Download Distribution of Furamidine Analogues in Tumor Cells: Targeting of

[CANCER RESEARCH 62, 7219 –7229, December 15, 2002]
Distribution of Furamidine Analogues in Tumor Cells: Targeting of the Nucleus or
Mitochondria Depending on the Amidine Substitution1
Amélie Lansiaux, Farial Tanious, Zohar Mishal, Laurent Dassonneville, Arvind Kumar, Chad E. Stephens, Qiyue Hu,
W. David Wilson, David W. Boykin, and Christian Bailly2
Institut National de la Santé et de la Recherche Médicale U-524 et Laboratoire de Pharmacologie Antitumorale du Centre Oscar Lambret, 59045 Lille, France [A. L., L. D.,
C. B.]; Centre National de la Recherche Scientifique, 94800 Villejuif, France [Z. M.]; and Department of Chemistry and Laboratory for Chemical and Biological Sciences,
Georgia State University, Atlanta, Georgia 30303 [F. T., A. K., C. E. S., Q. H., W. D. W., D. W. B.]
ABSTRACT
INTRODUCTION
Diphenylfuran diamidines represent an important class of DNA minor
groove binders of high therapeutic interest as antiparasitic or antitumor
agents depending on the compounds structures. To exert their cytotoxic
action, the compounds must first get into the cell and reach the nuclear
compartment where the main target, DNA, is located. The forces that
drive the drugs into cell nuclei, as well as the influence of the molecular
structures on the cell distribution, are not known. To address these issues,
we took advantage of the fluorescence of the molecules to analyze their
intracellular distribution profiles in tumor cells of different origins (B16
melanoma, MCF7 mammary adenocarcinoma, A549 lung carcinoma,
HT29 colon carcinoma, LNCaP, and PC3 prostatic carcinoma) by epifluorescence and confocal microscopy. A homogeneous series of synthetic
bis-substituted alkyl or phenyl amidine and reverse amidine derivatives of
furamidine was used to dissect the molecular mechanisms that control the
distribution of the drugs into the cytoplasm or the nucleus of the cells. The
amidine (DB75) and the various N-alkyl derivatives were found to accumulate selectively in the cell nuclei. This is also the case for a guanidine
derivative but not for the phenyl-substituted compound DB569, which
essentially localizes in cytoplasmic granules. Similar cytoplasmic patterns
were observed with a reverse amidine analogue and a pyridine-substituted
compound indicating that the presence of aromatic rings on the terminal
side chain is the limiting factor that restricts the uptake of the compounds
in the nuclear compartment. The use of different organelle-selective fluorescent probes, such as JC-1 and chloromethyl-X-rosamine, both specific
to mitochondria and neutral red considered as a lysosome-selective probe,
suggests that DB569 preferentially accumulates in mitochondria. Competition experiments with the antitumor drug daunomycin reveal that the
diphenylfurans are attracted into the nuclei by the DNA. The DNA minor
groove-drug interactions provide the driving force that permits massive
accumulation of the fluorescent molecules in the nuclei. The DNA binding
properties of the diphenylfuran derivatives were investigated by DNase I
footprinting and surface plasmon resonance biosensor experiments to
measure sequence selectivity and binding affinities, respectively. Furamidine and its phenyl-substituted analogue that accumulate in the cell nuclei
and mitochondria, respectively, share a common selectivity for AT sites
and bind equally tightly to these sites. Therefore, it is possible to modulate
the intracellular distribution of the furamidine derivatives without affecting their DNA binding and sequence recognition properties. The introduction of aromatic substituents on diphenylfuran diamidines represents
a novel strategy to control the intracellular compartmentalization of these
DNA binding agents and directs them to mitochondria. This drug design
strategy may prove useful to trigger drug-induced apoptosis.
Furamidine (DB75; Fig. 1) is a DNA minor groove binder endowed
with a pronounced selectivity for sequences containing consecutive
AT bp (1–3). In therapeutic terms, DB75 is the lead compound of an
important class of antimicrobial and antiparasitic agents. This diphenylfuran derivative has shown potent activities against several pathogen microorganisms such as Cryptosporidium parvum (4), Pneumocystis carinii (5), and Trypanosoma sp (6, 7). An amidoxime prodrug
of furamidine (8) is currently undergoing Phase II clinical trials
against human African trypanosomes. The mechanism by which the
drug enters the parasites is poorly known. For pentamidine (Fig. 1), a
diamidine largely used for the treatment of African trypanosomiasis,
specific transporters have been identified recently, such as the
adenosine-sensitive pentamidine transporter-1 and the low-affinity
pentamidine transporter-1 (9). These transporters can be inhibited by
other diamidines including propamidine, berenil, and stilbamidine
(Fig. 1), and, therefore, it is reasonable to think that furamidine can
use the same system to accumulate in the parasite. At present, no such
diamidine transporter has been identified in human cells.
In addition to their antiparasitic properties, certain diphenylfuran
dications display a useful spectrum of antitumor activities. For example, the imidazoline analogue furimidazoline (DB60), has shown
significant antiproliferative activities against various tumor cell lines,
including cells resistant to cisplatinum (10). But here again, the
mechanism by which these diamidine drugs enter the cells is poorly
understood. These charged compounds enter cells rapidly, and a pure
diffusion process seems unlikely (11). Normal and malignant cells
possess a high affinity transport system that controls the entry of
polyamines (putrescine, spermidine, and spermine) and guanidines
[e.g., methylglyoxal bis(guanylhydrazone); Ref. 12]. This membrane
transport system may well be exploited by furamidine and its analogues to cross the plasma membrane.
We have described recently the cell distribution profiles of a series
of mono, bis, and tetracationic analogues of furamidine (DB75).
Unexpectedly, we found that the presence of two or four positive
charges favored nuclear uptake, whereas the loss of one of the two
cationic side chains of DB75-type compounds, with a remaining
mono-amidine or -imidazoline group, was detrimental to the nuclear
binding presumably as a result of a reduced affinity for DNA (11).
This key finding prompted us to screen additional furamidine analogues for nuclear versus cytoplasmic staining. Here we present our
data with a series of synthetic bis-substituted alkyl or phenyl amidine
and reverse amidine derivatives of DB75. For all of the compounds
shown in Fig. 1, the distribution in malignant cells was evaluated by
epifluorescence microscopy taking advantage of the intrinsic blueemitting fluorescence of the compounds. High-resolution images were
also collected by confocal microscopy, and the relation to DNA
binding and cytotoxicity is discussed. Competition experiments with
the intercalating drug daunomycin suggest that DNA binding provides
the driving force that attract the compounds in the nucleus.
Received 6/12/02; accepted 10/17/02.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
1
Supported by research grants from the Ligue Nationale Contre le Cancer (Equipe
labellisée LA LIGUE; to C. B.), from NIH Grant GM61587 (to W. D. W. and D. W. B.),
the Georgia Research Alliance, and a Gates Foundation Grant.
2
To whom requests for reprints should be addressed, at Institut National de la Santé
et de la Recherche Médicale U-524, IRCL, Place de Verdun, 59045 Lille, France. Phone:
33-320-16-92-18; Fax: 33-320-16-92-29; E-mail: [email protected].
7219
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
at room temperature to yield a yellow solid (0.40 g, 90%). The bis-imidoyl
chloride was used directly in the next step without additional characterization.
A suspension of the bis-imidoyl chloride (0.40 g, 0.00081mol) in CH2Cl2
(50 ml, dry) was saturated with dry ammonia gas while cooling in an ice bath.
The mixture was stirred at room temperature for 24 h, and the solvent was
removed under reduced pressure. The residue was treated with ice-water, and
the pH of the suspension was adjusted to a value of 10 by adding 30% NaOH.
