<0.001.
Inverse relationship between airway inflammation and Rothia species in patients with neutrophilic airways disease
Two culture-independent approaches were applied to measure Rothia species in induced sputum samples from patients with bronchiectasis, a neutrophilic airways disease. The first approach (measuring R. mucilaginosa absolute load) indicated that R. mucilaginosa load significantly inversely correlated with levels of IL-8 (figure 6a) and IL-1β (figure 6b), as well as with those of MMP-1 (figure 6c) and MMP-8 (figure 6d). There were no significant correlations between R. mucilaginosa and MMP-9 (figure 6e) or neutrophil percentage in sputum (figure 6f). In the second approach, we determined the relative abundance of Rothia and found it also inversely correlated with levels of IL-8, IL-1β, MMP-1 and MMP-8. The correlation between Rothia relative abundance and MMP-9 was also significant. There was again no correlation between Rothia abundance and neutrophil percentage (supplementary figure S7).
Correlation of absolute load of Rothia mucilaginosa with pro-inflammatory parameters in induced sputum samples from bronchiectasis patients: a) interleukin (IL)-8, b) IL-1β, c) matrix metalloproteinase (MMP)-1, d) MMP-8, e) MMP-9 and f) neutrophils. Data points represent individual induced sputum samples. Correlation coefficients and p-values were calculated on ranked values when Spearman's test was used (rs).
Discussion
We report that common members of the lung microbiota belonging to the genus Rothia exert anti-inflammatory properties in vitro and that for at least one (R. mucilaginosa) this effect is also apparent in vivo. Being frequently part of the oral microbiota and only rarely causing lung infection [54, 55], R. mucilaginosa is often considered part of salivary contamination of samples from the lower airways. However, comparison of the microbiome of the oral cavity with the lung microbiome suggests that R. mucilaginosa is not a contaminant [37, 56, 57]. The core lung microbiome varies interindividually, yet Rothia species are detected in COPD and asthmatic patients, in paediatric and adult individuals with CF [23, 33, 35, 36, 58], and also in the lower airways of healthy individuals [59]. Nevertheless, little is known about its role in chronic lung disease [24, 60].
The presence of bacteria in the lower airways in chronic lung disease is conventionally considered as a contributor to the pathophysiology [13–15]. Results obtained in the present study strongly suggest that the presence of R. mucilaginosa in the airways is beneficial as it not only inhibited the production of P. aeruginosa-induced pro-inflammatory cytokines, it also showed an inhibitory effect on the IL-8 response induced by LPS and by another important respiratory pathogen, S. aureus. The anti-inflammatory effect was confirmed for a large collection of Rothia species and was also observed under microaerobic conditions. Since Rothia species are facultative anaerobes, it is likely that these microorganisms will colonise niches with low oxygen levels found in the lung mucus of patients with chronic lung disease [53]. Nevertheless, a minimal dose of Rothia was needed to obtain an anti-inflammatory effect in vitro for all doses of P. aeruginosa tested, suggesting a dose-dependent effect. The dose used most frequently throughout our study (MOI 10:1) is justified as a model for patients with chronic lung disease since a total bacterial MOI of ∼100:1 has been reported for CF patients [61], and the relative abundance of Rothia varies between 1% and 40% [62, 63]. Also, in the BLESS cohort that was analysed in our study, the mean relative abundance of Rothia was 9%, supporting the physiological relevance of the dose used. Nevertheless, we cannot rule out that at some point, high loads of Rothia may result in a net pro-inflammatory effect in the lung environment, regardless of its anti-inflammatory properties. Very high doses of R. mucilaginosa (up to MOI 1000:1) still exerted an anti-inflammatory effect in vitro; therefore, further research is needed to evaluate the load of Rothia that can be tolerated in vivo without causing adverse effects.
R. mucilaginosa was also able to reduce the LPS-triggered production of pro-inflammatory cytokines in lung homogenates of mice, using an alginate bead model. The levels of several cytokines in mice exposed to empty alginate beads were similar to those of LPS-treated mice, which is in agreement with a previous study reporting inflammatory responses in mouse lungs after instillation with sterile alginate beads [64]. Nevertheless, R. mucilaginosa caused a clear decline in the inflammatory response of lung tissue regardless of the stimulus and the degree of the anti-inflammatory effect was strain dependent.
