Stereospecific metabolism of R- and S-warfarin by human hepatic cytosolic reductases

Article abstract of DOI:10.1124/dmd.117.075929

Coumadin (rac-warfarin) is the most commonly used anticoagulant in the world; however, its clinical use is often challenging because of its narrow therapeutic range and interindividual variations in response. A critical contributor to the uncertainty is variability in warfarin metabolism, which includes mostly oxidative but also reductive pathways. Reduction of each warfarin enantiomer yields two warfarin alcohol isomers, and the corresponding four alcohols retain varying levels of anticoagulant activity. Studies on the kinetics of warfarin reduction have often lacked resolution of parent-drug enantiomers and have suffered from coelution of pairs of alcohol metabolites; thus, those studies have not established the importance of individual stereospecific reductive pathways. We report the first steady-state analysis of R- and S-warfarin reduction in vitro by pooled human liver cytosol. As determined by authentic standards, the major metabolites were 9R,11S-warfarin alcohol for R-warfarin and 9S,11S-warfarin alcohol for S-warfarin. R-warfarin (V_max 150 pmol/mg per minute, K_m 0.67 mM) was reduced more efficiently than S-warfarin (V_max 27 pmol/mg per minute, K_m 1.7 mM). Based on inhibitor phenotyping, carbonyl reductase-1 dominated R-and S-warfarin reduction, followed by aldo-keto reductase-1C3 and then other members of that family. Overall, the carbonyl at position 11 undergoes stereospecific reduction by multiple enzymes to form the S alcohol for both drug enantiomers, yet R-warfarin undergoes reduction preferentially. This knowledge will aid in assessing the relative importance of reductive pathways for R- and S-warfarin and factors influencing levels of pharmacologically active parent drugs and metabolites, thus impacting patient dose responses.

Full text of DOI:10.1124/dmd.117.075929

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Title Page
Stereospecific Metabolism of R- and S-Warfarin by Human Hepatic Cytosolic Reductases
Barnette, Dustyn A., Johnson, Bryce P., Pouncey, Dakota L., Nshimiyimana, Robert.
Desrochers, Linda P., Goodwin, Thomas E., Miller, Grover P.
Department of Biochemistry and Molecular Biology, University of Arkansas for Medical
Sciences, Little Rock, AR 72205, USA (DAB, DLP, GPM)
Department of Chemistry, University of Central Arkansas, Conway, AR 72035, USA (BPJ)
Department of Chemistry, Hendrix College, Conway, AR 72032, USA (RN, LPD, TEG)
1

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Running Title Page
Stereospecific R- and S-Warfarin Reduction in the Cytosol
Address for Correspondence
Grover P. Miller, PhD
Department of Biochemistry and Molecular Biology
University of Arkansas for Medical Sciences
4301 W. Markham, Slot 516.
Little Rock, AR 72205, USA;
Telephone: 501.526.6486;
Fax: 501.686.8169;
Email: MillerGroverP@uams.edu
Number of Text Pages: 14
Number of Tables: 1
Number of Figures: 6
Number of References: 39
Number of words in the Abstract: 239
Number of words in the Introduction: 700
Number of words in the Discussion: 1217
Abbreviations
WAR, warfarin; WAROH, warfarin alcohol; HOWAR, hydroxywarfarin; FFA, flufenamic acid;
INDO, indomethacin; QUE, quercetin; AKR, aldo-keto reductase; CBR, carbonyl reductase
2

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Abstract
Coumadin (rac-warfarin) is the most commonly used anticoagulant in the world, yet its clinical
use is often challenging due to a narrow therapeutic range and inter-individual variations in
response. A critical contributor to the uncertainty is variability in warfarin metabolism, which
includes mostly oxidative but also reductive pathways. Reduction of each warfarin enantiomer
yields two warfarin alcohol isomers, and the corresponding four alcohols retain varying levels of
anticoagulant activity. Studies on the kinetics of warfarin reduction have often lacked resolution
of parent drug enantiomers and suffered from co-elution of pairs of alcohol metabolites; thus,
those studies fail to establish the importance of individual, stereospecific reductive pathways.
We report the first steady-state analysis of R- and S-warfarin reduction in vitro by pooled human
liver cytosol. As determined by authentic standards, the major metabolites were 9R,11S-
warfarin alcohol for R-warfarin, and 9S,11S-warfarin alcohol for S-warfarin. R-warfarin (Vmax
150 pmol/mg/min, Km 0.67 mM) was reduced more efficiently than S-warfarin (Vmax 27
pmol/mg/min, Km 1.7 mM). Based on inhibitor phenotyping, carbonyl reductase-1 dominated R-
and S-warfarin reduction followed by aldo-keto reductase-1C3 and then other members of that
family. Overall, the carbonyl at position 11 undergoes stereospecific reduction by multiple
enzymes to form the S alcohol for both drug enantiomers, yet R-warfarin undergoes reduction
preferentially. This knowledge will aid in assessing the relative importance of reductive
pathways for R- and S-warfarin and factors influencing levels of pharmacologically active parent
drugs and metabolites thus impacting patient dose-responses.
3

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Introduction
Coumadin is a commonly used blood thinner worldwide that consists of a 1:1 mixture of R-
and S-warfarin (Figure 1) (Ansell et al., 2008). The anticoagulant effect of warfarin reflects
inhibition of VKORC, which plays a critical role in vitamin K cycling and the maturation of
coagulation factors. Though very efficacious, warfarin can be difficult to optimally treat patients
due to its narrow therapeutic range and inter-individual variations in response. Doses have to be
individually established for each patient based on how they respond to the medication. Failure
to achieve and maintain an optimal dose for patients increases risks for potentially life-
threatening adverse events due to under dosing (strokes) and over dosing (bleeding) (Wysowski
et al., 2007). Knowledge of the causes of this variability in dose-responses can then lead to
better management strategies.
Inter-individual variations in patient dose-responses to warfarin often reflect the importance
of drug metabolism(Jones and Miller, 2011). Cytochromes P450 oxidize R- and S-warfarin into
hydroxywarfarins, and as major routes for drug clearance, those metabolic pathways have
dominated the focus of most studies. Through minor, less characterized pathways, both drugs
also undergo reduction at the carbonyl at position 11 to yield warfarin alcohols. Warfarin
reduction introduces a second chiral center, making two of warfarin alcohol diastereomers for
each warfarin enantiomer. For R-warfarin, the two metabolites are 9R,11R- or 9R,11S-warfarin
alcohol (RR- or RS-warfarin alcohol), and for S-warfarin, they are 9S,11S- and 9S,11R- warfarin
alcohol (SS- or SR-warfarin alcohol). Warfarin alcohols retain pharmacological activity that is 6-
fold lower than that for the parent drugs (Gebauer, 2007); however, clearance of RS-warfarin
alcohol is 2.5-fold lower than the other alcohols indicating selective accumulation of the
metabolite in plasma based on stereochemistry (Lewis, 1973), which increases their relative
potency during maintenance dosing (Haque et al., 2014). According to a previous report
(Haque et al., 2014), warfarin alcohols are the most potent of the warfarin metabolites,
4