The resulting solid was filtered, washed with H2O, and dried under vacuum to
yield the free base (0.34 g, 91%). The product was used directly in the next step
without additional characterization. The free base was converted to salt by
taking up 0.34 g (0.00075 mol) in dry ethanol (40 ml) saturated with HCl and
stirring for 2 h. The volume of solvent was decreased to ⬃10 ml under reduced
pressure. Dry ether was added to the solution to cause precipitation of the salt,
which was filtered and recrystallized from Ethanol (dry)-ether(dry) to yield
yellow solid (0.28 g, 72%), mp 251–253°C. 1H NMR (DMSO-d6) was 11.88
(s, 2H), 10.01 (s, 2H), 8.99 (s, 2H), 8.12(m, 10H), and 7.53(m, 10H). 13C NMR
(DMSO-d6) was 169.2, 162.2, 152.5, 134.8, 134.2, 129.0, 128.2, 126.9, 125.5,
123.7, and 111.7. Anal. Calcd for C30H24N4O,2HCl,0.5H2O: C, 66.69; H, 5.37;
N, 10.37. Found: C, 66.36; H, 5.19; N, 10.22.
Cell Cultures and Survival Assay. The human prostatic carcinoma
LNCaP and PC3 cells were kindly provided by Dr. Nicole Pommery (Faculty
of Pharmacy, Lille, France). These cells were maintained in RPMI 1640
containing 10% fetal bovine serum, 26 mM NaH2CO3 (pH 7.4), 1% L-glutamine, and antibiotics (1% penicillin-streptomycin) in a 37°C incubator supplied with 5% CO2 until they reached a density corresponding to ⬃106 cells
onto 75 cm2 dishes. The colon adenocarcinoma HT29 cell line was a generous
gift of Dr. Daniele Demarquay (Institut H. Beaufour, Les Ulis, France). B16
melanoma cells were obtained from Dr. Marie-Claire De Pauw (University of
Fig. 1. Chemical structures of a few diamidines and the diphenylfuran derivatives used
Liège, Belgium). HT29, B16, MCF7, and A549 cells were maintained as
in this study. Hydroxystilbamidine (OHSA) bears an OH group on the phenyl ring (ⴱ).
monolayers in 150-cm2 culture flasks using culture medium consisting of
DMEM-glutaMAX medium supplemented with 10% fetal bovine serum, penicillin (100 IU/ml), and streptomycin (100 ␮g/ml). Cells were grown in a
humidified atmosphere at 37°C under 5% CO2. In the exponential phase, the
MATERIALS AND METHODS
doubling time of these melanocytes ranges from 15 to 18 h, and the confluence
5
2
Drugs and Chemicals. Daunomycin was purchased from Sigma. DiOC6,3 stage was achieved at a cell density of 3 ⫻ 10 cells/cm . B16 cells were
JC-1, NBD-ceramide, neutral red, and PI were from Molecular Probes. The harvested by trypsinization and plated 20 h before treatment with the test drug.
The cytotoxicity of the drugs was assessed using a cell proliferation assay
synthesis of 9 of the 10 diphenylfuran derivatives used in this study has been
described previously: DB75 (6); DB181, DB226, DB244, DB249, DB417 developed by Promega (CellTiter 96 AQueous one solution cell proliferation
4
(13); DB613, DB667, and DB673 (14). The procedure for the synthesis of the assay). Briefly, 2 ⫻ 10 exponentially growing cells were seeded in 96-well
microculture plates with various drug concentrations in a volume of 100 ␮l.
new molecule DB569 is described below.
Chemistry. A suspension of 2,5-bis(4-carboxyphenyl)furan (13; 0.5 g, After 72 h incubation at 37°C, 20 ␮l of 3-(4,5-dimethylthiazol-2-yl)-5-(30.0016 mol), dry benzene (30 ml), thionyl chloride (2.0 ml, 0.027 mol), and carboxymethoxy-phenyl)-2-(4-sulfonyl)-2H-tetrazolium (15) were added to
anhydrous dimethylformamide (2 drops) was refluxed for 4 h. The solvent was each well, and the samples were incubated for an additional 3 h at 37°C. Plates
removed under vacuum, and the residue was triturated with benzene; the were analyzed on a Labsystems Multiskan MS reader at 492 nm.
Fluorescence and Confocal Microscopy. The cells (20,000 cells/cm2)
benzene was removed under vacuum and the residue triturated with dry ether.
The orange solid was filtered and dried under vacuum to yield diacid chloride were incubated at 37°C with the test compound, usually at 2 ␮M for 18 h unless
otherwise stated. The medium was removed, and cells were rinsed with
(0.5 g, 89%), mp ⬎300°C. 1H NMR (CDCl3): 8.17 (d, 4H, J ⫽ 7.6), 7.86 (d,
ice-cold PBS (10 min) before the fixation with a 2% paraformaldehyde
4H, J ⫽ 7.6), and 7.02 (s, 2H).
solution for 20 min at ⫹4°C. After washing, the cells were incubated with
Aniline (0.37 g, 0.004 mol) was added to the diacid chloride from above
fluorescent probes DiOC6 (20 nM) for 5 min at 37°C in the dark, washed again
(0.35 g, 0.001 mol) in 50 ml of dry CH2Cl2, and the mixture was allowed to
stir at room temperature for 2.5 h. The solvent was removed under reduced with PBS, and then incubated with a solution of PI (0.2 ␮g/ml) for 5 min at
pressure, and water was added. The solid was filtered and washed with 10% room temperature. A drop of antifade solution was added, and the treated
HCl, 10% NaHCO3, and water. After drying, the solid was recrystallized from portion of the slide was covered with a glass coverslip. The fluorescence of the
DMF-H2O to give a yellow solid (0.37 g, 80%), mp 355–356°C. 1H NMR drug was detected and localized by fluorescence microscopy using a Zeiss
(DMSO-d6) was 10.14 (s, 2H), 8.07 (d, 4H, J ⫽ 8.0), 7.98 (d, 4H, J ⫽ 8.0), microscope with a ⫻63 or ⫻100 oil objective. Images were captured using the
7.78 (d, 4H, J ⫽ 7.5), 7.36 (t, 4H, J ⫽ 7.5), 7.26 (s, 2H), and 7.11(t, 2H, software Quips Smart Capture (Vysis). Alternatively, the fluorescence was
J ⫽ 7.0). 13C NMR (DMSO-d6) was 160.0, 147.9, 134.2, 128.9, 127.6, 123.6, detected by confocal microscopy using a Leica DMIRBE microscope con123.5, 118.8, 118.5, 115.7, and 105.3. Anal. Calcd for C30H22N2O3: C, 78.61; trolled by a Leica TCS-NT workstation (Leica Microsystems, Bensheim,
Germany) with a 63 ⫻ 1.32 NA oil objective, equipped with a 75 mW
H, 4.80; N, 6.11. Found: C, 78.21; H, 4.60; N, 6.35.
A mixture of 2,5-bis[4-(N-phenylcarbamoyl)phenyl]furan, from above (0.41 argon-krypton and Coherent Innova-90-UV laser lines. The emission signal
was observed through a dichroic mirror (DD488/568) followed by a filter set
g, 0.00089mol), thionyl chloride (0.27 ml, 0.0036 mol), DMF (3 drops), and
CH2Cl2 (50 ml, dry) were allowed to reflux overnight. The solution was (RSP 580, BF 530/30, BP 600/30, and BP460/30 for UV excited probes). The
distilled under reduced pressure to remove solvent. The residue was triturated optical sections were obtained in the Z axis and stored on the computer with
with dry ether. The solid was filtered (under nitrogen) and dried under vacuum a scanning mode. In all of the cases, the operating conditions were such that
detectable images could not be obtained for cell samples not treated with drugs.
DNA Thermal Melting. Thermal melting experiments were conducted
3
The abbreviations used are: DiOC6, 3,3-dihexyloxacarbocyanine iodide; MES,
with Cary 3 or Cary 4 spectrophotometers interfaced to microcomputers as
4-morpholinepropanesulfonic acid; JC-1, tetrachloro-tetraethylbenzimidazolcarbocyanine
described previously (16). MES10 buffer {10 mM [2-(N-morpholino) ethaneiodide; SPR, surface plasmon resonance; PI, propidium iodide; RU, response unit; OHSA,
sulfonic acid] and 1 mM EDTA (pH 6.25) with 0.1 M NaCl} was used in the
2-hydroxy-4,4⬘-diamidino stilbene.
7220
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
experiments with poly(deoxyadenylate-deoxythymidylic acid). A thermistor
fixed into a reference cuvette was used to monitor the temperature. In Tm
experiments DNA was added to buffer in 1-cm path length reduced volume
quartz cells, and the concentration was determined by measuring the absorbance at 260 nm. The experiments were generally conducted at a concentration
of 5 ⫻ 10⫺5 M DNA bp and a ratio of 0.6 compound:bp of DNA.
