Synthesis and evaluation of functionalized isoindigos as antiproliferative agents

Article abstract of DOI:10.1016/j.bmc.2009.09.008

A series of functionalized isoindigos structurally related to meisoindigo (1-methylisoindigo), a therapeutic agent used for the treatment of a form of leukemia, were synthesized and evaluated for antiproliferative activities on a panel of human cancer cells. Two promising compounds (1 -phenpropylisoindigo and 1-(p-methoxy-phenethyl)-isoindigo) that were more potent than meisoindigo and comparable to 6bromoindirubin-3'-oxime on leukemic K562 and liver HuH7 cells were identified. Structure-activity relationships showed the importance of keeping one of the lactam NH in an unsubstituted state. Substitution of the other lactam NH with aryl or arylalkyl side chains retained or improved activity in most instances. An intact exocyclic double bond was also essential, possibly to maintain planarity and rigidity of the isoindigo scaffold. None of the compounds were found to inhibit CDK2 in an in vitro assay, in spite of reports linking the antiproliferative activities of meisoindigo and other isoindigos to CDK2 inhibition. Hence, these functionalized isoindigos disrupted cell growth and proliferation by other mechanistic pathways that did not involve CDK2 inhibition.

Full text of DOI:10.1016/j.bmc.2009.09.008

Bioorganic & Medicinal Chemistry 17 (2009) 7562–7571
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and evaluation of functionalized isoindigos as antiproliferative agents
Xi Kai Wee a, Wee Kiang Yeo a, Bing Zhang b, Vincent B. C. Tan b, Kian Meng Lim b,
Tong Earn Tay b, Mei-Lin Go a,
*
a Department of Pharmacy, Faculty of Science, National University of Singapore, 18 Science Drive 4, Singapore 117543, Singapore
b Nanoscience and Nanotechnology Initiative, Faculty of Engineering, National University of Singapore, 2 Engineering Drive 3, Singapore 117581, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 July 2009
Revised 3 September 2009
Accepted 4 September 2009
Available online 11 September 2009
A series of functionalized isoindigos structurally related to meisoindigo (1-methylisoindigo), a therapeu-
tic agent used for the treatment of a form of leukemia, were synthesized and evaluated for antiprolifer-
ative activities on a panel of human cancer cells. Two promising compounds (1-phenpropylisoindigo and
1-(p-methoxy-phenethyl)-isoindigo) that were more potent than meisoindigo and comparable to 6-
bromoindirubin-30-oxime on leukemic K562 and liver HuH7 cells were identified. Structure–activity rela-
tionships showed the importance of keeping one of the lactam NH in an unsubstituted state. Substitution
of the other lactam NH with aryl or arylalkyl side chains retained or improved activity in most instances.
An intact exocyclic double bond was also essential, possibly to maintain planarity and rigidity of the iso-
indigo scaffold. None of the compounds were found to inhibit CDK2 in an in vitro assay, in spite of reports
linking the antiproliferative activities of meisoindigo and other isoindigos to CDK2 inhibition. Hence,
these functionalized isoindigos disrupted cell growth and proliferation by other mechanistic pathways
that did not involve CDK2 inhibition.
Keywords:
Isoindigo
Antiproliferative activity
CDK2 inhibition
SAR
Molecular modeling
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
widely investigated as an anti-cancer lead, isoindigos have at-
tracted less attention.
The indigoids are a group of isomeric bisindoles that have re-
cently emerged as a promising scaffold for anticancer activity.
Interest in the indigoids was kindled when indirubin was identified
as the active ingredient of a traditional Chinese recipe (Danggui
Longhui Wan) that was used for the treatment of chronic myelog-
enous leukemia (CML).1,2 The anti-tumor activity of indirubin and
some of its analogs (notably indirubin-30-oximes) were subse-
quently attributed to the inhibition of cyclin-dependant kinases
(CDK).3 Co-crystallized structures of CDK2 and various indirubin
analogs showed that these compounds interacted with the ATP
binding site via hydrogen bonds to key amino acids in the hinge re-
gion.3,4 Unfortunately, indirubin was not considered a good candi-
date for drug development because its poor water solubility would
adversely affect oral bioavailability. Moreover, there were reports
of gastrointestinal side effects associated with its use.2 Hence,
more drug-like substitutes of indirubin were sought and some
promising candidates identified were the functionalized indiru-
Meisoindigo (1-methylisoindigo) is used as an indirubin substi-
tute for the treatment of CML in China.12 It has a multi-targeting
profile that included inhibition of DNA biosynthesis and microtu-
bule assembly in tumor cells,10 induction of leukemic cell differen-
tiation and c-myb gene downregulation,11 and inhibition of
angiogenesis.12 NaturaTM is an example of an isoindigo that is deriv-
atized with a sugar moiety. Its anti-cancer agent activity was
attributed to inhibition of several cyclin dependent kinases (CDK
2, 4, and 6).13 The presence of the protected sugar moiety was pro-
H
N
O
H
N
HO
3'
O
N
O
NH
O
N
1
bin-30-oximes,5–7 meisoindigo,2,8–12 and NaturaTM (1-b-
D-triacetyl-
5
O
O
xylopyranosylisoindigo)13 (Fig. 1). Meisoindigo and NaturaTM are
derivatives of isoindigo (3,30-bisindole) which is a regioisomer of
indirubin (3,20-bisindole). Unlike the indirubin scaffold which is
OAc
R
N
1
N
6
H
CH3
7
OAc
AcO
R = H or 5-Br
Natura™
Indirubin-3'-oximes
Meisoindigo
* Corresponding author. Tel.: +65 6516 2654; fax: +65 6779 1554.
E-mail address: phagoml@nus.edu.sg (M.-L. Go).
Figure 1. Structures of indirubin 30-oximes, meisoindigo and NaturaTM
.
0968-0896/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2009.09.008

X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
7563
posed to enhance the interaction with the ATP binding site of the
kinase, since isoindigo was essentially inactive as a CDK inhibi-
tor.14 When several 1-glycosylisoindigos were evaluated as CDK
inhibitors, it was found that only compounds with O-benzyl groups
on the sugar hydroxyl (OH) residues inhibited CDK while most
entities with O-acetyl or unprotected OH groups on the sugar moi-
ety were inactive.14 Notwithstanding their CDK inhibitory activi-
ties, the O-benzyl 1-glycosylisoindigos had only weak
antiproliferative activities,15 possibly due to their poor aqueous
solubilities which would affect cellular penetration.14 As these
compounds had molecular weights that exceeded 500 Da and lipo-
philicities (evaluated by C log P) that were greater than 5, their
poor permeabilities are anticipated from property profiling rules
like the Lipinski’s rule of five.16 The same would be true for Natu-
raTM which was also poorly soluble in water.13 Thus, designing
pharmacologically active isoindigo derivatives with a desirable
solubility–lipophilicity balance that would translate to an accept-
able level of cellular permeability poses a formidable challenge.
