journa l homepage: www.e lsev ier .com/ locate / tox le t
irect toxicity effects of sulfo-conjugated troglitazone on human hepatocytes
udipta Saha, Lee Sun New, Han Kiat Ho, Wai Keung Chui, Eric Chun Yong Chan ∗
epartment of Pharmacy, Faculty of Science, National University of Singapore, 18 Science Drive 4, Singapore 117543, Singapore
r t i c l e i n f o
rticle history:eceived 18 November 2009eceived in revised form 3 February 2010ccepted 15 March 2010vailable online 20 March 2010
eywords:roglitazone
a b s t r a c t
Troglitazone (TGZ), an orally active hypoglycemic agent, was found to be associated with severe drug-induced liver failure and was withdrawn from the market in 2000. Although the exact mechanism is notclear, it has been postulated that the formation of its major sulfo-conjugated metabolite (TGZS) playsan important role in its toxicity. TGZS inhibits bile salt export pump (BSEP) that causes accumulation ofbile salts in liver. High concentration of bile salts causes cell death and mitochondrial dysfunction viadetergent properties. One question arises whether TGZS has direct toxicity effect on human liver cells inaddition to BSEP inhibition. In this study, both TGZ and chemically synthesized TGZS were incubated with
roglitazone sulphateuman hepatocyteslutathione
n vitro cell toxicityxidative stress
normal human hepatocytes (THLE-2 cells) for measuring their cytotoxicity in vitro using the MTT assay.Glutathione (GSH) and protein carbonyl (PC) assays were further performed to measure the oxidativestress generated by these two compounds during incubation with THLE-2 cells. The results from thisstudy indicated that TGZS (EC50 = 21.74 ± 5.38 �M) was more toxic than TGZ (EC50 = 41.12 ± 4.3 �M) inTHLE-2 cells. The GSH and PC data further confirmed that TGZS produced greater oxidative stress inTHLE-2 cells as compared to TGZ. In conclusion, our study demonstrated for the first time that TGZS has
uman
direct toxicity effect on h
. Introduction
Troglitazone (TGZ, Fig. 1), a 2,4-thiazolidinedione (TZD) antidi-betic drug, was developed and used for the treatment of typeI diabetes mellitus (non-insulin dependent diabetes mellitus)Parker, 2002). It lowers the glucose concentration in blood byxerting partial agonistic action on the peroxisome proliferatorctivated receptor gamma (PPAR�) (Henry, 1997). Upon PPAR�
ctivation, TGZ increases serum triglycerides level, alters thyroidetabolism (Miller et al., 1997; Sherman et al., 1999; Rizvi et
l., 1999) and inhibits phosphoenolpyruvate carboxylase enzymehich is mainly responsible for hepatic gluconeogenesis (Camp et
l., 2000). Several cases of fulminant hepatic failure were reportedith the clinical use of TGZ, leading to its withdrawal from thearket in 2000 (Chojkier, 2005). The exact nature of TGZ hep-
totoxicity has not yet been completely established but severalechanisms had been proposed (Smith, 2003). These include the
ormation of electrophilic reactive intermediates, binding to PPAR�,ausation of mitochondrial injury, inhibition of bile salt export
ump (BSEP) by TGZ and/or TGZS and a combination of mul-iple mechanisms and varying host factors (Kawai et al., 1997;ostrubsky et al., 2000). In vitro and in vivo experiments suggested
hat TGZ is mainly metabolized by cytochrome 3A4 (CYP3A4) and
2C8 (CYP2C8) enzymes (Sahi et al., 2000; Ramachandran et al.,1999; He et al., 2001). The metabolism of TGZ in liver includes oxi-dation (phase I metabolism), sulfation and glucuronidation (phaseII metabolism). The glucuronide metabolite of TGZ is found inhuman plasma at very low concentration while the major metabo-lites of TGZ in humans are the sulphate (TGZS, 70%, Fig. 1) andquinone (10%) conjugates, respectively (Loi et al., 1997, 1999).
While it was traditionally perceived that the phase II TGZSmetabolite is a detoxifying species, recent evidence suggested thatit might have toxicological potential through the binding and inhi-bition of BSEP (Funk et al., 2001). As BSEP plays an important rolein removing bile salts from liver cells using energy in the form ofATP (Kullak-Ublick et al., 2000), the accelerated accumulation ofbile salts may lead to cholestasis and subsequent hepatocyte apop-tosis. High levels of bile salt have been shown to induce cell deathand mitochondrial dysfunction due to their detergent properties(Delzenne et al., 1992; Gores et al., 1998). The cholestatic potentialof TGZ and TGZS had been demonstrated previously using in vitroand in vivo rat models (Funk et al., 2001).
