Experiments were conducted to characterize the metabolism of
ezlopitant alkene (CJ-12,458), an active metabolite of ezlopitant, in
human liver microsomes. In incubations with human liver microsomes and
cofactors required for cytochrome P450 (CYP) activity, CJ-12,458 was
converted to two metabolites: a diol (CP-611,781) and a 1° alcohol
(CP-616,762). In human liver microsomes, apparent
KM values of 5.4 and 8.5 µM were
determined for the formation of diol and 1° alcohol metabolites,
respectively. High KM activities were also
observed for formation of these metabolites; however, the aforementioned low KM activities accounted
for greater than 90% of the total intrinsic clearance. In pooled human
liver microsomes, formation of both metabolites was partially inhibited
by both quinidine and ketoconazole, suggesting that CYP2D6 and CYP3A
enzymes are involved in the metabolism of CJ-12,458. This evidence was corroborated through the use of heterologously expressed CYP enzymes and correlation analysis with a panel of human liver microsomes. The
data suggest that CYP2D6 is quantitatively more important than CYP3A in
the metabolism of CJ-12,458 by a factor of about 2 to 1. The conversion
of an alkene to a 1° alcohol represents a novel biotransformation
reaction. Incubations using 18O2,
2H2O, [2H5]CJ-12,458,
and [2H]NADPH were conducted and the 1° alcohol product
was characterized by ion trap mass spectrometry. From these data, a
mechanism for this reaction is proposed involving epoxidation, an
exocyclic hydride shift, and reduction at the benzylic position.
 |
Introduction |
Ezlopitant (Fig.
1) is a member of a new potential class
of agents that act by antagonizing the
substance P receptor. As such, this and other similar
agents may be useful in the treatment of disorders in which substance P
is involved, such as pain, inflammation, emesis, and psychiatric
disease (McLean, 1996
). It was found in both in vitro and in vivo
experiments (Reed-Hagen et al., 1999; Obach, 2000) that ezlopitant is
oxidatively metabolized to two pharmacologically equipotent
metabolites, CJ-12,458 (ezlopitant alkene) and CJ-12,764 (Fig. 1).
While the latter metabolite was a major metabolite in both human and
animal species, the former was found in greater relative abundance in
humans. Further experimentation found that both CYP3A4 and CYP2D6
catalyzed the metabolism of ezlopitant to these two metabolites in
human liver microsomes, with the former contributing a greater share of
the total activity (Obach, 2000).
The present investigations were undertaken to characterize the
metabolism of the major pharmacologically active metabolite, CJ-12,458,
in human liver microsomes. The primary objectives of these experiments
were to identify the major in vitro metabolites of CJ-12,458, describe
the enzyme kinetics of metabolism to gain an appreciation of the
kinetics of these reactions in vivo, and identify the human
CYP1 enzyme(s) responsible for CJ-12,458
metabolism. During these investigations, it was shown that the alkene
moiety of CJ-12,458 was metabolized to a 1° alcohol product (Fig.
2). Thus, a fourth objective of this
investigation was to gather evidence in support of a mechanistic
hypothesis for this unusual biotransformation reaction.
Experimental Procedures
Materials.
CJ-12,458 methanesulfonate salt and CJ-12,764 (Fig. 1) were obtained as
previously described (Reed-Hagen et al., 1999). CP-611,781 and
CP-616,762 (Fig. 2) were synthesized as described below. NADPH was from
Sigma (St. Louis, MO).
N-Methylmorpholine-N-oxide,
OsO4, and LiBH4 were from
Aldrich (Milwaukee, WI). Human liver microsomes and microsomes from Sf9
cells containing heterologously expressed recombinant CYP enzymes were
from a microsome bank maintained in the Drug Metabolism Department of
Pfizer (Groton, CT).
Identification of CJ-12,458 Metabolites in Human Liver
Microsomes.
A reaction mixture containing pooled human liver microsomes (1.0 mg/ml), CJ-12,458 (10 µM), MgCl2 (3.3 mM), and
NADPH (1.3 mM) in 5.0 ml of 25 mM
KH2PO4, pH 7.5, was
incubated at 37°C for 20 min in a shaking water bath open to the air.
A control incubation was also prepared that contained all components
except NADPH. The reaction was terminated with NaOH (0.1 ml/10 M) and
extracted with ethyl acetate (5 ml). The extract was dried under
N2 and reconstituted in HPLC mobile phase.
The HPLC system contained a Waters Symmetry C18
column (Waters, Milford, MA; 3.9 × 150 mm; 5-µ particle size)
equilibrated in 20 mM of acetic acid, pH 4 (with
NH4OH) containing 14.5%
CH3CN at a flow rate of 0.8 ml/min. After
injection, this mobile phase composition was maintained for 3 min,
followed by a linear gradient to 95% CH3CN at 25 min. The column effluent was analyzed by inline UV and atmospheric
pressure ionization mass spectrometry detection. The wavelength
used was 275 nm. The mass spectrometer (Sciex API100, Thornhille,
Ontario) contained an atmospheric pressure chemical ionization
interface operated in the positive ion mode with the heated nebulizer
temperature set at 500°C. The orifice voltage was varied to generate
little or extensive fragmentation of ions. Other state file settings
were adjusted to optimize the signal for the protonated molecular ion
of CJ-12,458.
