Faculty Bibliography

INTRODUCTION: Coumadin (R/S-warfarin) is the most widely prescribed oral anticoagulant in the world; nevertheless, its clinical use is complicated by unpredictability in dose requirements to achieve and maintain optimal anticoagulation. Variations in warfarin metabolism among patients contribute to unpredictability in therapeutic responses. Studying the clinical relevance of warfarin metabolism poses a significant analytical challenge. Warfarin is given to patients as an equal mixture of R and S enantiomers. Both drugs undergo extensive metabolism through different pathways to generate > 20 structurally similar isomeric metabolites. AREAS COVERED: The article discusses how analytical methods have evolved to effectively resolve and quantify individual metabolites. The authors also discuss how the application of these methods has identified clinically relevant metabolic pathways for warfarin and fostered the investigation of clinical biomarkers for patient responses to therapy. The article additionally presents the power of these methods and how aspects of warfarin metabolism have led to the use of warfarin as a phenotyping probe for multiple drug metabolizing enzymes. EXPERT OPINION: Progress in these areas has been hampered by shortcomings in analytical methods and a narrow focus on one metabolic pathway. Recent advances in liquid chromatographic-mass spectral methods can rapidly analyze most warfarin metabolites. It is now possible to effectively assess alternate metabolic pathways and expand biomarker analyses for clinical and phenotyping applications.

Coumadin (R/S-warfarin) is a commonly prescribed anticoagulant for over approximately 20 million Americans. Although highly efficacious, positive clinical outcomes during warfarin therapy depend on maintaining a narrow therapeutic range for the drug. This goal is challenging due to large inter-individual variability in patient response, which has been attributed to diversity in drug metabolism. Warfarin is given as a racemic mixture and evidence suggest differences of R and S-warfarin in their therapeutic activities and metabolism. Previous investigation of warfarin metabolism has been hampered by the inability to quantify the individual enantiomers. To overcome this limitation a multi-mode LC-MS/MS method is reported. This strategy combines phenyl based reverse phase chromatography with chiral phase chromatography prior to quantitation by liquid chromatography tandem mass spectrometry. This approach was made possible through advances in UPLC technology producing narrow peaks suitable for transferring to a second column. The reported method separated individual R and S enantiomers of hydroxywarfarin and warfarin. All four possible isomers of 10-hydroxywarfarin were resolved to reveal unprecedented insights into the stereo-specific metabolism of warfarin. Characterization of the method demonstrated that it is robust and sensitive with inter-day coefficients of error between <7% and a detection limit of 2 nM in sample or 10 fmol on column for each analyte. Individual metabolites may be suitable surrogate biomarkers or predictive markers that predict warfarin dose, adverse interactions, or other important clinical outcomes during anticoagulant therapy. Consequently, the metabolite profiles obtained through this dual phase UPLC-MS/MS method are expected to increase our understanding of the role warfarin metabolism plays in patient response to therapy and yield new strategies to improve patient outcomes.

Effective coumadin (R/S-warfarin) therapy is complicated by inter-individual variability in metabolism. Recent studies have demonstrated that CYP3A isoforms likely contribute to patient responses and clinical outcomes. Despite a significant focus on CYP3A4, little is known about CYP3A5 and CYP3A7 metabolism of warfarin. Based on our studies, recombinant CYP3A4, CYP3A5 and CYP3A7 metabolized R- and S-warfarin to 10- and 4'-hydroxywarfarin with efficiencies that depended on the individual enzymes. For R-warfarin, CYP3A4, CYP3A7, and CYP3A5 demonstrated decreasing preference for 10-hydroxylation over 4'-hydroxylation. By contrast, there was no regioselectivity toward S-warfarin. While all enzymes preferentially metabolized R-warfarin, CYP3A4 was the most efficient at metabolizing all reactions. Individuals, namely African-Americans and children, with higher relative levels of CYP3A5 and/or CYP3A7, respectively, compared to CYP3A4 may metabolize warfarin less efficiently and thus may require lower doses and be at risk for adverse drug-drug interactions related to the contributions of the respective enzymes.