Although TGZS has been demonstrated to inhibit BSEP, its directtoxic effect on human liver cells has not been investigated. It isimportant to ascertain this property of TGZS because it is a major
metabolite of TGZ and its mechanism of direct hepatocyte toxicityis still unclear. While it had been determined that TGZS showedno cytotoxicity in human hepatoma cell lines such as HepG2,HLE, HLF and HuH-7 cells (Yamamoto et al., 2001), its effect onphysiologically relevant normal human hepatocytes has not been
136 S. Saha et al. / Toxicology Letters 195 (2010) 135–141
(R = –H
ewtcth
2
2
(s(piSwlMwrVBL
2
p
2
pagb
2
oc3iotpaswgwt
Fig. 1. Structures of TGZ
stablished. In this study, TGZ and chemically synthesized TGZSere evaluated for their in vitro toxicity and oxidative stress poten-
ials using metabolically active normal human hepatocytes, THLE-2ells (Andrea et al., 1993; Saha et al., 2010). It is demonstrated forhe first time that TGZS has direct toxic effect on normal humanepatocytes.
. Materials and methods
.1. Chemicals and reagents
HPLC-grade acetonitrile (ACN) was purchased from Tedia Company Inc.,Fairfield, OH, USA). Fibronectin and collagen were purchased from BD Bio-ciences (Woburn, MA, USA). TGZ was purchased from Cayman Chemical CompanyAnn Arbor, MI, USA). Formic acid (99% purity) and glacial acetic acid wereurchased from VWR International Ltd., (Leicestershire, UK). Dimethyl sulfox-
de (DMSO, ACS grade) was obtained from Panreac Quimica SA (Barcelona,pain). Ethylenediaminetetraacetic acid disodium salt dihydrate (Na2EDTA·2H2O)ith > 98% purity was obtained from Duchefa Biochemie B. V. (Haarlem, The Nether-
ands). Dichloromethane (DCM) and ethyl acetate (EtOAc) were purchased fromerck (Merck Pte. Ltd., Singapore). Bronchial epithelium growth media providedith supplements (BEGM BulletKit, CC-3170) was purchased from Cambrex Corpo-
ation (Rutherford, NJ, USA). THLE-2 cell line was purchased from ATCC (Manassas,A, USA). Water was purified using a Milli-Q water purification system (Millipore,edford, MA, USA). All other reagents were purchased from Sigma–Aldrich (Saintouis, MO, USA).
.2. Preparation of TGZS
TGZS was prepared as outlined in the supplementary data (Figs. 1 and 2, sup-lementary data sheet).
.3. THLE-2 cells
For all experiments, THLE-2 cells were maintained in the logarithmic growthhase in BEGM supplemented with 10% foetal bovine serum (FBS) under humidifiedir with 5% CO2 at 37 ◦C. Cells were cultured in plates pre-coated with colla-en I (2.9 mg/mL), fibronectin (1 mg/mL) and bovine serum albumin (1 mg/mL) inronchial epithelial basal media (BEBM) according to ATCC guidelines.
.4. Sulfotransferase enzyme inhibition assay
THLE-2 cells were seeded into a 6-well plate (1.0 × 107 cells per well). 100 �Mf either TGZ alone or TGZ and quercetin (QR) in combination were added to theells and the mixtures were incubated under a humidified condition with 5% CO2 at7 ◦C. QR is a phenol sulfotransferase enzyme inhibitor and its concentration used
n each incubation was 50 �M. The final concentration of the test compounds andrganic solvent were 50 �M and 0.2% (v/v), respectively. After 24 h (t24), the reac-ion was terminated by addition of 2 mL of ice cold ACN:water (50:50). Initial timeoint samples (t0) were prepared by terminating the reaction immediately after
dding the test compound to the THLE-2 cells with same quenching solvent. Theseamples were centrifuged at 13,000 rpm at 4 ◦C for 10 min and the supernatantsere removed and evaporated to dryness at 35 ◦C for 8 h under gentle flow of nitro-
en gas using the Turbovap LV (Caliper Life Science, Hopkinton, MA, USA). Theyere each reconstituted with 50 �L of ACN:water (50:50), vortex-mixed and cen-
rifuged at 13,000 rpm at 4 ◦C for 2 min. 5 �L of the supernatant was subjected to
) and TGZS (R = –SO3H).
UPLC/MS/MS analysis. Negative controls were prepared using incubated samplesthat either contained THLE-2 cells and media or test compounds and media only.
