Cholestasis impairs gut microbiota development and bile salt hydrolase activity in preterm neonates

Abstract

Cholestasis refers to impaired bile flow from the liver to the intestine. In neonates, cholestasis causes poor growth and may progress to liver failure and death. Normal bile flow requires an intact liver-gut-microbiome axis, whereby liver-derived primary bile acids are transformed into secondary bile acids. Microbial bile salt hydrolase (BSH) enzymes are responsible for the first step, deconjugating glycine- and taurine-conjugated primary bile acids. Cholestatic neonates often are treated with the potent choleretic bile acid ursodeoxycholic acid (UDCA), although interactions between UDCA, gut microbes, and other bile acids are poorly understood. To gain insight into how the liver-gut-microbiome axis develops in extreme prematurity and how cholestasis alters this maturation, we conducted a nested case-control study collecting 124 stool samples longitudinally from 24 preterm infants born at mean 27.2 ± 1.8 weeks gestation and 946 ± 249.6 g, half of whom developed physiologic cholestasis. Samples were analyzed by whole metagenomic sequencing, in vitro BSH enzyme activity assays optimized for low biomass fecal samples, and quantitative mass spectrometry to measure the bile acid metabolome. In extremely preterm neonates, acquisition of the secondary bile acid biosynthesis pathway and BSH genes carried by Clostridium perfringens are the most prominent features of early microbiome development. Cholestasis interrupts this developmental pattern. BSH gene abundance and enzyme activity are profoundly reduced in cholestatic neonates, resulting in decreased quantities of unconjugated bile acids. UDCA restores total fecal bile acid levels in cholestatic neonates, but this is due to a 522-fold increase in fecal UDCA. A majority of bile acids in early development are atypical positional and stereo-isomers of bile acids. We report novel associations linking isomeric bile acids and BSH activity to neonatal growth trajectories. These data highlight deconjugation of bile acids as a key microbial function that is acquired in early neonatal development and impaired by cholestasis.

Keywords: Microbiome; bile acids; bile salt hydrolase; cholestasis; growth; neonate; premature infant; ursodeoxycholic acid.

Figure 1.

Development of the preterm gut microbiome in control neonates. (a) Observed OTUs increase with age over the first 6 weeks of life (n = 18–25). (b) The gut microbiome changes at both phylum and genus levels in the first 6 weeks of life (n = 64). (c) At the species level, Clostridium perfringens is most significantly increased and Staphylococcus epidermidis and Gemella sanguinis are most significantly decreased with increasing PMA (n = 13–31). (d) LEfSe analysis identified secondary bile acid biosynthesis as the most enriched pathway in control neonates 32–40 weeks PMA relative to control neonates 25–28 weeks PMA (n = 20–31). (e) The abundance of the secondary bile acid biosynthesis pathway increases with PMA (n = 13–31). BA, bile acid; LEfSe, linear discriminate analysis effect size; OTU, operational taxonomic unit; PMA, post-menstrual age.

Figure 2.

Cholestasis disrupts microbiota maturation. (a) Position on PC Axis 1 increases linearly with increasing PMA in control neonates; this pattern is absent in cholestatic neonates (n = 49–64). (b) LEfSe analysis identified secondary bile acid biosynthesis as the most enriched pathway in control neonates compared to cholestatic neonates 32–40 weeks PMA (n = 13–17). Relative abundances of (c) the secondary bile acid biosynthesis pathway, (D) the BSH gene, and (e) the BSH-carrying bacterium Clostridium perfringens are reduced in cholestatic neonates at 32–40 weeks PMA (n = 13–17). (f) Schematic of BSH activity assay. (g) Glycine deconjugation is preferred over taurine deconjugation in stool from control neonates; BSH activity is reduced in stool from cholestatic neonates (n = 5–6). (h) Total fecal unconjugated bile acids are reduced in cholestatic neonates, although the quantity of fecal conjugated bile acids is unchanged between groups (n = 16–18). BA, bile acid; BSH, bile salt hydrolase; LDA, linear discriminate analysis; LEfSe, linear discriminate analysis effect size; PC, principal coordinate; PMA, post-menstrual age.