Interestingly, previous studies have shown a negative correlation between the relative abundance of Rothia species and sputum inflammatory markers such as neutrophil elastase, IL-8 and IL-1β in CF patients [65, 66]. A recent report found the relative abundance of the genus Rothia in expectorated sputum to be a positive predictor of lung function in patients with CF [67]. Not only were we able to confirm a negative correlation between Rothia species and inflammation for the BLESS cohort at the relative abundance level, but we also showed that the absolute abundance of R. mucilaginosa was negatively correlated with inflammatory markers (based on IL-8 and IL-1β levels) and mediators of airway remodelling (MMP-1 and MMP-8 levels). Levels of MMP-1, MMP-8 and MMP-9, which degrade extracellular matrix proteins [68], are increased in the airways of patients with neutrophilic airways disease, are inversely correlated with lung function [51] and contribute to irreversible airway damage [68]. We did not find an association between Rothia and neutrophil abundance. This is potentially attributable to relatively low levels of variation in neutrophil percentages between patients (all ≥70%). Assessment of Rothia abundance in other cohorts of neutrophilic airways disease is needed to confirm a possible anti-inflammatory effect of Rothia species in patients and to determine the extent to which this might impair the progression of structural lung disease.
While several studies have associated an increased abundance of Rothia with positive health outcomes, it was recently demonstrated that the relative abundance of Rothia species was higher in a cohort of patients with a range of chronic lung diseases (severe asthma, COPD and bronchiectasis) [69] and in patients with tuberculosis compared with healthy individuals [70]. It remains to be determined whether the increased Rothia relative abundance also implies an actual change in the absolute microbial load of that species or merely reflects a decreased abundance of other (more dominant) species. Nevertheless, the study of Hong et al. [70] indicated that R. mucilaginosa may serve as an anchor species for Mycobacterium tuberculosis and 20 other taxa. It is possible that metabolic cross-feeding could alter the co-occurrence of other pathogenic or nonpathogenic species in the lung environment. For R. mucilaginosa, it was demonstrated that its metabolic products can cross-feed P. aeruginosa [71]. While we did not observe R. mucilaginosa-induced growth of P. aeruginosa in our study, long-term interspecies interactions are worth exploring further to evaluate whether commensal species such as R. mucilaginosa may influence overall community composition, in addition to exerting an anti-inflammatory activity. Furthermore, whether an altered Rothia abundance in the described and present studies is a driver of disease/disease markers or a consequence of an altered lung physiology (e.g. due to altered airway clearance or differential levels of reactive oxygen species, pro-inflammatory cytokines or mucins) remains a question to be fully addressed. The obtained data in animals suggest that supplementation of Rothia in the lungs may diminish the pro-inflammatory response to external stimuli, yet further exploration of the causal relation between Rothia abundance and inflammation in patients is needed.
The discovery of Rothia as an anti-inflammatory genus in the lung microbiota raises the question whether other nonpathogenic commensals in the respiratory tract, in particular originating from the oral cavity, may have immunomodulatory properties. Recent studies in animals [72] or in vitro [73] support this hypothesis. For example, pulmonary aspiration of oral human commensals (Prevotella melaninogenica, Veillonella parvula and Streptococcus mitis) in mice induces a T-helper cell type 17 inflammatory phenotype that diminishes susceptibility to Streptococcus pneumoniae. Our data on cytokine production and on NF-κB activation indicate that R. mucilaginosa prevents NF-κB signalling, probably at the level of IKK activation. Indeed, IL-8, IL-1β, IL-6 and tumour necrosis factor-α are regulated by NF-κB through binding of pathogen-associated molecular patterns to Toll-like receptors (TLRs) [74]. This is of particular interest since the NF-κB pathway mediates the elevated innate immune response in patients with chronic lung diseases [75–84]. Cell-free supernatant exhibited anti-inflammatory activity, indicating that R. mucilaginosa secretes (an) anti-inflammatory mediator(s) that directly interferes with the NF-κB pathway or induces host mediators that resolve inflammation. Since the anti-inflammatory activity was observed using both the pure TLR4 ligand LPS as well as different microbial species as the pro-inflammatory trigger, we can exclude that antimicrobial activity of R. mucilaginosa is causing the effect. Research to identify the immunomodulatory mediator(s) will be necessary to further understand how Rothia species modulate the pro-inflammatory response.