In this report, we describe our efforts to address this problem by
synthesizing a series of isoindigos that have functionalized arylal-
kyl or alkyl substituents on the ring nitrogen. To maintain a phar-
maceutically friendly profile, efforts were made to restrict the size
and lipophilicity of the synthesized compounds. The compounds
were evaluated for effects on the growth of several human cancer
cell lines. As CDK2 inhibition has been widely associated with the
growth inhibitory properties of isoindigos,13–15 we carried out in
silico docking to assess their potential as CDK2 inhibitors and fol-
lowed this up with experimental determination of CDK2 inhibitory
activity.
methyl in meisoindigo (2) with arylalkyl substituents of varying al-
kyl chain lengths (3–6) and introducing different substituents to
the phenyl ring of the phenethyl side chain (7–11). Analogs of
meisoindigo with an additional N-methyl group (12) and a reduced
exocyclic double bond (13) were also targeted for synthesis. In
addition, indirubins (14, 15) were included for comparison to cor-
responding isoindigos, as well as indirubin 30-oximes (16, 17)
which are known CDK inhibitors.3 In keeping with commonly em-
ployed property profiling rules,16,17 these compounds had molecu-
lar weights (6500), C log P values (65), total number of H bond
donors and acceptors (612) and rotatable bonds (610) that were
within threshold values.
2.2. Synthesis of test compounds
The isoindigos 1–5, 7, 8 were synthesized by an acid-catalyzed
aldol condensation of an isatin (or N-substituted isatin) with an
oxindole under microwave-assisted conditions (200 °C, 4 bar).
The microwave-assisted method gave yields that were comparable
to those obtained by refluxing20,21 but had shorter reaction times.
Except for 1-phenylisatin which was purchased, the other 1-substi-
tuted isatins were synthesized by reacting isatin with the appro-
priate alkyl halide in the presence of potassium carbonate in
anhydrous DMF with microwave heating22 (Scheme 1). An alterna-
tive approach was to react isoindigo with the alkyl halide under
similar conditions to give the 1-substituted isoindigo. Isoindigos
6, 9–11 were prepared in this way (Scheme 2). In the case of 9,
the phenolic hydroxyl group in 4-(2-bromoethyl)phenol was pro-
tected with dihydropyran, reacted with isoindigo and subse-
quently deprotected. Compound 11 was obtained by first
reducing 2-bromoethyl-4-nitrobenzene to give the aniline, fol-
lowed by acetylation of the amino function to give the correspond-
ing amide and then reaction with isoindigo.
2. Results
1,10-Dimethylisoindigo (12) was prepared by reacting isoindigo
with methyl iodide (ratio of 1:2). 3,30-Dihydroisoindigo (13) was
obtained as a mixture of isomers by catalytic reduction of isoindi-
go. The indirubins (14, 15) were prepared by reacting isatin or 1-
methylisatin with 3-acetoxyindole.4 The indirubin oximes (16,
17) were purchased. Compounds used for biological evaluation
2.1. Design of functionalized isoindigos
The structures of the functionalized isoindigos investigated in
this study are given in Table 1. As the compounds were designed
to be structural analogs of meisoindigo, modifications were re-
stricted to the N-substituent. These were the replacement of N-
Table 1
Structures of synthesized compounds and their C log P values
H3C
H
1'
H
N
N
N
O
O
R1
O
6
3'
X
H
H
3'
3
O
O
O
N
N
N
H
N
1
N
R
1
H
R
O
CH3
14-17
1-11
13
12
Compd
R
C log Pa
Compd
R
X
R1
C log Pa
1
2
3
4
5
6
7
8
H
–CH3
–C6H5
–CH2 C6H5
–(CH2)2 C6H5
–(CH2)3 C6H5
–(CH2)2 C6H4 (p-CH3)
–(CH2)2 C6H4 (p-OCH3)
–(CH2)2 C6H4 (p-OH)
–(CH2)2 C6H4 (p-NO2)
–(CH2)2 C6H4 (p-NHCOCH3)
2.0
2.7
2.9
4.0
4.8
5.2
5.3
4.7
3.2
4.6
3.8
12
13
14
15
16
17
—
—
H
–CH3
H
—
—
H
H
NOH
NOH
—
—
H
H
H
Br
2.7
3.4
2.6
3.2
2.4
3.5
H
9
10
11
a
C log P values were determined on ChemDraw Ultra 10.0, CambridgeSoft, Cambridge, MA.

7564
X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
H
N
O
O
O
(i)
(ii)
O
O
O
N
N
N
H
R1
R1
R1 =
CH3 - ( ),
R1 =
CH3 - (2a),
2
4
4a
C6H5CH2- ( ),
C6H5CH2- ( ),
C6H4(CH2)2 -(5),
pCH3C6H4(CH2)2 -( ),
8
pCH3OC6H4(CH2)2-( )
C6H4(CH2)2 -(5a),
7
pCH3C6H4(CH2)2 -(7a),
8a
pCH3OC6H4(CH2)2-(
)
Scheme 1. Reagents and conditions: (i) alkyl halide, K2CO3, anhydrous DMF, MW, 150 °C, 5–15 min; (ii) 2-oxindole, CH3COOH, HCl, MW, 200 °C, 15–30 min.
H
N
H
N
O
O
O
(i)
(ii)
O
O
O
N
H
N
H
N
R1
R1=
C6H5(CH2)3 - (6),
pOHC6H4(CH2)2 -(9),
pNO2C6H4(CH2)2 -(10),
pCH3CONHC6H4(CH2)2-(11)
Scheme 2. Reagents and conditions: (i) 2-oxindole, CH3COOH, HCl, MW, 200 °C, 15–30 min; (ii) alkyl bromide, K2CO3, anhydrous DMF, MW, 150 °C, 5–15 min.
were at least 95% pure as determined by combustion analyses or
reverse phase HPLC (Supplementary data).
HCT116, hepatoma HuH7) and normal lung fibroblasts (IMR-90).
The normal cell line was included to indicate the selective activity
of the compounds. The antiproliferative IC50 values of 1–17 are
listed in Table 2.
