Peptide-bond modified glutathione conjugate analogs modulate GSTπ function in GSH-conjugation, drug sensitivity and JNK signaling

Article abstract of DOI:10.1016/j.bcp.2005.11.003

Glutathione S-transferase π (GST, E.C.2.5.1.18) overexpression contributes to resistance of cancer cells towards cytostatic drugs. Furthermore, GSTπ is involved in the cellular stress response through inhibition of Jun N-terminal-kinase (JNK), a process t

Full text of DOI:10.1016/j.bcp.2005.11.003

b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/biochempharm
"
Peptide-bond modified glutathione conjugate analogs
modulate GSTp function in GSH-conjugation,
drug sensitivity and JNK signaling
*
Danny Burg, Joey Riepsaame, Chantal Pont, Gerard Mulder, Bob van de Water
Division of Toxicology, Leiden/Amsterdam Center for Drug Research, Leiden University,
P.O. Box 9502, 2300 RA Leiden, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
Glutathione S-transferase p (GST, E.C.2.5.1.18) overexpression contributes to resistance of
cancer cells towards cytostatic drugs. Furthermore, GSTp is involved in the cellular stress
response through inhibition of Jun N-terminal-kinase (JNK), a process that can be modulated
by GST inhibitors. GSH conjugates are potent GST inhibitors, but are sensitive towards g-
glutamyltranspeptidase (gGT)-mediated breakdown. In search for new peptidase stable GST
inhibitors we employed the following strategy: (1) selection of a suitable (GST inhibiting)
peptide-bond isostere from a series of previously synthesized gGT stabilized GSH-analogs.
(2) The use of this peptidomimetic strategy to prepare a GSTp selective inhibitor. Two gGT
stable GSH conjugate analogs inhibited human GSTs, although non-selectively. One of
these, a urethane-type peptide-bond is well accepted by GSTs and we selected this mod-
ification for the development of a gGT stable, GSTp selective inhibitor, UrPhg-Et2. This
compound displayed selectivity for GSTp compared to a and m class enzymes. Furthermore,
the inhibitor reversed GSTp-mediated drug resistance (MDR) in breast tumor cells. In
addition, short-term exposure of cells to UrPhg-Et2 led to GSTp oligomerization and JNK
activation, suggesting that it activates the JNK-cJun signaling module through GSTp dis-
sociation. Altogether, we show the successful use of peptidomimetic glutathione conjugate
analogs as GST inhibitors and MDR-modifiers. As many MDR related enzymes, such as
MRP1, glyoxalase 1 and DNA-pk are also inhibited by GSH conjugates, these peptidomimetic
compounds can be used as scaffolds for the development of multi-target MDR drugs.
# 2005 Elsevier Inc. All rights reserved.
Received 25 August 2005
Accepted 2 November 2005
Keywords:
Glutathione S-transferase
Peptidomimetic
Multidrug resistance
cJun N-terminal kinase
Glutathione conjugate
Cytostatics
1.
Introduction
MAPEG) GST isoenzymes [5,6], which have different, but
partially overlapping substrate specificity towards electro-
Glutathione S-transferases (GSTs, E.C.2.5.1.18) play a pivotal
role in the protection of cells from injury by toxic chemicals
[1–4]. GSTs conjugate xenobiotic or endogenously formed
electrophiles with the abundant tripeptide glutathione (g-L-
glutamyl-L-cysteinyl-glycine, GSH). The mammalian GST
family consists of eight distinct classes of cytosolic/mitochon-
drial and four classes of membrane-bound (microsomal,
philic substrates. Overexpression of a, m and p class cytosolic
GSTs has been associated with resistance of tumor tissues to
cytostatic drugs [7], although various studies have shown that
GST overexpression by itself is not sufficient to generate a
multidrug resistant phenotype. Increased efflux of the formed
GSH conjugates by overexpression of multidrug resistance
associated protein (MRP1) is often also required [8]. GST is
* Corresponding author. Tel.: +31 715276223.
E-mail address: b.water@lacdr.leidenuniv.nl (B. van de Water).
0006-2952/$ – see front matter # 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2005.11.003

b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
269
thought to protect by direct detoxification of electrophilic
2.
Materials and methods
species, or by inactivation of reactive intermediates formed
during treatment with cytotoxic agents [9,10]. Furthermore,
GSTs were shown to regulate stress-activated signal transduc-
tion pathways: cJun NH2-terminal kinase (JNK/SAPK), which
plays a crucial role in both cell-survival and death pathways, is
partly regulated by protein–protein interactions with GSTp
[10,11]. Similar functions in regulating stress-activated signal-
ing kinases have also been reported for other GST family
members. GSTm, for example, mediates the activity of Ask1
and MEKK1 [12,13], two upstream components of the JNK/
SAPK signaling pathway. GSTv, in addition, regulates ryano-
dine receptors, which are main regulators of calcium home-
ostasis [12]. As ryanodine receptor mediated release of Ca2+
from internal stores precedes many crucial cellular processes,
GSTv may also have a regulatory function.
2.1.
Materials
Recombinant human GST isoenzymes A1-1, M1-1 and P1-1
were purchased from PanVera Corporation (Madison, WI,
USA). Cisplatin, thiotepa and doxorubicin were obtained from
Sigma (St. Louis, MO, USA). GST assays were performed in 96-
well plates using a Perkin-Elmer HTS7000 BioAssay Reader
(Perkin-Elmer, Boston, MA, USA). Phospho-state specific anti-
bodies, were obtained from Cell Signaling (Beverly, MA, USA).
2.2.
