Potent and fully noncompetitive peptidomimetic inhibitor of multidrug resistance P-glycoprotein

Article abstract of DOI:10.1021/jm100839w

N^α-Boc-l-Asp(OBn)-l-Lys(Z)-OtBu (reversin 121, 1), an inhibitor of the P-gp ABC transporter, was used to conceive compounds inhibiting the drug efflux occurring through the Hoechst 33342 and daunorubicin transport sites of P-gp, respectively H and R sites. Replacement of the aspartyl residue by trans-4-hydroxy-l-proline (4(R)Hyp) gave compounds 11 and 15 characterized by half-maximal inhibitory concentrations (IC₅₀) of 0.6 and 0.2 μM, which are 2-and 7-fold lower than that of the parent molecule. The difference in IC₅₀ between 11 and 15 rests on the carbonyl group of the peptidyl bond, reduced in 15. Those compounds are rather specific of P-gp, having no or limited activity on MRP1 and BCRP. 15 displayed no marked cytotoxicity up to 10-fold its IC₅₀. Importantly, 15 equally inhibited the Hoechst 33342 and daunorubicin effluxes through a typical noncompetitive inhibition mechanism, suggesting its binding to a site different from the H and R drug-transport sites.

Full text of DOI:10.1021/jm100839w

6720 J. Med. Chem. 2010, 53, 6720–6729
DOI: 10.1021/jm100839w
Potent and Fully Noncompetitive Peptidomimetic Inhibitor of Multidrug Resistance P-Glycoprotein
†
§
ꢀ
Charles Dumontet, Attilio Di Pietro, Joelle Paris, and Pierre Falson*
€
ꢀ §
Ophelie Arnaud,†,^ Ali Koubeissi,‡,^ Laurent Ettouati,*,‡ Raphael Terreux, Ghina Alame, Catherine Grenot,
†
‡
,†
€
†Laboratoire des Proteꢀines de Reꢀsistance aux Agents Chimiotheꢀrapeutiques, Equipe Labelliseꢀe Ligue 2009, Institut de Biologie et Chimie des
‡
ꢀ
Proteꢀines, UMR 5086 CNRS, Universiteꢀ Lyon 1, IFR 128 BioSciences Gerland Lyon-Sud, F-69367 Lyon, France, EA 3741 Ecosysteꢁmes et
Moleꢀcules Bioactives, Institut des Sciences Pharmaceutiques et Biologiques, Universiteꢀde Lyon, UniversiteꢀClaude Bernard Lyon 1, F-69373 Lyon,
France, §Universiteꢀ de Lyon, Universiteꢀ Claude Bernard Lyon 1, INSERM, U863, F-69373 Lyon, France, and Laboratoire de Cytologie
ꢀ
Analytique, Faculte de Medecine, Universite de Lyon, Universite Claude Bernard Lyon 1, INSERM U590, F-69373 Lyon, France.
ꢀ
ꢀ
ꢀ
^ Co-first-authors.
Received July 6, 2010
NR-Boc-L-Asp(OBn)-L-Lys(Z)-OtBu (reversin 121, 1), an inhibitor of the P-gp ABC transporter, was
used to conceive compounds inhibiting the drug efflux occurring through the Hoechst 33342 and
daunorubicin transport sites of P-gp, respectively H and R sites. Replacement of the aspartyl residue by
trans-4-hydroxy-L-proline (4(R)Hyp) gave compounds 11 and 15 characterized by half-maximal
inhibitory concentrations (IC50) of 0.6 and 0.2 μM, which are 2- and 7-fold lower than that of the
parent molecule. The difference in IC50 between 11 and 15 rests on the carbonyl group of the peptidyl
bond, reduced in 15. Those compounds are rather specific of P-gp, having no or limited activity on
MRP1 and BCRP. 15 displayed no marked cytotoxicity up to 10-fold its IC50. Importantly, 15 equally
inhibited the Hoechst 33342 and daunorubicin effluxes through a typical noncompetitive inhibition
mechanism, suggesting its binding to a site different from the H and R drug-transport sites.
Introduction
pocket, as early modeled in silico by Garrigos and Orlowski,13
and recently experimentally observed by Chang and collea-
gues by resolving the 3D structure of the mouse P-gp homo-
ABCa (ATP-binding cassette) transporters constitute a key
family1 in the field of public health. Proteins of this family are
involvedinseveralgenetic diseases2 andare responsible forthe
cellular multidrug resistance phenotype encountered during
chemotherapeutic treatments against cancer and viral diseases.3-5
In such a context, they represent a serious threat, since cancer cells
overexpress them to reduce drug concentration below its
cytotoxic threshold. Three ABC transporters are mainly res-
ponsible for this phenomenon: P-glycoprotein (P-gp/ABCB1),
which was the first identified protein involved in drug resis-
tance6 and remains the major agent involved in this phenome-
non; the multidrug resistance protein (MRP1/ABCC1), in-
volved in cells where P-gp is not overexpressed;7 the breast
cancer resistance protein (BCRP/ABCG2).8,9 Drug resistance
driven by these membrane proteins is emphasized first by their
intrinsic polyspecificity, with these pumps transporting a wide
spectrum of molecules, and second by common structural
characteristics of their drug-binding sites.
3
logue14 which displays a volume of 6000 A . Among these
˚
sites, two drug-binding regions have been functionally identi-
fied: one site called “R” from its capacity to bind rhodamine
123 and anthracyclines, such as doxorubicin or daunorubicin,
and a second one called “H”, which binds Hoechst 33342,
colchicine, and quercetin.15 Shapiro and Ling observed that
these drug-binding sites are distinct and are characterized by a
positive cooperativity, although they can partially overlap, since
rhodamine 123 above 2 μM quantitatively reduces Hoechst
33342 binding.15
Numerous multidrug-resistance reversing agents have been
described since the discovery of verapamil.16,17 Among them,
peptides are quite scarce. Valspodar/SDZ-PSC-833, a cyclos-
porin A derivative, is the best representative of compounds
that have ultimately reached phase III clinical trials but were
stopped because of pharmacokinetic interactions with antic-
ancer drugs and/or lack of specificity.18 Two hexacyclic pep-
tides, QZ59-SSS and QZ59-RRR, were recently cocrystallized
with the mouse P-gp.14 Besides these relatively large mole-
Four sites have been described for the binding of drugs
and modulators within P-gp10,11 (see also the recent review of
Colabufo and colleagues12). This suggests a large binding
cules, a series of short peptide P-gp inhibitors called reversins
19,20
~
was first described by Seprodi and colleagues.
They
*To whom correspondence should be addressed. For L.E.: phone,
þ33 (0)47 877 7082; fax, þ33 (0)47 877 7082; e-mail, laurent.ettouati@
univ-lyon1.fr. For P.F.: phone, þ33 (0)43 765 2916; fax, þ33 (0)47 272
2604; e-mail, p.falson@ibcp.fr.
a Abbrevations: ABC, ATP-binding cassette; P-gp/ABCB1, P-glyco-
protein; MRP1/ABCC1, multidrug resistance protein 1; BCRP/
ABCG2, breast cancer resistance protein; LDA, lithium diisopropyla-
mide; THF, tetrahydrofuran; Z, benzyloxycarbonyl; 3D-QSAR, 3D
quantitative structure-activity relationship; CoMSIA, comparative
molecular similarity indices analysis; LOO, leave one out; Bn, benzyl.
