Novel structure-activity relationships and selectivity profiling of cage dimeric 1,4-dihydropyridines as multidrug resistance (MDR) modulators

Article abstract of DOI:10.1016/j.bmc.2010.06.004

Synthesized series of cage dimeric 1,4-dihydropyridines have been systematically evaluated as MDR modulators in in vitro assays to investigate structure-dependent selectivity properties of inhibiting most cancer-relevant efflux pump proteins. Structure-ac

Full text of DOI:10.1016/j.bmc.2010.06.004

Bioorganic & Medicinal Chemistry 18 (2010) 4983–4990
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Novel structure–activity relationships and selectivity profiling of cage
dimeric 1,4-dihydropyridines as multidrug resistance (MDR) modulators
Claudius Coburger a, Jörg Wollmann a, Martin Krug a, Christiane Baumert a, Marianne Seifert a, Joséf Molnár b,
Hermann Lage c, Andreas Hilgeroth a,
*
a Institute of Pharmacy, Martin Luther University, 06120 Halle, Germany
b Department of Medical Microbiology, University of Szeged, Szeged, Hungary
c Institute of Pathology, Charité Hospital, Berlin, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Synthesized series of cage dimeric 1,4-dihydropyridines have been systematically evaluated as MDR
modulators in in vitro assays to investigate structure-dependent selectivity properties of inhibiting most
cancer-relevant efflux pump proteins. Structure–activity relationships of each P-glycoprotein (P-gp) and
multidrug resistance associated protein (MRP) 1 and MRP2 inhibition are discussed and prove to be
mainly determined by certain aromatic substitution patterns. The characterization of breast cancer resis-
tance protein (BCRP) inhibition results in the discovery of benzyloxy substituted derivatives as selective
P-gp inhibitors.
Received 14 January 2010
Revised 3 June 2010
Accepted 4 June 2010
Available online 9 June 2010
Keywords:
Structure–activity relationships (SAR)
Multidrug resistance (MDR)
P-Glycoprotein (P-gp)
Ó 2010 Elsevier Ltd. All rights reserved.
Multidrugresistanceassociatedprotein(MRP)
Breast cancer resistant protein (BCRP)
Selectivity
1. Introduction
efforts like a transcriptional control of the efflux pump protein for-
mation have not been successful so far.2,10,11
The main obstacle in cancer treatment of today is the phenome-
non of multidrug resistance (MDR).1,2 Occurring resistances against
anticancer drugs which are caused by single mutations are of less
importance because structurally varied drugs that show a different
binding mode to the anticancer target protein of the drug may com-
pensate such a single resistance problem. The MDR phenomenon
concerns numerous structurally different drug families and makes
cancer treatment to a significant problem in these cases.2,3 The main
strategy to overcome MDR has been the development of structurally
new drugs that are unlikely to be affected by the MDR phenome-
non.2 Even in the case of newest therapeutics like the tyrosine kinase
inhibitors or monoclonal antibodies such resistance developments
were observed.2,4–6 Main causative agents of the MDR phenomenon
have been transmembrane efflux pumps which transport drugs out
of the cells so that therapeutically necessary intracellular drug levels
are not reached.7,8 Most important cancer-relevant efflux pumps are
P-glycoprotein (P-gp), the multidrug resistance associated protein
(MRP) 1 and the breast cancer resistance protein (BCRP).8,9 The
development of inhibitors of the efflux pump activities remains
the main perspective strategy to overcome MDR because alternative
Most important problems in the clinical application of such
inhibitors have been massive side effects and drug interactions
which strongly limit their usage so far.12,13 The undesired side ef-
fects mainly result from a nonselective inhibition of the different
naturally occurring efflux pumps by such inhibitors.14 Most inhib-
itors affect not only a desired tumour-cell relevant overexpressed
efflux pump but also other structurally related ones which protect
normal cells from toxification. This nonselective inhibition in-
creases the toxicity of co-applicated anticancer drugs which are
usually transported out of those normal cells which express such
naturally occurring efflux pumps.15 So undesired side effects result
from the use of nonselective efflux pump inhibitors.
The actual strategy in the development of efflux pump inhibi-
tors concentrates on the development of selective inhibitors in
the case of P-gp, MRP1 and MRP2, because especially these efflux
pumps occur in many human tissues so that a selective inhibition
of one targeted efflux pump will result in reduced side effects in
therapy.8,15 However, the sole inhibition of P-gp will not exclude
the observation of side effects.
Actually a target-based development of inhibitors is not possible
because of missing highly solved structures of the different efflux
pump proteins. The only strategy is the investigation of inhibitor-
structure dependent effects on the efflux pump inhibition and
* Corresponding author. Tel.: +49 345 55 25168; fax: +49 345 55 27207.
E-mail address: andreas.hilgeroth@pharmazie.uni-halle.de (A. Hilgeroth).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.06.004

4984
C. Coburger et al. / Bioorg. Med. Chem. 18 (2010) 4983–4990
varied lipophilic and hydrophilic substituents in our cage dimeric
R
1,4-dihydropyridines (Fig. 1).
Y
The influence on structure–activity relationships of the P-gp
inhibition is discussed. The inhibition of MRP1 and MRP2 has been
determined. The selectivity profile of the cage dimers has been
strengthened by the characterization of the BCRP inhibition.
X
Y
OH
R
X
X
OH
Y
N
N
2. Results and discussion
2.1. Chemistry
HO
Y
X
R
OH
R
R = H, OMe, OBz, OH
X or Y = C or N
The 1,4-dihydropyridines 1 have been given by a primary cyclo-
addition reaction of an aromatic aldehyde, ethyl propiolate and,
finally, the benzylamine derivative in freshly distiled acetic acid
after 2 h of stirring under reflux heating except compound 1h
(Scheme 1). Compound 1h was yielded by the alkylation of the
N-unsubstituted 1,4-dihydropyridine given by the above cyclocon-
densation reaction with ammonium acetate instead of the
benzylamine derivative. The alkylation was carried out in dimethyl-
cage dimeric 1,4-dihydropyridine
target structure
Figure 1.
selectivity to characterize a potential inhibitor binding site of the
respective efflux pump protein.
propyleneurea. The 1,4-dihydropyridine was treated with
a
We develop a series of structurally varied cage dimeric 1,4-dihy-
dropyridines as P-gp inhibitors. Similar to recently published
quinine and stipiamide homodimers the cage dimers have been
derived from monomers, that is, 1,4-dihydropyridines as P-gp
inhibitors.16,17 Monomeric 1,4-dihydropyridines have been early
MDR modulators with some P-gp inhibiting activities.10 They had
the disadvantage of retaining calcium antagonistic activities as
the original pharmacological activities of this compound class.
Therefore, their application led to unfavourable side effects.10 We
sevenfold excess of sodium hydride and then with a 1.2-fold excess
of 4-picolyl bromide.
The cage dimers 2a–j resulted from a double [2+2]cycloaddition
reaction of two corresponding monomers 1 as exclusive reaction
products under excitation of the dihydropyridine and the conju-
gated carbamide ester chromophores, respectively, at irradiation
wavelengths >270 nm using Ultra VitaluxÒ lamps.18 The cage dimers
mainly crystallized from the solutions during the irradiation
procedurein THF and gavefinal yields of about 80% after evaporation
R1
R1
N
O
O
R3O
OR3
O
H
O
O
EtO
OEt
OH
O
O
i
ii
MeO
OMe
+
R2
N
H
1a-g, i, j: R3 = Et
NH2
iii
1h
: R3 = Me
R2
OR3
OR
3
OH
OH
O
R1
O
R1
R2
R2
N
iv
N
N
N
R1
R2
3a-k, m
R1
R2
O
R3O
O
OR3
HO
2a-g, i, j: R3 = Et
2h
: R3 = Me
v
3d
3l
Compd.