Cell lines
VCREMS and MCF7 human mammary carcinoma cells, were
cultured in DMEM, supplemented with 10% (v/v) fetal calf
serum, 100 U/l penicillin and 100 mg/l streptomycin. MTLn3
rat mammary adenocarcinoma cells were grown in Mem-a,
supplemented with 5% (v/v) fetal calf serum. All cells were
cultured at 37 8C in a humidified 5% CO2 atmosphere.
Altogether, these findings show that GST inhibitors,
besides their ability to modulate GST-mediated GSH-conjuga-
tion, can also influence cell signaling processes [14]. The use of
such compounds is illustrated by TLK199, an isoenzyme
selective GSTp inhibitor that is able to increase the cytostatic
effect of anti-cancer agents [15]. In addition, TLK199 initiates a
signal transduction cascade by dissociating GSTp from JNK,
which subsequently phosphorylates its downstream target
cJun [10]. In vivo, this process leads to increased myelopro-
liferation, which is currently being explored in a clinical
setting. These studies emphasize the importance of GST
inhibitors as pharmacological manipulators of signaling
cascades.
2.3.
Synthesis of GSH conjugate analogs
g-Glutamyl-(S-benzyl)cysteinyl-phenylglycine (gGCP, identi-
cal to TLK199), was synthesized according Lyttle et al. [25]. The
synthesis of peptidomimetic GS-EA analogs was recently
described [21]. Cys(EA)–Gly was synthesized in accordance
with these published methods. Non-esterified, phenylglycine
containing, compounds were obtained by similar methods
described in our original article [21]; deviations from these
protocols are the use of phenylglycine-OtBu, instead of
glycine-OtBu and S-alkylation with benzylbromide, instead
of ethacrynic acid (see Fig. 3A). Synthesis of the diethylester
derivatives was performed using serine ethyl ester, phenyl-
glycine ethyl ester [25] and S-benzylcysteine as building blocks
(Fig. 3B). All compounds were purified by HPLC, using a
separations system (Separations Analytical Instruments,
Hendrik Ido Ambacht, The Netherlands) equipped with a
Supelcosil SPLC-18-DB semi-preparative column and UV-
detection (Waters 490 detector) at 254 nm. A linear gradient
(10–95%) acetonitrile in H2O/0.1% TFA was used to purify end
products. All tested compounds were >95% pure, as deter-
mined by HPLC (column: Alltech Platinum C18, Alltech, Breda,
The Netherlands, UV detection at 210/254 nm) and LC–MS
(Alltech Platinum C18 column, Perkin-Elmer Sciex API 165
mass spectrometer).
GSH conjugates are commonly good inhibitors of GSTs in
vitro [16–18]. Although the g-glutamyl-cysteine peptide-bond
renders them resistant to most a-peptidases, GSH conjugates
are rapidly degraded by g-glutamyl transpeptidase (gGT) [19].
Cleavage by this enzyme, which has particularly high activity
in the kidney, precludes the use of GSH conjugates in vivo [20].
Recently, we synthesized a series of novel peptidomimetic
GSH conjugate analogs (Fig. 1) [21], some of which were
stabilized towards gGT-mediated breakdown. These GSH-
isosteres were conjugated to ethacrynic acid [22,23], thus
enabling us to evaluate the effect of peptide-backbone
modification on GST inhibition. Initial studies indicated that
these compounds inhibit GST activity in the rat liver S100
fraction, a mixture of mainly a and m class GSTs [21]. As GSH
conjugates are also good MRP1 inhibitors, our series of GS-EA
analogs was also tested for MRP1 inhibition. We showed that
our new peptidomimetics modulate resistance towards
methotrexate in MRP1 overexpressing ovarian carcinoma
cells [24]. In the current report, we characterize the GST
inhibition by peptidomimetic GS-EA analogs on human GST
isoenzymes to find a peptide-bond mimetic that is compatible
with GST inhibition. We selected a urethane-type peptide-
bond isostere for the design of a novel GSTp inhibitor, UrPhg
and its cell permeable diethyl ester derivative, UrPhg-Et2. We
find that the novel compound increases sensitivity of mam-
mary carcinoma cells towards cytostatics. In addition, UrPhg-
2.4.
Inhibition of recombinant GST isoenzymes
GST inhibition assays were performed according to Habig et al.
[26], modified for a 96-wells plate-reader. Recombinant GSTs
(10–20 ng/ml) were incubated with or without inhibitor in the
presence of 1 mM GSH at 37 8C in 0.1 M potassium phosphate
buffer pH 6.5, containing 1 mM EDTA. For the determination of
IC50 values, six different inhibitor concentrations were tested
(range 1–500 mM, depending on inhibitor efficacy). The reac-
tion was initiated by addition of 1-chloro-2,4-dinitrobenzene
(CDNB) in ethanol (final concentration 1 mM, 2% ethanol in
assay mix), after which the formation of GS-DNB was
spectrophotometrically monitored at 340 nm. Reaction rates
Et2 initiates
a JNK-mediated signal transduction cascade,
leading to phosphorylation of the transcription factor cJun.
Altogether, the data indicate that peptide-bond modified GSH
conjugates can be used to modulate various cellular processes,
such as GSTp-mediated MDR and MAP-kinase signaling.

270
b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
Fig. 1 – Peptide-bond modified GSH conjugate analogs. Deviations from the GSH-backbone are indicated in bold.

b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
271
Table 1 – Inhibition of human GST isoenzymes by GS-EA
were corrected for the non-enzymatic conjugate formation.
mimics
IC50 values were obtained by non-linear regression of the
inhibition data. Experiments were performed three times,
using quadruple samples per data point. Kinetic parameters of
GSTA1-1, GSTM1-1 and GSTP1-1 inhibition by GS-EA analogs
were determined at various GSH concentrations (0.1–4 mM)
with constant CDNB (1 mM). Ki values were obtained by the
Km,app method described by Kakkar et al. [27,28]. From the [S]
versus V0 curves, apparent Km and Vmax values were obtained
by non-linear regression (Graphpad Prism, Graphpad Soft-
ware, San Diego, CA, USA). The Ki value was determined by
extrapolation of the linear Km/Vmax versus [inhibitor] graph.