Figure 3.

Fecal bile acid profiles are altered during cholestasis. Bile acid stereoisomers dominate in early development, then primary unconjugated bile acids increase in abundance in control neonates. Bile acid stereoisomers are less abundant and bile acid deconjugation is impaired in cholestatic neonates (excluding samples obtained during UDCA treatment) (n = 14–26). α-MCA, alpha-muricholic acid; CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; GCA, glycocholic acid; GCDCA, glycochenodeoxycholic acid; GHyoCA, glycohyocholic acid; HyoCA, hyocholic acid; Isomer2-T-Triol, taurine conjugate of an unidentified trihydroxy-cholanoic acid; PMA, post-menstrual age; Tα-MCA, tauro-α-muricholic acid; TCA, taurocholic acid; TCDCA, taurochenodeoxycholic acid; THyoCA, taurohyocholic acid; THyoDCA, taurohyodeoxycholic acid; UDCA, ursodeoxycholic acid.

Figure 4.

UDCA treatment alters microbiota and fecal bile acid profiles. (a) PC analysis of bile acid quantification reveals that stools obtained during UDCA treatment cluster distinctively (n = 124). (b) UDCA treatment restores fecal bile acid concentrations to those of control infants 29–40 weeks PMA (n = 7–20). (c) There are higher concentrations of UDCA in treated samples compared to untreated samples from cholestatic neonates 29–40 weeks PMA (n = 7–20). (d) UDCA treatment shifts the fecal bile acid pool to predominately UDCA in neonates 29–40 weeks PMA (n = 7–20). (e) UDCA administration increases the relative abundance of Firmicutes and decreases the relative abundance of Proteobacteria in infants 29–40 weeks PMA (n = 5–20). (F) BSH-carrying Clostridium perfringens is increased with UDCA treatment, while BSH-carrying Bifidobacterium breve is decreased with UDCA treatment in cholestatic infants 29–40 weeks PMA (n = 5–20). BA, bile acid; CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; GCA, glycocholic acid; GCDCA, glycochenodeoxycholic acid; HyoCA, hyocholic acid; Isomer2-T-Triol, taurine conjugate of an unidentified trihydroxy-cholanoic acid; ω-MCA, omega-muricholic acid; PC, principal coordinate; TCA, taurocholic acid; TCDCA, taurochenodeoxycholic acid; THyoCA, taurohyocholic acid; UDCA, ursodeoxycholic acid.

Figure 5.

Fecal bile acids and BSH are associated with neonatal growth outcomes. (a) Schematic of BSH transformation of TCA and TCDCA into CA and CDCA, respectively. (b) In neonates 32–40 weeks PMA, the relative proportions of the BSH gene and of fecal cholic acid are associated positively with growth (n = 12, all groups). (c) In neonates 25–28 weeks PMA, the proportion of combined fecal TCA and TCDCA is associated negatively with growth, and the proportion of combined fecal Isomer2-T-Triol and GHyoCA is associated positively with growth (n = 12, all groups). BSH, bile salt hydrolase; CA, cholic acid; CDCA, chenodeoxycholic acid; GHyoCA, glycohyocholic acid; Isomer2-T-Triol, taurine conjugate of an unidentified trihydroxy-cholanoic acid; PMA, post-menstrual age; TCA, taurocholic acid; TCDCA, taurochenodeoxycholic acid.

Figure 6.

The most distinctive features of preterm gut microbiome development are increasing abundance of Clostridium perfringens and BSH, which increases total and relative proportions of unconjugated BAs. Cholestasis disrupts the acquisition of C. perfringens, BSH enzyme activity, and the capacity to form unconjugated bile acids. Enteral UDCA administration increases the abundance of C. perfringens and dramatically increases UDCA in the stool. BA, bile acid; BSH, bile salt hydrolase; UDCA, ursodeoxycholic acid. Created with BioRender.Com.