Department of Medicinal Chemistry, Merck Frosst Center for
Therapeutic Research, Pointe-Claire-Dorval, Quebec, Canada (C.L., N.C.,
L.A.T., D.A.N.-G., J.M.S., D.M., H.P., J.A.Y.); and Celltech
Chiroscience Ltd., Slough, Berkshire, United Kingdom (T.P., R.A.,
G.W.)
CDP-840 is a selective and potent phosphodiesterase type IV
inhibitor, whose in vitro metabolism profile was first investigated using liver microsomes from different species. At least 10 phase I
oxidative metabolites (M1-M10) were detected in the microsomal incubations and characterized by capillary high-performance liquid chromatography continuous-flow liquid secondary ion mass
spectrometry (CF-LSIMS). Significant differences in the microsomal
metabolism of CDP-840 were found between rat and other species. The
major route of metabolism in rat involved
para-hydroxylation on the R4 phenyl. This pathway was
not observed in human and several other species. The in vitro
metabolism profile of CDP-840 was further examined using freshly
isolated hepatocytes from rat, rabbit, and human. The hepatocyte
incubations indicated more extensive metabolism relative to that in
microsomes. In addition to the phase I oxidative metabolites observed
in microsomal incubations, several phase II conjugates were identified
and characterized by CF-LSIMS. Interspecies differences in phase II
metabolism were also found in these hepatocyte incubations. The major
metabolite in human hepatocytes was identified as the pyridinium
glucuronide, which was not detected in rat hepatocytes. Simple
structural modification on R4, such as p-Cl
substitution, greatly reduced the species differences in microsomal
metabolism. Furthermore, modifications on R3, such as the
N-oxide, eliminated the N-glucuronide
formation in human. These results not only helped in determining the
suitability of animal species used in the preclinical safety studies
but also provided valuable directions for the synthetic efforts in
finding backup compounds that are more metabolically stable.
 |
Introduction |
Cyclic phosphodiesterase
(PDE1)
enzymes catalyze the hydrolysis of the 3'-phosphoester bonds of cAMP
and cGMP to form the corresponding AMP and GMP, and therefore, are
involved in controlling the intracellular concentrations of cAMP and
cGMP (Torphy, 1998
). At least 11 mammalian PDE isozyme families have
been reported, each encoded by a distinct gene. They are distinguished
on the basis of their enzyme kinetics, substrate selectivity, and
tissue distribution. The type IV family of phosphodiesterases (PDE-IV)
is a high-affinity cAMP-selective isozyme, and has been found in almost
all cell types that have been implicated in asthma pathogenesis
(O'Brien, 1997). Selective PDE-IV inhibitors, therefore, could become
promising therapeutic agents for the treatment of asthma and a wide
range of other inflammatory diseases (Torphy et al., 1994). CDP-840
{R-[+]-4-[2-(3-cyclopentyloxy-4-methoxy phenyl)-2-phenyl ethyl] pyridine} is a potent and selective PDE-IV inhibitor (Hughes et al., 1997; Perry et al., 1998), and was in development for the treatment of asthma. The structure of CDP-840 is
shown in Fig. 1.
View larger version (9K):
[in this window]
[in a new window]
|
Fig. 1.
Chemical structure of CDP-840,
R-[+]-4-[2-(3-cyclopentyloxy-4-methoxy
phenyl)-2-phenyl ethyl]pyridine, CT2412, and CT2481.
The substructures are labeled as R1 (cyclopentyl), R2 (catechol phenyl
ring), R3 (pyridyl), and R4 (phenyl).
|
|
Drug metabolism studies during the early drug discovery stage are
becoming increasingly important. They provide information useful for
many aspects of drug discovery, such as drug design, pharmacokinetic
evaluation, and toxicity assessment. Biotransformations can be
performed in vitro with the microsomal or cytosolic fraction of liver
tissue as the enzyme source under the appropriate incubation conditions. Microsomal incubations are useful to pinpoint specific pathways (oxidation or glucuronidation), and can be used to gain an
understanding of potential interspecies differences in metabolism (Kumar et al., 1999). Hepatocyte incubations retain phase I and phase
II enzyme activities, and therefore, are useful to determine overall
metabolism, and mimic in vivo metabolism more accurately than
incubations with subcellular fractions (Placidi et al., 1997). In vitro
incubations with liver microsomes and/or hepatocytes can be used as a
means to predict potential biotransformations in humans and in those
species used for preclinical safety studies.
