Synthesis and biological evaluation of novel oxophenylarcyriaflavins as potential anticancer agents

Article abstract of DOI:10.1039/b801121d

We report the synthesis and biological evaluation of new oxophenylarcyriaflavins designed as potential anticancer agents. An efficient synthesis involving palladium-catalyzed Suzuki and Stille reactions is presented, without any indolic protective group. The central ring closure of the scaffold was performed through an electrophilic reaction on the position C-2 of the indole ring. The use of indole and 5-benzyloxyindole, along with substituted phenyl rings, generated three different scaffolds, which were successively exploited to modulate the structure. The cytotoxicity of the newly designed compounds on four cancer cell lines and activities against three kinases (CDK1, CDK5 and GSK3) were evaluated. Several compounds showed a marked cytotoxicity with IC₅₀ values in the sub-micromolar range, and induced important cell cycle perturbations, with a G2/M arrest. Some compounds revealed DNA binding properties and were found to inhibit topoisomerase-mediated DNA relaxation of supercoiled DNA, but these properties are not mandatory for a cytotoxic action. A novel lead compound (32) has been identified and warrants further investigations. The Royal Society of Chemistry 2008.

Full text of DOI:10.1039/b801121d

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Synthesis and biological evaluation of novel oxophenylarcyriaflavins as
potential anticancer agents†
Aure´lie Bourderioux,a Vale´rie Be´ne´teau,a Jean-Yves Me´rour,a Brigitte Baldeyrou,b Caroline Ballot,b
Ame´lie Lansiaux,b Christian Bailly,‡b Re´my Le Gue´vel,c Christiane Guillouzoc and Sylvain Routier*a
Received 21st January 2008, Accepted 3rd March 2008
First published as an Advance Article on the web 15th April 2008
DOI: 10.1039/b801121d
We report the synthesis and biological evaluation of new oxophenylarcyriaflavins designed as potential
anticancer agents. An efficient synthesis involving palladium-catalyzed Suzuki and Stille reactions is
presented, without any indolic protective group. The central ring closure of the scaffold was performed
through an electrophilic reaction on the position C-2 of the indole ring. The use of indole and
5-benzyloxyindole, along with substituted phenyl rings, generated three different scaffolds, which were
successively exploited to modulate the structure. The cytotoxicity of the newly designed compounds on
four cancer cell lines and activities against three kinases (CDK1, CDK5 and GSK3) were evaluated.
Several compounds showed a marked cytotoxicity with IC50 values in the sub-micromolar range, and
induced important cell cycle perturbations, with a G2/M arrest. Some compounds revealed DNA
binding properties and were found to inhibit topoisomerase-mediated DNA relaxation of supercoiled
DNA, but these properties are not mandatory for a cytotoxic action. A novel lead compound (32) has
been identified and warrants further investigations.
cytotoxic action seems relatively minor when compared to the NB-
506-type series. The data suggested that other signalling proteins,
particularly kinases, may play a role.
Introduction
Indolo[2,3-a]pyrrolo[3,4-c]carbazole alkaloids form a class of
compounds endowed with potent antitumor, antiviral and/or
In order to develop selective kinase inhibitors, fitting in the ATP
antimicrobial activities.1 This family has raised considerable
binding site, modifications of the aromatic heterocyclic indolocar-
attention because of the central role of these molecules in
bazole moiety appeared as a valid alternative to the synthesis of
the regulation of cell cycle progression and specific enzyme
glycosylated compounds. In addition, most of the aryl carbazoles
inhibition.2,3 Structure–activity relationships (SAR) in the indolo-
designed so far for kinase inhibition possess an unsubstituted
carbazole series have been extensively studied in the context of
maleimide whereas in general the most cytotoxic compounds bear
topoisomerase I inhibition and tumor cell killing. Compounds
hydrophilic side chains. It is noteworthy to mention also that
bearing a pyrroloindolocarbazole and possessing one N-glycosidic
closely related heteroarylcarbazoles (type III) have been described
as inhibitors of Cyclin D1/CDK4.14,15 Cyclin dependent kinases
(CDKs) were also targeted with indolocarbazoles such as bis
N-indolyl alkylated arcyriaflavins recently described as CDK1,
CDK2 and CDK4 inhibitors.16,17 In the NH maleimide series,
indole versus (hetero)arylcarbazole replacement represents also
an interesting strategy.
With this in mind, we envisaged the synthesis of different
naphthalenic compounds IV.18 The replacement of one of the
two indoles by a naphthalene ring produced highly cytotoxic
naphthocarbazoles IV, suggesting that the naphthalene is effec-
tively a suitable bio-isostere for indole. At the same time, we
developed a phenylcarbazole series V in order to broaden our
SAR knowledge.19 In both series, the SAR studies showed that
i) for the phenylcarbazoles of type V, unsubstituted maleimide
compounds were non cytotoxic and enhanced the inhibition of
CDK, ii) the presence of a basic side chain on the maleimide led to
cytotoxic agents with weak CDK inhibition, iii) the introduction
of a hydroxyl group on the indole in position 5 greatly influenced
the biological properties.
bond, such as the antibiotic rebeccamycin, generally function as
DNA topoisomerase I inhibitors.4 A few analogues such as NB-
506 and J-107088 (also known as Edotecarin) have entered clinical
trials for cancer treatment.5–7 More recently, fluorinated derivatives
of such molecules have been reported and their topoisomerase
I-dependent anticancer activity looks promising.8–10 We have
recently described the bioisosteric replacement of an indole
moiety by a 7-azaindole unit to afford the first symmetrical
and dissymmetrical 7-azaindolocarbazoles I and II (Fig. 1).11
In the same vein, the cytotoxic properties of the N-glycosylated
derivatives of I and II have been reported by others.12,13 For these
different molecules, the role of topoisomerase I inhibition in the
aInstitut de Chimie Organique et Analytique, Universite´ dꢀOrle´ans, CNRS
UMR 6005, Rue de Chartres, BP 6759, 45067, Orle´ans Cedex 2, France.