2.3. Antiproliferative activity of test compounds
Although the compounds were tested on different cancer cell
lines, several common features were observed in their antiprolifer-
ative activities. Notably, the test compounds were more active
The antiproliferative activities of the test compounds were
determined by the microculture tetrazolium assay using MTT as
the indicator dye.23,24 The compounds were tested on several hu-
man cancer lines (leukemic cell lines K562 and HL60, colon
against one leukemic cell line (K562: IC50 1.4–8.5
lM) than the
other (HL60: IC50 3.8–16.6 M). The hepatoma cells HuH7 were
l
Table 2
Anti-proliferative IC50 values of target compounds on human cancer (K562, HL60, HCT116, HuH7) and normal (IMR-90) cells
a
Compd
MTT Assay IC50
HCT116
(lM)
K562
HL60
HuH7
IMR-90
1
2
3
4
5
6
7
8
6.7 (5.6–7.4)
6.7 (5.7–7.8)
8.5 (8.2–8.9)
>30b
7.3 (6.7–7.9)
1.8 (1.5–2.2)
>30
1.4 (1.1–1.7)
>30
>30
16.6 (13.5–20.5)
12.2 (11.3–13.1)
10.4 (8.1–13.2)
>30b
11.9 (10.8–13.0)
3.8 (3.5–4.2)
>30
12.0 (11.4–12.7)
>30
>30
>30
28.9 (26.5–31.6)
18.3 (16.6–20.1)
13.2 (12.5–14.1)
9.0 (8.4–9.7)
12.0 (10.3–13.9)
8.2 (7.5–9.0)
19.0 (16.1–22.3)
7.5 (6.8–8.4)
19.9 (16.5–24.0)
13.3 (11.8–15.0)
13.3 (11.5–15.3)
>20b
16.3 (14.0–19.1)
7.7 (6.8–8.6)
6.0 (5.6–6.6)
4.9 (4.4–5.5)
6.8 (6.1–7.6)
3.9 (3.1–4.8)
>30b
15.4 (11.5–20.6)
15.4 (13.8–17.2)
10.9 (9.0–13.1)
12.2 (11.5–13.0)
7.7 (6.3–9.4)
>30
11.6 (10.8–12.4)
>30
>30
5.4 (4.9–6.0)
9
>30b
10
11
12
13
14
15
16
17
7.5 (4.5–12.3)
19.5 (18.0–21.1)
>20b
>30
>30
>30
>20b
>20b
>20b
>30
>30
>30
>30
>30
>30b
>30b
>30b
>30b
>30b
>20b
>20b
>20b
>20b
>20b
>30
1.3 (1.2–1.5)
15.7 (11.2–21.6)
5.4 (4.8–5.9)
12.0 (10.2–14.2)
5.2 (4.9–5.7)
—
18.2 (16.8–19.7)
1.9 (1.6–2.2)
6.2 (5.8–6.7)
a
Concentration required to reduce viability by 50% compared to control untreated cells after incubation. Results of at least 4 independent determinations. Ninety five
percent confidence limits are given in brackets.
b
Could not be determined due to poor solubility of compound in vehicle. Compounds had less than 50% decrease in cell viability at the highest concentration tested.
Precipitation of the compound was observed beyond this concentration.

X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
7565
also more susceptible to the isoindigos than the colon HCT116 cells
as seen from the larger number of compounds with determinable
IC50 values against HuH7 (1–11) than HCT116 (2–6, 8). In spite of
these differences, we noted that some compounds like 6 and 8
were consistently more potent than others across different cancer
cell lines while poor activity was always found for certain mem-
bers like 10–15.
Some general structure–activity trends could be deduced from
the antiproliferative activities. First, meisoindigo (2) was almost
as potent as isoindigo (1) on leukemic cells but it was significantly
more active than isoindigo on the solid tumor cell lines (IC50 HCT116
of H bonding between the docked compound and selected residues
on the apo-protein using the Protein Ligand Interaction Fingerprint
(PLIF) module in Molecular Operating Environment (MOE, Version
2008.10). The residues in the hinge region (Glu 81 and Leu 83) of
CDK2 were selected as these were widely reported to be involved
in H bonding (Glu 81 as H bond donor, Leu 83 as both H bond do-
nor and acceptor) with CDK2 inhibitors.19 Three scores, one for
each H bond interaction, were calculated for every compound, with
each score providing a measure of the probability of H bonding at
the stated residue. Summation of these scores gave the ‘HingeSum’
which reflected the total probability of H bonding at the 3 sites. As
each compound docked with different poses on the apo-protein,
only the docked pose with the highest HingeSum was evaluated
(Table 3).
15.4
lM, IC50
18.3 lM as compared to >30 lM and 28.9 lM
HuH7
for 1). Meisoindigo and isoindigo were also more potent than their
indirubin counterparts 14 and 15, particularly against leukemic
cells. Second, introducing another N-methyl group to meisoindigo
(to give 12) or reducing the exocyclic double bond (to give 13) sig-
nificantly reduced anti-proliferative activity. Clearly, activity de-
pended on the presence of at least one lactam NH and an intact
exocyclic double bond on the isoindigo scaffold. Third, replacing
the methyl group of meisoindigo with phenyl (3), benzyl (4), phen-
ethyl (5) or phenpropyl (6) side chain generally caused small in-
creases in activity except for 4 (inactive against leukemic cell
lines) and 6 (2–3 times more potent than meisoindigo). The anti-
proliferative activity of 6 was almost comparable to that of the
indirubin 30-oxime 17 whose outstanding activity had been re-
ported by others.7 Lengthening the side chain from methyl (2) to
phenpropyl (6) caused a progressive increase in C log P (Table 1)
and it may be that 6 owed its improved activity to its greater lipo-
philic character. Lastly, introducing substituents of different polar-
ities and electronic effects to the phenyl ring of the phenethyl
analog 5 did not yield compounds with improved activities. A nota-
ble exception was the p-methoxy analog 8 which was more active
The indirubin-30-oximes (16, 17) had higher docking scores
than most compounds, indicating strong H bonding with the active
site residues which was compatible with their potent CDK2 inhib-
itory activities.3,4 Interestingly, several isoindigos (3, 5, 8) had
scores that were in the range of the 30-oximes (16, 17) and this
raised the possibility that CDK inhibition could have contributed
to their effects on cell growth and proliferation. Hence we investi-
gated the CDK2 inhibitory activities of these compounds.
2.5. Inhibition of CDK2
The CDK2 inhibitory activities of 1–17 were evaluated by the
immobilized metal Ion affinity-based fluorescence polarization
(IMAP) assay.26 Briefly, the assay measures the change in fluores-
cence polarization of a peptide substrate of CDK2 on phosphoryla-
tion. The phosphorylated substrate binds to the IMAP binding
reagent and experiences a loss in molecular motion that translates
to an increase in florescence polarization. In the presence of a CDK2
inhibitor, phosphorylation of the substrate was interrupted leading
to a decrease in fluorescence polarization. The compounds were
than 5 on the leukemic K562 cells (IC50 K562 1.4
7.3 M for 5) and hepatoma HuH7 cells (IC50 K562 7.5
to 12.0 M for 5). The good activity of 8 did not appear to be linked
to its lipophilic character (C log P of 5 and 8 were comparable), its
l
M compared to
l
lM compared
l
evaluated at a fixed concentration of 10 lM.
