Effects of isoprenoid analogues of SDB-ethylenediamine on multidrug resistant tumor cells alone and in combination with chemotherapeutic drugs

Article abstract of DOI:10.1021/jm011010t

Multidrug resistance (MDR) mediated by P-glycoprotein (Pgp) remains the major obstacle for successful treatment of cancer. Inhibition of Pgp transport is important for higher efficacy of anticancer drugs. Lipophilic cationogenic amines with at least one tertiary N atom, such as verapamil, are classical PgP-blocking agents. In a search for novel accessible compounds potent against MDR tumor cells, we synthesized a series of arylalkylamines that contain isoprenoid side chains of different length. Two out of seven new analogues of the known N,N′-bis(3,4-dimethoxybenzyl)-N-solanesylethylenediamine (SDB-ethylenediamine), namely, compounds with C₁₀ and C₁₅ side chains, at low micromolar concentrations were preferentially toxic for several mammalian tumor cell lines that acquired MDR during prolonged drug selection. Moreover, at noncytotoxic concentrations, these compounds potently sensitized MDR cells to Pgp substrates vinblastine and adriamycin. We conclude that these analogues of SDB-ethylenediamine may have dual therapeutic advantage because (i) they are preferentially toxic for MDR cells when administered alone and (ii) they potentiate the cytotoxicity of Pgp-transported anticancer drugs.

Full text of DOI:10.1021/jm011010t

5330
J . Med. Chem. 2002, 45, 5330-5339
Effects of Isop r en oid An a logu es of SDB-Eth ylen ed ia m in e on Mu ltid r u g
Resista n t Tu m or Cells Alon e a n d in Com bin a tion w ith Ch em oth er a p eu tic Dr u gs
Tatyana A. Sidorova,*,† Albert G. Nigmatov,‡ Eugenia S. Kakpakova,§ Alla A. Stavrovskaya,§
Galina K. Gerassimova,† Alexander A. Shtil,§ and Edward P. Serebryakov*,‡
Laboratories of Biochemical Pharmacology and Tumor Cell Genetics, N. Blokhin Cancer Center, Russian Academy of Medical
Sciences, 24 Kashirskoye Shosse, 115478 Moscow, Russia, and N. D. Zelinsky Institute of Organic Chemistry, Russian
Academy of Sciences, 47 Leninsky Prospekt, 119991 Moscow, GSP-2, Russia
Received August 3, 2001
Multidrug resistance (MDR) mediated by P-glycoprotein (Pgp) remains the major obstacle for
successful treatment of cancer. Inhibition of Pgp transport is important for higher efficacy of
anticancer drugs. Lipophilic cationogenic amines with at least one tertiary N atom, such as
verapamil, are classical PgP-blocking agents. In a search for novel accessible compounds potent
against MDR tumor cells, we synthesized a series of arylalkylamines that contain isoprenoid
side chains of different length. Two out of seven new analogues of the known N,N′-bis(3,4-
dimethoxybenzyl)-N-solanesylethylenediamine (SDB-ethylenediamine), namely, compounds
with C10 and C15 side chains, at low micromolar concentrations were preferentially toxic for
several mammalian tumor cell lines that acquired MDR during prolonged drug selection.
Moreover, at noncytotoxic concentrations, these compounds potently sensitized MDR cells to
Pgp substrates vinblastine and adriamycin. We conclude that these analogues of SDB-
ethylenediamine may have dual therapeutic advantage because (i) they are preferentially toxic
for MDR cells when administered alone and (ii) they potentiate the cytotoxicity of Pgp-
transported anticancer drugs.
In tr od u ction
prenoids, N-(p-methylbenzyl)decaprenylamine12-14 and
N,N′-bis(3,4-dimethoxybenzyl)-N-solanesylethylenedi-
amine (SDB-ethylenediamine, 1),14-20 were found to
potently overcome the MDR phenotype in cell culture
as well as in in vivo tumor models. In particular,
diamine 1 became the subject of synthetic17,19-21 and
biomedical14-20,22-27 studies. Photoaffinity labeling ex-
periments showed that 1, like VRP, can directly interact
with Pgp and inhibit the binding of Pgp-transported
agents to the pump, thus blocking drug efflux.16,17
Because of structural similarities between 1 and VRP
(i.e., a tertiary amino group, two benzene rings con-
nected by an aliphatic chain, and a hydrocarbon side
chain), one may suggest that the isoprenoid chain of 1
is necessary for effective blocking of Pgp. Neither of the
two constituents of 1, N,N′-bis(3,4-dimethoxybenzyl)-
ethylenediamine (2), and solanesol (SolOH) (Figure 1)
sensitized the MDR human hepatoma PLC/COL or
Chinese hamster V79/ADM sublines to anticancer
drugs.23,24 Plausibly, among many conformations of
intact 1, only one or only a few can be important for its
MDR-reversing activity. Therefore, a search for optimal
structure of isoprenoid amines to circumvent the MDR
phenotype may have a significant clinical potential.
Multidrug resistance (MDR) of tumor cells remains
a major obstacle for successful chemotherapy of cancer
patients.1-6 This type of resistance is associated with
an increased efflux of antineoplastic drugs out of the
cells and subsequent decrease of intracellular concen-
tration of the drug.2 Active extrusion of drugs is medi-
ated by several molecular transporters; among them,
P-glycoprotein (Pgp), a 140-170 kDa adenosine 5′-
triphosphate (ATP)-dependent transmembrane pump,
is the most widely studied.1,3 Pgp is capable of effluxing
a broad variety of compounds of different chemical
structure, thereby rendering cells resistant to many
agents. Consequently, blocking of Pgp-mediated trans-
port can sensitize tumor cells to chemotherapeutic
drugs. The combination of Pgp-reversing agents with
conventional anticancer drugs resulted in an improved
therapeutic response, at least in some types of malig-
nancies.4-6
Generally, MDR-reversing compounds interfere with
Pgp function by the depleting intracellular ATP pool or
by interacting with the pump itself. Most agents of the
second type, e.g., verapamil (VRP),7-9 are cationogenic
lipophilic amines with at least one secondary or tertiary
N atom.10,11 Among those amines, two synthetic iso-
With the exception of a few synthetic long-chained
(C45-C50) isoprenoid amines that display synergistic
effects in combination with some anticancer drugs,12,19
no data have been reported on the cytotoxicity and
MDR-reversing effect of other isoprenoid analogues of
1, in particular, of those with shorter isoprenoid chains.
Considering the putative role of the length and confor-
mation of isoprenoid chains in interaction with plasma
membrane proteins, as well as greater commercial
availability of linear C10, C15, and C20 isoprenoids as
* To whom correspondence should be addressed. T.A.S.: Tel: +7-
(095)324-5517. Fax: +7(095)323-5333. E-mail: shtilaa@yahoo.com.
E.P.S.: Tel: +7(095)137-1379. Fax: +7(095)135-5328. E-mail: ser@
ioc.ac.ru.
† Laboratory of Biochemical Pharmacology, Russian Academy of
Medical Sciences.
‡ N. D. Zelinsky Institute of Organic Chemistry, Russian Academy
of Sciences.
§ Laboratory of Tumor Cell Genetics, Russian Academy of Medical
Sciences.
10.1021/jm011010t CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/26/2002

Isoprenoid Analogues of SDB-Ethylenediamine
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24 5331
F igu r e 1. Structures of MDR-reversing amines, VRP, and
SDB-ethylenediamine (1) and biologically inert constituents
of 1.
compared with higher polymer homologues, this lack of
information is surprising. We report here the synthesis
and biological characteristics (i.e., cytotoxicity and
MDR-reversing activity) of several close analogues of 1
(Figure 2). These analogues have one (subseries A, 3-5)
or two (subseries B, 1a , 4a , and 5a ) side chains. We
provide evidence that the compounds with relatively
short C10 and C15 isoprenoid chains efficiently killed
different mammalian MDR tumor cells in culture.
Moreover, these compounds synergized with conven-
tional chemotherapeutic drugs, thereby circumventing
pleiotropic drug resistance.
F igu r e 2. Novel N-monosubstituted (subseries A) and N,N′-
disubstituted (subseries B) analogues of SDB-ethylenediamine
(1) derived from diamine 2. Compounds of the type A (3-5)
contain a single isoprenoid chain attached to one of the
nitrogen atoms. Analogues of the type B (1a and 3a -5a )
contain two isoprenoid side chains of equal size attached to
each of the nitrogen atoms in the molecule of the parental
diamine 2.
Resu lts
Ch em istr y. Our interest was originally focused on
the impact of isoprenoid side chains (or, more generally,
of conformationally flexible large lipophilic groups) on
cytotoxicity and MDR-reversing activity of 1 and its
congeners. For the initial modification of 1, we were
primarily interested in the properties of different short
chain analogues of this compound. No modification of
the pharmacogenic template was undertaken. The
synthesis of novel compounds is shown in Scheme 1.