Inhibitor
IC50 (mM)
GSTA1-1
GSTM1-1
GSTP1-1
GS-EA
2.3 Æ 0.1
>500
0.6 Æ 0.2
>500
4.5 Æ 0.6
>500
Cys(EA)–Gly
Sarc(EA)
MeCys(EA)
AmHex(EA)
Red(EA)
4.5 Æ 0.2
>500
7.6 Æ 0.7
53 Æ 8
4.8 Æ 0.4
>500
>500
61 Æ 5
390 Æ 20
45 Æ 2
37 Æ 3
8.5 Æ 1.4
12 Æ 2
Ur(EA)
6.7 Æ 0.8
28 Æ 4
Shown are inhibitor concentrations at which half-maximal
enzyme activity was obtained (IC50, mean Æ S.E.M., n = 4).
2.5.
Immunoblotting
Confluent monolayers of MTln3 cells in 6-well culture dishes
were exposed to UrPhg-Et2 in complete medium at 37 8C. Cells
were then washed, scraped in ice-cold PBS and subsequently
lysed by ultrasonication in cold TSE (10 mM Tris, 250 mM
sucrose, 1 mM EGTA, pH 7.4), containing protease inhibitors
(50 mM Na3VO4, 10 mg/ml leupeptin, 10 mg/ml pepstatin, 1 mM
phenylmethylsulfonylfluoride). For GSTp blotting, non-redu-
cing and non-denaturing conditions were employed to
maintain its original (homodimer) state. (Phospho)JNK, (phos-
pho)Erk, (phospho)P38 and (phospho)cJun blots were sub-
jected to standard Western blotting sample pre-treatment
protocols. Twenty micrograms of total cellular protein was
separated by SDS-polyacrylamide gel electrophoresis and
subsequently transferred to PVDF membrane (Millipore, Bill-
erica, MA, USA). Blots were blocked with 5% BSA in TBS-T
(0.5 M NaCl, 20 mM Tris–HCl, 0.05%, v/v Tween 20; pH 7.4) and
probed for the appropriate proteins. The secondary HRP-
coupled antibody was visualized with ECL reagent (Amersham
Pharmacia Biotech, Piscataway, NJ, USA).
3.
3.1.
Results
Inhibition of human GSTs
Many alterations to the GSH peptide backbone have been
made, mainly by substitution of the amino acids; modifica-
tions of the internal peptide-bonds have rarely been made [30].
We used peptide-bond surrogates to potentially improve
inhibition characteristics and peptidase stability. To compare
the inhibition of human GST isoenzymes, IC50 values were
obtained for the series of GS-EA analogs (Table 1). IC50 values
could not be determined for Cys(EA)–Gly as hardly any
inhibition was measured at 500 mM, indicating that the g-
Glu portion is critical for binding efficiency. N-Methylation of
the Cys–Gly amide is very well accepted by GSTA1-1 and
GSTP1-1. Interestingly, inhibition of GSTM1-1 is decreased by
this methylation. Contrary to methylation of the Cys–Gly-
bond, N-methylation of the g-glutamyl peptide-bond in
MeCys(EA) is not accepted by these enzymes. Apparently,
the increased steric bulk prevents this compound from
binding to the GST G-site. The increased flexibility and the
loss of H-bonds by omission of the Cys–Gly peptide-bond in
Amh(EA) prevent the compound from binding correctly. The
reduced isostere, Red(EA) has lost some of its activity on the
human GSTs, but is still able to interact with the enzymes. The
urethane derivative Ur(EA) is a fairly good inhibitor of GSTA1-
1, being four-fold less potent than GS-EA.
2.6.
Antineoplastic drug cytotoxicity assays
MCF7 and VCREMS cells were seeded at 4 Â 103 cells/well in 96-
wells polyethylene culture dishes. After overnight attach-
ment, culture medium was replaced by Hanks’ Balanced Salt
Solution (HBSS: 137 mM NaCl, 5 mM KCl, 0.8 mM MgSO4,
0.4 mM Na2PO4, 0.4 mM KH2PO4, 1.3 mM CaCl2, 4 mM NaHCO3,
25 mM HEPES, 5 mM D-glucose; pH 7.4), containing the GST
inhibitor and various concentrations of cytostatic agents (from
freshly prepared DMSO stock solutions, maximal DMSO
concentration in final incubation volume 0.3%, v/v). Experi-
ments were performed three separate times with each point
determined in quadruple. After 4 h at 37 8C, cells were washed
twice with PBS, and subsequently incubated for 72 h in
complete culture medium (in absence of inhibitor). The cells
were then rinsed once with PBS and lysed by repeated freeze–
thaw cycles in 200 ml water, followed by homogenization on a
rotary shaker. Cell proliferation was measured by DNA
content, using Hoechst-33258 staining according to Rago
et al. [29]. In short: 50 ml Hoechst-33258 (20 mg/ml in TNE;
10 mM Tris, 1 mM EDTA, 0.2 M NaCl, pH 7.4) was added to 50 ml
lysate. Stained DNA was measured by spectrofluorometry
(excitation: 360 nm; emission 465 nm). A calibration curve of
calf thymus DNA was used to determine total DNA quantities.
Growth curves were calculated using Graphpad/Prism.
We next investigated the inhibition characteristics of two
gGT stable compounds, Red(EA) and Ur(EA) in more detail. As
differences in inhibition are presumably a result of altered
binding in the GST G-site, we determined inhibitory constants
towards GSH. As shown by Lineweaver–Burk plots (Fig. 2), the
two compounds inhibit GSTM1-1 and GSTP1-1 competitively;
GSTA1-1 displays a mixed-type inhibition. These data suggest
that the inhibitors compete with GSH for binding the G-site.
