Simvastatin hydroxy acid (SVA), the pharmacologically active form
of simvastatin (SV), is a potent inhibitor of
3-hydroxy-3-methylglutaryl (HMG)-coenzyme A reductase and is formed on
hydrolysis of the orally administered SV. In this article, we report
the structural characterization of two new dihydroxy glutathione
adducts and a trihydroxy derivative of SVA, all found in rat bile.
Metabolite I is 5'
,6'-dihydroxy-4'a
-glutathione-SVA, and
metabolite II is a pentanoic acid derivative of metabolite I. The two
identified GSH conjugates accounted for 16 and 9% in males and 11 and
5% in females of the total radioactivity (metabolites I and II,
respectively). Metabolite III is 3',5',6'-dihydrotriol-SVA
and accounts for 2% (male) and 4% (female) of the total dose in rats.
Of these three newly identified metabolites, only metabolite III was
also observed in dog bile.
 |
Introduction |
Simvastatin is
widely used in the treatment of hypercholesterolemia and
hypertriglyceridemia (Mauro, 1993
). Simvastatin hydroxy acid
(SVA1; Fig. 1), the
pharmacologically active form of simvastatin (SV), is a potent
inhibitor of 3-hydroxy-3-methylglutaryl (HMG)-coenzyme A reductase
(Duggan and Vickers, 1990) and is formed on hydrolysis of the orally
administered SV. In rodents, hydrolysis of SV to SVA in plasma is rapid
(Vickers et al., 1990a). Biotransformation of SV exhibits noticeable
species differences, and multiple biotransformation pathways have been
elucidated (Vickers et al., 1990a,b; Prueksaritanont et al., 2001). In
vivo SV metabolites structurally characterized thus far involved either
oxidations in the decalin ring or -oxidation products of the
heptenoic acid side chain. These include oxidations at the 6', 3', and
6'-CH3 sites in the decalin ring, and a pentanoic acid product resulting from the -oxidation of the heptenoic side chain of SV. In this study, we report the structural characterization of two new dihydroxy glutathione adducts and a trihydroxy derivative of
SVA, all found in rat bile. Only the trihydroxy-SVA was also observed
in dog bile. The three new metabolites of SVA combined constitute a
significant percentage of the total dose in male and female rats.
| |
Materials and Methods |
Radiochemicals.
[2"-14CH3]SVA (Fig. 1)
was synthesized at Merck Research Laboratories (Rahway, NJ), with a
specific activity of 55 µCi/µmol. CD3OD
(99.96% D) was obtained from Isotec, Inc. (Miamisburg, OH). All other
chemicals were of analytical or HPLC grade.
Animal Studies and Sample Collection.
All studies were reviewed and approved by the Merck Research
Laboratories Institutional Animal Care and Use Committee.
[14C]SVA was administrated i.v. (5 mg/kg) to
male (n = 4) or female (n = 2)
Sprague-Dawley rats. The dosing solution was prepared by mixing the
[14C]SVA with nonradiolabeled SVA to closely
achieve a 1:1 isotopic ratio of
14C/12C. Before dosing, a
bile duct cannula was surgically implanted in the rats. Rat bile
sampled at different time points was collected on dry ice. The frozen
bile was thawed after collection and then adjusted to pH 4.5 by adding
1 M ammonium acetate (~10% of total volume). Buffered rat bile was
immediately frozen on dry ice and stored in a 
70°C freezer until
further analysis. Buffered bile from 0 to 3 h postdose was pooled
and used for LC-radiomatic-MS and/or -MS/MS analysis.
To generate larger amounts of metabolites I and II for NMR studies, a
mixture of unlabeled and 14C-labeled SVA was
administrated orally (100 mg/kg; ~140 µCi/rat) to male
Sprague-Dawley rats (n = 2). Bile was collected and
processed in the same manner as described above.
A SVA dosing solution containing 1:1 ratio of
[2"-14C]SVA/[2"-12C]SVA
was administrated i.v. (0.2 mg/kg) to male (n = 3)
bile-duct-cannulated dogs. Dog bile was collected over a 24-h period
into a bag that was carried on the dog's back and processed in the
manner described above.
