Microbial Inflammatory Networks in Bronchiectasis Exacerbators With Pseudomonas aeruginosa

Chronic airway infection with Pseudomonas aeruginosa is a well-known risk factor for adverse outcomes in bronchiectasis.1Finch S. McDonnell M.J. Abo-Leyah H. Aliberti S. Chalmers J.D. A comprehensive analysis of the impact of pseudomonas aeruginosa colonization on prognosis in adult bronchiectasis.Ann Am Thorac Soc. 2015; 12: 1602-1611PubMed Google Scholar Among patients with bronchiectasis and chronic P aeruginosa infection, frequent exacerbators show higher mortality rates than nonfrequent exacerbators.2Araújo D. Shteinberg M. Aliberti S. et al.The independent contribution of Pseudomonas aeruginosa; infection to long-term clinical outcomes in bronchiectasis.Eur Respir J. 2018; 511701953Crossref Scopus (128) Google Scholar However, the presence of chronic P aeruginosa infection per se does not adequately explain the link to exacerbation frequency. Data on the use of “integrative microbiomics” in bronchiectasis demonstrate a specific Pseudomonas interactome in the prediction of exacerbation risk.3Mac Aogáin M. Narayana J.K. Tiew P.Y. et al.Integrative microbiomics in bronchiectasis exacerbations.Nat Med. 2021; 27: 688-699Crossref PubMed Scopus (61) Google Scholar The influence of airway inflammation has not been considered in this model; therefore, an integrated microbiomic inflammation approach is warranted.3Mac Aogáin M. Narayana J.K. Tiew P.Y. et al.Integrative microbiomics in bronchiectasis exacerbations.Nat Med. 2021; 27: 688-699Crossref PubMed Scopus (61) Google Scholar We hypothesize that, among patients with bronchiectasis with chronic P aeruginosa infection, a specific interaction network between Pseudomonas and its associated neutrophilic inflammation may contribute additionally to an explanation of the different exacerbation frequencies observed.4Oriano M. Gramegna A. Terranova L. et al.Sputum neutrophil elastase associates with microbiota and P. aeruginosa in bronchiectasis.Eur Respir J. 2020; 562000769Crossref PubMed Scopus (30) Google Scholar A pilot observational cross-sectional study was performed at the Policlinico Hospital, Milan, Italy, between 2017 and 2019. Adults (aged ≥ 18 years) with non-cystic fibrosis bronchiectasis and with chronic P aeruginosa infection were enrolled prospectively and consecutively. All patients provided samples before starting either chronic macrolides or inhaled antibiotics. Chronic Pseudomonas infection was defined based on a past clinical history of isolation of P aeruginosa in sputum culture on two or more occasions at least 3 months apart over a 1-year period.5Pasteur M.C. Helliwell S.M. Houghton S.J. et al.An investigation into causative factors in patients with bronchiectasis.Am J Respir Crit Care Med. 2000; 162: 1277-1284Crossref PubMed Scopus (544) Google Scholar All patients also had positive polymerase chain reaction detection of P aeruginosa in sputum. The study was approved by the local ethics committee (#255_2020), and all subjects provided written informed consent. Patients were enrolled during clinical stability with no history of antibiotic treatment within 4 weeks before the enrollment; sputum samples were taken from a single timepoint. Patients underwent clinical, radiologic, microbiologic, and functional evaluation and provided spontaneously expectorated sputum. Mucous plugs underwent DNA extraction for microbiota analysis,4Oriano M. Gramegna A. Terranova L. et al.Sputum neutrophil elastase associates with microbiota and P. aeruginosa in bronchiectasis.Eur Respir J. 2020; 562000769Crossref PubMed Scopus (30) Google Scholar and sputum supernatant was used to quantify TNF-α, IL1β, IL6, and IL10 with the use of enzyme-linked immunosorbent assay kits (R&D Systems) and active neutrophil elastase.6Oriano M. Terranova L. Sotgiu G. et al.Evaluation of active neutrophil elastase in sputum of bronchiectasis and cystic fibrosis patients: a comparison among different techniques.Pulm Pharmacol Ther. 2019; 59: 1Crossref Scopus (11) Google Scholar Patients were divided into two groups based on the number of exacerbations in the previous year: < 3 exacerbations/year (nonfrequent exacerbators) vs ≥ 3 exacerbations/year (frequent exacerbators).7Chalmers J.D. Aliberti S. Filonenko A. et al.Characterization of the “frequent exacerbator phenotype” in bronchiectasis.Am J Respir Crit Care Med. 2018; 197: 1410-1420Crossref PubMed Scopus (177) Google Scholar Bronchiectasis exacerbation was defined according to Hill et al.8Hill A.T. Haworth C.S. Aliberti S. et al.Pulmonary exacerbation in adults with bronchiectasis: a consensus definition for clinical research.Eur Respir J. 2017; 491700051Crossref PubMed Scopus (217) Google Scholar The bioinformatic analysis of microbiota and the statistical evaluation were performed with the use of the Quantitative Insights into Microbial Ecology pipeline for analysis of microbiome data.4Oriano M. Gramegna A. Terranova L. et al.Sputum neutrophil elastase associates with microbiota and P. aeruginosa in bronchiectasis.Eur Respir J. 2020; 562000769Crossref PubMed Scopus (30) Google Scholar Sequence data are available from the National Center for Biotechnology Information Sequence Read Archive (Bioproject: PRJNA684438). Microbiome datasets were converted to relative abundances, and microbes that were present in at least 1% abundance in at least three patients were included for further analysis. Association networks were generated with the use of a generalized boosted linear models, as previously described.3Mac Aogáin M. Narayana J.K. Tiew P.Y. et al.Integrative microbiomics in bronchiectasis exacerbations.Nat Med. 2021; 27: 688-699Crossref PubMed Scopus (61) Google Scholar Normalized cytokine and microbiome data were combined as input to infer the cytokine-microbial associations. A detailed description of