Disentangling the potential roles of the human gut mycobiome and metabolites in asthma

Although previous studies reported associations between gut bacterial dysbiosis and asthma,1, 2 few addressed the gut fungal microbiota (mycobiome) and metabolome in adult asthma patients. Here, we firstly explored the gut mycobiome of 21 healthy controls (HCs) and 38 asthma patients, with 12 receiving no ICS treatment (NT) and 26 receiving inhaled corticosteroids (ICS; Table S1, Supporting information include detailed methodology). All participants were omnivores. Nine and 16 patients in NT and ICS groups, respectively, experienced allergic rhinitis. Patients on ICS exhibited longer disease duration than those on NT. Compared with HC, patients prescribed NT and ICS showed worse lung function (decreased predicted percentage of forced expiratory volume in 1 s (FEV1%pre)). No significant differences in species coverage were observed among the groups. Compared with HC, patients on ICS showed significantly lower fungal richness (observed species), evenness (Pielou_e), and diversity (Shannon) (Fig. 1A). This is inconsistent with previously reported bacteriome results inplying indiscriminate bacterial richness or diversity in adult asthma patients.3, 4 These data suggest a different influence of asthma state on gut mycobiome and bacteriome; however, validation via large-scale studies is needed. Both the observed species and Shannon indices negatively correlated with disease duration and positively correlated with sputum neutrophil percentages (Fig. 1B). Principal coordinate analysis (PCoA) using Bray–Curtis distance observed significant clustering among all groups, indicating differing community composition (Adonis test, p = .026; Fig. 1C; Table S2). Figures 1D,E, and S1 showed the gut mycobiome composition at phylum, class, family, and genus levels. The dominant phyla were Ascomycota, Basidiomycota, and Mortierellomycota; the dominant genera were Malassezia and Aspergillus. Linear discriminant analysis effect size was used to investigate mycobiome biomarkers (discriminately enriched fungi) at various taxonomic ranks. This revealed that one class, three families, and three genera, including Fusarium, were discriminately enriched in HC. One phylum, one class, three families, and eight genera (Dictyophora, etc.) were enriched in the NT group, the polysaccharides of Dictyophora was demonstrated to have several physicochemical and biological properties.5 In the ICS group, one class, four families, and four genera were enriched (Fig. 1F,G), among which Russula is a valuable, medicinal fungus with anti-inflammatory bioactivities.6 The correlations between the mycobiome and clinical indices showed that disease duration correlated positively with the abundance of Ganoderma, Russula, and Senacina and negatively with the abundance of Saitozyma and Fusarium. Sputum neutrophils (%) were positively associated with the abundance of Mortierella, Didymella, Saitozyma, Penicillium, and Pyrenochaetopsis (Fig. 1H). Comparing the gut mycobiome in asthma patients with different disease severities against HC revealed that severe patients exhibited significantly lower fungal richness (observed species), evenness (Pielou_e), and diversity (Shannon), and mild-to-moderate patients showed significantly lower fungal diversity (Shannon) (Fig. S2A–2D). PCoA indicated different community compositions among HC and patients with mild-to-moderate and severe disease (Adonis test p = .008; Fig. S2E; Table S2). Metabolites synthesized in the gut possess bioactivities against diseases.7 To investigate the gut metabolome in asthma, 24 HC, and 23 and 31 patients in NT and ICS groups, respectively, were enrolled. Patients receiving ICS exhibited longer disease duration than those in NT group; patient in NT and ICS groups showed worse lung function (lower FEV1%pre and FEV1/forced vital capacity) than HC (Table S3). Orthogonal partial least-squares discriminant (OPLS-DA) analysis showed dots in dark blue clearly separated from green dots (NT vs. HC, and NT vs. ICS) in electrospray ionization positive (ESI+) and negative (ESI–) modes, indicating different metabolic profiles between HC and NT, or between NT and ICS groups (Fig. 2A,B). With variable importance in projection >1 and p < .05, nine (ESI+) and three (ESI–) discriminate metabolites were found between NT and HC groups (Fig. 2C, Table S4). For example, sphingosine levels were upregulated in the NT group. Sphingosine/sphingosine-1-phosphate was considered to trigger asthma.8 NT group also possessed decreased anti-inflammatory metabolite (e.g., vitamin A and L-fucose) levels.9, 10 Six (ESI−) and three (ESI+) differential metabolites were found between NT and ICS group (Fig. 2D, Table S5). Strong correlations were found between these metabolites (Fig. 2E,F). Additionally, Kyoto Encyclopedia of Genes and Genomes enrichment analysis of differential metabolites revealed apoptosis to be most affected in the NT group, followed by necroptosis (Fig. 2G); glycine, serine, and threonine metabolism were significantly affected in patients receiving ICS (Fig. 2H). Fungi, bacteria, and metabolites coexist in the gut and modulate the immune system; their interactions may influence human health.7 We previously profiled the gut bacteriome in asthma,4 we selected the same subsets from bacterial, fungal, and metabolic analysis to establish cross-domain associations and bacterial–fungal, bacterial–metabolite, and fungal–metabolite co-occurrence networks. We detected more connections within bacterial and fungal species, and between bacteria and fungi in the NT group than in HC, as evidenced by the former's higher density (.16 vs..05) and number of edges (1010 vs. 587). The figures were lower in the ICS group than that in NT group (.04 vs. .16 and 362 vs. 1010, respectively; Fig. S5–S7). Figure 3 showed that HC had a greater density of bacterial-metabolite connections than NT and ICS groups. HC and NT group exhibited distinct fungal–metabolite connections, whereas the networks in ICS group were markedly low (Fig. 4), suggesting complex relationships among bacteria, fungi, and metabolites. Moreover, the ICS treatment altered bacterial–metabolite and fungal–metabolite connections. Our study was limited by its relatively small-sample size, a lack of corrections for multiple testing and racially diverse participants, and an independent replication cohort. Because this was not a randomized ICS trial and non-independent variables of asthma were included in correlation analyses, confounders may account for the observed associations between the mycobiome and ICS or clinical indices (e.g., asthma severity). Therefore, these exploratory findings require validation through large-scale studies. In conclusion, asthma patients harboured altered gut fungal compositions, metabolic profiles, and associated metabolic pathways. ICS altered the mycobiome community and fungal/bacterial–metabolite connections. Our findings further our understanding of the gut mycobiome, metabolome, and fungal/bacterial–metabolite associations that may be critical in asthma pathogenesis and may therefore be therapeutic targets. The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported. the National Natural Science Foundation of China, Grant Numbers: 82170023 and 81970020; Shanghai Sailing Program, Grant Number: 20YF1428300; Shanghai Municipal Health Commission, Grant Number: 2019SY006; Shanghai Key Laboratory of Emergency Prevention, Diagnosis and Treatment of Respiratory Infectious Diseases, Grant Number: 20dz2261100 ; Shanghai Municipal Key Clinical Specialty, Grant Number: shslczdzk02202; Cultivation Project of Shanghai Major Infectious Disease Research Base, Grant Number: 20dz2210500; Shanghai Key Discipline for Respiratory Diseases, Grant Number: 2017ZZ02014; Shanghai Shenkang Hospital Development Center Clinical Science and Technology Innovation Project, Grant Number: SHDC12018102 Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.