Sample Treatment for HPLC-Radiomatic-MS Analysis.
HPLC separation was carried out on an Hewlett Packard 1100 binary pump
system (Palo Alto, CA) using a Betasil C18 column
(4.6 × 250 mm, 5 µm; Thermo Hypersil-Keystone,
Bellefonte, PA). Samples were injected via a PerkinElmer 200 autosampler (Norwalk, CT), maintained at 5°C. The mobile phase
consisted of 0.1% aqueous formic acid (solvent A) and acetonitrile
(solvent B) and was delivered at a constant flow rate of 1 ml/min. The
initial mobile phase consisted of 16% B, which remained unchanged for
6 min, and increased linearly to 60% over 39 min, then to 90% over 5 min, and finally held at 90% B for 10 min. The column was equilibrated
with initial mobile phase at 1 ml/min for 10 min before sample
injection. The liquid chromatograph was coupled to a Packard 500TR flow
scintillation analyzer (Packard BioScience, Meriden, CT), which was
equipped with a 176-µl lithium glass flow cell and calibrated for
monitoring 14C, followed by a UV detector
and a Finnigan MAT LCQ ion trap mass spectrometer (Thermo Finnigan MAT,
San Jose, CA). The analog output from the 14C
channel of the Packard 500TR radiomatic detector was coupled to the LCQ
analog input in order for radiochromatograms to be simultaneously
recorded on both the LCQ and Packard radiomatic detector.
For the LC-MS or LC-MS/MS analyses, a portion of the HPLC effluent was
introduced into the LCQ ionization source at a rate of 180 µl/min.
Mass spectral analyses were performed using electrospray ionization in the negative ion mode. The electrospray ionization ionizing voltage was 2.5 kV, and the heated capillary temperature was
160°C. The isolation width and collision energy values were set at 3 and 25%, respectively, for the MS/MS experiments. Detection and
identification of the metabolites were based on radioactive peak
monitoring, molecular ion detection, and visual recognition of the
14C/12C isotopic
"goal-posts" in the mass spectrum resulting from the 1:1 ratio of
14C/12C in the SVA dosing solution.
Metabolite Isolation for NMR.
Metabolite isolation for NMR was accomplished with two rounds of HPLC
purification. The initial purification was done under the HPLC
conditions described above. Fractions containing the metabolites were
collected, pooled, dried under a stream of nitrogen gas, individually
reconstituted in 200 µl of 0.1% formic acid, and repurified. The LC
conditions for the repurification included a 2.0 × 250-mm Betasil
column and a linear gradient from 23 to 50% B. Fractions were
collected based on a UV peak at 240 nm. The collected fractions were
again dried under a stream of N2 gas and used
immediately for NMR analysis. Metabolite III was isolated from male dog
bile using the same isolation procedures.
NMR Analysis.
NMR spectra were obtained using an Inova (11.7 T/500 MHz) 51-mm
narrow-bore spectrometer (Varian Instruments, Palo Alto, CA) equipped
with a 3-mm inverse detection probe (MIDG-3; Nalorac Corporation,
Martinez, CA). The metabolites were dissolved in 160 µl of chilled
CD3OD, and data was acquired at 3°C to minimize sample degradation. 1H 1D, 2D total correlation
spectroscopy (TOCSY),
[13C]1H heteronuclear
multiple quantum coherence (HMQC), and 2D transverse rotating
Overhauser effect spectroscopy (ROESY) data sets (Summers et al., 1986;
Hwang and Shaka, 1992) were collected for metabolites I and II. For
metabolite III, 1H 1D TOCSY and a 1D nuclear
Overhauser effect (NOE) difference spectra were collected.
| |
Results |
Identification of SVA Metabolites.
Metabolite I
The molecular ion region of a full scan MS spectrum (Fig.
2A) of metabolite I shows the presence of
a molecular ion [M H]
= 774 Da and a
corresponding 14C peak at 776 Da in an
approximately 1:1 ratio. MS/MS fragmentation of the peak at 774 Da
shows peaks indicative of a glutathione moiety attached to the
metabolite (Fig. 2B), including the following: ions at
m/z 306 (GS),
m/z 272 (GS H2S), m/z 254 (m/z 272 H2O), and
a neutral loss of 129 (m/z 645). The apparent
molecular weight of this metabolite corresponds to an addition of two
oxygen atoms and a molecule of GSH to the parent SVA.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 2.