Understanding the contribution of potentially beneficial species to net inflammation could enable us to better understand variation in disease severity and progression between patients with chronic lung disease. Prospective studies to assess the clinical utility of R. mucilaginosa as a marker of airway inflammation, and the extent to which it may provide a basis for the development of novel therapeutic interventions, are now required.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material ERJ-01293-2021.Supplement
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-01293-2021.Shareable
Acknowledgements
We thank Catherine Greene (Respiratory Medicine, Royal College of Surgeons in Ireland, Dublin, Ireland) for providing the CF cell lines, Matthew Parsek (Dept of Microbiology, University of Washington, Seattle, WA, USA) and Boo Tseng (School of Life Sciences, University of Nevada Las Vegas, Las Vegas, NV, USA) for providing P. aeruginosa CF127, and Joana Doci (Laboratory of Pharmaceutical Microbiology, Ghent University, Ghent, Belgium) for help with the experiments.
Footnotes
Author contributions: A. Crabbé conceptualised the study. A. Crabbé, C. Rigauts and T. Coenye are responsible for the overall study design. J. Aizawa, C. Rigauts, M. Govaerts and L. Ostyn performed the animal experiments. S.L. Taylor and S. Sims performed DNA sequencing, qPCR experiments and data analysis. C. Rigauts and L. Ostyn performed the in vitro experiments, and C. Rigauts analysed the data. L. Burr and J.L. Simpson were responsible for clinical sampling. G.B. Rogers, P. Cos, Y. Dondelinger, M. Sze, L. Vereecke, A. Bragonzi, C. Cigana, H. Van Acker and E. Vandeplassche provided critical guidance in the experimental set-up, analysis and/or interpretation of results. C. Rigauts, A. Willems, M.M. Tunney, G.B. Rogers and S.L. Taylor isolated bacterial strains used in this study. C. Rigauts, A. Crabbé, G.B. Rogers, S.L. Taylor and J. Aizawa wrote the manuscript. T. Coenye, P. Cos, M.M. Tunney, A. Willems, Y. Dondelinger, L. Vereecke and C. Cigana revised the manuscript.
Conflict of interest: C. Rigauts declares a patent application related to this work.
Conflict of interest: J. Aizawa has nothing to disclose.
Conflict of interest: S.L. Taylor has nothing to disclose.
Conflict of interest: G.B. Rogers has nothing to disclose.
Conflict of interest: M. Govaerts has nothing to disclose.
Conflict of interest: P. Cos has nothing to disclose.
Conflict of interest: L. Ostyn has nothing to disclose.
Conflict of interest: S. Sims has nothing to disclose.
Conflict of interest: E. Vandeplassche has nothing to disclose.
Conflict of interest: M. Sze has nothing to disclose.
Conflict of interest: Y. Dondelinger has nothing to disclose.
Conflict of interest: L. Vereecke has nothing to disclose.
Conflict of interest: H. Van Acker has nothing to disclose.
Conflict of interest: J.L. Simpson has nothing to disclose.
Conflict of interest: L. Burr has nothing to disclose.
Conflict of interest: A. Willems has nothing to disclose.
Conflict of interest: M.M. Tunney has nothing to disclose.
Conflict of interest: C. Cigana has nothing to disclose.
Conflict of interest: A. Bragonzi has nothing to disclose.
Conflict of interest: T. Coenye declares a patent application related to this work.
Conflict of interest: A. Crabbé declares a patent application related to this work.
Support statement: This work was funded by the Research Foundation Flanders (FWO). C. Rigauts is a recipient of an FWO-Strategic Basic Research (SB) fellowship (1S23117N). A. Crabbé is a recipient of an FWO Odysseus grant (G.0.E53.14N), and P. Cos, T. Coenye and A. Crabbé received an FWO research grant (G010119N). J. Aizawa received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant (722467 Print-Aid consortium). G.B. Rogers is supported by a Matthew Flinders Research Fellowship and a NHMRC Senior Research Fellowship. Y. Dondelinger is supported by a postdoctoral fellowship from the FWO (12T9118N). Funding information for this article has been deposited with the Crossref Funder Registry.
- Received May 5, 2021.
- Accepted September 10, 2021.
- Copyright ©The authors 2022.
This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org
References