2.5. MTT viability assay
MTT assay was performed as described previously (Marcsek et al., 2007).Briefly, THLE-2 cells were seeded into 96-well plate (1.0 × 104 cells per well). After24 h, the media was removed and 200 �L of TGZ and TGZS at varying concen-trations (1–125 �M) in media were added to each well and the samples wereincubated under a humidified condition with 5% CO2 at 37 ◦C. After 72 h, MTTsolution (1 mg/mL, 50 �L) was added to each well and incubated for another 4 h.After which, the media and excess MTT were removed and 150 �L of DMSO wasadded. The absorbance at 570 nm was recorded using the Infinite M200 microplatereader controlled by Magellan and i-control softwares (Tecan Group Ltd., Mannedorf,Switzerland). Samples containing media with and without cells were also analyzedand labelled as ‘control’ and ‘blank’, respectively. In a separate experiment, 50 �MQR was co-incubated with TGZ and TGZS at varying concentrations (1–125 �M)in THLE-2 cells (Virginia et al., 2004). In parallel, another cell viability assay wasperformed using various concentrations of QR (1–125 �M) in THLE-2 cells in theabsence of TGZ and TGZS. In addition, another MTT assay was also performed byspiking 20 �M of tocopherol acetate (TA, antioxidant) together with TGZ and TGZSat varying concentrations (1–125 �M) in THLE-2 cells (Patrick et al., 2005). Inde-pendently, we performed the cytotoxicity assay of TA (1–100 �M concentrations)in THLE-2 cells in the absence of TGZ and TGZS. All experiments were performedin triplicates. The final organic solvent concentration was 0.3% (v/v). Cell survival(% of control) was calculated relative to untreated control cells. The data were rep-resented as the mean ± S.D. of triplicate experiments to determine the EC50 of therespective compounds. Statistical differences were analyzed using one-way analy-sis of variance (one-way ANOVA) and paired T-test. P < 0.05 represented statisticallysignificant data.
2.6. GSH depletion assay
The details of our GSH assay using NEM as a derivatizing agent to preventauto-oxidation of GSH had been previously published (New and Chan, 2008). Theconcentrations of the GSH-NEM calibrants were 0.001, 0.01, 0.05, 0.1, 0.5, 1.0 and10.0 �M. Quality control (QC) samples were also prepared at concentrations of 0.003,0.1 and 8.0 �M, representing low, medium and high QC samples (LQC, MQC andHQC), respectively. Triplicate calibrants and QC samples were prepared at eachconcentration.
THLE-2 cells were seeded in 24-well plate (2.0 × 104 cells per well). After 24 h,1 mL of the respective concentration of TGZ and TGZS in BEGM (1–125 �M) wasadded to each well and incubated under a humidified condition with 5% CO2 at 37 ◦C.After 72 h incubation, the cells were washed with phosphate buffer saline (PBS, pH7.4) and trypsinized using 0.05% trypsin-EDTA. Subsequently, the excess trypsin wasinactivated with BEGM containing FBS. The suspended cells were then transferredinto clean tubes, centrifuged at 13,000 rpm for 2 min and the supernatant removed.After which, 200 �L of water was added to the cell samples and 50 �L of solutionA (mixture of NEM [250 mM] and Na2EDTA·2H2O [1.5 mg/mL] in water:methanol[3:2], pH 7.4) was added to it. The mixture was vortex-mixed at high speed for 20 s.
Then 50 �L of TCA (10%) was added and the mixture was centrifuged at 13,000 rpmfor 5 min at 4 ◦C. The supernatant was transferred into clean tubes and subjected toNEM derivatization and extraction as previously described (New and Chan, 2008).Total protein content of cell suspension was measured at 72 h using Pierce MicroBCATM (Bicinchoninic acid) Protein Assay Kit. GSH depletion (�M) per �g of proteinwas calculated at varying concentrations of TGZ and TGZS after 72 h.
S. Saha et al. / Toxicology Letters 195 (2010) 135–141 137
F mple2
2
Me(ta
uU0aoim(aT
tmwt
1rpottr
2
TpTwtr
ig. 2. (A) Representative LC/MS/MS chromatogram of TGZS in “24 h TGZ-THLE-2” sa” sample (QRS1, RT 3.09 min; QRS2, RT 6.68 min).
.7. LC/MS/MS conditions
The LC/MS/MS system consisted of an ACQUITY UPLC system (Waters, Milford,A, USA) interfaced with a hybrid triple quadrupole linear ion trap mass spectrom-
ter (QTRAP MS) equipped with TurboIonSpray electrospray ionization (ESI) source3200 QTRAP, Applied Biosystems, Foster City, CA, USA). The UPLC/QTRAP/MS sys-em was both controlled by Analyst 1.4.2 software (Applied Biosystems). The columnnd autosampler temperatures were maintained at 60 and 4 ◦C, respectively.
For the profiling of sulfotransferase enzyme inhibition assay of both TGZ and QRsing THLE-2 cells, chromatographic separations were performed on an ACQUITYPLC BEH C18 1.7 �m 100 mm × 2.1 mm i.d. column (Waters). The flow rate was set at.5 mL/min and the mobile phases consisted of solvent A (0.1% formic acid in water)nd solvent B (0.1% formic acid in ACN). The elution conditions were: linear gradientf 5–95% solvent B (0–9.50 min), isocratic at 95% solvent B (9.50–10.50 min) and thensocratic at 5% solvent B (10.51–12.00 min). Electrospray ionization negative (ESI-)
ode was employed throughout the experiments. The multiple reaction monitoringMRM) transition for TGZS was m/z 520.0 → 440.0 and single ion monitoring (SIM)t m/z 381.0 for quercetin sulphate (QRS). The MS conditions are summarized inable 1, supplementary data sheet.