LC-MS and MS/MS spectra of metabolite
I.
MS spectrum showing the molecular region of metabolite I (A). Peaks at
m/z 774 and 776 in approximately a 1:1
ratio is a result of a 1:1 ratio of 12C/14C in
the SVA dosing solution (see text for details). MS/MS spectrum of
metabolite I m/z peak at 774 (B).
|
|
Identifiable 1H and 13C
chemical shifts for metabolite I are listed in Table
1. Protons 4 and 6 in the dihydroxy
heptanoic acid side chain have signature chemical shifts in the SV
lactone (H4, 4.25 ppm; H6, 4.62 ppm) and SVA (H4, 4.08 ppm; H6, 3.69 ppm). Protons 4 and 6 of metabolite I have chemical shifts (
) of
4.25 and 4.64 ppm, and hence, the metabolite as characterized by NMR is
in the lactone form. The connectivity involving the heptanoic lactone
side chain protons can be fully traced in the metabolite TOCSY spectrum
(Fig. 4A).
View this table:
[in this window]
[in a new window]
|
TABLE 1
Identifiable chemical shifts of SVA and the metabolites
All chemical shifts are referenced to the CD2HOD peak at
3.31/48.1 ppm.
|
|
Relative to SVA, the proton chemical shifts of the dimethyl butyric
acid side chain of metabolite I were unaltered. The peaks corresponding
to the 3'- and 4'-protons appeared as a nonfirst-order multiplet in the
metabolite 1D 1H spectrum (Fig.
3) and were well resolved in the
[13C]1H HMQC spectrum
(not shown). The 4', 3', 2', 2'Me, 1', 8'a, 8', and 7' connectivity can
be traced in the 2D TOCSY spectrum (Fig. 4A), and hence, the site of oxidation and
glutathione addition can be narrowed to positions 4'a, 5', and 6'.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 4.
TOCSY (80-ms mixing time) (A) and transverse
ROESY (300-ms mixing time) spectra (B) of metabolite I.
Only key peaks are labeled in the spectrum.
|
|
The 6'-proton peak at 2.42 ppm in SVA is absent in the metabolite
spectra. The peak corresponding to the 6'Me is a doublet at 1.08 ppm in
SVA but moves downfield to 1.62 ppm and appears as a singlet in the
metabolite, suggesting an OH addition at the 6'-position and an OH or
SG in the neighboring position. The SVA peak at 5.49 ppm (5'H)
has disappeared and a new singlet is seen at 3.72 ppm in the metabolite
1H spectrum (Fig. 3). This new singlet had a
correlated 13C peak at 77.4 ppm in the HMQC
spectrum (not shown). The 1H and
13C chemical shifts rule out a GS addition at the
5'-position and instead indicate an OH addition. In the ROESY spectrum,
cross peaks are seen from the 5'H (3.72 ppm) to the 6'Me (1.62 ppm) and
the glutathione c' methylene (2.82 ppm, 3.22 ppm). In addition, ROE correlation peaks (Fig. 4B) are also seen from the 4'H (5.73 ppm) to the glutathione c' proton (3.22 ppm) and the 5'H (3.72 ppm).
These observations together indicate that the glutathione addition
occurs at the 4'a-position and reaffirms OH additions to the 5'- and
6'-positions. This assignment is also supported by the up-field shift
of the 8'a-peak from 2.33 ppm to 1.88 ppm, which would occur due to
loss of the 
4'a,5' double bond. In addition,
the ROESY data indicate that 6'Me, 5'H, and the 4'a-SG are in the
-configuration (syn with respect to each other).
Metabolite I has been shown to be in its hydroxy acid form in rat bile.
Under the LC-MS conditions, the hydroxy acid and lactonized forms of
metabolite I have retention times of 9.7 and 16.0 min, respectively.