Ur(EA) is
a more potent inhibitor of the human GST
isoenzymes than the reduced isostere, Red(EA). Both com-
pounds inhibited GSTM1-1 equally well, but Ki values (Table 2)
indicate a reduction in affinity compared to unmodified GSH.
GSTP1-1 required higher concentrations for inhibition.
Altogether, data indicate that the reduced and urethane
peptide-bond isosteres are well accepted by human a, m and p
class GSTs. Red(EA) and Ur(EA) can be regarded as lead
compounds for the development of novel, gGT stabilized GST
inhibitors. As a next step we, therefore, investigated whether

272
b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
Fig. 2 – Lineweaver–Burk representation of inhibition of human GSTs by Red(EA) and Ur(EA). Shown are reciprocal enzymatic
velocities (WS.E.M., V in mmol/min/mg protein) vs. 1/[GSH] (mMS1) of GSTA1-1 (A), GSTM1-1 (B) and GSTP1-1 (C) without
inhibitor (- - -), or with Red(EA) (—) or Ur(EA) (– – –). Symbols: (*) no inhibitor; (&) Red(EA) 10 mM; (~) Red(EA) 50 mM; (&) Ur(EA)
12.5 mM; (~) Ur(EA) 25 mM. Insets represent a magnified view of the graph.
the urethane peptide-bond isostere could be used in the design
phenylglycine moiety that interacts with a hydrophobic
of a more selective GSTp inhibitor.
pocket near the GSTp GSH binding site [31]. TLK199 owes its
selectivity to the absence of this lipophilic cavity in other GST
classes. To gain selectivity for GSTp we utilized this moiety,
together with TLK199’s benzyl thioether, to obtain selective
peptidomimetic inhibitors. For the synthesis of the non-
esterified phenylglycine-containing urethane analog, UrPhg,
shown in Fig. 3A, we essentially employed methods analogous
to the synthesis of Ur(EA) [21]. In short, acylation of
BocSerOtBu with bis(4-nitrophenyl)carbonate yields an o-
nitrophenyl building block, which is condensed with
Cys(Acm)–PhgOtBu. After removal of the Acm group, the
benzyl thioether function can be attached at the last synthesis
step, which is useful for screening various thioether moieties.
During synthesis, we observed racemization of the phenyl-
glycine moiety, as was also described for TLK199 [32]; the L-
and D-phenylglycine containing products were separated by
semi-preparative HPLC. Overall purity of the D-phenylglycine
product, the major peak eluting later from the column, was
approximately 90%. The L-phenylglycine product could not be
3.2.
Synthesis and evaluation of a phenylglycine
containing GST inhibitor
To date, the most potent competitive, isoenzyme selective
GSTp inhibitor is TLK199. This compound contains
a
Table 2 – Ki values for inhibition of human GST
isoenzymes by GSH conjugate analogs
Inhibitor
Ki values (mM)
GSTA1-1
GSTM1-1
GSTP1-1
Red(EA)
Ur(EA)
16 Æ 3
4 Æ 2
6 Æ 3
3 Æ 1
22 Æ 6
16 Æ 7
3 Æ 2
UrP-Bzl
29 Æ 4
16 Æ 2
Ki values (mean Æ S.E.M., n = 3) were obtained by non-linear
regression of Michaelis–Menten graphs.

b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
273
Fig. 3 – (A) Synthesis of UrPhg: (I) bis(4-nitrophenyl)carbonate, diisopropylethylamine (DiPEA), dimethylformamide (DMF); (II)
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), dichloromethane (DCM); (III) dioxane, 80 8C; (IV) I2, methanol (MeOH); (V)
trifluoroacetic acid (TFA)/H2O (98:2, v/v); (VI) tri-n-butylphosphine, n-propanol/H2O (3:1, v/v), benzylbromide,
sodiumbicarbonate, EtOH/H2O. (B) Synthesis of UrPhg-Et2: (I) di-tert-butyldicarbonate (Boc2O), triethylamine; (II) bis(4-
nitrophenyl)carbonate, DiPEA, DMF (III) dicyclohexylcarbodiimide (DCC), hydroxybenzotriazole (HOBt), DCM; (IV) TFA/DCM
(3:7); (V) dioxane, 80 8C; (VI) TFA/DCM (3:7).
obtained in acceptable purity. Initial inhibition experiments
indicated that the D-phenylglycine product is a better GSTp
inhibitor than its L-diastereomer (not shown); we only use the
D-product for our inhibition studies. UrPhg competitively
inhibits GSTA1-1, GSTM1-1 and GSTP1-1. Ki values (Table 2)
indicate that inclusion of the phenylglycine moiety indeed
provides selectivity for the p class enzyme. UrPhg is a fairly
potent GST inhibitor, displaying 4–10 fold selectivity for
GSTP1-1 over the m and a class enzymes. Although potent,
UrPhg is a weaker GSTp inhibitor than parent compound
TLK199, for which literature data and our own inhibition data
show a submicromolar (<1 mM) Ki [33].
numerous by-products. Thus, we employed a more direct
method, shown in Fig. 3B, to obtain UrPhg-Et2. Starting from D-
phenylglycine-ethyl ester, S-benzyl cysteine and serine-ethyl
ester, we constructed the diester derivative according to our
published methods [21]. The end-product was purified in good
yield without significant racemization, due to the use of Boc-
protected cysteine instead of Fmoc. We used the diethyl ester
derivative to evaluate the effects of GSTp inhibition on GSTp-
mediated drug resistance in breast tumor cells and on the
MAPK signaling pathways.
