Reduction of dimesna to mesna by the isolated perfused rat liver

Mesna is administered with ifosfamide and cyclophosphamide to reduce the incidence of hemorrhagic cystitis. In the present model of mesna metabolism and disposition, mesna is rapidly and irreversibly oxidized to dimesna in the plasma, passes unchanged through the liver, and is then reduced by the kidney and excreted. Our detection of a high ratio of mesna to dimesna in the plasma of clinical samples led us to reinvestigate the hepatic metabolism of mesna and dimesna. We perfused isolated rat livers from female Sprague Dawley rats with protein-free buffered solution containing dimesna at concentrations observed during therapy. In single-pass perfusions, each liver was perfused with up to three dimesna concentrations during consecutive 20-min periods. Recirculating perfusions were used to study single supratherapeutic concentrations of dimesna or mesna. Mesna and dimesna concentrations were measured by specific chromatographic procedures. Dimesna reduction, adjusted by the effluent flow rate and liver weight (0.4-58.5 nmol/min/g liver), correlated closely by linear regression (r = 0.98; n = 36) to the perfused dimesna concentration (4.2-249 microM), indicating a clearance of 0.20 ml/min/g liver. The concentration of dimesna that entered the liver closely matched the summed concentration of mesna and dimesna emerging in the effluent perfusate (single-pass experiments: slope, 0.98; intercept, -0.30; r = 1.00; n = 31). Only trace amounts of unidentified thiols were detected in the bile during recirculation of perfusates with 1 mM mesna or 250 microM dimesna. The effluent mesna concentration correlated inversely with the flow rate, which was consistent with a low extraction ratio in the perfusion model. These data suggested that the dimesna reduction rate was limited by hepatic uptake. Dimesna reduction was decreased by agents that deplete glutathione. Pretreatment of rats with up to 100 mg/kg ifosfamide did not impair hepatic dimesna reduction. In control experiments, dimesna was not reduced during recirculation through the apparatus without a liver. Mesna was oxidized to dimesna during oxygenation of the perfusate in the reservoir, but mesna injected directly into the perfusate just before entry into the liver passed unchanged into the effluent. Extrapolation of the dimesna clearance data from the perfusion model to humans suggests that hepatic dimesna reduction may counterbalance the rapid oxidation of mesna in plasma. The proposed equilibrium is consistent with clinical observations and suggests a new model for mesna metabolism and disposition.