Diamine 1 and its N,N ′-bis-alkylated analogue 1a were
obtained from diamine 217 and solanesyl bromide (6)
with 85 and 12% yield, respectively (Scheme 1, reaction
ii). Low molecular isoprenoid analogues 3-5 were
prepared by alkylating 2 via a slightly modified proce-
dure previously used for the synthesis of 1.17,21 Allylic
halides 7-9 used in these alkylations contained 2E and
2Z stereoisomers in a roughly 3:1 ratio; the same
stereomeric ratio was found in the specimens 3-5
(Scheme 1). Because N,N′-bis-alkylated products 1a and
3a -5a were also the subject of our study and their
separation from 1 and 3-5 was easy, no attempts at
selective mono- or bis-alkylation of 2 were made. Taking
advantage of low selectivity of alkylation (Scheme 1,
reaction iii), we obtained both types of isoprenoid amines
1, 3-5 and 1a , 4a , and 5a in acceptable yields from a
single chemical operation. It is noteworthy that when
alkylation of 2 was performed with C10-chloride 7, the
reaction yielded monoalkylated diamine 3 as almost a
single product, whereas the reactions of 2 with bromides
6, 8, and 9 were less selective. In the reaction of 2 with
6, the selectivity of Suga’s procedure28 was substantially
higher (1:1a ∼7:1) than the selectivity of conventional
methods.17,21
All novel analogues of SDB-diamine were obtained
as hydrophobic waxes or viscous gums. Free bases were
1
characterized by their Rf values and H NMR and MS-
FAB (mass spectrometry fast atom bombardment) spec-
tra (Table 1). For biological testing, free bases were
converted into dihydrochlorides 1‚2HCl, 3‚2HCl, 5‚
2HCl, 1a ‚2HCl, 4a ‚2HCl, and 5a ‚2HCl that appeared
to be poorly soluble in water and only sparingly soluble
in a dimethyl sulfoxide (DMSO)-H2O mixture (see
Experimental Section). The solubility in DMSO-H2O
tended to increase with the decrease of the number and
length of isoprenoid chains in the molecule. For the
compounds of subseries B, the solubility was below 0.5
mM even at 10% of DMSO in water, whereas for
dihydrochlorides of 3-5 it was above 1 mM at 2%

5332 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24
Sidorova et al.
Ta ble 1. Chemical Characteristics of SDB-Ethylenediamine (1), Its Close Analogues (2-5 and 1a -5a ), and Their Dihydrochlorides
chemical parameters
b
compds
method of synthesisa
yield (%)
Rf
molecular formula
Subseries A
elemental analysisc
molecular ion (MS-FAB)c
1
ii, iii
iii
85 (ii), 43 (iii)
0.27
0.11
0.25
0.22
C
C
C
C
C
C
C
C
65H100N2O4
65H102Cl2N2O4
30H44N2O4
30H46Cl2N2O4
35H52N2O4
35H54Cl2N2O4
40H66N2O4
40H68Cl2N2O4
973
1‚2HCl
3
3‚2HCl
4
4‚2HCl
5
5‚2HCl
C, 74.25; H, 9.94
C, 63.09; H, 8.21
C, 65.73; H, 8.68
C, 67.29; H, 9.72
84
49
54
496
iii
564
iii
624 (M - CH3)
Subseries B
1a
1a ‚2HCl
3a
4a
4a ‚2HCl
5a
ii, iii
12 (ii), 19 (iii)
0.67
C
C
C
C
C
C
C
110H172N2O4
110H174Cl2N2O4
40H60N2O4
50H76N2O4
50H78Cl2N2O4
60H104N2O4
1586
N, 1.83
iii
iii
<1
19
0.53
0.60
633
769
C, 71.06; H, 9.54
N, 2.69
iii
14
0.51
902 (M - CH3)
5a ‚2HCl
60H106Cl2N2O4
a
b
Relates to stages (ii) or (iii) in Scheme 1. On Silufol sheets using AcOEtsMeOH (4:1, v/v) as the eluent. c Corresponds to calculated
d
values with 0.1-0.5% accuracy. Using a Cratos 50-TS instrument at 8000 eV (see Experimental Section for details); the m/z values are
given.
Ta ble 2. Distribution Coefficients of 1‚2HCl and 3‚2HCl in
AcOEt-H2O System and Calculated Partition Coefficients
(Clog P) for 1-3
Sch em e 1. Synthesis of SDB-Ethylenediamine (1) and
Its Close Analoguesa
Log D
compds
pH 6.4
pH 4.0
Clog Pa
1‚2HCl 1.628 ( 0.02 (n ) 2) 1.552 ( 0.06 (n ) 3)
1:5.587
3‚2HCl 0.389 ( 0.03 (n ) 3) -0.426 ( 0.03 (n ) 3) 3:4.506
2‚2HCl not determined
not determined
2:3.782
a
Using CHEM 3D ULTRA, version 6.0, for 1-octanol/water
system.
(where monocation forms can appear) and at pH 4.0
(which favors the dication form) the measured values
of log D for the dihydrochloride of 3 were ∼1.2 and 1.9
log units lower than in the case of 1‚2HCl. Actually, the
C
org/Cwater ratio at pH 6.4 was 2.45 for the dihydrochlo-
ride of 3 and 42.5 for the dihydrochloride of 1, while
the Corg/Cwater ratio at pH 4.0 was 0.38 for 3‚2HCl and
35.7 for 1‚2HCl. Thus, the concentration of what is
presumably the dication form of 1 in the mildly acidic
aqueous phase is approximately 2 orders of magnitude
lower than that of 3‚2HCl.
These data are qualitatively similar to the values of
ClogP (for 1-octanol/water) of compounds 1 (∼5.59) and
3 (∼4.51). For their common template, molecule 2, Clog
P is ∼3.78.
Biologica l Testin g. To assess pharmacological po-
tency of our novel compounds, we investigated their
cytotoxicity and ability to reverse the MDR phenotype
in cultured mammalian tumor cell lines.
a
Reagents and conditions: (i) (H2NCH2)2, NaBH4, MeOH, room
temperature, 8 h. (ii) Li+[C10H8]- under Ar, room temperature, 1
h, and then 6 (1.04 equiv), THF, room temperature, 15 h. (iii)
Compounds 7-9 (0.5 equiv), THFsEt2O (1:1), room temperature,
2 h (for 2 + 6 f 1 + 1a in Et2O alone).
Ch a r a ct er iza t ion of t h e MDR P h en ot yp e in
Dr u g-Selected Syr ia n Ha m ster F ibr obla st Cells.
The HET-SR-2SC-LNM Syrian hamster embryo fibro-
blast cell line is a highly malignant variant obtained
after infection of embryos with the Rous sarcoma virus
followed by in vivo selection for the ability to produce
metastases.33,34 Furthermore, this cell line is intrinsi-
cally resistant to colchicine (CLC), a Pgp-transported
drug.35 To establish an isogenic subline with a Pgp-
mediated MDR, HET-SR-2SC-LNM cells were selected
for resistance to increasing concentrations of CLC.34,36
The resulting 2SC/20 subline (constantly maintained at
20 µg/mL CLC) was resistant to CLC as well as to Pgp-
DMSO in water. These data are in contrast with good
solubility of VRP in water (7 g/100 g, that is, 140 mM).29
To determine the hydrophilicity of dihydrochlorides
of 1 and 3, their partition coefficients between ethyl
acetate, a volatile solvent of nearly the same lipophi-
licity as 1-octanol,30 and aqueous phosphate buffers (0.1
m, pH 6.4 and pH 4.0) were determined (Table 2). As
expected for ionizable compounds, these coefficients
were pH-dependent. For such solutes, the apparent
partition or distribution coefficient (log D) is an ad-
equate characteristic.31,32 Table 2 shows that at pH 6.4

Isoprenoid Analogues of SDB-Ethylenediamine
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24 5333
Ta ble 3. MDR Phenotype in 2SC/20 Subline
intracellular drug uptake
[3H]VBL
(pmol/mg protein)
ADM
(pmol/30 min)
IC50 ( nM)a (IR)b
ADM
cell line
CLC
25 ( 2
VBL
18 ( 8
AraC
BLEO
-VRP
22 ( 1
+VRP
-VRP
+VRP
HET-SR-2SC-LNM
2SC/20
52 ( 11
772 ( 383
519 ( 160
29 ( 1c 235 ( 36 310 ( 14
23 ( 2c 132 ( 16 289 ( 20c
500 ( 12 (20) 246 ( 25 (12) 620 ( 130 (12) 1250 ( 250 (1.6) 517 ( 248 (1.0) 10 ( 2
a
IC50 calculated as the concentration that caused 50% inhibition of growth of respective cell line after 72 h of drug exposure (by MTT
b
test; see Experimental Section for details) and expressed as mean ( SE of three experiments. Index of resistance (IR) for each drug was
calculated as mean IC50 for 2SC/20 subline divided by mean IC50 for HET-SR-2SC-LNM cells. For uptake experiments, [3H]VBL (10 nM)
or ADM (10 µM) were added to the cells with or without 5 µM VRP for 30 min. The values represent mean ( SE of four experiments.c p
< 0.01 as compared with uptake of respective drug (VBL or ADM) without VRP (-VRP group).
