Structure-activity relationship studies on benzofuran analogs of propafenone-type modulators of tumor cell multidrug resistance

Article abstract of DOI:10.1021/jm960384x

A series of benzofurylethanolamine analogs of propafenone (1a) have been prepared and evaluated for multidrug resistance-reversing activity in two in vitro assay systems. As for propafenones, an excellent correlation of biological data with calculated lipophilicity values was found for benzofurans, whereby the latter generally had lower activity/lipophilicity ratios. Almost identical slopes of the regression lines were obtained for both propafenones and benzofurans. Multiple linear regression analysis of the complete data set yielded an equation with excellent predictive power (r²(cross-valid) = 0.968). Interaction measurements with artificial membranes indicated that the differences in activity between these two series of compounds are not due to differences in the interaction pattern with biological membranes.

Full text of DOI:10.1021/jm960384x

J . Med. Chem. 1996, 39, 4767-4774
4767
Str u ctu r e-Activity Rela tion sh ip Stu d ies on Ben zofu r a n An a logs of
P r op a fen on e-Typ e Mod u la tor s of Tu m or Cell Mu ltid r u g Resista n ce
Gerhard Ecker,*,† Peter Chiba,‡ Manuela Hitzler,‡ Diethard Schmid,‡ Klaus Visser,§ Hans Peter Cordes,§
J osef Cso¨llei,†,| J oachim K. Seydel,§ and Klaus-J u¨rgen Schaper§
Institute of Pharmaceutical Chemistry, University of Vienna, Althanstrasse 14, A-1090 Wien, Austria,
Institute of Medical Chemistry, University of Vienna, Waehringerstrasse 10, A-1090 Wien, Austria,
and Medicinal and Pharmaceutical Chemistry, Borstel Research Center, D-23845 Borstel, Germany
Received May 28, 1996X
A series of benzofurylethanolamine analogs of propafenone (1a ) have been prepared and
evaluated for multidrug resistance-reversing activity in two in vitro assay systems. As for
propafenones, an excellent correlation of biological data with calculated lipophilicity values
was found for benzofurans, whereby the latter generally had lower activity/lipophilicity ratios.
Almost identical slopes of the regression lines were obtained for both propafenones and
benzofurans. Multiple linear regression analysis of the complete data set yielded an equation
with excellent predictive power (r2
membranes indicated that the differences in activity between these two series of compounds
are not due to differences in the interaction pattern with biological membranes.
) 0.968). Interaction measurements with artificial
cross-valid
In tr od u ction
designed highly active modulators with limited side
effects are urgently required.
The development of multiple drug resistance repre-
sents an increasing problem in cancer treatment as well
as in antimicrobial therapy. Within the past decade
several mechanisms of pleiotropic drug resistance of
tumor cells have been identified.1 One type of multi-
drug resistance (MDR) has been shown to be mediated
by an energy dependent, membrane-bound efflux pump
termed P-glycoprotein (PGP).2 PGP represents a mem-
ber of the ATP-binding casette3 with low substrate
specificity. A broad range of cytostatic drugs such as
anthracyclines, epipodophyllotoxins, actinomycin D,
vinca alkaloids, colchicines, and taxol are eliminated via
PGP-mediated efflux.4 Recently, this mechanism of
drug resistance has attracted additional attention since
it was proposed to be involved in multidrug resistance
of Gram-negative bacteria5 as well as in fungi6-8 and
host-mediated resistance of Mycobacterium tuberculosis
against tuberculostatic drugs.9 Within the past few
years a variety of substances have been shown to inhibit
PGP-mediated drug efflux and thereby reestablish
sensitivity toward chemotherapeutic agents.10 These
include ion channel blockers such as verapamil,11 amio-
darone,12 propafenone,13 and some dihydropyridines,14
antipsychotic drugs like phenothiazines15 and thioxan-
thenes,16 cyclosporines,17 and some recently identified
compounds like the triazinopiperidine S 978818 and the
acridone carboxamide GF 120918.19 Preliminary results
obtained in clinical studies clearly demonstrate that
modulation of MDR might be a successful approach in
hematological malignancies, but serious side effects
often preclude optimal dosage of modulators.20 These
side effects are due to the modulator’s inherent phar-
macological effects including cardiac effects, immuno-
suppression, and nephrotoxicity. Therefore, specifically
Although the proposed mechanism of action of MDR
modulators is inhibition of PGP, little is known about
structure-activity relationships (SAR) of these com-
pounds.21 Despite considerable structural heterogene-
ity, most of the modulators share the properties of high
lipophilicity and a basic nitrogen atom. In addition,
evidence has been published that amphiphilic com-
pounds may also influence membrane surface density
and viscosity.22 Therefore, it seems likely that both
direct protein and indirect lipid interactions influence
the activity of PGP inhibitors.23
We recently identified a series of analogs of pro-
pafenone as effective inhibitors of PGP.24 Propafenone
is in clinical use as an antiarrhythmic agent due to its
ability to block cardiac sodium channels. The substance
also has weak â-adrenoreceptor-blocking activity.
Rhodamine-123 as well as daunomycin efflux studies
on a series of closely related structural homologs of
propafenone showed a highly significant correlation
between lipophilicity and MDR-reversing ability. Nev-
ertheless, altering the acyl side chain by reducing the
carbonyl group or changing the substituent’s position
from ortho to meta or para led to a decrease in activity,
for which lipophilicity was no longer the only determi-
nant.25
A series of benzofuran analogs with diminished flex-
ibility were synthesized and tested for their chemosen-
sitizing potency in both rhodamine-123 efflux studies
and daunomycin cytotoxicity assays in order to gain
further insights into structural features required for
good PGP-inhibitory activity of propafenone-type MDR
modulators. Additionally, membrane interaction mea-
surements of selected compounds with lipid vesicles
were performed using nuclear magnetic resonance
(NMR) spectroscopy or differential scanning calorimetry
(DSC) in order to determine which regions of the
molecules interact with biological membranes.
* Author to whom correspondence should be adressed. Tel: +43-1-
31336-8571. Fax: +43-1-31336-771. E-mail: ecker@speedy.pch.univie.
ac.at.
† Institute of Pharmaceutical Chemistry.