E-mail: sylvain.routier@univ-orleans.fr; Fax: +33 2 3841 7281; Tel: +33 2
3849 4853
bINSERM U-837, IRCL, Place de Verdun, 59045, Lille, France
cHoˆpital de Pontchaillou, INSERM U-522, 65033, Rennes Cedex, France
† Electronic supplementary information (ESI) available: Preparation and
characterization of compounds 4, 10–14, 16–22, 24, 26–28, 32, 33, 35, 36.
See DOI: 10.1039/b801121d
‡ Present address: Institut de Recherches Pierre Fabre, 3 rue des satellites,
31140 Toulouse, France.
We have identified a few studies on bis-indolic compounds such
as homoarcyriaflavin VI, arcyriacyanin A VII which were synthe-
sized but as yet no significant activities have been reported.20,21 In
2108 | Org. Biomol. Chem., 2008, 6, 2108–2117
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Fig. 1 Modifications of the indolocarbazole skeleton.
the present article, we report the design and the synthesis of new
hybrid derivatives type VIII containing: an indole, an unusual
seven-membered ring and a fused phenyl group. The central
tropone moiety in VIII was surmounted by a fused substituted
maleimide ring. In addition, we introduced molecular diversity by
using indole or 5-hydroxyindole for the left part of the molecule
and phenyl or dihydroxyphenyl rings for the right part to improve
the SAR in this new series.
1 and the commercially available 2-formylphenylboronic acid led
to 3 (Scheme 2). The reaction was carried out in the presence of
K2CO3 in a refluxing mixture of dioxane–water using Pd(OAc)2
as catalyst. After 6 h, compound 3 was obtained in the best yield
Chemistry
In this paper we report an improvement on the previously reported
methodology22 and a general synthesis (Scheme 1), which starts
from indoles and involves (i) the introduction of the maleimide
leading to A, (ii) a Suzuki or Stille cross coupling reaction
adding the 2-formylphenyl or the 2-carboxyphenyl groups, (iii)
the transformation of the aldehyde or the ester into carboxylic
acid, iv) an intramolecular electrophilic cyclization to afford
the attempted cycloheptatrienones (tropones) currently named
oxophenylarcyriaflavins.
Scheme 1 Retrosynthesis of oxophenyl-arcyriaflavins.
Scheme
2
Reagents and conditions: i) 2-formylphenylboronic acid
(1.5 eq.), Pd(OAc)2 (0.1 eq.), K2CO3 (1.8 eq.), dioxane–water 85 : 15,
rflx, from 1 to 3, 6 h, 72%, from 2 to 4, 3.5 h, 79%; ii) Pd(PPh3)4 (0.1 eq.),
toluene, rflx, from 5 to 6, Sn2Bu6 (1.2 eq.), 23 h, 47%, from 5 to 7, Sn2Me6
(1.2 eq.), 2 h, 97%, from 9 to 11, Pd(PPh3)4 (0.2 eq.), Sn2Me6 (1.5 eq.),
6 h, 81%; iii) from 9 to 10, Sn2Bu6 (1.2 eq.), PdCl2(PPh3)2 (0.1 eq.), LiCl
(0.5 eq.), toluene, rflx, 24 h, 54%; iv) PdCl2(PPh3)2 (0.2 eq.), CuI (0.1 eq.),
dioxane, rflx, from 1 and 6 (1.5 eq.) to 12, 6 h, 74%, from 1 and 10 (1.5 eq.)
to 13, 5.5 h, 78%, from 2 and 6 (1.5 eq.) to 14, 3 h 45 min, 75%.
A. Palladium catalyzed reactions
Following this strategy, we began the synthesis by preparing
compounds 1 and 2, classically obtained from indole or 5-
benzyloxyindole and 2,3-dibromo-N-methylmaleimide in the pres-
ence of LiHMDS in THF in a fairly good yield.18 First, a
palladium-catalyzed cross coupling Suzuki type reaction between
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2108–2117 | 2109
©

View Article Online
of 72%.18,22 Applying these conditions to the 5-benzyloxyindole
derivative 2 led to the desired compound 4 in a 79% yield.
An alternative method to insert a carbonyl group between the
indole and the phenyl rings consisted in the direct introduction
of the 2-carboxymethylphenyl unit using 5, 8 and 9 as starting
materials. We have previously prepared 6 and 7.22–24 By increasing
the amount of Pd(PPh3)4 and Sn2Me6 we obtained 11 from 9 in
an 81% yield. The compound 10 was obtained from 9 in best
yield (54%) by modifying the nature of the palladium catalyst
(PdCl2(PPh3)2) and adding LiCl.
tional hours, to the maleic anhydride compound 17 in a quan-
titative yield. Starting from 13 and 14, this reaction required
adaptation of reaction times but yielded 18 and 19 in average
to excellent yields.
In parallel, we cleaved the benzyl group of 14, 16 and 19 using
BBr3 at room temperature to generate respectively compounds 20,
21 and 22 in very good yields. Basic treatments of compounds 15,
16 and 21 led to 17, 19 and 22 respectively in satisfying yields.
C. Electrophilic intramolecular reaction
The Stille procedure applied to the bromo compound 1 and stan-
nylated derivative 6 (1.5eq.) was nextperformed with PdCl2(PPh3)2
and CuI in refluxing dioxane. After 6 h, the desired compound
12 was obtained in 74% yield. Under similar conditions, the
trimethyltin derivative 7 led to the same product 12 after 5 h in a
70% yield, and 10 afforded 13 in a 78% yield. Disappointingly, the
trimethylstannyl derivative 11 led to 13 in only 53% yield in the
best case. From 2 and 6, the desired compound 14 was isolated
without any difficulty in a fairly good yield (75%).
The ring closure leading to the central seven-membered cycle
through an electrophilic cyclization without any pre-activation
of the carboxylic acid function, was the next goal. Based on our
knowledge, mild conditions involving a large excess of BF3·Et2O
in refluxing DCE had to be applied.22 After a few hours, the
maleimide containing compounds 15 or 21 led to compound 23
and 24 in 89% and 60% yield, respectively. The same reaction
carried out starting from 16 led to an intractable mixture. The ben-
zyloxy group proved very sensitive toward Lewis acids (Scheme 4).