R1
R2
1a, 2a, 3a
1b, 2b, 3b
1c, 2c, 3c
1d, 2d, 3d
1e, 2e, 3e
1f, 2f, 3f
phenyl
phenyl
4-methoxyphenyl
4-benzyloxyphenyl
phenyl
3-benzyloxyphenyl
phenyl
phenyl
4-pyridyl
phenyl
phenyl
4-methoxyphenyl
4-methoxyphenyl
phenyl
4-benzyloxyphenyl
phenyl
3-benzyloxyphenyl
4-pyridyl
1g, 2g, 3g
1h, 2h, 3h
1i, 2i, 3i
phenyl
3-pyridyl
phenyl
phenyl
1j, 2j, 3j
1k, 2k, 3k
1l, 2l, 3l
3-hydroxyphenyl
4-hydroxyphenyl
phenyl
1m, 2m, 3m
4-hydroxyphenyl
Scheme 1. Reagents and conditions: (i) acetic acid, 100 °C, 2 h; (ii) DMPU, NaH, rt, picolylbromide, 30 min; (iii) THF, rt, k >270 nm; (iv) THF, À18 °C, LiAlH4; (v) TFA,
thioanisole, rt.

C. Coburger et al. / Bioorg. Med. Chem. 18 (2010) 4983–4990
4985
Table 1
Concentration dependent P-gp, MRP1, MRP2 and BCRP inhibiting properties of target compounds (3a–m)
Compd
R values
P-gp
MRP1a
10
MRP2
10
BCRP
1
lM
10
l
M
l
M
l
M
0.1
lM
1 l
Mb
3a
3b
3c
3d
3e
3f
3g
3h
3i
3j
3k
3l
4.1 0.3
3.5 0.6
3.1 0.5
3.3 1.1
2.3 0.3
65 8.1
117 1.7
1.5 0.1
2.4 0.1
1.2 0.1
1.2 0.1
1.3 0.4
1.5 0.1
4.0 1.4
n.d.e
29 3.1c
13 0.3c
46 8.0c
47 2.3
28 2.5
125 16
205 6.1
1.1 0.1
7.7 0.4
1.1 0.5
3.0 0.2
1.5 0.7
3.3 0.2
5.1 0.2
n.d.e
2.3 0.4
3.4 0.8
4.8 1.0
1.0 0.3
2.5 0.6
1.9 0.5
2.8 0.8
1.2 0.2
2.2 0.7
1.1 0.1
1.3 0.3
1.8 0.6
1.6 0.4
n.a.d
1.02 0.14
1.10 0.19
1.25 0.19
1.56 0.66
0.75 0.21
0.97 0.27
0.97 0.26
0.80 0.23
0.74 0.29
0.90 0.29
1.05 0.43
0.63 0.18
0.82 0.06
n.a.d
1.1 0.2
1.2 0.2
1.5 0.1
1.4 0.3
n.a.d
1.9 0.2
3.2 0.6
6.1 1.7
3.5 0.2
1.2 0.1
2.8 0.3
4.2 0.2
1.3 0.2
1.2 0.1
n.a.d
1.3 0.3
1.5 0.1
n.a.d
n.a.d
n.a.d
n.a.d
1.1 0.2
1.5 0.2
1.1 0.1
n.a.d
n.a.d
3m
n.a.d
Verapamil
Indomethacin
Fumitremogin c
n.a.d
4.1 0.4
0.75 0.14
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
1.9 0.4
6.0 1.1
a
b
c
Calculated R values are the quotients of each measured fluorescence intensity in the inhibitor-treated cells and of that in the inhibitor-untreated control.
Activity saturation concentration.
Activity saturation effects.
Not active.
d
e
Not determined.
of the solvent under atmosphere pressure. The pyridyl-containing
monomers 1h and j gave similar higher yields of the corresponding
dimers after evaporation whereas the oily compound 1i resulted
in a lower dimer yield with some side products being detectable
by tlc.
The introduction of a methoxy function into two phenyl resi-
dues of derivative 3a resulted in similar activities of derivative
3b at the lower concentrations and some reduced activity at the
higher concentrations if compared to the only phenyl substituted
derivative 3a. Two additional methoxy functions in compound 3c
mainly increased the inhibition at the higher inhibitor concentra-
tion. So as a final considering of the effect of methoxy functions
on the P-gp inhibition properties of the new compound class it
can be stated that the introduction of such methoxy functions is
of a significant favour for the P-gp inhibition. Methoxy functions
in MDR modulators have been discussed to serve as hydrogen bond
acceptor functions to amino acid residues of the protein backbone
of a potential P-gp binding region.9,20,21
The alcoholic target compounds 3a–k and m were given by a
lithium aluminium hydride reaction of the ester functions to
hydroxymethylene groups and a further reductive debenzylation
in the case of yielded compounds 3k and m. The reactions were
carried out in dried THF at low temperatures of À18 °C. The pyri-
dine nucleus of the pyridyl-containing derivatives remained un-
touched at that temperature. The total yields of almost about
60% of the target compounds 3 have been isolated from chloroform
solutions of the oily residues after hydrolysis and solvent evapora-
tion. The 4-hydroxyphenyl substituted alcoholic target dimer 3l
was yielded from the 4-benzyloxyphenyl derivative 3d by an acid
catalyzed debenzylation procedure using trifluoroacetic acid and
thioanisole.
Moreover, the observed SAR of increased numbers of methoxy
functions in our class of MDR modulators that result in higher
activities suggests an important hydrogen binding of the com-
pounds to a certain P-gp binding site.
Another common feature of MDR modulators that has been dis-
cussed to be of favour for the biological activities of generally
inhibiting transmembrane efflux pumps has been an increased
lipophilicity.9 Such an increased lipophilicity facilitates a mem-
brane penetration of an inhibitor-acting MDR modulator to reach
a transmembrane efflux pump protein binding site within the
membrane. An increased lipophilicity in P-gp inhibitors has been
discussed to directly influence the biological activity by a better
P-gp binding to the P-gp binding site which is accessible from
the membrane. However, we recently demonstrated that an in-
creased lipophilicity of 3-benzyloxy-1-aza-9-oxafluorenes as a no-
vel class of P-gp inhibitors did not lead to an increased biological
activity.22 Both possible lipophilic interactions and additionally
hydrogen bonding via hydrogen bond acceptor functions play a
central role in the P-gp binding of inhibitors. 1,4-Dihydropyridines
were demonstrated to bind to the P-gp binding site in the cytosolic
membrane leafleft.23 Benzyloxy instead of methoxy functions have
been introduced into our cage dimers to investigate the lipophilic-
ity influence on biological activity. However, the exact positioning
of the benzyloxy substituent within the phenyl residues was found
to play an important role.19 The positioning of the benzyloxy sub-
stituent into the 3-position of the benzyl ring resulted in a main
2.2. Structure–activity relationships (SAR) of the P-gp inhibition
The P-gp inhibition was determined as the extent of the cellular
uptake of the fluorescent P-gp substrate rhodamine 123 both into
the P-gp expressing mouse T lymphoma cell line and the not P-gp
expressing parental cell line under inhibitor treatment and with
each fluorescence related to the inhibitor-untreated control. The
P-gp expression resulted from retrovirus transfection with the hu-
man mdr1 gene so that it is assured that an increase in fluorescence
uptake can only result from a P-gp inhibition and a consequent low-
ered efflux from the P-gp expressing cell line compared to the
parental cell line.