Identification of metabolic pathways of drug candidates significantly
relies on innovations in analytical chemistry. The role of mass
spectrometry, especially liquid chromatography/mass spectrometry (LC/MS), in drug metabolite identification has become evident and
increasingly important in recent years (Vrbanac et al., 1992; Iwabuchi
et al., 1994; Davis and Baillie, 1995; Jackson et al., 1995). Capillary
HPLC/continuous-flow liquid secondary ion mass spectrometry (CF-LSIMS)
is a powerful LC/MS technique for metabolite identification (Moritz et
al., 1992; Onisko et al., 1994), and has been used effectively and
routinely in our laboratory for metabolism studies (Nicoll-Griffith et
al., 1993; Li et al., 1995). In this article, the metabolic profiles of
CDP-840 were investigated using liver microsomes and hepatocytes
prepared from different species. The metabolites were rapidly
characterized by CF-LSIMS, supplemented with the use of NMR analysis
and comparison to authentic standards wherever possible. Based on the
in vitro metabolism results obtained with CDP-840, several backup
compounds with simple structural modifications were evaluated, such as
p-Cl substitution on R4 phenyl (CT2412) and
N-oxide on R3 (CT2481) (Fig. 1). These simple modifications
significantly improved the metabolism profile and metabolic stability.
Our in vitro metabolism studies proved to be extremely useful, not only
for predicting in vivo metabolism in animal models and in humans but
also for guiding the medicinal chemistry efforts for structural
optimization of the lead compound CDP-840.
| |
Experimental Procedures |
Materials.
CDP-840 (Warrellow et al., 1997, example 16), CT2412, CT2481 were
synthesized at Celltech (Slough, UK). All synthetic metabolite standards were synthesized at Celltech or Merck Research laboratories. The synthesis and characterization of some of the metabolites were
described (Warrellow and Alexander, 1996; Warrellow et al., 1997). NADP
(Na+ salt) and glucose-6-phosphate dehydrogenase
were purchased from Sigma Chemical Co. (St. Louis, MO). All solvents
used were obtained from commercial sources and were of HPLC grade.
Microsomal Oxidative Incubations.
Hepatic microsomes were prepared from frozen livers (human, rat,
rabbit, ferret, rhesus monkey, etc.) according to a standard procedure
(Lu and Levin, 1972). Incubations with microsomes prepared from
different species were typically conducted under linear conditions with
80 µM CDP-840 (or CT2412, CT2481) and 0.5 mg of microsomal protein in
the presence of an NADPH-generating system as previously described
(Nicoll-Griffith et al., 1993). Incubations (0.5 ml) were conducted for
30 min at 37°C and were quenched by the addition of an equal volume
of acetonitrile. Precipitated proteins were removed by centrifugation
(10,000 rpm; Eppendorf centrifuge model 5415C) for 10 min. Supernatants
were analyzed by HPLC/UV and HPLC/CF-LSIMS. Blank incubations
containing no drug and control incubations containing boiled microsomes
were also conducted at the same time.
Microsomal Glucuronidation.
Glucuronidation was achieved using an incubation of 200 µM CDP-840
(or CT2412, CT2481) with 1.0 mg of microsomal protein for 30 min. The
incubation was conducted with 12.5 mM MgCl2, 12.5 mM uridine 5'-diphosphate glucuronic acid (UDPGA) and 20 mM
D-saccharic acid-1,4-lactone in a phosphate buffer at pH
6.6. The incubation mixture (0.5 ml) was quenched with an equal volume
of acetonitrile. Precipitated proteins were removed by centrifugation
for 10 min, and supernatants were analyzed by HPLC/UV and
HPLC/CF-LSIMS.
Hepatocyte Incubations.
Rat hepatocytes were isolated from male Sprague-Dawley rats by
collagenase perfusion of liver as described previously (Silva et al.,
1998). Fresh human liver tissue was obtained from consenting donors
undergoing partial hepatectomies and from unused liver portions from
patients undergoing liver transplants (St. Luc Hospital, Montreal,
Canada). The tissues used were morphologically healthy. Human
hepatocyte isolation was conducted by a two-step collagenase perfusion
of the liver sample as described by Silva et al. (1998). Rabbit
hepatocytes were prepared in a similar way to the human hepatocytes.
Incubations were conducted with 50 µM CDP-840 (CT2412, CT2481) and
2 × 106 isolated hepatocyte cells/ml. The
cell mixture was incubated in Krebs-Henseleit buffer containing 12.5 mM
HEPES (pH = 7.4) at 37°C under 95% air and 5%
CO2 for 3 h, and was quenched by the
addition of an equal volume of acetonitrile. Precipitated proteins were
removed by centrifugation for 10 min. Supernatants were analyzed by
HPLC/UV and HPLC/CF-LSIMS.
In Vivo Pharmacokinetics and Metabolism.
CDP-840 and CT2412 were administered p.o. at 20 mg/kg in 1% methocel
(pH = 2.0) to rats, and were also administered i.v. at 5 mg/kg in
saline to rats. CDP-840 was dosed p.o. at 10 mg/kg to rabbits, and 15 mg b.i.d. to humans (Harbinson et al., 1997). The plasma samples
obtained postdosing were quenched with an equal volume of acetonitrile,
and analyzed by HPLC/UV. Selected samples were analyzed by LC/MS.
Isolation of Metabolites.