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
contributing 15% of the anticoagulant effect of S-warfarin for the average patient at a
maintenance dosing. Consequently, knowledge of the impact of stereochemistry on R- and S-
warfarin reduction is critical for adequately assessing the clinical relevance of those pathways.
Recent studies using subcellular fractions from rats (Alshogran et al., 2015a) and humans
(Malátková et al., 2016) have provided limited insights on reductive pathways for warfarin due to
reliance on rac-warfarin and not the actual drug enantiomers as a substrate and the subsequent
lack of chiral resolution of metabolites from the reaction. First, the drug is administered as a 1:1
mixture of enantiomers, but those conditions do not reflect patient serum levels and mask actual
clearance of individual warfarin enantiomers. Reported R-warfarin levels in patients are on
average two-fold higher than S-warfarin(Breckenridge et al., 1974; Jones et al., 2010; Haque et
al., 2014). Knowledge of the metabolic kinetics for individual drug enantiomers would provide
perspective for assessing their relative significance in drug clearance. Moreover, the use of a
mixture of substrates in those studies masks possible drug-drug interactions between the
enantiomers, which were implicated in one of the studies(Malátková et al., 2016). Second, it is
not possible to extrapolate the metabolic kinetics for R- and S-warfarin from those for rac-
warfarin so that knowledge of the impact of stereochemistry on metabolic clearance of the drugs
remains unknown. Third, current studies on warfarin reduction failed to resolve all four
diastereomers generated during rac-warfarin reduction. Instead, reported kinetic parameters
reflect the formation of mixtures of metabolites, i.e. alcohol 1 (RR- and SS-warfarin alcohols)
and alcohol 2 (RS- and SR-warfarin alcohols), generated during rac-warfarin metabolism, and
hence provide no information on the kinetics for individual pathways leading to specific
metabolites. Taken together, the current findings provide only a qualitative perspective on
warfarin reduction and lack specific quantitative details on the individual pathways that would
determine their potential clinical relevance.
5

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Herein, we report the impact of stereochemistry on the efficiency of rac-, R-, and S- warfarin
reduction by pooled human liver cytosolic (HLC150) cell fractions. Due to the lack of
commercially available standards, we synthesized and purified all four diastereomeric warfarin
alcohols using green chemistry techniques. We then optimized reaction conditions to achieve
steady state conditions and carried out kinetic studies with R- and S-warfarin along with rac-
warfarin for comparison to findings from other studies. We assessed potential interactions
between the drug enantiomers by simulating rac-warfarin metabolism using parameters for the
individual drugs and compared them to those measured for rac-warfarin in our study and kinetic
values reported in the literature that used a limited pool of cytosol fraction from five cadavers
(Malátková et al., 2016). Lastly, we employed inhibitor phenotyping to identify possible
reductases responsible for warfarin metabolism in cytosolic fractions.
Materials and Methods
Materials. HPLC grade methanol, acetic acid, dimethyl sulfoxide (DMSO), and NADPH were
purchased from Thermo Fisher (Pittsburgh, PA) and pooled human liver cytosol (HLC150) from
BD Biosciences. Diethyl ether, hydrochloric acid, ethyl acetate, acetone, and acetic acid were
obtained from PHARMCO-AAPER (Brookfield, CT). Sodium borohydride and DMS0-d6 was
purchased from Sigma-Aldrich (St. Louis, MO). Rac-Warfarin, warfarin alcohol mixed isomers,
and hydroxywarfarins (internal standards) were obtained from Toronto Research Chemicals
(Toronto, Canada). Individual R- and S-warfarin stereoisomers were prepared in a previous
study(Preston Pugh et al., 2014), while the metabolites 9R,11R-; 9R,11S-; 9S,11R-; and
9S,11S-warfarin alcohols were synthesized and purified to >98% purity.
Synthesis of warfarin alcohols through green chemistry. We prepared authentic isomers of
all four diastereomeric warfarin alcohols by synthesizing R- and S-warfarin and then reducing
them individually followed by isolating the individual alcohols by thin layer chromatography and
cycles of recrystallization. First, R- and S-warfarin were prepared with greater than 99%
6

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
enantiomeric excess using a green chemistry procedure as described by us (Preston Pugh et
al., 2014) and others (Kim et al., 2006; Wong et al., 2010). Second, reduction of R- and S-
warfarin to warfarin alcohols followed a literature procedure (Trager, 1970; Chan et al., 1972).
The R- or S-warfarin (50 mg, 1.62 mmol) was stirred in one mL of deionized water with sodium
borohydride (7 mg, 1.85 mmol) at the ambient temperature for 2 hr. Excess sodium borohydride
was neutralized by addition of one mL of 0.5 M hydrochloric acid added dropwise with stirring.
The mixture was filtered and washed with water. The resulting white solid was allowed to air
dry. The reduction step for each warfarin enantiomer introduced a new chiral center resulting in
a pair of diastereomers, i.e. RS and RR warfarin alcohols for R-warfarin and the SS- and SR-
warfarin alcohols for S-warfarin. The typical yield of each pair of diastereomers was greater
than 70%.
Isolation of individual alcohol diasteromers was achieved through silica gel thin layer
chromatography (TLC) followed by recrystallization. For example, a diastereomer mixture of
two warfarin alcohols (either RS and RR, or SR and SS) was dissolved in the smallest amount
of ethyl acetate and dropped onto a Sorbtech Silica XHL TLC plate W/UV254 (glass backed,
250 µm thickness, 20 x 20 cm dimensions). A solvent mix of 197:3 diethyl ether:acetic acid was
used to effect a separation. The bottom bands of each separation were the RR–and SS-
warfarin alcohols, while the top bands were the RS and SR–warfarin alcohols. TLC identity of
the aforementioned diastereomers was based upon the previous research by others
(Supplemental Data, (Chan et al., 1972)). Each band was carefully scraped from the glass TLC
plate and ethyl acetate was added to dissolve the alcohols. After stirring, the mixture was
filtered over Celite and ethyl acetate was removed by rotary evaporation. The white solid was
dissolved in the smallest amount of warm acetone, and then drops of warm water were added
slowly until the solution remained cloudy. At this point the solution was heated again until clear
and allowed to cool at the ambient temperature until crystals appeared, at which point the
7