B. Synthesis of the carboxylic acids
The next step was aimed at generating a C-2 indolic keto function
by oxidation of the aldehydes 3 (Scheme 3). This reaction was
carried out with an aqueous solution of sodium chlorite and
sulfamic acid at 6 ◦C, by accurate control of the temperature
during the oxidant addition, and led to the acid 15 in a 81%
yield.22 Applying these conditions to compound 4 led to 16 in only
◦
a 59% yield. Lowering the temperature to 2 C (limits of freeze)
afforded 16 in 83% yield.
Scheme 4 Reagents and conditions: i) BF3. Et2O (40 eq.), DCE, rflx, from
15 to 23, 5.5 h, 89%, from 21 to 24, 2 h, 60%, from 17 to 25, 20 h, 70%,
from 18 to 26, 5.5 h, 32%, from 22 to 28, 19 h, 60% ii) BBr3 (25 eq.), DCE,
0 ◦C to rflx, 47 h, from 18 to 27, 88% iii) a) aq. KOH (45%, 40 eq.), acetone,
rt, b) HCl conc. to pH = 1, rt, from 23 to 25, a) 24 h, b) 12 h (overnight),
quant., from 24 to 28, a) 23 h and b) 2 days, 90%.
The intramolecular reaction performed with the anhydride 17,
required 20 h of reaction time to improve the yield of 25 to
70%. Compound 22 led to impure 28 in only 60% (as judged
by 1H NMR). Starting from 18, the reaction led, even after flash
chromatography to a mixture of compounds but a supplementary
recrystallization afforded only 26 in a 32% yield.
We next attempted to selectively demethylate 18 with BBr3
in dichloromethane. In all cases, a concomitant electrophilic
cyclization, leading to complex mixtures, occurred. Thus, we set
out for one pot cleavage and intramolecular reaction using a large
excess of BBr3 (25 eq.). At room temperature, after 18 h, only 25%
of the desired product 27 was isolated. Increasing the temperature
to reflux of dichloroethane led to the targeted compound in
a good yield (88%) after 47 h. Disappointed by the problems
associated with preparation of 28 from 22, we decided to treat
N-methylmaleimide 24 with alcoholic KOH solution for 23 h.
Scheme 3 Reagents and conditions: i) NH2SO3H (8.1 eq.), NaClO2
(2.2 eq.), dioxane–water, 1 min, from 3 to 15, 6 ◦C, 81%, from 4 to 16,
2 ◦C, 83%; ii) a) aq. KOH (45%, 60 eq.), acetone, rt, b) HCl conc. to pH =
1, rt; from 12 to 17, a) 20 h, b) overnight, quant., from 13 to 18, a) 2 h,
reflux and b) overnight, quant., from 14 to 19, a) KOH (40 eq.), 7 h and b)
overnight, 36%, from 15 to 17, a) KOH (40 eq.), 28 h, and b) 3 min, 91%,
from 16 to 19, a) 24 h, and b) 24 h, 95%, from 21 to 22, a) KOH (40 eq.),
25 h and b) 3 days, 94%; iii) BBr3 (25 eq.), CH2Cl2, 0 ◦C to rt, 5 min, from
14 to 20, 86%, from 19 to 22, quant., from 16 to 21, 83%.
The second method used was hydrolysis of the ester 12 using
a large excess of an aqueous KOH solution in acetone for 20 h
at room temperature. Then, acidification with a concentrated
hydrochloric acid solution (pH = 1) led, after a few addi-
2110 | Org. Biomol. Chem., 2008, 6, 2108–2117
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
After the acidification step the attempted compound 28 was
obtained in a 90% yield. Similarly, compound 23 furnished the
derivative 25 in a quantitative yield.
D. Synthesis of functionalized oxophenylarcyriaflavins
The first modification we attempted was the introduction of
a dimethylaminoethyl side chain on the maleimide or maleic
anhydride part of scaffold 23, 25, 27, 28 (Scheme 5, Table 1,
entries 1–10). Previously, we reported that, in boiling dimethy-
laminoethylamine, in a sealed tube and in the presence of DMF,
the reaction of compound 23 led only to the reduced compound
29 in 41% yield.22 Under similar conditions, the reaction of the
more reactive anhydride 25 afforded compound 31 in 59% yield
(entries 1,2). Application of these reaction conditions to our
new derivatives 27 and 28 was fruitless. These results prompted
us to investigate other conditions. In open glassware, in boiling
dimethylaminoethylamine, compound 30 was synthesized either
from maleimide 23 or from anhydride 25 in 48% and 86% yield,
respectively (entries 3, 4). All attempts starting from 27 failed
again.
In order to reduce the reaction time, we next investigated the
reactivity of 25, 27 and 28 in the presence of dimethylaminoethy-
lamine under microwave irradiation (Biotage apparatus).25 In
the above diamine, at 150 ◦C, compound 25 led to the desired
compound 31 after 15 min, but in only 10% yield. It is possible
that the poor solubility of 25 causes the low yield. Next, the
two reagents were adsorbed on a solid phase.26 From the furo
derivatives 25 and 28, the direct absorption on silica gel with
5.0 eq. of diamine led to the attempted products 31 and 32 in
an improved yield of 48% and 33% respectively. Starting materials
(30%) and reduced compounds 30 or 33 were also isolated during
the purification steps (entries 5, 6).
The next attempted variation dealt with the reaction of 25
with ammonia in the presence of DMF (entry 7).19 Refluxing for
67 h an aqueous solution of NH4OH (30%) containing 25 led
to compound 34 in 70% yield with no trace of reduction of the
tropone. Another solution to perform the NH insertion consisted
in the reaction of the anhydride 25 with melted NH4OAc. After
2 h, the derivative 34 was isolated in a best yield of 81%. The
major advantage of this method was its compatibility with the
Scheme 5 For reagents and conditions see Table 1.
presence of quinonic or quinonimine sensitive systems. So, we
next performed the reaction with the two anhydrides 27 and 28
to isolate compounds 35 and 36 in 65 and 73% yield, respectively
(entries 8–10).
