Identification of novel metabolic pathways of pioglitazone in hepatocytes: N-glucuronidation of thiazolidinedione ring and sequential ring-opening pathway

Article abstract of DOI:10.1124/dmd.109.031583

The metabolism of [¹⁴C]pioglitazone was studied in vitro in incubations with freshly isolated human, rat, and monkey hepatocytes. Radioactivity detection high-performance liquid chromatography analysis of incubation extracts showed the detectio

Full text of DOI:10.1124/dmd.109.031583

0090-9556/10/3806-946–956$20.00
DRUG METABOLISM AND DISPOSITION
Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics
DMD 38:946–956, 2010
Vol. 38, No. 6
31583/3584656
Printed in U.S.A.
Identification of Novel Metabolic Pathways of Pioglitazone in
Hepatocytes: N-Glucuronidation of Thiazolidinedione Ring and
Sequential Ring-Opening Pathway
Minoru Uchiyama, Thomas Fischer, Juergen Mueller, Minoru Oguchi, Naotoshi Yamamura,
Hiroko Koda, Haruo Iwabuchi, and Takashi Izumi
Drug Metabolism and Pharmacokinetics Research Laboratories, Daiichi Sankyo Co., Ltd., Tokyo, Japan (M.U., N.Y., H.K., H.I.,
T.I.); Drug Metabolism Department, Daiichi Sankyo Europe GmbH, Martinsried, Germany (T.F., J.M.); and Research Department
I, Daiichi Sankyo RD Associe Co., Ltd., Tokyo, Japan (M.O.)
Received December 8, 2009; accepted February 25, 2010
ABSTRACT:
The metabolism of [14C]pioglitazone was studied in vitro in incu- opened N-glucuronide (M2), thiazolidinedione ring-opened methyl
bations with freshly isolated human, rat, and monkey hepatocytes. sulfone amide (M3), thiazolidinedione ring N-glucuronide (M7),
Radioactivity detection high-performance liquid chromatography thiazolidinedione ring-opened methylmercapto amide (M8), and
analysis of incubation extracts showed the detection of 13 metab- thiazolidinedione ring-opened methylmercapto carboxylic acid
olites (M1–M13) formed in incubations with human hepatocytes. An (M11). In summary, based on the results from these studies, two
identical set of metabolites (M1–M13) was also detected in monkey novel metabolic pathways for pioglitazone in hepatocytes are pro-
hepatocytes. However, in rat hepatocytes, M1 through M3, M5 posed to be as follows: 1) N-glucuronidation of the thiazolidinedi-
through M7, M9 through M11, and M13 were also detected, but M4, one ring of pioglitazone to form M7 followed by hydrolysis to M2,
M8, and M12 were not detected. The structures of the metabolites and methylation of the mercapto group of the thiazolidinedione
were elucidated by liquid chromatography/tandem mass spec- ring-opened mercapto carboxylic acid to form M11; and 2) meth-
trometry using electrospray ionization. Novel metabolites of pio- ylation of the mercapto group of the thiazolidinedione ring-opened
glitazone detected using these methods included thiazolidinedione mercapto amide to form M8, oxidation of M8 to form M1, and
ring-opened methyl sulfoxide amide (M1), thiazolidinedione ring- oxidation of M1 to form M3.
The thiazolidinedione-containing drug pioglitazone (Fig. 1) is used idiosyncratic hepatotoxicity (Gale, 2001; Graham et al., 2003). The
in the treatment of type 2 diabetes mellitus (Chilcott et al., 2001).
Pioglitazone is a member of the thiazolidinedione class agents, glita-
zones, and increases insulin sensitivity in target tissues by way of
interaction with the peroxisome proliferator-activated receptor gamma
(PPAR␥) (Sood et al., 2000). Several metabolites of pioglitazone in
animals and humans have been reported (Krieter et al., 1994; Kiyota
et al., 1997; Maeshiba et al., 1997; Shen et al., 2003). Five primary
metabolic pathways are involved, and the most important cytochrome
P450 isoenzymes appear to be CYP2C8/9 and CYP3A4 (Eckland and
Danhof, 2000; Baba, 2001; Jaakkola et al., 2006; Muschler et al.,
2009).
generation of many reactive metabolites has been proposed in the
in vitro metabolic experiments with troglitazone to investigate the
relationship to the hepatotoxicity (Kassahun et al., 2001; Tettey et
al., 2001; Yamamoto et al., 2002; Smith, 2003; He et al., 2004;
Alvarez-Sanchez et al., 2006). By liquid chromatography/tandem
mass spectrometry (LC/MS/MS) analysis, Kassahun et al. (2001)
found that a sulfoxide, an ␣-ketoisocyanate, and a sulfenic acid of
troglitazone were generated as reactive intermediates that easily
react with glutathione in human liver microsomes or a human
recombinant CYP3A4 system. Based on the discovery of the
reactive metabolites and the presumed metabolic pathway, Kas-
sahun et al. (2001) also reported the possible contribution of the
reactive metabolites to the hepatotoxicity of troglitazone.
Another thiazolidinedione-containing drug, troglitazone, was
withdrawn from the market because of rare but serious cases of
For pioglitazone, the reactive intermediates of pioglitazone and
their presumed metabolic pathway, which is similar to that of
troglitazone, were also reported (Baughman et al., 2005). They
Article, publication date, and citation information can be found at
http://dmd.aspetjournals.org.
doi:10.1124/dmd.109.031583.
ABBREVIATIONS: PPAR␥, peroxisome proliferator-activated receptor ␥; LC/MS/MS, liquid chromatography/tandem mass spectrometry;
radio-HPLC, radioactivity detection high-performance liquid chromatography; HBSS, Hank’s balanced salt solution; SAM, S-adenosyl-L-
methionine; DMSO, dimethyl sulfoxide; HPLC, high-performance liquid chromatography; TLC, thin-layer chromatography; MS, mass
spectrometry; TFA, trifluoroacetic acid; Q-Tof, quadrupole time-of-flight; ESI, electrospray ionization; BMS-204352, (3S)-(ϩ)-(5-chloro-2-
methoxyphenyl)-1,3-dihydro-3-fluoro-6-(trifluoromethyl)-2H-indol-2-one; MK-0767, (Ϯ)-5-[2,4-dioxothiazolidin-5-yl)methyl]-2-methoxy-
N[[(4-trifluoromethyl)-phenyl]methyl]benzamide; MRL-A, (Ϯ)-5-[(2,4-dioxothiazolidin-5-yl)methyl]-2-methoxy-N-[[4-trifluoromethoxy)phenyl]
methyl]benzamide]; bucolome, 5-n-butyl-l-cyclohexyl-2,4,6-trioxoperhydropyrimidine.