consist of di- and tripeptide derivatives sharing common phy-
sicochemical and structural features, such as bulky aromatic
and/or alkyl groups. Among them, 1 (reversin 121) is an
aspartyllysine (Asp-Lys) dipeptide derivative displaying good
affinity and specificity for P-gp.21
The present work describes a novel class of P-gp ligands, of
which the most potent, 15, behaves as a noncompetitive
inhibitor for both the R and H drug-transport sites of P-gp,
r
pubs.acs.org/jmc
Published on Web 08/23/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18 6721
Table 1. First Series of Reversin Derivatives and P-Gp Modulation Activity
NIH3T3 transfected cells
K562/R7
% inh effa
% inh effa
IC50
μM
,
resistant cells
% inh effa (10 μM)
compd
R1
X
Y
R2
R3
(10 μM)
(2 μM)
1
-CH2COOBn
-CH2COOBn
-CH2COOBn
CdO
CdO
NH
CH2
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ Bn
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
(S)-(CH2)4-NHZ tBu
75.8 ( 1.6
35.7 ( 15.3
67.7 ( 7.0
63.7 ( 5.4
37.2 ( 13.7
97.8 ( 4.0
59.6 ( 14.7
61.8 ( 3.2
82.3 ( 1.9
49.2 ( 10.2
34 ( 10.6
49.1 ( 5.8
93.4 ( 13.3
106.5 ( 8.6
104.4 ( 10.4
71.2 ( 1.5
78.7 ( 8.3
1.41 ( 0.34
88.0 ( 0.6
30
17
2
ψ(CO-CH2)-Gly
-(CH2)2COOBn
-(CH2)2COOtBu
-(CH2)3COOBn
-(CH2)2OCOBn
CdO
CdO
CdO
CdO
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
3
7
37.0 ( 9.0
65.5 ( 1.3
80.4 ( 4.5
4
5
-(CH2)2OCOcHex CdO
-(CH2)3OCOcHex CdO
8
35.7 ( 15.2
6
-(CH2)2OBn
CdO
CdO
CH2
13
16
14
19
20
21
22
-CH2p-PhOtBu
-CH2p-PhOtBu
-CH2p-PhOtBu
-CH2p-PhOtBu
-CH2p-PhOtBu
-CH2p-PhOtBu
-CH2p-PhOtBu
CdO
CdO
H
H
H
H
H
tBu
tBu
tBu
tBu
tBu
38.0 ( 7.2
(RS)CHSO2Ph
1.55 ( 0.86
(RS)CHOH (RS)CHSO2Ph
ψ(E,CHdCH)
CH2
CH2
a The efficacy of inhibition of drug efflux mediated by P-gp (% inh eff) is indicated, estimated as described in the Experimental Section from
NIH3T3 mouse cells expressing specifically P-gp or human K562/R7 drug resistant cell line32 which mainly expresses P-gp, and for selected
compounds the same experiment was carried out with HEK 293 and BHK expressing BCRP and MRP1, respectively. The efficacy of inhibition was
initially measured at 10 μM of each compound and then at 2 μM for those displaying 80% efficacy at 10 μM. The concentration of half maximal
inhibition, IC50, was estimated for the most potent compounds as detailed in caption of Figure 1. All data are the mean of duplicate experiments
carried out twice.
which limits the possibility for such an inhibitor to be itself
transported by P-gp or another ABC transporter. We chose
1 as a starting template, modified the aspartic acid side
chain length and ester-protecting group, then replaced the
aspartic acid residue by side chain constrained derivatives,
and finally modified the peptide bond. We also evaluated
the inhibitory effect of tyrosylglycine dipeptide derivatives
closely related to 1, especially concerning the hydrophobic
protecting groups. Once generated, these derivatives were
assayed in vitro, looking for compounds displaying a better
inhibition efficacy than 1 and, among them, selecting those
behaving as noncompetitive inhibitors for the H and R
sites.
Hoffman and Tao.26 This strategy implies the condensation
of an amino acid derived β-keto enolate with an amino acid
derived 2-triflyloxy ester, giving rise to the corresponding
ketomethylene. It is stressed that triflyloxy ester synthesis
undergoes a configuration inversion. However, for the sake
of economy and development, L-lysine was used instead of its
enantiomer (Scheme 1). The aspartic acid derived β-keto
ester 25 was obtained in 32% yield by reaction of aspartic
acid imidazolyl ester with the allyl acetate derived enolate
generated by the reaction with LDA and allyl acetate in
THF. Unexpectedly, the allyl ester byproduct 26 was also
recovered in 21% yield. Concerning the synthesis of the
L-lysine-derived 2-triflyloxy ester 29, the Z-protected L-lysine-
derived R-hydroxy acid 27 was obtained in quantitative yield
by treatment of Z-protected L-lysine with sodium nitrite in
50% aqueous acetic acid.27 We chose a benzyl protection
group for the carboxylic acid of 27, as tert-butyl protection
would imply a lengthy procedure.28 Finally we prepared the
unstable triflyloxy ester 29 in 41% yield, according to the
method of Weber et al.29 Coupling of protected triflyloxy
ester 29 with aspartic acid derived β-keto ester 25 was carried
out in THF with sodium hydride as base for 3 h at room tem-
perature, and then the crude adduct was treated with Pd(0) to
finally obtain the ketomethylene derivative 30 in 26% overall
yield.30
Results and Discussion
Chemistry. The products described in this article and
presented in Tables 1-4 were synthesized by standard
procedures.22,23 Reference reversins 1 and NR-Boc-L-Glu-
(OBn)-L-Lys(Z)-OtBu (reversin 1092, 2), as well as other
dipeptide derivatives 3-14, were obtained by the mixed
anhydride method (Tables 1 and 2).24 Aminomethylene
derivatives 15 and 16 were synthesized using the classical
reductive amination strategy, starting from chiral pool-
derived aminoaldehydes (Tables 1 and 2).25 Synthesis of
ketomethylene derivative 17, bearing a Gly residue inserted
between Asp and Lys residues, was previously reported, as
well as the synthesis of β-ketosulfone 18,23 Tyr-Gly deriva-
tives 19-22,23 and derivatives of (2S)-[3-((1S),3-bis-benzy-
loxycarbonylpropylcarbamoyl)propionylamino]pentane-
dioic acid dibenzyl ester (reversin 213, 23) and 2422 (Tables 1,
3, and 4). To obtain further information on the peptide bond
effect on the activity of 1, we synthesized its ketomethylene
derivative 30 (Scheme 1) using a strategy developed by
Biological Results. The efficacy of the compounds to
prevent the daunorubicin, mitoxantrone, and Hoechst
33342 efflux carried out by P-gp was evaluated by flow
cytometry. The amount of a fluorescent anticancer drug
substrate remaining in the cell was quantified in MDR1-
transfected NIH3T3/P-gp cell line expressing P-gp, and a
P-gp-positive drug-selected human leukemic K562/R7 resis-
tant cell line. The former cellular model has the advantage to
selectively express P-gp while the latter one is a biological

6722 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
Arnaud et al.
Table 2. Second Series of Reversin Derivatives and P-gp Modulation Activity
NIH3T3 P-gp transfected cells
K562/R7
resistant cells
HEK293 BCRP
transfected cells
BHK MRP1
transfected cells
compd
R1
X
% inh effa (10 μM) % inh effa (2 μM)
IC50, μM
% inh effa (10 μM) % inh effa (10 μM) % inh effa (10 μM)
9
COBn
Bn
CdO
CH2
68.2 ( 10.2
106.9 ( 23.0
105.7 ( 13.4
99.7 ( 11.0
76.6 ( 10.1
88.2 ( 5.1
77.6 ( 1.5
94.9 ( 0.4
83.5 ( 3.0
79.8 ( 1.2
10.0 ( 7.5
42.3 ( 0.8
11.2 ( 4.6
16.6 ( 3.4
18.8 ( 10.0
15
10
11
12
0.22 ( 0.03
1.68 ( 0.28
0.73 ( 0.20
1.13 ( 0.34
9.0 ( 3.5
7.0 ( 2.0
COcHex CdO
Bn CdO
CH2cHex CdO
120.5 ( 35.3
82.2 ( 10.8
a The efficacy of inhibition of drug efflux mediated by P-gp (% inh eff) is indicated, estimated as described in the Experimental Section from
NIH3T3 mouse cells expressing specifically P-gp or human K562/R7 drug resistant cell line32 which mainly expresses P-gp, and for selected
compounds the same experiment was carried out with HEK 293 and BHK expressing BCRP and MRP1, respectively. The efficacy of inhibition was
initially measured at 10 μM of each compound and then at 2 μM for those displaying 80% efficacy at 10 μM. The concentration of half maximal
inhibition, IC50, was estimated for the most potent compounds as detailed in caption of Figure 1. All data are the mean of duplicate experiments
carried out twice.