‡ Institute of Medical Chemistry.
Ch em istr y
§ Borstel Research Center.
| Present adress: Ustav Chemickych Leciv, Veterinarni a Farma-
ceuticka Univerzita, 61242 Brno, Czech Republic.
X Abstract published in Advance ACS Abstracts, October 15, 1996.
Propafenone analogs 1a -n were synthesized as pre-
viously described.25 The benzofurans 2a ,b as well as
S0022-2623(96)00384-6 CCC: $12.00 © 1996 American Chemical Society

4768 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 24
Sch em e 1. Synthesis of Benzofurans 2a ,ba
Ecker et al.
and ME is the modulator efficacy.
ME‚c
y ) yi -
(1)
ED50 + c
Chemical structures and ED50 values of all compounds
are given in Table 1.
2. Cell Cytotoxicity Assa y. Although inhibition of
rhodamine-123 efflux is the more direct method for
measuring interaction with PGP, MTT-based cytotox-
icity assays represent a more general assay for the
modulation of MDR, which additionally accounts for
intracellular metabolism of modulators. Therefore, we
measured the chemosensitizing effect of the propafenone-
type modulators with respect to daunomycin using
CCRF-CEM vcr1000 cells. Data were processed using
the previously established method of combined simul-
taneous analysis of sigmoidal dose-response curve
families, which leads to highly accurate ED50 values of
modulators33 (Table 1). Correlation with data obtained
in the rhodamine-123 efflux studies gave an r value of
0.946 (Figure 1). Generally, cytotoxicity of all com-
pounds tested did not exceed 20% at the highest
modulator concentration used (10 µM).
Mem b r a n e In t er a ct ion Mea su r em en t s. 1.
Nu clea r Ma gn etic Reson a n ce Exp er im en ts. To
investigate the influence of membrane lipids on selected
modulators, spin-spin relaxation times in the presence
of different amounts of lecithin liposomes were mea-
sured as described previously.22 Interaction strength
is represented by the slope of the regression line
obtained by plotting line broadening vs lecithin concen-
tration (Figure 2) and is listed in Table 2. Within the
series of o-[(acylaryl)oxy]propanolamines, the strongest
interaction was observed with the -O-CH2- group and
somewhat less with protons of the central phenyl ring.
In all cases almost no interaction between lecithin
and the phenylethyl group was observed. Within the
series of benzofuran analogs the deshydroxy derivative
2d showed strong interaction with lecithin, indicat-
ing that the entire (phenylethyl)benzofuryl system is
involved. In contrast, the propafenone analogous
compound 2b showed an interaction pattern similar
to 1a with the strongest interaction near the C-OH
group.
a
(i) Ac2O, BF3; (ii) Br2/AlCl3; (iii) LiAlH4; (iv) dihydropyran,
p-TSA; (v) amine; (vi) etheral HCl.
the enantiomers of 2b were synthesized as outlined in
Scheme 1. Thus, 3-alkylbenzofuran 3 (which can be
prepared in three steps from appropriate o-hydroxyphe-
nones)26 was acetylated under Friedel-Crafts conditions
to give 2-acetyl-3-alkylbenzofuran 4, which was bromi-
nated with Br2 and reduced with NaBH4 to yield the
bromohydrin 5. Subsequent protection of the hydroxy
group with dihydropyran, reaction with an appropriate
amine, and cleavage of the protecting group with
hydrochloric acid directly yielded the hydrochlorides of
2a ,27b28 in fair overall yields. The enantiomers of 2b
were prepared by using mbe-lactole instead of dihydro-
pyran followed by chromatographic resolution of the
diastereoisomers.29
The deshydroxy analogs 2d ,e were prepared as de-
scribed recently30 (Scheme 2). o-Hydroxy-3-phenylpro-
piophenone was alkylated with epichlorohydrin and the
epoxide ring opened by reaction with hydrochloric acid
to give the chlorohydrin 6. Oxidation with oxalyl
chloride/DMSO and silica gel-mediated cyclization gave
the dihydrobenzofuran 7. Reduction with triethylsilane/
BF3‚Et2O yielded the 2-(chloroalkyl)benzofuran 8, which
was reacted with n-propyl- or isopropylamine to give the
desired deshydroxy analogs 2d ,e. Synthesis of the
arylpiperazine analog 2f was achieved via acidic dehy-
dration of 7 followed by reaction with (o-methoxyphe-
nyl)piperazine. Subsequent reduction of 2f with NaBH4
yielded 2c. Alternatively, 2f can also be prepared by
reaction of the corresponding 2-(bromoacetyl)benzofuran
(Scheme 1) with (o-methoxyphenyl)piperazine. Never-
theless, the synthesis outlined in Scheme 2 generally
gave higher overall yields.
2. Differ en tia l Sca n n in g Ca lor im etr y. Artificial
dipalmitoylphosphatidylcholine (DPPC) membranes were
used to investigate the influence of selected modulators
on membrane lipids. Changes in phase transition
properties in the presence of different concentrations
of modulators were measured according to Pajeva et al.22
Generally, the DPPC curve characteristics showed a
broadening of the main peak and a shift of Tmax and
Ttrans to lower temperatures. Peaks were analyzed by
a nonlinear least-squares fit using a Gaussian function.
The peak width in °C of a molar DPPC to modifier
concentration of 1:0.05 was taken as a measure of
interaction strength and is presented in Table 3. These
values closely correlate with the lipophilicity of the
modulators (Figure 3; r ) 0.997, n ) 6).