The conversion from the hydroxy acid to the lactone form is favored
under acidic conditions and thus has occurred during the purification,
which was carried out under acidic conditions. An aliquot of the NMR
sample was subjected to base hydrolysis and resulted in the hydroxy
acid form, as confirmed by its retention time and MS spectrum. Thus,
metabolite I has been identified as 5',6'-dihydroxy-4'a-glutathione-SVA and is shown in Fig. 1.
Metabolite II.
A full scan MS spectrum of metabolite II shows the presence of a
molecular ion [M H] = 714 Da and the
corresponding 14C peak at 716 Da in approximately
a 1:1 ratio. As in metabolite I, MS/MS fragmentation of the parent ion
showed the following peaks indicative of a GSH moiety on the
metabolite: ions at m/z 306 (GS), 272 (GS H2S), 254 (272 H2O), and a neutral loss of 129 (m/z 585). The molecular weight of the parent ion
for metabolite II corresponds to metabolite I minus 2 CHOH units.
Identifiable 1H and 13C
chemical shifts for metabolite II are listed in Table 1. In metabolite
II, the geminal methyl groups in the butyric acid ester moiety are
identifiable and intact. The chemical shifts of all the glutathione
protons are identifiable in the metabolite spectra. Also, the
4'-3'-2'-2'Me-1'-8'a-8'-7' connectivity can be traced in the 2D
TOCSY spectrum, and hence, the decalin ring is intact.
NMR features of the decalin ring for metabolite II are similar to those
of metabolite I. The key features are as follows: 1) the parent SVA
peak at 5.49 ppm (5'H) has disappeared, and a new singlet is seen at
3.71 ppm in the 1H spectrum, which has a
13C correlated peak at 77.5 ppm; 2) ROESY cross
peaks are seen from the singlet at 3.71 ppm (assigned to 5'H) to the
6'Me (1.61 ppm) and to the c' methylene (2.82, 3.21 ppm); and 3) ROESY
cross peaks are also seen from the 4'H (5.72 ppm) to the c' protons in
glutathione and the 5'H (3.71 ppm). These changes suggest that the
glutathione addition occurs at the 4'a-position and OH group additions
at the 5'- and 6'-positions. Therefore, the main ring of metabolite II
is structurally identical to that of metabolite I.
With respect to the parent SVA, the distinct chemical shifts of protons
4 and 6 in the di-hydroxy heptanoic side chain are absent in the
1H spectrum of metabolite II. Instead, the
following shorter connectivity is seen in the TOCSY spectrum:
2.20 ~1.61 ~1.11, 1.18 ~1.47, 1.56 2.26 ppm corresponding to 5-6-7-8-1' connectivity. The 1H and its 13C-correlated
peaks at 2.20 and 36.5 ppm, respectively, are consistent with a
CH2COOH moiety, and they are assigned to the
methylene labeled 5. The MS and NMR spectra together indicate that
metabolite II is a pentanoic acid derivative of
5',6'-dihydroxy-4'a-glutathione SVA adduct (Fig. 1).
Metabolite III.
A full scan MS spectrum of metabolite III shows a pseudo molecular ion
at m/z 485 ([M H]) and the corresponding
14C peak at m/z 487 in a
~1:1 ratio. MS/MS analysis of the parent ion showed a neutral loss of
116, attributed to the dimethyl butyric acid moiety (also seen with
SVA). The parent ion at 485 Da corresponds to an addition of
H2O and two oxygen atoms to SVA.
Identifiable 1H chemical shifts for metabolite
III are listed in Table 1. NMR peaks corresponding to protons 4 and 6 in the heptanoic acid side chain had of 4.25 and 4.62 ppm,
respectively. Hence, metabolite III was in its lactone form in the NMR
sample. Again, as seen with metabolite I, the lactonization occurred
during metabolite purification carried out under acidic conditions.
There seem to be practically no changes in either the chemical shifts or connectivities (in the TOCSY spectrum) of protons in the heptanoic lactone and the dimethyl butoxy side chain, indicating that no oxidation has occurred on either of these moieties.
