Altered Lung Microbiome and Metabolome Profile in Children With Pulmonary Arterial Hypertension Associated With Congenital Heart Disease

Abstract

Backgrounds: Pulmonary arterial hypertension (PAH) is characterized by progressive pulmonary vascular functional and structural changes, resulting in increased pulmonary vascular resistance and eventually right heart failure and death. Congenital Left-to-Right shunts (LTRS) is one type of congenital heart disease (CHD) and PAH associated with the congenital Left-to-Right shunt (PAH-LTRS) is a severe disease in children. However, changes in the lung microbiome and their potential impact on PAH-LTRS have not been not fully studied. We hypothesized that lung microbiota and their derived metabolites have been disturbed in children with PAH-LTRS, which might contribute to the progression and outcomes of PAH-LTRS.

Methods: In this study, 68 age- and sex-matched children of three different groups (patients with PAH-LTRS cohort, patients with LTRS but have no pathologic features of PAH cohort, and healthy reference cohort) were enrolled in the current study. Bronchoalveolar lavage fluid samples from these participants were conducted for multi-omics analysis, including 16S rRNA sequencing and metabolomic profiling. Data progressing and integration analysis were performed to identify pulmonary microbial and metabolic characteristics of PAH-LTRS in children.

Results: We found that microbial community density was not significantly altered in PAH-LTRS based on α-diversity analysis. Microbial composition analysis indicated phylum of Bacteroidetes was that less abundant while Lactobacillus, Alicycliphilus, and Parapusillimonas were significantly altered and might contribute to PAH in children with LTRS. Moreover, metabolome profiling data showed that metabolites involved in Purine metabolism, Glycerophospholipid metabolism, Galactose metabolism, and Pyrimidine metabolism were also significantly disturbed in the PAH-LTRS cohort. Correlation analysis between microbes and metabolites indicated that alterations in the microbial composition from the lung microbiota could eventually result in the disturbance in certain metabolites, and might finally contribute to the pathology of PAH-LTRS.

Conclusion: Lung microbial density was not significantly altered in patients with PAH-LTRS. Composition analysis results showed that the relative microbiome abundance was different between groups. Metabolome profiling and correlation analysis with microbiota showed that metabolome also altered in children with PAH-LTRS. This study indicated that pulmonary microbes and metabolites disturbed in PAH-LTRS could be potentially effective biomarkers and provides valuable perspectives on clinical diagnosis, treatment, and prognosis of pediatric PAH-LTRS.

Keywords: congenital heart disease; left to right shunt; lung; metabolome; microbiome; pulmonary arterial hypertension.

Figure 1

Pulmonary bacterial community diversity in different cohorts and Principal Coordinate Analysis (PCoA) clustering. (A) Venn diagram of OUTs detected in three cohorts. (B) The number of species identified in three groups. (C) α-diversity among three cohorts measured by the Chao, Shannon and PD whole tree. Results were present as means ± s.e.m. (D) PLS-DA was conducted to compare the overall similarities in the bacterial taxonomy based on β-diversity. Each principal component represents most of the variation between samples. PAH-LTRS, patients with pulmonary arterial hypertension associated with Left-to-Right shunt (n = 15); REF, healthy reference (n = 16); LTRS, patients with Left-to-Right shunts but without PAH (n = 16).

Figure 2

Taxonomic composition of the lung microbiota among three cohorts. (A) The relative frequency of top abundant taxa in each cohort at the phylum level. (B) The relative abundance analysis of top altered phylum in three pair of comparisons (means ± s.e.m.). (C) The relative frequency of top abundant taxa in each cohort at the genus level. (D) Analysis of the relative abundance of top fluctuated genera between any two different cohorts (means ± s.e.m.). Wilcoxon rank sum test, p < 0.05. PAH-LTRS, patients with pulmonary arterial hypertension associated with Left-to-Right shunt (n = 15); REF, healthy reference (n = 16); LTRS, patients with Left-to-Right shunts but without PAH (n = 16).

Figure 3

LEfSe analysis to visualize differences in bacterial taxa among different groups. (A) The phylogenetic distribution of significantly enriched bacteria in different groups were showed by the cladogram. Nodes with different color represent different microbial taxa that are significantly abundant in the groups and significantly contribute to the differences between cohorts. (B) LDA bar graphs were utilized to identify the microbiota taxa with potentially significant effects in the three groups. Linear discriminant analysis (LDA, score > 2.0) scores are shown on the x-axis and taxa with the larger LDA score, the greater effect they might have in this group. PAH-LTRS, patients with pulmonary arterial hypertension associated with Left-to-Right shunt (n = 15); REF, healthy reference (n = 16); LTRS, patients with Left-to-Right shunts but without PAH (n = 16).

Figure 4

Differential lung metabolites between different groups. (A) OPLS-DA was used to discriminate between different groups. (B) volcano plot of the differential metabolisms for any two groups (VIP > 1, P < 0.05). (C) Venn diagram showed the altered metabolites overlap among three pair of comparisions. PAH-LTRS, patients with pulmonary arterial hypertension associated with Left-to-Right shunt (n = 19); REF, healthy reference (n = 20); LTRS, patients with Left-to-Right shunts but without PAH (n = 20).

Figure 5

Differential metabolites clustering and KEGG enrichment. (A) Differential metabolites in REF vs. PAH-LTRS were clustered into 5 clusters. These sub-clusters had different expression trend in two groups. (B) KEGG enrichment of these differential metabolites in REF vs. PAH. (C) Three differential metabolites in LTRS vs. PAH-LTRS also shown with different abundance in two groups. (D) Three differential metabolites in LTRS vs. PAH-LTRS were annotated into two pathways without significantly enrichment. EIP, Environmental Information Processing; HD, Human Diseases; M, Metabolism; OS, Organismal Systems. PAH-LTRS, patients with pulmonary arterial hypertension associated with Left-to-Right shunt (n = 19); REF, healthy reference (n = 20); LTRS, patients with Left-to-Right shunts but without PAH (n = 20).

Figure 6

Significant correlation between lung bacteria and differential metabolites. Heatmap indicated positive (red) and negative (blue) correlations between metabolites and the microbiota at the genus level detected in REF compared with PAH-LTRS cohort (A) and LTRS compared with PAH-LTRS cohort (B). The legend shows correlation values from −1 to 1. *Represents significantly negative or positive correlations (*P < 0.05, **P < 0.01). PAH-LTRS, patients with pulmonary arterial hypertension associated with Left-to-Right shunt (n = 19); REF, healthy reference (n = 20); LTRS, patients with Left-to-Right shunts but without PAH (n = 20).