946

IDENTIFICATION OF NOVEL METABOLITES OF PIOGLITAZONE
947
7.5 Hz, PyCH2CH3), 2.93 (1H, t, J ϭ 12.7 Hz, ArCH2CH), 3.22 (2H, dt, J ϭ
3.2 and 9.5 Hz, PyCH2CH2O), 3.4 to 3.6 (1H, m, ArCH2CH), 3.75 (3H, d,
J ϭ 3.5 Hz, CO2CH3), 4.12 (1H, t, J ϭ 7.2 Hz, C5-H), 4.3 to 4.4 (3H, m,
PyCH2CH2O, C1-H), 5.3 to 5.4 (3H, m, C2-H, C3-H, C4-H), 5.9 to 6.0 (1H, m,
ArCH2CH), 6.85 (2H, dd, J ϭ 4.9 and 8.6 Hz, Ar-H), 7.10 (2H, dd, J ϭ 1.6
and 8.6 Hz, Ar-H), 7.18 (1H, d, J ϭ 7.8 Hz, Py-H), 7.45 (1H, dd, J ϭ 1.9 and
7.8 Hz, Py-H), and 8.39 (1H, d, J ϭ 1.9 Hz, Py-H).
3-((2R,3R,4S,5S,6S)-6-Allyloxycarbonyl-3,4,5-trihydroxy-tetrahy-
dropyran-2-yl)-5-{4-[2-(5-ethyl-2-pyridinyl)ethoxy]benzyl}-thiazoli-
dine-2,4-dione (3). A solution of 2 (4.26 g, 6.33 mmol), allyl alcohol (80 ml),
and 4 N hydrogen chloride–dioxane (80 ml) was stirred for 4 h at room
temperature and left standing for 3 days at room temperature. The reaction
mixture was evaporated to dryness under reduced pressure. The residue was
diluted with a 5:1 (by volume) mixture of methylene chloride and 2-propanol
(150 ml), added to water (150 ml), and treated with 1 N NaOH solution until
the aqueous layer was neutralized to pH 4. The organic layer was separated,
and the water layer was extracted with a 5:1 (by volume) mixture of methylene
chloride and 2-propanol. The combined organic layers were washed with brine
and dried over anhydrous sodium sulfate, after which the solvent was removed
by evaporation under reduced pressure. The residue was purified by column
chromatography with silica gel (200 g, 70 ϫ 300 mm) using a gradient elution
method, where a mixture of methylene chloride and methanol (volume ratios
ranging from 30:1 to 10:1, respectively) was used by increasing the percentage
of methanol gradually, to give 3.10 g (5.41 mmol) of 3: yield 85%; MS m/z 573
[M ϩ H]ϩ; 1H NMR (400 MHz, CDCl3) ␦ 1.24 (3H, dt, J ϭ 3.8 and 10.7 Hz,
PyCH2CH3), 2.62 (2H, q, J ϭ 7.6 Hz, PyCH2CH3), 3.12 (1H, br s, ArCH2CH),
3.1 to 3.3 (2H, m, PyCH2CH2O), 3.33 (1H, br s, ArCH2CH), 3.4 to 4.0 (4H,
m, C2-H, C3-H, C4-H, C5-H), 4.3 to 4.4 (2H, m, PyCH2CH2O), 4.4 to 4.5 (1H,
m, C1-H), 4.6 to 4.8 (2H, m, OCH2CH ϭ CH2), 5.0 to 5.2 (1H, m, ArCH2CH),
5.2 to 5.3 (1H, m, OCH2CH ϭ CH2), 5.3 to 5.4 (1H, m, OCH2CH ϭ CH2), 5.8
to 6.0 (1H, m, OCH2CH ϭ CH2), 6.83 (2H, dd, J ϭ 8.6 and 16.0 Hz, Ar-H),
7.11 (2H, dd, J ϭ 8.6 and 21.9 Hz, Ar-H), 7.20 (1H, d, J ϭ 7.8 Hz, Py-H), 7.48
(1H, dd, J ϭ 1.8 and 7.8 Hz, Py-H), and 8.3 to 8.4 (1H, m, Py-H).
3-((2R,3R,4S,5S,6S)-6-Carboxy-3,4,5-trihydroxy-tetrahydropyran-
2-yl)-5-{4-[2-(5-ethyl-2-pyridinyl)ethoxy]benzyl}-thiazolidine-2,4-
dione (M7). A mixture of 3 (1.00 g, 1.75 mmol), morpholine (0.20 ml, 2.29
mmol), tetrakis(triphenylphosphine)palladium (0.40 g, 0.35 mmol), and meth-
ylene chloride (30 ml) was stirred for 4 h at room temperature. The reaction
mixture was evaporated under reduced pressure. The residue was purified by
column chromatography with silica gel (50 g, 70 ϫ 300 mm) using a gradient
elution method, where a mixture of ethyl acetate, methanol, and water (volume
ratios ranging from 30:2:1 to 5:2:1, respectively) was used by increasing the
percentage of methanol and water gradually, to give 0.60 g of yellow solid.
This solid was purified using an SCL-10A HPLC system (Shimadzu Corpo-
ration, Kyoto, Japan). Separation was performed on an Inertosil ODS-3 column
(50 ϫ 500 mm, 5 ␮m; GL Sciences, Inc., Tokyo, Japan), and UV detection was
carried out at 254 nm. The mobile phase, consisting of water containing 0.1%
trifluoroacetic acid (TFA; solvent A) and acetonitrile (solvent B), was deliv-
ered at a flow rate of 50 ml/min. The gradient started at 10% solvent B and
increased linearly to 35% solvent B for 60 min. The HPLC purification gave
0.53 g of TFA salt M7: yield 47%; MS m/z 533 [M ϩ H]ϩ; MS/MS (product
ions) m/z 134, 357; 1H NMR (400 MHz, DMSO-d6) ␦ 1.23 (3H, t, J ϭ 7.5 Hz,
PyCH2CH3), 2.73 (2H, q, J ϭ 7.5 Hz, PyCH2CH3), 2.9 to 3.2 (1H, m,
ArCH2CH), 3.2 to 3.5 (5H, m, PyCH2CH2O, ArCH2CH, C3-H, C4-H), 3.71
(1H, dd, J ϭ 9.6 and 16.5 Hz, C2-H), 4.0 to 4.3 (1H, m, C5-H), 4.34 (2H, t, J ϭ
6.1 Hz, PyCH2CH2O), 4.7 to 5.1 (2H, m, ArCH2CH, C1-H), 5.42 (3H, br s,
OH), 6.8 to 6.9 (2H, m, Ar-H), 7.1 to 7.2 (2H, m, Ar-H), 7.79 (1H, dd, J ϭ 1.7
and 8.2 Hz, Py-H), 8.19 (1H, d, J ϭ 8.2 Hz, Py-H), 8.64 (1H, d, J ϭ 1.7 Hz,
Py-H), and 12.9 (1H, br s, CO2H).
FIG. 1. Chemical structure of [14C]pioglitazone.
have identified the oxidative thiazolidinedione ring-opened prod-
ucts of pioglitazone (M-X) in rat and human liver microsomes. In
addition, they have identified its ring-opened glutathione conju-
gates of pioglitazone (M-A and M-B) in rat and human liver
microsomes and suspensions of freshly isolated rat hepatocytes. In
this study, we performed a detailed characterization of the metab-
olites of pioglitazone by using hepatocyte incubation, radioactivity
detection high-performance liquid chromatography (radio-HPLC),
and LC/MS/MS techniques to understand further the potential
metabolic pathways of the drug.
Materials and Methods
Materials. Radiolabeled [thiazolidinedione-5-14C]pioglitazone (specific ra-
dioactivity, 20 mCi/mmol; radiochemical purity, Ͼ98%) was synthesized by
Blychem Ltd. (Billingham, UK). Collagenase H was purchased from Roche
Diagnostics (Indianapolis, IN). Hanks’ balanced salt solution (HBSS) was
purchased from Mediatech (Herndon, VA). EGTA, S-adenosyl-L-methionine
(SAM), dimethyl sulfoxide (DMSO), and acetobromo-␣-D-glucuronic acid
methyl ester were purchased from Sigma-Aldrich (St. Louis, MO). HEPES was
purchased from Dojindo Laboratories (Kumamoto, Japan). Pooled human liver
S9 from 15 donors was purchased from BD Gentest (Woburn, MA). All of the
other chemicals and solvents were of analytical or high-performance liquid
chromatography (HPLC) grade.