Table 3. Third Series of Reversin Derivatives and P-gp Modulation
Activity
Table 4. Derivative 18 and P-gp Modulation Activity
NIH3T3 P-gp transfected cells
% inh effa(10 μM)
compd
NIH3T3 P-gp transfected cells
% inh effa(10 μM)
18
49.1 ( 8.8
compd
R1
Bn
cHex
a The efficacy of inhibition of drug efflux mediated by P-gp (% inh
eff) is indicated, estimated as described in the Experimental Section
from NIH3T3 mouse cells expressing specifically P-gp or human
K562/R7 drug resistant cell line32 which mainly expresses P-gp, and
for selected compounds the same experiment was carried out with
HEK 293 and BHK expressing BCRP and MRP1, respectively. The
efficacy of inhibition was initially measured at 10 μM of each
compound and then at 2 μM for those displaying 80% efficacy at
10 μM. The concentration of half maximal inhibition, IC50, was
estimated for the most potent compounds as detailed in caption of
Figure 1. All data are the mean of duplicate experiments carried out
twice.
23
24
26.6 ( 7.7
29.3 ( 12.3
a The efficacy of inhibition of drug efflux mediated by P-gp (% inh
eff) is indicated, estimated as described in the Experimental Section
from NIH3T3 mouse cells expressing specifically P-gp or human
K562/R7 drug resistant cell line32 which mainly expresses P-gp, and
for selected compounds the same experiment was carried out with
HEK 293 and BHK expressing BCRP and MRP1, respectively. The
efficacy of inhibition was initially measured at 10 μM of each
compound and then at 2 μM for those displaying 80% efficacy at
10 μM. The concentration of half maximal inhibition, IC50, was
estimated for the most potent compounds as detailed in caption of
Figure 1. All data are the mean of duplicate experiments carried out
twice.
of 1.41 μM, a value close to the 88% obtained when using
the drug-resistant cell lines. Few compounds displayed an
efficacy below 40%, such as 30, 3, 13 (Table 1), and 23, 24
(Table 3), whereas most of them such as 2, 17, 4, 5, 16, 21, 22
(Table 1), 12 (Table 2), and 18 (Table 4) displayed an efficacy
similar to that of 1. Finally, nine dipeptides were found to be
more efficient than 1: 7, 8, 14, 19, 20 (Table 1), and most of
the compounds of Table 2 (9, 15, 10, 11). Efficacies obtained
from drug-selected cells were in the same range (Tables 1 and
2), showing that those compounds remain efficient in a
system closer to a physiopathological context than trans-
fected cells.
Structure-Activity Relationships. Several modifications
had no positive effect or were negative, e.g., replacement of
the amide bond by a ketomethylene isostere in 30 or exten-
sion of the aspartate side chain in 2 and 7. Extending the
length of the aspartate side chain slightly increased the
efficacy of 7, which however remained poorly active at 2 μM.
model close to an in vivo chemoresistance phenotype. All
drugs were assayed at 10 μM in the NIH3T3 cells, and the
most efficient ones were then tested in K562/R7 cells.
Compounds displaying more than 80% efficacy were as-
sayed at 2 μM, and the half-maximal inhibition concentra-
tion, IC50, was determined for the most active ones. A typical
experiment is displayed in Supporting Information Figure SI
1 to detail the experimental procedure. As shown, values
obtained with NIH3T3/P-gp cells can be higher than those
obtained with control cells (see panel C, at 0.8 and 1 μM 15),
giving a higher level of variability than observed with the
K562/R7 resistant cells. Transfected cells were used to
compare all compounds, as they express a unique transpor-
ter. The reference compound 1 (Table 1), developed by
Sarkadi and colleagues,21 displayed 75% inhibition at 10 μM
in our transfected-cell system, corresponding to an IC50

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18 6723
Scheme 1. Synthesis of Ketomethylene Isostere 30a
a Reagents and conditions: (i) (a) 50% aqueous acetic acid, room temp; (b) NaNO2, H2O, room temp, quant yield; (ii) BnBr, Et3N, acetone, room
temp, 6 h, 40%; (iii) Tf2O, CH2Cl2, 2,6-lutidine, -78 ꢀC, 4 h, to room temp, 41%; (iv) (a) CDI; (b) CH3CO2allyl, LDA, THF, -78 ꢀC, 2 h, 25 32% and 26
21%; (v) (a) NaH, THF, -20 ꢀC, 20 min, to room temp, 3 h; (b) Pd(OAc)2, PPh3, HCOOH, Et3N, THF, room temp, 24 h, 26%.
In a similar manner, derivatives obtained from 2, in which
the protecting group of the Asp carboxyl side chain was
replaced by a tert-butyl in 3 or in which the ester was replaced
by a reversed ester in 4, 5, and 8 or by a benzyl ether in 6, were
not more efficient than 1.
In another set of modifications, we replaced the aspartyl
residue by a tyrosine. This single replacement reduced the
overall efficacy of resulting reversins 13 and 16 to 34%,
whereas a surprising increase to about 100% was obtained
by additionally replacing the lysine moiety by a (RS)CHSO2-
PhCH2 group. In this series, the most efficient derivative, 19,
led to an IC50 of 1.55 μM, similar to that of 1. A deep reduc-
tion in efficacy was also observed for lengthened derivatives
23 and 24 (Table 3).
Conversely, striking effects were also observed by intro-
ducing a higher level of constraint in template 2, e.g.,
replacing the N-terminal Glu residue by 4(R)Hyp residue,
resulting in the products listed in Table 2. Apart from 9,
compounds of that series displayed an inhibition efficacy at
10 μM close to 100%, leading to IC50 values of 1.68, 1.13,
0.73, and 0.22 μM for 10, 12, 11, and 15, respectively. As
shown in Table 2, some of these improvements were due to
the replacement of the ester function of the N-terminal
residue by either a benzyl ether function in 11 and 15 or
cyclohexylmethylene ether in 12. A significant improvement
was obtained by a single reduction of the peptide bond in 15,
compared to 11, allowing it to reach the highest affinity with
an IC50 of 0.22 μM. This value is nearly 1 order of magnitude
lower than that of template 1 (Figure 1).
linear correlation coefficient q2 of 0.85, indicating that 3D-
QSAR calculations were robust enough to give an accurate
prediction of the percent inhibition by a candidate molecule.
Isopotential volumes of hydrophobic, electrostatic, and
steric fields were drawn as displayed in Figure 2, panels C
and D. Analysis of the hydrophobic field (panel C) showed
that this character is mainly favored (pink grids) at the
vicinity of the phenyl moiety of the Z protecting group,
associated with a favorable steric hindrance (panel D, yellow
grids). This illustrates the importance of bulky nonpolar
protecting groups in this area. A small unfavorable hydro-
phobic region can be observed close to the Z carbamate
group (panel C, black grids), coupled with a sterically favor-
able region that might be ascribed to HB-acceptor interac-
tions with polar P-gp residues. By contrast, the region en-
compassing the N-terminal Boc carbamate group appears
hydrophobically favorable and sterically unfavorable (pink
and green grids, respectively). The same trend is observed for
the C-terminal tert-butyl region, suggesting that in both
cases the local docking pockets of these groups should be
hydrophobic and of small size. Region (I) is hydrophobically
unfavorable and sterically favorable (black and yellow grids,
respectively), suggesting that it should not contain large
hydrophobic groups, as those introduced in 23 and 24. As
displayed in panel D, an electrostatic field appears unfavor-
able (blue grids) at the vicinity of the backbone NH group
but favorable (red grids) in the background, which could be
ascribed to the positive effects observed with the sulfone-
containing compound 18. Such an electrostatic contribu-
tion, limited to the peptide backbone and associated with
the global hydrophobic character of the inhibitor, suggests
that the corresponding hydrophilic volume could be em-
bedded in a hydrophobic environment and that the inhi-
bitor could bind to a hydrophobic pocket. Region II, at
the level of the benzyl-protecting group (Bn), appears
hydrophobically unfavorable and sterically favorable
(black and yellow grids, respectively), which could illus-
trate the positive impact of ester bridge shift on the inhi-
bition activity.