To prepare ~0.1 mg of metabolites M12 and M17 for NMR
characterization, human microsomal incubations in the presence of UDPGA were scaled up appropriately. The isolation was similar to that described previously (Li et al., 1995), using a preparative Waters Novapak C18 column (7.5 × 300 mm), and a
Waters 990 diode array detector. Separation of metabolites and parent
compound was carried out using a linear gradient of 60% A to 90% A
(A = CH3OH; B = 20 mM
NH4OAc, pH 5.0) in 30 min at a flow of 4 ml/min.
Metabolite peaks were collected manually, concentrated, and desalted
using 1 ml of BondElut C18 SPE cartridges, and
dried in a Heto-Vac CT110 vacuum centrifuge (Heto Lab Equipment,
Berkerod, Denmark).
CF-LSIMS.
A JEOL HX110A double focusing mass spectrometer (EB configuration;
JEOL, Boston. MA), equipped with a 10-kV LSIMS source and a cesium ion
gun was used in this study. The mass spectrometer was operated in the
CF-LSIMS mode, and was described in detail previously (Li et al.,
1995). Data acquisition was in positive ion mode and the mass
spectrometer was scanned at a rate of 4 s from
m/z 0 to 1000 Da. Separation of CDP-840 and its
metabolites was achieved using a KAPPA Hypersil BDS
C18 (0.30 × 100-mm) capillary column (Keystone
Scientific, Bellefonte, PA). The flow rate was 3 µl/min, which was
obtained by splitting the main flow (1 ml/min) using a Valco tee (Valco
Instruments, Houston, TX). A linear gradient was used from 50% A to
90% A over 40 min (A = methanol, B = 20 mM ammonium acetate
adjusted to pH 5.0 with acetic acid; each solvent contained 1.5%
glycerol). The incubation supernatant was diluted 5-fold with aqueous
mobile phase, and 10 µl was injected onto the capillary HPLC column.
The capillary column was flushed with 100% methanol whenever not used
for analysis.
MS/MS analysis was carried out using a B/E linked scan. The collision
gas used was helium and the pressure was adjusted such that the
intensity of the precursor ion was reduced by 50%.
NMR Characterization of Glucuronide Conjugates M12 and M17.
NMR spectra were acquired at 22°C on a Bruker AMX 500 equipped
with a 5-mm inverse broad-banded probehead. Samples were dissolved in
160 µl of dimethyl sulfoxide-d6 and
placed in a Shigemi symmetrical microtube matched to the solvent.
One-dimensional 1H and two-dimensional double
quantum-filtered correlation spectra, heteronuclear multiple quantum
correlation spectra, and heteronuclear multiple bond correlation
spectra were acquired with standard parameters.
| |
Results |
Oxidative Microsomal Metabolism of CDP-840.
The in vitro metabolism profiles of CDP-840 in humans and potential
safety animal species were investigated using microsomal preparations.
The incubation mixtures of CDP-840 with microsomal proteins from
various species under oxidative conditions were first analyzed by
analytical HPLC with UV-photodiode array detection. The percentage of
metabolism was calculated using the LC/UV peak area ratio of CDP-840 in
an incubation with a control incubation (without NADPH). The
rate of microsomal metabolism (nmol/min · mg) of CDP-840 in
different species is summarized in Table
1. The UV spectra of the metabolites were
similar to that of CDP-840, except for one metabolite, M5. LC/UV
profiles of these microsomal incubations presented not quantitative but
relative percentages of most of the metabolites generated. The
incubation samples were also analyzed using capillary HPLC CF-LSIMS,
which did not have UV on-line. The total ion trace has high background
consisting of glycerol adduct ions (Li et al., 1995), because glycerol
was added to the mobile phase and used as matrix for CF-LSIMS.
Reconstructed ion chromatograms of the molecular ions for the parent
and all metabolites were therefore presented, as shown in Fig.
2. The reconstructed positive ion
chromatograms showed very similar percentage of each metabolite
compared with the analytical HPLC/UV traces. The microsomal metabolism
profile of CDP-840 in rat differed significantly from other species. At
least 10 different metabolites were detected, named M1 through M10. The
LSI mass spectra of CDP-840 and three representative phase I
metabolites, M5, M6, and M9, are shown in Fig.
3, a-d, and the fragmentation patterns
of phase I metabolites are summarized in Table
2.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 2.
Reconstructed ion chromatograms (sum of
MH+ of CDP-840 and its metabolites) that represent the
metabolism profiles of CDP-840 in incubations from human microsomes
(a), rabbit microsomes (b), and rat microsomes (c).
|
|
The positive ion primary mass spectrum of CDP-840 (Fig. 3a) displayed
an intense molecular ion [M + H]+ at
m/z 374 and a minor glycerol adduct [M + H]+ at m/z 466. Three
unique fragment ions, A, B, and C, at m/z 281, 213, and 182 were detected. These three fragment ions represented the
sequential elimination of methyl pyridyl, cyclopentyl, and methoxy
moieties from the protonated molecule (Table 2). Collision-induced dissociation of the molecular ion (B/E linked scan) also produced the
same fragment ions (data not shown), and therefore, did not offer any
additional structural information. These characteristic fragmentations
were subsequently used to determine which of the substructures had
undergone metabolism.