Synthesis of Authentic Standards of Thiazolidinedione Ring N-Gluc-
uronide M7 and Thiazolidinedione Ring-Opened N-Glucuronide M2. The
authentic standards of M7 and M2 were synthesized as outlined in Fig. 2 using
the procedure described below. Mass spectra were recorded using a JEOL
(Tokyo, Japan) JMS-700V, JEOL JMS-700QQ, or JEOL JMS-BU30 mass
spectrometer. 1H NMR spectra were recorded using a JEOL JNM-AL400,
Mercury 400 (Varian, Inc., Palo Alto, CA), or Mercury 400Vx spectrometer
(Varian, Inc.) and were reported in parts per million (␦) downfield from the
internal standard tetramethylsilane (Me4Si). All of the NMR spectra were
consistent with the assigned structures. Column chromatography was per-
formed by using Kishida reagent, silica gel SK-34 (Kishida Chemical Co.,
Ltd., Osaka, Japan) with solvents as described below. Thin-layer chromatog-
raphy (TLC) analyses were performed with Merck reagent, silica gel 60 F254
(0.25-mm thickness; Merck, Darmstadt, Germany). Spots were visualized by
either UV light or iodine.
3-((2R,3R,4S,5S,6S)-3,4,5-Triacetoxy-6-methoxycarbonyl-tetrahy-
dropyran-2-yl)-5-{4-[2-(5-ethyl-2-pyridinyl)ethoxy]benzyl}-thiazoli-
dine-2,4-dione (2). Acetobromo-␣-D-glucuronic acid methyl ester (6.07 g,
15.3 mmol) was dissolved in acetonitrile (60 ml) and added to a solution
containing pioglitazone hydrochloride (1; 3.00 g, 7.64 mmol), cesium carbon-
ate (4.98 g, 15.3 mmol), and acetonitrile (90 ml) at 70°C (Baba and Yoshioka,
2007). The combined mixture was stirred for 3 h at 70°C. The reaction mixture
was filtered with a silica gel short column (60 g, 70 ϫ 300 mm; using a 1:1 by
volume mixture of ethyl acetate and acetonitrile), and the filtrate was concen-
trated by evaporation under reduced pressure. The residue was purified by
column chromatography with silica gel (300 g, 70 ϫ 300 mm) using a gradient
elution method, where a mixture of hexane and ethyl acetate (volume ratios
ranging from 1:1 to 1:4, respectively) was used by increasing the percentage of
ethyl acetate gradually to give 4.27 g (6.35 mmol) of 2: yield 83%; mass
(2S,3S,4S,5R,6R)-6-(1-Carboxy-2-{4-[2-(5-ethyl-2-pyridinyl)ethoxy]
phenyl}-ethylsul-fanylcarbonylamino)-3,4,5-trihydroxy-tetrahydropy-
ran-2-carboxylic acid (M2). A solution of M7 (1.00 g, 1.88 mmol), sodium
acetate (2.00 g, 24.4 mmol), and 3:1 mixture of water and acetonitrile (20 ml)
was stirred for 8 h at room temperature. The reaction mixture was left standing
for 12 days at room temperature. The reaction mixture was acidified to pH 3
by the addition of TFA. This mixture was purified using an SCL-10A HPLC
spectrometry (MS) m/z 673 [M ϩ H]ϩ; 1H NMR (400 MHz, CDCl3) ␦ 1.24 system (Shimadzu Corporation). Separation was performed on an Inertosil
(3H, t, J ϭ 7.5 Hz, PyCH2CH3), 1.9 to 2.1 (9H, m, OCOCH3), 2.63 (2H, q, J ϭ ODS-3 column (50 ϫ 500 mm, 5 ␮m; GL Sciences, Inc.), and UV detection

948
UCHIYAMA ET AL.
FIG. 2. Synthetic scheme of thiazolidinedione ring N-glucuronide M7 and thiazolidinedione ring-opened N-glucuronide M2.
was carried out at 254 nm. The mobile phase, consisting of water containing
0.1% TFA (solvent A) and acetonitrile (solvent B), was delivered at a flow rate
of 50 ml/min. The gradient started at 10% solvent B and increased linearly to
30% solvent B for 60 min. The HPLC purification gave 0.48 g of TFA salt M2:
yield 38%; MS m/z 551 [M ϩ H]ϩ; MS/MS (product ions) m/z 134, 286, 332,
The monkey hepatocytes were washed twice with HBSS and suspended in
HBSS at a viable cell density of 2 ϫ 106 cells/ml. As determined by the trypan
blue dye exclusion method, the cellular viabilities of the monkey hepatocytes
were 91 and 94%.
Incubation of [14C]Pioglitazone with Freshly Isolated Human, Rat, and
358, 375, 417; 1H NMR (400 MHz, DMSO-d6) ␦ 1.22 (3H, t, J ϭ 7.6 Hz, Monkey Hepatocytes. [14C]Pioglitazone (30 ␮M) was incubated with viable
PyCH2CH3), 2.74 (2H, q, J ϭ 7.6 Hz, PyCH2CH3), 2.89 (1H, dd, J ϭ 6.4 and human, rat, or monkey hepatocytes suspended with HBSS (2 ϫ 106 cells/ml)
13.7 Hz, ArCH2CH), 3.0 to 3.3 (3H, m, C2-H, C3-H, C4-H), 3.15 to 3.25 (1H, for 3 h at 37°C under 95% O2 and 5% CO2 in an incubator (incubation volume,
m, ArCH2CH), 3.34 (2H, t, J ϭ 6.2 Hz, PyCH2CH2O), 3.62 (1H, d, J ϭ 9.3 2.5 ml each; incubation numbers for humans, n ϭ 3; rat, n ϭ 1; monkeys, n ϭ
Hz, C5-H), 4.09 (1H, dd, J ϭ 6.4 and 8.8 Hz, ArCH2CH), 4.34 (2H, t, J ϭ 6.2 2, as collected in the above section). For human hepatocytes, the initial
Hz, PyCH2CH2O), 4.7 to 4.8 (1H, m, C1-H), 5.17 (3H, br s, OH), 6.84 (2H, d,
viabilities were 66 to 81%, and the remaining CYP3A4 activities (testosterone
J ϭ 8.6 Hz, Ar-H), 7.11 (2H, d, J ϭ 8.6 Hz, Ar-H), 7.80 (1H, d, J ϭ 8.1 Hz, 6␤-hydroxylation) after incubation for 3 h were 60 to 100% of the initial
Py-H), 8.20 (1H, d, J ϭ 8.1 Hz, Py-H), 8.65 (1H, d, J ϭ 1.7 Hz, Py-H), 9.00 activity. For rat and monkey hepatocytes, the initial viabilities were 93 and 91
(1H, dd, J ϭ 4.5 and 8.8 Hz, NH), and 12.7 (2H, br s, CO2H).
to 94%, respectively. The reactions were terminated by the addition of aceto-
Preparation of Freshly Isolated Human Hepatocytes. Human liver tis- nitrile (5 ml). The cells homogenized by sonication and the samples of the
sues from three donors were obtained from patients undergoing partial
hepatectomy for metastatic liver tumors of colorectal cancer. Experimental
procedures were performed according to the guidelines of the charitable
incubated hepatocytes were stored at approximately Ϫ80°C until use. For each
species (humans, rat, or monkeys), the samples were pooled and centrifuged at
21,600g for 5 min at 4°C (Himac CF15D; Hitachi Koki Co., Ltd., Tokyo,
state-controlled foundation Human Tissue and Cell Research (Regensburg, Japan), and each supernatant was carefully collected. Each supernatant was
Germany) with informed patient’s consent and approved by the local ethics evaporated under a nitrogen stream to an approximate final volume of 1.5 ml.
committee of the University of Regensburg, Germany (Thasler et al., Each concentrated sample was used for the analysis of the metabolites by
2003). The liver samples were made anonymous. The human hepatocytes
used were isolated and provided by Hepacult GmbH (Regensburg, Ger-
many) after being commissioned by Human Tissue and Cell Research and
radio-HPLC, LC/MS, and LC/MS/MS.
