Oral administration of pioglitazone inhibits pulmonary hypertension by regulating the gut microbiome and plasma metabolome in male rats

Abstract

The oral administrated thiazolidinediones (TZDs) have been widely reported to alleviate experimental pulmonary hypertension (PH). However, previous studies mainly focused on their beneficial effects on the cardiopulmonary vascular system but failed to determine their potential roles on gut microenvironment. This study aims to investigate the effects of pioglitazone, an oral TZD drug, on gut microbiome in classic PH rat models induced by hypoxia (HPH) or SU5416/hypoxia (SuHx-PH) and evaluate the therapeutic potential of supplementation of selective probiotics for experimental PH. Pioglitazone remarkably inhibited the PH pathogenesis in both models and reshaped the gut microbiome and plasma metabolome. Correlation analyses represented strong and unique association between the protective metabolites and bacteria genera (Roseburia, Lactobacillus, and Streptococcus) that were positively stimulated by pioglitazone. Supplementation of selective probiotics Roseburia intestinalis (R. intestinalis) partially attenuated SuHx-PH and rebuilt a novel gut microbiome and host metabolome. This study reports for the first time that oral administration of pioglitazone protects PH by regulating the gut microbiome and host metabolome, providing novel insights for the TZD drugs. The data also supports that modulation of gut microbiota by supplementation of selective probiotics could be a novel effective therapeutic strategy for the treatment of PH.

Keywords: gut microbiome; peroxisome proliferator‐activated receptor gamma; pioglitazone; plasma metabolome; pulmonary hypertension.

FIGURE 1

Pioglitazone inhibits experimental PH rat models induced by hypoxia or SuHx. (a) Flow chart showing the experimental design. (b–d) Trace (b) and analyzed graphs (c and d) representing the right ventricular systolic pressure (RVSP, c), and Fulton Index (d). Scatter plot values were mean ± SEM, n = 4–6 rats in each group. (e and f) Hematoxylin and eosin (H&E) staining images (e) and analyzed bar graph (f) of lung sections showing the thickening of the pulmonary arteries. (g and h) Representative immunofluorescent (IF) staining images (g) and analyzed bar graph (h) showing the immunofluorescent staining of both the smooth muscle cell marker alpha‐Actin (α‐SMC, green), together with DAPI (blue) which indicates the nucleus. Scatter plot values were mean ± SEM, n = 5 rats in each group. Scale bar indicates 50 μm (e) and 20 μm (g), respectively.

FIGURE 2

Altered gut microbiome in HPH and SuHx‐PH rats treated with or without pioglitazone. (a–c) α‐diversity indexes showing Ace (a), Chao (b), and Shannon (c) at genus level. (d and e) β‐diversity showing PCA (d) and PLS‐DA (e) of the microbiome profile. (f) Percentage bar graph showing the relative abundance of microbiota at genus level.

FIGURE 3

Altered plasma metabolome in HPH and SuHx‐PH rats treated with or without pioglitazone. (a and b) Volcano plots representing the plasma metabolites between “N” versus “S” (a) and “S” versus “SP” (b), where red and green indicated upregulated and downregulated groups. (c and d) KEGG pathway analysis for the identification of changed pathways between “N” versus “S” (c) and “S” versus “SP” (d). (e) β‐diversity showing PCA of the metabolome profile. (f) The k‐means clustering of metabolites.

FIGURE 4

Correlation analysis between gut microbiome and plasma metabolome. (a–d) Heatmaps of the Spearman's correlation coefficients between changes in Cluster 1 (a), Cluster 3 (b), Cluster 4 (c), and Cluster 6 (d) of metabolites and bacteria phyla caused by PH circumstances and pioglitazone intervention. (e and f) Heatmaps of the Spearman's correlation coefficients between changes in Cluster 1 (e) and Cluster 6 (f) of metabolites and bacteria genera caused by PH circumstances and pioglitazone intervention. The data were adjusted for body weight, fat mass, and waist‐to‐hip ratio. ⁺ p < 0.05, *FDR <0.1, and **FDR <0.05 as indicated. Pink and blue in the far‐right column indicated increased and decreased relative abundance, respectively. Only taxa with significant correlations (at least one based on FDR or two based on raw p value) were marked.

FIGURE 5

Supplementation of R. intestinalis effectively attenuates the pathogenesis in SuHx‐PH rats. (a) Flow chart showing the experimental design. (b–d) Trace (b) and analyzed graphs (c and d) representing the RVSP (c) and Fulton Index (d). Scatter plot values were mean ± SEM, n = 5–6 rats in each group. (e and f) H&E staining images (e) and analyzed bar graph (f) of lung sections showing the thickening of the pulmonary arteries, n = 5 rats in each group. Scale bar indicates 50 μm. (g–k) Echocardiography (g and i) and analyzed bar graph (h, j, and k) showing the indexes of right heart function, PAT/PET, RVEDWT, and RVESWT. Scatter plot values were mean ± SEM, n = 5–6 rats in each group.

FIGURE 6

Supplementation of R. intestinalis reshapes the gut microbiome in SuHx‐PH rats. (a–c) α‐diversity indexes showing Ace (a), Chao (b), and Simpson (c) at genus level. (d) β‐diversity showing the PLS‐DA of microbiome profile, n = 5–6 rats in each group. (e) Bar plot showing the relative abundance of Roseburia. (f) Percentage bar graph showing the relative abundance of microbiota at genus level. (g) Kruskal–Wallis H test bar plot representing the top fifteen mostly altered taxa among five groups; p values were listed as indicated. (h and i) Percentage bar graph showing the relative abundance of SCFA‐producing bacteria (h) and TMAO‐producing bacteria (i).

FIGURE 7

Supplementation of R. intestinalis rebuilds the host metabolome in SuHx‐PH rats. (a) Differences in β‐diversity were measured by PCA representing grouped microbiome profile, n = 5–6 rats in each group. (b) Venn diagram showing the shared and unique plasma metabolites in each group. (c and d) Volcano plots representing the plasma metabolites between “SR” versus “S” (c) and “SAR” versus “SA” (d). (e and f) KEGG enrichment analysis showing the enriched pathways in “SR” versus “S” (e) and “SAR” versus “SA” (f). (g) Proposed working model of the study.