DNase I Footprinting. The experimental procedures for the purification
and 32P-labeling of the DNA fragments, and the detailed protocol for the
DNase I cleavage experiments have been described previously (17, 18).
SPR. Experiments were performed in HBS-EP buffer (from BIACORE)
containing 0.01 M HEPES, 0.15 M NaCl, 3 mM EDTA, and 0.005% polysorbate
20 (v/v; pH 7.4). Three different DNA oligomers (high-performance liquid
chromatography purified and desalted; Midland Certified Reagent Co.) were
used in these studies. The 5⬘-biotin labeled hairpin duplexes are d(BiotinCGAAATTTCCTCTGAAATTTCG; A3T3 DNA), d(Biotin-CATATATATCCCCATATATATG; AT DNA), and d(Biotin-CGCGCGCGTTTTCGCGCGCG; GC DNA), and the hairpin loop sequences are italicized. The oligomer
concentrations were determined optically using extinction coefficients per mol
of strand at 260 nm determined by the nearest neighbor procedure.
SPR measurements were performed with a four-channel BIAcore 2000
system and streptavidin-coated sensor chips (SA from BIACORE as described
previously; Refs. 19 –21). Briefly, the chips were prepared for use by conditioning with three to five consecutive 1-min injections of 1 M NaCl in 50 mM
NaOH followed by extensive washing with buffer. 5⬘-Biotinylated DNA
samples (25 nM) in HBS buffer were immobilized on the flow cell surface by
noncovalent capture. Three flow cells were used to immobilize DNA oligomer
samples, and the fourth cell was left blank as a control. Samples of the
compounds were prepared in filtered and degassed MES 10 buffer by serial
dilution from stock solutions. Drug samples were injected from 7-mm plastic
vials with pierceable plastic crimp caps at a flow rate of 20 ␮l/min using the
KINJECT command. For the furan-DNA complexes a solution of 10 mM
glycine (pH 2.0) was used to dissociate all of the furan from DNA to regenerate
the surface. An array of different furan concentrations was used in each
experiment, and the results were analyzed as described below. The injection of
the compound (association) was followed by injection of running buffer
(compound dissociation). To reduce the probability of nonspecific binding to
the chip surface 50 ␮l/liter of surfactant P20 was added to the MES buffer. The
amount of DNA immobilized was ⬃350 response units in each flow cell. SPR
experiments were performed at 25°C in MES 10. With the SPR technique the
change in refractive index occurring at the surface of the sensor chip is
monitored. The change in refractive index in terms of RUs is proportional to
the amount of compound bound to the DNA.
To obtain the affinity constants the data generated were fitted to different
interaction models using Kaleidagraph for nonlinear least squares optimization
of the binding parameters using the following equation:
RU ⫽ RUmax ⫻ (K1 ⫻ Cfree ⫹ 2 ⫻ K1 ⫻ K2 ⫻ Cfree2)/
(1 ⫹ K1 ⫻ Cfree ⫹ K1 ⫻ K2 ⫻ Cfree2)
where K1 and K2 are macroscopic equilibrium constants for two types of
binding sites, RU is the SPR response at the steady state level, RUmax is the
maximum SPR response for binding one molecule per binding site, and Cfree
is the concentration of the compound in solution (22). Poor fits were obtained
with a single-site model (K2 ⫽ 0), and no significant improvement in the
quality of the fits was observed on going to a three-site model. RUmax can be
predicted using the following equation:
morphology. Fig. 2 shows fluorescence images of B16 cells exposed
to DB75 and symmetric N-alkyl amidine derivatives. The cells were
treated with each drug at 2 ␮M for 18 h at 37°C, washed, fixed with
2% paraformaldehyde, and then labeled with the dyes 20 nM DiOC6
and 0.2 ␮g/ml PI, which stain principally the cytoplasm and the
nucleus in green and red, respectively. PI binds strongly to nucleic
acids, whereas DiOC6 binds preferentially to mitochondria in the
cytoplasm. The three colors, blue, green, and red, can be easily
differentiated by fluorescence microscopy. No difference was observed between the unsubstituted parent compound DB75 and the
analogues bearing an isopropyl (DB181), isopentyl (DB226), cyclopentyl (DB244), cyclohexyl (DB249), or a methyl (DB417) group. In
all of the cases, the molecules are sequestered in the cell nuclei and
colocalize with PI. The massive accumulation of the blue fluorescence
in the nuclear compartment contrasts with the extensive green fluorescence detected in the cytoplasm. The distribution of the studied
compounds in the nuclear compartment is more uniform than that of
PI, which concentrates into a few bright red nucleolar-like bodies. As
discussed recently (12), the different nuclear distribution between
DB75 and PI can be attributed to their distinct DNA sequence selectivity. Nucleolar DNA is GC-rich and provides a good substrate for
the intercalating drugs like PI but not for the minor groove binder
DB75, which exhibits a high preference for AT-rich sequences (1–3).
The guanidine derivative DB673 also concentrates into cell nuclei
(Fig. 2). This compound is relatively weakly fluorescent compared
with DB75 and, therefore, a much longer exposure time is required to
detect the compound in the cells. In this case, the images are less well
resolved because the background fluorescent is higher. All of the
studied diphenylfuran derivatives fluoresce in blue (␭excmax ⫽ 360 –
390 nm; ␭emmax ⫽ 445– 460 nm), but the relative fluorescence intensity varies considerably with the compound structure. The exposure
time were adjusted to obtain images of equal intensity (Table 1).
DB75, its guanidine analogue DB673, and the five N-alkyl derivatives accumulate selectively in the cell nuclei. The situation is
radically different with the phenyl-substituted compound DB569. In
this case, the fluorescence is essentially localized in cytoplasmic
granules, whereas nuclei appear as dark cores (Fig. 2). The complete
absence of nuclear blue fluorescence in all of the B16 cells treated
with DB569 was a particularly striking observation, which prompted
us to test additional analogues. We then tested the reverse amidine
analogue DB613. The terminal phenyl rings are now connected to the
diphenylfuran core via a C(NH)NH connector compared with
the NHC(NH) spacer of DB569. The reversal of the orientation of the
amidine group does not influence the cell distribution profile; DB613
does not enter the nuclei, just as observed with DB569 (Fig. 2).
Similar cytoplasmic patterns were also observed with a compound
Table 1 Cytotoxicity and DNA binding
RFI
DB75
DB181
DB226
DB244
DB249
DB417
DB569
DB613
DB667
DB673
RUmax ⫽ (RUDNA/MWDNA) ⫻ MWcompound ⫻ RII
Where RUDNA is the amount of DNA immobilized in RUs, MW is molecular
weight of compound and DNA, respectively, and RII is the refractive index
increment ratio of compound to refractive index of DNA (22, 23).
RESULTS
Intracellular Distribution of the Diphenylfuran Derivatives.
B16 melanoma cells were used in the initial investigation of the
nuclear staining by the test compounds because of their flattened
a
77
111
93
182
82
65
14
17
15
15
ET (s)b
⌬Tm (°C)c
IC50 (␮M)d
0.5
4.7
9.0
1.5
1.5
8.6
4.5
13.9
20
7.0
26.2
28.1
20.4
⬎29
28
27.6
26.9
23.9
19.6
21.6
9.2 ⫾ 0.3
61.8 ⫾ 7
4.3 ⫾ 0.2
⬎100
8.5 ⫾ 0.2
⬎100
4.6 ⫾ 1.2
1.8 ⫾ 0.8
3.6 ⫾ 1.3
11.2 ⫾ 0.9
a
Relative fluorescence intensity (RFI) determined with a solution of the compound in
a BPE buffer (pH 7.0) with excitation and emission wavelengths of 360 nm and 460 nm,
respectively, or 390 nm and 480 nm for DB569 and DB613.
b
Exposure time required to obtain digital images of equal fluorescence intensity.
c
Variation in melting temperature, ⌬Tm ⫽ Tmcomplex ⫺ Tmpoly(dA)䡠(dT). Tm measurements were performed in MES10 buffer.
d
Drug concentration that inhibits B16 cell growth by 50% after 72-h incubation.