For the stability profiling of TGZS, the LC experimental setup was similar tohat we used for the profiling of sulfotransferase enzyme inhibition assay. ESI-
ode was employed throughout the experiments and MRM transition for TGZSas m/z 520.0 → 440.0. The MS conditions are summarized in Table 1, supplemen-
ary data sheet.In the measurement of GSH depletion assay, an ACQUITY UPLC HSS T3 1.8 �m
00 mm × 2.1 mm i.d. column (Waters) was used to achieve chromatographic sepa-ation. The mobile phases and optimized elution conditions utilized were discussedreviously (New and Chan, 2008). ESI+ mode and MRM were employed through-ut the experiments. The MRM transition for GSH-NEM was m/z 433.0 → 304.0 andhe MS conditions are summarized in Table 2, supplementary data sheet. Statis-ical differences were analyzed using one-way ANOVA and paired T-test. P < 0.05epresented statistically significant data.
.8. Protein carbonyl (PC) assay
PC assay was performed as described previously (Reznick and Packer, 1994).
HLE-2 cells were seeded at 5.0 × 104 cells per well in the 6-well plate which wasreviously coated with the fibronectin-collagen media. After 24 h, 2 mL of TGZ andGZS in media at varying concentrations (1, 25 and 50 �M) were added to eachell and incubated at 37 ◦C in a humidified condition with 5% CO2. After 48 h,
he media were removed and cells were washed with PBS. Then, the cells wereemoved from the surface of the plates using 0.05% trypsin-EDTA and the excess
(RT, 4.92 min). (B) Representative LC/MS/MS chromatogram of “24 h TGZ-QR-THLE-
trypsin was deactivated by adding media which contained FBS. Media containingcells were transferred into clean tube and centrifuged at 13,000 rpm for 2 min. Thesupernatant was removed, the cell pellets were reconstituted with 300 �L of waterand vortexed for 1 min. 150 �L of the cell suspension was transferred to a cleantube and precipitated using 500 �L of 10% TCA solution. Then the tubes were cen-trifuged at 13,000 rpm for 2 min and the supernatant was removed. Each cell pelletwas incubated with 500 �L of 0.2% 2,4-dinitrophenylhydrazine for 1 h with con-stant vortexing at every 5 min interval. Subsequently, 50 �L of 100% TCA solutionwas added and the mixture was vortexed and centrifuged at 13,000 rpm for 5 min.All the supernatant was removed and the cell pellet was washed with 500 �L ofethanol:EtOAc (50:50) mixture three times. Finally, the cell pellet was dissolvedin 600 �L of 6 M guanidine hydrochloride where 300 �L was further transferredto 96-well plate and absorbance was measured at 360 nm using the Infinite M200microplate reader. The remaining 150 �L of cell suspension were used for measur-ing total protein content using the Pierce Micro BCATM (Bicinchoninic acid) ProteinAssay Kit. PC content (�M) per microgram of protein was calculated at varying con-centrations of TGZ and TGZS. Statistical differences were analyzed using one-wayANOVA and paired T-test. P < 0.05 represented statistical significance between thetest groups.
2.9. Stability profile of TGZ and TGZS using BEGM
Stability profiling of the parent TGZ and synthesized TGZS at 1.0, 50.0 and150.0 �M were performed using BEGM at 37 ◦C up to 72 h. At various time points (0,24, 48 and 72 h), 1 mL methanol was added to 1 mL of incubation mixture. Subse-quently, each sample was centrifuged at 13,000 rpm for 5 min and the supernatantwas removed and evaporated to dryness at 40 ◦C for 8–10 h under gentle flow ofnitrogen gas using the Turbovap LV. Each dried residue was reconstituted with50 �L of methanol, vortex-mixed and centrifuged at 13,000 rpm for 2 min. 5 �L ofthe supernatant was subjected to UPLC/MS/MS analysis. Triplicate analyses wereperformed at each time point.
3. Results
3.1. Sulfotransferase enzyme inhibition assay
LC/MS/MS analysis of THLE-2 cells incubation with TGZ revealedthe presence of TGZS using MRM at m/z 520.0 → 440.0. The MRMexperiment confirmed the presence of metabolically generatedTGZS (retention time [RT] 4.92 min) which was absent in “0 h
1 y Letters 195 (2010) 135–141
TTttbiofSpstf
3
cowawwHardawf(t
f(Tibavcfl
odTwdcBtbowT
3
2d0aeR(
Fig. 3. (A) Percentage MTT cell viabilities of THLE-2 cells at varying concentrationsof TGZ and TGZS. Statistical differences were demonstrated between control and testcompounds (one-way ANOVA, P < 0.05) and between TGZ and TGZS (paired T-test,P < 0.05) at and above 10 �M concentration, respectively. (B) Percentage MTT cellviabilities of THLE-2 cells at varying concentrations of TGZ and TGZS co-incubatedwith 50 �M of QR. Statistical differences were observed between TGZ and TGZS interms of percentage cell viability at and above 50 �M (paired T-test, P < 0.05). (C)
38 S. Saha et al. / Toxicolog
GZ-THLE-2” and negative control samples but present in “24 hGZ-THLE-2” incubation. TGZS is the sulfo-conjugate of TGZ wherehe conjugation with bi-sulphate (–OSO3H) moiety occurred athe 6-hydroxyl (–OH) group of TGZ (Fig. 2A). TGZS was found toe absent when TGZ and QR (“24 h TGZ-QR-THLE-2”) were co-
ncubated with THLE-2 cells for 24 h (data not shown). The absencef TGZS in the “24 h TGZ-QR-THLE-2” sample indicated that sul-otransferase enzyme activity blocked by QR. As shown in Fig. 2B,IM at m/z 381.0 of the “24 h TGZ-QR-THLE-2” sample exhibited theresence of quercetin sulphate-1 (QRS1, RT 3.09 min) and quercetinulphate-2 (QRS2, RT 6.68 min). The presence of QRS1 and QRS2 inhe “24 h TGZ-QR-THLE-2” sample confirmed the competition of QRor the sulfotransferase enzyme.