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
solution was cooled in ice. When crystallization ceased, the supernatant liquid was removed by
pipet and the solid allowed to air dry at ambient temperature.
NMR peak assignments were made with 1H, 13C, COSY, DEPT, HMQC, and HMBC
techniques. RR-Warfarin alcohol: 1H NMR (DMSO-d6, 400 MHz): δ 1.21 (3H, d, J = 6.4 Hz, H-
14), 2.40 (2H, m, H-12a, H-12b), 3.66 (1H, ddq, J = 6.0, 6.4, 6.4 Hz, H-13), 4.62 (1H, dd, J =
7.8, 8.2 Hz, H-11), 7.13 (1H, dd, J = 7.3, 7.3 Hz, H-4’), 7.24 (2H, dd, J = 7.3, 7.3 Hz, H-3’, H-3’),
7.28 (1H, dd, J = 0.9, 8.1 Hz, H-8), 7.33 (1H, ddd, J = 0.9, 7.4, 7.8 Hz, H-6), 7.46 (2H, d, J = 7.3
Hz, H-2’, H-2’), 7.56 (1H, ddd, J = 1.4, 7.4, 8.1 Hz, H-7), 7.93 (1H, dd, J = 1.4, 7.8 Hz, H-5); 13C
NMR (DMSO-d6, 400 MHz): δ 23.9 (C-14), 38.5 (C-11), 41.6 (C-12), 67.2 (C-13), 110.2 (C-3),
117.4 (C-8), 117.8 (C-10), 124.1 (C-5), 125.1 (C-6), 127.2 (C-4’), 129 (C-3’), 129.3 (C-2’), 132.9
(C-7), 144.2 (C-1’), 153.8 (C-9), 162.7 (C-4), 165.1 (C-2); RS-Warfarin alcohol: 1H NMR
(DMSO-d6, 400 MHz): δ 1.22 (3H, d, J = 6.4 Hz, H-14), 2.09 (1H, ddd, J = 4.8, 9.4, 14.0 Hz, H-
12b), 2.60 (1H, ddd, J = 3.2, 11.1, 14.0 Hz, H-12a), 3.66 (1H, m, H-13), 4.67 (1H, dd, J = 4.8,
11.1 Hz, H-11), 7.12 (1H, dd, J = 7.3 Hz, H- 4’), 7.22 (2H, dd, J = 7.3, 7.8 Hz, H-3’), 7.27 (1H, d,
J = 7.9 Hz, H-8), 7.32 (1H, dd, J = 7.9, 8.0 Hz, H-6), 7.39 (2H, d, J = 7.8 Hz, H-2’), 7.55 (1H,
ddd, J = 1.4, 7.9, 7.9 Hz, H-7), 7.95 (1H, dd, J = 1.4, 8.0 Hz, H-5); 13C NMR (DMSO-d6, 400
MHz): δ 24.1 (C-14), 38.3 (C-11), 41 (C-12), 67.1 (C-13), 108.7 (C-3), 117.3 (C-8), 118 (C-10),
124.4 (C-5), 125.1 (C-6), 127 (C-4’), 128.9 (C-3’), 129 (C-2’), 132.9 (C-7), 144.7 (C-1’), 154 (C-
9), 164 (C-4), 164.9 (C-2).
Identification of initial rate conditions for warfarin reduction. Prior to steady-state studies,
control experiments with HLC150 were carried out to determine optimal co-solvents, reaction
buffers, reaction time, and protein concentration to ensure steady-state conditions and avoid
confounding effects from experimental reagents. Warfarin stocks were prepared in methanol
solutions to minimize co-solvent effects on reductase activity, as shown in our control
experiments (see Results). Individual warfarin reactions were prepared out with 0.25 mg/mL
pooled human liver cytosol (HLC150) from 150 donors, and initiated upon addition of 1 mM
8

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
NADPH (final). Reactions were incubated at 37°C with 350 rpm rotation. After 30 min,
reactions were quenched with an equal volume of acetonitrile containing, 1 µM rac-8-
hydroxywarfarin (final), an internal standard, which did not co-elute with warfarin or the warfarin
alcohols. Quenched reactions were centrifuged at 5°C at 2500 rpm for 15 min, and the
supernatant was removed for analysis by HPLC. Unless otherwise stated, all control reactions
were carried out using rac-warfarin and experimental effects monitored by measuring total
product formation by combining areas of alcohols 1 and 2 to take all four alcohol diastereomers
into account. All control reactions were carried out in duplicate.
Analysis of warfarin reduction by HPLC. The resolution and quantitation of individual drugs
and metabolites was achieved through high performance liquid chromatography (HPLC).
Samples were injected onto a Waters HPLC Breeze system equipped with a 4.6 x 150 mm
Zorbax Eclipse 3.5 µM XDB-C18 column (Agilent) heated to 45°C and analytes were resolved
over 13 min using a gradient method with 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic
acid (HEPES) /NaOH solution pH 6.80 and 25:75 acetonitrile:methanol and monitored by
fluorescence (excitation: 325 nm; emission: 393 nm). Peak areas were normalized to the
internal standard and then quantitated relative to known standards. Reaction rates were
reported as total product formed per reaction time per amount of protein (pmoles/min/mg
protein).
Steady state kinetics for R-, and S-, and rac-warfarin reduction. After optimizing reaction
conditions, we determined reaction kinetics for R-, and S-, and rac-warfarin. Substrate
concentrations were 1000, 500, 250, 200, 125, 100, 50, and 25 µM initially. S-Warfarin reactions
contained an additional reaction with 750 µM substrate. Reactions were performed in at least
triplicate for each substrate concentration. Reaction products were separated and analyzed by
HPLC and fluorescence detection as described previously. Alcohol peaks 1 and 2 or individual
alcohol diastereomers were quantitated and assessed individually. Warfarin alcohol formation
9