3.3.
UrPhg-Et2 induces GSTp oligomerization and
For cell-based assays, the charged carboxylate residues of
UrPhg have to be shielded by esters, as this dramatically
improves their cellular uptake [34]. Initial attempts to obtain
the diethyl ester derivative of UrPhg, UrPhg-Et2, using SOCl or
TMS-Cl in dry ethanol failed due to the generation of
MAPkinase signaling
GSTp inhibitors can dissociate the GSTp/JNK complex [10]. In
line with these studies, we investigated whether UrPhg-Et2
induces GSTp oligomerization and activation of JNK signaling

274
b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
Fig. 4 – UrPhg-Et2 induces GSTp oligomerization and JNK activation: (A) MTLn3 cells were exposed to 50 mM UrPhg-Et2 (a) or
25 mM gGCP (b) for indicated times. GSTp oligomerization state (upper panel) was determined by Western blotting under
non-denaturing conditions, JNK phosphorylation was analyzed by standard (denaturating) Western blotting. (B) MTLn3
cells were exposed for 4 h to indicated concentrations (in mM) of UrPhg-Et2, after which cell lysates were used for Western
blot analysis. (C) MTLn3 cells were exposed to increasing concentrations of UrPhg-Et2 for 4 h, after which lysates were
analyzed for cJun phosphorylation by Western blot. (D) Densitometric analysis of three separate (phospho)cJun Western
blot experiments. Shown are mean W S.E.M. of the p-cJun/cJun ratio, symbol ‘*’ indicates significant increase compared to
the control (Student t-test, d.f. = 4, p < 0.05)).
in MTLn3 cells. Under non-denaturating conditions, GSTp is
predominantly present as non-covalently linked homodimers;
the 46 kDa dimer was therefore the main band (Fig. 4A); this
band also includes covalent GSTp dimers formed by inter-
Table 3 – Modulation of the cytostatic effect of thiotepa,
cisplatin and doxorubicin by UrPhg-Et2
subunit disulfide bonding of GSTp monomers (Mw 23 kDa).
Additional bands at 21 and 37 kDa were detected, presumably
IC50 (mM)
corresponding to post-translationally modified GSTp [35].
After UrPhg-Et2 exposure, bands at 92 kDa and higher appear,
comprised of four or more GSTp subunits. After 16 h, GSTp
oligomerization is lost. In concordance with GSTp oligomer-
ization, UrPhg-Et2 exposed MTLn3 cells also display a strong
increase in JNK phosphorylation. Phospho-JNK levels return to
base-levels 16 h after exposure. Altogether this shows that
UrPhg-Et2 is able to induce GSTp oligomerization and JNK
activation, albeit transiently.
[UrPhg-Et2]
= 0
[UrPhg-Et2]
= 25
[UrPhg-Et2]
= 50
MCF7
Thiotepa
164 Æ 26
34 Æ 4
7 Æ 3
152 Æ 11
17 Æ 2
77 Æ 8
8 Æ 1
Cisplatin
Doxorubicin
5 Æ 0.8
2 Æ 0.5
VCREMS
Thiotepa
Cisplatin
Doxorubicin
1380 Æ 170
115 Æ 16
11 Æ 1
766 Æ 59
93 Æ 16
7 Æ 0.6
223 Æ 29
23 Æ 4
To assess the effects of short-term exposure to UrPhg-Et2,
we investigated phosphorylation of JNK, Erk and p38 by
immunoblot. Fig. 4B clearly shows that GSTp inhibition
concentration-dependently induces JNK phosphorylation.
4 Æ 0.6
Shown are IC50 values (mM, ÆS.E.M., n = 3).

b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
275
Fig. 5 – UrPhg-Et2 modulates thiotepa cytotoxicity in breast cancer cell lines. Shown are representative dose–response
curves (mean W S.E.M., n = 3) showing inhibition of MCF7 and VCREMS cell growth by thiotepa in the presence or absence of
0 (&), 25 mM (~) or 50 mM (!) UrPhg-Et2. Cells were exposed to thiotepa/UrPhg-Et2 during 4 h and cell proliferation was
determined after 72 h.
This is rather unexpected: although dissociation of mono-
meric GSTp facilitates the kinase function of JNK, its
phosphorylation must be caused by activation of SEK1 or
factors even further upstream in the JNK signaling module.
Activation of MAPK signaling seems restricted to the JNK
pathway; in contrast to JNK, the other MAPK family members
p38 and Erk are not phosporylated in response to UrPhg-Et2.
As shown, GSTp inhibition initiates JNK-mediated signal-
ing. The penultimate event in this cascade is JNK-induced
phosphorylation of transcription factors, such as cJun and
ATF2, which will subsequently regulate transcription of target
genes through binding to their respective promoters. As
shown in Fig. 4C and D, cJun phoshorylation increases during a
short-term exposure to UrPhg-Et2.
cytostatic effect of several anticancer agents. In addition, it
induces phosphorylation of JNK and its downstream target
cJun.