As seen from Figure 2, only the indirubin 30-oximes 16 and 17
significantly inhibited CDK2 (70–80% inhibition), in keeping with
their reported CDK2 inhibitory activities.3 None of the isoindigos,
including meisoindigo (2) and those with good docking scores (3,
5, 8) inhibited CDK2 to any significant extent. To confirm this find-
ing, meisoindigo (2) and the phenpropyl analog 6 were screened
for inhibitory activities against CDK2 and other kinases by a com-
electron donating effect (rp of OCH3 was between that of CH3 and
OH present in the less active analogs 7 and 9, respectively)25 or H
bonding capability (NO2 and OCH3 were H bond acceptors and not
H bond donors, but the NO2 analog 9 was inactive).
The test compounds also affected the proliferation of the non-
cancerous fibroblasts IMR-90, often with lower IC50 values than
those obtained against the cancer cell lines. Notably, the more
promising isoindigos 6 and 8 were modestly selective against
K562 cells by a 2–4-fold difference as seen from the ratios of their
IC50 values (IC50 IMR-90/IC50 K562). These levels of selectivities com-
pared favorably to those obtained for meisoindigo (2) and the
indirubin 30-oxime 17 (IC50 IMR-90/IC50 K562 < 2).
Table 3
Docking scores of test compounds
Compound
HingeSuma
GLU81 as
H donor
LEU83 as
H donor
LEU83 as
H acceptor
NaturaTM
1
2
3
4
5
6
7
0.35
0.34
0.33
0.60
0.47
0.63
0.38
0.44
0.64
0.45
0.47
0.46
0.22
0.79
0.55
0.63
0.72
0.13
0.14
0.20
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.34
0.00
0.33
0.00
0.16
0.00
0.00
0.00
0.00
0.22
0.21
0.34
0.24
0.26
0.38
0.21
0.19
0.00
0.00
0.04
0.06
0.05
0.34
0.21
0.20
0.13
0.38
0.27
0.28
0.15
0.19
0.26
0.24
0.29
0.12
0.22
0.42
0.49
0.43
0.38
2.4. Molecular docking
Meisoindigo and NaturaTM were reported to suppress cyclin D
mediated CDK4/6 activities of prostate cancer cells (LNCaP) at con-
centrations of 5 and 15
l
M.13 Almost complete inhibition was re-
ported at 15 M while at 5 l
l
M, NaturaTM was a slightly stronger
inhibitor than meisoindigo (60% vs 40% inhibition).13 Several O-
benzyl 1-glycosylisoindigos inhibited CDK2 at submicromolar
IC50 values and molecular modeling of one of the potent analogs
revealed two H bonds with the critical hinge region amino acids
(Glu 81, Leu 83) and hydrophobic interactions between the benzyl
groups on the sugar moiety and a hydrophobic pocket in the CDK2
active site.14 These findings prompted us to investigate the binding
potential of our compounds (Table 1) to the CDK2 active site.
Docking was conducted using a co-crystallized complex of
CDK2, cyclin A and adenosine-50-triphosphate (ATP) (PDB code:
1FIN).18 The interaction was scored by evaluating the probability
8
9
10
11
12
14
15
16
17
a
HingeSum is summation of hydrogen bonding probability scores with Glu81 as
H donor, Leu83 as H bond donor, and Leu83 as H bond acceptor. The latter were
calculated with the PLIF module in MOE.

7566
X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
100
90
80
70
60
50
40
30
20
10
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17
Compound
Figure 2. Percentage inhibition of CDK2 by test compounds (10
l
M) in the IMAP assay. Vertical bars represent SD for n = 3 determinations.
mercial service provider (KINOMEScanTM-ScanEdgeTM, Ambit Biosci-
ences, San Diego, CA). The compounds were tested at 10 M on a
ing from the reduction of the exocyclic double bond. Replacing 1-
l
methyl of meisoindigo with longer and bulkier side chains as those
found in 3, 5, and 6 generally maintained activities. Of the various
substituents that were introduced at the p-position of the 1-phen-
ethyl side chain, only the p-methoxy group (8) gave rise to good
activity. Both 6 and 8 were more lipophilic than meisoindigo as as-
sessed by their C log P values but more isoindigos with varied lipo-
philicities must be evaluated before the role of lipophilicity in
determining antiproliferative activity is established.
binding assay that quantified the ability of the compound to com-
pete with an immobilized active site directed ligand.27 The results
confirmed that meisoindigo and 6 did not inhibit CDK2 at 10 lM.
In fact, the 2 compounds failed to inhibit any kinase by more than
70% at this concentration. The kinase that was most susceptible to
meisoindigo was CDK7 (56% inhibition). More kinases were inhib-
ited by 6 (at 40–52% inhibition) and of these, 2/3rd were tyrosine
kinases and none were serine–threonine kinases. The results of
the screen for meisoindigo and 6 are given in Supplementary data.
In spite of reports linking meisoindigo with the inhibition of
CDK2,13 none of the isoindigos inhibited CDK2 on the IMAP assay.
However, as anticipated, the indirubin 30-oximes (16, 17) had sig-
nificant CDK2 inhibitory activities on the same assay. It is notable
that meisoindigo had negligible effects on all the kinases in the
3. Discussion
KINOMEScanTM except CDK7, which it inhibited by 56% at 10
lM.
There are limited reports on the anti-cancer activity of isoindigo
derivatives and those available to date centred on meisoindigo2,8–
Interestingly, CDK7 is known to regulate the activity of several
CDKs (including CDK2) through phosphorylation of a threonine
residue in the activation segment.29 However, CDK2 was reported
to be capable of autophosphorylation at Thr60 of its activation seg-
ment30 which may mean that if CDK7 was inhibited, CDK2 would
still retain its activity. Unlike meisoindigo, the phenpropyl analog
6 did not inhibit CDK7 but it affected a larger number (9) of ki-
nases, most of which were tyrosine kinases. The antiproliferative
activity of meisoindigo may arise from its multiple effects on cellu-
lar processes10,11,13 and this may also be true for the functionalized
isoindigos.
The lack of agreement between the in silico docking experi-
ments and the experimentally determined CDK2 inhibitory activi-
ties raised the question as why this discrepancy was observed. A
likely explanation may be that molecular docking was carried
out with a CDK2 protein (1FIN) that was co-crystallized with ATP
instead of an isoindigo derivative. A recent study showed that
when compounds were docked onto protein structures derived
from complexes co-crystallized with a different ligand (‘non-native
docking’), the results would be less reliable.28 The problem would
be further compounded if a rigid docking procedure was used, as in
the present case. Thus, unless a flexible docking protocol was ap-
plied to a co-crystallized isoindigo-CDK2 complex of good resolu-
tion (not available as yet), the docking results of functionalized
isoindigos should be interpreted with caution. Subsequently, we
12
and isoindigos that were derivatized with sugar residues like
NaturaTM13 and various 1-glycosylated isoindigos.14,15 A major con-
cern with these isoindigo derivatives was that they may not be suf-
ficiently drug-like for clinical advancement. In this report, we have
synthesized a series of functionalized isoindigos that had more
desirable drug-like/lead-like features as defined by commonly
used structure profiling guidelines. We found that fairly simple
modifications of meisoindigo resulted in compounds that had
greater antiproliferative activities against a panel of human cancer
cell lines. Among the more promising analogs was 1-phenpropylis-
oindigo (6) which was more potent than meisoindigo and compa-
rable to 6-bromoindirubin-30-oxime (17) on the cancer cell lines,
notably the leukemic K562 and hepatoma HuH7 cells. Compound
6 was also selectively more potent against leukemic K562 cells
than normal fibroblast cells, unlike meisoindigo and the indiru-
bin-30-oxime (17). Another promising candidate was 8 which had
a 1-(p-methoxy-phenylethyl) side chain. Like 6, this compound
was modestly selective for the leukemic K562 cells compared to
normal fibroblasts.