In one experiment, incubation mixtures were extracted under both basic
(as above) and neutral conditions to determine whether the metabolite
profile varied with differing sample work-up conditions.
After a preliminary proposal of metabolite structure, the metabolites
were synthesized (see below), and the HPLC behavior of metabolites in
an incubation mixture extract was compared with the HPLC behavior of
the authentic metabolite standards.
Chemical Synthesis of CP-611,781 (Diol Metabolite).
CJ-12,458 (28.8 mg of the methanesulfonate salt) was dissolved in water
(10 ml), followed by basification with NaOH (0.1 ml; 1 M) and
extraction into methyl t-butyl ether (MTBE, 3 ml) to generate the free base. The solvent was evaporated under
N2. Tetrahydrofuran (0.6 ml), 1-propanol
(0.2 ml), and H2O (0.1 ml) were added to the
residue, and the solution was transferred to a 5-ml round-bottom flask.
N-Methylmorpholine-N-oxide (9 mg) was added, and
the mixture was cooled on ice. OsO4 (0.022 ml of
a 2.5% solution in t-butyl alcohol) was added to the
cooled, stirring solution, the ice was removed, and the reaction was
stirred overnight, protected from light. TLC analysis had indicated
complete reaction. The reaction was terminated with NaOH (2 ml; 1 M)
and extracted with ethyl acetate (5 ml). The extract was evaporated
under N2. The residue was reconstituted in 1 ml
of 10% CH3CN in water, pH adjusted to neutral
with HCl, and purified by semipreparative HPLC using a Phenomenex
C18 column (10 × 750 mm; Phenomenex,
Torrance, CA) and a mobile phase of 20 mM acetic acid, pH 4 (with
NH4OH) containing 25%
CH3CN at a flow rate of 3 ml/min. Fractions were
collected, and those containing the diol product were pooled, basified
(NaOH), and extracted with ethyl acetate (5 ml). The extract was
evaporated under N2 to yield 7.3 mg of CP-611,781
(Fig. 3). Compound purity was checked by
TLC and HPLC, and mass spectral and 1H NMR data
supported the structure.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 3.
Synthetic schemes for CP-611,781 (diol
metabolite), CP-616,762 (1° alcohol metabolite), and
[2H5]CJ-12,458. TFA, trifluoroacetic
acid.
|
|
Chemical Synthesis of CP-616,762 (1° Alcohol Metabolite).
CJ-13,517 (30 mg of the methanesulfonate salt courtesy of Dr. K. Satake, Pfizer, Nagoya, Japan) was dissolved in water (1 ml), followed
by basification with Na2CO3
(1 ml; 1 M; pH 10.5) and extraction into MTBE (5 ml) to generate the
free base. The solvent was evaporated under N2.
Tetrahydrofuran (1 ml) was added to the residue, and the solution was
transferred to a 5-ml round-bottom flask and cooled on ice.
LiBH4 (7.5 mg) was added, and the stirring reaction mixture was permitted to come to room temperature.
Approximately 2 h later, a second portion of
LiBH4 was added, and the mixture was stirred
overnight. The reaction was terminated by careful addition of
H2O (3 ml) followed by extraction with
CH2Cl2 (5 ml). The extract
was washed with saturated NaCl, dried (MgSO4),
and filtered, and the solvent was removed under
N2 to yield 19 mg of pale oil. This material was
applied to a preparative silica TLC plate and developed with ethyl
acetate/triethylamine (100:1). The product band was scraped and eluted
with ethyl acetate/methanol (3:1), and the solvent was removed under
N2 to yield 8.9 mg of CP-616,762 (Fig. 3) as a
white solid. Purity was checked by TLC, and mass spectral data
supported the structure.
Chemical Synthesis of [2H5]CJ-12,458.
Hexadeuterated CJ-12,764 (Reed-Hagen et al., 2000) was dissolved in
CH3CN (2 ml), and trifluoroacetic acid (0.02 ml)
was added. The reaction was permitted to stand at room temperature for
46 h, after which the solvent was removed under
N2. The residue was reconstituted in 0.3 ml of
20% CH3CN in 20 mM acetic acid (adjusted to pH
4.0 with NH4OH), and the material was applied to
a semipreparative HPLC column (Keystone Hypersil ODS; 10 × 250 mm) equilibrated in 45.5% CH3CN in 20 mM
acetic acid (adjusted to pH 4.0 with NH4OH). The
flow rate was 4 ml/min. Elution was effected by maintaining the initial
mobile phase conditions for 10 min, followed by a linear gradient to
95% CH3CN at 18 min, and holding this mobile phase composition. The product eluted at 16 to 19 min, and fractions containing the product were pooled and evaporated under
N2. The resulting material was partitioned
between 1 M NaOH and MTBE, and the organic material was evaporated
under N2 to yield 1.9 mg of
[2H5]CJ-12,458. The
structure was confirmed with mass spectral data.