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
rates were plotted as a function of substrate concentration and analyzed by comparing the fit of
the data to the Michaelis-Menten equation (hyperbolic curve) and Hill equation (non-hyperbolic
curve) using in GraphPad Prism (San Diego, CA). The best fit was determined using extra sum-
of-squares F test.
Inhibitor phenotyping of R- and S-warfarin reduction. HLC150 phenotyping reactions were
carried out using the steady state conditions established in the control experiments (see
Results). We prepared all inhibitor stocks in methanol and assessed their effects against a
control reaction containing just methanol at the same final percent (5% methanol). R- or S-
Warfarin concentrations at 50 and 500 µM were used in the presence of three different inhibitors
to selectively inhibit individual and classes of cytosolic reductases (as reviewed in (Malátková
and Wsól, 2014)). Moreover, with broadly specific inhibitors, we employed concentrations about
5-fold above the reported IC50 to ensure target inhibition while minimizing off target effects. We
used 2 µM flufenamic acid (FFA), which inhibits all Aldo-keto Reductase family 1C members
(AKR1C). This concentration has also been shown to inhibit AKR1B10 activity (Zhang et al.,
2013); however, levels are typically too low to be relevant for our studies, and induction was
observed only in liver cancer (Laffin and Petrash, 2012). For indomethacin (IND), a high
concentration (180 µM) would similarly inhibit all AKR1C members and Carbonyl Reductase 1
(CBR1), while a lower concentration (5 µM) has a more selective effect on AKR1C3 with only
limited CBR1 inhibition expected. We targeted Carbonyl Reductase 1 (CBR1) activity for
inhibition using 60 µM quercetin. In addition, we combined inhibitors to assess the collective
contributions of AKR1C and CBR1 enzymes. AKR1B1 is also inhibited by indomethacin and
quercetin (Byrns and Penning, 2009), but this enzyme is only expressed at minimal levels in the
liver, except in cases of alcoholic liver disease (O'CONNOR et al., 1999). Thus, we do not
expect it to significantly contribute to metabolism in our study.
Results
10

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
HLC150 steady state kinetics for R-, S-, and rac-warfarin reduction. Initially, we carried out
control experiments to identify optimal conditions for steady-state studies as shown in Figure 2.
Warfarin reduction was more efficient in potassium phosphate over Tris-HCl buffer, and MgCl2
had no impact on reaction rates (Panel A). Variations in methanol co-solvent levels had no
effect on activity while higher DMSO levels inhibited the reaction (Panel B). Observed warfarin
alcohol levels increased essentially linearly with total protein concentration (Panel C). Over the
course of 60 min, product formation levels were essentially linear (Panel D). Based on these
studies, the standard conditions for a 30 min reaction with 0.25 mg/mL protein was 50 mM
potassium phosphate 7.4, no MgCl2, and 1% methanol as a co-solvent.
Stereospecificity played a significant role in the steady state reduction of warfarin
enantiomers and corresponding products from cytosolic reactions (Figure 3, Table 1). The R-
warfarin reactions yielded RS-warfarin alcohol as the major metabolite and RR-warfarin alcohol
as the minor one. Both reactions were best fit to a simple Michaelis-Menten kinetic model as
compared to a non-hyperbolic curve described by the Hill equation. The metabolic efficiency for
formation of the RS-warfarin alcohol was nearly 200-fold compared to the RR-warfarin alcohol
due almost exclusively the difference in rate of the reaction. Similarly, the reduction of S-
warfarin yielded the analogous metabolites, i.e. SS-warfarin alcohol and SR-warfarin alcohol.
The kinetic profile for SS-warfarin alcohol was fit to the Michaelis-Menten model, yet the
predicted Km was greater than the highest substrate concentration used in this study (1000 µM)
due to solubility limits. Thus, there is more error in the kinetic constants than predicted by a fit,
hence the exclusion of standard error from the Table. The SR-warfarin alcohol was consistently
measurable but remained at the limit of quantitation so that it was not possible to determine the
kinetics for its formation. Taken together, the rate of S-warfarin conversion to SS-warfarin
alcohol was about 9-fold lower than that for R-warfarin to RS-warfarin alcohol at 1000 µM,
11

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
indicating a strong preference for R-warfarin reduction and formation of S alcohols regardless of
substrate chirality.
In addition, we followed the experimental design reported by others(Alshogran et al., 2015a;
Malátková et al., 2016) to carry out steady-state studies with rac-warfarin for comparative
purposes. Unlike the previous experiments in this study, substrate concentration reflected an
initial 1:1 mixture of R- and S-warfarin, and the corresponding diasteromeric metabolites were
not resolved completely. We used the previously reported nomenclature(Alshogran et al.,
2015a; Malátková et al., 2016) labeling the mixture of RS/SR-warfarin alcohols as warfarin
alcohol 2 and the mixture of RR/SS-warfarin alcohols as warfarin alcohol 1. Both sets of data
were fit to the Michaelis-Menten model indicating the eight-fold difference in efficiency for the
alcohol 2 reaction was due mainly to Vmax and not Km for this mixed metabolite assessment of
activity (Figure 4, Panel C; Table 1). The lack of resolution of the substrates or metabolites with
this approach made it impossible to identify the actual major and minor metabolites for R- and
S-warfarin without relying on our initial enantiospecific steady-state kinetic studies.
Inhibitor phenotyping of cytosolic enzymes responsible for R- and S- warfarin reduction.
General and relatively specific reductase inhibitors indicated multiple enzymes likely contribute
to overall reduction of R- and S-warfarin in the cytosol. The overall patterns of inhibition were
similar between the formation of the corresponding major metabolites RS-warfarin alcohol and
SS-warfarin alcohol (Figure 5). The minor metabolites RR-warfarin alcohol and SR-warfarin
alcohol were measured at levels too low for quantitative kinetics. Substrate concentrations had
little effect on the patterns of inhibition for the compounds with two exceptions. First, the effect
of quercetin inhibition was consistently significant but reduced for S-warfarin under low
substrate conditions. Second, inhibition by low indomethacin concentration (5 µM) was reduced
at lower concentration for both warfarin enantiomers. By 180 µM indomethacin, all reduction
rates were decreased greater than 10-fold. These findings suggest a dominant role for AKR1C3
12