In spite of the small number of compounds investigated, some
useful structure–activity correlations were evident. The inactivity
of 1,10-dimethylisoindigo (12) highlighted the importance of keep-
ing at least one free lactam NH. Another compound 13 may owe its
lack of activity to the lost of planarity and increased flexibility aris-

X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
7567
investigated the interaction of meisoindigo and indirubin-30-oxime
(16) with CDK2 by molecular dynamics (MD) simulation (Supplemen-
tary data) and found that meisoindigo was held by weak and transient
H bonds to the CDK2 binding pocket. Its binding energy was consider-
ably smaller than that of oxime 16 which established stronger interac-
tions with the enzyme consistent with its strong CDK2 inhibitory
activity.3,4 Thus, the more rigorous MD simulation reflected the weak
CDK2 inhibitory activity of meisoindigo more accurately than molecu-
lar docking.
(2a) or dichloromethane (8a). The organic layer was dried using
anhydrous Na2SO4 and concentrated to give a residue which was
recrystallization with ethanol to give the substituted isatin. For
4a, 5a, and 7a, the suspension was shaken vigorously, cooled down
to 5–10 °C for approximately 1 h and then filtered under reduced
pressure. The residue was washed with ice-cold deionized water
and recrystallized with ethanol to give the desired product. Yields
and 1HNMR spectroscopic data of 2a, 4a, 5a, 7a, and 8a are given in
Supplementary data.
In conclusion, we have identified several functionalized isoindi-
gos with antiproliferative activities against a panel of cancer cell
lines. The most promising members (6 and 8) were more potent
than meisoindigo and comparable to 6-bromoindirubin-30-oxime
(16) on leukemic K562 cells and hepatoma HuH7 cells but had less
growth inhibitory activity against normal lung fibroblast cells. In
spite of docking simulations that supported interactions of several
isoindigos with CDK2, this was not borne out experimentally and
none of the isoindigos (including meisoindigo) inhibited CDK2 to
any significant extent. The antiproliferative activity of these com-
pounds may arise from disruption of other pathways or enzymes
involved in cell growth processes. Current reports on meisoindigo
suggest that it has multiple targets in cancer cells10–12 and the ana-
logs 6 and 8 may share a similar profile. If this were so, activity
against several targets combined with their pharmaceutically
friendly features would make functionalized isoindigos promising
lead structures for further investigation as anticancer agents.
4.3. Isoindigo (1)
Isatin (2 g, 13.6 mmol) and 2-oxindole (1.81 g, 13.6 mmol) were
dissolved in 15 ml glacial acetic acid in a 20 ml microwave vessel
and three drops of concd hydrochloric acid was added. The seal
vessel was subjected to microwave irradiation at 200 °C for
30 min. The resultant crude mixture was concentrated in vacuo.
Toluene (30 ml) was added to aid in the removal of glacial acetic
acid. Ethyl acetate (30 ml) was added to the residue, the suspen-
sion was stirred vigorously and allowed to stand at 4 °C, 1 h. The
suspension was filtered under reduced pressure, rinsed with
chilled ethyl acetate and dried in the oven. Compound 1 was ob-
tained as a dark violet powder. Yield: 91.2%; mp: >350 °C. 1H
NMR (300 MHz, DMSO-d6): d = 10.87 (s, 2H), 9.04 (d, J = 8.1 Hz,
2H), 7.33 (t, J = 7.5 Hz, 2H), 6.95 (t, J = 7.8 Hz, 2H), 6.84 (d,
J = 7.8 Hz, 2H); 13C NMR (75 MHz, DMSO-d6): d = 170.0 (CO–NH),
145.1, 134.4, 133.7, 130.3, 122.7, 122.2, 110.6; MS(APCI) m/z:
263.4 [(M+H+) calculated for C16H11N2O2, 263.3]. Anal. Calcd for
C16H10N2O2 (262.26): C, 73.27; H, 3.84. Found: C, 73.29; H, 4.01.
4. Experimental
4.1. General details for chemical synthesis
4.4. Meisoindigo (2)
All reagents were obtained from commercial suppliers and used
without further purification. Melting points were determined on a
Gallenkemp melting point apparatus and reported as uncorrected
values. Reactions were routinely monitored by thin layer chroma-
tography (TLC) using pre-coated plates (Silica Gel 60, F254, Merck)
and visualized with UV light. Flash column chromatography was
carried out using Silica Gel 60 (230–400 mesh, Merck). 1H and
13C NMR spectra were measured on Brucker ACF (DPX-300) (1H
at 300 MHz, 13C at 75 MHz) magnetic resonance spectrometer. 1H
chemical shifts were reported in ppm using TMS as internal stan-
dard. Coupling constants (J) were reported in hertz (Hz). Proton
decoupled 13C NMR spectra were reported in ppm (d) relative to
residual CHCl3 (d 77.0) and DMSO (39.5). Nominal mass spectra
were captured on a LcQ Finnigan MAT mass spectrometer with
chemical ionization (APCI) as probe and m/z values for the molec-
ular ion were reported. Microwave assisted synthesis was carried
out using a Biotage InitiatorTM microwave synthesizer. Purity of fi-
nal compounds were verified by either combustion analysis (C,
H) (Elementar Varlo Micro Cube) or by reverse phase HPLC on
two different eluting systems (isocratic mode) (Supplementary
data).
1-Methyl-isatin (2a) (150 mg, 0.9 mmol) and 2-oxindole
(149 mg, 1 mmol) were dissolved in 3 ml glacial acetic acid in a
5 ml microwave vessel and one drop of concentrated HCl was
added. The sealed vessel was subjected to microwave irradiation
at 200 °C for 30 min. The resultant crude mixture was concentrated
in vacuo and purified using flash chromatography to give the de-
sired product as dark violet powder. Yield: 40.6%; mp: 229 °C. 1H
NMR (300 MHz, DMSO-d6): d = 10.92 (s, 1H), 9.10 (dd, J1 = 7.8,
J2 = 2.1 Hz, 2H), 7.44 (t, J = 7.5 Hz, 1H), 7.36 (t, J = 7.5 Hz, 1H),
7.05 (t, J = 8.4 Hz, 1H), 7.03 (d, J = 8.1, 1H), 6.98 (t, J = 7.8 Hz, 1H),
6.85 (d, J = 7.8 Hz, 1H), 3.23 (s, 3H); 13C NMR (75 MHz, DMSO-
d6): d = 169.9 (CO–NH), 168.3 (CO–NH), 146.0, 145.3, 134.8,
133.9, 133.6, 133.2, 130.4, 130.0, 122.8, 122.6, 122.2, 121.9,
110.7, 109.5; MS(APCI) m/z: 277.7 [(M+H+) calculated for
C17H13N2O2, 277.3]. Anal. Calcd for C17H12N2O2 (276.29): C,
73.90; H, 4.38. Found: C, 73.52; H, 4.43. Purity of 2 was verified
by HPLC to be more than 95% (Supplementary data).
