---
title: "Airway microbiome–metabolome signature linked to pediatric severe asthma"
id: "plos-one-11-a-microbiome-metabolome-signature-associated-with-pediatric-severe-asthma"
canonical_url: "https://medichelpline.com/clinical-feed/plos-one-11-a-microbiome-metabolome-signature-associated-with-pediatric-severe-asthma"
content_type: "clinical_feed_article"
specialty: "Pediatrics"
source_name: "PLOS ONE (Medicine)"
source_url: "https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0358560"
published_at: "2026-09-18T14:00:00.000Z"
evidence_level: "Journal Feed"
license: "CC-BY-NC-4.0 / Informational Use"
---
# Airway microbiome–metabolome signature linked to pediatric severe asthma
## Provenance & Clinical Metadata
- **Canonical URL:** https://medichelpline.com/clinical-feed/plos-one-11-a-microbiome-metabolome-signature-associated-with-pediatric-severe-asthma
- **Specialty:** [Pediatrics](https://medichelpline.com/clinical-feed/pediatrics.md)
- **Primary Source:** PLOS ONE (Medicine)
- **Source URL:** [Original Journal Publication](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0358560)
- **Published At:** 2026-09-18T14:00:00.000Z
- **Evidence Rating:** Journal Feed
## Executive GIST (TL;DR)
- This study used paired bronchoalveolar lavage (BAL) **microbiome** and **metabolome** profiling to identify local signatures associated with pediatric severe asthma. - Samples came from 20 children with severe asthma and 10 non-asthmatic disease-controls; analyses included 16S rRNA gene amplicon sequencing and LC-HRMS metabolomics. - BALs from children with severe asthma demonstrated increased alpha-diversity and higher relative abundances of several taxa, including **Actinobacteriota**, **Streptococcus**, **Moraxella**, **Corynebacterium**, **Tropheryma**, and **Treponema** compared with controls. - Untargeted metabolomics showed perturbation of the polyamine pathway in severe asthma: reduced arginine and increased **spermine** and **spermidine** levels in BAL fluid. - Integrated multi-omics analyses found significant associations between **Streptococcus** abundance and elevated spermine and spermidine levels, indicating microbial–metabolic links. - Each dataset (microbiome or metabolome) could discriminate clinical phenotypes within severe asthma, including exacerbation frequency and co-occurring atopic dermatitis. - Unsupervised clustering of microbiome profiles revealed four distinct clusters that may correspond to different severe asthma endotypes. - Authors conclude that an airway-level **microbiome–metabolome** signature characterizes pediatric severe asthma and that multi-omics airway profiling can inform mechanistic understanding and future targeted therapies. - Data and raw datasets are publicly available via cited accession numbers and a prior data paper; funding and conflicts of interest were disclosed in the article.
## Clinical Analysis & Structured Key Points
A microbiome–metabolome signature associated with pediatric severe asthma | PLOS One Browse Subject Areas ? Click through the PLOS taxonomy to find articles in your field. For more information about PLOS Subject Areas, click here . Article Authors Metrics Comments Media Coverage Reader Comments Figures Figures Abstract Background Severe asthma is a heterogeneous condition encompassing multiple phenotypes. Understanding lung-specific mechanisms in children with severe asthma may enable the development of more precise therapeutic strategies. We previously reported that immune components in bronchoalveolar lavages (BALs) differentiate children with severe asthma from non-asthmatic disease-controls and, frequent from non-frequent exacerbators, among children with severe asthma. Objective To identify a local signature of severe asthma using complementary multi-omics analyses of BALs. A secondary objective was to evaluate whether bacterial taxa and metabolites discriminate severe asthma subtypes associated with distinct phenotypes or endotypes. Methods BAL microbiome and metabolome were investigated in 20 children with severe asthma and 10 non-asthmatic children using 16S rRNA gene amplicon sequencing and liquid chromatography coupled to high-resolution mass spectrometry (LC-HRMS), respectively. Data were analysed separately and through integrative multi-omics approaches. Results Compared with controls, BALs from children with severe asthma showed increased alpha-diversity, higher relative abundances of Actinobacteriota , Streptococcus , Moraxella , Corynebacterium , Tropheryma , and Treponema , and an altered polyamine pathway characterized by reduced arginine and increased spermine and spermidine levels. Integrated analyses revealed significant associations between Streptococcus and both spermine and spermidine. Independently, each dataset discriminated severe asthma phenotypes, notably exacerbation frequency and co-occurring atopic dermatitis. Unsupervised clustering of microbiome profiles identified four distinct clusters that may reflect severe asthma endotypes. Conclusions This study identifies