Substrate Saturation in Human Liver Microsomes.
CJ-12,458 (at concentrations ranging from 0.1-500 µM) was incubated
with pooled human liver microsomes (2.0 mg/ml),
MgCl2 (3.3 mM), and NADPH (1.3 mM) in 0.2 ml of
KH2PO4, pH 7.5, in
triplicate. The incubations were started by addition of NADPH and
shaken at 37°C for 10 min. The reactions were terminated by addition
of 0.1 ml of NaOH (0.1 M) and analyzed for CP-611,781 and CP-616,762 as
described below. Enzyme kinetic parameters were estimated by first
plotting the data on Eadie-Hofstee plots to determine mono- or
biphasicity. The data were then fit to the Michaelis-Menten equation by
nonlinear regression using Deltagraph v4.0.3 (SPSS, Chicago, IL).
Correlation between CJ-12,458 Metabolism and CYP-Specific Marker
Activities.
CJ-12,458 (2.5 µM) was incubated separately with liver microsomes
isolated from 11 donors. Incubations were conducted using conditions
and cofactors, and samples were analyzed as described above. Reaction
velocities for CJ-12,458 conversion to CP-611,781 and CP-616,762 were
correlated to standard CYP-specific marker activities for CYP1A2
(phenacetin O-deethylase), CYP2C9 (tolbutamide hydroxylase),
CYP2C19 (S-mephenytoin 4'-hydroxylase), CYP2D6 (bufuralol 1'-hydroxylase), and CYP3A4 (testosterone 6
-hydroxylase) that had
been previously determined using standard conditions.
Inhibition of CJ-12,458 Metabolism by CYP-Specific Inhibitors.
CJ-12,458 (2.5 µM) was incubated with pooled human liver microsomes
as described above, in the absence and presence of quinidine (0.01-10
µM), ketoconazole (0.01-10 µM), or sulfaphenazole (1.0-100 µM).
Incubations were done in duplicate.
Metabolism of CJ-12,458 by Heterologously Expressed Recombinant
CYP Isoforms.
CJ-12,458 (2.5 µM) was incubated in a 0.2-ml volume as described
above containing baculovirus-expressed CYP1A1 (GENTEST Supersomes; GENTEST, Woburn, MA), CYP2A6 (GENTEST -lymphoblastoid expression system), CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1 (with cytochrome b5 added), CYP3A4, and CYP3A5. Unless
otherwise noted, CYP enzymes were from a bank generated and maintained
in-house. Subsequent substrate saturation experiments were performed
with CYP1A1, CYP2D6, CYP3A4, and CYP3A5 using CJ-12,458 with
concentration ranges of 0.5 to 50 µM (CYP1A1), 0.1 to 10 µM
(CYP2D6), or 1.0 to 200 µM (CYP3A4 and CYP3A5). Incubations were done
in duplicate. Enzyme kinetic parameters were estimated by fitting the
data to the Michaelis-Menten equation using nonlinear regression.
Analysis of CP-611,781 (Diol) and CP-616,762 (1° OH) by
HPLC-MS.
Standard curves containing CP-611,781 and CP-616,762 in a 0.2-ml
control matrix (liver microsomes plus buffer, without cofactors) were
constructed with concentrations ranging from 2.0 to 1000 ng/ml.
Internal standard solution (50 µl of a 2.0-µg/ml aqueous solution
of CJ-11,957, an analog of CJ-12,458 in which the isopropene substituent is replaced with an ethyl substituent) was added to samples
and standards, followed by basification with NaOH (0.1 ml; 0.1 M).
Samples were extracted with MTBE (3 ml), the organic fraction was
evaporated under N2 at 30°C, and the residue
was reconstituted in HPLC mobile phase (75% aqueous/25%
CH3CN).
Reconstituted extracts were injected (40 µl) onto HPLC. The column
was a Waters Symmetry C18 (3.9 × 150 mm;
5-µ particle size) equilibrated in mobile phase consisting of 20 mM
acetic acid, pH 4 (with
NH4OH)/CH3CN (65:35). The
flow rate was 0.8 ml/min. The mobile phase composition was maintained
for 2 min followed by a linear gradient to 95%
CH3CN at 7 min. Detection was accomplished by
selected ion monitoring on a Sciex API100 mass spectrometer. The
effluent was introduced into an atmospheric pressure chemical ionization source with the heated nebulizer temperature set at 500°C
and the orifice voltage set at 35 V. Ions m/z
487, 471, and 441 were monitored for CP-611,781, CP-616,762, and
CJ-11,957 internal standard, respectively. Retention times were
2.1, 4.1, and 5.7 min, respectively.
Mechanistic Studies.
Incubations were conducted as described above using pooled human liver
microsomes, recombinant CYP2D6 or recombinant CYP3A4, and CJ-12,458 (10 µM) in a total incubation volume of 5 ml. The various conditions
described in Table 5 were used to address the mechanistic questions.