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
in the reactions and possible contributions from other AKR1C members and CBR1. Quercetin
mainly inhibits CBR1, and when present, reduction rates decreased by around 70-85% for RS-
and SS-warfarin alcohols. Addition of 5 µM indomethacin to the quercetin reaction did not
significantly impact the observed reduction rates; however by180 µM indomethacin there was a
further approximate 10% decrease in rates. Unlike indomethacin, flufenamic acid had little
effect on reduction, despite both being previously shown to inhibit AKR1C family enzymes
(Malátková and Wsól, 2014). This difference could be due to variability in AKR1C inhibition by
flufenamic acid depending on the substrate being metabolized, as shown in previous studies
(Malátková and Wsól, 2014). Overall, these results suggest that AKR1C family enzymes and
CBR1 participate in the reduction of R- and S-warfarin.
Discussion
Herein, we demonstrated that stereospecificity played an important role in the efficiency of
warfarin reduction and the corresponding metabolites generated in the process. This metabolic
pathway is low affinity and high capacity, although it is likely underutilized under in vivo
conditions. These insights from the kinetics differ from previous reports relying on rac-warfarin
and a failure to resolve parent drugs and their metabolites in the analyses. We replicated the
findings of others and demonstrated how the experimental design confounds an accurate
understanding of R- and S-warfarin reduction. Our subsequent phenotyping studies revealed
the probable cytosolic reductases responsible for stereospecific metabolism of warfarin.
Collectively, these findings provide a strong foundation for further studies, such as genotyping of
reductases, to explore the clinical relevance of R- and S-warfarin in patients undergoing
anticoagulant therapy.
Our data show that human cytosolic reductases more efficiently metabolized R-warfarin over
S-warfarin and preferentially generate S alcohols in the process, which is consistent with
previous studies involving both humans and animal models (Lewis, 1974; Hermans and
13

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Thijssen, 1992). According to our kinetics, warfarin reduction is a low affinity but high capacity
metabolic pathway that is likely underutilized under in vivo conditions. Maintenance dosing to
achieve a targeted therapeutic response results in low micromolar levels of R- and S-warfarin in
plasma (1518 to 4500 nM as compiled in (Haque et al., 2014)). The amounts accessible for
reduction are likely about 50-fold lower (mid-nanomolar range) due to high plasma protein
binding (O’Reilly, 1986). A possible limitation of this estimation is the lack of consideration of
warfarin accumulation in the liver. Studies with rats have shown liver warfarin levels are
possibly three-fold higher than that observed in plasma (Levy et al., 2003; Sun et al., 2015).
The similarities in physiology between rats and humans suggest that warfarin accumulation in
liver is possibly occurring in humans. Even taking this effect into consideration, in vivo levels of
R- and S-warfarin would not approach saturation, so that their reductive elimination would
present as essentially linear kinetics in patients.
Steady-state reduction of rac-warfarin by human liver cytosol yielded similar results to that
reported by others (Moreland and Hewick, 1975; Alshogran et al., 2015b; Malátková et al.,
2016), yet we were able to provide a more in-depth analysis of the data given our kinetics for
the individual drug enantiomers. Despite the use of different human liver cytosolic fractions,
warfarin alcohol 2 was the major metabolite while warfarin alcohol 1 was the minor one. The
Vmax and Km values for the alcohol 2 reaction were nearly identical between the studies, as
shown in Table 1. The corresponding values for alcohol 1 were within two-fold of one another;
that difference could be due to error in the noise in data for the minor pathway. In our case, we
can identify likely metabolites contributing to observed kinetic profiles for warfarin alcohols 1 and
2. The RS-warfarin alcohol is the primary contributor to the alcohol 2 and SS-warfarin alcohol to
the alcohol 1. SR- and RR-Warfarin alcohols would still contribute to observed mixtures of
alcohols as metabolites of the reaction, such that the observed kinetics for rac-warfarin do not
represent that for the individual warfarin enantiomers.
14

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Based on the identification of the metabolites, we assessed potential interactions between
the warfarin enantiomers that could occur when rac-warfarin served as a substrate. For
comparative purposes, we predicted the apparent alcohol 1 and 2 formation rates for rac-
warfarin if the R- and S-warfarin did not interact. Under those conditions, observed rates would
simply be the summation of the corresponding rates for metabolites contributing to alcohol 1
(RR- and SS-warfarin alcohols) and alcohol 2 (RS- and SR-warfarin alcohols) taken from our
steady-state data for the reactions with the individual drug enantiomers. As shown in Figure 6,
the predicted formation rates for warfarin alcohols from rac-warfarin were calculated by
summing individual warfarin alcohol rates from warfarin enantiomers as described previously.
The predicted kinetic profiles significantly differ from the experimentally measured values as
shown by the dashed lines versus the solid lines. When both drug enantiomers are present, the
apparent Vmax and Km are lower indicating R- and S-warfarin interact to impact their reduction to
the corresponding alcohols, as has been suggested previously (Chan et al., 1972).
Nevertheless, the quantitation of mixtures of the alcohol metabolites precludes the ability to
identify the mechanism of interaction and thus understand its impact on major and minor
pathways for R- and S-warfarin reduction. These findings further underscore the complications
for relying on rac-warfarin metabolic kinetics to understand its relevance to human metabolism
of the drug.
The inhibitor phenotyping indicated that AKR1C3 and CBR1 dominated reductive
metabolism of R- and S-warfarin in the cytosol, while AKR1C1/2 may have made minor
contributions. These major enzymes shared similar preference for R-warfarin over S-warfarin
as a substrate and generated S alcohols regardless of substrate chirality. This stereospecificity
for the cytosolic reductases results in the overall preference formation of the major RS-warfarin
metabolite followed by the minor SS-warfarin metabolite with trace levels of the other warfarin
alcohols. Carbonyl reducing enzymes demonstrate stereospecificity that depends on substrate
15