7221
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
Fig. 2. Fluorescence micrographs of B16 melanoma cells stained with the indicated compound
(2 ␮M each in blue), with PI (in red), or with
DiOC6 (in green). Images on the right side of the
figure show the overlay of the DBN compound
with PI (blue ⫹ red) or with DiOC6 (blue ⫹
green). The cells were incubated with the test drug
for 18 h, washed, fixed with 2% paraformaldehyde,
and then labeled with PI (0.2 ␮g/ml PI) or DiOC6
(20 nM) before the microscopy observation (⫻63).
Results are representative of three independent
stainings.
designated DB667 containing pyridine terminal groups in place of the
end phenyls of DB613 (data not shown). The presence of aromatic
rings on the terminal side chain appears as the limiting factor that
restricts the uptake of the compounds in the nuclear compartment.
The phenyl-amidine compound DB569 is only weakly fluorescent
in solution, just like the guanidine analogue DB673, which was found
in the nucleus. But in cells, the fluorescence detection of DB569 can
be easily obtained by adjusting the exposure time to about 4 –5 s,
exactly as with the bis-N-isopropyl derivative DB181, which is highly
fluorescent in solution (Table 1). A direct quantitative analysis of
compound uptake is impossible because of their different intrinsic
fluorescence properties; however, the facile identification of DB569
in cells suggests that this compound massively concentrates in the
cytoplasm. But the possibility that this weakly fluorescent compound
is transformed into a more fluorescent metabolite retained in the
cytoplasm must be kept in mind.
The distinct distribution profiles observed with DB75 and DB569
likely reflect the intrinsic properties of the compounds, but we also
considered the possibility that the differences were because of the
biochemical nature of the murine B16 melanoma cells. To answer this
question, we repeated the experiments using five other cancer cell
lines from human origin. Fig. 3 shows typical images obtained with
HT29 colon carcinoma cells, MCF7 mammary adenocarcinoma cells,
A549 lung carcinoma cells, and two prostate tumor cell lines, PC3 and
LNCaP. These latter human malignant cells express androgen receptors and are hormonally responsive but not metastatic. In contrast, the
metastatic PC3 cells are androgen-insensitive. In these cells, the loss
of androgen receptors is generally associated with the progression of
prostate cancer to the more invasive carcinoma. These human cell
lines gave identical results to the murine cell line. In each case, DB75
was found almost exclusively in the nuclei of the human cells,
whereas DB569 was clearly maintained outside the nuclei, in the
7222
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
Fig. 3. Intracellular distribution of furamidine
(DB75) and its N-phenyl derivative DB569 (2 ␮M
each) in murine B16 melanoma cells, human MCF7
mammary adenocarcinoma cells, human A549 lung
carcinoma cells, and human prostatic tumor cells PC3
and LNCaP, visualized by fluorescence microscopy
(⫻63) after 18 h incubation. The green and red images correspond to the subsequent incubation of the
same cells with DiOC6 (20 nM) and PI (0.2 ␮g/ml PI),
and the images on the right side correspond to the
superimposed blue ⫾ green ⫾ red fluorescence.
cytoplasm. Similar results were also obtained with leukemia cells
(P388 and K562; data not shown). Therefore, the different distribution
profiles observed with DB75 versus DB569 must reflect the chemical
properties of the drugs, not the specific capacities of the cells.
The optical sectioning capabilities of confocal microscopy was
particularly useful to precisely locate the drug in cells. Fig. 4 is a
single confocal slice showing a cross-section of HT29 cells treated
with the diphenylfuran DB75 or DB569, and counterstained with
DiOC6 and the nuclear specific dye TOTO-3. This dye is used to stain
DNA with far-red fluorescence under red (647 nm) excitation (24).
Obviously, DB75 is not evenly distributed over the cell. The blue
fluorescence concentrates into the nucleus with no detectable fluorescence in the cytoplasm. The direct observation is reinforced by the
fluorescence intensity profiles in Fig. 4, which show the distribution
of DB75, PI, and DiOC6 fluorescence along the XY axis indicated on
the photographs. The blue fluorescence colocalizes with the red fluorescence. The reverse situation was observed with DB569. In this
case, the blue fluorescence essentially follows the green fluorescence.
The use of such a high-resolution imaging system attests unambiguously that the two compounds locate in different cell compartments:
DB75 in the nuclei and DB569 in the cytoplasm.
The confocal images also help to identify the cytoplasmic structures
that stain with DB569. The localization of the fluorescence blue spots
coincides with that of the green probe. The two images obtained with
DB569 and DiOC6 are essentially superimposable and, therefore,
support the idea that the DB569 molecules accumulate in the mitochondria organelles. Additional experiments performed with the mitochondria-specific probe JC-1 substantiate this observation. This
lipophilic cation, commonly used to measure the variation of the
mitochondrial membrane potential (⌬⌿mt; Refs. 25, 26), gives red-
7223
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
Fig. 4. Confocal laser scanning microscopy of
HT29 colon carcinoma cells treated with 2 ␮M
DB75 or DB569. The red and green images correspond to the subsequent incubation of the same
cells with TOTO-3 (0.5 ␮M) and DiOC6 (20 nM),
and the images on the right side correspond to the
superimposed blue ⫹ green and blue ⫹ red fluorescence. The scales (20 ␮m) are indicated. The
lower images correspond to the superimposed tricolor images (blue ⫹ green ⫹ red). The graphs
shows the distribution of the diphenylfuran (blue),
DiOC6 (green), and PI (red) fluorescence along the
XY axis indicated in the photographs for both
DB75 and DB569.
fluorescent J-aggregates locally and spontaneously under high mitochondrial ⌬⌿mt, whereas the monomeric form fluoresces green (27).
Typical images of B16 cells treated with DB569 and counterstained
with JC-1 are presented in Fig. 5B. The blue (DB569) and green
(JC-1) images superimpose relatively well. All of the blue granules
are also colored in green suggesting that DB569 and JC-1 essentially
locate in the same organelles. In contrast, the overlay is not so good
when the cells were counterstained with neutral red (Fig. 5A), which
is considered as a lysosome-specific probe (28, 29). Close inspection
of the micrographs reveals that the blue spots corresponding to
DB569-containing structures contain little red dye. Lysosomes may
contain a limited amount of DB569 but a priori the majority of the
drug molecules are internalized in the mitochondria. A good correlation was also observed when the cells were colored with DB569 and
the red dye chloromethyl-X-rosamine (Mitotracker red), another
mitochondria-specific dye. In contrast, no correlation was observed
when using the green dye NDB-ceramide considered as a suitable dye
for the Golgi apparatus (data not shown). All of the results together
strongly suggest that DB569 essentially accumulates in mitochondria,
but the molecular basis of this interaction remains to be determined.
The mechanism by which DB569 traverses the mitochondrial membrane and is sequestered inside is unknown at present. The internal pH
of mitochondria probably plays a role in this sequestration (see
“Discussion”). The three compounds that select the mitochondria,
DB569, DB667, and DB613, are likely to have pKs three to four units
below that of DB75.
The reasons why DB75, but not DB569, is attracted in the nuclei
are not known at present, but we hypothesized that the molecules are
attracted by the DNA targets in the nuclear compartment. To examine
this aspect, competition experiments were performed with antitumor
drug daunomycin, which is a well-known DNA intercalating agent
and has the additional advantage to fluoresce in red. The specific
accumulation of daunomycin and related anthracyclines in cell nuclei
has been reported in many studies (30 –32). Using the confocal
equipment, there is no overlap between the blue fluorescence of DB75
and the red fluorescence of daunomycin. Fig. 5C shows representative
confocal micrographs of B16 cells stained in blue with DB75 and/or
in red with daunomycin. DB75 alone only stains the nuclei of the
murine cells (Fig. 5C, panel 1) whereas in the presence of daunomycin, the majority of DB75 molecules is detected in the cytoplasm (Fig.
5C, panel 2). A short treatment with daunomycin suffices to expel the
blue compound from the nuclei. Fig. 5C, panel 4 shows B16 cells that
have been incubated for 17.5 h with 8 ␮M DB75 and then for 30 min
with 8 ␮M daunomycin. In this case, the intercalating drug stains the
nuclei to a lower extent, but the short incubation period is sufficient
to limit the entrance of the diphenylfuran compound in the nuclei.