The only phenyl substituted derivative 3a was demonstrated to
show strong inhibition properties both at the lowered and the
higher inhibitor concentrations compared to verapamil used as
control (Table 1).19 Nifedipine as a known P-gp inhibitor of the
1,4-dihydropyridine type practically showed no activities at a con-
centration of 10 lM with a R value of 1.0 0.1. So already the only
phenyl substituted dimeric 1,4-dihydropyridine 3a is much more
active than a monomeric 1,4-dihydropyridine.

4986
C. Coburger et al. / Bioorg. Med. Chem. 18 (2010) 4983–4990
increase in activity of compound 3g. A similar increase was also
observed for derivative 3f if compared to 3d that was found as
active as the methoxy derivative 3c with the double number of
methoxy functions. So beside the hydrogen bond acceptor func-
tions also the additional benzyl substituents contribute to the bio-
logical activities presumably by a better membrane penetration as
discussed above. We also calculated lipophilicity constants as log P
values of our cage dimers (Table 2). All the four benzyloxy deriva-
tives show a mainly increased lipophilicity with log P values of
about 6.5. The observed importance of an exact positioning within
the molecular scaffold which led to the observed significant differ-
ences in the biological activities may be caused by hydrophobic
interactions with the amino acid residues of the respective P-gp
binding site.19
investigated within the various MDR modulator classes.25 We
introduced hydroxy functions into both the 3- and the 4-position
of the aromatic residues to investigate this influence. The activities
of the resulting derivatives 3k–m were poor and independent from
the positioning of the hydroxy functions. The calculated log P val-
ues of about 3.0 were similar to those of the methoxyphenyl
substituted derivatives 3b and c which showed mainly increased
activities. These results clearly suggest that hydroxy functions as
hydrogen bond donator functions are unfavourable as reported
previously.20
So finally, it can be concluded that the new compound class of
P-gp inhibitors addresses a certain hydrophobic binding site which
is sensitive for the exact positioning of the lipophilic aromatic res-
idues for binding interactions.
The results clearly show that both lipophilic properties and
hydrogen bond acceptor functions mainly attribute to the P-gp
inhibiting activities. The more lipophilic substitution of the benzyl-
oxy substituted derivatives 3d–g resulted in higher inhibitory
activities than the less lipophilic substitution in the corresponding
methoxy substituted derivatives 3b and c with calculated log P val-
ues of about 3.6. In the case of the less lipophilic methoxy deriva-
tives the increase in methoxy groups as hydrogen bond acceptor
functions led to activity data of compound 3c similar the those of
the benzyloxy substituted derivative 3d with a higher log P value.
As the methoxy functions as hydrogen bond acceptor functions
were found favourable for the P-gp inhibiting properties we alter-
natively introduced nitrogen atoms into the phenyl residues to get
more insight in the SAR. Moreover, the introduction of the proton-
able nitrogen functions was suggested to be of favour because
most P-gp inhibitors have protonable nitrogen functions within
their molecular skeletons.24 Surprisingly, an introduction of a 4-
picolyl residue in derivative 3h led to a main decrease in activity
if compared to the methoxybenzyl substituted derivative 3b. A
movement of the nitrogen atom from the 4- to the 3-position of
the benzyl residue practically caused no changes in activity. In
the case of the benzyloxy substituents the position within the ben-
zylic residue resulted in main changes in activity. Consequently, it
can be concluded from the SAR that a role of the nitrogen atoms as
hydrogen bond acceptor functions seems to be unlikely. Further-
more, it can be concluded that the activities of the benzyloxy deriv-
atives are mainly caused by their ability to undergo hydrophobic
interactions with amino acid residues of the respective P-gp bind-
ing site. When the nitrogen atom was presented in the 4-position
of the phenyl residue in compound 3i instead of the benzyl residue
increases in activity were observed. However, the activity was even
lower than that of the methoxy substituted derivative 3b.
2.3. MDR modulator properties towards MRP1 and MRP2
MRP1 and P-gp have a great overlap in their anticancer drug
profiles which means that most P-gp substrates are also MRP1 sub-
strates. Most inhibitors of P-gp also inhibit MRP1 so that a selectiv-
ity of inhibition is not given which leads to severe side effects by
such dual inhibitors.
So it was of great interest to investigate the ability of our novel
cage dimers to inhibit MRP1 to work out structural criteria for a
selective P-gp inhibition and no MRP1 inhibition.
Although MRP2 does not play a central role in the MDR it is a
naturally spread efflux pump in human tissues that should not
be affected by a selective P-gp inhibitor to reduce side effects un-
der co-application of anticancer drugs which are MRP2 substrates.
The MRP inhibiting properties of the cage dimers have been
characterized by the determination of the uptake of the MRP spe-
cific substrate carboxyfluorescein using the MRP1 overexpressing
ovarial carcinoma cell line A2780P and for the determination of
the MRP2 inhibiting properties using the MRP2 overexpressing
subline A2780RCIS which was induced to overexpress MRP2 after
treatment with cisplatin. The MRP substrate specificity of carboxy-
fluorescein ensured that changes in the extent of the fluorescent
carboxyfluorescein uptake exclusively resulted from an inhibition
of MRP1 in the parental cell line A2780P whereas the MRP2 spe-
cific extent of fluorescence uptake was given by substraction of
the value of the MRP1 specific extent of fluorescence uptake into
the parental cell line from the whole extent of fluorescence deter-
mined as MRP1 and MRP2 inhibition in the induced cell line
A2780RCIS.
The only phenyl substituted derivative 3a was an inhibitor of
MRP1 although its activity at 10 lM made only half of the activity
Another aspect in the SAR of MDR modulators is the influence of
hydrogen bond donator functions on the biological activities. So far
the influence of such functional groups has not been systematically
than that of the indomethacin control (Table 1). Verapamil is a
known inhibitor of both MRP1 and MRP2 at high concentrations
>100 l
M.26 We found a consequent increase in the carboxyfluores-
cein uptake with increasing methoxy functions in compounds 3b
and, finally, 3c so that compound 3c was a slightly better inhibitor
than indomethacin. So the extent of methoxy functions as hydrogen
bond acceptor functions plays a central role in the MRP1 inhibiting
properties suggesting a better binding to amino acid residues of the
MRP1 binding site with increasing numbers of such hydrogen bond
acceptor functions. This is an important finding in the SAR of MRP1
inhibitors because less is known about structural features which
strengthen the MRP1 inhibiting properties of inhibitors.
Within the inhibitory activities of the benzyloxy substituted
compounds interesting SAR could be found. Both phenyl substi-
tuted derivatives 3d and f were practically no MRP1 inhibitors,
whereas the benzyl substituted derivatives 3e and g showed weak
inhibition of MRP1. Within the group of the benzyloxy substituted
compounds the 3-substituted derivatives were more active than
the poorly or not active 4-substituted derivatives. In contrast to
the P-gp inhibiting properties which were found mainly increased
Table 2
Lipophilicity constant (log P value) calcu-
lation data of target compounds (3a–m)
Compd
log P value
3a
3b
3c
3d
3e
3f
3g
3h
3i
3.8 0.3
3.6 0.4
3.6 0.6
6.5 1.0
6.5 1.0
6.4 1.0
6.4 1.0
1.5 0.4
1.6 0.3
1.5 0.4
3.0 0.2
3.0 0.2
3.0 0.2
3j
3k
3l
3m

C. Coburger et al. / Bioorg. Med. Chem. 18 (2010) 4983–4990
4987
in case of the lipophilic benzyloxy substituted compounds the
more lipophilic substitution in these compounds is less favourable
for a binding to the potential MRP1 binding region. This suggests
that the MRP1 binding region of our compounds is less lipophilic
than the P-gp binding region. Another remarkable feature is the
fact that a more lipophilic compound character for a possibly bet-
ter membrane penetration to reach the potential MRP1 binding
site is not of favour for a MRP1 inhibiting compound if compared
to the structural necessities of a compound to inhibit P-gp.