Ta ble 4. Toxicity of VRP and Dihydrochlorides of 1-4 for
Ta ble 5. Toxicity of VRP and Dihydrochlorides of 1 and 3 for
HET-SR-2SC-LNM and 2SC/20 Cella
Parental and Multidrug Transporter-Expressing Cell Linesa
IC50 (µM)
IC50 (µM)
multidrug
cell line
transporter
VRP
1
3
cell line
VRP
1
3
4
2
McA RH 7777
McA RH 7777/0.4
mS
mS/0.5
K562
K562i/S9
COR/L23P
COR/L23R
150 ( 20
470 ( 60
66 ( 7
230 ( 14
100 ( 5
100 ( 3
65 ( 5
55 ( 4
55 ( 5
HET-SR-2SC-LNM 48 ( 3
2SC/20
80 ( 6 15 ( 2b 35 ( 1b
520 ( 33
58 ( 0.4 20 ( 2b 10 ( 1c 13 ( 2b,d 470 ( 60
Pgp
Pgp
Pgp
MRP
20 ( 5b
86 ( 12 40 ( 8
a
45 ( 7c
250 ( 8 125 ( 1
250 ( 6 125 ( 4
18 ( 4c
Cells were treated with indicated compounds for 72 h. Cell
viability was determined in an MTT test. Data are mean ( SE of
three experiments. b p < 0.001 as compared with IC50 of 1 for HET-
d
SR-2SC-LNM cells. c p < 0.01. p < 0.05 as compared with IC50
80 ( 3 >200
80 ( 4 >200
25 ( 1
13 ( 1b
of 1 for 2SC/20 cells.
a
transported drugs vinblastine (VBL) and adriamycin
(ADM) (resistance factors 20, 12, and 12, respectively)
as compared with the parental HET-SR-2SC-LNM cell
line (Table 3). However, both the parental- and the CLC-
selected cells were similarly sensitive to 1-â-D-arabino-
furanosylcytidine (ARA C) and bleomycin (BLEO),
agents that normally do not participate in the MDR
phenotype (Table 3).37 The 2SC/20 cells overexpress Pgp
as determined by Western blotting.36 Furthermore, the
Pgp-blocking drug VRP markedly increased intracellu-
lar accumulation of [3H]VBL and ADM in 2SC/20 cells
(Table 3). Interestingly, VRP also increased [3H]VBL
uptake by HET-SR-2SC-LNM cells to a moderate but
statistically significant extent (p < 0.01 as compared
with the -VRP group). These data indicate that 2SC/
20 cells express “typical” Pgp-mediated MDR. Together,
HET-SR-2SC-LNM and 2SC/20 cell lines are adequate
models to assess the potency of 1 and its novel iso-
prenoid analogues for highly malignant cells that dem-
onstrate intrinsic (i.e., prior to drug selection) as well
as acquired (i.e., associated with selection procedure)
resistance to several anticancer drugs.
Toxicity of SDB-Eth ylen ed ia m in e An a logu es for
MDR Tu m or Cells. We were interested in the ability
of our novel compounds to completely execute their
cytotoxic effect. For this purpose, we used prolonged (up
to 72 h) treatments and assessed late events of cell
death such as formazan conversion by mitochondrial
dehydrogenases (MTT test)38 or loss of the plasma
membrane integrity (Trypan blue exclusion test). The
cytotoxicity of newly synthesized analogues of 1 to HET-
SR-2SC-LNM and 2SC/20 cells was compared with the
effects of three reference compounds: VRP, dihydro-
chloride of 1, and dihydrochloride of the parental
diamine 2. Compound 1 and its short chain analogues
3 and 4 (all tested as the corresponding dihydrochlo-
rides) were toxic to HET-SR-2SC-LNM and 2SC/20 cells
at micromolar concentrations (Table 4). Most impor-
tantly, 1, 3, and 4 were even more toxic for the 2SC/20
subline than for HET-SR-2SC-LNM cells (indices of
resistance are 4, 1.5, and 2.7, respectively; see Experi-
Cells were treated with indicated compounds for 72 h. Cell
viability was assessed in an MTT test (see Experimental Section).
p < 0.01. c p < 0.05 as compared with respective parental cell
line.
b
mental Section), whereas 2SC/20 cells were more resis-
tant to VRP than HET-SR-2SC-LNM cells (Table 4). The
range of cytotoxicity for the 2SC/20 subline was 3 > 4
> 1 > VRP, indicating that 1 and its analogues 3 and 4
could be more potent than VRP in killing MDR tumor
cells. Moreover, both 3 and 4 were significantly more
potent than 1 for HET-SR-2SC-LNM cells (p < 0.001;
see Table 3). In agreement with previous data,22 di-
amine 2 demonstrated much poorer cytotoxicity (Table
4). Its derivatives with two side chains also demon-
strated much poorer cytotoxicity (1a , 4a , and 5a ); for
these compounds, the IC50 values were in the submil-
limolar range (data not shown).
We sought to expand the therapeutic potential of
novel compounds by testing the toxicity of 3, the most
accessible analogue, for other tumor cell lines that
express MDR transporters. Drug-selected mammalian
tumor cell lines of different species and tissue origin
were used as models (Table 5). The compound 3 proved
to be more toxic for Pgp positive rat hepatoma McA RH
7777/0.4 subline39 than to its parental Pgp negative
counterpart McA RH 7777 (Table 5). In contrast, McA
RH 7777/0.4 cells were ∼3-fold more resistant to VRP
than McA RH 7777 cells. SDB-Ethylenediamine 1 was
similarly toxic for McA RH 7777 and McA RH 7777/0.4
cells. Furthermore, 1 and 3 were more potent against
the Pgp-expressing human melanoma mS/0.5 subline40
as compared with the wild-type Pgp negative mS cell
line (Table 5). In contrast, the mS/0.5 subline showed
cross-resistance to VRP. The compound 3 was more toxic
for the MDR-related protein (MRP) positive, Pgp nega-
tive COR/L23R lung carcinoma cells41 than to the
parental COR/L23P counterparts (Table 5). Thus, the
monogeranyl derivative 3 could be efficient for a broader
panel of drug transporter-expressing tumor cells that
are otherwise resistant to several chemotherapeutics.

5334 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24
Sidorova et al.
Ta ble 6. Sensitization to Pgp-Transported Drugs by VRP and
Dihydrochlorides of 1, 3, and 4
HET-SR-2SC-LNM
2SC/20
compd
VBL
ADM
VBL
ADM
VRP
4 ( 0.3a
1 ( 0.4
8 ( 1b
2 ( 0.1
2 ( 0.1
2 ( 0.04
2 ( 0.1
11 ( 2
11 ( 1
1
3
4
4 ( 0.3
6 ( 0.3
11 ( 0.4b
12 ( 0.3b
12 ( 0.4b
12 ( 0.3b
6 ( 2c
a
Index of sensitization (IS) was calculated as mean IC50 for cells
incubated with VBL or ADM alone divided by mean IC50 for cells
incubated with one of these drugs plus VRP or 1, 3, or 4 (5 µM
b
each). p < 0.001. c p ) 0.05 as compared with IS by 1 for VBL or
ADM in respective cell line. The experiments were repeated four
times with <10% error.
F igu r e 3. Potentiation of ADM accumulation by VRP, 1‚2HCl,
and 3‚2HCl. The 2SC/20 cells were treated with 10 µM ADM
for indicated times in the presence or absence of VRP and
dihydrochlorides of 1 or 3 (5 µM each). Intracellular ADM
concentration was determined as described in the Experimen-
tal Section. Each treatment was performed in triplicate. One
representative experiment out of three (with essentially
similar results) is shown.
Drug-selected cells presumably accumulate various
changes in addition to up-regulation of multidrug
transporters. To discriminate between Pgp expression
per se and other selection-associated mechanism(s) that
render cells sensitive to the analogues of 1, we compared
the toxicity of these analogues for K562 and K562i/S9
cells. The latter subline expresses exogenous Pgp with-
out drug selection.42,43 Table 5 shows that there was no
preferential toxicity of VRP, 1, and 3 for K562i/S9 cells
as compared to K562 cells, suggesting that it is drug
selection (but not merely Pgp expression) that confers
higher sensitivity of cells to our novel compounds. Taken
together, our observations demonstrate that the ana-
logues of 1 with shorter isoprenoid chains not only
retain toxicity for the parental tumor cells but also can
be even more potent than 1 in killing cells selected for
Pgp- and MRP-mediated MDR.
SDB-Eth ylen ed ia m in e a n d Its Novel An a logu e
3 In cr ea se ADM Up ta k e by P gp -Exp r essin g Cells.
We next studied the ability of 1 and its congeners to
reverse MDR in tumor cells. Two activities were tested
as follows: (i) potentiation of the intracellular ac-
cumulation of anticancer drugs and (ii) increased cell
death after coincubation with chemotherapeutic drugs.
To address the first activity, we studied the uptake of
ADM in the presence or absence of 1 or 3. VRP was used
as a reference agent. As shown in Figure 3, 1 and 3
augmented the accumulation of ADM in 2SC/20 cells,
although the rate of uptake was different in the pres-
ence of 1 and 3. The effect of 1 was detectable only when
the cells were coincubated with this compound and
ADM for 4 h; however, after an additional 2 h with 1,
the intracellular amount of ADM decreased. In contrast,
3 and VRP markedly potentiated ADM accumulation
within 30 min (Figure 3). These results indicate that 1
and its short chain isoprenoid congeners can reverse
MDR by increasing intracellular content of the Pgp-
transported drug(s).