MDR -Mod u la t in g Act ivit y. 1. R h od a m in e-123
Efflu x Stu d ies. The rhodamine-123 assay is a well-
documented, direct, and reproducible functional assay
for measuring PGP dependent efflux.31 We therefore
measured the ability of our compounds to inhibit PGP-
mediated rhodamine-123 efflux in the resistant T-
lymphoblast cell line CCRF-CEM vcr1000.32 As shown
previously, similar results can be obtained by measuring
either rhodamine-123 or daunomycin efflux.25 The time
dependent linear decrease in mean fluorescence of cells
was determined in the presence of various concentra-
tions of modifier, and the initial efflux rates were
calculated by linear regression analysis. Correction for
simple diffusion was achieved by subtracting the efflux
rates observed in the parental line. ED50 values of all
modifiers were obtained from dose-response curves of
initial efflux rate vs modifier concentration. Data points
of at least two independently performed experiments
were fitted according to eq 1, where y is the initial efflux
rate determined as a function of modifier concentration
c, yi is the initial efflux rate in the absence of modulator,
3. Deter m in a tion of Lip op h ilicity. The log P
values were calculated according to the method of Ghose
et al.34 using the software package MOLGEN.35 As
previously shown, these values are in excellent agree-
ment with those obtained experimentally by an HPLC
method.25

SAR of Benzofuran Analogs of Propafenone
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 24 4769
Sch em e 2. Synthesis of Benzofurans 2c-fa
a
(i) Epichlorohydrin, NaOH; (ii) concentrated HCl; (iii) C2O2Cl2, DMSO; (iv) silica gel; (v) triethylsilane/BF3; (vi) amine; (vii) H3PO4;
(viii) amine, MeOH; (ix) NaBH4.
Resu lts a n d Discu ssion
The measured activity of the deshydroxy analogs 2d ,e
correlated well with the expected values. Thus, the
hydroxyl group does not seem to be involved in interac-
tion with PGP. Slight differences in activity of the
enantiomers of 2b in cytotoxicity assays (1.83 ( 0.13
for (S)-2b and 1.03 ( 0.01 for (R)-2b) were not found in
rhodamine-123 efflux studies. Since incubation times
are considerably longer in cytotoxicity assays (72 h) than
those of efflux studies (2 min), the slightly lower activity
of (S)-2b in cytotoxicity assays might be due to ste-
reospecific cellular metabolism of the compounds.
Previous experiments showed that propafenones car-
rying a carbonyl group are more active than the corre-
sponding hydroxy and methoxy derivatives. Therefore,
the hydroxyl group of the benzofuran 2c was converted
to a carbonyl group. An increase in the activity/
lipophilicity ratio was, however, not observed. Con-
versely, a decrease in activity in relation to the expected
value by a factor of 4 was noted. These data suggest
that the distance between the carbonyl group and the
basic nitrogen atom might be crucial, which agrees with
results previously obtained for propafenone analogs.25
Two desphenyl analogs (1m and 2a ) were synthesized
and tested in order to evaluate whether the benzyl group
influences activity by direct interaction with PGP. The
results demonstrate that the phenylethyl group influ-
ences activity mainly via its contribution to overall
lipophilicity of propafenones and benzofurans and prob-
ably not via π-π interaction of the phenyl ring with
aromatic amino acids of PGP.
As recently demonstrated an excellent correlation
between lipophilicity and PGP inhibition was found for
a homologous series of propafenone analogs. Using an
MDR1-transfected cell line, we showed that inhibition
of PGP is indeed the main mechanism of action of
propafenone-type MDR modulators. SAR studies indi-
cated that a hydrogen bond acceptor near the central
aromatic ring moiety seems to be crucial for maintaining
high MDR-modulating activity.25
In this paper we extend our studies to a series of
analogous, conformationally restricted benzofuran de-
rivatives. Pharmacological activity was determined
both in rhodamine-123 efflux studies and in daunomycin
cytotoxicity assays. The two assay systems correlate
excellently, indicating that the cellular interaction site
for benzofurans is indeed PGP. Figure 4 shows the
correlation of calculated lipophilicity values and the
chemosensitizing activity determined in rhodamine-123
efflux assays. Two regression lines were obtained for
propafenones (log(1/ED50) ) 0.82((0.03) log P - 3.21-
((0.11); r ) 0.994, n ) 13) and benzofurans (log(1/ED50)
) 0.96((0.10) log P - 4.94((0.46); r ) 0.982, n ) 5).
Both lines were virtually parallel. Multiple linear
regression analysis using an indicator variable for
benzofurans (IBf ) 1, else IBf ) 0) resulted in the
following, statistically highly significant, equation:
log(1/ED50) ) 0.86((0.03) log P - 1.16((0.06)IBf
-
3.33((0.13)
It has been proposed in the literature, that MDR-
modulating activity of amphiphilic drugs depends on
their ability to interact with membrane phospholipids.22
NMR and DSC were employed to characterize the
interaction of selected compounds with artificial lipid
membranes. NMR-interaction measurements with
lecithin liposomes showed that propafenone and ben-
zofuran analogs bearing a hydroxyl group only interact
in the vicinity of this substructure, whereas the deshy-
droxy benzofuran 2d showed strong interaction over the
whole (phenylethyl)benzofuryl region. However, this
difference in the interaction pattern of various benzo-
furans with phospholipid vesicles had no influence on
r ) 0.990, n ) 18, r2
) 0.968
cross-valid
Figure 5 shows a plot of predicted vs corresponding
observed log potency values (r ) 0.990). Analogous
results were obtained in cytotoxicity assays (pro-
pafenones, r ) 0.964; benzofurans, r ) 0.961; r (multiple
linear regression) ) 0.971). Generally, incorporation of
the ether oxygen of propafenone into a benzofuran
moiety results in a remarkable decrease in activity.
Since the two regression lines are parallel, this loss of
activity is compensated for in part by increasing lipo-
philicity. 2c is, for example, equipotent to 1e.

4770 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 24
Ecker et al.
Ta ble 1. Chemical Structure, Calculated Lipophilicity, and
Pharmacologic Activity of Compounds 1a -n and 2a -f
F igu r e 1. Correlation of ED50 values of modulators obtained
in rhodamine-123 efflux assays vs those obtained in cytotox-
icity assays (MTT); r ) 0.946, n ) 19.
F igu r e 2. Correlation of line broadening (expressed as peak
width in Hz) of the ArO-CH2- signal in 1h vs concentration of
lecithin liposomes (in mg/mL).
Ta ble 2. Interaction Strength with Lecithin Vesicles (n 1/2),
Calculated log P, and MDR-Modulating Activity of Selected
Propafenone Analogs
n 1/2a
log P
ED50
b
compd
1a
1d
1e
1h
1i
42.34
35.34
39.50
52.40
63.42
3.36
3.62
3.67
4.25
2.54
3.55
1.40
0.61
1.65
13.56
a
n 1/2 is represented as the slope of the correlation of peak
width of the ArO-CH2- signal vs concentration of lecithin liposomes
b
(see also Figure 2). ED50 values obtained in Rh123 efflux assays
are presented.
flexibility in comparison to propafenones or to the lack
of an appropriate hydrogen-bond acceptor.