From the pyridyl substituted compounds only the 4-pyridyl
substituted derivative 3i showed some MRP1 inhibiting activities
whereas the picolyl substituted compounds 3j and h were practi-
cally no MRP1 inhibitors. Similar results were found for the
hydroxyphenyl derivatives 3k–m with the 4-hydroxyphenyl
substituted derivative 3l showing weak activities.
activities decreased. All of the pyridyl substituted derivatives 3h–j
and the hydroxy substituted derivatives 3k–m were practically no
BCRP inhibitors at the given concentrations.
So we found interesting SAR of the BCRP inhibition profile of our
novel cage dimeric MDR modulators suggesting hydrogen bonding
to the amino acids of the protein backbone via the methoxy func-
tions and to hydrophobic regions via the benzyloxy substituents.
3. Conclusions
We systematically characterized SAR of a novel class of MDR
modulators towards P-gp, MRP1, MRP2 and BCRP. As highly solved
structures of the various efflux pump proteins do not exist essen-
tial SAR for an important selective efflux pump inhibition could
only be investigated by a variation of functional groups within
the relevant aromatic substituents which were found essential
for a good biological activity, that is, methoxy functions and nitro-
gen atoms which may act as hydrogen bond acceptor functions and
hydroxy functions as additional hydrogen bond donator functions.
Methoxy groups as hydrogen bond acceptor functions played an
important role in the inhibition activities towards all efflux pump
proteins. Whereas the P-gp inhibition was strengthened by more
lipophilic benzyloxy substituents, the MRP inhibition was mainly
reduced, so that selectively binding compounds 3d–g were found.
The derivatives 3f and g with a 3-positioned benzyloxy substituent
showed the comparable strongest differences in activity of P-gp
and MRP1/MRP2 inhibition, respectively, and thus are outstanding
selective efflux pump inhibitors.
The BCRP profiling suggested a hydrophobic binding site for an
exact inhibitor binding and led to the characterization of the
4-benzyloxy derivative 3e as overall selective P-gp inhibitor. An-
other selective P-gp inhibitor with no inhibition of MRP1, MRP2
and BCRP was the 4-pyridyl substituted derivative 3i.
So beside the discovery of important selective P-gp inhibitors,
novel insight in the SAR of MDR modulators was given by the
inhibitor binding towards the hitherto not characterized binding
sites of the efflux pumps MRP1, MRP2 and BCRP.
So our systematical investigations gave interesting insight in
the SAR of MRP1 inhibitors. Methoxy functions as hydrogen bond
acceptor functions play an important role in activity whereas more
lipophilic substituents are not favourable.
The investigations of the MRP2 inhibitions showed an increase
in the MRP2 inhibiting properties from 3a with no additional
methoxy function to 3b with two methoxy functions and, finally,
to 3c with four methoxy functions. So hydrogen bond acceptor
functions also play an important role in the inhibitor binding to
the MRP2 binding site. From the four benzyloxy substituted deriv-
atives 3d–g only the 4-phenyl substituted compound 3d had some
MRP2 inhibiting properties. Within the group of pyridyl substi-
tuted derivatives 3h–j only poor activities were found and within
the group of the hydroxyphenyl substituted derivatives only the
compound 3k showed some activities.
The SAR in the MRP2 inhibition profile within our new class of
MDR modulators are similar to those of the MRP1 inhibition with
an importance of methoxy functions as hydrogen bond acceptor
functions and unfavourable more lipophilic substitutions so that
generally the benzyloxy substituted derivatives that have been
strong P-gp inhibitors are poor or practically no inhibitors of nei-
ther MRP1 nor MRP2. Thus they turn out as ideal selective P-gp
inhibitors.
4. Experimental
4.1. Chemistry
2.4. BCRP inhibition profiling
If compared to P-gp or MRP1 the tissue spread of BCRP in hu-
mans is limited to liver and intestine.8,15 The expression of BCRP
has been reported in some solid tumours.27 However, the role of
BCRP in the emergence of MDR is yet under discussion.2 We deter-
mined the inhibition profile of our new P-gp inhibitors using the
fluorescent BCRP substrate mitoxantrone in comparison of the
fluorescence extent uptake into the inhibitor-treated BCRP induced
gastric carcinoma cell line EPG85-257RNOV and the non-BCRP-
overexpressing parental cell line EPG85-257P each related to the
untreated control. We used the fungal toxin fumitremorgin c as
inhibitor in relevant BCRP inhibiting concentrations up to satura-
tion effects (Table 2).
All the chemical agents used were either synthesized or have
been commercially available. Melting points were determined
using a Boetius melting desk microscope and are uncorrected. Pro-
ton NMR spectra were recorded on a Varian Gemini 2000 at 200 or
400 MHz, respectively. Chemical shifts are reported in ppm units
with tetramethylsilane as internal reference standard. Mass spec-
tra were recorded on an AMD 402 mass spectrometer named
AMD INTEGRA. Elemental analyses (C, H, N) were carried out with
a Leco analyzer apparatus (CHN-932) and the results were within
0.4% of the theoretical values. The synthesis of compounds 1a,b,
d–f, 2a,b,d–f and 3a,b,d–f has been demonstrated.18
The allover phenyl substituted compound 3a is only a poor BCRP
inhibitor at the higher concentrations if compared to the control.
Similar to the inhibition of MRP1 and MRP2, respectively, we found
an increase of BCRP inhibiting activities with an increasing number
of methoxy functions from derivative 3a–c. The benzyloxy substi-
tuted compounds show some BCRP inhibiting activities at higher
concentrations. Interestingly, the 4-benzyl substituted derivative
3e is practically no BCRP inhibitor. The positioning of the benzyloxy
substituent within the molecular scaffold is also of importance for
the BCRP inhibition which will result from an interaction with the
amino acids of the potential BCRP binding site. If presented in the
3-position of the benzyl substituent in derivative 3g we found mod-
erate activities and if moved to the 4-position in compound 3e the
4.1.1. General procedure for the preparation of the 4-aryl- and
4-heteroaryl-1,4-dihydropyridines (1c,i,j)
Either the aromatic or the heteroaromatic aldehyde (10 mmol),
ethyl propiolate (20 mmol) and the benzylamine derivative
(10 mmol) were dissolved in freshly distiled acetic acid (1 mL) un-
der heating.18 The solution was then refluxed for 2 h on a steam
bath. After cooling to rt sufficient water was added and extraction
was carried out with portions of chloroform (50 mL) for several
times. After drying over sodium sulfate and filtration the solvent
was removed in vacuum. The remaining oil was dissolved in dried
ethanol from which the 1,4-dihydropyridine crystallized on cooling.

4988
C. Coburger et al. / Bioorg. Med. Chem. 18 (2010) 4983–4990
4.1.1.1. Ethyl 1-(4-methoxyphenymethyl)-4-(4-methoxyphe-
nyl)-1,4-dihydropyridine-3,5-dicarboxylate (1c). Yellow crys-
tals, mp 101–103 °C (ethanol); yield 84%; MS (ESI) m/z 474
(M+Na+, 15%), 925 (2 M+Na+, 100%). 1H NMR (CDCl3) d 1.16 (t,
J = 7.3 Hz, 6H, COOCH2CH3), 3.74, 3.81(2 Â s, 6H, OCH3), 4.05 (q,
J = 7.3 Hz, 4H, COOCH2CH3), 4.49 (s, 2H, NCH2), 4.81 (s, 1H, 4-H),
6.71–6.92 (2 Â m, 4H, 3-, 5-H of Ph), 7.14–7.20 (2 Â m, 4H, 2-, 6-H
of Ph), 7.22 (s, 2H, 2-, 6-H). Anal. Calcd for C26H29NO6: C, 69.16; H,
6.47; N, 3.10. Found: C, 69.39; H, 6.64; N, 3.50.