4.5. 1-Phenyl-isoindigo (3)
1-Phenyl-isoindigo (3) was synthesized by a similar method as
2 starting from 1 mmol quantities of 1-phenyl-isatin and 2-oxin-
dole. Compound 3 was obtained as a dark red powder. Yield:
62.0%; mp: 273–274 °C. 1H NMR (300 MHz, DMSO-d6): d = 10.95
(s, 1H), 9.17 (d, J = 7.8 Hz, 1H), 8.97 (d, J = 8.1 Hz, 1H), 7.64–7.59
(m,2H), 7.53–7.50 (m, 3H), 7.38–7.33 (m, 2H), 7.10 (t, J = 7.8 Hz,
1H), 6.95 (t, J = 7.8 Hz, 1H), 6.86 (d, J = 7.8 Hz, 1H), 6.68 (d,
J = 8.1 Hz, 1H); 13C NMR (75 MHz, DMSO-d6): d = 168.8 (CO–NH),
166.7 (CO–NH), 144.8, 144.4, 134.4, 133.9, 133.1, 132.6, 131.9,
129.6, 129.5, 129.2, 128.4, 127.3, 122.3, 121.6, 121.2, 121.0,
109.7, 108.8; MS(APCI) m/z: 339.5 [(M+H+) calculated for
C22H15N2O2, 339.4]. Anal. Calcd for C22H14N2O2 (338.35): C,
78.09; H, 4.17. Found: C, 78.07; H, 4.14.
4.2. General method for the preparation of 1-methylisatin (2a),
1-benzylisatin (4a), 1-phenethylisatin (5a), 1-(p-methyl-
phenethyl) isatin (7a) and 1-(p-methoxy-phenethyl)-isatin (8a)
Anhydrous K2CO3 (400 mg) was added to a solution of the isatin
(2 mmol) dissolved in anhydrous DMF (1.5 ml) in a 5 ml micro-
wave vessel. The vessel was sealed and the contents stirred at
room temperature (27 °C) for 15 min, after which the alkyl halide
(in excess) was injected into the vessel, stirred for another 5 min
and then heated under pressure in a microwave reactor for
150 °C, 15 min. Deionized water (5 ml) was added to quench the
reaction and the suspension was extracted with ethyl acetate

7568
X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
4.6. 1-Benzyl-isoindigo (4)
NH), 168.0 (CO–NH), 145.2, 145.1, 136.4, 136.3, 134.9, 134.0,
133.7, 133.1, 130.3, 130.0, 129.8, 122.7, 122.6, 122.2, 121.9,
110.7, 109.7, 42.0, 33.5, 21.7; MS(APCI) m/z: 381.7 [(M+H+) calcu-
lated for C25H21N2O2, 381.4]. Anal. Calcd for C25H20N2O2 (380.43):
C, 78.93; H, 5.30. Found: C, 78.85; H, 5.29.
1-Benzyl-isoindigo (4) was synthesized by the method de-
scribed for
2 starting from 1-benzyl-isatin (4a) (0.84 mmol,
200 mg) and 2-oxindole (1.1 mmol, 135 mg, 1.2 equiv). Compound
4 was obtained as dark red crystals. Yield: 43.1%; mp: 238–241 °C.
1H NMR (300 MHz, DMSO-d6): d = 10.94 (s, 1H), 9.10 (dd,
J1 = 7.8 Hz, J2 = 3.3 Hz, 2H), 7.36–7.27 (m, 7H), 7.06–6.95 (m, 3H),
6.86 (d, J = 7.8 Hz, 1H), 5.02 (s, 2H); 13C NMR (75 MHz, DMSO-
d6): d = 169.9 (CO–NH), 168.4 (CO–NH), 145.4, 145.0, 137.3,
135.4, 134.1, 133.5, 132.9, 130.6, 130.2, 129.8, 128.5, 128.3,
123.0, 122.7, 122.3, 122.0, 110.7, 110.0, 43.8; MS(APCI) m/z:
353.7 [(M+H+) calculated for C23H17N2O2, 353.4]. Anal. Calcd for
C23H16N2O2 (352.28): C, 78.39; H, 4.58. Found: C, 78.27; H, 4.63.
4.10. 1-(p-Methoxy-phenethyl)-isoindigo (8)
1-(p-Methoxy-phenethyl)-isoindigo (8) was synthesized by a
similar method as 2 starting from 1-(p-methoxy-phenethyl)-isatin
(8a) (0.71 mmol, 200 mg) and 2-oxindole (0.85 mmol, 114 mg,
1.2 equiv). Yield: 34%; mp: 216–217 °C. 1H NMR (300 MHz,
DMSO-d6): d = 10.89 (s, 1H), 9.09 (d, J = 8.1 Hz, 1H), 8.97 (d,
J = 8.1 Hz, 1H), 7.41 (t, J = 7.5 Hz, 1H), 7.35 (t, J = 7.5 Hz, 1H), 7.20
(d, J = 8.4 Hz, 2H), 7.07 (d, J = 7.8 Hz, 1H), 7.03 (t, J = 7.8 Hz, 1H),
6.96 (t, J = 7.8 Hz, 1H,), 6.83–6.82 (m, 3H), 3.95 (t, J = 7.2 Hz, 2H,),
3.69 (s, 3H), 2.87 (t, J = 7.5 Hz, 2,); 13C NMR (75 MHz, DMSO-d6):
d = 169.9 (CO–NH), 168.0 (CO–NH), 158.9, 145.2, 145.1, 134.9,
133.9, 133.6, 133.2, 131.2, 130.9, 130.4, 130.1, 122.7, 122.6,
122.2, 121.9, 114.9, 110.7, 109.7, 56.0, 42.2, 33.1; MS(APCI) m/z:
397.4 [(M+H+) calculated for C25H21N2O3, 397.4]. Anal. Calcd for
C25H20N2O3 (396.43): C, 75.74; H, 5.08. Found: C, 75.48; H, 5.04.