The mass spectrum of M1 showed an abundant molecular ion at
m/z 360, which is 14 amu less than that of
CDP-840. Two diagnostic fragment ions (A and B) at
m/z 267 and 199 were also shifted 14 amu lower
relative to the fragment ions of CDP-840. This implied that M1 was an
O-desmethyl metabolite.
The primary LSI mass spectra of M2 to M9 all showed abundant molecular
ions at m/z 390, i.e., 16 amu higher than that of
CDP-840, indicating that they are different mono-oxygenated
metabolites. The mass spectrum of M2 showed fragment ions A and B at
m/z 297 and 229, also 16 amu higher than the
corresponding fragments 281 and 213 of CDP-840, suggesting that the
methyl pyridyl and cyclopentyl moieties were not altered. The exact
site of hydroxylation could not be assigned based on its LSI mass
spectrum or MS/MS spectrum. M2 was speculated to be the monohydroxy
metabolite on the C2 position of the ethyl link (Fig. 9), based on its
coelution and identical mass spectrum with an available authentic
standard (racemic). The chirality of this metabolite was not determined.
The primary LSI mass spectra of M3 and M8 were identical, both showed
fragment ions A and B at m/z 281 and 213, suggesting that oxidation occurred on the methyl pyridyl moiety. M8 was
confirmed as a monohydroxy metabolite on the C1 position of the ethyl
link (Fig. 9), based on coelution and similar mass spectrum with an authentic standard (chiral with C1: S and C2: R). M3 was likely to be
the diastereoisomer of M8. A notable minor fragment ion at
m/z 373 was present in the primary mass spectra
of M2, M3, and M8, but was not present in the corresponding product ion
MS/MS spectra (B/E linked scan) of [M + H]+.
This fragment ion is postulated to arise from the loss of OH radical
from [M + H]+ by the initial high-energy
ionization process, and seemed to be diagnostic of these two types of hydroxylation.
The primary mass spectrum of M4 showed strong fragment ions A and B at
m/z 297 and 213, indicating that the
hydroxylation was on the R1 cyclopentyl group. The exact position of
hydroxylation and the stereochemistry could not be determined based on
its mass spectrum. Similar mass spectra were obtained for M7 and M9
(Fig. 3d), suggesting that they were either positional isomers or
diastereoisomers of M4.
In the primary mass spectrum of M5 (Fig. 3b), fragment ions A and B at
m/z 281 and 213 suggested that oxidation was on
the methyl pyridyl moiety. Another abundant fragment ion at
m/z 374 was also detected. This fragment is
characteristic of an N-oxide, and was produced by the
high-energy ionization process, but was not present in the MS/MS
spectrum of [M + H]+. Metabolite M5 was
confirmed as the R3 N-oxide with the authentic standard.
This metabolite was the only one which showed different UV spectrum
with that of CDP-840. The extinction coefficient for the
N-oxide is about 3 times greater than that of CDP-840. When a standard solution containing equal concentrations of the
N-oxide and CDP-840 was analyzed by CF-LSIMS, the intensity
of MH+ of the N-oxide
(m/z 390) was about half of the
MH+ of CDP-840 (m/z 374),
but the N-oxide also produced a fragment ion at
m/z 374 with equal intensity with
MH+ at m/z 390. When the
extracted ion chromatograms for 374 and 390 were summed, the resulting
reconstructed ion chromatogram showed nearly the same response for the
two compounds. Therefore, we believe that the reconstructed CF-LSIMS
profiles shown in Fig. 2 represented the relative amount of metabolites generated.
The primary LSI mass spectrum of M6 (Fig. 3c), a major metabolite in
the rat microsomal incubation, showed abundant fragment ions at
m/z 229 and 197, similar to that of M2, except
for the absence of a fragment ion at m/z 373. This metabolite was confirmed as the R4 para-phenol, as it
coeluted with and showed a mass spectrum identical to the authentic standard.
The LSI mass spectrum of M10 showed an abundant molecular ion [M + H]+ at m/z 306, 68 amu
less than that of CDP-840. Fragment ion A at m/z
213 represented the loss of methyl pyridyl from [M + H]+. The spectrum was consistent with the
O-descyclopentyl metabolite, and was confirmed using an
authentic standard.
Metabolism of CDP-840 in Hepatocytes.
Incubations of CDP-840 were carried out with hepatocytes isolated from
rat, rabbit, and human livers. The reconstructed positive ion
chromatograms (summed MH+ of CDP-840 and all
metabolites) are shown in Fig. 4. The
rate of metabolism (nmol/h · 106 cells) in
these hepatocyte incubations was summarized in Table 1. Both phase I
oxidative metabolites (M1-M10) and phase II conjugates were detected.