Then, we used compounds 23 and 34 to introduce the dimethy-
laminoethyl side chain onto the free indolic nitrogen atom (entries
11, 12) in the presence of NaH and chloroethyldimethylamine in
DMF at 90 ◦C. From compound 23, the derivative 37 was isolated
in a 84% yield, whereas starting from 34, both nitrogen atoms
reacted with the chloroalkyl derivative and the compounds 31 and
Table 1 Substitution of derivatives, 23, 25, 27, 28 and 34
Entry
Starting material
Reagent
Conditions
Temperature
Time
Product (yield)
1
2
3
4
5
6
7
8
9
10
11
12
13
23
25
23
25
25
28
25
25
27
28
23
34
34
Excess of NH2(CH2)2N(CH3)2
Excess of NH2(CH2)2N(CH3)2
Excess of NH2(CH2)2N(CH3)2
Excess of NH2(CH2)2N(CH3)2
NH2(CH2)2N(CH3)2 (5 eq.)
NH2(CH2)2N(CH3)2 (5 eq.)
Aqueous NH4OH
Fused NH4OAc in excess
Fused NH4OAc in excess
Fused NH4OAc in excess
ClCH2CH2N(CH3)2 (2.5 eq.)
ClCH2CH2N(CH3)2 (2.5 eq.)
ClCH2CH2N(CH3)2 (2.5 eq.)
Sealed tube, DMF
Sealed tube, DMF
Open flask without solvent
Open flask without solvent
lwave, SiO2 (8.5 eq.)
lwave, SiO2 (8.5 eq.)
DMF
rflx
rflx
rflx
rflx
22 h
18 h
22 h
64 h
29 (41%)b
31 (59%)b
30 (48%)b
30 (86%)
150 ◦
150 ◦
rflx
C
C
15 min
31 (48%)b, 30 (15%)a
32 (33%)b, 33 (9%)a
34 (70%)b
67 h
2 h
3 h
130 ◦
130 ◦
130 ◦
C
C
C
34 (81%)b
35 (65%)b
3 h 30 min 36 (73%)b
NaH (3.8 eq.), DMF
NaH (3.8 eq.), DMF
K2CO3
90 ◦
90 ◦
65 ◦
C
C
C
9 h
24 h
24 h
37 (84%)b
31 (13%)b, 38 (30%)b
31 (33%)b, 38 (10%)b
a Estimated by 1H NMR. b Isolated product.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2108–2117 | 2111
©

View Article Online
38 were obtained in 13 and 30% yield respectively. Replacing NaH
for K2CO3 and performing the reaction at 65 ◦C slightly favored
the maleimide substitution and compound 31 was isolated in 33%
yield whereas only 10% of 38 was obtained (entry 13).
Biological results and discussion
The newly synthesized compounds were tested for DNA binding
and topoisomerase inhibition. Their cytotoxic properties were also
measured. The data are collected in Table 2. DNA interaction was
first estimated using a conventional melting temperature assay
with calf thymus DNA (CT-DNA, 42% GC). Unsurprisingly, the
molecules bearing a dimethylaminoethyl side chain were found to
stabilize CT-DNA whereas the other neutral molecules showed
little, if any, interaction with nucleic acids. Compounds 31 and
32 bearing the cationic chain on top of the maleimide ring
showed marked interaction with DNA, with DTm of 7 ◦C and
these two compounds were markedly cytotoxic toward HL60
human leukemia cells. Compound 32 showed the most cytotoxic
activity in the series, with an IC50 of 0.3 lM close to that of
the reference antitumor antibiotic rebeccamycin. The hydroxyl
group on the indole ring contributed modestly to DNA binding
but, more significantly, to cytotoxicity (and solubility as well).
Interestingly, by comparing compounds 30 and 31, as well as 32
and 33, it clearly appeared that the reduction of the keto group
of the central tropone ring strongly reduced the DNA binding
capacity and abolished the cytotoxic effects, suggesting that DNA
interaction may contribute (to some extent) to the biological
activity of 31 and 32. This conclusion is also supported by the
observation that the other neutral compounds that bind weakly,
or with no detectable binding to DNA, correspondingly showed
no significant cytotoxic properties. However, DNA interaction is
not sufficient to determine cytotoxicity since compound 38, with
two dimethylaminoethyl side chains on the maleimide and indole
rings, was found to be poorly cytotoxic towards HL60 cells despite
its superior propensity to DNA binding.
drug-DNA complex
Fig. 2 Melting temperature variation DTm (Tm
− TmDNA alone in
◦C) of poly(dAdT)2 after incubation with drugs at increasing drug/DNA.
Tm measurements were performed in BPE buffer, in 1 cm quartz cuvettes
at 260 nm with a heating rate of 1 ◦C min−1. The Tm values were obtained
from first-derivative plots.
DNA binding of four key molecules (30–32, 38) was investigated
further using the alternative polymer poly(dAdT)2. Under the
experimental conditions used (16 mM Na+), the helix-to-coil
transition occurs at 42 ◦C in the absence of drugs. The Tm was
shifted to higher temperatures with the four drugs and more
pronounced with the keto derivative 31 as compared to the
methylene analogue 30 (Fig. 2). Thus corroborating the previous
observation with CT-DNA, we can conclude that the keto group
is directly involved in DNA recognition (through H-bonding,
for example) and/or indirectly, by providing a more planar
chromophore. The DTm values reached 18 ◦C with compound
38, and the molecules rank order 38 > 31, 32 > 30.
The mode of binding to DNA was investigated by spectroscopic
methods. These compounds intercalate into DNA and this is
particularly clear with compound 38. DNA induces significant
shifts in the UV–Visible spectra of 38 with a marked decrease in
the extinction coefficient around 450 nm and a shift of the band to a
much longer wavelength (Fig. 3). The UV–Visible spectral changes
are more subtle with 32. Circular dichroism (CD) measurements
showed that a negative band centered at 350 nm appeared upon
addition of CT-DNA. This band reflects the orientation of the
tetracyclic chromophore bound to the double helix and is entirely
consistent with an intercalative binding process. Intercalating
Fig. 3 (left) Absorption and (right) CD DNA titration of compound 38
and 32 in BPE buffer for absorption and cacodylate buffer for CD DNA
titration. Aliquots of a concentrated calf thymus DNA solution were added
to 1 ml of a drug solution (20 lM for the absorbance and 50 lM for the
CD measurements). The drug/DNA ratios increased from 0 to 1.
agents usually (but not always) give this type of negative CD
signal whereas positive signals are commonly seen with minor
groove binders. The negative CD band was not observed with 32
but the unwinding data (see below) also support intercalation.