.2. MTT viability assay
THLE-2 cells were treated with TGZ and TGZS at various con-entrations (1–125 �M) for 72 h. As shown in Fig. 3A, cell viabilityf the hepatocytes was dose-dependent for both TGZ and TGZShere an increase in dose of each test compound was associ-
ted with a consistent decline of cell viability. It was shown thathen each compound was incubated with THLE-2 cells, toxicityas observed from 1 �M with respect to control in both cases.owever, the observed toxic effect of TGZS was greater than TGZfter 1 �M. The THLE-2 cell viability was 50% at 75 �M of TGZ andemained unchanged up to 125 �M. On the other hand, cell viabilityecreased dramatically when dosed with TGZS and was below 10%t and above 50 �M of TGZS. The calculated EC50 for TGZS and TGZere 21.76 ± 5.34 and 41.12 ± 4.3 �M, respectively. Statistical dif-
erences were demonstrated between control and test compoundsone-way ANOVA, P < 0.05) and between TGZ and TGZS (paired T-est, P < 0.05) at and above 10 �M concentration, respectively.
To inhibit the biotransformation of TGZ to TGZS by sulfotrans-erase enzyme in THLE-2 cells, we co-incubated both compoundsTGZ and TGZS) with 50 �M of QR. Our results demonstrated thatGZS was more toxic than TGZ at and above 50 �M when co-ncubated with QR (Fig. 3B). Statistical differences were observedetween TGZ and TGZS in terms of percentage cell viability at andbove 50 �M (paired T-test, P < 0.05). It was also confirmed that theiability of THLE-2 cells was not affected by QR alone up to a con-entration of 125 �M (data not shown). QR is a naturally occurringavonoid having hepatoprotective activity (Gulati et al., 1995).
To determine if the increased cytotoxicity of TGZS was linked toxidative stress, we co-incubated both treatments with the antioxi-ant, TA (Patrick et al., 2005). The results further demonstrated thatGZS was more toxic than TGZ above 25 �M when the compoundsere co-incubated with TA at 20 �M level (Fig. 3C). Statisticalifferences were observed between TGZS and TGZ in terms of per-entage cell viability at and above 25 �M (paired T-test, P < 0.05).y comparing Fig. 3A and C, it was clear that TGZS produced lessoxicity in the presence of TA as compared to TGZS alone. The via-ilities of THLE-2 cells were 20 and 50% in the absence and presencef TA at 25 �M of TGZS, respectively. As an experimental control,e confirmed that the viability of THLE-2 cells was not affected by
A alone up to a concentration of 100 �M (data not shown).
.3. GSH depletion assay
Determination of GSH–NEM adducts was performed in THLE-cells after 72 h incubation with TGZ and TGZS to measure any
epletion of GSH. A linear regression performed over a range of
.001–10 �M yielded a correlation coefficient (r2) of 0.9986. Theccuracy of the assay was found to be within 80–108%. The recov-ries of the GSH–NEM complex in QC samples were 60–80%. TheT for GSH–NEM complex was 3.13 min (Fig. 4A). GSH depletion�M) per �g of protein was measured at varying concentrations
Percentage MTT cell viabilities of THLE-2 cells at varying concentrations of TGZ andTGZS co-incubated with 20 �M of TA. Statistical differences were observed betweenTGZ and TGZS in terms of percentage cell viability at and above 25 �M (paired T-test,P < 0.05).
of TGZ and TGZS after 72 h (Fig. 4B). As shown in Fig. 4B, TGZSdepleted more GSH than TGZ after 50 �M. Statistical differenceswere demonstrated between control and test compounds (one-wayANOVA, P < 0.05) and between TGZ and TGZS (paired T-test, P < 0.05)at and above 25 �M, respectively.