F igu r e 4. Effects of dihydrochlorides of 1 and 3 on ADM-
induced death in 2SC/20 cells. Cells were treated with 10 µM
ADM alone (open bars) or with ADM + VRP (oblique hatching),
ADM + 1 (horizontal hatching), or ADM + 3 (closed bars) for
24 h followed by washing with PBS and incubation in drug
free medium for an additional 48 h. Cell viability was
determined by Trypan blue exclusion. VRP, 1, and 3 were at
5 µM each. Data are mean ( SE of three independent
experiments.
ethylenediamine analogues (at low micromolar concen-
trations) with anticancer drugs for 72 h as well as
cotreatment with SDB-ethylenediamine analogues plus
anticancer drugs for 24 h followed by cell washing and
further incubation in drug-free medium for an ad-
ditional 48 h. In the experiments with a 72 h coexposure,
geranyl derivative 3 and farnesyl derivative 4 (5 µM
each) sensitized 2SC/20 cells to Pgp-transported drugs
VBL and ADM as potently as VRP (Table 5). Further-
more, 3 and 4 were more efficient than VRP in sensitiz-
ing HET-SR-2SC-LNM cells to VBL and ADM (Table
6). Compounds 1 and 2 did not influence the survival
of the wild-type Syrian hamster embryo fibroblasts
treated with VBL or ADM. Neither VRP nor 1 or 3
sensitized 2SC/20 cells to ARA C or BLEO (not shown),
drugs that are not transported by Pgp. These data
indicate that a continuous (72 h) coexposure of 3 and 4
(at low micromolar concentrations) with Pgp-transport-
ed chemotherapeutics can reverse both intrinsic and
acquired Pgp-mediated drug resistance with an efficacy
similar or higher than that of 1 or VRP.
Finally, we addressed the ability of 1 and its conge-
ners to circumvent Pgp-mediated MDR in the course of
a shorter (24 h) coexposure with anticancer drugs. As
exemplified in Figure 4, VRP, 1, and 3 (all at 5 µM) were
not toxic for 2C/20 cells when administered alone for
24 h followed by withdrawal and further incubation for
an additional 48 h. ADM in combination with VRP or 3
Dr u g Resista n ce Rever sin g Activity of SDB-
Eth ylen ed ia m in e An a logu es. We further investi-
gated the capability of 1 and its close analogues to
overcome intrinsic (HET-SR-2SC-LNM cells) and ac-
quired (2SC/20 cells) drug resistance. Cell survival was
studied after both continuous coexposure of SDB-

Isoprenoid Analogues of SDB-Ethylenediamine
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24 5335
potently killed cells whereas 1 did not potentiate cell
death by ADM (Figure 4). This lack of synergistic
activity of 1 cannot be explained by insufficient intra-
cellular concentration of ADM because even by 4 h of
coincubation of 1 and ADM the level of intracellular
ADM was similar to that attained by combinations of
VRP + ADM or 3 + ADM (see Figure 3). It is tempting
to speculate that cotreatment of cells with 1 and ADM
somehow alters the intracellular distribution of the
latter drug, so ADM does not reach its targets. Whatever
the mechanism, the results establish the advantage of
3 over 1 in synergy with conventional chemotherapeu-
tics.
reversing agents to increase the uptake of the anticancer
drug is necessary but may not be sufficient for cell
killing.
The reason(s) for differential effects of 1 and 3 on
kinetics of intracellular ADM accumulation and poten-
tiation of ADM toxicity is unknown. For 1, the necessity
of prolonged treatment of cells for effective sensitization
may be partly due to high hydrophobicity and low
solubility of 1 in aqueous media. While VRP can form
true molecular solutions in water up to C ) 140 mM,
the surfactant-like 1 may exist in water as an equilib-
rium between molecular and micellar (colloidal) solu-
tions. The micellar fraction of 1 would serve as a pool
of free molecules (cations) capable of interfering with
Pgp. If the effective concentration of molecular 1 in
aqueous medium is low, the rate of this interaction
would be low too. Because the side chain in 3 is much
shorter, this compound should be relatively hydrophilic.
Consequently, the equilibrium between true and micel-
lar solutions would favor the molecular interaction of 3
with Pgp. This speculation is in agreement with experi-
mentally found log D values for dihydrochlorides of 1
and 3.
Discu ssion a n d Con clu sion s
Our results show that the modification of isoprenoid
chains in SDB-ethylenediamine 1 yielded readily ac-
cessible compounds with higher antitumor activity
against various tumor cell lines than the parental agent.
Two major biological characteristics of the novel con-
geners are particularly important. First, these novel
analogues were preferentially toxic for several MDR
tumor cells. Second, our novel compounds sensitized
MDR cells to conventional chemotherapeutic drugs. The
latter effect is due to the ability of novel compounds to
increase the intracellular uptake of chemotherapeutic
drugs that could be otherwise effluxed by multidrug
transporters.
Out of seven novel analogues of SDB-ethylenedi-
amine, the compounds 3 and 4 with relatively short
isoprenoid chains (C10 and C15, respectively) proved to
be the most efficient both when used in a single agent
treatment and in circumventing MDR when combined
with conventional chemotherapeutic drugs. The com-
pounds with longer or more saturated side chains were
less potent, suggesting that the length of isoprenoid side
chains and the extent of its saturation can be limiting
factors for pharmacological activity of this group of
antineoplastic drugs. Taking into consideration that the
analogues of 1 with shortened isoprenoid side chains
were synthesized using linalool (for 3 and 3a ) and
nerolidol (for 4 and 4a ) that are large-scale manufac-
tured chemicals, these MDR-reversing compounds war-
rant further investigation as potentially useful antin-
eoplastic agents.
These considerations might explain the differential
time course of the effects of 1 and 3 on intracellular
ADM concentration (Figure 3). The amount of ADM in
cells treated with ADM + 1 was low and transient,
peaking at 240 min and then declining by 360 min
whereas the MDR-reversing effect of 3 was fast (by 30
min of coexposure with ADM) and sustained for at least
360 min. One explanation for this difference is an
equilibrium between molecular and colloidal forms of 1
in water (see above), so the capability of 1 to interfere
with Pgp transport should be a function of time.
Alternatively (or in addition to), 1 could alter the
interaction of ADM with its intracellular targets, al-
lowing for easier washing of ADM out of the cell. Or
else 1 can influence compartmentalization of ADM. Most
importantly, 3, which is more hydrophilic due to shorter
hydrocarbon chains, increased ADM accumulation in a
faster and more stable manner than 1 (Figure 3). If this
fact is due to the predominance of true molecular form,
then the major principle of chemical modification of 1,
i.e., shortening its isoprene units for higher pharmaco-
logical potency, is validated.
Interestingly, 1 and 3, whose hydrocarbon chains are
longer and conformationally more flexible than the
isopropyl moiety of VRP, were more potent than VRP
for some of tested cell lines. This fact implies that the
specific features of the hydrocarbon side chain can
markedly influence the toxicity of Pgp blockers. The role
of structural and conformational modifications in phar-
macological function of hydrocarbon chains has since
been long-recognized in other areas of medicinal chem-
istry, e.g., for cannabinoids and ceramide analogues.44,45
The time course of biological effects and the range of
active concentrations for 3, 4, and VRP did not vary
dramatically. Indeed, 3 and VRP showed good syner-
gism as determined by the kinetics of ADM uptake by
MDR cells and viability analyses (Figure 3). In contrast,
the increase of ADM uptake by the prototypic compound
1 was slower (Figure 3); moreover, 1 did not potentiate
ADM toxicity for 2SC/20 cells even after a 24 h coex-
posure (Figure 4). These data suggest that use of MDR-
Considering the role of the isoprenoid side chain in
MDR-circumventing activity, it is worth noting that
monophytyl derivative 5 was inactive. Possibly, the
reduction of three double bonds in the acyclic C20
isoprenoid causes a modification of the shape (the
bulkiness) and an opportunity to establish an electronic
interaction (e.g., π-π interaction) with the target.
Reduced biological activity after saturation of double
bonds has been demonstrated earlier.44,45
Clinical use of MDR-reversing drugs is often limited
by their side effects such as toxicity for normal tis-
sues.4,46 It remains to be elucidated whether close
analogues of 1 retain the in vivo effects of their
prototype. Significantly, for cultured cells, 3‚2HCl and
4‚2HCl were potent at concentrations that were lower
than toxic doses of 1‚2HCl or VRP (Tables 4 and 5).
Moreover, low micromolar concentrations of 3‚2HCl and
4‚2HCl were sufficient for MDR reversal. These findings

5336 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24
Sidorova et al.
with Et2O. The ethereal layer was washed with water, dried
(Na2 SO4), and concentrated in vacuo to leave a brownish oil
that was fractionated by flash column chromatography on SiO2
(30 g, 40-100 µm) using a heptane-AcOEt gradient (0f25%
AcOEt) as the eluent. Less polar fractions afforded compound
1a (238 mg, 12%) as a yellowish viscous oil. TLC: Rf ) 0.67.
FAB-MS: [M]+ ) 1586 (C110H172N2O4). 1H NMR: 1.6-1.62 (m,
54 H, Me groups); 1.7 (s, 6H, Me groups); 1.9-2.2 (m, 64 H,
allylic CH2 groups); 2.58 (br s, 4 H, N-CH2CH2-N moiety);
3.08 (d, 4 H, J ) 6 Hz, two NCH2CHdC); 3.49 (br s, 4 H, two
NCH2Ar); 3,85 (s, 6H, two OMe); 3.87 (s, 6 H, two OMe); 5.12-
5.14 (overlap of two broad triplets with J ) 6 Hz, 16 H, internal
CH2CH)C); 5.30 (br t, 2 H, J ) 6 H, two NCH2CH)C), 6.7-
6.95 (m, 6 H, Ar-H). Further elution yielded diamine 1 as a
make the compounds 3 and 4 attractive candidates for
further studies.