4.7. 1-(Phenethyl)-isoindigo (5)
1-(Phenethyl)-isoindigo (5) was synthesized by the method de-
scribed for 2 starting from 1-(phenethyl)-isatin (5a) (0.80 mmol,
200 mg) and 2-oxindole (0.96 mmol, 127 mg, 1.2 equiv). Com-
pound 5 was obtained as dark red crystals. Yield: 49%; mp: 216–
217 °C. 1H NMR (300 MHz, DMSO-d6): d = 10.91 (s, 1H), 9.09 (d,
J = 7.8 Hz, 1H), 8.98 (d, J = 7.8 Hz, 1H), 7.40 (t, J = 7.5 Hz, 1H), 7.35
(t, J = 7.8 Hz, 1H), 7.29–7.18 (m, 5H), 7.08 (d, J = 7.8 Hz, 1H), 7.03
(t, J = 8.1 Hz, 1H), 6.96 (t, J = 7.8 Hz, 1H), 6.84 (d, J = 7.8 Hz, 1H),
4.00 (t, J = 7.5 Hz, 2H), 2.94 (t, J = 7.5 Hz, 2H); 13C NMR (75 MHz,
DMSO-d6): d = 169.9 (CO–NH), 168.1 (CO–NH), 145.2, 145.1,
139.4, 134.9, 134.0, 133.6, 133.1, 130.4, 130.1, 129.9, 129.4,
127.5, 122.7, 122.6, 122.2, 121.8, 110.7, 109.7, 41.9, 34.0; MS(APCI)
m/z 367.8 [(M+H+) calculated for C24H19N2O2, 367.4]. Anal. Calcd
for C24H18N2O2 (366.40): C, 78.67; H, 4.95. Found: C, 78.40; H, 4.95.
4.11. 1-(p-Hydroxy-phenethyl)-isoindigo (9)
p-Hydroxyphenethyl bromide (600 mg, 3 mmol), dihydro-2H-
pyran (411 ll, 4.5 mmol), and pyridinium p-toluenesulfonate
(PPTS, 75 mg, 0.3 mmol) were dissolved in 5 ml DCM and stirred
for 16 h at room temperature. The mixture was extracted with
DCM/brine, the organic fraction was dried over anhydrous sodium
sulfate, concentrated in vacuo, and the residue was purified by
flash chromatography with 20:1 hexane/ethyl acetate as eluting
solvent. The tetrahydropyranyl ether was obtained as a colorless li-
quid in 80% yield. It was then added to a previously stirred mixture
of K2CO3 (310 mg, 2.25 mmol) and 393 mg of isoindigo (390 mg,
1.5 mmol) in 2 ml anhydrous DMF contained in a 5 ml microwave
vessel, as described for 6. The mixture was irradiated by micro-
wave for 30 min at 150 °C and worked up as described for 6. The
tetrahydropyranyl protecting group was removed by refluxing
the product from column chromatography (155 mg) with 20 mg
of PPTS in 50 ml of ethanol for 3 h. The solution was concentrated
under reduced pressure and extracted using ethyl acetate/brine.
The organic fraction was dried (anhydrous Na2SO4), concentrated
in vacuo, and recrystallized using ethanol to give the desired prod-
uct. Yield: 34%; mp: 282–285 °C. 1H NMR (300 MHz, DMSO-d6):
d = 10.90 (s, 1H), 9.20 (br s, 1H), 9.08 (d, J = 8.1 Hz, 1H), 9.00 (d,
J = 7.8 Hz, 1H), 7.39 (t, J = 7.8 Hz, 1H), 7.34 (t, J = 7.5 Hz, 1H),
7.07–7.01 (m, 4H), 6.96 (t, J = 7.8 Hz, 1H), 6.84 (d, J = 7.5 Hz, 1H),
6.65 (d, J = 8.4, 2H), 3.91 (t, J = 7.2 Hz, 2H), 2.80 (t, J = 7.5 Hz, 2H);
13C NMR (75 MHz, DMSO-d6): d = 168.8 (CO–NH), 166.9 (CO–NH),
155.8, 144.14, 144.08, 133.8, 132.8, 132.5, 132.1, 129.7, 129.3,
129.0, 128.3, 121.6, 121.1, 120.8, 115.1, 109.6, 108.6, 41.2, 32.0.
MS(APCI) m/z: 383.4 [(M+H+) calculated for C24H19N2O3, 383.4].
Anal. Calcd for C24H18N2O3 (382.40): C, 75.38; H, 4.74. Found: C,
75.24; H, 4.75.
4.8. 1-(Phenpropyl)-isoindigo (6)
Anhydrous potassium carbonate (1 mmol) and isoindigo
(1 mmol) were dissolved in 1.5 ml of anhydrous DMF in a 5 ml
microwave vessel and allowed to stir at room temperature for
15 min. Phenpropyl bromide (167 ll, 1.1 mmol) was injected into
the mixture which was then subjected to microwave irradiation
for 30 min at 150 °C. The mixture was extracted with ethyl ace-
tate/brine, the organic fraction was dried (anhydrous Na2SO4), con-
centrated in vacuo and the residue purified with flash
chromatography (4:1 hexane/ethyl acetate as eluting solvent).
Compound 6 was recrystallized from ethanol. Yield: 42%; mp:
173–174 °C. 1H NMR (300 MHz, DMSO-d6): d = 10.91 (s, 1H), 9.08
(t, J = 7.5 Hz, 2H), 7.41 (t, J = 7.8 Hz, 1H), 7.35 (t, J = 7.5 Hz, 1H),
7.29–7.17 (m, 5H), 7.01 (t, J = 6.6 Hz, 2H), 6.97 (t, J = 7.8 Hz, 1H),
6.85 (d, J = 7.8 Hz, 1H), 3.80 (t, J = 6.9 Hz,2H), 2.67 (t, J = 7.5 Hz,
2H), 1.93 (p, J = 7.5 Hz, 2H); 13C NMR (75 MHz, DMSO-d6):
d = 168.9 (CO–NH), 167.1 (CO–NH), 144.2, 141.2, 133.8, 132.9,
132.6, 132.2, 129.5, 129.1, 128.3, 128.2, 125.9, 121.7, 121.6,
121.2, 120.9, 109.6, 108.5, 40.3, 32.5, 28.6; MS(APCI) m/z: 381.5
[(M+H+) calculated for C25H21N2O2, 381.4]. Anal. Calcd for
C25H20N2O2 (380.43): C, 78.93; H, 5.30. Found: C, 78.75; H, 5.25.