The formation of the phase II metabolites differed significantly
between the three species. These phase II metabolites were
characterized by CF-LSIMS, and in some cases were also confirmed by
NMR. The mass spectral fragmentation patterns of these phase II
metabolites are summarized in Table 2.
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 4.
Reconstructed ion chromatograms (sum of
MH+ of CDP-840 and its metabolites) that represent the
metabolism profiles of CDP-840 in incubations from rat hepatocytes (a),
rabbit hepatocytes (b), and human hepatocytes (c).
|
|
M11 was present in the rat and rabbit hepatocyte incubations, but not
in the human hepatocyte incubation. The primary LSI mass spectrum of
M11 showed an abundant molecular ion [M + H]+
at m/z 536. Fragment ions at
m/z 360, 267, and 199 were observed, and the
spectrum below m/z 400 was identical to that of
M1, the O-desmethyl metabolite. The mass difference between
[M + H]+ (536) and 360 is 176 (dehydrated
glucuronic acid), consistent with a glucuronide conjugate. The exact
site of glucuronide attachment (either an O-linked or
N-linked glucuronide) could not be determined from the LC/MS
or LC/MS/MS spectrum.
M12 was detected in both rabbit and human hepatocytes; in fact it was
the major metabolite in human hepatocytes. However, rat hepatocytes did
not produce M12. The primary LSI mass spectrum of M12 (Fig.
5a) showed an abundant molecular ion at
m/z 550. The spectrum below
m/z 400 was very similar to that of CDP-840, with
fragment ions at m/z 374, 281, and 213. The mass
difference between molecular ion (550) and 374 is 176, consistent with
a glucuronide conjugate. High-energy collision-induced dissociation of
[M + H]+ at m/z 550 also
produced similar fragment ions (data not shown). The pyridine nitrogen
is the only functional group in the CDP-840 structure to which a
glucuronide could be directly linked; therefore, it was proposed that
M12 was a pyridinium glucuronide based on its mass spectrum.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 5.
CF-LSI background-subtracted mass spectra of
phase II metabolites M12 (a), M15 (b), M17 (c), and M20 (d).
|
|
M14 and M15 were present only in the rabbit hepatocyte incubation. The
primary LSI mass spectra exhibited abundant molecular ions [M + H]+ at m/z 566. The mass
spectrum of M15 (Fig. 5b) below mass 400 amu resembled that of M9, the
monohydroxy cyclopentyl metabolite. M15 was postulated to be the
glucuronide conjugate of a hydroxy cyclopentyl, although the
possibility of a hydroxy cyclopentyl, pyridinium glucuronide could not
be ruled out. M14 was very minor, and its mass spectrum was too weak to
produce useful fragment ions. It was likely an isomer of M15.
Metabolite M19 was also formed only in the rabbit hepatocyte
incubation. Ions at m/z 386 and 403 corresponded
to [M + H]+ and [M + NH4]+. Fragment ions at
m/z 213 and 306 were detected, suggesting that the cyclopentyl group was lost. The mass difference between [M + H]+ 386 and fragment ion 306 was 80, implying a
sulfate conjugate (Fig. 10).
Several metabolites such as M6', M13, M16, M17, M18, and M20 were
detected in the rat hepatocyte incubation, but not in the human
hepatocyte incubation. The mass spectrum of M17 showed abundant molecular ions [M + H]+ at
m/z 566. The spectrum below mass 400 amu was
identical to that of M6, the para-phenol metabolite. M17
was, therefore, identified as R4 para-hydroxy glucuronide
conjugate (Fig. 5c).
The mass spectrum of M13 (data not shown) displayed a strong [M + H]+ at m/z 408, 34 amu
higher than that of CDP-840. Fragment ions at m/z
390, 297, and 229 were detected, consistent with a dihydrodiol on the
R4 phenyl. The mass spectrum of M16 (data not shown) showed a molecular
ion [M + H]+ at m/z 406, 32 amu higher than that of CDP-840, indicating a dioxygenated
metabolite, and was proposed to be an R4 catechol.
Metabolite M6', which eluted very closely with M6, showed a protonated
molecule at m/z 420, and major fragment ions at
m/z 327 and 259. The LSI mass spectrum of M6'
exhibited a very similar pattern to that of M6, except that all the
ions were shifted 30 amu higher. The increment of 30 amu could imply an
additional methoxy group on the R4. This metabolite was proposed to be
the R4 O-methyl catechol.
Metabolite M18 showed abundant molecular ion at
m/z 596, the LSI mass spectrum exhibited a very
similar pattern to that of M17, except that all the ions were shifted
30 amu higher. The increment of 30 amu would again imply an additional
methoxy group on the R4 phenyl (see structure in Fig. 10).
The LSI mass spectrum of metabolite M20 displayed a strong [M + H]+ at m/z 697. Fragment
ions at m/z 229, 281, 297, 374, and 390 were
detected. The spectrum suggested that M20 was a glutathione adduct
(Fig. 5d).