Monodeuterated NADPH (R- and S-enantiomers) were
prepared biosynthetically as described by Ottolina et al. (1989). For
analysis by ion-trap mass spectrometry, CP-616,762 was first purified
from extracted incubation matrices by HPLC. Fractions containing the
product were directly infused (5-20 µl/min) into a Finnigan LCQ ion
trap mass spectrometer (ThermoFinnigan, San Jose, CA) with an
ionspray interface operated in the positive ion mode. Various
potentials were applied to effect fragmentation and
MS3 analysis.
| |
Results |
Identification of Human Liver Microsomal Metabolites of
CJ-12,458.
After incubation of CJ-12,458 with pooled human liver microsomes, two
major metabolites were observed, eluting at
Rt = 11.1 and 12.8 min (Figs. 4 and
5). Mass spectra for these metabolites were obtained at low and high orifice voltages (Fig.
6). For the metabolite eluting at 11.1 min, the protonated molecular ion at low energy was
m/z 487, indicating an addition of 34 mass units to CJ-12,458 (m/z 453). Fragment ions of
m/z 469 (loss of H2O, most
likely from the benzylic position) and m/z 455 (loss of CH3OH) were observed. Under high-energy
conditions, fragment ions of m/z 276 (intact
diphenylquinuclidine) and m/z 195 (indicating two
OH on the isopropyl anisole) were observed. The diol metabolite was
proposed from these data, in light of the knowledge that epoxidation of
alkenes and subsequent hydrolysis is a common route of xenobiotic biotransformations. This was confirmed by comparing chromatographic and
mass spectral properties of the synthetic standard, CP-611,781, with
those of the metabolite. In addition to the sample workup described
under Experimental Procedures, an additional experiment was
conducted in which the incubation mixture was not basified prior to
extraction. This was done to determine whether the diol metabolite
arose via base hydrolysis of the corresponding epoxide, and would thus
potentially be an artifact of the sample work-up procedure. The results
demonstrated that the diol was present under both sample processing
conditions and was not artifactually formed from a putative epoxide
intermediate.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 4.
UV chromatograms of extracts of incubations
containing CJ-12,458 and human liver microsomes.
A, control incubation in which the reaction mixture was terminated at
time = 0. B, 20-min incubation. Note the peak at
Rt = 13.1 min in the control
incubation. This represents a UV-absorbing contaminant. CJ-12,458
eluted at 16.4 to 16.6 min; CP-611,781 and CP-616,762 eluted
at Rt = 11.1 and 12.8 min,
respectively.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 5.
Total ion chromatograms of CJ-12,458
incubation mixture extract and synthetic standards.
A, CJ-12,458 human liver microsomal incubation mixture. B, CJ-12,458
standard. C, CP-611,781 (diol) standard. D, CP-616,762 (1° alcohol)
standard.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 6.
Atmospheric pressure chemical ionization
mass spectra.
A, CP-611,781 (diol metabolite) mass spectrum at low orifice voltage.
B, CP-611,781 (diol metabolite) mass spectrum at high orifice voltage.
C, CP-616,762 (1° alcohol) mass spectrum at low orifice voltage. D,
CP-616,762 (1° alcohol) mass spectrum at high orifice voltage. amu,
atomic mass units.
|
|
The metabolite eluting at Rt = 12.8 min
possessed a protonated molecular ion of m/z 471, indicating an addition of 18 mass units to CJ-12,458. Fragment ions of
m/z 276 (intact diphenylquinuclidine) and
m/z 179 (indicating OH on the isopropyl anisole
moiety) were observed at high orifice voltage. Examination of the
chromatographic and mass spectral properties of CJ-12,764 (the benzylic
alcohol metabolite of ezlopitant) demonstrated that the CJ-12,458
metabolite was not this compound. The structure of a 1° alcohol was
proposed from these data and confirmed when chromatographic and
spectral properties were compared with the synthetic standard
CP-616,762. The metabolite CP-616,762 represents hydration of
CJ-12,458, which is an unusual biotransformation reaction.
Enzyme Kinetics of CJ-12,458 Metabolism in Human Liver Microsomes.
An initial experiment was conducted to determine appropriate microsomal
protein concentration and incubation time that yielded linear reaction
velocity conditions. All subsequent experiments were done using 2.0 mg/ml liver microsomal protein and a 10-min incubation time. The
Michaelis-Menten plot for CJ-12,458 metabolism is shown in Fig.
7. The data indicated more than one
enzyme involved in the metabolism to each metabolite and were thus
fitted to the function:
in which the individual parameters
Vmax and apparent
KM could be defined for the low
KM activity, but the high
KM activity could only be defined as a
composite intrinsic clearance value. A summary of the kinetic
parameters is shown in Table 1. The low
KM activity accounts for greater than 90%
of the overall conversion of CJ-12,458 to the two metabolites. The data
suggest that the formation of the diol metabolite is of a greater
magnitude (ca. 3-fold) than formation of the 1° alcohol.