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
and the particular enzyme (Hoffmann and Maser, 2007). CBR1 substrates range in molecular
mass from 207 g/mol for 4-methylnitrosamino-1-(3-pyridyl)-1-butanone (NNK) to 543 g/mol for
doxorubicin (Malatkova et al., 2010), which includes the mass of warfarin (308 g/mol). The
enantiospecificity of these reactions has only been reported for NNK in which there is a more
than 90% preferential formation of the S-alcohol (Breyer-Pfaff et al., 2004), as observed for R-
and S-warfarin reduction in this study. Nevertheless, stereospecificity does vary among
individual members of the AKR1C family (Rižner and Penning, 2014).
As further evidence for their role in R- and S-warfarin metabolism, recombinant AKR1C3
and CBR1 readily catalyzed the reduction of rac-warfarin compared to AKR1B1, 1C1, 1C2, and
1C4 along with CBR3 (Malátková et al., 2016); however, the use of a mixture of R- and S-
warfarin and lack of resolution of substrates and metabolites obscured the stereospecificity of
substrate recognition and metabolite formation by the reductases. When considered with
respect to our observations, the importance and consequence of AKR1C3 and CBR1 in R- and
S-warfarin becomes more clearly the cause for preferential metabolism of R-warfarin and
formation of its RS-warfarin alcohol metabolite as major reductive pathways in humans.
Concluding Remarks. The kinetics for R- and S-warfarin metabolism by these reductases
could impact clearance of the parent drugs and formation of pharmacologically active
metabolites in patients undergoing anticoagulant therapy. For the average patient, R-warfarin
will primarily undergo reduction compared to S-warfarin and thus impact its contributions to
anticoagulation (Breckenridge et al., 1974) as well as the resulting less pharmacologically active
warfarin alcohol metabolites (Lewis, 1973). The relevance of warfarin reduction on patient dose
response variability will depend on genetics, personal habits, and environmental factors that
influence reductase activity (Kassner et al., 2008). Xenobiotics including drugs and pollutants
interact with aryl hydrocarbon receptor (AhR) to bind to xenobiotic response elements (XRE)
that regulate expression and activity of drug metabolizing enzymes including CBR1 (Lakhman et
16

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
al., 2007). By contrast, AKR1C are likely regulated through antioxidant regulatory pathways
(Burczynski et al., 1999) stimulated by habits suggest as smoking, which induces AKR1C3
levels 15-fold (Nagaraj et al., 2006). That effect could explain why smokers are resistant to
warfarin anticoagulation by increasing drug metabolism. Similarly, genetic variants of AKR1C3
(Mindnich and Penning, 2009) and CBR1 (Malatkova et al., 2010) would increase or decrease
their roles in warfarin metabolism and hence patient responses, so if shown relevant, the
polymorphisms could be incorporated into current pharmacogenetic-based dosing algorithms
(Shaw et al., 2015) to improve their predictability. Aside from those unexplored possibilities, the
relative contribution of warfarin reduction to the more dominant oxidative pathways remains
unknown so that their collective effect on R- and S-warfarin metabolism requires further
investigation. In sum, knowledge of the stereospecificity and efficiency of warfarin reduction
provides a critical advance in understanding the overall relevance of metabolism that could
impact patient dose-responses.
17

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Authorship Contributions
Participated in Research Design: Barnette, Goodwin, Miller
Conducted Experiments: Barnette, Johnson, Pouncey, Nshimiyimana, Desrochers
Contributed new reagents or analytic tools: Desrochers, Goodwin, Miller
Performed Data Analysis: Barnette, Johnson, Pouncey, Desrochers, Goodwin, Miller
Wrote or contributed to writing of the manuscript: Barnette, Johnson, Pouncey, Goodwin, Miller
18

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
References
Alshogran OY, Naud J, Ocque AJ, Leblond FA, Pichette V, and Nolin TD (2015a) Effect of Experimental
Kidney Disease on the Functional Expression of Hepatic Reductases. Drug Metab Dispos 43:100-
106.
Alshogran OY, Urquhart BL, and Nolin TD (2015b) Downregulation of Hepatic Carbonyl Reductase Type 1
in End-Stage Renal Disease. Drug Metab Lett 9:111-118.
Ansell J, Hirsh J, Hylek E, Jacobson A, Crowther M, and Palareti G (2008) Pharmacology and management
of the vitamin K antagonists: American College of Chest Physicians Evidence-Based Clinical
Practice Guidelines. Chest 133:160S - 198S.
Breckenridge A, Orme M, Wesseling H, Lewis RJ, and Gibbons R (1974) Pharmacokinetics and
pharmacodynamics of the enantiomers of warfarin in man. Clin Pharmacol Ther 15:424-430.
Breyer-Pfaff U, Martin H-J, Ernst M, and Maser E (2004) ENANTIOSELECTIVITY OF CARBONYL REDUCTION
OF 4-METHYLNITROSAMINO-1-(3-PYRIDYL)-1-BUTANONE BY TISSUE FRACTIONS FROM HUMAN
AND RAT AND BY ENZYMES ISOLATED FROM HUMAN LIVER. Drug Metab Dispos 32:915-922.
Burczynski ME, Lin H-K, and Penning TM (1999) Isoform-specific Induction of a Human Aldo-Keto
Reductase by Polycyclic Aromatic Hydrocarbons (PAHs), Electrophiles, and Oxidative Stress:
Implications for the Alternative Pathway of PAH Activation Catalyzed by Human Dihydrodiol
Dehydrogenase. Cancer Res 59:607-614.
Byrns MC and Penning TM (2009) Type 5 17β-hydroxysteroid dehydrogenase/prostaglandin F synthase
(AKR1C3): Role in breast cancer and inhibition by non-steroidal anti-inflammatory drug analogs.
Chem Biol Interact 178:221-227.
Chan KK, Lewis RJ, and Trager WF (1972) Absolute configurations of the four warfarin alcohols. J Med
Chem 15:1265-1270.
19