A 3D quantitative structure-activity relationships (3D-
QSAR) analysis was then carried out on the reversin deriva-
tives. As detailed in the Experimental Section, models were
first energetically minimized and chains were then drawn in a
linear conformation. Since those molecules have a great
degree of freedom, a linear conformation was chosen for
all of them. They were first aligned on the peptide backbone,
and then alignment of each side chain was manually carried
out by rotating dihedral angles. Each model was then mini-
mized, setting the upper energy threshold to 20 kcal mol-1
.
3
The resulting alignment is displayed in Figure 2A. A data-
base was generated from all models and treated by the
CoMSIA method to carry out 3D-QSAR studies (see Ex-
perimental Section for details). Values of inhibition efficacy
measured at 10 μM of each compound from Tables 1-4
were used for calculations with the LOO (leave one out)
cross-validation method. The plot of predicted vs observed
inhibitions values is shown in Figure 2B (close circles), with a
Specificity, Reversing Efficacy, and Toxicity of Peptides
Derivatives. The compounds listed in Table 2 produced
only low or even no inhibition of BCRP/ABCG2 and
MRP1/ABCC1 when tested at 10 μM on cells expressing
those ABC transporters. They thus appear rather specific
toward P-gp. This is useful for preserving the physiological
activity of ABC transporters not concerned by the drug
resistance.

6724 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
Arnaud et al.
We then characterized the inhibition properties of the
most active compound 15. As shown in Figure 3, the specific
expression of P-gp in NIH3T3 cells led to a resistance index
of 8.5 for daunorubicin (circles vs triangles) that remains
limited but significant enough to observe a further sensitiza-
tion effect. Further addition to this drug-resistant cell line of
1 μM 15, corresponding to 5-fold its IC50, restored its initial
sensitivity to the anticancer drug (diamonds), showing the
efficacy of 15 to reverse the drug resistance phenotype. At
this concentration, no marked cytotoxicity was observed as
shown in Figure 4. Indeed, viability of either cell expressing
P-gp (squares) or control ones (circles) was not significantly
altered up to 1 μM. A much pronounced toxicity was, how-
ever, observed above 2 μM, a concentration that is 10 times
higher than the corresponding IC50.
Inhibition Mechanism of 15. As mentioned above, the R
and H drug-transport sites were identified on P-gp, the
former involved in the binding of rhodamine 123 and
anthracyclines such as daunorubicin while the latter binds
Hoechst 33342, colchicine, and quercetin.15 We examined
the behavior of 15 toward these drug-transport sites, moni-
toring with flow cytometry the variation of intracellular
retention of daunorubicin and Hoechst 33342 in cells expres-
sing P-gp. As detailed in the Experimental Section, sub-
strates were used in the range of 0-20 μM for daunoru-
bicin and 0-10 μM for Hoechst 33342. 15 was added at fixed
concentrations of 0.11, 0.22, 0.33, and 1 μM. The resulting
retention rates of substrates are displayed in Figure 5 as a
direct function of substrate concentrations (panels A and C)
Figure 1. Reversion of P-gp-mediated mitoxantrone efflux by 1
and 15. The inhibition efficacy of peptide derivatives was assayed by
flow cytometry, quantifying the intracellular amount of mitoxan-
trone as described in the Experimental Section. Values were deter-
mined by using eq 1 and fitted with eq 2, as described in the
Experimental Section. Squares and circles correspond to 1 and 15,
respectively.
Figure 2. 3D-QSAR of the reversin-derived compounds CT. (A) Superimposed molecules of the series. Molecules were superimposed as
detailed in the text. The regions labeled I and II are detailed in the corresponding paragraph in the Results and Discussion section. (B) Predicted
versus experimental inhibition efficacy values. Calculated values (closed circles) were obtained with the CoMSIA model by the LOO method as
described in the Experimental Section, with a correlation coefficient q2cv of 0.85, a line intercept of 16.29, and a slope of 0.77. Calculated values
(open circles) were obtained with the CoMSIA model without validation method (calibration), as described in the Experimental Section, with a
linear fit correlation coefficient r2 of 0.99, a line intercept of 0.40, and a slope of 0.99. (C, D) Analyses of hydrophobic (C) and electrostatic and
steric (D) clusters. They were carried out as described in the Experimental Section and displayed is the structure of the most efficient compound
15, the remaining derivatives being shown as a ghost. Favorable and unfavorable characters were as follows: pink and black clusters,
respectively, for hydrophobicity, red and blue clusters for the electrostatic field and yellow and green for the steric field.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18 6725
Figure 4. Cytotoxicity of 15. The compound was added at the
indicated concentrations. Complete cell survival (100%) was esti-
mated as the number of cells in the absence of compound. Squares
and circles correspond to experiments carried out with NIH3T3/P-
gp and NIH3T3 cells, respectively.
Figure 3. P-gp-expressing cells chemosensitization by 15. Chemo-
sensitization was assayed and quantified as described in the Experi-
mental Section, using daunorubicin as anticancer drug at indicated
concentrations and with 15 added at 1 μM. Complete cell survival
(100%) was estimated as the number of cells in absence of anticancer
drug. Squares and circles correspond to control NIH3T3 cells that
do not express P-gp incubated with and without 15, respectively.
Triangles and diamonds correspond to the same conditions with
NIH3T3/P-gp cells.
solution of NR-Boc-Asp(OBn)-OH (404.5 mg, 1.25 mmol) in
dry THF (6 mL) under argon atmosphere. The resulting solu-
tion was stirred for 1 h at the same temperature and used for the
next step without further purification. Meanwhile, a solution of
lithium enolate of allyl acetate was made by addition of 2.5 M
BuLi solution in hexane (1.75 mL, 4.375 mmol) to a solution of
diisopropylamine (615 μL, 4.375 mmol) in dry THF (6 mL) at
0 ꢀC under argon atmosphere. The solution was stirred for
10 min at 0 ꢀC and then cooled to -78 ꢀC. A solution of allyl
acetate (472 μL, 4.375 mmol) in THF (3 mL) was then added
dropwise, and the resulting solution was stirred for 15 min at
-78 ꢀC. The above imidazole solution was added dropwise to
the pale-yellow solution of lithium enolate at -78 ꢀC under
argon atmosphere. The resulting mixture was stirred for 2 h and
then quenched at -78 ꢀC with 1 M HCl (20 mL) and extracted
with ethyl acetate (3 ꢀ 30 mL). The organic extracts were com-
bined, washed with saturated sodium bicarbonate solution (2 ꢀ
40 mL) and brine (40 mL), dried over anhydrous sodium sulfate,
filtered, and concentrated to provide the crude product as a
yellow oil which was purified by flash chromatography on silica
gel (petroleum ether/ethyl acetate 8:2) to afford 163 mg of 25
(32% yield) as a beige solid. Mp 48-51 ꢀC. IR (film, cm-1): 3371
(νNH), 2978 (νCH), 1733 (νCdO), 1456 (νjCdC), 1392 1367 (δCH3
t-Bu), 1167 (νO-C-C asym). 1H NMR (300 MHz, CDCl3, δ): 1.38
(s, -C(CH3)3 Boc, 9H), 2.74 (dd, J = 4.6 and 17.8 Hz, CHβ,
1H), 2.98 (dd, J = 4.8 and 17.3 Hz, CHβ, 1H), 3.59 (s, CH2-
CO2allyl, 2H), 4.53 (m, CHR, 1H), 4.55 (dd, J = 1.8 and 6.7 Hz,
CH2-CHdCH2, 2H), 5.05 (s, CH2Ph, 2H), 5.23 (m, CHdCH2,
2H), 5.57 (d, J = 9.0 Hz, NH, 1H), 5.83 (m, CHdCH2, 1H), 7.28
(m, ArH, 5H).