Based on the structural analogy with rebeccamycin, the com-
pounds were tested as potential inhibitors of topoisomerases
I and II. In all cases, none of the compound was found to
act as a “poison” capable of stabilizing topoisomerase–DNA
complexes, as can be commonly detected with the reference
drugs camptothecin (for topoisomerase I) or etoposide (for
topoisomerase II) (data not shown). An interference with the
catalytic activity of the enzyme was observed in some cases,
but this only reflects the DNA-binding capacity. As shown in
Fig. 4, concentration-dependent inhibition of the relaxation of
a supercoiled plasmid DNA with topoisomerase I was detected
2112 | Org. Biomol. Chem., 2008, 6, 2108–2117
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 2 In vitro cytotoxic effect IC50 (lM); Fibro = human diploid skin fibroblastic cells (normal cell line); HL60 human leukemia cells (tumor cell line),
Caco2 = human colon carcinoma cells (tumor cell line);27 Huh7 = human hepatocarcinoma cells (tumor cell line);28 F1 = RLEC for rat biliary epithelial
cells clone F1 (tumor cell line)29
a
Compound
DTm
Fibroc
HL60b
Cacoc
Huh7c
F1c
23 R1 = H
0
0
>25
>100
10
2
>25
10
20
24 R1 = OH
10
27
30
29
ND
>25
ND
20
4
4
25 R1 = R3 = R4 = H
0
0
0
0
>25
10
>25
>25
52
50
>100
>100
2
10
>25
>25
>25
5
>25
>25
0.2
10
0.4
>25
26 R1 = H, R3 = R4 = OMe
27 R1 = H, R3 = R4 = OH
28 R1 = OH, R3 = R4 = H
31 R1 = H
6.8
7.1
10
0.6
1.22
0.31
1
0.5
2
1
0.4
0.07
32 R1 = OH
30 R1 = H
2.5
ND
10
>100
72
ND
3
30
30
10
0.7
4
33 R1 = OH
37 R = CH3
ND
9.9
10
>25
ND
17
10
7
2
6
2
2
38 R = CH2CH2N(CH3)2
34 R1 = R3 = R4 = H
1.2
0.9
0
10
>25
>25
22
5.81
37
0.3
10
>25
>25
0.3
10
20
2
20
35 R1 = H, R3 = R4 = OH
36 R1 = OH, R3 = R4 = H
References
Roscovitin
Rebeccamycin
ND
ND
20
ND
ND
0.16
2
ND
3
ND
5
ND
◦
ND = not determined.a Variations in melting temperature, DTm = Tm
− Tm
in C. Tm measurements were performed in BPE buffer,
drug-DNA complex
CTDNA alone
◦
in 1 cm quartz vials at 260 nm with a heating rate of 1 C min−1. The Tm values were obtained from first-derivative plots. b Drug concentrations that
inhibit circulating HL60 leukemia cells from solid tumors growth by 50% after 72 h of incubation. Roscovitin and Rebeccamycin were used as internal
standard for cell assays. c Drug concentrations that inhibit circulating adherent cells from solid tumors growth by 50% after 48 h of incubation. Roscovitin
and Rebeccamycin were used as internal standard for cell assays.
on poly-acrylamide gels. The unwinding of DNA resulting from
the insertion of the drug between base pairs is clearly evidenced
with several compounds, in particular with 31 and 38. In these
cases, the topoisomer population is shifted to the top part of
the gel at 20 lM (relaxation) and at higher concentrations, the
DNA becomes positively supercoiled because of intercalation and
it starts to migrate faster in the gel. Together with the CD data,
this result effectively attests that these compounds intercalate into
DNA.
The cytotoxicity was evaluated further using a pair of murine
leukemia cells, either sensitive or resistant to the antitumor drug
camptothecin (CPT) which is a reference topoisomerase I poison.
The resistance of the P388CPT5 cells has been attributed to the
expression of a deficient form of topoisomerase I as a result of
a mutation in the top1 gene of these cells.30 The two mutations
(Gly361Val and Asp709Tyr) in conserved regions of the top1 gene
strongly diminish the sensitivity of the cells to CPT, by a factor
>100 (Table 3).
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2108–2117 | 2113
©

View Article Online
Table 3 Cytotoxicity on sensitive and resistant P388 cell line
Compound
P388 (IC50/lM)
P388CPT5 (IC50/lM)
RRIa
Camptothecin
Rebeccamycin
0.031
0.19
3.310
0.71
107
3.7
31
32
0.70
0.231
1.02
0.521
1.46
2.26
a Relative resistance index: IC50(CPT-resistant)/IC50(CPT-sensitive).
circulating HL-60 cells and P388 cells as well as with adherent
cancer cells derived from solid tumors, such as human colon
carcinoma Caco2 cells and human hepatocarcinoma Huh7 cells
(Table 2). In addition, very actively proliferating transformed rat
biliary epithelial cells were also found to be exquisitely sensitive to
this molecule, with an IC50 of 700 nM (Table 2). Meanwhile, this
compound was also found to be active onto non-tumoral poorly
proliferating human skin fibroblastic cells, suggesting distinct
signaling pathways that warrant further investigation.
In parallel, the effect on the cell cycle of P388 and P388CPT5
leukemia cells of DNA-binding compounds 30–32, 38 was in-
vestigated (Fig. 5). Compounds 31 and 32 induced a strong
accumulation of the cells in the G2 + M phase at 1 lM, as observed
with camptothecin, whereas 38 and 30 showed no significant effect
at the same concentration. A similar trend was observed with the
topoisomerase I-mutated cell line P388CTP5. It is clear that the
mutation of the top1 gene does not impact on the activity of the
two oxophenylarcyriaflavin compounds.
Fig. 4 Inhibition of topoisomerase I-mediated DNA plasmid pUC19 in
the presence of graded concentrations of drugs. Plasmid DNA (120 ng,
lane DNA) was incubated for 45 min at 37 ◦C with drug, then 30 min at
50 ◦C with 4 units of topoisomerase I (topogen Inc) in the absence (lane
topoI) or in the presence of drug at the indicated concentrations (lM).