In the present study, we used VRP, the classical
MDR-reversing drug structurally related to 1, as a
major reference agent for MDR-reversing effects of our
novel compounds. Although a detailed comparison of the
potency of novel derivatives of 1 with other Pgp blockers
was not the major goal of this work, the efficient doses
of our compounds can be matched with the reported
potency of other MDR-modulating drugs. The leading
agent 3 reversed MDR at concentrations (5 µM, Table
6) similar to those used in the experiments with SDZ
PSC 833, a nonimmunosuppressive cyclosporin cur-
rently in clinical trials.47 These results are encouraging
in terms of the possibility to synthesize relatively
inexpensive and potent agents for MDR circumvention.
Importantly, dihydrochlorides of 1, 3, and 4 were more
toxic to cells selected for MDR than to their wild-type
counterparts, whereas no difference was found for the
cells that express Pgp without selection (Table 6).
Preferential toxicity against MDR cells has been dem-
onstrated for Pgp blockers of different chemical struc-
ture.48 This implies that Pgp is important for the
physiology of resistant cells, and its inhibition is clini-
cally relevant. Regardless of the fact that in drug-
selected cells, mechanisms additional to drug transport-
ers can cooperate to ensure cell survival, selection for
MDR can confer sensitivity to compounds such as short
chain isoprenoid analogues of 1. The phenomenon of
collateral sensitivity, “the price paid by the cell” for
adaptation to the toxic environment, indicates that the
functionality of some death pathways is preserved even
after prolonged drug exposure.49,50 The agents that are
preferentially toxic for MDR cells may be valuable tools
for dissecting mechanisms of apoptosis and/or necrosis
in drug resistant cancer. Directed targeting of these
mechanisms should improve clinical outcome in patients
with chemotherapy resistant malignancies.
viscous oil (1.034 g, 85%). TLC: Rf ) 0.27. FAB-MS: [M]+
)
973 (C65H100N2O4). 1H NMR: 1.6 (br s, 27 H, Me-groups); 1.70
(s, 3 H); 1.95-2.15 (m, 32 H, allylic CH2 groups); 2.62 and 2.70
(dt, 4 H, NCH2CH2N′); 3.08 (d, 2 H, J ) 6 Hz, NCH2CHdC);
3.50 (s, 2 H, NCH2Ar); 3.7 (s, 2 H, N′CH2Ar); 3.8 (br s, 1H,
NH); 3.86 (s, 6 H, two OMe); 3.88 (s, 6 H, two OMe); 5.1 (br t,
8 H, CH2CH dC); 5.3 (br t, 1 H, J ) 6 Hz, NCH2CH)C); 6.7-
6.95 (m, 6 H, Ar-H).
Syn t h esis of Com p ou n d s 4, 5 a n d 4a , 5a . Gen er a l
P r oced u r e.17,21 To a magnetically stirred solution of diamine
2 (720 mg, 2 mM) in dry THF-Et2O (4:1.5 mL), a solution of
an isoprenoid bromide 8 or 9 (1 mM) in dry Et2O (2 mL) was
added at ∼20 °C within 45 min, and the stirring was continued
for an additional 3 h. The reaction mass was diluted with water
(5 mL) and extracted with CHCl3 (2 × 5 mL). The extract was
thoroughly washed with a saturated aqueous solution of Na2-
CO3 and water, dried (Na2SO4), and concentrated (25-30 °C,
40 Torr) to the permanent weight of the residue. Flash column
chromatography on SiO2 using heptane-AcOEt (20f100%)
and then AcOEt-MeOH (5f25%) gradients as eluents af-
forded first bis-alkylated diamines (4a or 5a ) and subsequently
the corresponding monoalkylated diamines (4 or 5). Finally,
the excess of starting diamine 2 (mp 102-103 °C) was eluted.
All products were isolated as free bases. Prolonged chroma-
tography resulted in somewhat darker specimens, although
no contaminants were revealed by routine 1H NMR control.
N,N′-Bis(3,4-d im et h oxyb en zyl)-N-(2E/Z,6E)-fa r n esyl-
eth ylen ed ia m in e (4). Yield, 141 mg (50%). A brownish,
glassy gum. Rf 0.25. FAB-MS: [M]+ )564 (C35H52N2O4). 1H
NMR: 1.60-1.62 (overlap of three singlets, 9 H, Me groups);
1.68 (s, 12-H3); 1.95-2.12 (m, 8 H, allylic CH2); 2.62-2.70 (dt,
4 H, J ∼6 Hz, N-CH2CH2-N′ moiety); 3.08 (d, 2 H, NCH2-
CH)); 3.5 and 3.66 (two s, 4 H, ArCH2N and ArCH2N′); 3.82
(br s, 1 H, NH); 3.88 and 3.90 (two s, 6 H + 6 H, OMe groups);
5.1 (br t, 2 H, CH2CHd); 5.31 (br t, 1 H, NCH2CHd); 6.75-
6.95 (m, 6 H, Ar-H).
N,N′-Bis(3,4-d im et h oxyb en zyl)-N,N′-d i-(2E/Z,6E)-fa r -
n esyleth ylen ed ia m in e (4a ). Yield, 36.5 mg (19%). A brown-
ish solid foam. Rf 0.60. FAB-MS: [M + 1]+ )769 (C50H76N2O4).
1H NMR: 1.61-1.63 (overlapping singlets, 18 H, internal Me);
1.69 (s, 6 H, terminal Me); 1.93-2.12 (m, 16 H, dCCH2); 2.59
(m, 4 H, N-CH2CH2-N); 3.07 (d, 4 H, J ) 6.5 Hz, NCH2CHd
); 3,49 (br s, 4 H, NCH2Ar); 3.85 and 3.87 (s + s, 12 H, OMe);
5.12 (br t, 4 H, CH2CHd); 5.3 (br t, 2 H, NCH2CHd); 6.73-
6.96 (m, 6 H, Ar-H).
Exp er im en ta l Section
N,N′-Bis(3,4-dim eth oxyben zyl)eth ylen ediam in e (2). The
mp 102-104 °C (acetone) was obtained according to ref 17
(Scheme 1). Solanesyl bromide (6) was prepared from crystal-
line SolOH (mp 38-41 °C, acetone at -10 °C) according to a
stereoselective procedure.51 The same method was used for
converting racemic isophytol (a technical grade product puri-
fied by preparative thin-layer chromatography (TLC) on SiO2)
into phytyl bromide (9) and for converting (6E)-nerolidol52 into
(2E/Z,6E)-farnesyl bromide (8). “Geranyl” chloride (7) was
obtained from linalool and concentrated HCl by analogy with
the preparation of farnesyl chlorides from nerolidols.51 Chemi-
cal purity of all synthesized compounds was controlled by TLC
on Silufol sheets using AcOEt-MeOH (4:1, v/v) for developing,
by 1H NMR spectroscopy (CDCl3, Bruker AM-300 instrument),
and by FAB mass spectrometry (Cratos 50-TS instrument,
8000 eV, xenon as a reagent gas, glycerol as a solvent; in the
case of 1a , a mixture of glycerol with thioglycerol was used).
SDB-Eth ylen ed ia m in e (1) a n d N,N′-Bis(3,4-d im eth oxy-
ben zyl)-N,N′-d isola n esyleth ylen ed ia m in e (1a ). To a solu-
tion of naphthalene (100 mg, 0.78 mM) in absolute tetrahy-
drofuran (THF, 2 mL), lithium cuttings (10 mg, 1.1 mM) were
added at ∼20 °C under the atmosphere of argon, and the
mixture was magnetically agitated for 1 h. Then, diamine 2
(450 mg, 1.25 mM) was slowly added, the stirring was
continued for 2 h, and the resulting amide was gradually
quenched with bromide 6 (900 mg, 1.3 mM). The reaction
mixture was agitated for 1 h at ∼20 °C, left overnight,
decomposed with MeOH, diluted with water, and extracted
N,N′-Bis(3,4-d im et h oxyb en zyl)-N-(2E/Z)-p h yt ylet h yl-
en ed ia m in e (5). Yield, 344 mg (54%). A brownish gum with
Rf 0.22. FAB-MS: 624 ([M - CH3] ion. 1H NMR: 0.9-1.12
(overlapping doublets, 12 H, CH-Me); 1.24-1.53 (m, 17 H, CH2
and CH); 1.62 >1.66 (two s, 3 H, Me-Cd); 1.97 (m, 2 H, allylic
CH2); 2.46-2.6 (m, 4 H, N-CH2CH2-N′); 2.97 (d, 2 H, J ) 6
Hz, NCH2CHd); 3.42 (s, 2 H, ArCH2N); 3.50 (s, 2 H, ArCH2N′);
3.83 and 3.87(s + s, 12 H, OMe); 5.23 (t, 2 H, NCH2CHd);
6.75-6.97 (m, 6 H, Ar-H).