Inhibition of CJ-12,458 Metabolism by P450 Isoform-Specific
Inhibitors.
The effects of P450 specific inhibitors quinidine (CYP2D6),
ketoconazole (CYP3A), and sulfaphenazole (CYP2C9) on the metabolism of
CJ-12,458 in human liver microsomes were investigated. The substrate
concentration used was 2.5 µM so as to be below the low
KM values and thus reflect the low
KM activity. Inhibition curves are shown in
Fig. 8. Complete inhibition was not
observed with any of the inhibitors, thus the data were fit to the
function:
in which the IC50 represents the inflection
point (not a true 50% inhibition since complete inhibition would not
be observed), and the value of 100 
(AB) is the maximum inhibition
observed at an infinite inhibitor concentration. A summary of
inhibition parameters is listed in Table
2. Both quinidine and ketoconazole demonstrated inhibition, with IC50 values in the
range of those reported in the literature. Sulfaphenazole had no
effect.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 8.
Inhibition of CJ-12,458 metabolism in human
liver microsomes.
Top panel, quinidine (CYP2D6 inhibitor). Bottom panel, ketoconazole
(CYP3A inhibitor). Data were fit by nonlinear regression
[r2 = 0.982 and 0.986 for quinidine
inhibition of CP-611,781 ( ) and CP-616,762 ( ), respectively, and
r2 = 0.980 and 0.886 for ketoconazole
inhibition of CP-611,781 and CP-616,762, respectively].
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 2
Inhibition of human liver microsomal ezlopitant alkene (CJ-12,458)
metabolism by cytochrome P450 isoform-specific inhibitors
|
|
Correlation of CJ-12,458 Metabolism with Cytochrome P450 Marker
Activities.
The rate of metabolism of CJ-12,458 (at 2.5 µM) to diol and 1°
alcohol metabolites was measured in liver microsomes from 11 individual
donors. These rates were correlated to rates of standard P450 marker
substrate activities (Table 3). Of the
five marker activities, the metabolism of CJ-12,458 only correlated
with bufuralol 1'-hydroxylase, the CYP2D6-specific activity.
Correlation plots are shown in Fig. 9.
Also, the formation rates of CP-611,781 and CP-616,762 were highly
correlated to each other (r2 = 0.928)
supporting the notion that both activities are catalyzed by the same
P450 isoform(s).
View this table:
[in this window]
[in a new window]
|
TABLE 3
Correlation of ezlopitant alkene (CJ-12,458) metabolism to cytochrome
P450-specific marker activities in a panel of human liver microsomes
|
|
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 9.
Correlation between CJ-12,458 metabolism and
bufuralol 1'-hydroxylase activities in a panel of human liver
microsomes.
Top panel, CP-616,762 (1° alcohol) formation. Bottom panel,
CP-611,781 (diol) formation.
|
|
Metabolism of CJ-12,458 by Heterologously Expressed Recombinant
Human Cytochrome P450 Enzymes.
In an initial experiment, CJ-12,458 (2.5 µM) was incubated with
several recombinant human P450 enzymes. Of those examined, only CYP1A1,
CYP2D6, CYP3A4, and CYP3A5 demonstrated a substantial capability to
metabolize CJ-12,458. Other P450 enzymes tested that do not appear to
metabolize CJ-12,458 to a significant extent include CYP1A2, CYP2A6,
CYP2C9, CYP2C19, and CYP2E1.
A subsequent experiment was undertaken to describe the enzyme kinetics
of CJ-12,458 metabolism by the four P450 enzymes initially demonstrated
to possess activity toward this substrate. Each demonstrated monophasic
Michaelis-Menten kinetics (Fig. 10). A
summary of enzyme kinetic parameters is listed in Table
4. The lowest
KM(app) was observed for CYP2D6 (
1
µM). For each enzyme, the KM(app) values for the two metabolic routes were nearly identical.
View this table:
[in this window]
[in a new window]
|
TABLE 4
Enzyme kinetic parameters for ezlopitant alkene (CJ-12,458) metabolism
by heterologously expressed recombinant human cytochrome P450 enzymes
|
|
Mechanistic Studies.
When the incubation of CJ-12,458 with human liver microsomes was
conducted without NADPH, neither CP-611,781 nor CP-616,762 was
observed. In incubations of CP-611,781 with human liver microsomes, no
CP-616,762 was observed, suggesting that the 1° alcohol did not arise
from the diol; however, incubation of CP-616,762 did give rise to a
small amount of CP-611,781.
When incubations of CJ-12,458 were conducted under an
18O2 atmosphere, mass
spectrometric analysis showed that the diol metabolite possessed one
18O while the 1° alcohol metabolite possessed
only 16O, indicating that the origin of this
oxygen atom in the 1° alcohol was not from dioxygen (Fig.
11). When the incubation was conducted in 80% 2H2O, no deuterium
was observed in the diol metabolite, and approximately 40% of the 1°
alcohol metabolite possessed a single deuterium atom (Fig. 11).