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Gebauer M (2007) Synthesis and structure-activity relationships of novel warfarin derivatives. Bioorg
Med Chem 15:2414-2420.
Haque JA, McDonald MG, Kulman JD, and Rettie AE (2014) A cellular system for quantitation of vitamin K
cycle activity: structure-activity effects on vitamin K antagonism by warfarin metabolites. Blood
123:582-589.
Hermans JJ and Thijssen HH (1992) Stereoselective acetonyl side chain reduction of warfarin and
analogs. Partial characterization of two cytosolic carbonyl reductases. Drug Metab Dispos
20:268-274.
Hoffmann F and Maser E (2007) Carbonyl Reductases and Pluripotent Hydroxysteroid Dehydrogenases
of the Short-chain Dehydrogenase/reductase Superfamily. Drug Metab Rev 39:87-144.
Jones DR, Kim SY, Boysen G, Yun CH, and Miller GP (2010) Contribution of Three CYP3A Isoforms to
Metabolism of R- and S-Warfarin. Drug Metab Lett 4:213-219.
Jones DR and Miller GP (2011) Assays and applications in warfarin metabolism: what we know, how we
know it and what we need to know. Expert Opin Drug Metab Toxicol 7:857-874.
Kassner N, Huse K, Martin H-J, Gödtel-Armbrust U, Metzger A, Meineke I, Brockmöller J, Klein K, Zanger
UM, Maser E, and Wojnowski L (2008) Carbonyl Reductase 1 Is a Predominant Doxorubicin
Reductase in the Human Liver. Drug Metab Dispos 36:2113-2120.
Kim H, Yen C, Preston P, and Chin J (2006) Substrate-Directed Stereoselectivity in Vicinal Diamine-
Catalyzed Synthesis of Warfarin. Organic Lett 8:5239-5242.
Laffin B and Petrash M (2012) Expression of the Aldo-Ketoreductases AKR1B1 and AKR1B10 in Human
Cancers. Front Pharmacol 3. 104.
Lakhman SS, Chen X, Gonzalez-Covarrubias V, Schuetz EG, and Blanco JG (2007) Functional
Characterization of the Promoter of Human Carbonyl Reductase 1 (CBR1). Role of XRE Elements
20

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
in Mediating the Induction of CBR1 by Ligands of the Aryl Hydrocarbon Receptor. Mol
Pharmacol 72:734-743.
Levy G, E. Mager D, Cheung WK, and Jusko WJ (2003) Comparative pharmacokinetics of coumarin
anticoagulants L: Physiologic modeling of S-warfarin in rats and pharmacologic target-mediated
warfarin disposition in man. J Pharm Sci 92:985-994.
Lewis R, Trager, WF, Chan, KK, Breckenridge, A, Orme, M, Roland, M, Schary, W (1974) Warfarin.
Stereochemical aspects of its metabolism and the interaction with phenylbutazone. J Clin Invest
53:1607-1617.
Lewis R, Trager, WF, Robinson, AJ, Chan, KK (1973) Warfarin metabolites: the anticoagulant activity and
pharmacology of warfarin alcohols. J Lab Clin Med 81:925-931.
Malatkova P, Maser E, and Wsol V (2010) Human Carbonyl Reductases. Current Drug Metab 11:639-658.
Malátková P, Sokolová S, Chocholoušová Havlíková L, and Wsól V (2016) Carbonyl reduction of warfarin:
Identification and characterization of human warfarin reductases. Biochem Pharmacol 109:83-
90.
Malátková P and Wsól V (2014) Carbonyl reduction pathways in drug metabolism. Drug Metab Rev
46:96-123.
Mindnich RD and Penning TM (2009) Aldo-keto reductase (AKR) superfamily: Genomics and annotation.
Human Genomics 3:362-370.
Moreland TA and Hewick DS (1975) Studies on a ketone reductase in human and rat liver and kidney
soluble fraction using warfarin as a substrate. Biochem Pharmacol 24:1953-1957.
Nagaraj NS, Beckers S, Mensah JK, Waigel S, Vigneswaran N, and Zacharias W (2006) Cigarette smoke
condensate induces cytochromes P450 and aldo-keto reductases in oral cancer cells. Toxicol Lett
165:182-194.
21

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
O'CONNOR T, IRELAND LS, HARRISON DJ, and HAYES JD (1999) Major differences exist in the function
and tissue-specific expression of human aflatoxin B1 aldehyde reductase and the principal
human aldo-keto reductase AKR1 family members. Biochem J 343:487-504.
O’Reilly R (1986) Warfarin metabolism and drug-drug interactions. , in: The new dimensions of warfarin
prophylaxis: advances in experimental medicine and biology (Wessler SB, CG Nemerson, Y ed),
pp 205-212, Plenum, New York, NY.
Preston Pugh C, Pouncey DL, Hartman JH, Nshimiyimana R, Desrochers LP, Goodwin TE, Boysen G, and
Miller GP (2014) Multiple UDP-glucuronosyltransferases in human liver microsomes
glucuronidate both R- and S-7-hydroxywarfarin into two metabolites. Arch Biochem Biophys
564:244-253.
Rižner TL and Penning TM (2014) Role of aldo–keto reductase family 1 (AKR1) enzymes in human steroid
metabolism. Steroids 79:49-63.
Shaw K, Amstutz U, Kim RB, Lesko LJ, Turgeon J, Michaud V, Hwang S, Ito S, Ross C, Carleton BC, and
Group tCCR (2015) Clinical Practice Recommendations on Genetic Testing of CYP2C9 and
VKORC1 Variants in Warfarin Therapy. Ther Drug Monitor 37:428-436.
Sun H, Zhang T, Wu Z, and Wu B (2015) Warfarin is an Effective Modifier of Multiple UDP-
Glucuronosyltransferase Enzymes: Evaluation of its Potential to Alter the Pharmacokinetics of
Zidovudine. Journal of Pharm Sci 104:244-256.
Trager W, Lewis, RJ, Garland, WA (1970) Mass spectral analysis in the identification of human
metabolites of warfarin. J Med Chem 13:1196-1204.
Wong TC, Sultana CM, and Vosburg DA (2010) A Green, Enantioselective Synthesis of Warfarin for the
Undergraduate Organic Laboratory. J Chem Educ 87:194-195.
Wysowski DK, Nourjah P, and Swartz L (2007) Bleeding complications with warfarin use: a prevalent
adverse effect resulting in regulatory action. Arch Intern Med 167:1414-1419.
22

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Zhang L, Zhang H, Zhao Y, Li Z, Chen S, Zhai J, Chen Y, Xie W, Wang Z, Li Q, Zheng X, and Hu X (2013)
Inhibitor selectivity between aldo–keto reductase superfamily members AKR1B10 and AKR1B1:
Role of Trp112 (Trp111). FEBS Lett 587:3681-3686.
23

