Characterization of the Gut Microbiota and Mycobiota in Italian Pediatric Patients With Primary Sclerosing Cholangitis and Ulcerative Colitis

Abstract

Background: Primary sclerosing cholangitis (PSC) is a chronic, fibroinflammatory, cholestatic liver disease of unknown etiopathogenesis, often associated with inflammatory bowel diseases. Recent evidence ascribes, together with immunologic and environmental components, a significant role to the intestinal microbiota or its molecules in the PSC pathogenesis.

Methods: By metagenomic sequencing of 16S rRNA and ITS2 loci, we describe the fecal microbiota and mycobiota of 26 pediatric patients affected by PSC and concomitant ulcerative colitis (PSC-UC), 27 patients without PSC but with UC (UC), and 26 healthy subjects (CTRLs).

Results: Compared with CTRL, the bacterial and fungal gut dysbiosis was evident for both PSC-UC and UC groups; in particular, Streptococcus, Saccharomyces, Sporobolomyces, Tilletiopsis, and Debaryomyces appeared increased in PSC-UC, whereas Klebsiella, Haemophilus, Enterococcus Collinsella, Piptoporus, Candida, and Hyphodontia in UC. In both patient groups, Akkermansia, Bacteroides, Parabacteroides, Oscillospira, Meyerozyma and Malassezia were decreased. Co-occurrence analysis evidenced the lowest number of nodes and edges for fungi networks compared with bacteria. Finally, we identified a specific patient profile, based on liver function tests, bacterial and fungal signatures, that is able to distinguish PSC-UC from UC patients.

Conclusions: We describe the gut microbiota and mycobiota dysbiosis associated to PSC-UC disease. Our results evidenced a gut imbalance, with the reduction of gut commensal microorganisms with stated anti-inflammatory properties (ie, Akkermansia, Bacteroides, Parabacteroides, Oscillospira, Meyerozyma, and Malassezia) and the increase of pathobionts (ie, Streptococcus, Saccharomyces, and Debaryomyces) that could be involved in PSC progression. Altogether, these events may concur in the pathophysiology of PSC in the framework of UC.

Keywords: dysbiosis; gut microbiota; gut mycobiota; primary sclerosing cholangitis; ulcerative colitis.

Figure 1.

Biodiversity of the bacterial gut microbiota. The bacterial subset of gut microbiota was analyzed in PSC-UC, UC, and CTRL groups. Alpha diversity was estimated by Shannon (A) and Chao1 (B) indexes. Median, percentiles and minimum and maximum values of indexes were plotted; Kruskal-Wallis test was used for comparisons. Beta diversity was assessed by principal coordinate analysis of Bray-Curtis distance (C). The fraction of diversity captured by the coordinate is reported as a percentage. The PERMANOVA test was used for group comparisons. Linear Discriminant Analysis Effect Size (LEfSe) was calculated at phylum (D) and genus (E) levels. The P values were FDR adjusted (P < .05) and LDA score was >2.

Figure 2.

Biodiversity of the fungal gut mycobiota. The fungal subset of gut mycobiota was analyzed in PSC-UC, UC, and CTRL. Alpha diversity was estimated by Shannon (A) and Chao1 (B) indexes. Median, percentiles and minimum and maximum values of indexes were plotted; Kruskal-Wallis test was used for comparisons. Beta diversity was assessed by principal coordinate analysis of Bray-Curtis distance (C). The fraction of diversity captured by the coordinate is reported as a percentage. The PERMANOVA test was used for comparisons. Linear Discriminant Analysis Effect Size (LEfSe) was calculated at phylum (D) and genus (E) levels. The P values were FDR adjusted (P < .05) and LDA score was >2.

Figure 3.

Correlations amongst microbes, fungi, and liver function features. Network analysis of intestinal microbiota (A) and mycobiota (B) represented by Pearson’s correlation between PSC-UC, UC, and CTRL groups. The nodes refer to genera; the edges represent the correlation values (filtered for adjusted P < .05) between genera (blue lines, negative correlations; red lines, positive correlations). Nodes are colored according to their relative abundance in each group. Spearman’s correlation amongst bacteria, fungi, and liver function features (C). Red blocks: positive Spearman’s coefficient with P ≤ .005; blue blocks: negative Spearman’s coefficient with P ≤ .005.

Figure 4.

Principal component analysis (PCA) plot. Red dots: PSC-UC samples; blue dots: UC samples; black dots: loadings.

Figure 5.

Metabolic biomarkers associated with PSC-UC, UC, and CTRL. A linear discriminant effect size (LeFse) analysis have been performed (α = 0.05, logarithmic LDA score threshold = 2.0).