Results of DSC-interaction measurements between
modulators and phosphatidylcholine vesicles closely
correlate with calculated lipophilicity values of both
propafenones and benzofurans. This is shown for
selected compounds in Figure 3. Data indicate that
within the series of compounds tested, calculated lipo-
philicity of the molecules seems to be a good predictor
of partitioning into uncharged membranes. A close
correlation between DSC data and biological activity is
found within both homologous series. Comparison of
the nearly equilipophilic compounds 1j (propafenone
analog) and 2d (benzofuran derivative) shows that in
this series of compounds the DSC method can, how-
a
Data points of at least two independently performed experi-
ments were used to determine the ED50 values. Generally,
b
interexperimental variation was below 10%. Data points were
determined in triplicate as described previously;24 interexperi-
mental variation was below 15%.
their MDR-modulating activity. Thus, the decrease in
activity of benzofurans may be related to a decrease in

SAR of Benzofuran Analogs of Propafenone
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 24 4771
Ta ble 3. Interaction Strength with Phosphatidylcholine
Vesicles (b 1/2), Calculated log P, and MDR-Modulating
Activity of Selected Compounds
b
compd
b 1/2a
log P
ED50
1.65
1h
1i
1j
1n
2b
2d
1.64
1.18
2.08
0.67
1.83
1.95
3.67
2.54
4.93
0.94
4.07
4.71
13.56
0.16
na
13.61
2.90
a
b 1/2 is represented as the peak width in °C of the main
transition peak at a molar DPPC to modifier concentration of
1:0.05. b ED50 values obtained in Rh123 efflux assays are presented.
F igu r e 5. Correlation of predicted vs observed MDR-
modulating activity (expressed as log(1/ED50) of modulators
in Rh123 efflux assays) using the equation: log(1/ED50) ) 0.86
log P - 1.16IBf - 3.33; r ) 0.990, n ) 18.
tions were obtained both within the series of pro-
pafenone analogs as well as within the series of benzo-
furans with the regression lines being virtually parallel.
Multiple linear regression analysis of combined sets
leads to an equation with high predictive power
(r2
) 0.968). Thus, incorporation of the ether
cross-valid
oxygen of propafenone into a benzofuran moiety and
simultaneous removal of the carbonyl group leads to a
remarkable decrease in activity, whereby lipophilicity
retains its predictive character. Results obtained for
the enantiomers of 2b and for the deshydroxy analogs
2d ,e show that the hydroxyl group does not seem to
contribute to PGP interaction. Although it has been
previously reported that a carbonyl group near to an
aromatic system seems to be important for MDR-
modulating activity, oxidation of the hydroxyl group led
to a decrease of activity. This suggests that the distance
between the carbonyl group and the basic nitrogen
seems of utmost importance. Synthesis of desphenyl
derivatives of propafenones as well as benzofurans
demonstrates that the phenyl ring contributes to phar-
macological activity only by influencing lipophilicity.
Membrane interaction measurements using NMR spec-
troscopy show that propafenone analogs as well as
benzofurylethanolamines mainly interact within the
region of the aryl ether oxygen, whereas the deshydroxy
benzofuran derivative 2d exhibits a strong interaction
over the whole (phenylethyl)benzofuryl region. Never-
theless, this difference was not reflected in the MDR-
modulating ability of the compounds. DSC measure-
ments showed that the strength of interaction with
phosphatidylcholine, the main plasma membrane com-
ponent, is a good predictor for the lipophilicity of the
molecules. Thus, for this set of compounds, calculated
lipophilicity of the molecules seems to correspond to
their membrane partitioning coefficients. Further mem-
brane interaction studies, employing fluorescein leakage
assays and fluorescence spectroscopy, are currently in
progress.
F igu r e 3. Correlation of interaction strength of selected
compounds with phosphatidyl vesicles (expressed as peak
width in °C at a molar DPPC to modifier concentration of
1:0.05) vs calculated log P values; r ) 0.997; n ) 6.
F igu r e 4. Correlation of MDR-modulating activity (expressed
as log(1/ED50) values of modulators determined in Rh123 efflux
assays) and calculated log P values: (9) propafenones 1a -n ,
(O) benzofurans 2a -e, and (2) 2f.
ever, not be used to predict activity for nonhomologous
molecules.
Con clu sion s
A series of benzofurylethanolamine analogs of the
class Ic antiarrhythmic agent propafenone was synthe-
sized. These substances were evaluated as modulators
of multidrug resistance using both a cytotoxicity assay
and rhodamine-123 efflux studies. Lipophilicity of the
molecules was calculated and correlated to their
chemosensitizing activity. Highly significant correla-
Exp er im en ta l Section
Ch em istr y. Melting points were determined on a Kofler
melting point apparatus and are uncorrected. Infrared spectra
were recorded as thin films on salt disks on a Perkin Elmer
298 spectrophotometer. Mass spectra were performed on a
Shimadzu QP 1000 spectrometer by G. Reznicek (Institut fu¨r
Pharmakognosie, University of Vienna, Vienna, Austria). GC/

4772 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 24
Ecker et al.
MS spectra were performed by L. J irovetz (Institut fu¨r
Pharmazeutische Chemie, University of Vienna, Vienna,
Austria) on an HP-5890A GC equipped with an HP-5970 MSD
and a 59970 ChemStation data system. NMR spectra were
recorded on a Bruker AC 80 spectrometer and a Varian Unity
plus 300 system, using tetramethylsilane as internal standard.
Microanalyses were done by J . Theiner (Institut fu¨r Phys-
ikalische Chemie, University of Vienna, Vienna, Austria).
Satisfactory C, H, N, and Cl analyses ((0.4%) were obtained
for all hydrochlorides.