Incubation of M7 with Human Liver S9. Incubation of M7 (authentic
standard) with pooled human liver S9 was performed as follows. The incuba-
Daiichi Sankyo Europe GmbH (Martinsried, Germany). Cells were isolated tion mixture contained 10 ␮M M7 (dissolved in methanol) and 1 mg/ml protein
using a modified two-step EGTA-collagenase perfusion procedure (Weiss from human liver S9 in 500 ␮l of potassium phosphate (50 mM), pH 7.4, in the
et al., 2003). The human hepatocytes were washed twice with HBSS and
presence or absence of 2 mM SAM. The incubations were carried out for 2 h
suspended in HBSS at a viable cell density of 2 ϫ 106 cells/ml. As at 37°C in a shaking water bath. Control incubation without human liver S9
determined by the trypan blue dye exclusion method, the cellular viabilities
of the human hepatocytes were 66, 73, and 81%.
and with SAM alone was carried out under otherwise similar conditions. Of the
500-␮l incubation outtake, 50-␮l aliquots were taken and quenched by adding
Preparation of Freshly Isolated Rat and Monkey Hepatocytes. All the acetonitrile (100 ␮l). After centrifugation at 21,600g for 5 min at 4°C, the
experimental procedures were performed in accordance with the in-house
guidance of the Institutional Animal Care and Use Committee of Daiichi
Sankyo Co., Ltd. (Tokyo, Japan). Freshly isolated rat and monkey hepatocytes
were obtained by the standard method (Molde´us et al., 1978). The rat hepa-
tocytes were prepared from a male Sprague-Dawley rat (Crl:CD; 7 weeks old;
body weight, 170 g). The rat hepatocytes were washed twice with HBSS and
suspended in HBSS at a viable cell density of 2 ϫ 106 cells/ml. The cellular
viability of the rat hepatocytes was 93%, as determined by the trypan blue dye
exclusion method. The monkey hepatocytes were prepared from two male
cynomolgus monkeys (Macaca fascicularis; 5 years old as the estimated age).
supernatant was collected and diluted with 50 ␮l of water and analyzed by
LC/MS and LC/MS/MS.
Radio-HPLC Analysis. Samples from the hepatocyte incubations were
analyzed by radio-HPLC. The radio-HPLC analysis was conducted using an
L-6000 HPLC system (Hitachi, Ltd., Tokyo, Japan) combined with a Radi-
omatic 525TR radioactivity detector (PerkinElmer Life and Analytical Sci-
ences, Waltham, MA). Chromatographic separation was performed on a YMC-
Pack ODS-A column (6.0 ϫ 150 mm, 5 ␮m; YMC Co., Ltd., Kyoto, Japan) at
ambient temperature. The mobile phase, consisting of water containing 0.01%
TFA (solvent A) and acetonitrile containing 0.01% TFA (solvent B), was

IDENTIFICATION OF NOVEL METABOLITES OF PIOGLITAZONE
949
delivered at a flow rate of 1 ml/min. The gradient started at 15% solvent B,
increased linearly to 50% solvent B for 30 min, increased linearly to 90%
solvent B for 5 min, and then held at 90% solvent B for 5 min. The column
effluent was monitored using an L-4000 UV detector (UV at 225 nm; Hitachi,
Ltd.) and a radioactivity detector with a 3-ml/min flow rate for the Ultima-Flo
M liquid scintillator (PerkinElmer Life and Analytical Sciences).
LC/MS/MS analyses were performed using collision energy of 20 eV and
xenon as the collision gas. Chromatographic separation was performed on a
YMC-Pack ODS-A column (1.5 ϫ 150 mm, 5 ␮m; YMC Co., Ltd.); the
column temperature was maintained at 30°C, and UV detection was carried out
at 225 nm. The mobile phase consisted of water containing 0.01% TFA
(solvent A) and acetonitrile containing 0.01% TFA (solvent B). The gradient
LC/MS and LC/MS/MS Analyses. The conditions described below were started at 15% solvent B, increased linearly to 50% solvent B for 30 min,
used for the LC/MS and LC/MS/MS analyses of pioglitazone metabolites
formed during incubation with human, rat, and monkey hepatocytes. These
increased linearly to 90% solvent B for 5 min, and was then held at 90%
solvent B for 5 min. The flow rate was set at 0.1 ml/min, and the elution flow
analyses were performed using a quadrupole time-of-flight (Q-Tof) mass from HPLC was introduced into the Q-Tof mass spectrometer ionization
spectrometer (Waters, Manchester, UK) with an L-7000 HPLC system (Hita- source through an ESI interface.
chi, Ltd.) consisting of an intelligent pump (model L-7100), a column oven
(model L-7300), a chromatointegrator (model D-7500), and a UV detector
The conditions described below were used for the LC/MS and LC/MS/MS
analyses of the generated products of M7 in human liver S9. These analyses
(model L-7400S). The LC/MS analysis was conducted using electrospray were performed using a Waters (Manchester, UK) Q-Tof Premier mass spec-
ionization (ESI) in the positive ion mode. The capillary voltage and cone
voltage were set at 3300 and 45 V, respectively. The source temperature and
desolvation gas temperature were 120 and 300°C, respectively. The mass range
from m/z 50 to 1000 was acquired with an integration time of 1 s. The
trometer with a Waters (Milford, MA) ACQUITY UPLC system consisting of
a binary solvent manager, a sample manager, a photodiode array detector, and
a column heater. The LC/MS analysis was conducted using ESI in the positive
ion mode. The capillary voltage and cone voltage were set at 3300 and 40 V,
FIG. 3. Representative radiochromatograms after
incubation of [14C]pioglitazone with freshly iso-
lated human, rat, and monkey hepatocytes.
[
14C]Pioglitazone (30 ␮M) was incubated with
freshly isolated human, rat, and monkey hepato-
cytes for 3 h at 37°C. Radio-HPLC analysis was
conducted as described under Materials and
Methods.

950
UCHIYAMA ET AL.
TABLE 1
Positive ion LC/MS/MS data of metabolites of pioglitazone in freshly isolated human, rat, and monkey hepatocytes
Freshly Isolated Hepatocytes
ϩ
Metabolites
Retention Time
͓M ϩ H͔
Product Ions
Human
Rat
Monkey
min
m/z
m/z
M1
M2
M3
M4
M5
M6
M7
M8
7.8
12.2
13.2
14.5
15.0
15.2
16.8
17.8
18.2
19.4
23.3
23.9
30.1
361
551
377
373
373
387
533
345
373
373
346
355
371
119, 134, 254 (B), 280, 297
134, 286, 332 (B), 358, 375, 417
119, 134, 254 (B), 280, 297
132, 150 (B), 240, 256
132, 135, 150 (B), 238, 355
146, 164 (B)
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
N.D.