Now the DB75 molecules are distributed equally in the nucleus and
the cytoplasm of the tumor cells. Similar results were obtained with
the human HT29 cell line (data not shown). Daunomycin was found
exclusively in the nuclei, but surprisingly DB75 was confined to the
cytoplasm in the presence of the anthracycline competitor. The high
affinity of the intercalating drug for DNA must restrict the access of
the minor groove binder. These competition experiments strongly
support the hypothesis that DNA is responsible for the staining of the
nuclei by DB75.
DNA Binding. The DNA binding properties of the diphenylfuran
derivatives was investigated by a combination of spectroscopic, biosensor, and biochemical methods. Melting temperature (Tm) experi-
7224
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
Fig. 5. Fluorescence micrographs of B16 melanoma cells incubated for 18 h with 2 ␮M DB569
and counterstained (A) with neutral red (50 nM,
15mn, 37°C) or (B) with the green fluorescent dye
JC-1 (10 ␮M, 15mn, 37°C). C, confocal laser scanning microscopy of B16 melanoma cells labeled
with DB75 in the absence and presence of daunomycin. Panels 1 and 3 refer to B16 murine cells
incubated with 8 ␮M DB75 and 8 ␮M daunomycin,
respectively. Panel 2 refers to B16 cells treated
simultaneously with 8 ␮M DB75 and 8 ␮M daunomycin for 18 h. In panel 4, the B16 cells were
incubated with 8 ␮M DB75 for 17.5 h and then with
8 ␮M daunomycin for 30 min before the observation. The scales (20 ␮m) are indicated. The cells
shown are representative of the entire population.
ments were performed to evaluate the relative affinities of the compounds for poly(deoxyadenylate-deoxythymidylic acid). The ⌬Tm
values (⌬Tm ⫽ Tmcomplex ⫺ TmDNA) measured with each compound,
in MES10 buffer at a ratio of 0.6 compound:bp of DNA, are collected
in Table 1. All of the compounds bind strongly to DNA with ⌬Tm
values of ⬃20°C or more. The diamidine DB75 and its phenylamidine analogue stabilize the double helix to the same extent, as does
the reverse amidine derivative DB613.
A DNase I footprinting study was performed with the two most
representative compounds of the present series, the lead compound
DB75 and the phenyl-substituted analogue DB569, which exhibit
distinct cell distribution properties, with a marked nuclear accumulation for DB75 as opposed to a cytoplasmic staining with DB569.
Three DNA restriction fragments of 117, 160, and 265 bp, each
uniquely 3⬘-end radiolabeled, were prepared and subjected to limited
cleavage by the nuclease in the presence of different concentrations of
DB75 and DB569. Well-resolved footprints were detected with both
compounds. The substitution of the amidine ends with a phenyl group
does not perturb the capacity of the drug to recognize specific sequences. The footprints are located at the same positions and are
equally intense with both compounds (Fig. 6). The binding curves
obtained with DB75 and DB569 superimpose very well, and the
sequences protected from DNase I cleavage all correspond to AT-rich
sequences, whereas GC-rich sequences are often cut more readily by
the enzyme in the presence of the drugs. There is no doubt that the two
compounds share a common selectivity for AT sites and bind equally
well to these sites.
A quantitative analysis of the drug-DNA interaction was performed
by SPR biosensor experiments. The compounds were passed through
three sensor chip flow cells containing DNAs with an A3T3 minor
groove-binding site, an alternating AT DNA sequence, or an alternating GC DNA sequence. The compounds were also passed through a
control flow cell with no DNA, and the response in this flow cell was
subtracted to generate the response sensorgrams. An example set of
sensorgrams (RUs versus time) for DB417 binding to the A3T3
oligomer hairpin are shown in Fig. 7A. With these DNAs and compounds the sensorgrams can be divided into three phases: (a) association of the furan with DNA that results in an increase of RU; (b) a
steady state region during the solution injection where the rate of
compound binding is balanced by dissociation from the complex and
there is no change in RU with time; and (c) dissociation of the furan
from surface during buffer flow with a decrease in bound compound
and observed RU. As can be seen from Fig. 7, the interaction with
DB417 is slow at low concentrations and increases in association rate
with an increase in the furan concentrations as expected for a bimolecular interaction. Also as expected for a first order reaction, the
dissociation reactions are concentration independent. The sensorgrams have a slow dissociation rate up to ⬃30 RU, but exhibit a
second, faster dissociation for the sensorgram region above ⬃30 RU.
The sensorgrams below 30 RU are at concentrations below 1 ␮M and
are characteristic of specific interactions. The region above 30 RU is
only reached at higher concentrations (above 1 ␮M concentration) and
represents a weak secondary binding mode for these compounds.
It was possible to reach a steady-state plateau for all of the furan
derivatives in Fig. 1 in ⬃200 s at the lowest concentrations used in
these experiments. Average fitting of the sensorgrams at the steady
state level was performed over a 30-s time period with the BIAevaluation 3.1 software. The binding constants for all of the compounds to
the three different DNA sequences were determined by fitting RU at
the steady state as a function of concentration, by using the equations
in “Materials and Methods,” and results for DB417 are shown in Fig.
7B. All of the compounds required a two-site binding equation for the
best fit. As expected from the sensorgrams the compounds had a
single strong binding site in the DNAs with AT bp as well as a weaker
7225
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
Fig. 6. Differential cleavage plots comparing the susceptibility of three DNA fragments to DNase I cutting in the presence of the diphenyl-furan derivatives: (F) DB75 and (E)
DB569 (2 ␮M each). Negative values correspond to a ligand-protected site, and positive values represent enhanced cleavage. Vertical scales are in units of ln(fa) ⫺ ln(fc), where fa
is the fractional cleavage at any bond in the presence of the drug and fc is the fractional cleavage of the same bond in the control, given closely similar extents of overall digestion.
Each line drawn represents a 3-bond running average of individual data points, calculated by averaging the value of ln(fa) ⫺ ln(fc) at any bond with those of its two nearest neighbors.
secondary binding site. The primary binding constants are collected in
Table 2. The secondary binding exhibits fast dissociation from DNA
and has binding constants that are 10 –100 times lower than the strong
binding constants. The GC sequence exhibited much weaker binding
and lower RUs at the same concentration than the AT containing
sequences (Fig. 7; Table 2). This is the expected result for AT-specific
minor groove complex formation. The GC binding constants are only
slightly greater than the nonspecific binding observed for the AT
DNA duplexes. However, even with the AT DNAs, the compounds
show significant structure dependent variations in K values (Table 2).
The rates of binding of the furans are slow enough at low concentration to allow determination of association and dissociation kinetics
constants. Curves in the strong binding region were fit by a global
analysis procedure and equilibrium constants determine by the ratio of
association/dissociation constants. In all of the cases the equilibrium
constants determined from the rate constants were in good agreement
with those determined by the steady state procedure. The rates for
binding and dissociation with the GC DNA sequence are too fast to
determine by the SPR-biosensor method.
Cytotoxicity. The cytotoxic potential of the diphenylfuran derivatives was evaluated using a tetrazolium-based survival assay. The
drug concentrations required to inhibit cell growth by 50% after
incubation in the culture medium for 72 h are indicated in Table 1.
DB75 is weakly cytotoxic to the B16 murine melanoma cells, and the
substitution of the amidines with a methyl, an isopropyl, and a
cyclopentyl group abolishes the cytotoxic potential. The introduction
of cyclohexyl groups or the replacement of the amidine with guanidine groups has no effect on cytotoxicity. The substitution of the
amidine with isopentyl or a phenyl group only decreases the IC50 by
a factor of 2. The most cytotoxic compound in the series is the
phenyl-reverse amidine derivative DB613, but nevertheless these diphenylfuran derivatives must be considered as weak cytotoxic agents,
and no direct relationship with the distribution profile can be proposed. DB226, which seems to freely enter the nuclei, is equally toxic
as DB569, which apparently remains confined in the cytoplasmic
compartment of the cells. however, it should be noted that a priori the
absence of nuclear DB569 fluorescence does not exclude the possibility that a low concentration of the drug (which is weakly fluorescent) could exist in the nuclei, with the drug molecules inserted in the
minor groove of nuclear DNA. It is important to mention that on
binding to the minor groove of DNA, the fluorescence of these
bis(amidinophenyl)furan derivatives decreases by about 20 –25%, and
this effect contributes to underestimate the extent of nuclear uptake.