Glutathione conjugation, catalyzed by GST, is a major event
in the sequestration of electrophiles. GSH conjugates are then
readily excreted by members of the MRP family of efflux-
pumps. Both GST and MRP are frequently overexpressed in
multidrug resistant tumors, acting either alone or in concert
[8]. GSH conjugates are potent inhibitors of both these GSH
dependent enzymes and, by disturbing GSTp–JNK interac-
tions, can also influence gene transcription. A major drawback
of GSH conjugates is their inherit lability towards gGT, which
limits their use in vivo. To address this problem, we employed
peptide-bond surrogates to fortify the g-Glu–Cys amide. We
recently reported the inhibition of MRP1 by these compounds
and demonstrated the ability of Ur(EA) to abolish MRP1-
mediated resistance to methotrexate in ovarian carcinoma
cells [24]. For our present study, the effects of peptide
backbone alterations on GST inhibition was evaluated with
a panel of recombinant human GSTs. The peptidomimetic
GSH-analogs were conjugated to ethacrynic acid (EA), as GS-
EA is a potent inhibitor these GSTs. Molecular modeling
studies showed that the EA moiety binds to the hydrophobic
binding site (H-site) of GSTs in a similar fashion for all GS-EA
mimics (not shown); differences in binding efficacy, therefore,
reflects altered affinity for the peptide portion of the
molecules. The shape and electronic configuration of the
GSH binding site (G-site) is relatively conserved among the
different GST isoenzymes; GSH is bound in strikingly similar
fashion in different isoenzymes. This is surprising because
there are only few similarities in the amino acid residues that
form the G-site [37]. For GSH, the main contributors to G-site
docking are the terminal glycine carboxyl and the zwitterionic
glutamic acid residue [38,39]. Furthermore, H-bonding inter-
actions with the GSH amides are essential for binding and
keeping GSH in an optimal bioactive conformation. As a result,
alterations in the GSH peptide backbone can dramatically alter
binding characteristics, as was seen for some of our GS-EA
mimics (for example, MeCys(EA)). A urethane-type amide
proved to be well accepted by GSTs. As Ur(EA) also potently
3.4.
Modulation of antineoplastic drug cytotoxicity
through GSTP1-1 inhibition
UrPhg competitively inhibits GSTp, induces JNK signaling and
may also modulate GSTp-mediated drug resistance. Hereto we
used MCF7 derived VCREMS human breast tumor cells, which
display increased resistance towards a variety of anticancer
agents [36]. VCREMS cells express high levels of GSTp and p-
glycoprotein, both of which are low in MCF7. As model
antineoplastic agents, we used thiotepa, cisplatin and doxor-
ubicin. We exposed cells for 4 h to the cytostatics in the
presence of UrPhg-OEt2 and determined cell proliferation after
72 h. IC50 values, shown in Table 3, were determined from
dose–response curves (as shown for thiotepa in Fig. 5) to assess
the effect of the GST inhibitor on antineoplastic drug
cytotoxicity. The multidrug resistant VCREMS cell line was
highly resistant to thiotepa and, to lesser extend, to cisplatin
and doxorubicin, when compared to parent MCF7 cells. Co-
incubation with UrPhg-OEt2 enhanced cytotoxicity of all three
cytostatics in both cell lines.
4.
Discussion
In the present study, we evaluated inhibitory potency of a
series of peptidomimetic GSH conjugate analogs towards
recombinant human GST isoenzymes. One of these peptido-
mimetic linkages was then used to obtain a GSTp selective
inhibitor. We showed that this compound promotes the
inhibits MRP1 [24], it is
a potential candidate for the
development of a multiple target MDR-modifier. We utilized
the urethane peptide-bond mimic of Ur(EA) together with a
phenylglycine moiety to obtain a GSTp selective inhibitor.

276
b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
Elevated expression of GSTp, either alone or coordinately
All together, our data show that GSH conjugate analogs that
contain peptide-bond isosteres can be used as isoenzyme
selective GST inhibitors. The improved peptidase stability may
be beneficial in vivo. Such compounds are of pharmaceutical
interest, as they can manipulate anticancer drug sensitivity
and cell biological processes. Using our peptide-bond mimetics,
selective inhibitors can be devised for GST, MRP1, glyoxalase,
DNA-dependent protein kinase (DNA-pk) and presumably also
for numerous other GSH-dependent enzymes.
with MRP1, generates resistance towards alkylating antineo-
plastic drugs, but also towards other drug classes. This is a
controversial observation; GSTp-mediated resistance has also
been reported for etoposide, doxorubicin and other cytostatics
which are not substrates for GSTp. For our experiments, we
used VCREMS cells, which display elevated amounts of GSTp
and p-glycoprotein as a result of mutagenesis with ethyl
methane sulfonate and subsequent selection with vincristin
[36]. The parent cell line MCF7 expresses low levels of GSTp.
VCREMS cells were originally characterized as resistant
towards doxorubicin, vincristin and etoposide, but not to
cisplatin. In our experiments, we actually did observe a
decreased sensitivity towards this platinum drug. In general,
coexposure with UrPhg-Et2 sensitizes the breast cancer cells to
thiotepa, cisplatin and doxorubicin. As this effect is also seen
in MCF7 cells, questions arise whether the increased drug
sensitivity is due to obstruction of GST enzyme activity. The
effects of UrPhg-Et2 on JNK-mediated signaling provides a
more likely explanation. Studies have shown that cisplatin
and doxorubicin induced cell killing is dependent on pro-
longed JNK activity; JNK inhibition delays the onset of cisplatin
induced apoptosis [40]. UrPhg-Et2 may induce the dissociation
of GSTp from JNK and facilitate its kinase function, thereby
improving the apoptotic response after cytostatic insult.