ϩ
ϩ
ϩ
N.D.
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
134, 357 (B)
119, 134 (B), 240, 254, 280, 328
119, 134, 150, 240, 268 (B), 284, 313, 355
133, 150, 210 (B), 222, 238, 239, 284, 355
119, 134 (B), 254, 270, 286,
132 (B), 210, 238, 239, 284
106, 148 (B), 254
M9
M10
M11
M12
M13
N.D.
ϩ
B, base peak; ϩ, metabolite detected; N.D., not detected.
respectively. The source temperature and desolvation gas temperature were
120 and 300°C, respectively. The mass range from m/z 50 to 1000 was
acquired with an integration time of 0.5 s. The LC/MS/MS analysis was
performed using collision energy of 25 eV and argon as the collision gas.
Chromatographic separation was performed on a YMC-UltraHT Pro C18
column (2.0 ϫ 50 mm, 2 ␮m; YMC Co., Ltd.); the column temperature was
maintained at 40°C, and UV detection was carried out at 225 nm. The mobile
phase consisted of water containing 0.01% TFA (solvent A) and acetonitrile
containing 0.01% TFA (solvent B). The gradient started at 10% solvent B and
increased linearly to 20% solvent B for 9 min. The flow rate was set at 0.2
ml/min, and the elution flow from UPLC was introduced into the Q-Tof
ionization source through an ESI interface.
set of metabolites (M1–M13) was also detected in monkey hepato-
cytes. In rat hepatocytes, M1 through M3, M5 through M7, M9
through M11, and M13 were also detected. M4, M8, and M12 were
not detected in rat hepatocytes in this study. M1, M2, M3, M7, M8,
and M11 were novel metabolites of pioglitazone, whereas M4, M5,
M6, M9, M10, M12, and M13 were previously reported metabolites
as M-VI, M-IV, M-V, M-VII, M-II, M-IX, and M-III, respectively
(Krieter et al., 1994; Kiyota et al., 1997; Maeshiba et al., 1997; Shen
et al., 2003) (Table 1).
Structure Analysis of M8. The positive ion LC/MS spectrum of
M8 showed a protonated molecule [M ϩ H]ϩ at m/z 345. The positive
ion LC/MS/MS spectrum of the precursor ion [M ϩ H]ϩ at m/z 345
and the proposed fragmentation scheme of M8 are shown in Fig. 4.
Product ions at m/z 119, 134, 240, 254, 280, and 328 were obtained.
Results
Metabolic Profiles of Pioglitazone in Freshly Isolated Human,
Rat, and Monkey Hepatocytes. Representative radiochromatograms The most intense product ion at m/z 134 was formed via the loss of
after incubation of [14C]pioglitazone with freshly isolated human, rat, 211 Da from the ion [M ϩ H]ϩ at m/z 345. The product ions at m/z
and monkey hepatocytes are shown in Fig. 3. Thirteen metabolite 119 and 134 indicated that the 2-(5-ethyl-2-pyridine)ethyl moiety was
peaks in human hepatocytes were detected and designated M1 through not modified. On the other hand, the product ions at m/z 254, 280, and
M13 according to their order of retention time in HPLC. An identical 328 from the ion [M ϩ H]ϩ at m/z 345 suggested thiazolidinedione
FIG. 4. Positive ion LC/MS/MS spectrum of the ion [M ϩ H]ϩ at
m/z 345 and the proposed fragmentation scheme of M8.

IDENTIFICATION OF NOVEL METABOLITES OF PIOGLITAZONE
951
FIG. 5. Positive ion LC/MS/MS spectrum of the ion [M ϩ H]ϩ at
m/z 361 and the proposed fragmentation scheme of M1.
ring metabolism. The product ion at m/z 254 was formed via the at m/z 361. The product ion at m/z 254 was formed via the
elimination of the methylmercapto and the amide groups from the ion elimination of the methyl sulfoxide and the amide groups from the
[M ϩ H]ϩ at m/z 345. Based on these results, M8 was proposed to be ion [M ϩ H]ϩ at m/z 361. Based on these results, M1 was proposed
a thiazolidinedione ring-opened methylmercapto amide (Fig. 4).
Structure Analysis of M1. The positive ion LC/MS spectrum of produced by the oxidation of the sulfur atom of M8 (Fig. 5).
M1 showed a protonated molecule [M ϩ H]ϩ at m/z 361, which was
Structure Analysis of M3. The positive ion LC/MS spectrum of
to be a thiazolidinedione ring-opened methyl sulfoxide amide
16 Da higher than that of M8. The positive ion LC/MS/MS M3 showed a protonated molecule [M ϩ H]ϩ at m/z 377, which was
spectrum of M1, which was obtained by collision-induced disso- 16 Da higher than that of M1. The positive ion LC/MS/MS spectrum
ciation of the ion [M ϩ H]ϩ at m/z 361, and the proposed of the precursor ion [M ϩ H]ϩ at m/z 377 and the proposed fragmen-
fragmentation scheme are shown in Fig. 5. Product ions at m/z 119, tation scheme of M3 are shown in Fig. 6. Product ions at m/z 119, 134,
134, 254, 280, and 297 were obtained. The most intense product 254, 280, and 297 were obtained. The most intense product ion at m/z
ion at m/z 254 was formed via the loss of 107 Da from the ion 254 was formed via the loss of 123 Da from the ion [M ϩ H]ϩ at m/z
[M ϩ H]ϩ at m/z 361. The product ions at m/z 119 and 134 377. The product ions at m/z 119 and 134 indicated that the 2-(5-
indicated that the 2-(5-ethyl-2-pyridine)ethyl moiety was not mod- ethyl-2-pyridine)ethyl moiety was not modified. On the other hand,
ified. On the other hand, the product ions at m/z 254, 280, and 297 the product ions at m/z 254, 280, and 297 from the ion [M ϩ H]ϩ at
provided strong evidence for metabolic transformation of the thia- m/z 377 suggested thiazolidinedione ring metabolism. The product ion
zolidinedione ring. The product ion at m/z 297 was formed via the at m/z 297 was formed via the elimination of the methyl sulfone group
elimination of the methyl sulfoxide group from the ion [M ϩ H]ϩ from the ion [M ϩ H]ϩ at m/z 377. The product ion at m/z 254 was
FIG. 6. Positive ion LC/MS/MS spectrum of the ion [M ϩ H]ϩ at
m/z 377 and the proposed fragmentation scheme of M3.

952
UCHIYAMA ET AL.