(2S)-tert-Butoxycarbonylaminosuccinic Acid 1-Allyl Ester 4-
Benzyl Ester (26). The product 26 is a byproduct issued from the
synthesis of β-keto allyl ester 25. This product (95 mg, 21%
yield) was isolated by flash chromatography on silica gel
(petroleum ether/ethyl acetate 8:2) as a yellow oil. IR (film,
cm-1): 3354 (νNH), 2935 (νCH), 1731 (νCdO), 1447 (νjCdC), 1391
1366 (δCH3 t-Bu), 1152 (νO-C-C asym). 1H NMR (300 MHz,
CDCl3, δ): 1.46 (s, -C(CH3)3 Boc, 9H), 2.89 (dd, J = 4.7 and
17.5 Hz, CHβ, 1H), 3.08 (dd, J = 5.0 and 16.7 Hz, CHβ, 1H),
4.60-4.68 (m, CHR and CH2-CHdCH2, 3H), 5.14 (s, CH2Ph,
2H), 5.28 (m, CHdCH2, 2H), 5.50 (d, J = 7.9 Hz, NH, 1H), 5.87
(m, CHdCH2, 1H), 7.36 (m, ArH, 5H).
6-Benzyloxycarbonylamino-(2S)-hydroxyhexanoic Acid (27).
Commercial H-Lys(Z)-OH (1.012 g, 3.614 mmol) was dissolved
in 50% aqueous acetic acid (120 mL). A solution of sodium
nitrite (2 g, 29 mmol) in water (20 mL) was then added at 0 ꢀC,
and the reaction mixture was stirred for 40 min at the same
temperature and then quenched with water (140 mL) and
extracted with ethyl acetate (2 ꢀ 140 mL). The organic extracts
or in Lineweaver-Burk double reciprocal plots31 (panels B
and D). As shown, the maximal retention rates appeared much
more affected than the corresponding KM. Fitting of the Line-
weaver-Burk plots tend to a common intercept on the negative
x-axis, [drug]-1. Table 5 summarizes the kinetic parameters
deduced from the direct plots by using the Michaelis and
Menten equation. As shown, the Michaelis constant of each
drug did not change while the corresponding maximal reten-
tion rates increased with the inhibitor concentration. All these
data suggest that 15 behaves as a noncompetitive inhibitor
toward both substrates. They allow to estimate KI values of
about 0.5 and 1.5 μM (Table 5), suggesting existence of a
second, noncompetitive binding site for 15 at high concentra-
tion. Further docking experiments on P-gp will be carried out
based on the mouse P-gp 3D structure14 to localize the putative
inhibitor binding sites.
Experimental Section
Chemistry. Solvents and reagents were of reagent grade and
used without further purification. Melting points were taken on
an Electrothermal 9200 apparatus or a Kofler bench and are
uncorrected. Specific rotation [R]D was recorded on Bellingham
þ Stanley Ltd. ADP 220, Jasco P-1010 or Roussel-Jouan 71
polarimeters. IR spectra were determined on Perkin-Elmer
Spectrum One FT-IR or Perkin-Elmer 1600 series FT-IR spec-
trometers (KBr film). 1H, 13C, and 19F NMR spectra were
measured on Bruker AC 200 (1H 200.11 MHz; 13C 50.29 MHz),
Bruker ALS 300 (1H 300.17 MHz), Bruker AM 300 (1H 300.17
MHz), or Bruker DRX 300 (13C 75.44 MHz; 19F 282.23 MHz)
spectrometers (residual protonated solvent signal as internal
reference). Elemental analysis was performed on all tested
compounds by the Service Central d’Analyse, Vernaison,
France, and the results, unless otherwise indicated in the Ex-
perimental Section, are within (0.4% of theoretical values.
Electrospray mass spectra were recorded on a Thermo-Finnigan
LCQ-Advantage in dichloromethane. Column chromatography
˚
was carried out with Merck silica gel 60 (40-63 A) or SDS silica
˚
gel 60 (35-70 A).
(3S)-tert-Butoxycarbonylamino-4-oxohexanedioic Acid 6-Al-
lyl Ester 1-Benzyl Ester (25). Carbonyldiimidazole (220 mg,
1.312 mmol) was added at room temperature to a stirred

6726 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
Arnaud et al.
Figure 5. Compound 15 is a noncompetitive inhibitor of H and R drug-transport sites. Experiments were carried out as described in the
Experimental Section and in Figure 1, using daunorubicin (A, B) and Hoechst 33342 (C, D) as transported drugs.15 15 was either omitted (0%
inhibition, circles) or added at 0.11, 0.22, 0.33, and 1 μM corresponding to 25% (bottom-up triangles), 50% (squares), 75% (diamonds), and
100% (bottom-down triangles) inhibition respectively, as deduced from Figure 2. The drug retention rate corresponding to the amount of drug
(expressed as arbitrary units of fluorescence, a.u.) remaining inside 105 cells after 1 h was plotted as a function of drug concentration (A, C) and
fitted with the Michaelis and Menten equation data. Small symbols in panel C correspond to the same series of experiments, carried out with
control NIH3T3 cells. Data are also displayed in Lineweaver-Burk double-reciprocal plots31 (B, D). Experiments were carried out in duplicate
in two independent series.
Table 5. Kinetic Inhibition Parameters of 15a
(20 mL), and triethylamine (600 μL, 4.27 mmol) was added.
Benzyl bromide (510 μL, 4.27 mmol) was then added dropwise,
and the reaction mixture was stirred for 6 h at room tempera-
ture. The solvent was evaporated, and the obtained residue was
dissolved in ethyl acetate, washed with saturated sodium bicar-
bonate solution and brine (twice each one), dried over anhy-
drous sodium sulfate, filtered, and concentrated. The yellow oily
residue was purified by flash chromatography on silica gel
(petroleum ether/ethyl acetate 6:4) to yield 516 mg of 28 (40%
yield) as a yellow oil. IR (film, cm-1): 3365 (νNH and νOH), 3066,
3034 (νjCH), 2937 (νCH), 1738 1699 (νCdO), 1587, 1538, 1455
(νjCdC), 1251 (νC-C-O asym), 1126 (νO-C-C asym). 1H NMR (300
MHz, CDCl3, δ): 1.27-1.89 (m, CH2β, CH2γ, CH2δ, 6H), 3.17
(q, J = 6.4 Hz, CH2N, 2H), 4.23 (m, CHR, 1H), 4.75 (br, NH,
1H), 5.10 (s, CH2Ph, 2H), 5.20 (AB, J = 12.4 Hz, CH2Ph, 2H),
7.36 (m, 2ArH, 10H).
[15], μM
Rmax
KM, μM
KI, μM
Daunorubicin Substrate
0
0.11
547 ( 42
16.7 ( 1.1
16.7 ( 1.2
16.4 ( 1.4
15.9 ( 1.7
16.3 ( 1.2
656 ( 46
860 ( 65
1230 ( 58
1808 ( 77
0.66 ( 0.01
0.60 ( 0.01
0.59 ( 0.04
1.43 ( 0.08
0.22
0.33
1.0
Hoechst 33342 Substrate
0
15491 ( 562
0.23 ( 0.06
0.24 ( 0.12
0.27 ( 0.09
0.25 ( 0.07
0.11
0.22
1.0
20485 ( 1199
26098 ( 1278
38349 ( 987
0.45 ( 0.04
0.54 ( 0.06
1.68 ( 0.08
a Maximal retention rates (Rmax) and Michaelis constants (KM) of
daunorubicin and Hoechst 33342, as well as inhibition constants (KI),
were estimated from Figure 5, panels A and C, as described in the Experi-
mental Section.