Glucuronidation of CDP-840 in Microsomes.
A proposed pyridinium glucuronide metabolite M12 was detected in rabbit
and human hepatocytes. If the proposed structure is correct, the same
metabolite might be formed in the incubation of CDP-840 with microsomal
proteins from various species in the presence of UDPGA. Indeed, M12 was
detected in the incubations with human, rabbit, rhesus monkey, and
guinea pig microsomes in the presence of UDPGA at rates of 1.23, 0.17, 0.03, and 0.02 nmol/min · mg of protein. This metabolite was not
detected in the incubations with rat, dog, mouse, or ferret microsomes,
nor was it not detected in the rat hepatocyte incubation.
M12 was isolated from the human microsomal incubation by
semipreparative HPLC. The structure was identified as 
-linked
pyridinium glucuronide (Fig. 6; Table
3) by NMR. The linkage between the glucuronide anomeric carbon and the pyridyl nitrogen was confirmed by
long-range correlation experiments. In particular, a three-bond correlation between the glucuronide anomeric proton (H-26) and the two
carbons adjacent to the pyridyl nitrogen (C-11 and C-12) was observed.
In addition, the reverse correlation between the two protons adjacent
to the pyridyl nitrogen (H-11 and H-12) and the glucuronide anomeric
carbon (C-26) was also observed, confirming the presence of the
pyridinium glucuronide species depicted in Fig. 6. A coupling constant
of 8.6 Hz was observed at the anomeric proton, indicating that the
glucuronide linkage has the configuration.
When the authentic standard of M6 (R4 para-phenol) was
incubated with rat microsomal protein in the presence of UDPGA, a
metabolite was also detected. This metabolite had the same retention
time and LSI mass spectrum as those of M17 in rat hepatocytes. M17 was
isolated from the microsomal incubation by prep-HPLC, and NMR studies
confirmed its structure as R4 para-O-glucuronide
with the configuration.
In Vitro Metabolism of CT2412 and CT2481.
CT2412 and CT2481 are two synthetic analogs (see structures in Fig. 1)
of CDP-840. Both have a para-Cl substitution on the R4
phenyl, and CT2481 also has a simple modification on R3, i.e., N-oxide. The in vitro metabolism profiles of these two
analogs were investigated using microsomal preparations from rat and
rhesus monkey. The microsomal oxidative metabolism profiles of CT2412 in rat and rhesus monkey are shown in Fig.
7. The profiles were clearly similar in
the two species. The metabolites were identified analogously to those
described for CDP-840 metabolites. They were descyclopentyl (1),
hydroxy cyclopentyl (2 and 3), N-oxide (4), and desmethyl
(5) metabolites. When CT2412 was incubated with microsomes under
glucuronidation conditions, the N-glucuronide metabolite of
CT2412 was detected in human microsomes (3%) and rabbit microsomes
(3%), but to a lesser extent compared with that of CDP-840. The
microsomal oxidative metabolism profiles of CT2481 in rat and rhesus
monkey were also identical (data not shown), with descyclopentyl,
hydroxy cyclopentyl, and desmethyl being the major metabolites. The
reduction of N-oxide to free pyridyl was also observed, but
was very minor based on LC/UV and LC/MS data. In the glucuronidation
studies, the N-glucuronide metabolite was not detected for
CT2481 in human or rabbit microsomes supplemented with UDPGA.
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 7.
Microsomal metabolism profiles of CT2412 in
rat and rhesus monkey by HPLC/UV (at 245 nm).
Metabolite peaks were assigned by LC/MS, descyclopentyl (1), hydroxy
cyclopentyl (2 and 3), N-oxide (4), and desmethyl (5).
|
|
The metabolism profiles of CT2412 and CT2481 in rat hepatocytes were
also investigated, and were similar to those in microsomes with
addition of phase II glucuronide conjugates (+O+gluc, data not shown).
The phase II metabolism, however, was significantly reduced compared
with that of CDP-840 in rat hepatocytes.
In Vivo Pharmacokinetics of CDP-840 and CT2412.
Rat plasma concentrations of CDP-840 and CT2412 following p.o. or i.v.
dosing were determined by HPLC with UV detection, and were plotted in
Fig. 8. The levels of CDP-840 in rat
plasma after 20-mg/kg p.o. dosing were not detectable. In the
i.v.-dosed (5 mg/kg) rats, CDP-840 had a plasma concentration of 4.4 µM (average of three rats) at 5-min postdosing, and no detectable
levels of CDP-840 were found after 2 h. In healthy male volunteers
at 16 mg b.i.d dosing with CDP-840, the half-life of CDP-840 was
determined to be 6 h. For CT2412, plasma levels in p.o.-dosed rats
were detected with a Cmax of 0.44 µM at
2 h. In i.v.-dosed rats, a similar concentration (3.7 µM) of
CT2412 at 5-min postdosing was observed compared with that of CDP-840,
however, the levels of CT2412 could be detected up to 4 h (0.3 µM) postdosing.