The 6'Me peak at 1.08 ppm (doublet) in SVA has shifted downfield to
1.23 ppm and appears as a singlet, indicating hydroxylation at the
6'-carbon. The SVA peaks at 5.79 ppm (3'H) and 5.49 ppm (5'H) have
disappeared. Instead, new peaks are seen at 3.86 ppm (multiplet) and
3.79 ppm (singlet), indicative of two CHOH groups. The TOCSY spectrum
of metabolite III shows the following connectivity: 5.85 3.86 1.94, 0.85 1.72 2.40 5.30 1.77, 2.17 ppm corresponding to protons 4'-3'-2',
2'Me-1'-8'a-8'-7' in the decalin ring. The 2' has moved up-field by
0.5 ppm in the metabolite spectrum, a change that can be accounted for
by the loss of the 3',4' double bond and
hydroxylation at the 3'-position. The remaining CHOH singlet at 3.79 ppm is assigned to the 5'H. A difference NOE spectrum obtained by
irradiating the 6'Me singlet at 1.23 ppm shows a positive NOE peak at
3.79 ppm (5'H), thereby confirming the above assignment and also that
the 5'-OH is in the -configuration. The 3'- and 5'-methine chemical
shifts are consistent with the previously identified
3',5'-dihydrodiol-SV (Prueksaritanont et al., 1997). Therefore,
metabolite III has been identified as 3',5',6'-dihydrotriol-SVA, as shown in Fig. 1.
Excretion of Metabolites I, II, and III in Bile.
For a 5 mg/kg i.v. dose in rats, 87% (in male;
n = 4) and 83% (in female; n = 2) of
the total radioactivity is recovered in bile collected over 0 to 8 h. Metabolite I accounts for 16% (male) and 11% (female) and
metabolite II for 9% (male) and 5% (female) of the total dose (Table
2). Metabolite III was a minor
metabolite, accounting for 2% (male) and 4% (female) of the total
radioactivity. The two previously identified metabolites (Vickers et
al., 1990a,b) were also seen, namely, 6'-OH-SVA at 11 and 4% and
3'-OH-SVA at 20 and 12% of the total radioactivity in male and female
rats, respectively.
View this table:
[in this window]
[in a new window]
|
TABLE 2
Recovery of SVA metabolites (% of total dose) in bile
Values are from pooled samples: male rats (n = 4),
female rats (n = 2), and male dogs (n = 3).
|
|
For a 0.2 mg/kg i.v. dose in male dogs (n = 3),
glutathione adducts (metabolites I and II) were not seen. Of the 45%
total radioactivity recovered during the 24-h collection period,
metabolite III accounts for 4% of the total dose. Other significant
oxidative metabolites found in rat and dog bile are listed in Table 2.
| |
Discussion |
Hydrolysis of SV to SVA in rodent plasma is rapid (Vickers et al.,
1990a). Hence, metabolism of SVA is expected to give a similar
metabolite profile as seen for SV in rats. In this article, structural
characterizations of three new metabolites of SVA that constitute a
significant percent (29% male and 19% female) of the total dose in
rats are described.
Analogous to metabolite I, the 5',6'-dihydroxy-4'a-glutathione
adduct of pravastatin sodium (sodium salt of pravastatin hydroxy acid)
has been previously reported to be formed both in vitro and in vivo in
rats (Nakamura et al., 1991; Komai et al., 1992). Metabolite I could be
formed by two possible mechanisms. The first possibility is an initial
oxidation of SVA to 6'OH-SVA, followed by formation of a
4'a,5'-epoxide and nucleophilic attack of glutathione at the
4'a-carbon from the sterically less hindered -side. Alternatively,
the epoxide may be formed first followed by oxidation at the
6'-position. The former possibility seems to be more likely since
6'OH-SVA is a significant in vivo metabolite and also no
5'-OH,4'a-glutathione adduct of SV or SVA is observed.