N,N′-Bis(3,4-d im et h oxyben zyl)-N,N′-d i-(2E/Z)-p h yt yl-
eth ylen ed ia m in e (5a ). Yield, 130 mg (14.2%). A brownish
gum with Rf 0.51. FAB-MS: 902 ([M - CH3] ion. 1H NMR:
0.9-1.1 (m, 24 H, Me); 1.25-1.55 (m, 34 H, CH2 and CH);
1.63 > 1.67 (d, 4 H, J ) 1.5 Hz, Me-Cd); 2.5 (m, 4 H,
N-CH2CH2-N′); 2.9-3.05 (br d, J ∼6 Hz, NCH2CHd);

Isoprenoid Analogues of SDB-Ethylenediamine
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24 5337
3.45 (s, 4 H, ArCH2N); 3.83-3.86 (s + s, 12 H, OMe); 5.3
(t, 2 H, NCH2CHd); 6.8-7.0 (m, 6 H, Ar-H).
3 in AcOEt (washed to neutrality and redistilled before use)
was mixed with an equal volume of buffer and shaken on a
reciprocal shaker for 1 h at room temperature. The mixture
was left for 12 h in a separation funnel to attain a clean
exfoliation. The two phases were carefully separated, and the
organic phase was transferred into a calibrated flask and
concentrated to permanent weight at 45-50 °C (bath)/3 Torr.
The weight of the residue was determined using an OHAUS
Explorer balance (readability 0.1 mg). The mass of the solute
in the organic layer was calculated as mAcOEt ) w - c, where
w is the weight of the residue and c is the standard weight of
inorganic contaminant(s) from the aqueous buffer (0.2 mg at
pH 6.4 and 0.3 mg at pH 4.0). The content of dihydrochlorides
The reaction of diamine 2 with solanesyl bromide 6 in pure
Et2O (∼20 °C, 45 min + 3 h) proceeded similarly to give
compounds 1 and 1a with yields 43 and 12%, respectively.
N,N′-Bis(3,4-d im eth oxyben zyl)-N,N′-(2E/Z)-ger a n yleth -
ylen ed ia m in e (3). A solution of diamine 2 (720 mg, 2 mM)
in dry THF (4 mL) and a solution of (2E/Z)-geranyl chloride
(173 mg, 1 mM) in dry Et2O (2 mL) were mixed and refluxed
for 4 h. The reaction mixture was diluted with water (5 mL)
and extracted with CHCl3 (2 × 5 mL). The extract was washed
with an aqueous solution of Na2CO3 (until gas production
stopped) and water, dried (Na2SO4), and concentrated in vacuo.
The column chromatography on SiO2 using the AcOEt-MeOH
(0f25%) gradient as an eluent gave 6 mg of a yellowish oil
with Rf 0.53 that was tentatively identified as the compound
3a (yield < 1%). Further elution afforded the compound 3 as
a yellowish gum with Rf 0.11. Yield, 417 mg (84%). FAB-MS:
of 1 or 3 in the aqueous phase was calculated as mwater
m0 - mAcOEt, where m0 is the mass of the analyzed specimen.
)
Cell Lin es a n d Dr u gs. The HET-SR-2SC-LNM Syrian
hamster embryo fibroblast cell line (a gift of G. I. Deichman,
Moscow, Russia) and its isogenic Pgp positive variant 2SC/20
(refs 33 and 34) were propagated in Dulbecco’s modified Eagle’s
medium (Gibco BRL, Grand Island, NY) supplemented with
5% fetal bovine serum (Gibco), 2 mm L-glutamine, 100 U/mL
penicillin, and 100 µg/mL streptomycin at 37 °C, 5% CO2 in a
humidified atmosphere. Other cell models included human
melanoma cell line mS, its MDR variant mS/0.5, and rat
hepatoma McA RH 7777 and its MDR subline McA 7777/0.4.
The mS/0.5 and McA 7777/0.4 variants were obtained by us
after a stepwise selection of respective wild-type cell lines for
Pgp-mediated MDR.39,40 In addition, we used K562 human
leukemia cell line (American Type Culture Collection, Manas-
sas, VA) and its subline K562/i-S9. The latter variant expresses
Pgp after infection of K562 cells with a retrovirus carrying
full-length human MDR1 cDNA followed by flow cytometry-
based sorting of Pgp positive cells (gift of I. B. Roninson,
Chicago).42,43 The K562/i-S9 cells have not been selected with
any drug. For studies of Pgp-unrelated drug resistance, we
used human lung carcinoma cell line COR L23/P (parental)
and its ADR-selected subline COR L23/R (an MRP but not Pgp-
mediated resistance41 (provided by P. Twentyman, Cambridge,
U.K.). All nonfibroblast cell lines were maintained in RPMI-
1640 supplemented as above. The selective agent was with-
drawn from cultures for 2 days prior to the experiments. The
cells were routinely tested and found free from mycoplasma.
The cultures in the logarithmic phase of growth were used in
all experiments.
Dihydrochloride of SDB-ethylenediamine and dihydrochlo-
rides of 1a , 2-5, 4a , 5a , and VRP (Sigma Chemical Co., St.
Louis, MO) were dissolved in DMSO. ARA C (Upjohn, Bel-
gium), BLEO (Nippon Kayaki Co., Ltd., J apan), VBL (G.
Richter, Hungary), and ADM (Bristol Myers Squibb Pharma-
ceuticals, Princeton, NJ ) were reconstituted in deionized water.
All chemicals were prepared as 100-1000× stock solutions
immediately before the experiments. When the cells were
treated with the compounds dissolved in DMSO, the final
concentration of the solvent in the medium was 0.1%. In our
preliminary experiments, this dose of DMSO did not influence
the activity of Pgp or MRP as compared with untreated cells.
This dose also did not cause any discernible cytotoxicity,
growth arrest, or morphological changes in cells within the
time frame of experiments (not shown). [3H]VBL (specific
activity, 2-10 Ci/mM) was purchased from Amersham, U.K.
1
[M]+ ) 496 (C30H44N2O4). H NMR: 1.60 (s, 6 H, internal Me
groups); 1.68 (s, 3 H, 8-H3); 1.95-2.15 (m, 4 H, allylic CH2);
2.62 and 2.7 (t, 4 H, N-CH2CH2-N′); 3.08 (d, J ) 6 Hz, NCH2-
CHd); 3.5 and 3.66 (s + s, 4 H, NCH2Ar and N′CH2Ar); 3.83
(br s, 1 H, NH); 3.87-3.88 (s, 12 H, OMe); 5.10 (br t, 1 H, J )
6 Hz, CHd); 5.3 (br t, 1 H, J ) 6.5 Hz, NCH2CHd); 6.75-6.95
(m, 6 H, Ar-H).
Tr a n sfor m a tion of F r ee Ba ses in to Cor r esp on d in g
Dih yd r och lor id es. Gen er a l P r oced u r e. In contrast to
diamine 2, its isoprenylated derivatives were insoluble in
water as free bases. To increase the solubility of compounds
1, 1a , 3, 4, 4a , 5, and 5a in DMSO-water, their dihydrochlo-
rides were prepared by adding an anhydrous ethereal solution
of a base to a saturated solution of dried gaseous HCl in dry
Et2O at 5-10 °C. This was followed by either collecting the
resulting precipitates on a porous (4-5.5 mm) Pyrex funnel
or by evaporating the solutions in vacuo. The specimens thus
obtained were dried for 4 h in a rotary evaporator at 45-50
°C/5 Torr (oil bath) and kept over CaCl2 in a desiccator.
Dihydrochlorides of 1, 1a (mp ∼40 °C), 3 (mp ∼110 °C,
decomposed; creamy-white solid from cold EtOH), and 4a (mp
∼104 °C, decomposed) were yellowish, paraffin-like substances.
Dihydrochlorides of 4 and 5 were brownish solid foams with
imprecise melting temperatures. The solubility of dihydro-
chlorides (mol/L) in DMSO-H2O at 20 °C was determined
nephelometrically by diluting their stock solutions in DMSO
with distilled water until faint turbidity appeared. The results
are shown below.
DMSO:H2O
ratio (v/v)
in the
base in [base‚2HCl]
solvent
1a
5a
4a
1
5
4
3
10:90
2:98
2.5×10-5 3×10-4 5×10-4 6×10-4 nd
nd
nd
nda
nd
nd
nd
3×10-3 >1.6×10-3 1.8×10-3
a
nd, not determined.
The dihydrochloride of parental diamine 2 was soluble in
water, DMSO, and EtOH. When stored in sealed tubes in a
refrigerator at -15 °C for 6-8 months, dihydrochlorides of 1,
1a , 3, 4, and 4a slowly decayed with the formation of artifacts
detectable by TLC and 1H NMR (diminished intensity of
signals attributed to dCH at δ ∼5.2 and allylic methyl groups
at δ ∼1.6-1.7), which may be due to isomerization or polym-
erization induced by traces of HCl. After 2 years of storage,
the decay was nearly complete (TLC and 1H NMR data).
Deter m in a tion of th e Distr ibu tion Coefficien ts. The
distribution coefficients of dihydrochlorides of 1 and 3 between
ethyl acetate and two aqueous phases were determined by the
shake-flask method. The aqueous phases consisted of (i) 0.1
M phosphate buffer prepared from Na2HPO4 and NaH2PO4 in
distilled water, pH 6.4, and (ii) 0.1 M phosphate buffer made
from Na2HPO4 and H3PO4 in distilled water, pH 4.0. The
aqueous and organic phases were saturated with each other.
