Use of High Resolution Accurate Mass Spectrometry with an Exogenous Mixture of GSH and Stable Isotope-Labled GSH to Trap Reactive Metabolites of Troglitazone in Human Hepatocytes

CYP-mediated biotransformation of some drugs can lead to the formation of reactive metabolites, the binding of which to intracellular and/or extracellular macromolecules may be associated with idiosyncratic toxicity. Troglitazone (TGZ) is one such drug and is frequently used as a model compound in studies of reactive metabolites, most of which have been conducted with subcellular fractions such as liver microsomes rather than intact cell-based systems. In this study, metabolites of TGZ were generated by incubation with cryopreserved human hepatocytes in suspension and trapped in the presence of an exogenous mixture of unlabeled and stable isotope-labeled (SIL; [1,2 13C,15N]-glycine) glutathione (GSH/SIL-GSH). Samples were analyzed using liquid chromatography-high resolution accurate mass spectrometry (LC-HRAMS) on an LTQ Orbitrap instrument. GSH conjugates (trapped reactive metabolites) were detected based on a characteristic mono-isotopic pattern (peaks separated by 3.0037 mass units). A total of fi ve conjugates, two of which have not previously been described in the literature, were detected and partially characterized based on accurate mass measurements and MS/MS data. The addition of exogenous GSH/SIL-GSH increased the amount of GSH conjugates formed; the magnitude of the effect was conjugate-specifi c. Furthermore, the mono-isotopic intensity ratios of the conjugates varied from 1:0.15 (apparently for conjugates formed mostly inside the cells, where there was no SIL-GSH) to near the 1:1 ratio applied (for conjugates formed mostly outside the cells). Results obtained with hepatocytes from different donors suggest that not all GSH conjugates are formed in all cases, consistent with the idiosyncratic nature of drug-induced liver toxicity. The results of experiments in which hepatocytes were exposed to exogenous SIL-GSH, then separated from the medium by centrifugation through oil, clearly indicate that most conjugation with exogenous GSH occurs extracellularly. The fi nding that GSH conjugates are formed extracellularly raises the question of which transporter(s) is/are involved in the effl ux of TGZ and its metabolites. It also suggests a facile mechanism by which reactive drug metabolites could be presented on the extracellular surface of liver cells, leading to the formation of immunogenic neoantigens as has been proposed by others [1,2]. Materials and Methods All solvents (from VWR) were HPLC grade. Troglitazone (TGZ) was purchased from Toronto Research Chemicals. Unlabeled glutathione (GSH, 99% pure) was from Sigma, and stable isotope-labeled (SIL)-GSH ([1,2-13C, 15N]glycine, isotopic purity >90%), was from Cambridge Isotope Laboratories. Incubation, Sample Preparation, and Bioanalysis Human cryopreserved hepatocytes (mixed gender, pool of three donors) were obtained from XenoTech. The hepatocytes were thawed and prepared according to the vendor’s instructions, pooled into Krebs Henseleit buffer (pH 7.4) at 1.5x106 cells/mL, and after addition of either SIL-GSH or a mixture of GSH/SIL-GSH (1:1) to a fi nal concentration of 5 mM, were kept on ice prior to the experiments. Incubation was initiated by adding TGZ (50 μM fi nal) in DMSO, or the same volume of DMSO alone (solvent control); fi nal organic content 0.1%. Reaction tubes were mixed briefl y and incubated in a shaking water bath at 37°C. Aliquots (400 μL) of samples containing SIL-GSH only were taken at 5, 10, 30, and 120 min (n=2) and centrifuged through oil to separate cells and supernatant. The supernatant was removed and treated with MeCN (plus warfarin as internal standard [IS]); the remainder of the sample was frozen ( 80°C) and the pellet was recovered and treated with MeCN (containing IS). After centrifugation, clear supernatants were transferred, evaporated to dryness in a stream of nitrogen, and reconstituted in 30% MeOH (40 μL). After centrifugation, aliquots were transferred for sample analysis using LC-HRAMS. Control samples, containing a mixture of GSH/SIL-GSH with and without TGZ or TGZ only, were taken at t=120 min, and treated as described above. LC-MS Analysis Samples were analyzed using an RP-LC coupled with High Resolution Accurate Mass Spectrometry (HRAMS) detection implemented on LTQ-Orbitrap XL instrument in positive mode. The HRAMS survey scan was 300-800 Th at resolution 30,000. The manufacturer’s default settings were used for automatic gain control. MetWorks software (Thermo) was used for LC-HRAMS data processing. Results and Discussion The metabolism of TGZ, leading to reactive metabolites, has been studied and reported previously in microsomal preparations (e.g., [3]). Our recent paper [4] describes the application of a GSH/SIL-GSH mixture for trapping reactive metabolites of TGZ in cryopreserved hepatocytes. We also hypothesized two mechanisms responsible for the observed non-uniform incorporation of exogenous GSH in the detected conjugates: 1) spatial differences in both uptake of exogenous GSH mixture and formation/trapping of reactive metabolites; 2) different effl ux rates of reactive metabolites, resulting in different levels of extracellular trapping. Here we report preliminary evidence in support of the latter mechanism. Figure 1 (left-hand panel) shows extracted ion chromatograms (XICs) of GSH conjugate M6 detected in a t=120 min sample. In the SIL-GSH-supplemented sample, after oil fi ltration the molecular ion of M6 was detected as the higher-mass SIL-GSH conjugate (764.2566; +3.0037) only in the extracellular medium (panel D), and the molecular ion of unlabeled M6 (761.2511, conjugated with endogenous GSH) was not detected in either fraction. In fact, no conjugates with endogenous GSH were detected in the cell pellet (panel C). These results could be elaborated via effl ux of the precursor of M6 followed by its trapping with extracellular GSH. However, both SIL and unlabeled M6 can be seen in the GSH/SIL-GSH-supplemented total sample (B); only unlabeled M6 was observed in the non-supplemented total sample (A). The unlabeled M6 (conjugated with endogenous GSH) in panel A (and a portion of that in panel B) most likely forms during sample processing Results