Letter to the Editor IN VITRO STUDY OF MYCOPHENOLIC ACID GLUCURONIDATION

IN VITRO STUDY OF MYCOPHENOLIC ACID GLUCURONIDATION In the July issue of Drug Metabolism and Disposition was published an article on the metabolism of the immunosuppressive drug mycophenolic acid (MPA) by Basu et al. (2004). Recently, the metabolism of MPA has been extensively studied since it might contribute to the interindividual variation of its pharmacokinetics and presumably to some of the drug-drug interactions and side effects observed under mycophenolate mofetil therapy. Identification of the enzymes primarily involved in the metabolism of MPA will clearly be helpful. However, we have some comments and suggestions regarding this paper, which investigates the pathways of MPA glucuronidation. The interest of studying MPA metabolism in the esophagus is unclear to us, since tablets of the ester mycophenolate mofetil or enteric coated tablets of mycophenolate sodium pass very rapidly through the esophagus, the wall of which is covered with a thick mucus, so that any contact of MPA with the esophagus mucus cells is unlikely or, at most, negligible. To assess the main isoforms involved in MPA glucuronidation, the authors used recombinant UGTs expressed in COS-1, which is one of the best tools, and Western blot for the determination of UGT protein levels, allowing normalization of the activities observed, which was lacking in some previous reports. However, we think that the analytical method used (thin-layer chromatography) is poor. In particular, both identification and determination of the acyl-glucuronide of MPA (AcMPAG) would have required a more sensitive and specific method, such as liquid chromatography-mass spectrometry, liquid chromatographytandem mass spectrometry, or at least high-performance liquid chromatography-UV [Shipkova et al. (2000) previously published a rapid and sensitive method using high-performance liquid chromatography-UV for the determination of both MPAG and AcMPAG]. Moreover, in the study by Basu et al. (2004), the lack of pure AcMPAG did not allow clear identification and quantitation of this metabolite, and its determination was based on the difference in MPA levels before and after hydrolysis, which is an indirect method with poor performance. Moreover, no validation data were provided for this procedure. Basu et al. (2004) carried out incubation experiments at different pH for the different UGT isoforms. The pH of the incubation medium should in principle parallel that of the cell cytoplasm in vivo. Whereas it is acknowledged that substantial differences can be observed in the metabolic rate at different pH and that it may be useful in identification experiments, when comparing the activity of different recombinant isoforms, it is strongly recommended to use the same conditions for all the enzymes to allow unbiased comparison (Venkatakrishnan et al., 2001). Based on the linear increase of the metabolic rate over a large range of MPA concentration, the authors described the kinetics of MPA glucuronidation as atypical. “Atypical kinetics” generally refers to either enzyme autoactivation, which leads to a sigmoidal profile described by the Hill equation, or autoinhibition, resulting in a convex profile due to a decreasing Vmax at high substrate concentrations (Venkatakrishnan et al., 2001). In the present case, the profile presented could probably be adequately fitted using the Michaelis-Menten model, and the only “atypical” finding is a low Km due to a high Vmax. The high in vivo MPA-phenyl-glucuronide concentration is well known and has been described in many reports (Bullingham et al., 1998; Johnson et al., 1999), and it is obvious that this feature of MPA metabolism explains the high dosage requirement that Basu et al. (2004) state as “not understood.” For Basu et al. (2004), the postulated stimulation of UGT activity with increasing MPA concentration would translate clinically in different metabolic rates of MPA at 2or 3-g doses of MPA (which, by the way, do not correspond to actual doses as 2 g of mycophenolate mofetil correspond to 1.44 g of MPA), leading to “adverse pharmacokinetics for free MPA.” However, it is unclear to us how the authors could relate their in vitro concentration range (1.6–2.4 mM), supposed to reflect concentrations in the vicinity of the enzymes, to such oral doses. Also, such an increase in the metabolic rate of MPA with increasing dose has not been reported in the different pharmacokinetic and concentration-studies conducted so far (Bullingham et al., 1998; van Gelder et al., 1999). Finally, even if this phenomenon occurred, it would have a small impact on free-MPA pharmacokinetics, because any decrease in the binding of MPA to plasma albumin would then be compensated for by increased MPA clearance, resulting in almost unchanged free-MPA levels, as was described in pediatric renal transplant recipients, where MPA metabolic rate was high in the first weeks of treatment and strongly decreased between months 1 and 6, with no change in free-MPA concentrations and AUC (Weber et al., 1999). Finally, contrary to the authors, we think that it cannot be inferred from in vitro studies that the natural dietary curcumin will decrease MPA glucuronidation by reversible inhibition, and we doubt it would be of any clinical interest anyway.