FIG. 7. Mass spectrometric comparison of M7 and the authentic standard M7: positive ion LC/MS/MS spectrum of M7 (a), extracted ion chromatogram of the ion [M ϩ
H]ϩ at m/z 533 of M7 (b), positive ion LC/MS/MS spectrum of authentic standard M7 (c), extracted ion chromatogram of the ion [M ϩ H]ϩ at m/z 533 of the authentic
standard M7 (d), and the proposed fragmentation scheme of M7 (e).
formed via the elimination of the methyl sulfone and the amide groups Fig. 8. Product ions at m/z 134, 286, 332, 358, 375, and 417 were
from the ion [M ϩ H]ϩ at m/z 377. Based on these results, M3 was obtained. The product ion at m/z 375 was formed via the loss of 176
proposed to be a thiazolidinedione ring-opened methyl sulfone amide Da from the ion [M ϩ H]ϩ at m/z 551, corresponding to the glucu-
produced by further oxidation of the sulfur atom of the metabolite M1 ronic acid moiety. Therefore, it was suggested that M2 was a gluc-
(Fig. 6).
uronide metabolite. The product ion at m/z 417 was formed via the
Structure Analysis of M7. The positive ion LC/MS spectrum of elimination of a part of the glucuronic acid moiety from the ion [M ϩ
M7 showed a protonated molecule [M ϩ H]ϩ at m/z 533. The positive H]ϩ at m/z 551. The product ions at m/z 286, 332, and 358 suggested
ion LC/MS/MS spectrum of the precursor ion [M ϩ H]ϩ at m/z 533 thiazolidinedione ring metabolism. Furthermore, the positive ion LC/
and the proposed fragmentation scheme of M7 are shown in Fig. 7. MS/MS spectrum (Fig. 8a) and HPLC retention time (Fig. 8b) of M2
Product ions at m/z 134 and 357 were obtained. The most intense were identical to those of the authentic standard M2 (Fig. 8, c and d).
product ion at m/z 357 was formed via the loss of 176 Da from the ion Based on these results, M2 was identified as a thiazolidinedione
[M ϩ H]ϩ at m/z 533, corresponding to the glucuronic acid moiety. ring-opened N-glucuronide, a hydrolyzed metabolite of the thiazo-
The product ion at m/z 134 was formed via the elimination of the lidinedione ring N-glucuronide M7 (Fig. 8e).
2-(5-ethyl-2-pyridine)ethyl moiety from the ion at m/z 357. Further-
Structure Analysis of M11. The positive ion LC/MS spectrum of
more, the positive ion LC/MS/MS spectrum (Fig. 7a) and HPLC M11 showed a protonated molecule [M ϩ H]ϩ at m/z 346. The
retention time (Fig. 7b) of M7 were identical to those of the authentic positive ion LC/MS/MS spectrum of the precursor ion [M ϩ H]ϩ at
standard M7 (Fig. 7, c and d). Based on these results, M7 was m/z 346 and the proposed fragmentation scheme of M11 are shown in
identified as a thiazolidinedione ring N-glucuronide (Fig. 7e).
Fig. 9. Product ions at m/z 119, 134, 254, 270, and 286 were obtained.
Structure Analysis of M2. The positive ion LC/MS spectrum of The most intense product ion at m/z 134 was formed via the loss of
M2 showed a protonated molecule [M ϩ H]ϩ at m/z 551, which was 212 Da from the ion [M ϩ H]ϩ at m/z 346. The product ions at m/z
18 Da higher than that of the thiazolidinedione ring N-glucuronide 119 and 134 indicated that the 2-(5-ethyl-2-pyridine)ethyl moiety was
M7. The positive ion LC/MS/MS spectrum of the ion [M ϩ H]ϩ at not modified. On the other hand, the product ions at m/z 254 and 286
m/z 551 and the proposed fragmentation scheme of M2 are shown in from the ion [M ϩ H]ϩ at m/z 346 suggested thiazolidinedione ring

IDENTIFICATION OF NOVEL METABOLITES OF PIOGLITAZONE
953
FIG. 8. Mass spectrometric comparison of M2 and the authentic standard M2: positive ion LC/MS/MS spectrum of M2 (a), extracted ion chromatogram of the ion [M ϩ
H]ϩ at m/z 551 of M2 (b), positive ion LC/MS/MS spectrum of authentic standard M2 (c), extracted ion chromatogram of the ion [M ϩ H]ϩ at m/z 551 of the authentic
standard M2 (d), and the proposed fragmentation scheme of M2 (e).
metabolism. The product ion at m/z 286 was formed via the elimina- product ions at m/z 119 and 134 indicated that the 2-(5-ethyl-2-
tion of the methyl and the carboxylic acid groups from the ion [M ϩ pyridine)ethyl moiety was not modified. On the other hand, the
H]ϩ at m/z 346. The product ion at m/z 254 was formed via the product ions at m/z 254 and 286 from the ion [M ϩ H]ϩ at m/z 332
elimination of the methylmercapto and the carboxylic acid groups suggested thiazolidinedione ring metabolism. The product ion at
from the ion [M ϩ H]ϩ at m/z 346. Based on these results, M11 was m/z 286 was formed via the elimination of the carboxylic acid
proposed to be a thiazolidinedione ring-opened methylmercapto car- group from the ion [M ϩ H]ϩ at m/z 332. The product ion at m/z
boxylic acid (Fig. 9).
254 was formed via the elimination of the mercapto and the
Metabolism of M7 in Human Liver S9. The metabolism of M7 carboxylic acid groups from the ion [M ϩ H]ϩ at m/z 332. Based
was studied using human liver S9. Incubation of M7 with human liver on these results, S9P1 was proposed to be a thiazolidinedione
S9 in the presence of SAM yielded a thiazolidinedione ring-opened ring-opened mercapto carboxylic acid (Fig. 11).
N-glucuronide M2, a newly generated product in human liver S9
Discussion
(S9P1), and a thiazolidinedione ring-opened methylmercapto carbox-
ylic acid M11 (Fig. 10). Incubation of M7 with human liver S9 in the
In this study, we detected 13 metabolites (M1–M13) in an in vitro
absence of SAM yielded M2 and S9P1 (data not shown). M2 and metabolic system of pioglitazone with freshly isolated human, rat, and
S9P1 also resulted from incubation of M7 without human liver S9 and monkey hepatocytes, and their structures were elucidated by LC/MS/
with SAM alone (data not shown).
MS. Furthermore, metabolites M2, S9P1, and M11 were detected in
Structure Analysis of S9P1. The positive ion LC/MS spectrum an in vitro metabolic system with human liver S9 using M7 (authentic
of S9P1 showed a protonated molecule [M ϩ H]ϩ at m/z 332. The standard) as a substrate. Based on the structures of the elucidated
positive ion LC/MS/MS spectrum of the precursor ion [M ϩ H]ϩ metabolites, the in vitro metabolic pathways for pioglitazone are
at m/z 332 and the proposed fragmentation scheme of S9P1 are proposed, as shown in Fig. 12.
shown in Fig. 11. Product ions at m/z 119, 134, 254, 270, and 286
Novel metabolic pathways for pioglitazone, N-glucuronidation
were obtained. The most intense product ion at m/z 134 was formed of the thiazolidinedione ring of pioglitazone to form M7 and
via the loss of 198 Da from the ion [M ϩ H]ϩ at m/z 332. The hydrolysis of M7 to form M2, were observed in hepatocytes. As a

954
UCHIYAMA ET AL.
FIG. 9. Positive ion LC/MS/MS spectrum of
the ion [M ϩ H]ϩ at m/z 346 and the proposed
fragmentation scheme of M11.
result of the presence of these metabolites, the metabolic pathway atomic mass units. However, they have not found N-glucuronide of
via the thiazolidinedione ring N-glucuronide M7 and subsequent pioglitazone. On the other hand, we have focused on the exhaustive
hydrolysis to the thiazolidinedione ring-opened N-glucuronide M2 investigation to find minor metabolites in freshly isolated hepato-
is proposed. This thiazolidinedione ring-opening pathway is dis- cytes using [14C]pioglitazone (30 ␮M).
tinct from the known oxidative thiazolidinedione ring-opening
Moreover, in the human liver S9 study using M7, we inferred that M7
pathway for troglitazone (Kassahun et al., 2001) and pioglitazone is metabolized to S9P1 via M2 and consequent S-methylation to M11
(Baughman et al., 2005). Baughman et al. (2005) investigated the immediately in the presence of SAM. These are also novel metabolic
metabolites related to reactive metabolite formation using nonla- pathways for pioglitazone. We have recently reported similar metabolic
beled pioglitazone (10 ␮M) in freshly isolated human hepatocytes pathways in another thiazolidinedione ring-containing drug, rosiglitazone
by the detection method of a generic neutral loss method of 129 (Uchiyama et al., 2010). Thus, the metabolic pathways of ring opening
FIG. 10. Extracted ion chromatograms of the metabolites after incubation of M7 with human liver S9 in the presence of SAM. [M ϩ H]ϩ at m/z 551 of thiazolidinedione
ring-opened N-glucuronide M2 (a), [M ϩ H]ϩ at 332 of S9P1 (b), and [M ϩ H]ϩ at m/z 346 of thiazolidinedione ring-opened methylmercapto carboxylic acid M11 (c).