CH2-CH2-, -CH2-N-, piperazine CH2), 3.88 (s, 3H, -OCH3), 4.65
(m, 1H, -CH(O)), 6.85-7.55 (m, 13H, arom H); 13C NMR
(chloroform-d) δ 26.89, 35.88 (Ph-CH2-CH2-), 50.74, 53.03
(piperazine CH2), 55.35 (-OCH3), 61.00 (-CH2-N-), 61.14 (-CH-
(OH)), 111.17, 111.35, 116.78, 118.22, 119.46, 120.99, 122.28,
123.07, 124.30, 126.19, 128.41, 128.65, 128.79, 141.08, 141.40,
151.53, 152.22, 154.29 (arom C); MS (70 eV) 456 (M+, 0.2),
205 (100), 190 (18), 70 (71). Anal. (C29H32N2O3) C, H, N.
2c h yd r och lor id e: mp 167-170 °C. Anal. (C29H34N2O3-
Cl2) C, H, N, Cl.
Gen er a l P r oced u r e for th e Syn th esis of th e P r o-
pafen on e An alogs 1f,n . An appropriate epoxide (17.7 mmol)21
was dissolved in 20-30 mL of amine and refluxed for 6 h. The
mixture was evaporated to dryness and the oily residue
purified via column chromatography (silica gel, CH2Cl2/
methanol/concentrated NH4OH, 200/10/1). Formation of the
hydrochlorides was carried out by dissolving the amine in
ethyl acetate and adding a 1 M solution of HCl in diethyl ether.
The hydrochloride was filtered off and purified via crystal-
lization.
P h a r m a cology. Cell Lin es a n d Cu ltu r e Con d ition s.
The CCRF-CEM T-lymphoblast cell line as well as the
resistant line were obtained as described previously.25 Cells
were kept in RPMI1640 medium supplemented with 10%
fetal calf serum under standard culture conditions. The
resistant CCRF vcr1000 cell line was kept in the continuous
presence of 1000 ng/mL vincristine. The selecting agent was
washed out at least 1 week prior to the experiments. PGP
expression was shown to be stable for at least 1 month after
washout of the selective agent as shown by flow cytometry
using the MRK16 antibody (Behring Institut GesmbH, Vienna,
Austria) and by cytotoxicity and efflux experiments (data not
shown). The cell line used in our studies was selected in the
presence of increasing doses of vincristine without prior
mutagenization. This cell line has been chosen on the basis
of distinct PGP expression and does not show the mutation at
codon 185. In addition, no significant contribution of other
factors to MDR could be observed (V. Gekeler, unpublished
data).
MTT Assa y. The assay is dependent on the cellular
reduction of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltet-
razolium bromide; Sigma Chemical Co., St. Louis, MO) in
mitochondria of viable cells to water insoluble formazan. The
assays were performed in 96-well plates essentially as de-
scribed by Mosmann,36 with the exception that water insoluble
formazan granules were dissolved in 2-propanol containing
0.04 N HCl. Absorbance was read spectrophotometrically
using an EL311 Biotek microtiter plate reader (Biotek Instru-
ments Inc., Highland Park, VT). Data points for different
modulator concentrations are corrected for modulator toxicity,
which generally did not exceed 20% at 10 µM.
Rh od a m in e Efflu x Stu d ies. Rhodamine efflux studies
were performed in a modification of published methods.31 Cells
were pelleted, the supernatant was removed by suction, and
the cells were resuspended at a density of 1 × 106/mL in
RPMI1640 medium containing rhodamine-123 (Sigma Chemi-
cal Co., St. Louis, MO) at a final concentration of 0.2 µg/mL.
Cell suspensions were incubated at 37 °C for 15 min. Tubes
were chilled on ice and pelleted at 500g in an Eppendorf 5403
centrifuge (Eppendorf, Germany). Supernatants were re-
moved, and the cell pellet was resuspended in medium which
was prewarmed to 37 °C and contained either no modulator
or chemosensitizer at various concentrations ranging from 16
nM to 500 µM, depending on solubility and expected potency
of the modifier. Eight concentrations (serial dilution 1:2.5)
were tested for each modulator. After 30, 60, 90, and 120 s
aliquots of the incubation mixture were transferred to tubes
containing an equal volume of ice cold stop solution (RPMI1640
medium containing verapamil at a final concentration of 10
µg/mL); 0 time points were done by immediately pipetting
Rh123-preloaded cells into ice cold stop solution. Parental
CCRF-CEM cells were used as controls for simple plasma
membrane diffusion, whereby initial Rh123 fluorescence levels
were adjusted to be equal to initial levels observed in resistant
cells. Samples drawn at the respective time points were kept
in an ice water bath and measured within 1 h on a Becton
Dickinson facscalibur flow cytometer (Becton Dickinson, Vi-
enna, Austria). Viable cells were gated on the basis of forward
and side scatter; 5000 gated events were accumulated for
the determination of mean fluorescence values. Time depend-
ent decrease in mean fluorescence values was linear over
time for at least 2 min and is expressed as the percentage
of 0 time points, to allow comparison of independent experi-
ments.
1-[2-[3-(Ben zyla m in o)-2-h yd r oxyp r op oxy]p h en yl]-3-
p h en yl-1-p r op a n on e (1f): yield 65%; 1H NMR (chloroform-
d) δ 2.66-3.01 (m, 4H, -CH2-NH-, -OH), 2.99 (t, 2H, J ) 7.8
Hz, -CH2-Ph), 3.28 (t, 2H, J ) 7.8 Hz, -CH2-CO), 3.67 (d, 1H,
J ) 13.5 Hz, -N-CHa-Ph), 3.76 (d, 1H, J ) 13.5 Hz, -N-CHb-
Ph), 3.97-4.05 (m, 3H, -O-CH2-CH(O)-), 6.91 (d, 1H, J ) 7.8
Hz, arom H), 6.97 (t, 1H, J ) 7.8 Hz, arom H), 7.13-7.31 (m,
10H, arom H), 7.39 (dt, 1H, J ) 1.5, 7.8 Hz, arom H), 7.63
(dd, 1H, J ) 1.5, 7.8 Hz); 13C NMR (chloroform-d) δ 30.12
(PhCH2-), 44.97 (COCH2-), 51.23, 53.57 (-CH2-N-CH2-), 68.03
(CH(OH)), 71.25 (O-CH2-), 112.97, 125.84, 127.03, 127.99,
128.26, 128.30, 128.34, 130.07, 133.29, 139.67, 141.42, 157.54
(arom C), 201.40 (CO); IR (cm-1) 1670 (CO). Anal. (C25H27
NO3) C, H, N.