In Vivo Metabolism of CDP-840.
The plasma samples obtained after p.o. dosing to rat, rabbit, and human
were analyzed. The major circulating metabolites observed in rat plasma
were R4 para-phenol (M6) and N-oxide (M5). In
rabbit plasma, the hydroxy cyclopentyl (M9) was the major metabolite. In human plasma, pyridinium glucuronide (M12), N-oxide (M5)
and hydroxy cyclopentyl (M9) were detected as major metabolites.
| |
Discussion |
The in vitro metabolism profiles of CDP-840 were evaluated using
hepatic microsomes and freshly isolated hepatocytes from different
species. With the combination of HPLC/UV, LC/MS, NMR, and available
synthetic standards, in vitro phase I and phase II metabolites of
CDP-840 were characterized. Primary CF-LSI mass spectra of CDP-840 and
its metabolites all exhibited abundant molecular ions and unique
fragment ions, allowing the sites of metabolism to be easily and
quickly pinpointed to a particular substructure. NMR data and synthetic
standards provided confirmation of metabolite structures. In our study
of CDP-840 metabolism, many synthetic standards had been or were
rapidly synthesized, making some of the metabolite identification much
more straightforward.
The phase I oxidative metabolism profile of CDP-840 was evaluated using
microsomal proteins from various species. Up to 10 oxidative
metabolites were detected. The reconstructed CF-LSIMS ion chromatograms
of all MH+ of metabolites and CDP-840 showed very
similar profiles and peak ratios as the analytical HPLC/UV traces. The
pyridyl nitrogen is likely the site for protonation, and therefore the
CF-LSIMS responses for the metabolites and CDP-840 were believed to be very similar. The sum of the reconstructed ion chromatogram (Fig. 2)
reflected not quantitative but relative amount of metabolites generated. Several sites of CDP-840 were found to undergo
biotransformation. The R1 cyclopentyl group was the major site for
metabolism in most species. Several hydroxy cyclopentyl metabolites
were detected. On the R2 catechol ring, both desmethyl and
descyclopentyl metabolites were observed. Similar biotransformations
were also reported for rolipram (Krause and Kuhne, 1992, 1993), a
PDE-IV inhibitor from Schering, which has the same catechol moiety. On
the R3 pyridyl, N-oxidation was detected, and on the R4
phenyl, para-hydroxylation was observed. The oxidative
metabolism of CDP-840 in human microsomes was similar to those in
hepatic microsomes from rhesus monkey, guinea pig, dog, ferret, rabbit,
and mouse, although the relative percentages of each metabolite varied
in these different species. The metabolism profile in rat microsomes,
however, differed significantly from these species. The major site of
metabolism in rat involved the para-hydroxylation (M6) on
the R4 phenyl ring. M6 was detected as a minor metabolite in rabbit
microsomes, and was not detected in the microsomal incubations of
CDP-840 with human and other species. The oxidative microsomal
metabolism pathways of CDP-840 are summarized in Fig.
9.
Phase II metabolism of CDP-840 was evaluated using hepatocytes. Freshly
isolated hepatocytes retain phase I and phase II enzyme activities, and
therefore, should provide a better correlation with in vivo metabolism
(Placidi et al., 1997; Nicoll-Griffith et al., 1999). Rat hepatocytes
are used as our primary in vitro system for checking overall metabolic
stabilities of synthetic compounds at the early drug discovery stage,
and proved to be very useful to determine the overall metabolism (both
phase I and phase II simultaneously). Because significant interspecies difference in metabolism were found between rat and other species from
microsomal studies, hepatocytes isolated from different species were
prepared, and their metabolism profiles were compared. Metabolism profiles of CDP-840 in rat, rabbit, and human hepatocytes indicated more extensive metabolism relative to that in microsomes, and provided
evidence of additional interspecies differences in the phase II
metabolism. The in vitro phase II metabolism pathways of CDP-840 are
summarized in Fig. 10.
In the human hepatocyte incubation of CDP-840, the phase II metabolism
dominated, and the major metabolite was confirmed as R3 pyridinium
glucuronide (M12). It was also found to be the major circulating
metabolite of CDP-840 in human plasma. This metabolite was not detected
in the rat hepatocyte incubation. The formation of pyridinium
glucuronides has been reported in the literature for tripelennamine
(Yeh, 1991) and nicotine (Byrd et al., 1992). The food pyrolysis
product 2-amino-1-methy-6-phenylimidazo[4,5-b] pyridine also forms
pyridinium glucuronides in human and rabbit microsomes under UDPGA, but
not in rat microsomes (Styczynski et al., 1993).