4.1.3.1. Tetraethyl 3,9-bis(4-methoxyphenylmethyl)-6,12-bis(4-
methoxyphenyl)-3,9-diazahexacyclo[6.4.0.02.7.04.11.05.10]dode-
cane-1,5,7,11-tetracarboxylate (2c). White solid; mp 238–
241 °C (ethanol); yield 75%; MS(ESI) m/z 903 (M+H+, 20%), 925
(M+Na+, 100%). 1H NMR (CDCl3)
d 1.01 (t, J = 7.1 Hz, 12H,
COOCH2CH3), 3.73, 3.78 (2 Â s, 12H, OCH3), 3.97 (q, J = 7.1 Hz, 8H,
COOCH2CH3), 4.21 (s, 2H, 6-, 12-H), 4.22 (s, 4H, 2-, 4-, 8-, 10-H),
4.39 (s, 4H, NCH2), 6.60–6.82 (2 Â m, 8H, 3-, 5-H of ArH), 7.12–
7.24 (2 Â m, 8H, 2-, 6-H of ArH). Anal. Calcd for C52H58N2O12: C,
69.16; H, 6.47; N, 3.10. Found: C, 69.08; H, 6.54; N, 3.48.
4.1.1.2. Ethyl 1-benzyl-4-(4-pyridyl)-1,4-dihydropyridine-3,5-
dicarboxylate (1i). Brownish oil (petrol ether); yield 8%; MS
(ESI) m/z 415 (M+Na+, 100%). 1H NMR (CDCl3) d 1.15 (t, J = 7.2 Hz,
6H, COOCH2CH3), 4.04 (q, J = 7.2 Hz, 4H, COOCH2CH3), 4.57 (s, 2H,
NCH2), 4.88 (s, 1H, 4-H), 7.13–7.41 (m, 9H, 2-, 6-H of pyr, ArH of
Ph, 2-, 6-H), 8.45 (m, 2H, 3-, 5-H of pyr). Anal. Calcd for
4.1.3.2. Tetramethyl 3,9-bis(4-picolyl)-6,12-diphenyl-3,9-diaza-
hexacyclo[6.4.0.02.7.04.11. 05.10]dodecane-1,5,7,11-tetracarboxyl-
ate (2h). White solid; mp 255–262 °C (ethanol); yield 83%;
MS(ESI) m/z 729 (M+H+, 100%). 1H NMR (CDCl3) d 3.54 (s, 12H,
COOCH3), 4.24 (s, 4H, 2., 4-, 8-, 10-H), 4.25 (s, 2H, 6-, 12-H), 4.48
(s, 4H, NCH2), 7.11–7.15 (m, 14H, ArH of Ph, 2-, 6-H of pyr), 8.48
(m, 4H, 3-, 5-H of pyr). Anal. Calcd for C42H40N4O8: C, 69.22; H,
5.53; N, 7.69. Found: C, 69.19; H, 5.55; N, 7.65.
C23H24N2O4: C, 70.39; H, 6.16; N, 7.13. Found: C, 69.99; H, 5.87;
N, 7.49.
4.1.1.3. Ethyl 4-phenyl-1-(3-picolyl)-1,4-dihydropyridine-3,5-
dicarboxylate (1j). Brown prisms, mp 161–164 °C (diethyl ether);
yield 6%; MS (ESI) m/z 393 (M+H+, 100%). 1H NMR (CDCl3) d 1.15 (t,
J = 7.0 Hz, 6H, COOCH2CH3), 4.05 (q, J = 7.0 Hz, 4H, COOCH2CH3),
4.62 (s, 2H, NCH2), 4.89 (s, 1H, 4-H), 7.12–7.24 (m, 7H, ArH of Ph,
2-, 6-H), 8.41 (m br, 1H, 5-H of pyr), 7.68 (d, J = 8.5 Hz, 1H, 6-H
of pyr), 8.61 (m br, 2H, 2-, 4-H of pyr). Anal. Calcd for
4.1.3.3. Tetraethyl 3,9-dibenzyl-6,12-bis(4-pyridyl)-3,9-diaza-
hexacyclo[6.4.0.02.7.04.11. 05.10]dodecane-1,5,7,11-tetracarboxyl-
ate (2i). White solid; mp 225–233 °C (methanol); yield 13%;
MS(ESI) m/z 785 (M+H+, 100%). 1H NMR (CDCl3)
d 0.96 (t,
J = 7.0 Hz, 12H, COOCH2CH3), 3.95 (q, J = 7.0 Hz, 8H, COOCH2CH3),
4.26 (s, 4H, 2-, 4-, 8-, 10-H), 4.27 (s, 2H, 6-, 12-H), 4.47 (s, 4H,
NCH2), 7.19–7.29 (m, 14H, ArH of Ph, 2-, 6-H of pyr), 8.20 (m, 4H,
3-, 5-H of pyr). Anal. Calcd for C46H48N4O8: C, 70.39; H, 6.16; N,
7.13. Found: C, 70.55; H, 6.48; N, 7.19.
C23H24N2O4: C, 70.39; H, 6.16; N, 7.13. Found: C, 70.15; H, 5.95;
N, 6.88.
4.1.2. Methyl 4-phenyl-1-(4-picolyl)-1,4-dihydropyridine-3,5-
dicarboxylate (1h)
4.1.3.4. Tetraethyl 3,9-bis(3-picolyl)-6,12-diphenyl-3,9-diaza-
hexacyclo[6.4.0.02.7.04.11.05.10 ]dodecane-1,5,7,11-tetracarboxyl-
ate (2j). White solid; mp 249–250 °C (ethanol); yield 49%;
The N-unsubstituted methyl 4-phenyl-1,4-dihydropyridine-3,5-
dicarboxylate28 (0.85 g, 3.1 mmol) was dissolved in a minimum
volume of dry dimethylpropyleneurea. A sevenfold molar excess
of sodium hydride as a petrol ether (bp 30–50 °C) washed suspen-
sion in oil (80%) was added. After stirring at 50 °C, the solution was
cooled to rt and a threefold molar excess of 4-picolyl bromide
hydrobromide was added in portions over a period of 30 min. After
additional stirring at rt the excess of sodium hydride was hydroly-
sed with ice water. The precipitating oil was filtered and dissolved
in chloroform after washing with water for several times. After re-
moval of the dried organic layer the oily residue was purified by
column chromatography over silica gel with an eluent mixture of
chloroform/acetic acid ethyl ester/methanol (75:25:5). After re-
moval of the eluent mixture in vacuum the compound crystallized
from diethyl ether in white prisms, mp 209–211 °C; yield 38%; MS
(ESI) m/z 365 (M+H+, 100%). 1H NMR (CDCl3) d 3.16 (s, 6H,
COOCH3), 4.61 (s, 2H, NCH2), 4.93 (s, 1H, 4-H), 7.14–7.29 (m, 9H,
ArH of Ph, 2-, 6-H of pyr, 2-, 6-H), 8.65 (m, 2H, 3-, 5-H of pyr). Anal.
Calcd for C21H20N2O4: C, 69.22; H, 5.53; N, 7.69. Found: C, 69.35; H,
5.83; N, 7.36.
MS(ESI) m/z 785 (M+H+, 100%). 1H NMR (CDCl3)
d 0.97 (t,
J = 7.2 Hz, 12H, COOCH2CH3), 3.99 (q, J = 7.2 Hz, 8H, COOCH2CH3),
4.24 (s, 4H, 2-, 4-, 8-, 10-H), 4.25 (s, 2H, 6-, 12-H), 4.48 (s, 4H,
NCH2), 7.10–7.20 (m, 12H, ArH of Ph, 5-H of pyr), 7.59 (d,
J = 7.3 Hz, 2H, 6-H of pyr), 8.51 (d, J = 8.5 Hz, 2H, 4-H of pyr), 8.55
(s, 2H, 1-H of pyr). Anal. Calcd for C46H48N4O8: C, 70.39; H, 6.16;
N, 7.13. Found: C, 70.25; H, 6.28; N, 7.33.