-
1f h yd r och lor id e: mp 132-137 °C (ethyl acetate). Anal.
(C25H28NO3Cl) C, H, N, Cl.
1-[2-[2-Hyd r oxy-3-(4-m or p h olin yl)p r op oxy]p h en yl]-1-
p r op a n on e (1n ): yield 44%; 1H NMR (chloroform-d) δ 1.19
(t, 3H, J ) 6.7 Hz, -CH3), 2.33-2.88 (m, 7H, -CH2-N-(CH2)2-,
-OH), 3.04 (qu, 2H, J ) 6.7 Hz, -CH2CO), 3.76 (t, 4H, J ) 4.8
Hz, -CH2-O-CH2-), 4.02-4.29 (m, 3H, O-CH2-CHO)-), 6.93-
7.72 (m, 4H, arom H); 13C NMR (chloroform-d) δ 8.46 (-CH3),
36.73 (-CH2CO), 53.75 (-CH2-N-), 61.12 (-N-(CH2)2-), 65.42
(-CH(OH)), 66.89 (-CH2-O-CH2-), 71.03 (Ar-O-CH2-), 112.81,
120.96, 128.67, 130.07, 133.04, 157.45 (arom C), 203.26 (CO);
IR (cm-1) 1670 (CO); MS (70 eV) 293 (M+, 1.3), 128 (14), 100
(100). Anal. (C16H23NO4) C, H, N.
1n h yd r och lor id e: mp 155-160 °C (ethyl acetate). Anal
(C16H24NO4) C, H, N, Cl.
2-[4-(2-Meth oxyp h en yl)-1-p ip er a zin yl]-1-[3-(2-p h en yl-
eth yl)-2-ben zofu r yl]eth a n on e (2f). Chloroethanone (9) (0.5
g, 1.7 mmol)26 was dissolved in 10 mL of toluene, and 0.64 g
of 1-(2-methoxyphenyl)piperazine (3.4 mmol) was added. The
reaction mixture was stirred for 5 h at 60 °C, filtered, and
evaporated to dryness. Crystallization with 2-propanol gave
1
0.53 g (70%) of 2f as colorless crystals: mp 138-140 °C; H
NMR (chloroform-d) δ 2.82-2.93 (m, 4H, -N-(CH2)2-), 2.97
(t, 2H, J ) 6.5 Hz, Ph-CH2), 3.16-3.27 (m, 4H, -(CH2)2-N-),
3.39 (t, 2H, J ) 6.5 Hz, -CH2-Bf), 3.87 (s, 3H, -OCH3), 3.98
(COCH2-), 6.86-7.56 (m, 13H, arom H); 13C NMR (chloroform-
d) δ 26.45, 30.90 (-CH2-CH2-), 50.51, 53.94 (piperazine CH2),
55.34 (-OCH3), 64.49 (COCH2-), 111.13, 112.15, 118.30, 121.01,
121.51, 122.93, 123.31, 126.04, 128.22, 128.32, 128.41, 128.56,
128.93, 141.33, 147.18, 152.28, 153.95 (arom C), 189.20 (CO);
IR (cm-1) 1690 (CO); MS (70 eV) 454 (M+, 17), 205 (100), 190
(23), 70 (45). Anal. (C29H30N2O3) C, H, N.
2f h yd r och lor id e: mp 197-198 °C (ethanol). Anal.
(C29H31N2O3Cl) C, H, N, Cl.
r-[[4-(2-Met h oxyp h en yl)-1-p ip er a zin yl]m et h yl]-3-(2-
p h en yleth yl)-2-ben zofu r a n m eth a n ol (2c). Ethanone 2f
(0.5 g, 1.1 mmol) was dissolved in 10 mL of methanol, and
0.05 g of NaBH4 was added. The reaction mixture was diluted
with water and extracted twice with CH2Cl2. The combined
organic layers were washed with water, dried over Na2SO4,
and evaporated to dryness. Crystallization from cyclohexane
gave 0.46 g (93%) of 2c: mp 99-100 °C; 1H NMR (chloroform-
d) δ 2.07 (d, 1H, J ) 12 Hz, -OH), 2.54-3.19 (m, 14H, Ph-
Mem br a n e In ter a ction Mea su r em en ts. 1. NMR Mea -
su r em en ts. P r ep a r a tion of Lip osom es for NMR Exp er i-
m en ts. BBPS was used for liposome preparation. Samples

SAR of Benzofuran Analogs of Propafenone
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 24 4773
(3) Higgins, C. F.; Hiles, I. D.; Salmonel, G. P. C.; Gill, D. R.; Downie,
J . A.; Evans, I. J .; Holland, I. B.; Gray, L.; Buckel, S. D.; Bell,
A. W.; Hermodson, M. A. A family of related ATP-binding
subunits coupled to many distinct biological processes in bac-
teria. Nature 1986, 323, 448-450.
(4) Kellen, J . A. Multidrug Resistance. In Reversal of multidrug
resistance in cancer; Kellen, J . A., Ed.; CRC Press: Boca Raton,
FL, 1994; pp 1-19.
containing 0.01 mg/µL of D2O were sonicated using a Branson
B-12 sonifier (Branson Sonic Power Co., Damburg, CT) three
times for 30 s at 40 W. This led to a desired increase in
temperature to 35 °C. After centrifugation the liposome
preparation was allowed to equilibrate for 24 h at room
temperature. This stock solution was diluted for the interac-
tion studies as indicated.
(5) Li, X.-Z.; Ma, D.; Livermore, D. M.; Nikaido, H. Role of efflux
pump(s) in intrinsic resistance of Pseudomonas aeruginosa:
Active efflux as a contributing factor to â-lactam resistance.
Antimicrob. Agents Chemother. 1994, 38, 1742-1752.
(6) Parkinson, T.; Falconer, D. J .; Hitchcock, C. A. Fluconazole
resistance due to energy-dependent drug efflux in Candida
glabrata. Antimicrob. Agents Chemother. 1995, 39, 1696-1699.