The metabolism profile of CDP-840 in rat hepatocytes was very different
from that in human hepatocytes. The extent of metabolism was also
greater in rat compared with that in human or rabbit. Several
metabolites (M6', M13, M16, M17, M18, and M20) were detected in rat
(some were also in rabbit) hepatocytes, but were not present in human
hepatocytes. These metabolites were probably formed through an epoxide
intermediate on the R4 phenyl. The epoxide could isomerize nonenzymatically to the R4 para-phenol (M6). It could also
convert to a R4 dihydrodiol (M13) possibly by epoxide hydrases. The
stable diol could further dehydrate enzymatically to yield the
para-phenol (M6), or be oxidized by a dehydrogenase to
generate a diphenolic metabolite (M16). The epoxide could also be
susceptible to conjugation (M20) with glutathione by glutathione
S-epoxide transferase. The formation of aromatic epoxides
and further metabolized products such as phenols, dihydrodiols, and
glutathione conjugates is well documented in the literature, as in the
case of naphthalene (Jerina et al., 1970) and
monohalogenobenzenes (Parke, 1968). The R4 diphenolic metabolite (M16)
could undergo O-methylation to form M6'.
O-Methylation of catechol metabolites formed by the
epoxide-dihydrodiol pathway is in the literature for diphenylhydantoin
(Glazko, 1973). The reaction was mediated by catechol
O-methyl transferase (Bakke, 1970). Both M6' and M6
(R4 para-phenol) could also further conjugate with
glucuronic acid to form the corresponding glucuronides M17 and M18.
In the development of any potential therapeutic drug, the preclinical
safety studies are crucial to evaluate efficacy, pharmacokinetics, and
toxicity. It is very important that the safety species produce the same
metabolites as those found in humans. Rats are typically used as the
early preclinical safety species. The profound interspecies differences
in the in vitro metabolism of CDP-840 between rat and human would make
the preclinical safety studies in the rat invalid. Species differences
in metabolism became a critical issue, and one of the important
criteria to select potential backup compounds was to eliminate these
differences. CT2412 was quickly identified, it has similar potency
against PDE-IV (GST-met248A assay: 6.7 nM) compared with that of
CDP-840 (4.3 nM). The para-Cl substitution on R4 eliminated
the R4 para-phenol metabolite and other epoxide-mediated metabolites in rat. This simple substitution improved the metabolic stability in rat, and most importantly, similar metabolism profiles of
CT2412 were observed in the microsomal incubations with rat and rhesus
monkey. However, in the glucuronidation studies, the N-glucuronide metabolite was still detected in human
microsomes and rabbit microsomes at a rate of 0.1 nmol/min · mg
of protein. This metabolite would probably be expected in human in
vivo, but not in rat. Therefore, it was important to eliminate this
metabolite, and CT2481, which had N-oxide on R3, was again
quickly identified. CT2481 is also equally potent against PDE-IV
(GST-met248A assay: 9.2 nM). The N-oxide effectively blocked
the N-glucuronide metabolite formation in human. Among
several N-oxide compounds we investigated, the reduction of
N-oxide to free pyridyl appeared to be a very minor
biotransformation pathway both in vitro and in vivo; therefore. N-Glucuronidation is unlikely to occur in vivo with the
N-oxide compounds.
In the in vivo pharmacokinetic studies of CDP-840 and CT2412 in rats,
it was found that CDP-840 had extremely poor bioavailability in rats
(0%). No plasma levels of CDP-840 were detected in the p.o.-dosed
rats, and CDP-840 also had very short half-life in the i.v.-dosed rat.
These results are consistent with the fact that CDP-840 was extensively
metabolized in rat from the in vitro studies. In healthy male
volunteers with 16 mg of CDP-840 dosed b.i.d., the half-life was
determined to be 6 h; this is consistent with the in vitro
results, which showed less metabolism of CDP-840 in human than in rat.
For CT2412, the improvement in metabolic stability in vitro was also
reflected in the pharmacokinetics in vivo. CT2412 has improved
bioavailability (11%), and longer half-life in rats. The analysis of
plasma samples obtained after p.o. dosing of CDP-840 to rat, rabbit,
and human showed that the predominant metabolites detected in plasma
were also the major metabolites formed in the hepatocyte incubations. A
good correlation between the in vitro and in vivo metabolism was
established. In vitro techniques, especially the hepatocyte incubations
can be therefore used to predict the in vivo metabolism of potential backup compounds in animal models and in humans.
In summary, the extent of metabolism and significant interspecies
differences in metabolism were the two main issues for CDP-840. The in
vitro metabolism studies described here not only helped in determining
the suitability of animal species used in preclinical safety studies
but also were extremely useful in identifying backup compounds and
directing the synthetic efforts. Simple structural modifications on
CDP-840, such as R4 para-Cl substitution and R3
N-oxide, greatly improved the metabolic profile and
stability. As a result, metabolism guided studies have been greatly
accelerated for lead finding and optimization of drug candidates.
We acknowledge Dr. Marc Bilodeau from the Hospital Saint-Luc, Montreal
for providing human liver tissue. We also thank Yves Girard for useful
discussions and laboratory animal resources of Merck Frosst for
obtaining the plasma samples from rat and rabbit.
![[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
