• Potential conflict of interest: Nothing to report.


We thank Prof. Alan F. Hofmann for his comments and the opportunity for a deeper discussion on the role of glucuronidation as a conjugating pathway for bile acids.

Several reports revealed significant levels of bile acid glucuronides in human serum1-3 and urine2, 4-6 in healthy subjects3, 6 or cholestatic patients.1-5 In particular, CDCA-glucuronide was identified as one of the most abundant bile acid glucuronides in human serum and urine.2, 3 On the basis of these observations, our study aimed to determine which type of glucuronide was formed in the presence of human liver microsomes and to identify which of the 18 human UDP-glucurosonyltransferase (UGT) enzymes were involved in CDCA glucuronidation. We reported that human liver microsomes mainly produce CDCA-24glucuronide, and we identified UGT1A3 as the human enzyme implicated in the formation of this glucuronide conjugate.7 Although our systematic study highlights the role of specific UGTs, we did not explore the physiological role of this conjugation reaction nor demonstrated the relative occurrence of this pathway compared to other conjugation pathways for bile acids.

In his letter, Hofmann indicates that endogenous bile acids are more actively conjugated with amino acids than with a glucuronosyl group. Based on our current knowledge and the available literature, we agree that glucuronidation is likely not the major conjugation pathway for bile acids. Nevertheless, Hofmann indicates that bile acid acyl glucuronidation and amidation are mutually exclusive, because both conjugation reactions involve the 24 carboxyl group of CDCA. Our work further indicates that tauro-CDCA or glyco-CDCA are not substrates for UGT1A3,8 suggesting that the presence of glycine or taurine prevents C24 glucuronidation of bile acids. However, a major distinction between acyl glucuronidation and amidation concerns their respective consequences on the biological activity of bile acids. Glyco-CDCA and tauro-CDCA as well as CDCA share a comparable efficacy to activate the bile acid sensor farnesoid X receptor.9 By contrast, we clearly demonstrated that UGT1A3-catalyzed glucuronidation abolishes such an action of CDCA.8 This observation supports that the 2 conjugation pathways lead to distinct physiological consequences for bile acids and that only glucuronidation may be considered as an inactivating process of biologically active bile acids. Glucuronidation, more than amidation, may have profound physiological consequences on the ability of bile acids to modulate gene expression.

On the other hand, we also demonstrated that human UGT1A3 expression and C24-glucuronidation activity was induced in human hepatocytes exposed to pharmacological activators of the peroxisome proliferator-activated pathway receptor-α (PPARα).8 These observations were subsequently confirmed in vivo in a transgenic mouse model expressing the human UGT1 locus.7 Because glucuronidation is considered a detoxification pathway for bile acids, we proposed that cholesterol-lowering drugs acting through PPARα (i.e., fibrates), may be considered in the clinic to enhance the inactivation and detoxification of bile acids such as CDCA.

In conclusion, our observations suggest that the low level of bile acid glucuronidation as it occurs under normal conditions can be induced in the presence of pharmacological activators of nuclear receptors, such as PPARα. Because glucuronidation inactivates bile acids, such induction may have profound consequences on the maintenance of the nuclear receptor–mediated biological effects of bile acids. However, and in accordance with Hofmann's letter, the physiological significance of these observations remains to be established in humans, as is clearly stated in our article.

Olivier Barbier*, Jocelyn Trottier*, * CHUQ-CHUL Research Centre, Québec, Canada.