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Footnotes
Financial support for undergraduate research was provided in part by the Summer
Undergraduate Research Fellowship program sponsored by the Biochemistry and Molecular
Biology Department at the University of Arkansas for Medical Sciences (BPJ and DLP), the
UAMS Summer Undergraduate Research Program to Increase Diversity in Research under the
National Institutes of Health National Heart, Lung and Blood Institute award [R25HL108825]
(DLP). In addition, financial support was provided to GPM through a Grant-in-Aid from the
SouthWest Affiliate of the American Heart Association [13GRNT16960043] and a pilot study
grant provided by the National Institutes of Health [UL1 TR000039] awarded to the University of
Arkansas for Medical Sciences along with grants from the National Institutes of Health National
Library of Medicine [R01LM012222 and R01LM012482]. The content of this manuscript is
solely the responsibility of the authors and does not necessarily represent the official views of
the University of Arkansas for Medical Sciences, the American Heart Association, or National
Institutes of Health.
TEG gratefully acknowledges funding from the Hendrix College Odyssey and Rwanda
Presidential Scholars Programs, as well as from John and Laura Byrd. Funding was provided by
the U.S. National Science Foundation for the NMR spectrometer [1040470].
24

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
Legends for Figures
Figure 1. Reduction of R- and S- warfarin introduces a second chiral center to the
molecule. The molecular structure of warfarin contains a single chiral center (letter designation
in blue italics). The reduction of R- and S- warfarin introduces a second chiral center to yield a
warfarin alcohol so that there are four possible metabolites.
Figure 2. Control experiments to assess optimal conditions for steady state studies.
Panel A, impact of buffer on activity. Total warfarin alcohol formation rates over 30 min
measured for 500 µM warfarin, 1 mM NADPH and 0.5 mg/mL protein reactions in 50 mM
potassium phosphate pH 7.4 (KPi) with or without 5 mM MgCl2 or 50 mM Tris hydrochloride pH
7.4 (Tris-HCl) with or without 5 mM MgCl2. Panel B, impact of co-solvent on activity. Total
warfarin alcohol formation rates measured for reactions with 50 µM warfarin, 1 mM NADPH and
0.5 mg/mL protein at 0.5, 1.0, 2.0, or 5.0 % methanol or dimethylsulfoxide (DMSO) co-solvent in
50 mM potassium phosphate pH 7.4. Panel C, impact of protein levels on activity. After 40 min
reaction, total warfarin alcohol levels measured for 50 µM warfarin and 1 mM NADPH reactions
with 0.25, 0.5 and 1.0 mg protein/mL pooled human liver cytosol in 50 mM potassium phosphate
pH 7.4. Panel D, impact of time on activity. Total warfarin alcohol levels measured for reactions
of 50 µM warfarin at 10, 20, 40, and 60 min after initiation with NADPH in 50 mM potassium
phosphate pH 7.4.
Figure 3. Kinetic profiles for steady-state reduction of R- and S-warfarin by human liver
cytosol. Reduction of R- and S-warfarin as described in Materials and Methods yielded two
alcohols for each set of reactions. Panel A, kinetic profile for R-warfarin reduction showing the
major (RS-warfarin alcohol, dark blue circle) and minor (RR-warfarin alcohol, light blue circle)
metabolites. Both sets of data were fit to the Michaelis-Menten model and the resulting kinetic
constants reported in Table 1. Panel B, expanded view of kinetic profile for R-warfarin reduction
to RS-warfarin alcohol and its corresponding fit of the data. Panel C, kinetic profile for S-warfarin
25

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
reduction showing the major metabolites (SS-warfarin alcohol, dark red circle). SR-warfarin
alcohol was not quantifiable due to signal at the limit of detection. Data for SS-warfarin alcohol
were fit to Michaelis-Menten model and resulting kinetic constants reported in Table 1.
Figure 4. Kinetic profile for steady-state reduction of rac-warfarin by human liver cytosol.
Reduction of rac-warfarin as described in Materials and Methods yielded unresolved mixtures of
metabolites. Panel A, kinetic profile for alcohol 1 (RR- and SS-warfarin alcohol, light purple
circle) and alcohol 2 (RS- and SR-warfarin alcohols, dark purple circle). Panel B, expanded view
of kinetic profile for rac-warfarin reduction to alcohol 1 and its corresponding fit of the data. Both
sets of data were fit to the Michaelis-Menten model and the resulting kinetic constants reported
in Table 1.
Figure 5. Inhibitor phenotyping of steady state reduction of R- and S-warfarin by HLC150.
Reactions were carried out with two substrate concentrations, 50 µM (Panels A and B) and 500
µM (Panels C and D), 0.25 mg/mL protein, and 1 mM NADPH with inhibitors for specific cytosol
reductases at 37 ºC and pH 7.4. Panels A and C depict selective inhibition of R-warfarin to RS-
warfarin alcohol. Panels B and D show selective inhibition of S-warfarin to SS-warfarin alcohol.
Reaction activities were normalized to the average control rate measured for reactions in the
absence of inhibition. Rates were reported as percentages of the control rate. The reported
activities are averaged from a minimum of five individual experiments ± standard error.
Asterisks indicate statistical significance of inhibition compared to control as calculated using
Student’s t-test (* p<0.05, ** p<0.005; *** p < 0.0005).
Figure 6. Assessment of potential interactions between R- and S-warfarin during
metabolism. After experimentally determining rac-warfarin kinetics for alcohol 1 and 2
formation (light and dark purple circles, respectively), we predicted the apparent alcohol 1 (light
purple circles) and 2 (dark purple circles) formation rates by combining our steady-state data for
the reactions with the individual drug enantiomers (Figure 3). The predicted versus
26

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
DMD # 75929
experimentally measured alcohol 1 and 2 rates were plotted together for comparison. Panel A,
experimentally determined kinetic profile for rac-warfarin reaction (circles, solid lines) as well as
that predicted by metabolism of individual enantiomers (squares, dashed lines). Panel B,
expanded view of kinetic profile for rac-warfarin reduction to alcohol 1 and the predicted kinetic
profile. Fit of the estimated data to Michaelis-Menten model is shown in dashed line for
comparison to that for actual rac-warfarin reactions in solid line.
27

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
Figures
Figure 1
29

DMD Fast Forward. Published on June 23, 2017 as DOI: 10.1124/dmd.117.075929
This article has not been copyedited and formatted. The final version may differ from this version.
Figure 2
30