6-Benzyloxycarbonylamino-(2S)-trifluoromethanesulfonylox-
yhexanoic Acid Benzyl Ester (29). Alcohol 28 (302.7 mg, 0.816
mmol) was dissolved in CH2Cl2 (5 mL) and then cooled to -78 ꢀC
(acetone). 2,6-Lutidine (265 μL, 2.284 mmol) was added dropwise
followed by triflic anhydride (276 μL, 1.632 mmol). The resulting
solution was allowed to warm to room temperature and the
reaction mixture was stirred under argon for 4 h. An amount of
5 g of silica gel was added directly to the reaction before
evaporation of the solvent at room temperature under reduced
pressure. Purification by flash chromatography on silica gel
(petroleum ether/ethyl acetate 8:2) afforded 170 mg of 29 (41%
yield) as a brownish oil. IR (film, cm-1): 3349 (νNH), 3068, 3036
(νjCH), 2942 (νCH), 1760, 1715 (νCdO), 1587, 1456, 1417 (νjCdC),
1326 (νSO2 asym), 1246 (νC-C-O asym), 1144 (νSO2 sym), 698 (νCF3).
were combined, washed with 0.1 M HCl solution (200 mL),
dried over anhydrous sodium sulfate, filtered, and concentrated
to provide the crude product 27 (1 g, quantitative yield) as a
yellow oil which was dried under vacuum in order to eliminate
the remaining acetic acid and used without further purification.
IR (film, cm-1): 3454 3335 (νNH, νOH, and νCOOH), 2925 (νCH),
1683 (νCdO), 1586, 1539, 1456 (νjCdC), 1260 (νC-OH), 1127
1
(νO-C-C asym), 1102 (νC-N). H NMR (300 MHz, CDCl3, δ):
1.40-1.59 (m, CH2γ, CH2δ, 4H), 1.82 (m, CH2β, 2H), 3.22 (q, J =
6.4 Hz, CH2N, 2H), 4.28 (m, CHR, 1H), 4.94 (br, NH, 1H), 5.11 (s,
CH2Ph, 2H), 5.40-5.60 (br, COOH, 1H), 7.36 (s, ArH, 5H).
6-Benzyloxycarbonylamino-(2S)-hydroxyhexanoic Acid Ben-
zyl Ester (28). Acid 27 (1 g, 3.56 mmol) was dissolved in acetone
1H NMR (300 MHz, CDCl3, δ): 1.36-1.59 (m, CH2γ, CH2δ
4H), 2.03 (m, CH2β, 2H), 3.17 (q, J = 6.5 Hz, CH2N, 2H), 4.71
,

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18 6727
(br, NH, 1H), 5.11 (s, CH2Ph, 2H), 5.14 (m, CHR, 1H), 5.27 (AB,
J = 12.0 Hz, CH2Ph, 2H), 7.36 (m, 2ArH, 10H). 19F NMR (282
MHz, CDCl3, δ): -75 (CF3).
the same media without drug; cells were maintained on ice until
analysis by flow cytometry.
Flow Cytometry. The fluorescence of intracellular daunoru-
bicin or mitoxantrone was measured with a Facscan flow
cytometer (Becton Dickinson, Mountain View, CA). Daunor-
ubicin and mitoxantrone were excited at 488 nm, and the
fluorescence emission was monitored at 650 nm. The amount
of Hoechst 33342 was measured by a LSRII flow cytometer, at
350 nm excitation and 450 nm emission. The compound effi-
ciency was estimated by using eq 1,
Nr-Boc-(S)-Asp(OBn)-ψ(CO-CH2)-(S)-Lys(Z)-OBn (30). A
solution of 25 (119 mg, 0.294 mmol) in dry THF (2 mL) was
added dropwise to a stirred suspension of NaH (14 mg of 60% in
oil, 0.353 mmol) in dry THF (2 mL) at -20 ꢀC under argon.
The mixture was stirred for 20 min. Then the alkyl triflate 29
(162.7 mg, 0.323 mmol) in dry THF (1 mL) was added. The
resulting solution was allowed to warm to room temperature
and stirred for 3 h. After being quenched with 1 M HCl (5 mL),
the reaction mixture was then extracted with ethyl acetate (3 ꢀ
10 mL). The organic extracts were combined, washed with satu-
rated sodium bicarbonate solution (15 mL) and brine (15 mL),
dried over anhydrous sodium sulfate, passed through a short
pad of silica gel, and concentrated to provide a yellow oil.
Without further purification, this oil (223 mg, 0.294 mmol)
was dissolved in dry THF (2 mL) and added dropwise to a
stirred solution of palladium acetate (1.6 mg, 7.23.10-3 mmol),
PPh3 (3.9 mg, 0.0147 mmol), formic acid (22 μL, 0,588 mmol),
and Et3N (105 μL, 0.735 mmol) in dry THF (2 mL) at room tem-
perature under argon. The mixture was stirred overnight. After
the solvent was evaporated under reduced pressure, the oily
residue was purified by flash chromatography on silica gel
(petroleum ether/ethyl acetate 7:3) to afford 52 mg of 30 (26%
% efficacy ¼ 100 ꢀ ðFA - FBGÞ=ðFe - FBGÞ
ð1Þ
where FA and FBG (background) correspond to the intracel-
lular fluorescence of cells expressing P-gp incubated with (FA)
or without (FBG) the fluorescent transported drug, in the
presence of each tested compound, Fe corresponding to FA
measured on control cells that do not express P-gp. Assays were
performed in duplicate. Mean and standard error (SE) were
calculated from at least three independent experiments, and
curves were fitted with Sigmaplot using the eq 2:
% efficacy ¼ að1 - expð - b½inhibitor; μ MꢁÞÞ
a and b parameters being adjusted by the fit.
ð2Þ
Cytotoxicity Assays. Cells were seeded in 200 μL of growth
medium at a density of 2500 cells per well in a 96-well plate and
were allowed to establish for 24 h at 37 ꢀC in humidified 5%
CO2. The medium was then removed, replaced by 100 μL of
growth medium, completed with 100 μL of each tested com-
pound at increasing concentrations in 1% DMSO. Cells were
incubated for 72 h at 37 ꢀC in humidified 5% CO2. The cyto-
toxicity of each tested compounds was then evaluated with a
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
(MTT) colorimetric assay (Sigma-Aldrich).
Chemosensitization Assays. Experiments were carried as for
cytotoxicity assays, incubating control or P-gp-expressing cells
with increasing concentrations of anticancer drugs as indicated
and a concentration of the tested modulator corresponding to
5 times its IC50.
yield) as a brown oil. [R]25 -10.0 (c 0.19, CH2Cl2). IR (film,
D
cm-1): 3360 (νNH), 3034 (νjCH), 2933 (νCH), 1715 (νCdO), 1587,
1456 (νjCdC), 1367 (δCH3 t-Bu), 1251 (νC-C-O asym), 1165
(νO-C-C asym). 1H NMR (300 MHz, CDCl3, δ): 1.37 (s, C(CH3)3
Boc, 9H), 1.10-1.72 (m, 3CH2 Lys, 6H), 2.45-2.91 (m, CH2CO,
CH2CO2Bn, 4H), 3.05 (q, J = 6.7 Hz, CH2N, 2H), 4.38 (m,
CHR, 1H), 4.62 (m, CHR, 1H), 4.97-5.16 (m, 3CH2Ph, 6H), 5.45
(d, J = 8.3 Hz, NH, 1H), 7.22-7.29 (m, 3ArH, 15H). 13C NMR
(CDCl3, 75 MHz): 28.7 (3C), 24.4, 30.0, 31.7 (3C), 36.1 (C), 40.3
(C), 40.9 (C), 41.1 (C), 56.3 (C), 67.0, 67.1, 67.2 (3C), 80.7 (C),
128.5, 128.7, 128.8, 128.9, 129.0, 130.2 (15C), 135.7, 136.3, 137.0
(3C), 155.8, 156.7 (2C), 171.4, 175.1 (2C), 206.9 (C). MS-ESI
(m/z): 1371.2 [2M þ Na]þ, 697.3 [M þ Na]þ. HRMS-ESI (m/z):
[M þ Na]þ calculated for C38H46N2O9Na, 697.3101; found,
697.30991.