Regulation of JNK-mediated signaling through its interaction
with GSTp provides an explanation for many of the con-
troversies regarding GSTp expression and MDR [41]. The GST–
JNK interaction can be perturbed either by oxidative stress or
by binding of an inhibitor in the GSTp active site, which
induces a conformational change of the enzyme. We showed
that UrPhg-Et2 induces GSTp oligomerization, which may
represent a fraction of the enzyme that has dissociated from
JNK. In addition, GSTp cluster formation is accompanied by
JNK activation, which eventually leads to cJun phosphoryla-
tion. Unknown is why JNK becomes phosphorylated. Adler
et al. showed that phosphorylation of JNK by MKK4 is not
inhibited by GSTp, although forced expression of GSTp
reduces the phosphorylation level of MKK4 itself [10]. An
excess of GSTp may block the MKK4-JNK module, which can
be cleared by GSTp inhibitors. In this scenario, UrPhg-Et2 could
be able to induce JNK phosphorylation. Remarkably, UrPhg-Et2
does not promote p38 and Erk phosphorylation. HL60 cells,
chronically exposed to TLK199, show elevated levels of
phospho-Erk [42]. This was also seen in embryonic fibroblast
derived from GSTpÀ/À mice. Our data suggest that Erk
activation by GSTp inhibitors is not seen after a short-term
exposure, and may therefore reflect a secondary response to
the initial JNK activation. Indeed, UrPhg-Et2 induced JNK
activity gives rise to phosphorylation of cJun, which is also
responsible for the immunoproliferative effect of TLK199 in
vivo. We, therefore, speculate that UrPhg-Et2 might cause a
similar response in vivo, which may be clinically relevant.
The mechanism behind this event is not completely under-
stood, although it has been shown that deregulation of JAK/
STAT signaling is a crucial event for increased myeloprolifera-
tion in GSTpÀ/Àmice [43]. Gene expression profiling after
GSTp inhibitor exposure may provide more information
regarding the consequences of GSTp inhibition on (tumor)
cell biology.
Acknowledgements
This work was financially supported by the Dutch Cancer
Society, grant RUL2002-2741. We thank Dr. Roland Wolff
(Ninewells Hospital and Medical School, Dundee, Scotland) for
providing VCREMS cells.
r e f e r e n c e s
[1] Armstrong RN. Structure, catalytic mechanism, and
evolution of the glutathione transferases. Chem Res Toxicol
1997;10:2–18.
[2] Tsuchida S, Sato K. Glutathione transferases and cancer.
Crit Rev Biochem Mol Biol 1992;27:337–84.
[3] Whalen R, Boyer TD. Human glutathione S-transferases.
Semin Liver Dis 1998;18:345–58.
[4] Mannervik B, Danielson UH. Glutathione transferases-
structure and catalytic activity. CRC Crit Rev Biochem
1988;23:283–337.
[5] Hayes JD, Pulford DJ. The glutathione S-transferase
supergene family: regulation of GST and the contribution of
the isoenzymes to cancer chemoprotection and drug
resistance. Crit Rev Biochem Mol Biol 1995;30:445–600.
[6] Board PG, Coggan M, Chelvanayagam G, Easteal S, Jermiin
LS, Schulte GK, et al. Identification, characterization, and
crystal structure of the Omega class glutathione
transferases. J Biol Chem 2000;275:24798–806.
[7] Tew KD. Glutathione-associated enzymes in anticancer
drug resistance. Cancer Res 1994;54:4313–20.
[8] Morrow CS, Smitherman PK, Townsend AJ. Combined
expression of multidrug resistance protein (MRP) and
glutathione S-transferase P1-1 (GSTP1-1) in MCF7 cells and
high level resistance to the cytotoxicities of ethacrynic acid
but not oxazaphosphorines or cisplatin. Biochem
Pharmacol 1998;56:1013–21.
[9] Satoh K. Weak electrophile selective characteristics of the
rat preneoplastic marker enzyme glutathione S-transferase
P-form, GST-P (7-7): a theory of linear free energy
relationships for evaluation of the active site
hydrophobicity of isoenzymes. Carcinogenesis
1998;19:1665–71.
[10] Adler V, Yin Z, Fuchs SY, Benezra M, Rosario L, Tew KD,
et al. Regulation of JNK signaling by GSTp. EMBO J
1999;18:1321–34.
[11] Wang T, Arifoglu P, Ronai Z, Tew KD. Glutathione S-
transferase P1-1 (GSTP1-1) Inhibits c-Jun N-terminal kinase
(JNK1) signaling through interaction with the C terminus. J
Biol Chem 2001;276:20999–1003.
[12] Dulhunty A, Gage P, Curtis S, Chelvanayagam G, Board P.
The glutathione transferase structural family includes
a nuclear chloride channel and a ryanodine receptor
calcium release channel modulator. J Biol Chem
2001;276:3319–23.

b i o c h e m i c a l p h a r m a c o l o g y 7 1 ( 2 0 0 6 ) 2 6 8 – 2 7 7
277
[13] Ryoo K, Huh SH, Lee YH, Yoon KW, Cho SG, Choi EJ.
parameters and inhibition mechanism. J Pharmacol Exp
Ther 2000;293:861–9.
[29] Rago R, Mitchen J, Wilding G. DNA fluorometric assay in
96-well tissue culture plates using Hoechst 33258 after
cell lysis by freezing in distilled water. Anal Biochem
1990;191:31–4.
[30] Lucente G, Luisi G, Pinnen F. Design and synthesis
of glutathione analogues. Farmaco 1998;53:721–35.
[31] Koehler RT, Villar HO, Bauer KE, Higgins DL. Ligand-based
protein alignment and isozyme specificity of glutathione
S-transferase inhibitors. Proteins 1997;28:202–16.
[32] Keppler D, Leier I, Jedlitschky G, Konig J. ATP-dependent
transport of glutathione S-conjugates by the multidrug
resistance protein MRP1 and its apical isoform MRP2. Chem
Biol Interact 1998;111–112:153–61.
Negative regulation of MEKK1-induced signaling by
glutathione S-transferase mu. J Biol Chem 2004;279:
43589–94.
[14] Burg D, Mulder GJ. Glutathione conjugates and their
synthetic derivatives as inhibitors of glutathione-
dependent enzymes involved in cancer and drug
resistance. Drug Metab Rev 2002;34:821–63.
[15] Morgan AS, Ciaccio PJ, Tew KD, Kauvar LM. Isozyme-
specific glutathione S-transferase inhibitors potentiate
drug sensitivity in cultured human tumor cell lines.