Positive ion LC/MS analysis was conducted as described under Materials and Methods.

IDENTIFICATION OF NOVEL METABOLITES OF PIOGLITAZONE
955
FIG. 11. Positive ion LC/MS/MS spectrum of
the ion [M ϩ H]ϩ at m/z 332 and the proposed
fragmentation scheme of S9P1.
FIG. 12. Proposed in vitro metabolic pathways
for pioglitazone. Bold indicates that metabolites
M1, M2, M3, M7, M8, M11, and S9P1 were
novel metabolites of pioglitazone.
via N-glucuronidation of the thiazolidinedione ring are assumed to be al., 2005), bucolome (5-n-butyl-1-cyclohexyl-2,4,6-trioxoperhydropyri-
common metabolic reactions to the thiazolidinedione ring.
midine, BCP) (Mohri et al., 1985), and nirvanol (5-ethyl-5-phenylhydan-
As examples of N-glucuronidation to analogs related to thiazo- toin) (Maguire et al., 1982) for N-glucuronidation of cyclic amide com-
lidinediones, N-glucuronides of MaxiPost [(3S)-(ϩ)-(5-chloro-2- pounds, and that of PNU-107859 for N-glucuronidation of cyclic
methoxyphenyl)-1,3-dihydro-3-fluoro-6-(trifluoromethyl)-2H-indol-2- thioamide compound (Kuo et al., 1999) have been reported. For all cases,
one, BMS-204352; Bristol-Myers Squibb Co., Stamford, CT] (Zhang et N-glucuronides were found to be major metabolites and stable. The

956
UCHIYAMA ET AL.
effectiveness of pioglitazone in the treatment of type 2 diabetes mellitus. Clin Ther 23:1792–
1823; discussion 1791.
Eckland DA and Danhof M (2000) Clinical pharmacokinetics of pioglitazone. Exp Clin Endo-
crinol Diabetes 108 (Suppl 2):234–242.
Gale EA (2001) Lessons from the glitazones: a story of drug development. Lancet 357:1870–
1875.
Graham DJ, Green L, Senior JR, and Nourjah P (2003) Troglitazone-induced liver failure: a case
study. Am J Med 114:299–306.
He K, Talaat RE, Pool WF, Reily MD, Reed JE, Bridges AJ, and Woolf TF (2004) Metabolic
activation of troglitazone: identification of a reactive metabolite and mechanisms involved.
Drug Metab Dispos 32:639–646.
Jaakkola T, Laitila J, Neuvonen PJ, and Backman JT (2006) Pioglitazone is metabolised by
CYP2C8 and CYP3A4 in vitro: potential for interactions with CYP2C8 inhibitors. Basic Clin
Pharmacol Toxicol 99:44–51.
Karanam BV, Hop CE, Liu DQ, Wallace M, Dean D, Satoh H, Komuro M, Awano K, and
Vincent SH (2004) In vitro metabolism of MK-0767 [(Ϯ)-5-[(2,4-dioxothiazolidin-5-
yl)methyl]-2-methoxy-N-[[(4-trifluoromethyl) phenyl]methyl]benzamide], a peroxisome pro-
liferator-activated receptor ␣/␥ agonist. I. Role of cytochrome P450, methyltransferases, flavin
monooxygenases, and esterases. Drug Metab Dispos 32:1015–1022.
nonenzymatic hydrolysis of synthetic M7 to M2 was observed (half-life
of approximately 0.7 h) in buffer solution, pH 7.4, during incubation at
37°C (data not shown). The chemical mechanism of ring opening of
thiazolidinedione of pioglitazone is unknown. However, after N-
glucuronidation of the thiazolidinedione ring, the formation of hydrogen
bonding between the hydroxyl group at 2-position of the glucuronic acid
moiety and the carbonyl group at 4-position of the thiazolidinedione ring
would be proposed. As a result, the thiazolidinedione ring may increase
electrophilicity and be easily attacked by nucleophile and hydrolyzed. In
addition, if the chemical reactivity of M7 is very high, there is the
possibility of detecting glutathione conjugates, but no glutathione conju-
gates were observed in this study.
The known oxidative thiazolidinedione ring opening and subsequent
pathways via reactive intermediate generation of troglitazone have been
proposed based on identification of glutathione conjugate metabolites
using human liver microsomes and cDNA-expressed cytochrome P450
isoforms (Kassahun et al., 2001). Similar metabolic pathways have been
reported for pioglitazone (Baughman et al., 2005) and the thiazolidinedi-
one derivatives MK-0767 [(Ϯ)-5-[(2,4-dioxothiazolidin-5-yl)methyl]-2-
methoxy-N-[[(4-trifluoromethyl)-phenyl]methyl]benzamide] (Karanam
et al., 2004; Liu et al., 2004; Kochansky et al., 2006) and MRL-A
[(Ϯ)-5-[(2,4-dioxothiazolidin-5-yl)methyl]-2-methoxy-N-[[(4-trifluoro-
methoxy)phenyl]methyl]benzamide] (Reddy et al., 2005). In this study,
we have identified M8, M1, and M3 as new metabolites of pioglitazone
via the thiazolidinedione ring-opened mercapto amide. We estimate that
these metabolites are sequentially generated by the oxidative thiazo-
lidinedione ring opening and subsequent pathways, which agree well
with the previous proposed metabolic activation scheme of pioglitazone
(Baughman et al., 2005).
Kassahun K, Pearson PG, Tang W, McIntosh I, Leung K, Elmore C, Dean D, Wang R, Doss G,
and Baillie TA (2001) Studies on the metabolism of troglitazone to reactive intermediates in
vitro and in vivo. Evidence for novel biotransformation pathways involving quinone methide
formation and thiazolidinedione ring scission. Chem Res Toxicol 14:62–70.
Kiyota Y, Kondo T, Maeshiba Y, Hashimoto A, Yamashita K, Yoshimura Y, Motohashi M, and
Tanayama S (1997) Studies on the metabolism of the new antidiabetic agent pioglitazone.
Identification of metabolites in rats and dogs. Arzneimittelforschung 47:22–28.
Kochansky CJ, Rippley RK, Yan KX, Song H, Wallace MA, Dean D, Jones AN, Lasseter K,
Schwartz J, Vincent SH, et al. (2006) Absorption, metabolism, and excretion of [14C]MK-0767
(2-methoxy-5-(2,4-dioxo-5-thiazolidinyl)-N-[[4-(trifluoromethyl)phenyl] methyl]benzamide)
in humans. Drug Metab Dispos 34:1457–1461.
Krieter PA, Colletti AE, Doss GA, and Miller RR (1994) Disposition and metabolism of the
hypoglycemic agent pioglitazone in rats. Drug Metab Dispos 22:625–630.