Reaction was stopped with SDS and treatment with proteinase K. The
DNA was analyzed by agarose gel electrophoresis. The gel was stained
with ethidium bromide and photographed under UV light. Nck: nicked;
Rel: relaxed; Sc: supercoiled.
Kinase inhibition
It is well known that (hetaryl)pyrrolo[3,4-c]carbazole derivatives
can induce selective protein kinase inhibition.14–17,32 So we next
tested the compounds for potential inhibition of CDK1, CDK5
and glycogen synthase kinase-3 (GSK3) to evaluate a possible
selectivity (Table 4).19 The non cytotoxic compound 36 showed a
significant inhibitory activity (IC50 CDK1 = 1.3 lM, IC50 CDK5 =
0.9 lM IC50 GSK3 = 2.2 lM), without any selectivity, thus
indicating the major role of the unsubstituted maleimide group in
In contrast, these mutations showed little effect on the cytotoxic
potential of 31 and 32, as is the case for rebeccamycin. Topoiso-
merase I is not the main target of these compounds. Nevertheless,
the use of these additional pair of cell lines confirmed the cytotoxic
bioactivity of 31–32, with IC50 < 1.0 lM. Noteworthy, the higher
cytotoxicity of 32 compared to 31 was observed in all cases, with
Fig. 5 Cell cycle distribution in CEM cells treated for 24 h with graded concentrations of 30–32, 38. Cells were analyzed with the FACScan flow
cytometer. Data are the result of two independent experiments.
2114 | Org. Biomol. Chem., 2008, 6, 2108–2117
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 4 Inhibition tests on CDK1, CDK5 and GSK3
IC50/lM
Compound
CDK1
CDK5
GSK3
23 R1 = H
>10
75
ND
10
>10
>100
24 R1 = OH
25 R1 = R3 = R4 = H
4.5
>10
40
14
>10
30
45
>10
5.7
26 R1 = H, R3 = R4 = OMe
27 R1 = H, R3 = R4 = OH
28 R1 = OH, R3 = R4 = H
19
>100
>100
31 R1 = H
16
8.2
12
32 R1 = OH
>10
>10
>10
30 R1 = H
4
9
20
10
38 R = CH2CH2N(CH3)2
2.6
4.8
34 R1 = R3 = R4 = H
4.3
1.3
3.9
0.9
18
2.2
36 R1 = OH, R3 = R4 = H
this activity. The lack of the hydroxyl group impaired the inhibition
indicating that the indolic hydroxyl group of 36 gives a favourable
interaction with the enzyme catalytic site (compound 34, IC50
CDK1 = 4.3 lM, IC50 CDK5 = 3.9 lM IC50, GSK3 = 18 lM). The
selectivity of CDKs versus GSK3 was enhanced when the indolic
hydroxy group was absent. Surprisingly the last unsubstituted
maleimide compound 35, which is hydroxylated on the phenyl
ring, was inactive against all the tested kinases but exhibited strong
cytotoxic activities against Huh7 and F1 cell lines. This result
indicates also that kinase inhibition and cytotoxicity were not
correlated. Compound 36 constitutes an interesting scaffold from
which more potent and more selective inhibitors could potentially
be designed.
ring or on the maleimide moiety were successfully achieved
to provide access to a new potential anticancer agent family.
Some molecules proved to be potent cytotoxic agents able to
interfere with the cell cycle of cancer cells. DNA binding, but
not topoisomerase inhibition, seems to play a role in the cytotoxic
action. CDK1, CDK5 and GSK3 are apparently not major targets
for these compounds but the oxophenylarcyriaflavin scaffold may
be further exploited to generate kinase inhibitors.
Experimental section
A. Chemistry
1H and 13C NMR spectra were recorded on a Bruker Avance DPX
250 instrument using CDCl3 or DMSO-d6. The chemical shifts are
reported in ppm (d scale) and all J values are in Hz. The following
abbreviations are used: singlet (s), doublet (d), doubled doublet
(dd), triplet (t), multiplet (m), quaternary carbon (Cq). Melting
points are uncorrected. IR absorptions were recorded on a Perkin
Elmer PARAGON 1000 PC and values were reported in cm−1. MS
spectra (Ion Spray) were performed on a Perkin Elmer Sciex API
300 or on Avatar 320 equipped using an ATR (Ge) technique.
Conclusion
To sum up, in this paper we designed new polysubstituted
oxophenylarcyriaflavins starting from indoles using efficient syn-
theses. Preparation of these compounds was achieved using a
Suzuki or a Stille procedure whereas formation of the central
tropone ring required an intramolecular electrophilic reaction.
Different substitutions either on the indolic nucleus, on the phenyl
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2108–2117 | 2115
©

View Article Online
HRMS were performed by the Centre Re´gional de Mesures
Physiques de lꢀOuest (CRMPO, Rennes) on a high resolution
mass spectrometer with double focalisation Varian Mat 311 using
electronic impact. Monitoring of the reactions was performed
using silica gel TLC plates (silica Merck 60 F254). Spots were
visualized by UV light at 254 and 356 nm. Flash chromatography
columns were performed using silica gel 60 (0.040–0.063 mm,
Merck). Microwave experiments were performed on a Biotage
Initiator apparatus. For compounds 4, 12–22 in the 1H NMR data
the Hꢀ refers to phenyl ring protons. Synthesis and experimental
data of compounds 3, 15, 23, 25, 29–31, 34, 37, 38 were previously
reported.22
compounds, ranging from 0.1 to 25 lM in a final volume of 80 ll
of culture medium. They were fixed with 4% paraformaldehyde
solution and nuclei were stained with Hoechst 3342 and counted
using automated imaging analysis (Simple PCI software).