Ten milliliters of a 1 mm solution of dihydrochlorides of 1 or
Cytotoxicity Assa ys. The toxicity of chemotherapeutic
drugs and novel compounds was determined in a MTT test38
or by direct count of cells after drug exposure. For the MTT
test, cells ((3-4) × 103 in 100 µL of culture medium) were
plated into a 96 well plate (Becton Dickinson, Franklin Lakes,
NJ ). After 16 h, cells were either mock-treated (vehicle control)
or treated with increasing concentrations of VBL, ADM, BLEO,
or ARA C (in duplicate) in the absence or presence of 1‚2HCl
or its close analogues for 72 h. In the experiments with
leukemia cells and lung carcinoma cells, the initial cell
densities were 3 × 104 and 5 × 103 in 200 µL of culture
medium, respectively. After the drug exposure was completed,
20 µL of aqueous MTT solution (Sigma; 2 mg/mL) was added
into each well for an additional 3 h. Formazan was dissolved

5338 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24
Sidorova et al.
(12) Ikezaki, K.; Yamaguchi, T.; Miyazaki, C.; Ueda, H.; Kishie, T.;
Tahara, Y.; Hiroyasu, K.; Takahashi, T.; Fukawa, H.; Komiyama,
S.; Kuwano, M. Potentiation of anticancer agents by new
synthetic isoprenoids. I. Inhibition of the growth of cultured
mammalian cells. J . Natl. Cancer Inst. 1984, 73, 895-901.
(13) Yamaguchi, T.; Ikezaki, K.; Kishiye, T.; Tahara, Y.; Koyama,
H.; Takahashi, T.; Fukawa, H.; Komiyama, S.; Akiyama, S.;
Kuwano, M. Potentiation of anticancer agents by new synthetic
isoprenoids. II. Inhibition of the growth of transplantable murine
tumors. J . Natl. Cancer Inst. 1984, 73, 903-907.
(14) Nakagawa, M.; Akiyama, S.; Yamaguchi, T.; Shiraishi, N.; Ogata,
J .; Kuwano, M. Reversal of multidrug resistance by synthetic
isoprenoids in the KB human cancer cell line. Cancer Res. 1986,
46, 4453-4457.
(15) Yamaguti, T.; Nakagawa, M.; Shiraishi, N.; Yoshida, T.; Kiyosue,
T.; Arita, M.; Akiyama, S.; Kuwano, M. Overcoming drug
resistance in cancer cells with synthetic isoprenoids. J . Natl.
Cancer Inst. 1986, 76, 947-953.
(16) Akiyama, S.; Cornwell, M. M.; Kuwano, M.; Pastan, I.; Gottes-
man, M. M. Most drugs that reverse multidrug resistance inhibit
photoaffinity labeling of P-glycoprotein by a vinblastine ana-
logue. Mol. Pharmacol. 1988, 33, 144-147.
(17) Akiyama, S.; Yoshimura, A.; Kikuchi, H.; Sumizawa, T.; Kuwano,
M.; Tahara, Y. Synthetic isoprenoid photoaffinity labeling of
P-glycoprotein specific to multidrug-resistant cells. Mol. Phar-
macol. 1989, 36, 730-735.
(18) Ogawa, O.; Kishi, I.; Tahara, Y.; Sugita, M. SDB-Ethylenedi-
amine malate as anticancer activity potentiator. Eur. Pat. Appl.
EP 355604 1990; Chem. Abstr. 1991, 114, 95153m.
(19) Inomata, K.; Takahashi, T.; Inoue, H.; Yanai, M.; Yamazaki, H.;
Suzuki, M.; Takasawa, T.; Kawamura, K.; Oshida, N.; Ikemoto,
H.; Kishie, T. Preparation of terpene ethylenediamines as
multidrug resistance inhibitors. Eur. Pat. Appl. EP 781552 1997;
Chem. Abstr. 1997, 127, 121902.
(20) Tahara, Y.; Komatsu, Y.; Koyama, H.; Kubota, R.; Yamaguchi,
T.; Takahashi, T. Isoprenylamine derivatives and their phar-
maceutical compositions. Fr. Demande FR 2505826 1982; Chem.
Abstr. 1983, 98, 215838.
(21) Tawara, Y.; Mishima, Y. Preparation of N-solanesyl-N,N′-bis-
(3,4-dimethoxybenzyl)ethylenediamine. J apan Pat. J P 03/02150
1991; Chem. Abstr. 1991, 114, 247565.
(22) Suzuki, H.; Tomida, A.; Nishimura, T. Cytocidal Activity of a
synthetic isoprenoid, N-solanesyl-N,N′-bis-(3,4-dimethoxyben-
zyl)ethylenediamine, and its potentiation of antitumor drugs
against multidrug-resistant and sensitive cells in vitro. J pn. J .
Cancer Res. 1990, 50, 298-303.
(23) Tomida, A.; Tatsuta, T.; Suzuki, H. Novel mechanism of N-solane-
syl-N,N′-bis-(3,4-dimethoxybenzyl)ethylenediamine in potentia-
tion of antitumor drug action on multidrug-resistant and
sensitive Chinese hamster cells. J pn. J . Cancer Res. 1991, 51,
127-133.
(24) Tomida, A.; Suzuki, H. Synergistic effect in culture of bleomycin-
group antibiotics and of N-solanesyl-N,N′-bis(3,4-dimethoxyben-
zyl)ethylenediamine, a synthetic isoprenoid. J pn. J . Cancer Res.
1990, 50, 1184-1190.
(25) Akiyama, S.; Shiraishi, N.; Yoshimura, A.; Nakagawa, M.;
Yamaguchi, T.; Shimada, T.; Kuwano, M. Enhancement by a
synthetic isoprenoid of the toxicity of conjugates of epidermal
growth factor with Pseudomonas exotoxin. Biochem. Pharmacol.
1987, 36, 861-868.
(26) Ogihara, A.; Aiba, K.; Shibata, H.; Inoue, K.; Horikoshi, N.;
Hirano, A.; Yamazaki, H.; Yoshida, K.; Kuraishi, Y. An inves-
tigation of mechanisms of enhanced cytotoxicity of 5-fluorouracil
by a synthetic isoprenoid for human colon cancer cell lines. Proc.
Am. Assoc. Cancer Res. 1994, 35, abstract A1910.
in acidified DMSO, and the absorbency at λ ) 540 nm was
measured on a Flow Multiscan plate reader (LKB, Sweden).
In some experiments, cells (5 × 104 in 1 mL of medium) were
treated for 24 h with ADM in the absence or presence of tested
isoprenoid analogue, washed with ice-cold phosphate-buffered
saline (PBS), pH 7.4, and further incubated in drug-free
medium for 48 h. The cells were detached from plastic using
1 mM ethylenediaminetetraacetic acid disodium salt in PBS
and counted in a hemocytometer, and the percentage of viable
cells (by Trypan blue exclusion) was calculated.
Dr u g Up ta k e Stu d ies. The effects of 1‚2HCl and its
isoprenoid analogues on intracellular accumulation of VBL or
ADM were determined in the in vitro drug uptake assay.36
HET-SR-2SC-LNM or 2SC/20 cells (5 × 105 per well in a 24
well plate) were treated with [3H]VBL (0.037 MBq, 10 nM) in
1 mL of culture medium in the absence or presence of 1‚2HCl
or its analogues for 30 min at 37 °C, 5% CO2. Cells were then
washed three times with PBS and lysed in 0.5 N NaOH
overnight, and radioactivity was counted on a â-counter (LKB,
Sweden). Data were normalized by total protein content as
determined by the Lowry method.53 For ADM uptake, cells
were incubated with 10 µM ADM for 30-360 min, then lysed
in 1 mL of 0.3 N HCl in 50% ethanol, and centrifuged (100g,
30 min). The fluorescence of supernatants was assayed on a
Specord M-40 spectrophotometer (excitation at λ ) 470 nm,
emission at λ ) 585 nm). All treatments were performed in
triplicate. Statistical analysis was performed using Student’s
t-test. Data are expressed as mean ( SE of 3-4 experiments.
Ack n ow led gm en t. We are grateful to G. I. Deich-
man (N. Blokhin Cancer Center, Moscow, Russia), I. B.
Roninson (University of Illinois, Chicago, U.S.A.), and
P. R. Twentyman (MRC Centre, Cambridge, U.K.) for
generous gifts of cell cultures. This study was supported
by Grants 96-15-97461 and 00-15-97347 of the Federal
Program of Supporting the Leading National Scientific
Schools and 99-04-48617 of the Russian Foundation for
Basic Research.
Refer en ces
(1) Roninson, I. B. P-Glycoprotein-mediated drug resistance: Puzzles
and perspectives. In Molecular and Cellular Biology of Multidrug
Resistance in Tumor Cells; Roninson, I. B., Ed.; Plenum Press:
New York and London, 1991; pp 395-402.
(2) Dano, K. Active outward transport of daunorubicin in resistant
Ehrlich ascites tumor cells. Biochim. Biophys. Acta 1973, 323,
466-483.
(3) Gottesman, M. M.; Pastan, I.; Ambudkar, S. V. P-Glycoprotein
and multidrug resistance. Curr. Opin. Genet. Dev. 1996, 6, 610-
617.
(4) Volm, M. Multidrug resistance and its reversal. Anticancer Res.
1998, 18 (4C), 2905-2918.
(5) Sheehan, D.; Meade, G. Chemical modification of chemotherapy
resistance. Biochem. Soc. Trans. 2000, 28, 27-32.
(6) Dorr, R.; Karanes, C.; Spier, C.; Grogan, T.; Moore, J .; Wein-
berger, B.; Schiller, G.; Pearce, T.; Litchman, M.; Dalton, W.;
Roe, D.; List, A. Phase I/II study of the P-glycoprotein modulator
PSC 833 in patients with acute myeloid leukemia. J . Clin. Oncol.