Kuo MS, Yurek DA, Mizsak SA, Prairie MD, Mattern SJ, and DeKoning TF (1999) Isolation and
identification of a major metabolite of PNU-107859, an MMP inhibitor from the biliary fluid
of rats. J Pharm Sci 88:705–708.
Liu DQ, Karanam BV, Doss GA, Sidler RR, Vincent SH, and Hop CE (2004) In vitro metabolism
of MK-0767 [(Ϯ)-5-[(2,4-dioxothiazolidin-5-yl)methyl]-2-methoxy-N-[[(4-trifluoromethyl)-
phenyl] methyl]benzamide], a peroxisome proliferator-activated receptor ␣/␥ agonist. II.
Identification of metabolites by liquid chromatography-tandem mass spectrometry. Drug
Metab Dispos 32:1023–1031.
Maeshiba Y, Kiyota Y, Yamashita K, Yoshimura Y, Motohashi M, and Tanayama S (1997)
Disposition of the new antidiabetic agent pioglitazone in rats, dogs, and monkeys. Arzneimit-
telforschung 47:29–35.
Besides those described above, the known metabolic pathways for
pioglitazone (Krieter et al., 1994; Kiyota et al., 1997; Maeshiba et al.,
1997; Shen et al., 2003) were also confirmed in this study: the
generation of M4 (M-VI), M5 (M-IV), M6 (M-V), M9 (M-VII), M10
(M-II), M12 (M-IX), and M13 (M-III). Thus, because these known
metabolites observed in vivo have been determined in this study, these
hepatocyte systems seem to reflect the in vivo metabolism of piogli-
tazone. However, because the ratios of the peak area of M7 and M2
to M5, a major metabolite in Fig. 3, were 0.06 and 0.12 for human
hepatocytes, there may be a low amount of these metabolites formed
in vivo.
Maguire JH, Butler TC, and Dudley KH (1982) 5-Ethyl-5-phenylhydantoin N-glucuronide, the
major urinary metabolite of 5-ethyl-5-phenylhydantoin (Nirvanol) in the dog. Drug Metab
Dispos 10:595–598.
Mohri K, Uesugi T, and Kamisaka K (1985) Bucolome N-glucuronide: purification and identi-
fication of a major metabolite of bucolome in rat bile. Xenobiotica 15:615–621.
Molde´us P, Ho¨gberg J, and Orrenius S (1978) Isolation and use of liver cells. Methods Enzymol
52:60–71.
Muschler E, Lal J, Jetter A, Rattay A, Zanger U, Zadoyan G, Fuhr U, and Kirchheiner J (2009)
The role of human CYP2C8 and CYP2C9 variants in pioglitazone metabolism in vitro. Basic
Clin Pharmacol Toxicol 105:374–379.
Reddy VB, Karanam BV, Gruber WL, Wallace MA, Vincent SH, Franklin RB, and Baillie TA
(2005) Mechanistic studies on the metabolic scission of thiazolidinedione derivatives to
acyclic thiols. Chem Res Toxicol 18:880–888.
Shen Z, Reed JR, Creighton M, Liu DQ, Tang YS, Hora DF, Feeney W, Szewczyk J, Bakhtiar
R, Franklin RB, et al. (2003) Identification of novel metabolites of pioglitazone in rat and dog.
Xenobiotica 33:499–509.
Some toxicological concerns have been expressed regarding the oxi-
dative thiazolidinedione ring-opening reaction of thiazolidinedione-con-
taining drugs (Kassahun et al., 2001; Tettey et al., 2001; Smith, 2003; He
et al., 2004; Baughman et al., 2005; Alvarez-Sanchez et al., 2006). For
this new ring-opening reaction via N-glucuronidation of pioglitazone,
further studies are needed to investigate the toxicological evaluation.
Smith MT (2003) Mechanisms of troglitazone hepatotoxicity. Chem Res Toxicol 16:679–687.
Sood V, Colleran K, and Burge MR (2000) Thiazolidinediones: a comparative review of
approved uses. Diabetes Technol Ther 2:429–440.
Tettey JN, Maggs JL, Rapeport WG, Pirmohamed M, and Park BK (2001) Enzyme-induction
dependent bioactivation of troglitazone and troglitazone quinone in vivo. Chem Res Toxicol
14:965–974.
Thasler WE, Weiss TS, Schillhorn K, Stoll PT, Irrgang B, and Jauch KW (2003) Charitable
state-controlled foundation human tissue and cell research: ethic and legal aspects in the
supply of surgically removed human tissue for research in the academic and commercial sector
in Germany. Cell Tissue Bank 4:49–56.
Uchiyama M, Oguchi M, Yamamura N, Koda H, Fischer T, Mueller J, Izumi T, and Iwabuchi
H (2010) Identification of novel metabolites of rosiglitazone in freshly isolated human, rat, and
monkey hepatocytes by liquid chromatography/tandem mass spectrometry. J Mass Spectrom
Soc Jap 58:1–11.
Acknowledgments. We thank Dr. Osamu Okazaki, Drug Metabo-
lism and Pharmacokinetics Research Laboratories, Daiichi Sankyo
Co., Ltd., for careful reading of the document and scientific expertise.
Weiss TS, Pahernik S, Scheruebl I, Jauch KW, and Thasler WE (2003) Cellular damage to
human hepatocytes through repeated application of 5-aminolevulinic acid. J Hepatol 38:476–
482.
Yamamoto Y, Yamazaki H, Ikeda T, Watanabe T, Iwabuchi H, Nakajima M, and Yokoi T (2002)
Formation of a novel quinone epoxide metabolite of troglitazone with cytotoxicity to HepG2
cells. Drug Metab Dispos 30:155–160.
Zhang D, Krishna R, Wang L, Zeng J, Mitroka J, Dai R, Narasimhan N, Reeves RA, Srinivas NR,
and Klunk LJ (2005) Metabolism, pharmacokinetics, and protein covalent binding of radio-
labeled MaxiPost (BMS-204352) in humans. Drug Metab Dispos 33:83–93.
References
Alvarez-Sanchez R, Montavon F, Hartung T, and Pa¨hler A (2006) Thiazolidinedione bioactiva-
tion: a comparison of the bioactivation potentials of troglitazone, rosiglitazone, and pioglita-
zone using stable isotope-labeled analogues and liquid chromatography tandem mass spec-
trometry. Chem Res Toxicol 19:1106–1116.
Baba A and Yoshioka T (2007) An improved chemo-enzymatic synthesis of 1-␤-O-acyl glucu-
ronides: highly chemoselective enzymatic removal of protecting groups from corresponding
methyl acetyl derivatives. J Org Chem 72:9541–9549.
Baba S (2001) Pioglitazone: a review of Japanese clinical studies. Curr Med Res Opin 17:166–
189.
Baughman TM, Graham RA, Wells-Knecht K, Silver IS, Tyler LO, Wells-Knecht M, and Zhao
Z (2005) Metabolic activation of pioglitazone identified from rat and human liver microsomes
and freshly isolated hepatocytes. Drug Metab Dispos 33:733–738.
Address correspondence to: Takashi Izumi, Drug Metabolism and Pharmaco-
kinetics Research Laboratories, Daiichi Sankyo Co., Ltd., 1-2-58 Hiromachi, Shina-
gawa-ku, Tokyo, 140-8710, Japan. E-mail: izumi.takashi.ka@daiichisankyo.co.jp
Chilcott J, Tappenden P, Jones ML, and Wight JP (2001) A systematic review of the clinical