The fungal microbiota modulate neonatal oxygen-induced lung injury

Abstract

Background: The immature lungs of very preterm infants are exposed to supraphysiologic oxygen, contributing to bronchopulmonary dysplasia (BPD), a chronic lung disease that is the most common morbidity of prematurity. While the microbiota significantly influences neonatal health, the relationship between the intestinal microbiome, particularly micro-eukaryotic members such as fungi and yeast, and lung injury severity in newborns remains unknown.

Results: Here, we show that the fungal microbiota modulates hyperoxia-induced lung injury severity in very low birth weight premature infants and preclinical pseudohumanized and altered fungal colonization mouse models. Instead of fungal communities dominated by Candida and Saccharomyces, the first stool microbiomes of infants who developed BPD had less interconnected community architectures with a greater diversity of rarer fungi. After using a pseudohumanized model to show that transfer to the neonatal microbiome from infants with BPD increased the severity of lung injury, we used gain and loss of function approaches to demonstrate that modulating the extent of initial neonatal fungal colonization affected the extent of BPD-like lung injury in mice. We also identified alterations in the murine intestinal microbiome and transcriptome associated with augmented lung injury.

Conclusions: These findings demonstrate that features of the initial intestinal fungal microbiome are associated with the later development of BPD in premature neonates and exert a microbiome-driven effect that is transferable and modifiable in murine models, which suggests both causality and a potential therapeutic strategy. Video Abstract.

Keywords: Chronic lung injury; Fungal microbiome; Gut microbiome; Multikingdom microbiome; Mycobiome; Preterm infant; Very low birthweight; Yeast.

Fig. 1

The fungal gut microbiota is altered in infants that later develop bronchopulmonary dysplasia. Intestinal microbiome analysis from the first true stool from 102 very low birthweight premature infants (birthweight < 1500 g). A Fungal alpha diversity of infants with moderate to severe bronchopulmonary dysplasia or death prior to 36 weeks of gestation (BPD) to those without (NoBPD) as quantified by the Shannon (p = 0.0010, Mann–Whitney) and Simpson indices (p = 0.0021, Mann–Whitney). Values represent means ± SEM. B Bacterial alpha diversity. C Fungal beta diversity. Principal coordinates analysis (PCoA) of Bray–Curtis dissimilarity displaying differences in fungal beta diversity (p = 0.040, permutational multivariate analysis of ANOVA, PERMANOVA. p = 0.44, permutational multivariate analysis of dispersion, PERMDISP). Ellipses indicate the 95% confidence interval. D Biplot of principal components analysis (PCA) with fungal community composition driven by amplicon sequence variants (ASVs) aligning to the genera Candida and Aureobasidium. PCA is shown in the inset. E Bacterial beta diversity. PCoA of Bray–Curtis dissimilarity displaying bacterial beta diversity (p = 0.06109, PERMANOVA. p = 0.37, PERMDISP). F Biplot of PCA of bacterial community composition. PCA is shown in the inset. G and H SPIEC-EASI network analysis shows that BPD infants formed sparser multikingdom interaction networks than NoBPD infants (fewer fungal than bacterial edges but a similar number of nodes, and lower edge density). See also Fig. S2–3

Fig. 2

Sex-specific differences in premature infants that will develop BPD. Pre-determined analysis of 102 very low birthweight premature infants. A Fungal alpha diversity in infants in infants assigned male (AMAB) or female at birth (AFAB) with NoBPD or BPD, as quantified by the Shannon and Simpson indices. Values represent means ± SEM. Significance testing by two-way Mann–Whitney. B Bacterial alpha diversity. Significance testing by two-way Mann–Whitney. C Fungal beta diversity. Principal coordinates analysis (PCoA) of Bray–Curtis dissimilarity showing differences in the beta diversity of the mycobiome for each sex and disease state (p = 0.041, PERMANOVA, p = 0.5957, PERMDISP). Ellipses indicate the 95% confidence interval. D Bacterial beta diversity. PCoA of Bray–Curtis dissimilarity of the bacterial microbiome (p = 0.32, PERMANOVA. P = 0.64, PERMDISP). E–H SPIEC-EASI network analysis shows that BPD infants AMAB formed the sparsest networks (fewer edges, but a similar number of nodes, lower edge density). See also Fig. S6–7