4.1.4. General procedure for the preparation of the alcoholic di-
meric 4-aryl- and 4-heteroaryl-1,4-dihydropyridines (3c,h–k,m)
Cage dimers 2 (0.07 mmol) were dissolved in 20 mL of dry THF
under heating. After stirring at rt for 1 h the solution was cooled
down to À18 °C and a 1.1-fold molar excess of lithium aluminium
hydride (1 M solution in THF) was added dropwise. After addi-
tional stirring at the low temperature the solution was diluted with
ice water at 0 °C and extracted with chloroform for several times.
The organic layer was then dried over sodium sulfate and filtered.
After evaporation to dryness the oily residue was dissolved in chlo-
roform. Then diethyl ether and petrol ether (bp 30–50 °C) were
added dropwise until turbidity. The resulting precipitating crystal-
line alcoholic compounds 3 were recrystallized from methanol.
4.1.3. General procedure for the preparation of the cage dimeric
alkyl 4-aryl- or 4-heteroaryl-1,4-dihydropyridine-3,5-dicarbox-
ylates (2c,h–j)
4.1.4.1. 1,5,7,11-Tetrakishydroxymethyl-3,9-bis(4-methoxyphe-
nylmethyl)-6,12-bis(4-methoxyphenyl)-3,9-diazahexacyclo-
[6.4.0.02.7.04.11.05.10]dodecane (3c). White crystals; mp 238–
241 °C; yield 75%; MS(ESI) m/z 735 (M+H+, 62%), 757 (M+Na+,
100%). 1H NMR (DMSO-d6) d 2.89 (s, 4H, 2-, 4-, 8-, 10-H), 3.06
(ABX, J = 10.2, 4.4 Hz, 4H, CHBOH), 3.15 (ABX, J = 10.2, 4.4 Hz, 4H,
CHAOH), 3.58 (s, 2H, 6-, 12-H), 3.70, 3.74 (2 Â s, 12H, OCH3), 4.03
(s, 4H, NCH2), 4.35 (t, J = 4.4 Hz, 4H, OH), 6.51–6.68 (m, 4H, 3-, 5-
H of C6-,C12-Ph), 6.86–6.89 (m, 4H, 3-, 5-H of NCH2Ph), 7.16–
7.71 (m, 8H, 2-, 6-H of Ph). Anal. Calcd for C44H50N2O8: C, 71.91;
H, 6.86; N, 3.81. Found: C, 71.68; H, 6.66; N, 3.57.
Monomeric derivatives 1 (1.2 mmol) have been dissolved in
30 mL of dry THF in a quartz flask. Irradiation of the solutions
was carried out under nitrogen atmosphere using Ultra Vitalux
lampsÒ at wavelengths k >270 nm with an irradiation distance
of 60 cm.18 During irradiation most of the cage dimers crystal-
lized from the solutions and irradiation was continued until no
more starting compound could be detected by tlc. During evapo-
ration of the solvent under atmosphere pressure the isolated
yields of the cage dimers 2 were collected and recrystallized from
alcohol.

C. Coburger et al. / Bioorg. Med. Chem. 18 (2010) 4983–4990
4989
4.1.4.2. 1,5,7,11-Tetrakishydroxymethyl-3,9-bis(4-picolyl)-6,
12-diphenyl-3,9-diazahexacyclo[6.4.0.02.7.04.11.05.10]dodecane
(3h). White crystals; mp 291–296 °C; yield 65%; MS(ESI) m/z 617
(M+H+, 100%). 1H NMR (DMSO-d6) d 2.94 (s, 4H, 2-, 4-, 8-, 10-H),
3.12–3.20 (m, ABX, 8H, CHABOH), 3.66 (s, 2H, 6-, 12-H), 4.17 (s,
4H, NCH2), 4.53 (t, J = 4.5 Hz, 4H, OH), 7.00–7.17 (m, 8H, 2- or 6-
H, 3-, 4-, 5-H of Ph), 7.30–7.50 (m, 4H, 2-, 6-H of pyr), 7.70 (d,
J = 7.5 Hz, 2H, 2- or 6-H of Ph), 8.84 (m, 2H, 3-, 5-H of pyr). Anal.
Calcd for C38H40N4O4: C, 74.00; H, 6.54; N, 9.08. Found: C, 73.66;
H, 6.35; N, 8.84.
(0.4 mL) and the solution was stirred at rt until no more of the
starting compound was detectable by tlc. Then the solvent was re-
moved in vacuum and the resulting oil was taken up in methanol
from which compound 3l crystallized in yellow prisms under cool-
ing, mp 245–252 °C; yield 32%; MS(ESI) m/z 647 (M+H+, 100%). 1H
NMR (DMSO-d6) d 2.90 (s, 4H, 2-, 4-, 8-, 10-H), 3.09 (ABX, J = 10.5,
4.5 Hz, 4H, CHBOH), 3.16 (ABX, J = 10.5, 4.5 Hz, 4H, CHAOH), 3.52 (s,
2H, 6-, 12-H), 4.09 (s, 4H, NCH2), 4.33–4.35 (m, 4H, CH2OH), 6.35
(d, J = 8.6 Hz, 2H, 3-, 5-H of Ph), 6.49 (d, J = 8.2 Hz, 2H, 3-, 5-H of
Ph), 7.03 (d, J = 8.2 Hz, 2H, 2-, 6-H of Ph), 7.23–7.30 (m, 10H, ArH
of Bz), 7.54 (d, J = 8.6 Hz, 2H, 2-, 6-H of Ph), 8.87 (s, 2H, Ph-OH).
Anal. Calcd for C40H42N2O6: C, 74.28; H, 6.55; N, 4.33. Found: C,
74.46; H, 6.23; N, 4.68.
4.1.4.3. 3,9-Dibenzyl-1,5,7,11-tetrakishydroxymethyl-6,12-bis
(4-pyridyl)-3,9-diazahexacyclo[6.4.0.02.7.04.11.05.10]dodecane
(3i). White solid; mp 228–237 °C; yield 55%; MS(ESI) m/z 617
(M+H+, 100%). 1H NMR (DMSO-d6) d 2.94 (s, 4H, 2-, 4-, 8-, 10-H),
3.06 (ABX, J = 10.2, 4.5 Hz, 4H, CHBOH), 3.16 (ABX, J = 10.2, 4.5 Hz,
4H, CHAOH), 3.64 (s, 2H, 6-, 12-H), 4.11 (s, 4H, NCH2), 4.41 (t,
J = 4.5 Hz, 4H, OH), 7.21–7.30 (m, 12H, ArH of Ph, 2- or 6-H of
pyr), 7.76 (d, J = 7.8 Hz, 2H, 2- or 6-H of pyr), 8.28 (d, J = 7.6 Hz,
4H, 3-, 5-H of pyr). Anal. Calcd for C38H40N4O4: C, 74.00; H, 6.54;
N, 9.08. Found: C, 73.76; H, 6.66; N, 9.11.
4.2. Cell culture
TheL5178YmouseTlymphomacelllinewhichwasagiftfromthe
National Cancer Institute (NCI) has been infected with the PHA
mdr-1/A retrovirus as described.29,30 P-gp expressing cells have been
selected by culturing the infected cells in medium which contained
60 ng/mL of colchicine. Both the L5178Y mdr subline and the paren-
tal cell line L5178 Y were grown in McCoys 5A medium with 10%
4.1.4.4. 1,5,7,11-Tetrakishydroxymethyl-3,9-bis(3-picolyl)-6,12-
diphenyl-3,9-diazahexacyclo[6.4.0.02.7.04.11.05.10]dodecane (3j).