(7) Clark, F. S.; Parkinson, T.; Hitchcock, C. A.; Gow, N. A. R.
Correlation between rhodamine 123 accumulation and azole
sensitivity in Candida species: Possible role for drug efflux in
drug resistance. Antimicrob. Agents Chemother. 1996, 40, 419-
425.
(8) Sanglard, D.; Kuchler, K.; Ischer, F.; Pagani, J .-L.; Monod, M.;
Bille, J . Mechanisms of resistance to azole antifungal agents in
Candida albicans isolates from AIDS patients involves specific
multidrug transporters. Antimicrob. Agents Chemother. 1995,
39, 2378-2386.
(9) Gollapudi, S.; Reddy, M.; Gangadharam, P.; Tsuruo, T.; Gupta,
S. Mycobacterium tuberculosis induces expression of P-glyco-
protein in promonocytic U1 cells chronically infected with HIV
type 1. Biochem. Biophys. Res. Commun. 1994, 199, 1181-1187.
(10) Ford, J . M.; Hait, W. N. Pharmacology of drugs that alter
multidrug resistance in cancer. Pharmacol. Rev. 1990, 42, 155-
199.
(11) Tsuruo, T.; Iida, H.; Tsukagoshi, S.; Sakurai, Y. Overcoming of
vincristine resistance in P 388 leukemia in vivo and in vitro
through enhanced cytotoxicity of vincristine and vinblastine by
verapamil. Cancer Res. 1981, 41, 1967-1972.
NMR Mea su r em en ts. Drugs were dissolved in D2O or
phosphate buffer (pH 5.0). The final pH was adjusted to pH
5.0. The drug concentration was 4 mM and stayed constant
for the time of the experiment. A chemical shift of the
resonance signals and changes in pH were not observed.
Liposomes were added in 3-5 µL portions to 500 µL of solution.
Acetone served as the standard to control field homogeneity.
The experiments were performed with an AM 360L spectrom-
eter (Bruker, Darmstadt, Germany) at a probe temperature
of 23 °C. Data acquisition included 32 scans, 32K fid, sweep
width at 4098 Hz, and homonuclear presaturation to depress
the H2O signal; peak broadening (change in 1/T2) at half-peak
height as a function of liposome concentration was determined
from the NMR spectra. For calculation of peak half-widths, a
curve-fitting program was employed (M. Wiese, Borstel, Ger-
many). This program allows to fit mixed-type Gaussian and
Lorentzian curve shapes.
2. DSC. P r ep a r a tion of Lip osom es for DSC Exp er i-
m en ts. Liposome suspensions were prepared in phosphate
buffer (1/15 M, pH 7.4) using conditions that give multilamel-
lar vesicles.37 Briefly, DPPC/drug mixtures were prepared by
mixing appropriate amounts of the drug dissolved in either
methanol or chloroform/methanol (2:1, v/v) and DPPC dis-
solved in chloroform. All solvents were used in sufficient
quantities to ensure complete dissolution. The solvents were
evaporated under argon at 30 °C, and the samples were placed
in a vacuum desiccator overnight at 4 °C. Phosphate buffer
was added to dried samples, and the samples were vortexed
for 2 h at 60 °C. In all preparations the incubation temper-
ature was higher than the main phase transition temperature
of DPPC (about 42 °C). The vortex intensity used was 1200-
(12) Chauffert, B.; Martin, M.; Hamman, A.; Michel, M. F. Amio-
daron-induced enhancement of doxorubicin and 4′-deoxydoxo-
rubicin cytotoxicity to rat colon cancer cells in vitro and in vivo.
Cancer Res. 1986, 46, 825-830.
(13) Chiba, P.; Ecker, G.; J a¨ger, W.; Freudhofmayer, E.; Gekeler, V.;
Fleischhacker, W. A new group of class I antiarrhythmic drug
analogs reestablishes daunomycin sensitivity in multidrug-
resistant T-lymphoblasts. Proc. Am. Assoc. Cancer Res. 1993,
34, 321.
1300 min-1
. The lipid concentration was 5 mg/mL in all
(14) Tsuruo, T.; Iida, H.; Tsukagoshi, S.; Sakurai, Y. Potentiation of
vincristine and adriamycine in human hematopoietic tumor cell
lines by calcium antagonists and calmodulin inhibitors. Cancer
Res. 1983, 43, 2267-2272.
(15) Ford, J . M.; Prozialeck, W. C.; Hait, W. N. Structural features
determining activity of phenothiazines and related drugs for
inhibition of cell growth and reversal of multidrug resistance.
Mol. Pharmacol. 1989, 35, 105-115.
(16) Ford, J . M.; Bruggemann, E. P.; Pastan, I.; Gottesman, M. M.;
Hait, W. N. Cellular and biochemical characterization of thiox-
anthenes for reversal of multidrug resistance in human and
murine cell lines. Cancer Res. 1990, 50, 1748-1756.
(17) Twentyman, P. R. Cyclosporins as drug resistance modifiers.
Biochem. Pharmacol. 1992, 43, 109-117.
(18) Atassi, G.; Pierre, A.; Kraus-Berthier, L.; Leonce, S.; Regnier,
G.; Dhainaut, A. S 9788 corrects adriamycin and vincristine
resistance on human and murine tumor cells in vitro and on P
388 leukemia in vivo. J . Cancer Res. Clin. Oncol. 1991, 117
(Suppl. 3), 108.
(19) Hyafil, F.; Vergely, C.; Du Vignaud, P.; Grand Perret, T. In vitro
and in vivo reversal of multidrug resistance by GF 120918, an
acridonecarboxamide derivative. Cancer Res. 1993, 53, 4595-
4602.
(20) Raderer, M.; Scheithauer, W. Clinical trials of agents that
reverse multidrug resistance. Cancer 1993, 72, 3553-3563.
(21) Ecker, G.; Chiba, P. Structure-activity-relationship studies on
modulators of the multidrug transporter P-glycoprotein - an
overview. Wien. Klin. Wochenschr. 1995, 107, 681-686.
(22) Pajeva, I. K.; Wiese, M.; Cordes, H.-P.; Seydel, J . K. Membrane
interactions of some catamphiphilic drugs and relation to their
multidrug-resistance-reversing ability. J . Cancer Res. Clin.