a distinct airway microbiome–metabolome signature associated with pediatric severe asthma. Enrichment of specific bacterial taxa, particularly Streptococcus , together with altered polyamine metabolic pathway, highlights microbial–metabolic interactions potentially involved in disease pathophysiology. The ability of microbiome and metabolome profiles to independently and jointly discriminate clinical phenotypes underscores the relevance of multi-omics approaches for diagnosis and follow-up of severe asthma. Our findings support the importance of airway-level profiling to improve mechanistic understanding of severe asthma and inform on future targeted therapeutic strategies. Citation: Briard M, Guillon B, Venot E, Grauso M, Hennequet-Antier C, Bruneau A, et al. (2026) A microbiome–metabolome signature associated with pediatric severe asthma. PLoS One 21(9): e0358560. https://doi.org/10.1371/journal.pone.0358560 Editor: Farah Al-Marzooq, UAE University: United Arab Emirates University, UNITED ARAB EMIRATES Received: April 24, 2026; Accepted: September 2, 2026; Published: September 18, 2026 Copyright: © 2026 Briard et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability: The microbiota and metabolite datasets underlying this study have been published as a data paper (accession number: 10.1016/j.dib.2025.112359 ), which provides full details of BAL fluid sampling, measurements, and data pre-treatment and curation. The datasets generated and analyzed during the current study are available from the data.inrae.fr database (metabolites accession number: 10.57745/1L8VRI ; microbiota accession number: 10.57745/LL3TFW ). All other data are within the manuscript and its Supporting Information files. Funding: This work was supported by a grant from the French National Research Agency ANR (SevAsthma-children, grant no. ANR-18-CE14-0011-01, to MLM, GL, KAP and MT). The funding institution had no role in the design of the study; in the collection, analyses of data; in the writing of the article, or in the decision to publish the results. Competing interests: The authors have declared that no competing interests exist. Introduction Asthma is a chronic inflammatory airway disease marked by hyperresponsiveness to irritants, allergens, or viral infections (Global Initiative for Asthma. Global strategy for asthma management and prevention. www.ginasthma.org (2022)). Traditionally, asthma has been classified into T2-high (eosinophilic) and T2-low (non-eosinophilic) endotypes [ 1 , 2 ]. More recent studies, however, reveal involvement of additional pathways, including T1, T2, and T17 responses, underscoring immune complexity beyond a simple T2 paradigm [ 3 – 5 ]. In some individuals, asthma remains uncontrolled despite high-dose inhaled corticosteroids (ICS), additional controllers, correct medication use, and elimination of modifiable risk factors such as tobacco exposure, defining severe asthma (SA) [ 6 ]. SA is clinically heterogeneous and accounts for a disproportionate share of morbidity, healthcare use, and impaired quality of life [ 7 – 10 ]. Although biologics have emerged as promising therapies for SA, their optimal use requires understanding SA phenotypes (clinical traits), each of which results from different endotypes (underlying molecular mechanisms) [ 1 , 11 ]. Because many patients remain unresponsive and local–systemic signatures correlate poorly [ 3 ], deeper tissue-level characterization and biomarker discovery are needed. Contrary to long-held assumptions, the lower respiratory tract harbours a dynamic microbial community that interacts with the host epithelial barrier and immune system. The lower airway microbiome influences immune homeostasis and airway inflammation [ 12 ] and differs from that of the upper airway [ 13 ]. An imbalance between symbiotic and pathogenic bacterial strains in the lung may lead to altered immune development and inappropriate inflammatory responses [ 14 , 15 ]. Indeed, altered lung microbiota composition has been associated with susceptibility to and/or severity of childhood respiratory diseases, including asthma [ 16 – 20 ]. Studies show that the bronchial and sputum microbiomes differ in asthmatic versus healthy adults, with higher abundance of Actinobacteriota , Haemophilus , Moraxella , Staphylococcus , and Streptococcus [ 21 – 24 ]. In contrast, data on the lower airway microbiota in children with asthma remain scarce; most pediatric studies have focused on the upper airway, though these studies suggest comparable microbial shifts [ 17 – 20 ]. One study investigated the upper and lower airway microbiomes in children with SA and found significantly different communities between the two locations, and a negative association between Actinomyces and transcripts related to inflammation [ 25 ]. Metabolites in bronchoalveolar lavage (BAL) fluid reflect the combined biochemical activities of microbes and host cells, providing more direct insight into airway physiology, inflammation and remodeling than microbial taxonomy