3.4. Protein carbonyl (PC) assay
Determination of PC content was performed in THLE-2 cells after48 h incubation with TGZ and TGZS as a surrogate for reactive-oxygen-mediated (ROS) protein oxidation. For each compound, PCcontent (�M) per �g of protein of THLE-2 cells was increased atall concentrations greater than10 �M with respect to control. As
S. Saha et al. / Toxicology Letters 195 (2010) 135–141 139
F .13 min). (B) GSH depletion in THLE-2 cells at varying concentrations of TGZ and TGZS.S e-way ANOVA, P < 0.05) and between TGZ and TGZS (paired T-test, P < 0.05) at and above2
sba
3
hisca3tdccdt
4
coutti
(tTi
ig. 4. (A) Representative LC/MS/MS chromatogram of GSH–NEM complex (RT, 3tatistical differences were demonstrated between control and test compounds (on5 �M, respectively.
hown in Fig. 5A and B, statistical differences were demonstratedetween control and test compounds (one-way ANOVA, P < 0.05)nd between TGZ and TGZS (paired T-test, P < 0.05) at 50 �M.
.5. Stability profile of TGZ and TGZS using BEGM
To evaluate the contribution of the sulfo-conjugate to directepatotoxicity and mitigate the possible confounding factors com-
ng from parental TGZ through hydrolysis of TGZS, we performedtability assessment of TGZS at 1.0, 50.0 and 150.0 �M in our cellulture conditions. The stability study demonstrated that TGZS atbove mention concentrations were stable up to 72 h in BEGM at7 ◦C (Fig. 3A, supplementary data sheet). The levels of TGZS amonghe different time points (0, 24, 48 and 72 h) were not significantlyifferent. Separately, another experiment was also performed toheck the stability of parent TGZ at 1.0, 50.0 and 150.0 �M in cellulture conditions. The levels of TGZ were not also significantlyifferent among the different time points (Fig. 3B, supplemen-ary data sheet).
. Discussion
TGZ was withdrawn from the market in 2000 due to its idiosyn-ratic liver toxicity (Chojkier, 2005). While several mechanismsf toxicity had been proposed for TGZ (Smith, 2003), it remainsnclear if the sulfo-conjugated metabolite (TGZS) of TGZ has directoxic effect on human liver cells. In the current project, the aim waso ascertain and establish the role of TGZS in causing direct toxicityn human hepatocytes.
To achieve this goal, TGZS was synthesized where 6-hydroxyl–OH) group of TGZ was replaced by bi-sulfate (–OSO3H) group. Inhis study, the cytotoxicities of TGZS and TGZ were compared usingHLE-2 cells, normal human hepatocytes with metabolizing capac-ty. The MTT assay results confirmed that TGZS was more toxic in
Fig. 5. (A) Protein carbonyl assay of THLE-2 cells at varying concentrations of TGZand TGZS. (B) Protein carbonyl assay of THLE-2 cells at varying concentrations ofTGZ and TGZS with respect to control. Statistical differences were demonstratedbetween control and test compounds (one-way ANOVA, P < 0.05) and between TGZand TGZS (paired T-test, P < 0.05) at 50 �M.
1 y Lette
hnorTntmQa
ptTtwdcfrclttstbt7Twa
7aTeatwi
eKcilwcxstomBtoiaumh(tcd
40 S. Saha et al. / Toxicolog
uman hepatocytes than TGZ. This data suggested a novel mecha-ism of toxicity of TGZ where TGZS is a direct toxicant in humans. Inrder to elucidate the cytotoxic effects of TGZ and TGZS more accu-ately, MTT assay was further performed where the sulfonation ofGZ to TGZS in THLE-2 cells was inhibited using QR, a specific phe-ol sulfotransferase inhibitor (Virginia et al., 2004). QR preventedhe sulfonation of TGZ which was confirmed by LC/MS/MS experi-
ents. MTT data using co-incubation of our test compounds withR further supported that TGZS was more toxic than TGZ in thebsence of biotransformation of TGZ to TGZS during incubation.