Coumadin (R/S-warfarin) anticoagulant therapy poses a risk to over 50 million Americans, in part due to interpersonal variation in drug metabolism. Consequently, it is important to understand how metabolic capacity is influenced among patients. Cytochrome P450s (P450 or CYP for a specific isoform) catalyze the first major step in warfarin metabolism to generate five hydroxywarfarins for each drug enantiomer. These primary metabolites are thought to reach at least 5-fold higher levels in plasma than warfarin. We hypothesized that hydroxywarfarins inhibit the hydroxylation of warfarin by CYP2C9, thereby limiting enzymatic capacity toward S-warfarin. To test this hypothesis, we investigated the ability of all five racemic hydroxywarfarins to block CYP2C9 activity toward S-warfarin using recombinant enzyme and human liver microsomes. We initially screened for the inhibition of CYP2C9 by hydroxywarfarins using a P450-Glo assay to determine IC(50) values for each hydroxywarfarin. Compared to the substrate, CYP2C9 bound its hydroxywarfarin products with less affinity but retained high affinity for 10- and 4'-hydroxywarfarins, products from CYP3A4 reactions. S-Warfarin steady-state inhibition studies with recombinant CYP2C9 and pooled human liver microsomes confirmed that hydroxywarfarin products from CYP reactions possess the capacity to competitively inhibit CYP2C9 with biologically relevant inhibition constants. Inhibition of CYP2C9 by 7-hydroxywarfarin may be significant given its abundance in human plasma, despite its weak affinity for the enzyme. 10-Hydroxywarfarin, which has been reported as the second most abundant plasma metabolite, was the most potent inhibitor of CYP2C9, displaying approximately 3-fold higher affinity than S-warfarin. These results indicate that hydroxywarfarin metabolites produced by CYP2C9 and other CYPs may limit metabolic capacity toward S-warfarin through competitive inhibition. Subsequent processing of hydroxywarfarins to secondary metabolites, such as hydroxywarfarin glucuronides, could suppress product feedback inhibition, and therefore could play an important role in the modulation of metabolic pathways governing warfarin inactivation and elimination.

The widely prescribed anticoagulant, Coumadin (racemic R/S-warfarin), Bristol-Myers Squibb Company, Clinton, NY has a narrow therapeutic range and wide interindividual response due, in part, to drug metabolism. Early identification of hydroxywarfarins (OHWARs), especially S-7-OHWAR, as major metabolites fostered studies characterizing cytochrome P450s responsible for those reactions. Nevertheless, phase II metabolism by sulfotransferases and, especially uridine diphosphate (UDP)-glucuronosyltransferases (UGTs), marks the next chapter in warfarin inactivation and clearance. Rodents converted OHWARs to glucuronides (O-GLUC), including high levels of 4'-, 7-, and 8-O-GLUC. Similarly, humans generated significant levels of glucuronides following treatment with warfarin. 7-O-GLUC was a major metabolite, while 6- and 8-O-GLUC were minor ones. Surprisingly, warfarin glucuronidation accounted for up to 13% of metabolites. This capacity in humans derives from several UGTs, as shown by studies with recombinant enzymes and racemic warfarin and OHWARs. 7-OHWAR was a high-affinity substrate for UGT1A1, compared to other UGTs. UGT1A1 and UGT1A10 also glucuronidated 6-OHWAR. Of five UGT1A enzymes, UGT1A10 was approximately 7-fold more efficient than the rest. Broad substrate specificity for UGT1A10 derives, in part, from an active site-binding motif, specifically F90-M91-V92-F93. Unlike glucuronidation, less is known about sulfonation of warfarin and its metabolites, except that low capacities are shown by rats and, possibly, humans. Collectively, phase I and II metabolic steps create pathways for inactivating and eliminating warfarin that require elucidation. These findings will ultimately enrich our understanding of warfarin metabolism and facilitate the interpreting of metabolic profiles of patients. This knowledge will possibly avoid complications during warfarin therapy related to metabolism by personalizing therapy for the patient.

As a step toward exploring a targeted metabolomics approach to personalized warfarin (Coumadin) therapy, we developed a liquid chromatography-tandem mass spectrometry (LC-MS/MS) method capable of quantifying specific enantiomeric (R and S) contributions of warfarin (WAR) and the corresponding hydroxywarfarins (OH-WAR) and glucuronides (-GLUC) generated by cytochrome P450s (CYP) and UDP-glucuronosyltransferases (UGTs), respectively. Evaluation of quality control samples and three commercially available human samples showed that our analytical approach has the ability to measure 24 unique WAR metabolites in human urine. Evaluation of the human data also provides new insights for evaluating WAR toxicity and begins characterizing important UGT metabolic pathways responsible for WAR detoxification. Data revealed the significance of specific metabolites among patients and the corresponding enzymatic capacity to generate these compounds, including the first report of direct WAR glucuronidation. On the basis of total OH-WAR levels, (S)-7-OH-WAR was the predominant metabolite indicating the significance of CYP2C9 in WAR metabolism, although other CYP2C enzymes also contributed to clearance of this isomer. (R)-WAR hydroxylation to OH-WARs was more diverse among the patients as reflected in varying contributions of CYP1A2 and multiple CYP2C enzymes. There was wide variation in the glucuronidation of WAR and the OH-WARs with respect to the compounds and patients. 6- and 7-OH-WAR were primarily (>70%) excreted as glucuronides unlike 4'-OH-WAR and 8-OH-WAR. For all patients, UGT1A1 is likely responsible for 6-O-GLUC production, although UGT1A10 may also contribute in one patient. 7-O-GLUC levels reflected contributions from potentially five different UGT1A enzymes. In all cases, WAR, 4'-OH-WAR, 8-OH-WAR, and the corresponding glucuronides were minor metabolites with respect to the others. Taken together, these data suggest that both P450 and UGT reactions contribute to the generation of excretable products in human urine, thereby generating complex metabolic networks.