Kinetics. Kinetic parameters for P-gp inhibition by com-
pound 15 were determined by incubating 105 NIH3T3/P-gp
and NIH3T3 cells with various concentrations of daunorubicin
(5, 10, 15, and 20 μM) in the presence of four to five concentra-
tions (0, 0.11, 0.22, 0.33, and 1 μM) of 15. The residual intra-
cellular amount of drugs remaining inside cells after 1 h of
incubation with or without 15 was quantified by fluorescence
coupled to flow cytometry at each concentration of 15, with data
proportional to a drug retention rate. It was plotted as a
function of drug concentration (A, C) and fitted with the
Michaelis and Menten equation R = Rmax[D]/([D] þ KM) where
R is the drug retention rate, Rmax the maximal drug retention
rate, [D] the drug concentration, and KM the Michaelis constant
of drug efflux. The inhibition constant of 15, KI was deduced
from the equation Rmax0 = Rmax/(1 þ [I]/KI), where Rmax0 corre-
sponds to the maximal drug retention rate calculated at each
modulator concentration [I]. Each value is the mean of two
duplicate experiments.
Biology. Cell Cultures and Drug Efflux Assays. The inhibi-
tion of P-gp activity by the studied compounds was evaluated by
their ability to prevent the efflux of daunorubicin, mitoxan-
trone, and Hoechst 33342 (Sigma-Aldrich) in either the K562/
R7 drug-selected human cell line32 or the NIH3T3 G185 trans-
fected cell line (ATCC).33 They were also assayed on human
embryonic kidney (HEK) 293 cells transfected with BCRP/
ABCG234 and baby hamster kidney (BHK) 21 expressing
MRP1.35 K562/R7 cells (1 ꢀ 106 cells) were incubated for 1 h at
37 ꢀC in 1 mL of RPMI 1640 medium containing a final con-
centration of 10 μM daunorubicin, in the presence or absence of
10 μM reversin derivatives. After incubation, cells were washed
with ice-cold PBS and kept on ice until analysis by flow cytometry.
Cyclosporin A was used as a positive control (mean inhibitory
activity of 100%). NIH3T3 drug-sensitive parental cell line and its
derivative cell line transfected with human MDR1-G185,
NIH3T3/P-gp, derived from NIH Swiss mouse embryo cultures,
were grown in Dulbecco’s modified Eagle’s medium (PAA
laboratories) supplemented with 10% fetal bovine serum (PAA
laboratories) and 1% penicillin/streptomycin (PAA Laborato-
ries). The NIH3T3/P-gp growth medium was also supplemented
with 60 ng/mL colchicine. Cells were incubated at 37 ꢀC in
humidified 5% CO2. NIH3T3 and NIH3T3/P-gp were incubated
for 30 min at 37 ꢀC under 5% CO2 in the presence of 10 μM mitox-
antrone or various concentrations of either daunorubicin or
Hoechst 33342 (as indicated in Figures) as the P-gp transport
substrate, with or without reversin derivatives. Compounds were
solubilized in DMSO, the final concentration of which was limited
to 0.5%. After incubation, cells were washed with phosphate
buffer saline (PBS) (PAA Laboratories) and incubated for 1 h in
Bioinformatics. 3D-QSAR. Molecules were modeled with the
Sybyl molecular-modeling package using the force field and
partial charges of MMFF94.36-39 The dielectric constant was
set to 80 to simulate the solvent, and the electrostatic cut-off was
˚
set to 16 A. In CoMSIA, the steric, electrostatic, hydrophobic,
hydrogen bond donor, hydrogen bond acceptor, and donor-
acceptor similarity index fields were calculated, with a default
˚
attenuation factor of 0.3 in a grid box extended by adding 6 A to
the largest model, in all directions of the space. A common probe
˚
atom was used, with 1 A radius and charge, hydrophobicity, and
hydrogen bond properties of þ1. The indices were evaluated
according to the usual CoMSIA protocol with a 2.0 kcal mol-1
column filtering value. The predictive power of the models was
3

6728 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18
Arnaud et al.
evaluated by LOO cross-validation (SYBYL 6.9-7.3 versions;
Tripos Inc.) based on the cross-validated coefficient (q2), dis-
played in Figure 2B, finding an optimal component number of
10. The maximal error of prediction is about 22% of inhibition
with a mean error value of 5.2% ( 6.1.
(14) Aller, S. G.; Yu, J.; Ward, A.; Weng, Y.; Chittaboina, S.; Zhuo, R.;
Harrell, P. M.; Trinh, Y. T.; Zhang, Q.; Urbatsch, I. L.; Chang, G.
Structure of P-glycoprotein reveals a molecular basis for poly-
specific drug binding. Science 2009, 323, 1718–1722.
(15) Shapiro, A. B.; Ling, V. Positively cooperative sites for drug
transport by P-glycoprotein with distinct drug specificities. Eur.
J. Biochem. 1997, 250, 130–137.
(16) Avendano, C.; Menendez, J. C. Inhibitors of multidrug resis-
tance to antitumor agents (MDR). Curr. Med. Chem. 2002, 9,
159–193.
(17) Teodori, E.; Dei, S.; Martelli, C.; Scapecchi, S.; Gualtieri, F. The
functions and structure of ABC transporters: implications for the
design of new inhibitors of Pgp and MRP1 to control multidrug
resistance (MDR). Curr. Drug Targets 2006, 7, 893–909.
(18) Szakacs, G.; Paterson, J. K.; Ludwig, J. A.; Booth-Genthe, C.;
Gottesman, M. M. Targeting multidrug resistance in cancer. Nat.
Rev. Drug Discovery 2006, 5, 219–234.
ꢀ
Acknowledgment. We thank Lea Payen (Institut des Sciences
Pharmaceutiques et Biologiques, Lyon, France) for critical dis-
cussion and Philippe Lawton (Institut des Sciences Pharmaceu-
tiques et Biologiques, Lyon, France) for proofreading. This work
was supported by the Centre National de la Recherche Scienti-
fique (CNRS) and University of Lyon 1 (UMR5086), and the
French Ministry of Research (EA3741, Lyon 1). It was funded by
the National Research Agency (Grants ANR-06-BLAN_0420,
ANR-06-PCVI-0019-01, ANR piribio09_444706), the Associa-
tion pour la Recherche sur le Cancer (ARC), and the Ligue
Nationale Contre le Cancer (Labellisation 2009). The Ph.D.
€
~
ꢀ
ꢀ
ꢀ
(19) Seprodi, J.; Mezo, I.; Vadasz, Z.; Szabo, K.; Sarkadi, B.; Teplan, I.
Peptides Derivatives against Multidrug Resistance: Mayflower
Scientific Ltd.: Kingswinford, U.K., 1998.
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
(20) Vadasz, Z.; Szabo, K.; Sarkadi, B.; Teplan, I.; Mak, M.; Miklos, I.;
€
€
Gyorffy, E.; Seprodi, J. New Hydrophobic Peptide-like Com-
pounds Modulating MDR1-ATPase Activity. Proceedings of the
25th European Peptide Symposium; Bajusz, S., Hudecz, F., Eds.;
ꢀ
fellowship of Ophelie Arnaud was from the French Ministry of
Research.
ꢀ
ꢀ
Akademiai Kiado: Budapest, Hungary, 1999; pp 640-641.
(21) Sharom, F. J.; Yu, X.; Lu, P.; Liu, R.; Chu, J. W.; Szabo, K.;
Muller, M.; Hose, C. D.; Monks, A.; Varadi, A.; Seprodi, J.;
Sarkadi, B. Interactionof the P-glycoprotein multidrug transporter
(MDR1) with high affinity peptide chemosensitizers in isolated
membranes, reconstituted systems, and intact cells. Biochem.
Pharmacol. 1999, 58, 571–86.
Supporting Information Available: Synthetic procedures and
analytical data for compounds 1-24; example of estimation of
the inhibition efficacy by flow cytometry. This material is
available free of charge via the Internet at http://pubs.acs.org.
(22) Koubeissi, A.; Raad, I.; Ettouati, L.; Guilet, D.; Dumontet, C.;
Paris, J. Inhibition of P-glycoprotein-mediated multidrug efflux by
aminomethylene and ketomethylene analogs of reversins. Bioorg.
Med. Chem. Lett. 2006, 16, 5700–5703.