Cell cycle analysis
For flow cytometric analysis of DNA content, 0.7 × 106 cells
in exponential growth were treated with graded concentrations
of the tested drug for 24 h and then washed with 1 mL of
PBS. After centrifugation, the cell pellet was resuspended in 1 mL
of cold ethanol for 24 h at −0 ◦C. The ethanol was removed,
and the pellet was washed with 1 mL PBS and then incubated for
30 min in a solution containing 50 lg mL−1 PI and 100 lg mL−1
RNase. Samples were analyzed on a Becton Dickinson FACScan
flow cytometer using CellQuest software, which was also used
to determine the percent of cells in the different phases of the
cell cycle. PI was excited at 488 nm, and fluorescence analyzed at
620 nm on channel Fl-2.
For the preparation and characterization of compounds 4, 10–
14, 16–22, 24, 26–28, 32, 33, 35, 36, see the ESI.
B. Biological studies
Topoisomerase inhibition assays. These were performed as
previously described.31
DNA binding measurements
Acknowledgements
Melting curves were measured with an Uvikon 943 spectropho-
tometer coupled to a Nesab RTE111 cryostat. Titrations of
the drug with DNA, covering a large range of drug/DNA-
phosphate ratios (D/P), were performed by adding aliquots of a
concentrated drug solution to a constant DNA solution (20 lM).
Tm measurements were performed in BPE buffer pH 7.1 (6 mM
Na2HPO4, 2 mM NaH2PO4, 1 mM EDTA). The temperature
inside the cuvette (10 mm path length)◦was increased over the
range 20–100 ◦C with a heating rate of 1 C min−1. The “melting”
temperature Tm was taken as the mid-point of the hyperchromic
transition.
This research was supported by grants from the Ligue Nationale
Contre le Cancer (Comite´ du Nord) and the Institut de Recherches
sur le Cancer de Lille (A.L.), from the Canceropoˆle Grand Ouest,
the Re´gion Centre and the Association pour la Recherche contre
le Cancer (S.R.). We thank Dr Laurent Meijer and Olivier Lozach
(CNRS Roscoff) for performing the kinase assays.
Notes and references
1 C. Bailly, Targetting DNA and Topoisomerase I with Indolocarbazole
Antitumor Agents, in Small Molecule DNA and RNA Binders, Ed.
M. Demeunynck, C. Bailly and W. Wilson, Wiley-VCH, Weinheim,
2003, vol. 2, pp. 538–575.
Cell culture and survival assay
2 J. Pines, Nat. Cell Biol., 1999, 1, E73–E79.
Human CEM and CEMC2 leukemia cells were obtained from the
American Tissue Culture Collection. Cells were grown at 37 C
3 D. O. Morgan, Annu. Rev. Cell Dev. Biol., 1997, 13, 261–291.
4 M. Prudhomme, Eur. J. Med. Chem., 2003, 38, 123–140.
5 B. H. Long, W. C. Rose, D. M. Vyas, J. A. Matson and S. Forenza, Curr.
Med. Chem.: Anti-Cancer Agents, 2002, 2, 255–265.
◦
in a humidified atmosphere containing 5% CO2 in RPMI 1640
medium, supplemented with 10% fetal bovine serum, 4.5 g L−1
glucose, 10 mM HEPES, 1 mM sodium pyruvate, penicillin
(100 IU mL−1), and streptomycin (100% g mL−1). The cytotoxicity
on these cells of the tested compounds was assessed using a cell
proliferation assay developed by Promega (CellTiter 96 Aqueous
one solution cell proliferation assay). Briefly, 2 × 104 exponentially
growing cells were seeded in 96-well microculture plates with
graded drug concentrations in a volume of 100 lL. After 72 h
6 T. Yoshinari, M. Ohkubo, K. Fukasawa, S. Egashira, Y. Hara, M.
Matsumoto, K. Nakai, H. Arakawa, H. Morishima and S. Nishimura,
Cancer Res., 1999, 59, 4271–4275.
7 H. Arakawa, M. Morita, T. Kodera, A. Okura, M. Ohkubo, H.
Morishima and S. Nishimura, Jpn. J. Cancer Res., 1999, 90, 1163–1170.
8 B. N. Balasubramanian, D. R. St. Laurent, B. H. Long, C. Bachand, F.
Beaulieu, W. Clarke, M. Deshpande, J. Eummer, C. R. Fairchild, D. B.
Frennesson, R. Kramer, F. Y. Lee, M. Mahler, A. Martel, B. N. Naidu,
W. C. Rose, J. Russell, E. Ruediger, M. G. Saulnier, C. Solomon, K. M.
Stoffan, H. Wong, K. Zimmermann and D. M. Vyas, J. Med. Chem.,
2004, 47, 1609–1612.
◦
incubation at 37 C, 20 lL of the tetrazolium dye was added to
each well, and the samples were incubated for a further 2 h at
37 ◦C. Plates were analyzed on a Labsystems Multiskan MS (type
352) reader at 492 nm.
9 M. G. Saulnier, B. N. Balasubramanian, B. H. Long, D. B. Frennesson,
E. Ruediger, K. Zimmermann, J. T. Eummer, D. R. St. Laurent, K. M.
Stoffan, B. N. Naidu, M. Mahler, F. Beaulieu, C. Bachand, F. Y. Lee,
C. R. Fairchild, L. K. Stadnick, W. C. Rose, C. Solomon, H. Wong, A.
Martel, J. J. Wright, R. Kramer, D. R. Langley and D. M. Vyas, J. Med.
Chem., 2005, 48, 2258–2261.
10 M. G. Saulnier, D. R. Langley, D. B. Frennesson, B. H. Long, S. Huang,
Q. Gao, D. Wu, C. R. Fairchild, E. Ruediger, K. Zimmermann, D. R. St.
Laurent, B. N. Balasubramanian and D. M. Vyas, Org. Lett., 2005, 7,
1271–1274.
11 S. Routier, N. Ayerbe, J. Y. Me´rour, G. Coudert, C. Bailly, A. Pierre, B.
Pfeiffer, D. H. Caignard and P. Renard, Tetrahedron, 2002, 58, 6621–
6630.