The bright blue fluorescence detected in the nuclei of the cells treated
with DB75 likely reflects a massive nuclear accumulation.
No correlation between DNA binding strength and cytotoxicity
could be established. The graph in Fig. 8 shows the distribution of
drug equilibrium binding constants for the A3T3 sequence versus IC50
values. The four compounds with K values ⬎108 M⫺1 are poorly
cytotoxic, whereas those with K values ⬍5 ⫻ 107 M⫺1 are more
potent at reducing the growth of the melanoma cells.
DISCUSSION
This study provides three important findings. First, it confirms that
furamidine (DB75) selectively accumulates in the nuclei of tumor
7226
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
cells. This conclusion was already drawn from our recent fluorescence
study using the murine B16 cell line (11). Here the use of different
human cell lines validates this finding, and the specific sequestration
of DB75 in cell nuclei was monitored by confocal microscopy to
produce high resolution images. Second, the results indicate that the
extent of nuclear uptake by the diphenylfuran derivatives is directly
dependent on the compound structure. Substitution of the amidine
groups with an alkyl chain does not markedly change the distribution
profiles of the drugs in malignant cells. All of the N-alkyl bis(amidi-
Fig. 8. Equilibrium DNA binding constants (K values for the A3T3 sequence) versus
cytotoxicity values (IC50) are plotted for all 10 compounds. The compounds can be
divided in two groups, (f) with high and (F) low K values.
Fig. 7. SPR binding data. A, SPR sensorgrams for the interaction of DB417 with the
hairpin DNA oligonucleotide containing the A3T3 sequence. The concentrations are from
1 nM (lowest) to 5 ␮M (highest). Data were collected at 25°C in MES 10 buffer with a flow
rate of 20 ␮l/min. B, binding plots (SPR response versus drug concentration) used to
determine the affinity constants for DB417 complexed with the three different hairpin
duplexes [A3T3], [AT]4, and [CG]4. Above 0.5 ␮M the RU values remain practically
unchanged. RU values were obtained from the steady-state region of the SPR sensorgrams.
Table 2 DNA binding constants determined by SPR
Experiments were conducted at 25°C in MES10 buffer. All K values are for the
strongest binding site and are in M⫺1 units. The K values are all times 106 and can be
directly compared.
DNA
Drug
A 3 T3
AT
CG
DB75
DB181
DB226
DB244
DB249
DB417
DB569
DB613
DB667
DB673
110
170
28
190
31
98
39
8.1
7.5
35
34
140
12
110
16
75
19
2.7
1.9
18
0.8
0.5
0.4
0.9
0.5
1.1
0.7
0.6
0.2
9.3
nophenyl)furan derivatives tested here mainly distribute in the nuclear
compartment where their DNA substrate is located. In contrast, the
three derivatives substituted with aromatic rings, phenyl, or pyridine,
exhibit distinct distribution profiles with a specific accumulation in
cytoplasmic vesicles, most likely the mitochondria. This is important
information for future drug design in this series. The incorporation of
aromatic rings on the furamidine side chain may provide a suitable
delivery strategy to target drugs specifically to mitochondria rather
than in the nuclei. Because of the apparently specific non-nuclear
uptake, a compound like DB569 has the potential to be used for
targeting mitochondrial DNA. Third, the competition studies with
daunomycin suggest that the diphenylfurans are attracted into the
nuclei by the DNA. The minor groove binders apparently have no
intrinsic properties that drive them into the nuclear compartment, and
the energetics for transport against the concentration gradient are
driven by binding to the nucleic acid targets. DNA is clearly a magnet
for these compounds and the DNA minor groove-drug interactions
provide the driving force that permits massive accumulation of the
fluorescent molecules in the nuclei. In the absence of the competing
intercalating drug, DB75 is exclusively sequestered in the nuclear
compartment, whereas incubation of the cells with daunomycin profoundly altered the drug compartmentalization by releasing DB75
from the nuclei and shifting it to the cytoplasm.
On entering the cells, furamidine and analogs are attracted by the
nucleus where the main known target, DNA, is located. Apparently,
these highly polar biscationic compounds have little or no interaction
with cell membranes; no staining of the plasma membrane or nuclear
envelope was detected. At first sight, DNA attracts the compounds,
thus forming a reservoir of molecules in the nuclear compartment.
Cationic furamidine-type molecules can be bound to the nucleic acids
in the cell nuclei but they could also be attracted by the relatively high
negative electric potential across the mitochondrial membranes. For
example, certain rhodamine compounds (rhodamine 123, rhodamine
6G, and rhodamine 3B) that are positively charged at physiological
pH are able to stain mitochondria specifically, whereas uncharged
fluorescein compounds do not (33). Moreover, the fact that mitochondria contain a significant amount of nucleic acids could also play a
role in directing the drug molecules to this compartment. Therefore,
one can adopt the view that there may exist a competition or an
exchange between nucleus and mitochondrion for the uptake of furamidine-type compounds. Depending on the structure of the terminal
groups and the relative affinity for the target, the compound will be
driven to the nucleus (e.g., DB75) or to the mitochondrion (e.g.,
DB569).
The cytotoxic measurements corroborate our previous study showing that nuclear uptake is not associated with cytotoxicity (11). The
7227
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
biological activity may rather depend on the capacity of these compounds to recognize specific sequences in DNA and to interfere with
the correct expression of some critical genes. Interestingly, the three
compounds that seem to accumulate in mitochondria are the most
cytotoxic compounds in the series. Mitochondrial targeting with
phenyl-substituted diphenylfuran diamidines may be more efficient
than nuclear targeting to prevent tumor cell growth
The DNA binding data are important because they show that it is
possible to modulate the intracellular distribution of the furamidine
derivatives without affecting their DNA binding and sequence recognition properties. The footprinting data agree well with the SPR
results to indicate that the diamidine DB75 and its phenyl-substituted
analogue DB659 both selectively recognize AT-rich sequences in
DNA. Linear dichroism measurements (data not shown) indicated that
the two compounds bind to the minor groove of DNA. The introduction of aromatic substituents may thus represent a useful strategy to
control the intracellular compartmentalization of these DNA-binding
agents.
At present, it is unclear whether DB569 is blocked from the nucleus
or selectively attracted to the mitochondrion. Blocking seems likely,
because the nucleus contains more DNA binding sites. There may be
a specific mechanism (e.g., a transporter or an efflux pump) that
restricts the nuclear transport of aromatic amidines. At first sight, the
lack of nuclear uptake of a compound like DB569 would seem
detrimental to its pharmacological effect. One would like to see a
DNA-targeted drug entering freely the nucleus to reach its target. But
in fact, the lack of nuclear binding and the preferential entrance into
mitochondria may be advantageous. As noted above, mitochondria
contain a significant proportion of nucleic acids and play a pivotal role
in the propagation of the apoptotic signal by which most anticancer
drugs kill cells. Mitochondrion is currently viewed as a valid and
exploited target for cancer chemotherapy (34). Several examples can
be cited. The cationic rhodacyanine dye MKT-077 (1-ethyl-2-{[3ethyl-5-(methylbenzothiazolin-2-ylidene)-4-oxothiazolidin-2-ylidene]
methyl}pyridinium chloride) exhibits anticarcinoma activity based on
selective mitochondria accumulation (35–38). Similarly, the DNA
binding drug ditercalinium, a dicationic bisintercalator, induces mitochondrial damages (39). Several reports suggest that mitochondria
alterations in carcinomas by different antitumor drugs could play a
role in drug-induced cytotoxicity. DNA-binding drugs like ethidium
bromide, novobiocine, ellipticine, and paraquat all induce mitochondrial alterations (40 – 43). Therefore, a new strategy can be developed
to target mitochondrial DNA using diphenylfuran analogues of
DB569.