Oncol. 1996, 122, 27-40.
(23) Wadkins, R. M.; Houghton, P. J . The role of drug-lipid interac-
tions in the biological activity of modulators of multidrug
resistance. Biochim. Biophys. Acta 1993, 1153, 225-236.
(24) Chiba, P.; Burghofer, S.; Richter, E.; Tell, B.; Moser, A.; Ecker,
G. Synthesis, pharmacologic activity, and structure-activity
relationships of a series of propafenone-related modulators of
multidrug resistance. J . Med. Chem. 1995, 38, 2789-2793.
(25) Chiba, P.; Ecker, G.; Schmid, D.; Drach, J .; Tell, B.; Goldenberg,
S.; Gekeler, V. Structural requirements for activity of pro-
pafenone type modulators in PGP-mediated multidrug resis-
tance. Mol. Pharmacol. 1996, 49, 1122-1130.
samples, and changes in pH were not observed. The experi-
ments were done at lipid:drug molar ratios from 1:0 (control)
to 1:0.1 in most cases.
DSC Mea su r em en ts. All measurements were performed
with a highly sensitive differential scanning microcalorimeter
(Setaram, France), equipped with a digital interface and a data
acquisition allowing automatic data collection; 200 µL aliquots
of the liposome suspension and of the reference solution
(phosphate buffer) were added to the relevant pans and heated
at a rate of 0.5 °C/min. The sensitivity range setting was 50
µV. Temperature intervals varied from 0 to 70 °C for different
phospholipids and phospholipid/drug mixtures. Every experi-
ment was performed in duplicate, and two runs were recorded
for selected samples to ensure that the calorimetric response
of the system was stable. Peaks were analyzed by a nonlinear
least-squares fit using a Gaussian function. The peak width
of a molar DPPC to modifier concentration of 1:0.05 was taken
as a descriptor of interaction strength.
Ca lcu la tion of Lip op h ilicity. The log P values were
calculated according to the method of Ghose and Crippen34
using the software package MOLGEN.35 Molecules were
generated using the builder function and energetically mini-
mized using the optimization function. Conformationally
independent lipophilicity values were calculated.
Su p p or tin g In for m a tion Ava ila ble: NMR spectra of 1a
and 2b in the absence and presence of lecithin liposomes (1
page). Ordering information is given on any current masthead
page.
Refer en ces
(1) Fan, D.; Beltran, T. J .; O’Brien, C. A. Reversal of multidrug
resistance. In Reversal of multidrug resistance in cancer; Kellen,
J . A., Ed.; CRC Press: Boca Raton, FL, 1994.
(2) J uliano, R. L.; Ling, V. A surface glycoprotein modulating drug
permeability in Chinese hamster ovary cell mutants. Biochim.
Biophys. Acta 1976, 455, 152-162.

4774 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 24
Ecker et al.
(26) Nielek, S.; Lesiak, T. Chemistry of thiazole, I: Synthesis and
properties of 2,3,5,6-tetrahydro-6-(3-methyl-benzofuran-2-yl)-
imidazo[2,1-b]thiazole. Chem. Ber. 1982, 115, 1247-1251.
(27) Ecker, G.; Mohr, E.; Geyer, R.; Fleischhacker, W. Enantiosepa-
ration of propafenone-type modulators of multidrug resistance
on cyclodextrin based chiral stationary phases. Sci. Pharm. 1996,
64, 1-11.
(28) Ecker, G.; Fleischhacker, W.; Noe, C. R. New benzofurane-type
antiarrhythmic compounds related to propafenone. Heterocycles
1994, 38, 1247-1256.
(32) Gekeler, V.; Frese, G.; Noller, A.; Handgretinger, R.; Wilisch,
A.; Schmidt, H.; Muller, C.; Dopfer, R.; Klingbiel, T.; Diddens,
H.; Probst, H.; Niethammer, D. Mdr1/P-glycoprotein, topo-
somerase and glutathione-S-transferase gene expression in
primary and relapsed state adult and childhood leukemias. Br.
J . Cancer 1992, 66, 507-517.
(33) Ecker, G.; Chiba, P.; Schaper, K.-J . Estimation of chemosensi-
tizing activity of modulators of multidrug resistance via com-
bined simultaneous analysis of sigmoidal dose-response curves.
J . Pharm. Pharmacol., submitted.
(29) Ecker, G.; Fleischhacker, W.; Helml, T.; Noe, C. R.; Scasny, S.;
Lemmens-Gruber, R.; Studenik, C.; Marei, H.; Heistracher, P.
Improved synthesis and pharmacologic activity of the enanti-
omers of a new benzofurane type antiarrhythmic compound.
Chirality 1994, 6, 329-336.
(30) Ecker, G.; Fleischhacker, W.; Helml, T.; Noe, C. R.; Studenik,
C.; Schade, B.; Heistracher, P. Synthesis and pharmacodynamic
activity of 2-(3-(2-phenylethyl)benzofuran-2-yl)-N-propyl-etha-
namine. Arch. Pharm. 1995, 328, 343-348.
(31) Lee, J . S.; Paull, K.; Alvarez, M.; Hose, C.; Monks, A.; Grever,
M.; Fojo, A. T.; Bates, S. E. Rhodamine efflux patterns predict
P-glycoprotein substrates in the National Cancer Institute drug
screen. Mol. Pharmacol. 1994, 46, 627-638.
(34) Ghose, A. K.; Pritchett, A.; Crippen, G. M. Atomic physicochem-
ical parameters for three dimensional structure directed quan-
titative structure-activity relationships III: modeling hydropho-
bic interactions. J . Comput. Chem. 1988, 9, 80-90.
(35) P. Baricic and M. Mackov, distributed by Milan Hudecek, P.
Horova 18, 841 07 Bratislava, Slowakia.
(36) Mosmann, T. Rapid colorimetric assay for cellular growth and
survival: application to proliferation and cytotoxicity assays. J .
Immunol. Methods 1983, 65, 55-63.
(37) New, R. R. C., Ed. Liposomes: a practical approach; Oxford
University Press: New York, 1990.
J M960384X