However, the presence of this deuterium was not due to chemical
exchange, because incubation of the 1° alcohol and CJ-12,458 in
2H2O did not result in
incorporation of deuterium into these compounds. Ion trap mass
spectrometric analysis of the 1° alcohol (to
MS3) showed that fragmentation by loss of water
could be observed. When the 1° alcohol product containing a single
deuterium atom was subjected to this analysis, the loss of water
occurred with a loss of 18 mass units (H2O), not
19 mass units (DHO), suggesting that the position of deuteration is not
the benzylic position.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 11.
Molecular ions of CP-616,762 (1° alcohol)
generated in incubations of CJ-12,458 with human liver microsomes under
various conditions.
A, incubation under an 18O2 atmosphere. B,
incubation in 80% 2H2O. C, using
[2H5]CJ-12,458 as substrate. D, using
[4-2H]NADPH as cofactor.
|
|
Metabolism of
[2H5]CJ-12,458 by human
liver microsomes resulted in the formation of the 1° alcohol
metabolite in which 52% of the molecules retained all five deuterium
atoms and 48% of the molecules had one hydrogen atom in place of a
deuterium. Ion trap MS3 analysis showed that
neither tetradeuterated nor pentadeuterated product possessed deuterium
on the benzylic position. Furthermore, a synthetic standard of
[2H2]CP-616,762 in which
the two deuterium atoms are on the 1° position was analyzed by
MS3 to verify the proposed fragmentation pathway.
It was confirmed that the fragmentation arising from loss of water did
indeed occur by loss of the benzylic hydrogen and the 1° alcohol.
To determine whether the benzylic hydrogen in the 1° alcohol product
came from NADPH, incubations were conducted in which [4R-2H]- and
[4S-2H]NADPH were used as the
cofactor. With either stereoisomer of deuterated NADPH, 24% of the
product 1° alcohol contained a single deuterium atom. When the 1°
alcohol was analyzed by MS3, the deuterium atom
was found to be incorporated at the benzylic position. A summary of the
findings from these mechanistic experiments is presented in Table
5.
Further mechanistic experiments were done using recombinant CYP3A4 and
CYP2D6. When incubations were conducted under an
18O2 atmosphere, the
identical results regarding isotopic incorporation were observed as
with the human liver microsomes. The 1° alcohol product only
possessed 16O2, indicating
the source was water (Fig. 12).
However, incubations with biosynthesized
[4S-2H]NADPH yielded different
results. The 1° alcohol metabolite formed by CYP3A4 and deuterated
NADPH was 66% incorporated with deuterium, whereas the corresponding
value for CYP2D6 was merely 6%.
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 12.
Molecular ions of CP-616,762 (1° alcohol)
generated in incubations of CJ-12,458 with recombinant CYP enzymes
under various conditions.
A, incubation with CYP3A4 under an 18O2
atmosphere. B, incubation with CYP3A4 using [4-2H]NADPH
as cofactor. C, incubation with CYP2D6 under an
18O2 atmosphere. D, incubation with CYP2D6
using [4-2H]NADPH as cofactor.
|
|
| |
Discussion |
CJ-12,458 is a major circulating metabolite in human subjects
receiving ezlopitant and is also a major metabolite of ezlopitant in
incubations with human liver microsomes and heterologously expressed
recombinant human CYP enzymes (Obach, 2000). It arises by the unusual,
but not unprecedented, biotransformation route of aliphatic
dehydrogenation. Since CJ-12,458 is also a pharmacologically active
metabolite of ezlopitant, an understanding of its biotransformation pathways and enzymes involved is important.
Evidence was presented that demonstrated the identities of the two
major human liver microsomal metabolites of CJ-12,458. The diol,
CP-611,781, is proposed to arise from an unisolable epoxide
intermediate. It is not known whether the epoxide is transformed to the
diol via enzymatic means (i.e., microsomal epoxide hydrolase) or
whether it spontaneously hydrates at neutral pH. Interestingly, no
evidence of formation of diastereomeric diols was seen. This may
indicate that the hydrolysis of the epoxide is enzymatically catalyzed,
explaining the lack of chromatographically resolvable diastereomeric
diols. However, the lack of observation of diastereomers is not proof
of this. The second metabolite is unusual in that it represents an
alkene hydration reaction (CP-616,762, referred to also as the 1°
alcohol). The structure of this metabolite was confirmed by comparing
chromatographic and mass spectrometric properties of the metabolite
with the synthetic standard. There is no precedent for hydration of an
alkene as a P450-catalyzed reaction. Since the formations of diol and
1° alcohol have nearly identical KM(app)
values for all P450s tested, it is likely that both could arise from a
common intermediate (in which the dominating kinetic microconstants in
KM occur prior to the step of partitioning between diol and 1° alcohol). The resulting intermediate would require reduction, which could come from P450 reductase or other NADPH
utilizing reductases present in microsomes. However, the identity of
the enzyme responsible for the reduction step was not determined.