At this point it is interesting to refer also to the literature on the
cellular uptake and distribution of diamidines in tumor cells. The
amidine antibiotic hydroxystilbamidine (OHSA), which is the active
constituent of the neuronal tracer Fluoro-Gold (44), was shown to
distribute in the cytoplasm and the nuclei of tumor cells by the early
1950s (45– 47). OHSA is an AT-specific DNA minor groove binder
(48 –50), exactly like the diphenylfuran derivatives used here, and it is
supposed to enter human cells by passive diffusion. It has been
proposed that the uncharged form of weakly basic compounds is in
equilibrium across cell membranes, and that the charged form does
not cross membranes readily (51). As a result, it is expected that if a
weak base became protonated intracellularly (e.g., in an acidic compartment such as a lysosome or a mitochondria) it would accumulate
in that compartment (51, 52). The maximum extent to which a
compound can be accumulated into any compartment is equal to the
ratio of hydrogen ion concentrations inside and outside the compartment (51). Considering that the pH gradient across mitochondria is
considerably less than that across lysosome (44), the mitochondrial
accumulation of a compound like DB569 is, at first sight, disfavored,
but the high affinity of the compound for DNA may well alter the
distribution to favor its sequestration in the mitochondria. The hypothesis that nuclear DNA holds DB75 whereas mitochondrial DNA
holds DB569 is currently investigated through the use of specific cells
such as the HeLa-Rho tumor cell line containing DNA-depleted
mitochondria (53).
ACKNOWLEDGMENTS
We thank the Service Commun d’Imagerie Cellulaire de l’IMPRT (IFR114)
for access to the fluorescence microscope. W. D. W. acknowledges an Institut
National de la Santé et de la Recherche Médicale “Poste Orange” fellowship
for research in Lille at Institut National de la Santé et de la Recherche Médicale
U-524.
REFERENCES
1. Trent, J. O., Clark, G. R., Kumar, A., Wilson, W. D., Boykin, D. W., Hall, J. E.,
Tidwell, R. R. Blagburn, B. L., and Neidle, S. Targeting the minor groove of DNA:
crystal structures of two complexes between furan derivatives of berenil and the DNA
dodecamer d(CGCGAATTCGCG)2. J. Med. Chem., 39: 4554 – 4562, 1996.
2. Wilson, W. D., Tanious, F. A., Ding, D., Kumar, A., Boykin, D. W., Colson, P.,
Houssier, C., and Bailly, C. Nucleic acid interactions of unfused aromatic cations:
Evaluation of proposed minor-groove, major-groove and intercalation binding modes.
J. Am. Chem. Soc., 120: 10310 –10321, 1998.
3. Mazur, S., Tanious, F. A., Ding, D., Kumar, A., Boykin, D. W., Simpson, I. J., Neidle,
S., and Wilson, W. D. A thermodynamic and structural analysis of DNA minor
groove complex formation. J. Mol. Biol. 300: 321–337, 2000.
4. Blagburn, B. L., Drain, K. L., Land, T. M., Moore, P. H., Lindsay, D. S., Kumar, A. J.,
Shi, J., Boykin, D. W., and Tidwell, R. R. Dicationic furans inhibit development of
Cryptosporidium Parvum in HSD/ICR suckling mice. J. Parasitol., 84: 851– 856,
1998.
5. Boykin, D. W., Kumar, A., Spychala, J., Zhou, M., Lombardi, R. L., Wilson, W. D.,
Dykstra, C. C., Jones, S. K., Hall, J. E., Tidwell, R. R., Laughton, C., Nunn, C. M.,
and Neidle, S. Dicationic diarylfurans as anti-Pneumocystis carinii agents. J. Med.
Chem., 38: 912–916, 1995.
6. Das, B. P., and Boykin, D. W. Synthesis and antiprotozoal activity of 2, 5-bis(4guanylphenyl)furans. J. Med. Chem., 20: 531–536, 1977.
7. Steck, E. A., Kinnamon, K. E., Rane, D. S., and Hanson, W. L. Leishmania donovani.
Plasmodium berghi, Trypanosoma rhodesiense: Antiprotozoan effects of some amidine types. Exp. Parasitol., 52: 404 – 413, 1981.
8. Boykin, D. W., Kumar, A., Bender, B. K., Hall, J. E., and Tidwell, R. R. AntiPneumocystis activity of bis-amidoximes and bis-O-alkylamidoximes prodrugs.
Bioorg. Med. Chem. Lett., 6: 3017–3020, 1996.
9. De Koning, H. P. Uptake of pentamidine in Trypanosoma brucei brucei is mediated
by three distinct transporters: implications for cross-resistance with arsenicals. Mol.
Pharmacol., 59: 586 –592, 2001.
10. Neidle, S., Kelland, L. R., Trent, J. O., Simpson, I. J., Boykin, D. W., Kumar, A., and
Wilson, W. D. Cytotoxicity of bis(phenylamidinium)furan alkyl derivatives in human
tumour cell lines: Relation to DNA minor groove binding. Bioorg. Med. Chem., 7:
1403–1408, 1997.
11. Lansiaux, A., Dassonneville, L., Facompré, M., Kumar, A., Stephens, C. E., Bajic,
M., Tanious, F., Wilson, W. D., Boykin, D. W., and Bailly, C. Distribution of
furamidine analogues in tumor cells: influence of the number of positive charges.
J. Med. Chem., 45: 1994 –2002, 2002.
12. Ekelund, S., Nygren, P., and Larsson, R. Guanidino-containing drugs in cancer
chemotherapy: biochemical and clinical pharmacology. Biochem. Pharmacol., 61:
1183–1193, 2001.
13. Boykin, D. W., Kumar, A., Xiao, G., Wilson, W. D., Bender, B. C., McCurdy, D. R.,
Hall, J. E., and Tidwell, R. R. 2, 5-Bis[4-N-alkylamidino)phenyl]furans as antiPneumocystis carinii agents. J. Med. Chem., 41: 124 –129, 1998.
14. Stephens, C. E., Tanious, F., Kim, S., Wilson, W. D., Schell, W. A., Perfect, J. R.,
Franzblau, S. G., and Boykin, D. W. Diguanidino and “reversed” diamidino 2,
5-diarylfurans as antimicrobial agents. J. Med. Chem., 44: 1741–1748, 2001.
15. Cory, A. H., Owen, T. C., Barltrop, J. A., and Cory, J. G. Use of an aqueous soluble
tetrazolium/formazan assay for cell growth assays. Cancer Commun., 3: 207–212,
1991.
16. Wilson, W. D., Tanious, F. A., Fernandez-Saiz, M., and Rigl, C. T. Evaluation of
drug-nucleic acid interactions by thermal melting curves. Methods Mol. Biol., 90:
219 –240, 1997.
17. Bailly, C., and Waring, M. J. Comparison of different footprinting methodologies for
detecting binding sites for a small ligand on DNA. J. Biomol. Struct. Dyn., 12:
869 – 898, 1995.
18. Bailly, C. Tardy, C., Wang, L., Armitage, B., Hopkins, K., Kumar, A., Schuster,
G. B., Boykin, D. W., and Wilson, W. D. Recognition of ATGA sequences by the
unfused aromatic dication DB293 forming stacked dimers in the DNA minor groove.
Biochemistry, 40: 9770 –9779, 2001.
19. Wang, L., Bailly, C., Kumar, A., Ding, D., Bajic, M., Boykin, D. W., and Wilson,
W. D. Specific molecular recognition of mixed nucleic acid sequences: An aromatic
dication that binds in the DNA minor groove as a dimer. Proc. Natl. Acad. Sci. USA,
97: 12–16, 2000.
7228
CELLULAR DISTRIBUTION OF FURAMIDINE ANALOGUES
E20. Wang, L., Carrasco, C., Kumar, A., Stephens, C. E., Bailly, C., Boykin, D. W., and
Wilson, W. D. Evaluation of the influence of compound structure on stacked-dimer
formation in the DNA minor groove. Biochemistry, 40: 2511–2521, 2001.
21. Lacy, E. R., Le, N. M., Price, C. A., Lee, M., and Wilson, W. D. Influence of a
terminal formamido group on the sequence recognition of DNA by polyamides.