White crystals; mp 280–291 °C; yield 68%; MS(ESI) m/z 617
(M+H+, 100%). 1H NMR (DMSO-d6) d 2.95 (s, 4H, 2-, 4-, 8-, 10-H),
3.10 (ABX, J = 10.4, 4.5 Hz, 4H, CHBOH), 3.19 (ABX, J = 10.4, 4.5 Hz,
4H, CHAOH), 3.67 (s, 2H, 6-, 12-H), 4.14 (s, 4H, NCH2), 4.52 (t,
J = 4.5 Hz, 4H, OH), 6.96 (dd, J = 7.2, 5.0 Hz, 2H, 5-H of pyr), 7.04–
7.15 (m, 6H, 3-, 4-, 5-H of Ph), 7.29 (d, J = 7.9 Hz, 2H, 2- or 6-H of
Ph), 7.66 (d, J = 7.2 Hz, 2H, 6-H of pyr), 7.71 (d, J = 8.1 Hz, 2H, 2-
or 6-H of Ph), 8.45 (d, J = 5.0 Hz, 2H, 4-H of pyr), 8.51 (s, 2H, 2-H
of pyr). Anal. Calcd for C38H40N4O4: C, 74.00; H, 6.54; N, 9.08.
Found: C, 73.86; H, 6.52; N, 8.69.
heat-inactivated horse serum, L
-glutamine(2 mM) and antibiotics.31
The A2780P cell line was derived from an ovarian cancer32 and
its subline A2780RCIS was the result of treatment with increasing
amounts of the anticancer agent and MRP2 substrate cisplatin. The
EPG85-257RNOV cell line is a mitoxantrone resistant subcell line
derived from EPG85-257 by induction under exposure to increas-
ing concentrations of mitoxantrone.33
The human gastric carcinoma cell line EPG85-257RNOV, the
ovarian cancer cell line A2780 and the cisplatin-resistant cell line
A2780RCIS were cultivated in Leibovitz L15 medium (Bio Whittak-
er, Verviers, Belgium) supplemented by 10% FCS (Biochrom AG,
Berlin, Germany), 1 mM Ultraglutamine (Bio Whittaker, Verviers,
Belgium), 1.1 g/L NaHCO3, 1% minimal essential vitamins,
0.225 g/L glucose, 80 IE/L insulin (InsumanÒ Rapid, Hoechst Marion
Roussell, München, Germany), 5000 KIE TrasylolÒ (Bayer AG,
Leverkusen, Germany), 2.5 mg/mL transferrin and 6.25 mg/L fetuin
in a humidified atmosphere of 5% CO2 at 37 °C as described.34 The
cell culture medium for EPG85-257RNOV was supplemented with
4.1.4.5. 3,9-Dibenzyl-1,5,7,11-tetrakishydroxymethyl-6,12-bis(3-
hydroxyphenyl)-3,9-diazahexacyclo[6.4.0.02.7.04.11.05.10]dode-
cane (3k). White crystals; mp 179–184 °C; yield 12%; MS(ESI) m/
z 647 (M+H+, 100%). 1H NMR (DMSO-d6) d 2.91 (s, 4H, 2-, 4-, 8-, 10-
H), 3.07–3.10 (m, ABX, 4H, CHBOH), 3.14–3.18 (m, ABX, 4H,
CHAOH), 3.60 (s, 2H, 6-, 12-H), 4.16 (s, 4H, NCH2), 4.32–4.43 (m,
4H, CH2OH), 6.44, 6.51 (2 Â d, J = 8.1 Hz, 2H, 4-H of Ph), 6.71–
6.76 (m, 2H, 5-H of Ph), 7.21 (d, J = 7.6 Hz, 1H, 6-H of Ph), 7.29–
7.32 (m, 11H, ArH of Bz, 2-H of Ph), 7.39 (d, J = 7.6 Hz, 1H, 6-H of
Ph), 7.46 (s, 1H, 2-H of Ph), 8.78, 8.87 (2 Â s, 2H, Ph-OH). Anal.
Calcd for C40H42N2O6: C, 74.28; H, 6.55; N, 4.33. Found: C, 73.97;
H, 6.35; N, 3.99.
mitoxantrone (0.2 lg/mL) to ensure BCRP overexpression and for
stable MRP2 overexpression the medium for A2780RCIS was sup-
plemented with cisplatin (10 ng/mL).
4.3. MDR reversal assays for P-gp, MRP1, MRP2 and BCRP
inhibition
For the reversal of the P-gp-mediated mdr the cultured cells
were adjusted to a concentration of 2 Â 106/mL, then resuspended
in serum free McCoys 5A medium and distributed into 0.5 mL ali-
quots in Eppendorf centrifuge tubes. The test compounds were
added from stock solutions (1.0 mg/mL). The samples were incu-
bated for 10 min at rt. After addition of the P-gp substrate rhoda-
4.1.4.6.
3,9-Bis(4-hydroxybenzyl)-1,5,7,11-tetrakishydroxym-
ethyl-6,12-diphenyl-3,9-diazahexacyclo[6.4.0.02.7.04.11.05.10
]
dodecane (3m). White crystals; mp 241–246 °C; yield 25%;
MS(ESI) m/z 647 (M+H+, 100%). 1H NMR (DMSO-d6) d 2.90 (s, 4H,
2-, 4-, 8-, 10-H), 3.05–3.08 (m, ABX, 4H, CHBOH), 3.14–3.16 (m,
ABX, 4H, CHAOH), 3.65 (s, 2H, 6-, 12-H), 4.00 (s, 4H, NCH2), 4.38–
4.42 (m, 4H, CH2OH), 6.68, (d, J = 8.4 Hz, 4H, 3-, 5-H of Bz), 6.94–
7.01 (m, 2H, 3-, 5-H of Ph), 7.05 (t, J = 7.2 Hz, 2H, 4-H of Ph),
7.09–7.14 (m, 6H, 2-, 6-H of Bz, 3-, 5-H of Ph), 7.24 (d, J = 7.0 Hz,
2H, 2-, 6-H of Ph), 7.83 (d, J = 7.6 Hz, 2H, 2-, 6-H of Ph), 9.17 (s,
2H, Ph-OH). Anal. Calcd for C40H42N2O6: C, 74.28; H, 6.55; N,
4.33. Found: C, 74.33; H, 6.48; N, 4.14.
mine 123 (5.2 lM final concentration) the cells were incubated
for additional 20 min at 37 °C. Then they were washed twice and
resuspended in 0.5 mL phosphate-buffered saline (PBS). For analy-
sis the fluorescence of 1 Â 104 cells was measured by flow cytom-
etry with a Becton Dickinson FACScan instrument.
For the reversal of the MRP1- and MRP2-mediated mdr, A2780
cells were adjusted to a concentration of 1 Â 105 cells per mL and
seeded out in 6-well plates to a concentration of 2 Â 105 cells per
mL. The medium was removed after 24 h and fresh medium con-
taining several concentrations of tested compounds was added.
After an incubation time of 20 min at 37 °C carboxyfluoresceindi-
4.1.5. 3,9-Dibenzyl-1,5,7,11-tetrakishydroxymethyl-6,12-bis(4-
hydroxyphenyl)-3,9-diazahexacyclo[6.4.0.02.7.04.11.05.10]dode-
cane (3l)
Cage dimer 3d (20 mg, 0.02 mmol) was dissolved in a mixture
of thioanisole (0.12 mL, 0.1 mmol) and trifluoroacetic acid
acetat (2
lM final concentration) as a fluorescent substrate for
MRP1 and MRP2 was added and cells were incubated for further

4990
C. Coburger et al. / Bioorg. Med. Chem. 18 (2010) 4983–4990
40 min at 37 °C. Next, cells were washed for two times with icecold
PBS, resuspended in 0.5 mL icecold PBS and stored on ice until
measurement.