alone [ 26 , 27 ]. Advances in untargeted and targeted metabolomic approaches, including liquid chromatography coupled to high-resolution mass spectrometry (LC–HRMS), have enabled detection of lipid mediators, amino acids, and microbial by-products in BALs, highlighting pathways central to immune responses and airway physiology [ 28 ]. To date, pediatric BAL metabolomics, particularly in children with SA, remains rare, limiting our understanding of how host–microbe interactions manifest at the functional level in diseased lung tissue. Indeed, attempts to identify biomarkers of childhood asthma using metabolomics, have focused primarily on blood, urine, and exhaled breath condensate samples [ 29 ]. We found only two studies investigating the BAL metabolome in childhood asthma [ 28 , 30 ], of which one did not compare its data to non-asthmatic controls [ 30 ]. The other study indicated that children with persistent wheezing, but not SA, had higher abundances of choline, oleamide, butyrylcarnitine, palmitoylethanolamide, and various phosphatidylcholines, suggesting alterations in local metabolism [ 28 ]. Recently, we showed that a large set of immune components in BAL fluids differentiated children with severe asthma (SA) from disease-control subjects [ 3 ], and that distinct immune signatures further discriminated frequent from non-frequent exacerbators among children with SA [ 31 ]. Given the scarcity of studies of the BAL microbiome and metabolome in pediatric SA, we hypothesized that integrative microbiome–metabolome profiling would provide new insights into the pathophysiology of pediatric SA. To test this hypothesis, we used 16S rRNA gene amplicon sequencing and untargeted LC-HRMS metabolomics to identify microbial taxa and metabolites, respectively, associated with SA, and exacerbation frequency or atopic dermatitis (AD). We then applied the Data Integration Analysis for Biomarker discovery using Latent cOmponents (DIABLO [ 32 ]) method to investigate the link between taxa and metabolites. Unsupervised clustering of these datasets further revealed SA subgroups, which may correspond to distinct SA endotypes. This highlights the need for validation in larger patient cohorts and for assessing their association with individual trajectories and response to treatment. Materials and Methods Study population and sample processing Twenty school-aged children with severe asthma (SA) were included. Patients were mainly from the Ile-de-France region, and followed in the department of paediatric pulmonology and allergy at Necker Hospital (Paris, France). Ten children, with chronic respiratory disorders unrelated to asthma who required endoscopy, were also recruited as age-matched disease-control subjects (hereafter referred to as non-asthmatic, NA). NA subjects had chronic respiratory disorders, including ciliary dyskinesia (n = 2), viral pulmonary sequelae (n = 3), or non-cystic fibrosis bronchiectasis (n = 5). Clinical data, patient characteristics and ethical statements are described in our previous studies [ 3 , 31 , 33 ] and summarized in S1 Table . Patients included in this study were from the CLASSE (Cellules Lymphoïdes Innées dans l’ASthme Sévère de l’Enfant) cohort. They were recruited in the department of Paediatric Pulmonology and Allergy at Necker Hospital (Paris, France) between 05/07/2016 and 14/06/2018. Written and oral information about the study was provided to patients and their parents/guardians, and institutional ethical approval (ref 2016-03-07 RNI) and written informed consent from parents/guardians were obtained. The BAL microbiota and metabolite datasets underlying this study have been published as a data paper [ 34 ], which provides full details of BAL fluid sampling, measurements, and data pre-treatment and curation. The datasets generated and analyzed in the current study are available at data.inrae.fr (metabolites [ 35 ]; microbiota: [ 36 ]). Here, we summarize the key aspects relevant to the present analysis. Microbiota analyses DNA extraction from BAL fluids dry pellets was performed using the QIAamp Power Fecal DNA kit and following the manufacturer’s instructions (Qiagen; Courtaboeuf, France). Extraction blanks were processed alongside all BAL samples by performing the complete DNA extraction protocol on tubes containing no biological material. These blanks served as negative controls to monitor for potential contamination introduced during extraction or laboratory handling, or by reagents. BAL bacterial load was measured by real-time qPCR on BAL DNA samples diluted 1/10 using the BACT1369F forward primer (5’-CGGTGAATACGTTCCCGG-3’), the PROK1492R reverse primer (5’-GGCTACCTTGTTACGACTT-3’) and the TM1389F probe (5’-6-FAM-CTTGTACACACCGCCCGTC-3’), as previously described [ 37 , 38 ] (amplification efficiency 93%). Data processing of bacterial 16S rRNA gene amplicon sequences and statistical analyses were performed in RStudio (R version 4.4.1). Raw ASV counts, taxonomy, and sample metadata were imported into phyloseq (version 1.50.0). To