(23) Charrier, C.; Ettouati, L.; Paris, J. New application of the Julia
olefination for the synthesis of Tyr-Gly E-alkene and carba isostere
pseudopeptides. Tetrahedron Lett. 1999, 40, 5705–5707.
(24) Lai, M. Y. H.; Brimble, M. A.; Callis, D. J.; Harris, P. W. R.; Levi,
M. S.; Sieg, F. Synthesis and pharmacological evaluation of
glycine-modified analogues of the neuroprotective agent glycyl-
l-prolyl-l-glutamic acid (GPE). Bioorg. Med. Chem. 2005, 13, 533–
548.
(25) Martinez, J.; Bali, J. P.; Rodriguez, M.; Castro, B.; Magous, R.;
Laur, J.; Lignon, M. F. Synthesis and biological activities of some
pseudo-peptide analogs of tetragastrin: the importance of the
peptide backbone. J. Med. Chem. 1985, 28, 1874–1879.
(26) Hoffman, R. V.; Tao, J. Synthesis of Cbz-Protected Ketomethy-
lene Dipeptide Isosteres. In Peptidomimetics Protocols; Kazmierski,
W. M., Ed.; Methods in Molecular Medecine, Vol. 23; Humana Press:
Totowa, NJ, 1999; pp 106-121.
(27) Shin, I.; Lee, M.; Lee, J.; Jung, M.; Lee, W.; Yoon, J. Synthesis of
optically active phthaloyl D-aminooxy acids from L-amino acids or
L-hydroxy acids as building blocks for the preparation of aminooxy
peptides. J. Org. Chem. 2000, 65, 7667–75.
(28) Yang, D.; Li, B.; Ng, F. F.; Yan, Y. L.; Qu, J.; Wu, Y. D. Synthesis
and characterization of chiral N-O turns induced by alpha-
aminoxy acids. J. Org. Chem. 2001, 66, 7303–12.
(29) Weber, I.; Potier, P.; Thierry, J. Triflates as synthons for the
synthesis of lysine analogues. Tetrahedron Lett. 1999, 40, 7083–
7086.
(30) Hoffman, R. V.; Maslouh, N.; Cervantes-Lee, F. Highly stereo-
selective syntheses of syn- and anti-1,2-amino alcohols. J. Org.
Chem. 2001, 67, 1045–1056.
(31) Lineweaver, H.; Burk, D. The determination of enzyme dissocia-
tion constants. J. Am. Chem. Soc. 1934, 56, 658–666.
References
(1) Dean, M.; Hamon, Y.; Chimini, G. The human ATP-binding
cassette (ABC) transporter superfamily. J. Lipid Res. 2001, 42,
1007–1017.
ꢀ
^
(2) Mourez, M.; Jehanno, M.; Hofnung, M.; Dassa, E. Role, fonc-
ꢁ
tionnement et structure des transporteurs a ATP binding cassette
(ABC). Med. Sci. 2000, 16, 386–94.
(3) Kim, R. B. Drug transporters in HIV Therapy. Top. HIV Med.
2003, 11, 136–139.
(4) Wang, X.; Furukawa, T.; Nitanda, T.; Okamoto, M.; Sugimoto,
Y.; Akiyama, S.; Baba, M. Breast cancer resistance protein (BCRP/
ABCG2) induces cellular resistance to HIV-1 nucleoside reverse
transcriptase inhibitors. Mol. Pharmacol. 2003, 63, 65–72.
(5) Weiss, J.; Rose, J.; Storch, C. H.; Ketabi-Kiyanvash, N.; Sauer, A.;
Haefeli, W. E.; Efferth, T. Modulation of human BCRP (ABCG2)
activity by anti-HIV drugs. J. Antimicrob. Chemother. 2007, 59,
238–245.
(6) Juliano, R. L.; Ling, V. A surface glycoprotein modulating drug
permeability in Chinese hamster ovary cell mutants. Biochim.
Biophys. Acta, Biomembr. 1976, 455, 152–162.
(7) Cole, S. P.; Bhardwaj, G.; Gerlach, J. H.; Mackie, J. E.; Grant,
C. E.; Almquist, K. C.; Stewart, A. J.; Kurz, E. U.; Duncan, A. M.;
Deeley, R. G. Overexpression of a transporter gene in a multidrug-
resistant human lung cancer cell line. Science 1992, 258, 1650–1654.
(8) Doyle, L. A.; Yang, W.; Abruzzo, L. V.; Krogmann, T.; Gao, Y.;
Rishi, A. K.; Ross, D. D. A multidrug resistance transporter from
human MCF-7 breast cancer cells. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 15665–15670.
(9) Litman, T.; Brangi, M.; Hudson, E.; Fetsch, P.; Abati, A.; Ross,
D. D.; Miyake, K.; Resau, J. H.; Bates, S. E. The multidrug-
resistant phenotype associated with overexpression of the new
ABC half-transporter, MXR (ABCG2). J. Cell Sci. 2000, 113 (Part
11), 2011–2021.
(32) Comte, G.; Daskiewicz, J.-B.; Bayet, C.; Conseil, G.; Viornery-
Vanier, A.; Dumontet, C.; Di Pietro, A.; Barron, D. C-Isoprenyla-
tion of flavonoids enhances binding affinity toward P-glycoprotein
and modulation of cancer cell chemoresistance. J. Med. Chem.
2001, 44, 763–768.
(33) Cardarelli, C. O.; Aksentijevich, I.; Pastan, I.; Gottesman, M. M.
Differential effects of P-glycoprotein inhibitiors on NIH3T3 cells
transfected with wild-type (G185) or mutant (V185) multidrug
transporters. Cancer Res. 1995, 55, 1086–1091.
(34) Robey, R. W.; Steadman, K.; Polgar, O.; Morisaki, K.; Blayney,
M.; Mistry, P.; Bates, S. E. Pheophorbide a is a specific probe
for ABCG2 function and inhibition. Cancer Res. 2004, 64, 1242–
1246.
(35) Chang, X.-B.; Hou, Y.-X.; Riordan, J. R. ATPase activity of
purified multidrug resistance-associated protein. J. Biol. Chem.
1997, 272, 30962–30968.
(10) Pascaud, C.; Garrigos, M.; Orlowski, S. Multidrug resistance
transporter P-glycoprotein has distinct but interacting binding
sites for cytotoxic drugs and reversing agents. Biochem. J. 1998,
333 (Part 2), 351–358.
(11) Martin, C.; Berridge, G.; Higgins, C. F.; Mistry, P.; Charlton, P.;
Callaghan, R. Communication between multiple drug binding sites
on P-glycoprotein. Mol. Pharmacol. 2000, 58, 624–632.
(12) Colabufo, N. A.; Berardi, F.; Cantore, M.; Contino, M.; Inglese,
C.; Niso, M.; Perrone, R. Perspectives of P-glycoprotein modulat-
ing agents in oncology and neurodegenerative diseases: pharma-
ceutical, biological, and diagnostic potentials. J. Med. Chem. 2009,
53, 1883–1897.
ꢀ
(13) Garrigues, A.; Loiseau, N.; Delaforge, M.; Ferte, J.; Garrigos, M.;
ꢀ
Andre, F.; Orlowski, S. Characterization of two pharmacophores
on the multidrug transporter P-glycoprotein. Mol. Pharmacol.
2002, 62, 1288–1298.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 18 6729
(36) Halgren, T. A. Merck molecular force field. I. Basis, form, scope,
parameterization, and performance of MMFF94. J. Comput.
Chem. 1996, 17, 490–519.
(37) Halgren, T. A. Merck molecular force field. II. MMFF94 van der
Waals and electrostatic parameters for intermolecular interactions.
J. Comput. Chem. 1996, 17, 520–552.
(38) Halgren, T. A. Merck molecular force field. III. Molecular geo-
metries and vibrational frequencies for MMFF94. J. Comput.
Chem. 1996, 17, 553–586.
(39) Halgren, T. A.; Nachbar, R. B. Merck molecular force field. IV.
conformational energies and geometries for MMFF94. J. Comput.
Chem. 1996, 17, 587–615.