The enzyme kinetics of human liver microsomal metabolism of CJ-12,458
were measured, and the data showed that at least two enzymes are
involved in the formation of both CP-611,781 and CP-616,762. One
kinetically distinguishable activity possesses a low apparent Michaelis
constant (KM = 5.4 and 8.5 µM) while the
second has a KM so high as to be
unmeasurable. The low KM activity accounts for greater than 90% of the total intrinsic clearance and is thus likely to be of greater significance in vivo.
Three lines of evidence support roles for CYP2D6 and CYP3A4 in the
liver microsomal metabolism of CJ-12,458: correlation analysis, CYP-specific inhibition experiments, and recombinant CYP enzymes. The
extent to which each of these enzymes participates in the clearance of
CJ-12,458 can be estimated from these data. The correlation analysis
suggests a major role for CYP2D6, since the rate of CJ-12,458 metabolism was highly correlated with CYP2D6-specific activity. The use
of the specific inhibitors quinidine and ketoconazole provides a more
quantitative projection. Quinidine inhibited CJ-12,458 metabolism with
IC50 values of about 0.1 µM, a value consistent with the known inhibitory potency of this agent on CYP2D6. Fitting the
data to an inhibition curve function yielded a maximum value that
quinidine could inhibit: 60%. This suggests that CJ-12,458 metabolism would have a 60% contribution from CYP2D6. The ketoconazole data, in which the IC50 value was similar to the
known potency of this compound for CYP3A enzymes, yielded a projected
maximum inhibition of 20 to 30%. Thus, another 20 to 30% of CJ-12,458 hepatic metabolism would be due to CYP3A. The third line of evidence comes from the use of recombinant heterologously expressed P450 enzymes. On a per nanomole of P450 basis, the intrinsic clearance for
CYP2D6-mediated CJ-12,458 metabolism was 17-fold greater than that of
CYP3A4. Estimates of P450 enzyme expression in human liver place CYP3A4
at about 30% of total CYP and CYP2D6 at about 4% (Rodrigues, 1999).
Thus, normalizing these activity values to hepatic expression levels,
the 17-fold difference is reduced by 7.5-fold to an estimate of about
2.3-fold greater intrinsic clearance for CYP2D6. This would then
suggest that approximately 70% of CJ-12,458 metabolism is
CYP2D6-mediated and the remaining 30% is CYP3A4 mediated
values in
good agreement with those estimated from the inhibition data.
The contributions of CYP1A1 and 3A5 to the metabolism of CJ-12,458 are
more difficult to gauge. Both of these P450 enzymes are found primarily
in extrahepatic tissues. CYP1A1 is primarily associated with human lung
(Saarikoski et al., 1998; Mollerup et al., 1999) and intestine (Paine
et al., 1999), and is an inducible P450 enzyme. It is more typically
associated with metabolism of toxic environmental agents than with
drugs. CYP3A5 is expressed in intestine and liver in some individuals
and could contribute to metabolism during first pass (Kolars et al.,
1994; Gervot et al., 1996). However, since CJ-12,458 is not an orally
administered agent, but rather a metabolite formed in situ, intestinal
CYP3A5 would probably not contribute significantly to CJ-12,458
metabolism in vivo.
The hydration of CJ-12,458 represents a novel metabolic reaction. The
evidence obtained using stable isotopes is consistent with the
mechanism shown in Fig. 13. The 1°
alcohol and diol metabolites are proposed to arise from a common
epoxide intermediate, with the latter formed via hydrolysis of this
intermediate. The presence of 16O in the 1°
alcohol product when the incubation was conducted under an atmosphere
of 18O2 supports the
existence of the aldehyde intermediate that would be able to exchange
with solvent water via the hydrated form. This occurred in incubations
in human liver microsomes as well as with recombinant CYP3A4 and
CYP2D6. The proposed intermediary aldehyde has precedence, as it was
observed as a product in the in vitro metabolism of styrene by
cytochrome P450 (Mansuy et al., 1984) and it was proposed as an
intermediate in the metabolism of 
-methylstyrene in the rat (De
Costa et al., 2001). Through the use of biosynthesized deuterium
labeled NADPH, it was established that the addition of the reducing
equivalent (possibly as a hydride ion) occurs at the benzylic position
in human liver microsomal incubations. This would require an initial
hydride shift from the epoxide intermediate, a step precedented in the
well known NIH shift (Guroff et al., 1967; Boyd et al., 1972). However,
while the NIH shift is a 1,2-hydride shift, the mechanism proposed for the hydride shift in CJ-12,458 metabolism invokes a five-membered ring
transition state to the ortho position on the phenyl ring. This would then permit addition of the hydride to the benzylic position, yielding the result observed with deuterated NADPH. In
incubations using deuterated NADPH, 24% of the 1° alcohol product possessed deuterium when formed in human liver microsomes, whereas incorporation of deuterium was 66 and 6% when the 1° alcohol was formed by recombinant CYP3A4 and CYP2D6, respectively. Thus, the mechanism may exhibit some differences, depending on the CYP enzyme involved, and suggests a role for CYP itself in this biotransformation reaction. In consideration of these different extents of incorporation of deuterium from deuterated NADPH by CYP3A4 and CYP2D6, and the estimated contributions of these two enzymes in human liver microsomes, the results are consistent with the observation of 24% incorporation of deuterium when incubations are conducted with human liver microsomes and deuterated NADPH as cofactor.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 13.