the methods and codes used is reported at https://github.com/Jayanth-kumar5566/inflammatory-networks-in-bronchiectasis. Among the 47 patients (70.2% female; median age, 63 years; interquartile range [IQR], 53 to 71 years) with chronic P aeruginosa infection who were enrolled in the study, 15 (31.9%) were nonfrequent and 32 (68.1%) were frequent exacerbators. No significant differences were detected between the two study groups in terms of demographics, clinical characteristics, and chronic infection with pathogens other than P aeruginosa from both a standard microbiology and a molecular biology point of view, although a higher number of patients with a percent predicted FEV1 < 35 were identified among frequent vs nonfrequent exacerbators. In addition, no differences in microbiome characteristics were detected between the study groups in terms of alpha diversity in sputum, beta diversity in sputum, and microbiota composition, as determined with the use of the DESeq algorithm (Fig 1A, 1B). Co-occurrence analysis reported a similar number of interactions of Pseudomonas with the other genera in the two study groups. Codependence among Pseudomonas and the other bacteria was characterized by a negative association and was more common in nonfrequent exacerbators compared with frequent ones. Negative associations imply that a decrease of an independent genus is associated with the increase of the dependent genera (Fig 1C, 1D). In terms of airway inflammation, a lower concentration of IL-10 (257651.6 [172936.2-314698.0] pg/mL vs 466653.5 [249492.7-658117.8] pg/mL; P = .034) and IL-1B (13823.5 [10946.5-20081.7] pg/mL vs 28992.3 [20022.2-55691.6] pg/mL; P = .006) has been detected in frequent vs nonfrequent exacerbators. The inflammatory biomarkers-microbiome network in frequent exacerbators showed a direct association between active neutrophil elastase and Pseudomonas and was characterized by interactions among inflammatory biomarkers and a low number of bacteria involved (Fig 1F). In comparison with the network of frequent exacerbators, that of nonfrequent exacerbators showed a higher number of bacteria involved and an association with different inflammatory biomarkers (Fig 1E). These biomarkers were not all linked one to the others, although IL-1B and IL-10 developed their own network with a co-exclusion of both cytokines. Our study suggests the presence of different patterns of microbiome-inflammation interactions in patients with bronchiectasis with chronic P aeruginosa infection and potentially suggests the reason why some patients frequently exacerbate while others do not. The cycle in the frequent exacerbators' network along with the complexity and codependence of relations in the nonfrequent exacerbators' network might explain one of the mechanisms that determines the exacerbation risk of these patients. The nonfrequent exacerbators' network shows a higher number of genera with codependence with Pseudomonas and is associated with the involvement of a higher number of bacteria compared with frequent exacerbators. These interactions might represent a more stable environment that is prone to adapt to perturbations. Because these interactions are derived with the use of relative abundance data (ie, only taxonomy and its abundance), the impact of “dependent” genera would mean that an increase in an “abundance” of dependent genera causes a decrease in the abundance of Pseudomonas. This analysis cannot specifically pin point whether the virulence changes but can only provide inferences on the abundance (the input dataset used). This reduction in the abundance of Pseudomonas could be due to various factors, which include cross-microbial metabolite interaction (such as inhibition or pathogenicity modulation) and competition for microbial resources (competition for an ecologic niche). Frequent exacerbators' inflammation network includes a low number of bacteria and shows a direct interaction among cytokines. These interactions might represent an unstable environment that is more prone to dysregulation. Although IL-10 and IL-1B interact in a co-exclusive way in the network, quantitative data show lower concentrations of both cytokines in frequent exacerbators. This result underlines the need to investigate inflammation as a complex network, not as the single contribution of the components. Notably, dysregulation of cytokine interaction in these patients may support recent evidence on the efficacy of long-term azithromycin use in frequent exacerbators with chronic Pseudomonas infection.9Chalmers J.D. Boersma W. Lonergan M. et al.Long-term macrolide antibiotics for the treatment of bronchiectasis in adults: an individual participant data meta-analysis.Lancet Respir Med. 2019; 7: 845-854Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar Our study is limited by its single-center and cross-sectional design, which hindered the assessment of temporal stability of these interactions, which could be an avenue of future work. We also recognize that sputum samples are not fully representative for lower airways. Other potential mechanisms behind the frequent exacerbator status, such as the contribution of viruses and fungi along with other potential triggers of exacerbation such as air pollution, have not been considered. Finally, the present experience confirms the importance of the exploration of host-pathogen interactions through integrative analysis that may detect patterns that are not appreciated by conventional microbiology and/or microbiome analysis alone.3Mac Aogáin M. Narayana J.K. Tiew P.Y. et al.Integrative microbiomics in bronchiectasis exacerbations.Nat Med. 2021; 27: 688-699Crossref PubMed Scopus (61) Google Scholar This research is supported by the Singapore Ministry of Health's National Medical Research Council under its Clinician Scientist Award (MOH-000710) (S. H. C).