Metabolite II is possibly a product of the -oxidation pathway
starting with metabolite I. Alternatively, metabolite II can also be
formed by initial -oxidation of the dihydroxy heptanoic acid side
chain of SVA to pentanoic acid, followed by events paralleling the
formation of metabolite I. Neither mechanism can be ruled out since
both the pentanoic acid derivative of SVA (Prueksaritanont et al.,
2001) and metabolite I are observed. An analog of metabolite II has not
been reported previously in any of the structurally similar statins
(pravastatin, lovastatin, or SV/SVA). However, -oxidation pathways
yielding the pentanoic acid analog are a common feature in pravastatin
and other statins (Arai et al., 1988; Vickers et al., 1990b; Vyas et
al., 1990; Everett at al., 1991; Komai et al., 1992; Halpin et al.,
1993; Le Couteur et al., 1996; Black et al., 1999;
Prueksaritanont et al., 2001). Therefore, although not reported, it is
possible that the pentanoic acid derivative similar to metabolite II is
also formed from pravastatin.
Metabolite III, the 3',5',6'-dihydrotriol derivative of SVA has
not been previously reported from either SV/SVA or lovastatin. It has
been shown to form in vivo from the pravastatin hydroxy acid in humans
and in rats (Everett et al., 1991; Komai et al., 1992). As mentioned
above, the initial site of oxidation could be the 6' site, followed by
formation of 4'a,5'-epoxide and hydrolysis, as proposed by
Nakamura et al. (1991), leading to the dihydrotriol-SVA. Based on the
NOE experiment, the 6'Me and 5'H are syn to each other (in
the 6'-OH, 5'-OH configuration) but the configuration at the
3'-position could not be conclusively determined. In the case of
pravastatin, its dihydrodiol was shown to be in the 3'-OH, 5'-OH,
6'-OH configuration by comparison with a synthetic standard
(Nakamura et al., 1991).
Qualitatively, all significant bile metabolites of SV that were found
in male rats were also seen in the females. However, quantitatively,
they were formed in greater amounts in males versus females.
Metabolites I and II together account for 25 and 16% of total i.v.
dose, whereas the 3'-OH metabolite accounted for 20 and 12% of the
i.v. dose in male and female rats, respectively. These gender
differences in the metabolism of SV are consistent with previously
published in vitro data (Ohtawa and Uchiyama, 1992), which
showed a higher rate of metabolism in males than in female rats. In
vivo, a similar trend was observed, and the gender differences have
been attributed to the different P450 isozymes responsible (CYP3A in
female and CYP2C11 in male rats) for the metabolism of SV (Ishigami et
al., 2001).
Neither of the glutathione adducts (metabolites I and II) were observed
in dogs dosed with SV or SVA. It is noteworthy that the glutathione
conjugates of SVA have not been observed in humans following SV or SVA
administration (unpublished data on file, Merck Research
Laboratories). Also, glutathione adducts analogous to metabolites I and
II have not been previously reported in dogs for lovastatin and pravastatin.
Metabolite III was found to be inactive in the HMG-coenzyme A reductase
inhibition assay. Metabolites I is expected to be inactive based on
previous studies of the 3',5'-dihydroxy diol metabolite of SV
(Prueksaritanont et al., 1997). Metabolite II is also expected to be
inactive due to the loss of the dihydroxy heptanoic acid side chain,
which is necessary for HMG-coenzyme A reductase activity (Duggan and
Vickers, 1990).
We thank Dr. A. Jones, Dr. C. Raab, G. Gatto, and N. Yu for the
synthesis and purification of [14C]SVA and P. Deluna and
J. Brunner for help in the animal studies. R.S. thanks Drs. B. Arison,
K. Regal, and J. Yergey for many fruitful discussions.
Dr. Raju Subramanian,
Department of Drug Metabolism, WP75A-203, Merck Research Laboratories,
West Point, PA 19486. E-mail: raju_subramanian{at}merck.com
Abbreviations used are:
SVA, simvastatin hydroxy
acid;
SV, simvastatin;
HMG, 3-hydroxy-3-methylglutaryl;
HPLC, high-performance liquid chromatography;
LC, liquid chromatography;
MS, mass spectrometry;
MS/MS, tandem mass spectrometry;
1D, one
dimensional;
2D, two dimensional;
TOCSY, total correlation
spectroscopy;
HMQC, heteronuclear multiple quantum coherence;
ROESY, rotating Overhauser effect spectroscopy;
NOE, nuclear Overhauser
effect;
GSH, glutathione.
![[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