12 C. Marminon, A. Pierre, B. Pfeiffer, V. Perez, S. Le´once, P. Renard and
M. Prudhomme, Bioorg. Med. Chem., 2003, 11, 679–687.
Skin diplo¨ıd fibroblastic cells were provided by BIOPREDIC
International Company (Rennes, France). Caco-2 cells and Huh7
cells were obtained from the ECAC collection. Cells were grown
according to ECAC recommendations. RLEC-F1 clone is derived
from an established rat biliary epithelial cell line as previously
described.28 The toxicity test of the compounds on these cells
was as follows: 4 × 103 cells were seeded in 96 multiwell plates
and left for 24 h for attachment, spreading and growing. Then,
they were exposed for 48 h to increasing concentrations of the
2116 | Org. Biomol. Chem., 2008, 6, 2108–2117
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
13 C. Marminon, A. Pierre, B. Pfeiffer, V. Perez, S. Le´once, A. Joubert,
C. Bailly, P. Renard, J. Hickman and M. Prudhomme, J. Med. Chem.,
2003, 46, 609–622.
14 C. Sanchez-Martinez, C. Shih, M. M. Faul, G. Zhu, M. Paal, C.
Somoza, T. Li, C. A. Kumrich, L. L. Winneroski, Z. Xun, H. B. Brooks,
B. K. R. Patel, R. M. Schultz, T. B. DeHahn, C. D. Spencer, S. A.
Watkins, E. Considine, J. A. Dempsey, C. A. Ogg, R. M. Campbell,
B. A. Anderson and J. Wagner, Bioorg. Med. Chem. Lett., 2003, 13,
3835–3839.
15 T. A. Engler, K. Furness, S. Malhotra, C. Sanchez-Martinez, C. Shih,
W. Xie, G. Zhu, X. Zhou, S. Conner, M. M. Faul, K. A. Sullivan,
S. P. Kolis, H. B. Brooks, B. Patel, R. M. Schultz, T. B. DeHahn,
K. Kirmani, C. D. Spencer, S. A. Watkins, E. L. Considine, J. A.
Dempsey, C. A. Ogg, N. B. Stamm, B. D. Anderson, R. M. Campbell, V.
Vasudevan and M. L. Lytle, Bioorg. Med. Chem. Lett., 2003, 13, 2261–
2267.
16 M. M. Faul, T. A. Engler, K. A. Sullivan, J. L. Grutsch, M. T. Clayton,
M. J. Martinelli, J. M. Pawlak, M. LeTourneau, D. S. Coffey, S. W.
Pedersen, S. P. Kolis, K. Furness, S. Malhotra, R. S. Al-Awar and J. E.
Ray, J. Org. Chem., 2004, 69, 2967–2975.
21 M. Brenner, G. Mayer, A. Terpin and W. Steglich, Chem.–Eur. J., 1997,
3, 70–74.
22 A. Bourderioux, S. Routier, V. Be´ne´teau and J.-Y. Me´rour, Tetrahedron,
2007, 63, 9465–9475; A. Bourderioux, S. Routier, V. Be´ne´teau and J.-Y.
Me´rour, Tetrahedron Lett., 2005, 46, 6071–6074.
23 A. Kalivretenos, J. K. Stille and L. S. Hegedus, J. Org. Chem., 1991, 56,
2883–2894.
24 K. C. Nicolaou, D. Hepworth, N. P. King, M. Raymond, V. Finlay, R.
Scarpelli, M. Manuela, A. Pereira, B. Bollbuck, A. Bigot, B. Werschkun
and N. Winssinger, Chem.–Eur. J., 2000, 6, 2783–2800.
25 T. Vidal, A. Petit, A. Loupy and R. N. Gedye, Tetrahedron, 2000, 56,
5473–5478.
26 A. Mortoni, M. Martinelli, U. Piarulli, N. Regalia and S. Gagliardi,
Tetrahedron Lett., 2004, 45, 6623–6627.
27 L. Mahraoui, A. Rodolosse, A. Barbat, E. Dussaulx, A. Zweibaum,
M. Rousset and E. Brot-Laroche, Biochem. J., 1994, 298, 629–633.
28 H. Nakabayashi, K. Taketa, K. Miyano, T. Yamane and J. Sato, Cancer
Res., 1982, 42, 3858–3863.
29 M. Rialland, A. Corlu, G. Ilyin, F. Cabillic, I. Lamy and C. Guguen-
Guillouzo, Cytokine, 2000, 12, 951–959.
17 G. Zhu, S. E. Conner, X. Zhou, C. Shih, T. Li, B. D. Anderson, H. B.
Brooks, R. M. Campbell, E. Considine, J. A. Dempsey, M. M. Faul,
C. Ogg, B. Patel, R. M. Schultz, C. D. Spencer, B. Teicher and S. A.
Watkins, J. Med. Chem., 2003, 46, 2027–2030.
18 S. Routier, P. Peixoto, J.-Y. Me´rour, G. Coudert, C. Bailly, A. Pierre, S.
Le´once and D.-H. Caignard, J. Med. Chem., 2005, 48, 1401–1413.
19 S. Routier, J.-Y. Me´rour, N. Dias, A. Lansiaux, C. Bailly, O. Lozach and
L. Meijer, J. Med. Chem., 2006, 49, 789–799.
30 J. Madelaine, S. Prost, A. Naudin, G. Riou, F. Lavelle and J. F. Riou,
Biochem. Pharmacol., 1993, 45, 339–348.
31 C. Bailly, Methods Enzymol., 2001, 340, 610–623.
32 G. Zhu, S. E. Conner, X. Zhou, H.-K. Chan, C. Shih, T. A. Engler, R. S.
Al-awar, H. B. Brooks, S. A. Watkins, C. D. Spencer, R. M. Schultz,
J. A. Dempsey, E. L. Considine, B. R. Patel, C. A. Ogg, V. Vasudevan
and M. L. Lytle, Bioorg. Med. Chem. Lett., 2004, 14, 3057–3061; G.
Zhu, S. Conner, X. Zhou, C. Shih, H. B. Brooks, E. Considine, J. A.
Dempsey, C. Ogg, B. Patel, R. M. Schultz, C. D. Spencer, B. Teicher
and S. A. Watkins, Bioorg. Med. Chem. Lett., 2003, 13, 1231–1235.
20 S. Mahboobi, T. Burgemeister, S. Dove, S. Kuhr and A. Popp, J. Org.
Chem., 1999, 64, 8130–8137.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 2108–2117 | 2117
©