The gut microbiome in end-stage lung disease and lung transplantation

Abstract

Gut dysbiosis has been associated with impaired outcomes in liver and kidney transplant recipients, but the gut microbiome of lung transplant recipients has not been extensively explored. We assessed the gut microbiome in 64 fecal samples from end-stage lung disease patients before transplantation and 219 samples from lung transplant recipients after transplantation using metagenomic sequencing. To identify dysbiotic microbial signatures, we analyzed 243 fecal samples from age-, sex-, and BMI-matched healthy controls. By unsupervised clustering, we identified five groups of lung transplant recipients using different combinations of immunosuppressants and antibiotics and analyzed them in relation to the gut microbiome. Finally, we investigated the gut microbiome of lung transplant recipients in different chronic lung allograft dysfunction (CLAD) stages and longitudinal gut microbiome changes after transplantation. We found 108 species (58.1%) in end-stage lung disease patients and 139 species (74.7%) in lung transplant recipients that were differentially abundant compared with healthy controls, with several species exhibiting sharp longitudinal increases from before to after transplantation. Different combinations of immunosuppressants and antibiotics were associated with specific gut microbial signatures. We found that the gut microbiome of lung transplant recipients in CLAD stage 0 was more similar to healthy controls compared to those in CLAD stage 1. Finally, the gut microbial diversity of lung transplant recipients remained lower than the average gut microbial diversity of healthy controls up to more than 20 years post-transplantation. Gut dysbiosis, already present before lung transplantation was exacerbated following lung transplantation.IMPORTANCEThis study provides extensive insights into the gut microbiome of end-stage lung disease patients and lung transplant recipients, which warrants further investigation before the gut microbiome can be used for microbiome-targeted interventions that could improve the outcome of lung transplantation.

Keywords: dysbiosis; end-stage lung disease; gut microbiome; immunosuppressive medication; lung transplantation.

Fig 1

The gut microbiome before and after lung transplantation is different than that of healthy controls. (a) Violin plots show the Shannon diversity index in healthy controls (HC) (gray), end-stage lung disease (ESLD) patients (orange), and lung transplant recipients (LTR) (steel blue). P-values are shown above each comparison. (b) PCA scatter plot with samples as dots: HC in gray, ESLD patients in orange and LTR in blue. The three larger circles represent the centroid of each group. The distance is the Aitchison distance, and samples closer to each other have more similar gut microbial community compositions. (c) Cells in the heatmap show the coefficient from a linear regression model that corresponds to the average difference (in log fold change) between either (i) ESLD patients compared with HC, (ii) LTR compared with HC, or (iii) LTR compared with ESLD patients. Cells in red and purple represent positive and negative log fold changes, respectively. Positive values indicate that the focal species has a higher abundance in the former group compared with the latter group, and vice versa for negative values. Cells in gray indicate species for which the comparison was not significant after FDR-correction.

Fig 2

Short-term dynamic changes of clr abundances of species with interesting alteration patterns. (a) Species with continuous enrichment between consecutive timepoints. (b) Species with a large expansion in three months after transplantation but either decreasing or becoming stagnant in their clr abundances as time since transplantation went by. (c) Species with a sharp decrease in three months after transplantation, of which the clr abundances in ESLD patients were similar to those in healthy controls. As a reference, each facet includes the average clr abundance (±SE) of the focal species in the healthy controls (healthy controls were not included in the linear model).

Fig 3

Effects of different medication regimens on gut microbiome diversity and microbial composition. (a) Violin plots show the Shannon diversity index in healthy controls (HC) and lung transplant recipients (LTR) on different medication regimens (MR1–MR5, see Table 3 for details). (b) Heatmap depicts the log fold changes (i.e., effect sizes) between all pairwise combinations of different medication regimens. Cells in red represent positive log fold change and those in purple negative log fold change, which means that, compared with the latter group in each comparison, the focal species has higher (red) or lower (purple) relative abundance in the former group. Cells in gray indicate that the focal species showed no significant difference in the corresponding comparison.

Fig 4

Association between gut dysbiosis and chronic lung allograft dysfunction. Violin plot shows the average Aitchison distances of lung transplant recipients in different chronic lung allograft dysfunction (CLAD) stages to healthy controls.