J. Am. Chem. Soc., 124: 2153–2163, 2002.
22. Davis, T. M., and Wilson, W. D. Determination of the refractive index increments of
small molecules for correction of surface plasmon resonance data. Anal. Biochem.,
284: 348 –353, 2000.
23. Davis, T. M., and Wilson, W. D. Surface plasmon resonance biosensor analysis of
RNA-small molecule interactions. Methods Enzymol., 340: 22–51, 2001.
24. Suzuki, T., Fujikura, K., Higashiyama, T., and Takata, K. DNA staining for fluorescence and laser confocal microscopy. J. Histochem. Cytochem., 45: 49 –53, 1997.
25. Salvioli, S., Ardizzoni, A., Franceschi, C., and Cossarizza, A. JC-1, but not DiOC6(3)
or rhodamine 123, is a reliable fluorescent probe to assess ⌬⌿ changes in intact cells:
implications for studies on mitochondrial functionality during apoptosis. FEBS Lett.,
411: 77– 82, 1997.
26. Kluza, J., Lansiaux, A., Wattez, N., Mahieu, C., Osheroff, N., and Bailly, C.
Apoptotic response of HL-60 human leukemia cells to the antitumor drug TAS-103.
Cancer Res., 60: 4077– 4084, 2000.
27. Smiley, S. T., Reers, M., Mottolahartshorn, C., Lin, M., Chen, A., Smith, T. W.,
Steele, G. D., and Chen, L. B. Intracellular heterogeneity in mitochondrial membrane
potentials revealed by a J-aggregate forming lipophilic cation. Proc. Natl. Acad. Sci.
USA, 88: 3671–3675, 1991.
28. Bulychev, A., Trouet, A., and Tulkens, P. Uptake and intracellular distribution of
neutral red in cultured fibroblasts. Exp. Cell Res., 115: 343–355, 1978.
29. Sousa, C., Sa e Melo, T., Geze, M., Gaullier, J. M., Maziere, J. C., and Santus, R.
Solvent polarity and pH effects on the spectroscopic properties of neutral red:
application to lysosomal microenvironment probing in living cells. Photochem. Photobiol., 63: 601– 607, 1996.
30. Egorin, M. J., Hildebrand, R. C., Cimino, E. F., and Bachur, N. R. Cytofluorescence
localization of adriamycin and daunomycin. Cancer Res., 34: 2243–2245, 1974.
31. Gervasoni, J. E., Jr., Fields, S. Z., Krishna, S., Baker, M. A., Rosado, M., Thuraisamy,
K., Hindenburg, A. A., and Taub, R. N. Subcellular distribution of daunorubicin in
P-glycoprotein-positive and –negative drug-resistant cell lines using laser-assisted
confocal microscopy. Cancer Res., 51: 4955– 4963, 1991.
32. Marquardt, D., and Center, M. S. Drug transport mechanism in HL60 cells isolated for
resistance to adriamycin: Evidence for nuclear drug accumulation and redistribution
in resistant cells. Cancer Res., 52: 3157–3163, 1992.
33. Johnson, L. V., Walsh, M. L., and Chen, L. B. Localization of mitochondria in living
cells with rhodamine 123. Proc. Natl. Acad. Sci. USA, 77: 990 –994, 1980.
34. Costantini, P., Jacotot, E., Decaudin, D., Kroemer, G. Mitochondrion as a novel target
of anticancer chemotherapy. J. Natl. Cancer Inst., 92: 1042–1053, 2000.
35. Koya, K., Li, Y., Wang, H., Ukai, T., Tatsuta, N., Kawakami, M., Shishido, T., and
Chen, L. B. MKT-077, a novel rhodacyanine dye in clinical trials, exhibits anticarcinoma activity in preclinical studies based on selective mitochondrial accumulation.
Cancer Res., 56: 538 –543, 1996.
36. Weisberg, E. L., Koya, K., Modica-Napolitano, J., Li, Y., and Chen, L. B. In vivo
administration of MKT-077 causes partial yet reversible impairment of mitochondrial
function. Cancer Res., 56: 551–555, 1996.
37. Modica-Napolitano, J. S., Koya, K., Weisberg, E., Brunelli, B. T., Li, Y., and Chen,
L. B. Selective damage to carcinoma mitochondria by the rhodacyanine MKT-077.
Cancer Res., 56: 544 –550, 1996.
38. Modica-Napolitano, J. S., Brunelli, B. T., Koya, K., and Chen, L. B. Photoactivation
enhances the mitochondrial toxicity of the cationic rhodacyanine MKT-077. Cancer
Res., 58: 71–75, 1998.
39. Fellous, R., Coulaud, D., El Abed, I., Roques, B. P., Le Pecq, J. B., Delain, E., and
Gouyette, A. Cytoplasmic accumulation of ditercalinium in rat hepatocytes and
induction of mitochondrial damages. Cancer Res., 48: 6542– 6549, 1988.
40. Soslau, G., and Nass, M. M. Effects of ethidium bromide on the cytochrome content
and ultrastructure of L1210 cell mitochondria. J. Cell Biol., 51: 514 –524, 1971.
41. Nass, M. M. Analysis of methylglyoxal bis(guanylhydrazone)-induced alterations of
hamster tumor mitochondria by correlated studies of selective rhodamine binding,
ultrastructural damage. DNA replication, and reversibility. Cancer Res., 44: 2677–
2688, 1984.
42. Downes, C. S., Ord, M. J., Mullinger, A. M., Collins, A. R., and Johnson, R. T.
Novobiocin inhibition of DNA excision repair may occur through effects on mitochondrial structure and ATP metabolism, not on repair topoisomerases. Carcinogenesis (Lond.), 6: 1343–1352, 1985.
43. Ueda, T., Hirai, K., and Ogawa, K. Effects of paraquat on the mitochondrial structure
and Ca-ATPase activity in rat hepatocytes. J. Electron Microsc. (Tokyo), 34: 85–91,
1985.
44. Wessendorf, M. W. Fluoro-Gold: composition, and mechanism of uptake. Brain Res.,
553: 135–148, 1991.
45. Snapper, I., Schneid, G., Greenspan, E., and Lieben, F. On the deposition of stilbamidine and 2-hydroxystilbamidine in cytoplasm and nuclei of different organs and
tumors. Bull. N. Y. Acad. Med., 26: 269 –270, 1950.
46. Snapper, I., Schneid, G., and Kurnik, N. The presence of diamidines in myeloma cells
after treatment with stilbamidine and 2-hydroxystilbamidine. Acta Haematol., 3:
129 –134, 1950.
47. Snapper, I., Schneid, G., Lieben, F., Gerber, I., and Greenspan, E. Influence of nucleic
acid upon the fluorescence of nuclei and cytoplasm after injection of aromatic
diamidines. J. Lab. Clin. Med., 37: 562–574, 1951.
48. Festy, B., and Daune, M. Hydroxystilbamidine. A nonintercalating drug as a probe of
nucleic acid conformation. Biochemistry, 12: 4827– 4834, 1973.
49. Festy, B., Sturm, J., and Daune, M. Interaction between hydroxystilbamidine and
DNA. I. Binding isotherms and thermodynamics of the association. Biochim. Biophys. Acta, 407: 24 – 42, 1975.
50. Gresh, N., and Pullman, B. A theoretical study of the nonintercalative binding of
berenil and stilbamidine to double-stranded (dA-dT)n oligomers. Mol. Pharmacol.,
25: 452– 458, 1984.
51. de Duve, C., de Barsy, T., Poole, B., Trouyet, A., Tulkens, P., and Van Hoof, F.
Lysosomotropic agents. Biochem. Pharmacol., 23: 2495–2531, 1974.
52. Ohkuma, S., and Poole, B. (1981) Cytoplasmic vacuolation of mouse peritoneal
macrophages and the uptake into lysosomes of weakly basic substances. J. Cell. Biol.,
90: 656 – 664, 1981.
53. Marchetti, P., Susin, S. A., Decaudin, D., Gamen, S., Castedo, M., Hirsch, T.,
Zamzami, N., Naval, J., Senik, A., and Kroemer, G. Apoptosis-associated derangement of mitochondrial function in cells lacking mitochondrial DNA. Cancer Res., 56:
2033–2038, 1996.
7229