Cancer Chemother Pharmacol 1996;37:363–70.
[16] Chen WJ, deSmidt PC, Armstrong RN. Stereoselective
product inhibition of glutathione S-transferase. Biochem
Biophys Res Commun 1986;141:892–7.
[17] Ong LK, Clark AG. Inhibition of rat liver glutathione S-
transferases by glutathione conjugates and corresponding
L-cysteines and mercapturic acids. Biochem Pharmacol
1986;35:651–4.
[18] Clark AG, Debnam P. Inhibition of glutathione S-
transferases from rat liver by S-nitroso-L-glutathione.
Biochem Pharmacol 1988;37:3199–201.
[19] Hanigan MH. Gamma-glutamyl transpeptidase, a
glutathionase: its expression and function in carcino-
genesis. Chem Biol Interact 1998;111–112:333–42.
[20] Ouwerkerk-Mahadevan S, Mulder GJ. Inhibition of
glutathione conjugation in the rat in vivo by analogues
of glutathione conjugates. Chem Biol Interact 1998;111–
112:163–76.
[21] Burg D, Filippov DV, Hermanns R, van der Marel GA, van
Boom JH, Mulder GJ. Peptidomimetic glutathione analogues
as novel gGT stable GST inhibitors. Bioorg Med Chem
2002;10:195–205.
[22] Ploemen JH, van Ommen B, van Bladeren PJ. Inhibition of
rat and human glutathione S-transferase isoenzymes by
ethacrynic acid and its glutathione conjugate. Biochem
Pharmacol 1990;40:1631–5.
[23] Awasthi S, Srivastava SK, Ahmad F, Ahmad H, Ansari GA.
Interactions of glutathione S-transferase-pi with
ethacrynic acid and its glutathione conjugate. Biochim
Biophys Acta 1993;1164:173–8.
[24] Burg D, Wielinga P, Zelcer N, Saeki T, Mulder GJ, Borst P.
Inhibition of the multidrug resistance protein 1 (MRP1) by
peptidomimetic glutathione-conjugate analogs. Mol
Pharmacol 2002;62:1160–6.
[25] Lyttle MH, Hocker MD, Hui HC, Caldwell CG, Aaron DT,
Engqvist GA, et al. Isozyme-specific glutathione-S-
transferase inhibitors: design and synthesis. J Med Chem
1994;37:189–94.
[33] Flatgaard JE, Bauer KE, Kauvar LM. Isozyme specificity
of novel glutathione-S-transferase inhibitors. Cancer
Chemother Pharmacol 1993;33:63–70.
[34] Ouwerkerk-Mahadevan S, Tirona RG, Ripping RA, Ploemen
JH, van Bladeren PJ, Pang KS, et al. Inhibition of glutathione
conjugation by glutathione analogues in the perfused rat
liver. Effect of esterification on the potency of gamma-L-
glutamyl-alpha-(D-2-aminoadipyl)-N-2-heptylamine. Drug
Metab Dispos 1997;25:1137–43.
[35] Shen H, Tsuchida S, Tamai K, Sato K. Identification of
cysteine residues involved in disulfide formation in the
inactivation of glutathione transferase P-form by hydrogen
peroxide. Arch Biochem Biophys 1993;300:137–41.
[36] Whelan RD, Waring CJ, Wolf CR, Hayes JD, Hosking LK,
Hill BT. Over-expression of p-glycoprotein and glutathione
S-transferase pi in MCF-7 cells selected for vincristine
resistance in vitro. Int J Cancer 1992;52:241–6.
[37] Wilce MC, Parker MW. Structure and function of
glutathione S-transferases. Biochim Biophys Acta
1994;1205:1–18.
[38] Oakley AJ, LoBello M, Battistoni A, Ricci G, Rossjohn J, Villar
HO, et al. The structures of human glutathione transferase
P1-1 in complex with glutathione and various inhibitors at
high resolution. J Mol Biol 1997;274:84–100.
[39] Oakley AJ, Rossjohn J, LoBello M, Caccuri AM, Federici G,
Parker MW. The three-dimensional structure of the human
Pi class glutathione transferase P1-1 in complex with the
inhibitor ethacrynic acid and its glutathione conjugate.
Biochemistry 1997;36:576–85.
[40] Krilleke D, Ucur E, Pulte D, Schulze-Osthoff K, Debatin KM,
Herr I. Inhibition of JNK signaling diminishes early but not
late cellular stress-induced apoptosis. Int J Cancer
2003;20(107):520–7.
[41] Tew KD, Ronai Z. GST function in drug and stress response.
Drug Resist Updat 1999;2:143–7.
[26] Habig WH, Pabst MJ, Jakoby WB. Glutathione S-
transferases-The first enzymatic step in mercapturic acid
formation. J Biol Chem 1974;249:7130–9.
[27] Kakkar T, Boxenbaum H, Mayersohn M. Estimation of
Ki in a competitive enzyme-inhibition model: comparisons
among three methods of data analysis. Drug Metab Dispos
1999;27:756–62.
[42] Ruscoe JE, Rosario LA, Wang T, Gate L, Arifoglu P,
Wolf CR, et al. Pharmacologic or genetic manipulation of
glutathione-S-transferase pi influences cell proliferation
pathways. J Pharmacol Exp Ther 2001;298:339–45.
[43] Gate L, Majumdar RS, Lunk A, Tew KD. Increased
myeloproliferation in glutathione S-transferase pi-deficient
mice is associated with a deregulation of JNK and Janus
kinase/STAT pathways. J Biol Chem 2004;279:8608–16.
[28] Kakkar T, Pak Y, Mayersohn M. Evaluation of a minimal
experimental design for determination of enzyme kinetic