Based on earlier reports that TGZ mediated hepatotoxicity inart through generating oxidative injury (Smith, 2003), we raisedhe question whether the observed cytotoxicity was related toGZS-induced oxidative stress in the biological system. To answerhis question, MTT assay was performed by co-incubating 20 �M TAith TGZ and TGZS at various concentrations in THLE-2 cells. Theata demonstrated that cell viability was improved when TGZS waso-incubated with TA. The partial salvation of human hepatocytesrom the toxic effects of TGZS by an antioxidant indicated indi-ectly that TGZS might have exerted oxidative stress on the THLE-2ells. In line with this observation, the GSH depletion assay estab-ished that TGZS depleted more GSH than TGZ, further underscoringhe oxidative stress potential of TGZS. To further characterize theargets of oxidative stress, we performed PC assay which demon-trated that TGZS generated greater oxidation on cellular proteinshan TGZ. Collectively, TGZS appeared to cause human liver toxicityy inducing oxidative stress. Finally, the stability study confirmedhat both TGZ and TGZS were stable in BEGM at 37 ◦C up to2 h. This finding was important to rule out the possibility ofGZS undergoing deconjugation to TGZ during incubation. Hence,e were able to attribute all observed toxicity effects to TGZS
ccurately.The human blood concentration of TGZ is 0.3–0.5 �g/mL or
–12 �M after oral administration (Loi et al., 1999). Our MTT, GSHnd PC assays demonstrated that TGZS was more toxic towardsHLE-2 cells at and above 10 �M concentration as compared to par-nt TGZ. As the pharmacological and in vitro concentrations of TGZnd TGZS are in the same dynamic order of magnitude, it suggestedhat the investigated in vitro concentrations of TGZS (1–125 �M)ere clinically relevant and associated with the human physiolog-
cal concentrations.This study yielded a surprising outcome in the light of some
arlier works. Using primary cultures of human hepatocytes,ostrubsky et al. (2000) proposed that TGZ was responsible forytotoxicity rather than its metabolite TGZS. It was reported thatnhibition of sulfation in these hepatocytes resulted in an accumu-ation of parent TGZ and cytotoxicity. Nonetheless, their findings
ere not confirmed using pure chemically synthesized TGZS. Sulfo-onjugation is an essential metabolic pathway for many drugs,enobiotics and endogenous substances. While the main role ofulfo-conjugation is to enhance the hydrophilicity of drugs and aidheir excretion from the body, it has been reported that sulfationf certain allylic alcohols and polycyclic aromatic hydrocarbonsay produce metabolites with increased toxicity (Chou et al., 1998;
anoglu, 2000; King et al., 2000). This paper reported for the firstime that TGZS, a sulfo-conjugate of TGZ, exerted direct toxic effectsn human hepatocytes, possibly via oxidative stress induction. Its noted that some phase II metabolites such as indoxyl sulphate,
sulfate metabolite of 3-hydroxyindole, causes renal failure vianknown mechanism (Banoglu and King, 2002). Similarly, the toxicechanism of TGZS is unknown. However, as both TGZ and TGZS
ad been shown to bind to similar biological receptor such as BSEPFunk et al., 2001), it is possible that these compounds may bindo other similar biological receptors leading to oxidative stress andell death. As TGZS is a major metabolite of TGZ in humans, theifferential extent of phase II sulfation of TGZ in diabetic patients
rs 195 (2010) 135–141
may be accountable for the idiosyncratic nature of TGZ-inducedhepatotoxicity.
In summary, our present study reinforced the concept that hepa-totoxicity of TGZ is possibly related to multiple mechanisms (Smith,2003). Sulfonation at the 6-hydroxyl group of the chroman ring ofTGZ is a potential cause of TGZ-induced hepatotoxicity. In addi-tion to bile salt accumulation via inhibition of BSEP, we confirmedthat TGZS is a direct human hepatotoxicant. Prospective studiesinvolving other normal human hepatocytes, collagen-sandwichedhepatocytes, human liver slices and animal models are needed toconfirm the interplay between the dual toxic mechanisms of TGZS.
Conflict of interest
The authors declare that there are no conflicts of interest.
Acknowledgements
This work was supported by the National University ofSingapore (NUS) grants R-148-050-088-101/133 and R-148-000-100-112 to E.C.Y.C. The UPLC system was kindly supported by NUSgrant R-279-000-249-646. S.S. is supported by the NUS graduatescholarship.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.toxlet.2010.03.010.
Banoglu, E., 2000. Current status of the cytosolic sulfotransferases in the metabolicactivation of promutagens and procarcinogens. Curr. Drug Metab. 1, 1–30.
Banoglu, E., King, R.S., 2002. Sulfation of indoxyl by human and rat aryl (phenol)sulfotransferases to form indoxyl sulfate. Eur. J. Drug Metab. Pharmacokinet. 27,135–140.
Camp, H.S., Li, O., Wise, S.C., Hong, Y.H., Frankowski, C.L., Shen, X., Vanbogelen, R., Leff,T., 2000. Differential activation of peroxisome proliferator-activated receptor-gamma by troglitazone and rosiglitazone. Diabetes 49, 539–547.
Chojkier, M., 2005. Troglitazone and liver injury: in search of answers. Hepatology41, 237–246.
Delzenne, N.M., Calderon, P.B., Taper, H.S., Roberfroid, M.B., 1992. Comparative hep-atotoxicity of cholic acid, deoxycholic acid and lithocholic acid in the rat: in vivoand in vitro studies. Toxicol. Lett. 61, 291–304.
Funk, C., Ponelle, C., Scheuermann, G., Pantze, M., 2001. Cholestatic potential oftroglitazone as a possible factor contributing to troglitazone-induced hepato-toxicity. In vivo and in vitro interaction at the canalicular bile salt export pump(Bsep) in the rat. Mol. Pharmacol 59, 627–635.
Gores, G.J., Miyoshi, H., Botla, R., Aguilar, H.I., Bronk, S.F., 1998. Induction ofmitochondrial permeability transition as a mechanism of liver injury duringcholestasis: a potential role for mitochondrial proteases. Biochim. Biophys. Acta1366, 167–175.
Gulati, R.K., Agarwal, S., Agrawal, S.S., 1995. Hepatoprotective studies on Phyllanthusemblica Linn. and quercetin. Indian J. Exp. Biol. 33, 261–268.