Fig. 3

Microbiome-driven clustering analysis of heterogeneity in the mycobiome. A Principal Coordinates Analysis (PCoA) plot showing a five-cluster Dirichlet-Multinomial Mixture (DMM) model clustering of the mycobiome, n = 102. In Cluster 1, 3 infants developed BPD, and 18 did not (NoBPD). In Cluster 2: 6 BPD, 7 NoBPD, Cluster 3: 9 BPD and 3 NoBPD, Cluster 4: 5 BPD and 6 NoBPD, Cluster 5: 1 BPD and 5 NoBPD. B Fungal community diversity analyses indicate the greatest microbial richness in Cluster 3, as quantified by the Shannon diversity index. Values represent means ± SEM. Significance testing by ANOVA. C Average relative abundance of fungi in each DMM cluster. D–F A two-cluster DMM model in NoBPD infants, with the greatest richness in Cluster 2, n = 64. G–I A two-cluster DMM model in BPD infants with the highest microbial richness in Cluster 2, n = 38. Clust1–5, Cluster 1–5. Ellipses indicate the 95% confidence interval. See also Fig. S8–10

Fig. 4

Fecal microbiota transfer from infants with BPD augments lung injury in mice. A Experimental schematic showing pregestational fecal microbiota transfer to colonize neonatal mice with the intestinal microbiota from human infants with BPD or without (NoBPD). Hyperoxia (HO) or normoxia (NO) exposure was performed from day 3 to day 14 of life (P3–14). Tissue harvest and subsequent analyses were performed at P14. Schematic created using BioRender.com. Total sample size is 18–24 neonatal mice from three independent litters per experimental group. B Gut fungal microbiome analysis showing alpha diversity as quantified by the Shannon diversity index and a principal coordinates analysis (PCoA) plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 12 mice/group). C Gut bacterial microbiome analysis showing alpha diversity as quantified by the Shannon diversity index and a principal coordinates analysis (PCoA) plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 12 mice/group). D Lung fungal microbiome analysis showing alpha diversity as quantified by the Shannon diversity index and a principal coordinates analysis (PCoA) plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 12 mice/group). E Lung bacterial microbiome analysis showing alpha diversity as quantified by the Shannon diversity index and a principal coordinates analysis (PCoA) plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 12 mice/group). F Representative hematoxylin and eosin confocal micrographs of the distal lung of 14-day-old neonatal mice born to pseudohumanized mice with fecal microbiota transport from BPD-associated microbiomes (BPD-FMT) or NoBPD-associated microbiomes (NoBPD-FMT). Scale bars, 100 µm. G Histomorphological analysis and forced oscillometry. Morphometry is quantified by mean linear intercept and radial alveolar count. Forced oscillometry is quantified by resistance and compliance. Significance testing by two-way ANOVA. H Representative confocal microscopy images of immunofluorescence of alpha-smooth muscle actin and von Willebrand’s factor showing vascular changes consistent with early pulmonary hypertension and reduced vascular density of 20–50 µm vessels. Scale bars, 20 and 50 µm. Significance testing by two-way ANOVA. I–K Hyperoxia exposure induces different transcriptomic profiles in the lungs in BPD-FMT or NoBPD-FMT mice by RNA-seq (n = 4 mice/group). I Volcano plots of the hyperoxia-induced transcriptomic alterations in BPD-FMT or NoBPD-FMT mice. J Venn diagram showing differential gene expression in BPD-FMT and NoBPD-FMT as compared to SPF controls. K Heatmaps of the top 10 most differentially expressed genes, either unique BPD-FMT, shared or unique to NoBPD. FiO₂, fraction of inspired oxygen. Values represent means ± SEM. See also Fig. S11–12