The BCRP-overexpressing EPG85-257RNOV cells were adjusted
to a concentration of 0.5 Â 106 cells per mL, distributed into
0.5 mL aliquots in Eppendorf centrifuge tubes and test compounds
were added in several concentrations. After an incubation time of
References and notes
1. Gottesman, M. M.; Ling, V. FEBS Lett. 2006, 580, 998.
2. Lage, H. Cell. Mol. Life Sci. 2008, 65, 3145.
3. Livingston, D. M.; Silver, D. P. Nature 2008, 451, 1066.
4. Mahon, F. X.; Belloc, F.; Lagarde, V.; Chollet, C.; Moreau-Gaudry, F.; Reiffers, J.;
Goldman, J. M.; Melo, J. V. Blood 2003, 101, 2368.
5. Matsui, H.; Takeshita, A.; Naito, K.; Shinjo, K.; Shigeno, K.; Maekawa, M.;
Yamakaw, Y.; Tanimito, M.; Kobayashi, M.; Ohnishsi, K.; Ohno, R. Leukemia
2002, 16, 813.
6. Bonavida, B. Oncogene 2007, 26, 3629.
7. Huang, Y.; Sadee, W. Cancer Lett. 2006, 239, 168.
8. Choi, C.-H. Cancer Cell Int. 2005, 5, 30.
9. Sharom, F. J. Pharmacogenomics 2008, 9, 105.
10. Avendanó, C.; Menéndez, J. C. Med. Chem. Rev. 2004, 1, 419.
11. Baumert, C.; Hilgeroth, A. Anticancer Agents Med. Chem. 2009, 9, 415.
12. Sheps, J. A.; Ling, V. Eur. J. Physiol. 2007, 453, 545.
13. Dean, M.; Rzhetsky, A.; Allikments, R. Genome Res. 2009, 11, 1156.
14. Leschziner, G. D.; Andrew, T.; Pirnohamed, M.; Johnson, M. R.
Pharmacogenomics J. 2007, 7, 154.
15. Tain, Q.; Zhang, J.; Chan, E.; Duan, W.; Zhou, S. Drug Dev. Res. 2005, 64, 1.
16. Pires, M. M.; Emmert, D.; Hrycyna, A. Ch.; Chmielewski, J. Mol. Pharmacol. 2009,
75, 92.
17. Sauna, Z. E.; Andrus, M. B.; Turner, T. M.; Ambudkar, S. V. Biochemistry 2004, 43,
2262.
18. Hilgeroth, A. Mini-Rev. Med. Chem. 2002, 2, 235.
19. Coburger, C.; Wollmann, J.; Baumert, C.; Krug, M.; Molnár, J.; Lage, H.;
Hilgeroth, A. J. Med. Chem. 2008, 51, 5871.
20. Seelig, A. Eur. J. Biochem. 1998, 251, 252.
21. Ferte, J.; Kühnel, J.-M.; Chapuis, G.; Rolland, Y.; Lewin, G.; Schwaller, M. A. J.
Med. Chem. 1999, 42, 478.
22. Krug, M.; Voigt, B.; Baumert, Ch.; Lüpken, R.; Molnár, J.; Hilgeroth, A. Eur. J.
Med. Chem. 2010, 45, 2683.
23. Ferry, D.; Boer, R.; Callaghan, R.; Ulrich, W. R. Int. J. Clin. Pharmacol. Ther. 2000,
38, 130.
24. Dhainaut, A.; Regnier, G.; Tizot, A.; Pierre, A.; Leonce, S.; Guilbaud, N.; Kraus-
Berthier, L.; Atassi, G. J. Med. Chem. 1996, 39, 4099.
25. Wiese, M.; Pajeva, I. K. Curr. Med. Chem. 2001, 8, 685.
26. Shuhong, Z.; Morris, M. E.. In Drug Delivery and Applications; Wang, B., Siahaan,
T., Soltero, R., Eds.; John Wiley & Sons, 2005; Vol. 1, pp 381–411.
27. Diestra, J. E.; Scheffer, G. L.; Català, I.; Maliepaard, M.; Schellens, J. H.; Scheper,
R. J.; Germà-Liuch, J. R.; Izquierdo, M. A. J. Pathol. 2002, 198, 213.
28. Chennat, T.; Eisner, U. J. Chem. Soc., Perkin Trans. 1 1975, 10, 926.
29. Weaver, J. L.; Szabo, D.; Pine, P. S.; Gottesmann, S.; Goldberg, A.; Aszalos, A. Int.
J. Cancer 1993, 54, 456.
20 min at 37 °C mitoxantrone (20 lM final concentration) was
added and cells were incubated for further 60 min at 37 °C. Next,
medium was discarded and the cells were trypsinized, washed
with icecold PBS for two times and resuspended in 0.5 mL icecold
PBS for analysis.
Fluorescence of carboxyfluorescein and mitoxantrone was ana-
lysed with a Becton Dickinson FacsCalibur flow cytometer that was
equipped with an argon laser. The fluorescence of 10,000 events
was logarithmically measured at a laser excitation wavelength of
488 nm through a 530 nm (carboxyfluorescein) and a 670 nm
(mitoxantrone) bandpass filter.
Results of flow cytometry were calculated as R values. In the
case of MRP1 these calculated R values each result from dividing
the measured fluorescence intensity in the inhibitor-treated cells
by the measured fluorescence intensity in the inhibitor-untreated
cells. Thus the R value means the quotient of the fluorescence
intensity in the inhibitor-treated cells and the fluorescence in the
inhibitor-untreated cells. In the cases of P-gp and BCRP the result-
ing quotient has been divided by the quotient of the measured
fluorescence intensities in the inhibitor-treated parental cell line
and the untreated parental cell line. In the case of MRP2 the R val-
ues were given by a substraction of the value of the MRP1 specific
extent of fluorescence uptake into the parental cell line from the
whole extent of fluorescence determined as MRP1 and MRP2 inhi-
bition in the induced cell line A2780RCIS.
4.4. Log P value calculations
30. Cornwell, M. M.; Pastan, I.; Gottesmann, M. M. J. Biol. Chem. 1987, 262,
2166.
31. Voigt, B.; Coburger, C.; Molnár, J.; Hilgeroth, A. Bioorg. Med. Chem. 2007, 15,
5110.
32. Eva, A.; Robbins, K. C.; Andersen, P. R.; Srinivasan, A.; Tronick, S. R.; Reddy, E. P.;
Ellmore, N. W.; Galen, A. T.; Lautenberger, J. A.; Papas, T. S.; Westin, E. H.;
Wong-Staal, F.; Gallo, R. C.; Aaronson, S. A. Nature 1982, 295, 116.
33. Dietel, M.; Arps, H.; Lage, H.; Niendorf, A. Cancer Res. 1990, 50, 6100.
34. Lage, H.; Jordan, A.; Scholz, R.; Dietel, M. Int. J. Hyperthermia 2000, 16, 291.
35. http://www.vcclab.org/lab/alogps.
Molecules were constructed and log P values were determined
with nine programs (ALOGPs, AC logP, AB/LogP, miLogP, ALOGP,
MLOGP, KOWWIN, XLOGP2 and XLOGP3).35 The resulting mean
values were given by the ALOGPS 2.1 program.35
Acknowledgements
The authors gratefully acknowledge the financial support of
their work by the EU within the COST B16 action, the BMBF and
the DFG.