He, K., Woolf, T.F., Kindi, E.K., Fielder, A.E., Talaat, R.E., 2001. Troglitazone quinine for-mation catalyzed by human and rat CYP3A: an atypical CYP oxidation reaction.Biochem. Pharmacol. 62, 191–198.
Henry, R.R., 1997. Thiazolidinones. Endocrinol. Metab. Clin. North Am. 26, 553–573.Kawai, K., Kawasaki-Tokui, Y., Odaka, T., Tsuruta, F., Kazui, M., Iwabuchi, H., 1997.
Disposition and metabolism of the new oral antidiabetic drug troglitazone inrats, mice and dogs. Arzneim. Forsch. Drug Res. 104, 1507–1514.
King, R.S., Teitel, C.H., Kadlubar, F.F., 2000. In vitro bioactivation of N-hydroxy-2-amino-�-carboline. Carcinogenesis 21, 1347–1354.
Kostrubsky, V.E., Sinclair, J.F., Ramachandran, V., Venkataramanan, R., Wen, Y.H.,Kindt, E., Galchev, V., Rose, K., Sinz, M., Strom, S.C., 2000. The role of conjugation
oi, C.M., Randiniis, E.J., Vassos, A.B., Kazierad, D.J., Koup, J.R., Sedman, A.J., 1997.Lack of effect of type II diabetes on the pharmacokinetics of troglitazone in amultiple-dose study. J. Clin. Pharmacol. 37, 1114–1120.
oi, C.M., Alvey, C.W., Randinitis, E.J., Sedman, A.J., Koup, S.R., 1999. Steady-statepharmacokinetics and dose proportionality of troglitazone and its metabolites.J. Clin. Pharmacol. 39, 920–926.
arcsek, Z.L., Kocsis, Z., Szende, B., Tompa, A., 2007. Effect of formaldehyde andreservatrol on the viability of vero, HepG2 and MCF-7 cells. Cell Biol. Int. 31,1214–1219.
iller, V.A., Benedetti, F.M., Rigas, J.R., Verret, A.L., Pfister, D.G., Strarus, D., Kris,M.G., Crisp, M., Heyman, R., Loewen, G.R., Truglia, J.A., Warrell, R.P., 1997. Initialclinical trial of a selective retinoid X receptor ligand LGD1069. J. Clin. Oncol. 15,790–795.
ew, L.S., Chan, E.C.Y., 2008. Evaluation of BEH C18, BEH HILIC, and HSS T3 (C18)column chemistries for the UPLC–MS–MS analysis of glutathione, glutathione
disulfide, and ophthalmic acid in mouse liver and human plasma. J. Chromatogr.Sci. 46, 209–214.
arker, J.C., 2002. Troglitazone: the discovery and development of a novel therapy forthe treatment of Type 2 diabetes mellitus. Adv. Drug Deliver. Rev. 54, 1173–1197.
atrick, Y.M., Thomas, N., Jan, F., Elisabeth, S., Gerald, R., Luca, B., 2005. Comparativequantification of pharmacodynamic parameters of chiral compounds (RRR- vs.
rs 195 (2010) 135–141 141
all-rac-a tocopherol) by global gene expression profiling. J. Plant Physiol. 162,811–817.
Ramachandran, V., Kostrubsky, V.E., Komoroski, B.J., Zhang, S., Dorko, K., Esplen,J.E., 1999. Troglitazone increases cytochorme P-450 3A protein and activity inprimary cultures of human hepatocytes. Drug Metab. Dispos. 27, 1194–1199.
Reznick, A.Z., Packer, L., 1994. Oxidative damage to proteins: spectrophotometricmethod for carbonyl assay. Methods Enzymol. 233, 357–363.
Rizvi, N.A., Marshall, J.L., Dahut, W., Ness, E., Truglia, J.A., Loewen, G., Gill, G.M., Ulm,E.H., Geiser, R., Jaunakais, D., Hawkins, M.J., 1999. A phase I study of LGD1069 inadults with advanced cancer. Clin. Cancer Res. 5, 1658–1664.
Saha, S., New, L.S., Ho, H.K., Chui, W.K., Chan, E.C.Y., 2010. Investigation the role ofthe thiazolidinedione ring of troglitazone in inducing hepatotoxicity. Toxicol.Lett. 192, 141–149.
Sahi, J., Hamilton, G., Sinz, M., Barros, S., Huang, M., Lesko, M.J., 2000. Effect of trogli-tazone on cytochrome P450 enzymes in primary cultures of human and rathepatocytes. Xenobiotica 30, 273–284.
Sherman, S.I., Gopal, J., Haugen, B.R., Chiu, A.C., Whaley, K., Nowlakha, P., Duvic,M., 1999. Central hypothyroidism associated with retinoid X receptor-selectiveligands. E. Engl. J. Med. Chem. 340, 1075–1079.
Yamamoto, Y., Nakajima, M., Yamazaki, H., Yokoi, T., 2001. Cytotoxicity and apop-tosis produced by troglitazone in human hepatoma cells. Life Sci. 70, 471–482.