Fig. 5

Inhibiting fungal colonization reduces hyperoxia-induced lung injury in mice. A Experimental schematic showing prenatal fluconazole exposure to inhibit neonatal fungal colonization. Hyperoxia (HO) or normoxia (NO) exposure was performed from day 3 to day 14 of life (P3–14). Tissue harvest and subsequent analyses were performed at P14. Schematic created using BioRender.com. Total sample size is 18–24 neonatal mice from three independent litters per experimental group. B Gut fungal microbiome analysis, showing alpha diversity as quantified by the Shannon diversity index and a PCoA plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney and PERMANOVA (n = 8 mice/group). C Gut bacterial microbiome analysis, showing alpha diversity as quantified by the Shannon diversity index and a principal coordinates analysis (PCoA) plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney and PERMANOVA (n = 8 mice/group). D Lung fungal microbiome analysis, showing alpha diversity as quantified by the Shannon diversity index and a PCoA plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 8 mice/group). E Lung bacterial microbiome analysis, showing alpha diversity as quantified by the Shannon diversity index and a PCoA plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 8 mice/group). F Representative hematoxylin and eosin confocal micrographs of the distal lung of 14-day-old neonatal mice born to mouse dams exposed to fluconazole or conventional specific pathogen-free (SPF), vehicle-exposed, control mice. Scale bars, 100 µm. Histomorphological analysis and forced oscillometry. Morphometry is quantified by mean linear intercept and radial alveolar count. Forced oscillometry is quantified by resistance and compliance. Significance testing by two-way ANOVA. G Representative immunofluorescence of alpha-smooth muscle actin. Quantification of vascular density of 20–50 µm vessels, and echocardiography quantification of right ventricular systolic pressure. Scale bars, 100 µm. Significance testing by two-way ANOVA. H Flow cytometric quantification of group 2 innate lymphoid cells (ILC2), expressed as a percentage of CD45⁺ cells. ILC2 were defined as live, lineage negative, CD45⁺, CD90.2⁺, CD127⁺, IL-33R (ST2)⁺, and GATA3⁺ cells. Significance testing by two-way ANOVA (n = 5–7 mice/group). I–K Hyperoxia exposure alters the lung transcriptome in the lungs in fluconazole-exposed or SPF mice by RNA-seq (n = 4 mice/group). I Volcano plot of the hyperoxia-induced transcriptomic alterations in fluconazole-exposed mice and SPF controls. J Venn diagram showing differential gene expression in fluconazole-exposed neonatal mice compared to SPF controls. K Heatmaps of the top 10 most differentially expressed genes either unique to SPF controls, shared, or unique to fluconazole-exposed mice. FiO₂, fraction of inspired oxygen. Values represent means ± SEM. See also Fig. S14–17, Table S4

Fig. 6

Amplifying fungal colonization augments hyperoxia-induced lung injury in mice. A Experimental schematic showing cefoperazone-facilitated intestinal colonization with the mouse commensal yeast, Candida tropicalis (C. trop). Hyperoxia (HO) or normoxia (NO) exposure was performed from day 3 to day 14 of life (P3-14). Tissue harvest and subsequent analyses were performed at P14. Total sample size is 18–24 neonatal mice from three independent litters per experimental group. B Gut fungal microbiome analysis, showing alpha diversity as quantified by the Shannon diversity index and a principal coordinates analysis (PCoA) plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 8 mice/group). C Gut bacterial microbiome analysis, showing alpha diversity as quantified by the Shannon diversity index and a principal coordinates analysis (PCoA) plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 8 mice/group). D Lung fungal microbiome analysis, showing alpha diversity as quantified by the Shannon diversity index and a PCoA plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 8 mice/group). E Lung bacterial microbiome analysis, showing alpha diversity as quantified by the Shannon diversity index and a PCoA plot showing Bray–Curtis dissimilarity. Significance testing by Mann–Whitney or PERMANOVA (n = 8 mice/group). F Representative hematoxylin and eosin confocal micrographs of the distal lung of 14-day-old neonatal mice born to mouse dams colonized by C. trop or conventional specific pathogen-free (SPF), vehicle-exposed, control mice. Scale bars, 100 µm. Histomorphological analysis and forced oscillometry. Morphometry is quantified by mean linear intercept, radial alveolar count. Forced oscillometry is quantified by resistance and compliance. Significance testing by two-way ANOVA. G Representative immunofluorescence of alpha-smooth muscle actin. Quantification shows vascular density of 20–50 µm vessels and echocardiography quantification of right ventricular systolic pressure. Scale bars, 100 µm. Significance testing by two-way ANOVA. H Flow cytometric quantification of group 2 innate lymphoid cells (ILC2), expressed as a percentage of CD45⁺ cells. ILC2 were defined as live, lineage negative, CD45⁺, CD90.2⁺, CD127⁺, IL-33R (ST2)⁺, and GATA3⁺ cells. Significance testing by two-way ANOVA (n = 5–7 mice/group). I–K Hyperoxia exposure alters the lung transcriptome in the lungs in fluconazole-exposed or SPF mice by RNA-seq (n = 4 mice/group). I Volcano plots of the hyperoxia-induced transcriptomic alterations in Ctrop mice and specific pathogen-free (SPF) controls. J Venn diagram showing unique gene regulation in Ctrop mice compared to SPF controls. K Heatmaps of the top 10 most differentially expressed genes either unique to SPF controls, shared, or unique to Ctrop mice. FiO₂, fraction of inspired oxygen. Schematic created using BioRender.com. Sample size is 18–24 neonatal mice from three independent litters per experimental group. Values represent means ± SEM. See also Fig. S18–19 and Table S4