2001, 19, 1589-1599.
(27) Takahashi, Y.; Urade, M.; Kishimoto, H.; Arimoto, T.; Yanag-
isawa, T.; Yoshioka, W. Combined effect of bleomycin and SDB-
ethylenediamine, a synthetic isoprenoid, on oral squamous
carcinoma cell lines and transplantable nude mouse tumors. Gan
To Kagaku Ryoho 1995, 22, 883-887.
(28) Suga, K.; Watanabe, S.; Fujita, T.; Pan, T. P. N-Alkylation of
primary and secondary amines by alkyl halides and lithium
naphthalene. Bull. Chem. Soc. J pn. 1969, 42, 3606-3609.
(29) Baky, H.; Kirsten, E. B. Verapamil. In Pharmacological and
Biochemical Properties of Drug Substances; Goldberg, D., Ed.;
American Pharmaceutical Association: Washington, DC, 1981;
Vol. 3, pp 226-261.
(7) Tsuruo, T.; Iida, H.; Tsukagoshi, S.; Sakurai, Y. Overcoming of
vincristine resistance in P388 leukemia in vivo and in vitro
through enhanced cytotoxicity of vincristine and vinblastine by
verapamil. Cancer Res. 1981, 41, 1967-1972.
(8) Yusa, K.; Tsuruo, T. Reversal mechanism of multidrug resistance
by verapamil: Direct binding of verapamil to P-glycoprotein on
specific sites and specific transport of verapamil outward across
the plasma membrane of K562/ADM cells. Cancer Res. 1989,
49, 5002-5006.
(9) Safa, A. R. Photoaffinity labeling of multidrug resistance-related
P-glycoprotein with photoactive analogues of verapamil. Proc.
Natl. Acad. Sci. U.S.A. 1988, 85, 7187-7191.
(10) Klopman, G.; Srivastava, S.; Kolossvary, I.; Epand, R. F.; Ahmed,
N.; Epand, R. M. Structure-activity study and design of mul-
tidrug-resistant reversal compounds by a computer automated
structure evaluation methodology. Cancer Res. 1992, 52, 4121-
4129.
(30) Leo, A. J . Relationship between partitioning solvent systems.
In Biological CorrelationssThe Hansch Approach; Advanced
Chemistry Series; American Chemical Society: Washington, DC,
1972; Vol. 114; pp 51-60.
(31) Franke, U.; Munk, A.; Wiese, M. Ionization constants and
distribution coefficients of phenothiazines and calcium channel
antagonists determined by a pH-metric method, and correlation
(11) Abraham, I.; Hester, J . B. Use of amines to sensitize multidrug-
resistant cells. PCT Int. Appl. WO 92/18089 1992; Chem. Abstr.
1993, 118, 52416q.

Isoprenoid Analogues of SDB-Ethylenediamine
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 24 5339
with calculated partition coefficients. J . Pharm. Sci. 1999, 88,
89-95.
(43) Mechetner, E. B.; Schott, B.; Morse, B. S.; Stein, W. D.; Druley,
T.; Davis, K. A.; Tsuruo, T.; Roninson, I. B. P-Glycoprotein
function involves conformational transitions detectable by dif-
ferential immunoreactivity. Proc. Natl. Acad. Sci. U.S.A. 1997,
94, 12908-12913.
(44) Keimowitz, A. R.; Martin, B. R.; Razdan, R. K.; Crocker, P. J .;
Mascarella, S. W.; Thomas, B. F. QSAR Analysis of ∆8-THC
analogues: Relationship of side-chain conformation to cannab-
inoid receptor affinity and pharmacological potency. J . Med.
Chem. 2000, 43, 59-70.
(45) Igarashi, J .; Thatte, H. S.; Prabhakar, P.; Golan, D. E.; Michel,
T. Calcium-independent activation of endothelial nitric oxide
synthase by ceramide. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
12583-12588.
(46) Ford, J . M.; Hait, W. N. Pharmacology of drugs that alter multi-
drug resistance in cancer. Pharmacol. Rev. 1990, 42, 155-199.
(47) Komarov, P. G.; Shtil, A. A.; Holian, O.; Tee, L.; Buckingham,
L.; Mechetner, E. B.; Roninson, I. B.; Coon, J . S. Activation of
the LRP (lung resistance-related protein) gene by short-term
exposure of human leukemia cells to phorbol ester and cytara-
bine. Oncol. Res. 1998, 10, 185-192.
(32) Pearce, H. L.; Winter, M. A.; Beck, W. T. Structural character-
istics of compounds that modulate P-glycoprotein-associated
multidrug resistance. Adv. Enzyme Regul. 1990, 30, 357-373.
(33) Deichman, G. I.; Topol, L. Z.; Kluchareva, T. E.; Matveeva, V.
A.; Zakamaldina, T. A.; Uvarova, E. N.; Tatosyan, A. G. Cluster-
ing of discrete cell properties essential for tumorigenicity and
metastasis. III. Dissociation of the properties in N-ras-trans-
fected RSV-SR-transformed cells. Int. J . Cancer 1992, 51, 903-
908.
(34) Shtil, A.; Shushanov, A.; Moynova, E.; Stavrovskaya, A. fre-
quency of metastasis in Syrian hamster tumor cells selected for
low levels of “typical” multidrug resistance. Exp. Toxicol. Pathol.
1994, 46, 257-262.
(35) Minor, P. D.; Roscoe, D. H. Colchicine resistance in mammalian
cell lines. J . Cell Sci. 1975, 17, 381-396.
(36) Erokhina, M. V.; Shtil, A. A.; Shushanov, S. S.; Sidorova, T. A.;
Stavrovskaya, A. A. Partial restoration of the actin cytoskeleton
in transformed Syrian hamster fibroblasts selected for low
levels of “typical” multidrug resistance. FEBS Lett. 1994, 341,
295-298.
(37) Beck, W. T.; Danks, M. K. Characteristics of multidrug resistance
in human tumor cells. In Molecular and Cellular Biology of
Multidrug Rresistance in Tumor Cells; Roninson, I. B., Ed.;
Plenum Press: New York and London, 1991; pp 3-55.
(38) Mossman, T. Rapid Colorimetric assay for cellular growth and
survival: Application to proliferation and cytotoxicity assays.
J . Immunol. Methods 1983, 65, 55-63.
(48) Lehne, G.; De Angelis, P.; den Boer, M.; Rugstad, H. E. Growth
inhibition, cytokinesis failure and apoptosis of multidrug-
resistant leukemia cells after treatment with P-glycoprotein
inhibitory agents. Leukemia 1999, 13, 768-778.
(49) Blagosklonny, M. V. Drug-resistance enables selective killing of
resistant leukemia cells: Exploiting of drug resistance instead
of reversal. Leukemia 1999, 13, 2031-2035.
(50) Shtil, A. A.; Turner, J . G.; Durfee, J .; Dalton, W. S.; Yu, H.
Cytokine-based tumor cell vaccine is equally effective against
parental and isogenic multidrug-resistant myeloma cells: the
role of cytotoxic T-lymphocytes. Blood 1999, 93, 1831-1837.
(51) Grigor’eva, N. Ya.; Avrutov, I. M.; Semenovsky, A. V.; Odinokov,
V. N.; Akhunova, V. R.; Tolstikov, G. A. Synthons for the syn-
thesis of acyclic polyisoprenoids from isomeric ethyl farnesoates.
Bull. Acad. Sci. USSR, Div. Chem. Sci. 1979, 38, 353-359.
(52) Nigmatov, A. G.; Serebryakov, E. P.; Yanovskaya, L. A. Improved
procedure for producing geranyl esters of (4E/Z,8E)- and
(4E/Z,8Z)-farnesylacetic acids. Khim.-Farm. Zh. 1987, 21,
854-858 (in Russian).
(39) Alekseevskaya, O. D.; Anfimova, M. L.; Djuraeva, F. H.; Somova,
O. V.; Stavrovskaya, A. A.; Stromskaya, T. P.; Shtil, A. A.;
Eraiser, T. L. Alterations of embryo-specific protein expression
in multidrug resistant tumor cells. Her. Cancer Res. Center Russ.
Acad. Med. Sci. 1993, 2, 21-30 (in Russian).
(40) Stromskaya, T. P.; Filippova, N. A.; Rybalkina, E. Yu.; Egudina,
S. V.; Shtil, A. A.; Eliseenkova, A. V.; Stavrovskaya, A. A.
Alterations in melanin synthesis in human melanoma cells
selected in vitro for multidrug resistance. Exp. Toxicol. Pathol.
1995, 47, 157-166.
(41) Twentyman, P. R.; Fox, N. E.; Wright, K. A.; Bleechen, N. M.
Derivation and preliminary characterization of adriamycin
resistant lines of human lung cancer cells. Br. J . Cancer 1986,
53, 529-537.
(42) Chaudhary, P. M.; Roninson, I. B. Expression and activity of
P-glycoprotein, a multidrug efflux pump, in human hematopoi-
etic stem cells. Cell 1991, 66, 85-94.
(53) Lowry, O.; Rosebrough, N.; Farr, A.; Randall, R. Protein
measurement with the Folin phenol reagent. J . Biol. Chem.
1951, 193, 265-270.
J M011010T