reduce sparsity and low-abundance noise, ASVs detected in fewer than 10% of samples and with relative abundance below 0.01% were excluded from downstream analyses. For diversity and ordination, feature tables were total-sum scaled (TSS) to relative abundance. Alpha diversity indices (observed richness, Chao1, Shannon, inverse Simpson) were computed with estimate_richness and group comparisons used the Wilcoxon rank-sum test. Beta diversity was assessed using Bray–Curtis dissimilarities calculated from total-sum scaled ASV counts. Ordinations were performed by principal coordinates analysis (PCoA) using the ordinate function in the phyloseq R package. Group differences in microbial community composition were tested using permutational multivariate analysis of variance (PERMANOVA; adonis2 function in vegan v2.7-2) with 999 permutations. Homogeneity of group dispersions was evaluated prior to PERMANOVA using the betadisper and permutest functions (vegan). Graphical representations of ordinations were generated in R with ggplot2, following the code structure and workflow described by van Beveren et al. [ 39 ]. Differential abundance analyses were performed using multiple complementary methods. For compositional approaches, ALDEx2 (v1.38.0; centered log-ratio transformation denom = “all”, 250 Monte Carlo instances, Welch’s t-test, effect sizes with 95% CI) and ANCOM-BC2 (v2.8.0; fixed-effects model fix_formula=”asthma”, prv_cut = 0.1, lib_cut = 1000, structural zeros struc_zero = TRUE, neg_lb = TRUE, global test global = TRUE, EM control tol = 1e-5, max_iter = 100, BH-adjusted p-values, alpha = 0.05) were applied. Count-based methods included DESeq2 (v1.46.0; median-of-ratios size-factor normalization, Wald test, parametric dispersion fit, log-fold change shrinkage via apeglm) and edgeR (v4.4.0; TMM normalization, negative-binomial GLM, quasi-likelihood F-test). Zero-inflated modelling was performed with metagenomeSeq (v1.43.0; CSS normalization percentile = 0.5, zero-inflated Gaussian model fitZig, BH-adjusted p-values). Multivariable linear modelling was performed with MaAsLin2 (v1.15.1; prevalence ≥10% min_prevalence = 0.1, no minimum abundance filter, log transform, linear model with fixed effects for asthma). Finally, LinDA (MicrobiomeStat v1.2; CLR transform, pseudo-count = 0.5, prevalence filter 10% prev.filter = 0.1, outlier trimming is.winsor = TRUE, outlier.pct = 0.03, adaptive bias correction adaptive = TRUE, BH-adjusted p-values, alpha = 0.05) was applied. Multiple testing correction was performed with the Benjamini–Hochberg method, and amplicon sequence variants were considered differentially abundant when the adjusted p-values were less than 0.05. To balance sensitivity and false discovery risk in exploratory analyses, we used an adjusted p-value cutoff of 0.20 (rather than the conventional 0.05) in the initial candidate selection step. This liberal threshold was deliberately chosen to maximize detection of potential signals across diverse statistical frameworks (ALDEx2, ANCOM-BC2, DESeq2, edgeR, MaAsLin2, MetagenomeSeq, and LinDA), each of which may capture distinct aspects of the data due to differences in assumptions, normalization, or modeling approaches. By requiring consensus (taxon adjusted p-value 1 and a nominal p-value 1 for metabolites), using block.splsda() with two latent componen
## Related Clinical Research

- [OM-85 bacterial lysate for primary prevention of preschool wheezing: randomized placebo-controlled](https://medichelpline.com/clinical-feed/pubmed-42705251.md) (DOI: 10.1016/S0140-6736(26)01443-1)
- [Age-Dependent Differences in Disease Risk for People with Down Syndrome](https://medichelpline.com/clinical-feed/medrxiv-14-effects-of-age-on-the-association-between-down-syndrome-and-common-disease.md)
- [BMI Trajectories From Birth to Adolescence and Their Association With Asthma in the Leicester Coho](https://medichelpline.com/clinical-feed/medrxiv-18-body-mass-index-trajectories-from-birth-to-adolescence-and-their-association.md)
- [DOHaD Reframed: Socioeconomic Drivers Exceed Maternal Behaviors for Child Health](https://medichelpline.com/clinical-feed/plos-medicine-0-expanding-the-developmental-origins-of-health-and-disease-beyond-maternal.md)
- [Retinal changes on OCT may predict atrial fibrillation risk years before diagnosis](https://medichelpline.com/clinical-feed/medical-news-today-0-how-changes-in-the-eyes-may-help-spot-afib-risk-years-early.md)

## Navigation
- [← Back to Pediatrics Feed](https://medichelpline.com/clinical-feed/pediatrics.md)
- [← All Clinical Specialties](https://medichelpline.com/clinical-feed.md)
## Medical & Regulatory Disclaimer

> [!CAUTION]
> MedicHelpline content is structured for research, educational, and professional discovery purposes. It does not constitute individual medical advice, clinical diagnosis, or treatment recommendations.
> Always verify dosing, contraindications, and regulatory alerts against official product labeling and primary regulatory sources before clinical decision-making.