Proposed mechanism of conversion of
CJ-12,458 to CP-616,762.
The designations of specific atoms of oxygen and hydrogen are as
follows: Oa, oxygen derived from air; Ow,
oxygen derived from water; Ha, hydrogen derived from the
alkene; Hr, hydrogen derived from the two
ortho positions of the benzene ring; Hw,
hydrogen derived from water; Hn, hydrogen derived from
NADPH. CJ-12,458 (1) undergoes epoxidation to generate
(2). The epoxide (2) is either hydrolyzed to
yield CP-611,781 (3) or undergoes a hydride shift to yield
intermediate aldehyde (4), which is a mixture of two
regioisomers. The oxygen of the aldehyde (4) is exchangeable
with water, via an hemiacetal intermediate. The proton on the
ortho position of the ring would be acidic and exchangeable,
as it is electronically linked to the position alpha to the aldehyde
carbonyl. Exchange of the oxygen and acidic proton would yield
(5). A hydride is added from NADPH at the benzylic
position to yield the product 1° alcohol
(6, CP-616,762).
|
|
It is difficult to rationalize the data concerning incorporation of
deuterium from deuterated water into 50% of the 1° alcohol product
and the loss of one deuterium in 50% of the 1° alcohol product when
beginning with deuterated CJ-12,458 as the substrate. It is
hypothesized that this results from exchange of protons from the
ortho position after hydride shift has occurred. These protons should possess some acidic character, being electronically linked to both the carbonyl of the aldehyde and the methoxy group. Thus, one of these protons would have been derived from the terminal position of the alkene, and they would be subject to exchange with
solvent. This would yield the results that
[2H5]CJ-12,458 could
yield [2H4]CP-616,762 in
H2O and that CJ-12,458 could yield
[2H1]CP-616,762 in
2H2O. However, the fact
that only 50% of the product underwent exchange with solvent is
difficult to rationalize, yet internally consistent since the result
was 50% in both of these "opposite" cases. This could be due to a
rate limitation on the exchange relative to the reduction step.
Alternately, this could be due to the fact that the two
ortho positions are not equivalent, and that the basic
benzylic nitrogen could participate as a general base catalyst in the
solvent exchange of the protons on one side via a five-membered ring
structure (Fig. 14). However, proof for this explanation is not presently available.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 14.
Five-membered ring structure wherein the
benzylic nitrogen can affect the acidity of the proton on ring position
6.
|
|
In conclusion, it was shown from these experiments that CJ-12,458 is
metabolized in human liver microsomes to two compounds, CP-611,781 and
CP-616,762, and that this metabolism is mediated by CYP2D6 and CYP3A4.
These data suggest that the pharmacokinetics of CJ-12,458 may exhibit
differences between CYP2D6 extensive and poor metabolizers (the latter
representing approximately 5-10% of the Caucasian population). Since
ezlopitant itself is primarily metabolized by CYP3A4 (to CJ-12,458 and
CJ-12,764), poor metabolizer subjects would be able to metabolically
generate CJ-12,458 from a dose of ezlopitant but may have a decreased
capability to subsequently clear this metabolite. The conversion of the
alkene to a 1° alcohol observed in human liver microsomes and
recombinant CYP enzymes is an interesting reaction. Some evidence was
presented to support a mechanistic hypothesis for this reaction;
further experimentation is required to fully elucidate the mechanism of
this metabolic transformation.
Technical assistance in the operation of the Finnigan LCQ from Loretta
Gernert and helpful discussions with Dr. Deepak Dalvie, Dr. Alfin Vaz,
and Dr. Neal Castagnoli, Jr. on the mechanistic work are gratefully acknowledged.
R. S. Obach, Drug
Metabolism Department, Pfizer, Inc., Groton, CT 06340. E-mail:
obachrs{at}groton.pfizer.com
Abbreviations used are:
CYP, cytochrome P450;
HPLC, high-pressure liquid chromatography;
TLC, thin-layer
chromatography;
MTBE, methyl-t-butyl ether;
MS, mass
spectrometry.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
![v=<FR><NU>V<SUB><UP>max</UP>(<UP>1</UP>)</SUB> · [S]</NU><DE>K<SUB>M(1)</SUB>+[S]</DE></FR>+<UP>CL′</UP><SUB><UP>int</UP>(<UP>2</UP>)</SUB>](/content/vol29/issue7/fulltext/1057/img001.gif)
![% <UP>control activity</UP>=<UP>A</UP>−<FR><NU><UP>B · </UP>[<UP>I</UP>]</NU><DE>[<UP>I</UP>]+<UP>IC<SUB>50</SUB></UP></DE></FR>](/content/vol29/issue7/fulltext/1057/img002.gif)