4.9. 1-(p-Methyl-phenethyl)-isoindigo (7)
4.12. 1-(p-Nitro-phenethyl)isoindigo (10)
1-(p-Methyl-phenethyl)-isoindigo (7) was synthesized by a
similar method as 2 starting from 1-(p-methyl-phenethyl)-isatin
(7a) (0.75 mmol, 200 mg) and 2-oxindole (0.91 mmol, 121 mg,
1.2 equiv). Compound 11 was obtained as dark red crystals. Yield:
49%; mp: 242–243 °C. 1H NMR (300 MHz, DMSO-d6): d = 10.90 (s,
1H), 9.08 (d, J = 7.8 Hz, 1H), 8.94 (d, J = 8.1 Hz, 1H), 7.41 (t,
J = 7.5 Hz, 1H), 7.35 (t, J = 7.8 Hz, 1H), 7.16–7.00 (m, 6H), 6.95 (t,
J = 7.8 Hz, 1H), 6.84 (d, J = 7.5 Hz, 1H), 3.96 (t, J = 7.2 Hz, 2H), 2.88
(t, J = 7.5 Hz, 2H); 13C NMR (75 MHz, DMSO-d6): d = 169.9 (CO–
1-(p-Nitro-phenethyl)isoindigo (10) was synthesized by the
method described for 6 starting from isoindigo (1) (5 mmol,
1310 mg) and 1-(p-nitro-phenethyl)-isatin (6 mmol, 1380 mg,
1.2 equiv). Yield: 33%; mp: 241–242 °C. 1H NMR (300 MHz,
DMSO-d6): d = 10.90 (s, 1H), 9.09 (d, J = 7.8 Hz, 1H), 8.85 (d,
J = 7.8 Hz, 1H), 8.13 (d, J = 8.4 Hz, 2H), 7.55 (d, J = 8.4 Hz, 2H),
7.41 (t, J = 7.5 Hz, 1H), 7.33 (t, J = 7.5 Hz, 1H), 7.11 (d, J = 7.8 Hz,
2H), 7.03 (t, J = 7.8 Hz, 1H), 6.90 (t, J = 7.5 Hz, 1H), 6.83 (d,

X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
7569
J = 7.5 Hz, 1H), 4.07 (t, J = 6.9 Hz, 2H), 3.10 (t, J = 6.9 Hz, 2H); 13C
NMR (75 MHz, DMSO-d6): d = 168.8 (CO–NH), 167.0 (CO–NH),
146.9, 146.2, 144.2, 143.8, 133.9, 132.9, 132.5, 131.9, 130.3,
129.1, 129.0, 123.4, 121.8, 121.5, 121.1, 120.8, 109.6, 108.7, 42.0,
32.7; MS(APCI) m/z: 412.7 [(M+H+) calculated for C24H17N3O4,
412.4]. Anal. Calcd for C24H16N3O4 (411.41): C, 70.07; H, 4.16.
Found: C, 68.64; H, 4.46. Purity of 10 was verified by HPLC to be
more than 95% (Supplementary data).
added to quench the reaction and the suspension was shaken vig-
orously and left to stand in the refrigerator at 4 °C for 1 h. The sus-
pension was filtered under reduced pressure, rinsed with excess
water and dried in the oven. Recrystallization with ethanol 12 as
dark violet powder. Yield: 95%; mp: 274–275 °C. 1H NMR
(300 MHz, DMSO-d6): d = 9.20 (d, J = 8.1 Hz, 2H), 7.37 (t, J =
7.8 Hz, 2H), 7.07 (t, J = 8.1 Hz, 2H), 6.78 (d, J = 7.8 Hz, 2H), 3.29 (s,
6H); 13C NMR (75 MHz, DMSO-d6): d = 164.8, 146.2, 133.9, 130.1,
122.9, 121.8, 109.6, 100.6; MS(APCI) m/z: 291.7 [(M+H+) calculated
for C18H15N2O2, 291.3]. Anal. Calcd for C18H14N2O2 (291.32): C,
74.47; H, 4.86. Found: C, 74.43; H, 4.87.
4.13. N-(4-(2-Bromoethyl)phenyl)acetamide
4-Nitrophenethyl bromide (1 g, 4.3 mmol) was dissolved in a
solution of 16 ml glacial acetic acid and 7.5 ml methanol. The solu-
tion was heated to 65 °C in an oil bath, and iron powder (870 mg)
was added in 2 batches, 30 min apart. The mixture was then stirred
for 2.5 h at 65 °C. It was then concentrated in vacuo, with toluene
(30 ml) added to aid in the removal of acetic acid. DCM (50 ml) was
added to the residue, followed by concd ammonia solution and
extraction with DCM/brine. The organic solvent was dried with
anhydrous sodium sulfate and concentrated to give a colorless li-
quid which was then dissolved in dry DCM (4 ml) and triethyl-
4.16. 3, 30-Dihydroisoindigo (13)
Isoindigo (1, 350 mg, 1.3 mmol) was dissolved in 20 ml of etha-
nol and 35 mg of 10% Pd/C was added. Hydrogen gas was bubble
into the solution and stirred at room temperature (27 °C) for 2 h.
The suspension was filtered and the solvent removed under re-
duced pressure to give a pale yellow powder which gave two spots
on the TLC. Compound 13 was obtained as a mixture of isomers
which was not separated. The methyl protons at 3 and 30 positions
were detected at 4.16 (major isomer) and 4.07 (minor isomer) in
the ratio of 2:1 in the 1H NMR (300 MHz). 13C NMR (75 MHz,
DMSO-d6): d = 177.2, 176.1, 143.2, 142.7, 128.19, 128.12, 127.4,
125.9, 123.3, 122.9, 121.3, 121.2, 109.4, 109.3, 45.7, 45.6; MS(APCI)
m/z 265.5 [(M+H+) calculated for C16H13N2O2, 265.3]. Anal. Calcd
for C16H12N2O2 (264.28): C, 72.72; H, 4.58. Found: C, 72.78; H,
4.61. Compound 13 was tested as a mixture of stereoisomers.
amine (600 ll, 2.2 equiv). Acetyl chloride (300 ll, 2.2 equiv) was
added to the solution, followed by refluxing for 1 h. NaOH
(0.1 M) was added to quench the reaction. The reaction mixture
was then extracted with DCM/brine, washed successively with
0.1 M NaHCO3 and 0.1 M HCl, the organic fraction was dried (anhy-
drous Na2SO4), concentrated in vacuo and the residue purified by
flash chromatography with 2:1 hexane/ethyl acetate as eluting sol-
vent. Target compound was obtained as a beige powder. Yield:
31%; mp: 136–137 °C. 1H NMR (300 MHz, CDCl3): d = 7.45 (d,
J = 8.4 Hz, 2H), 7.26 (s, 1H), 7.16 (d, J = 8.1 Hz, 2H), 3.54 (t,
J = 7.5 Hz, 2H), 3.12 (t, J = 7.5 Hz, 2H), 2.17 (s, 3H).
4.17. Determination of anti-proliferative activity by MTT assay
The cell lines K562 (human chronic myelogenous leukemia
lymphoblast), HL60 (human acute promyelocytic leukemia myelo-
blast), HCT116 (human colon carcinoma) and IMR-90 (normal hu-
man lung fibroblast) were purchased from American Type Culture
Collection (ATCC, Rockville, MD) and grown in media recom-
mended by ATCC. HuH7 (human hepatoma) cells were a gift from
Dr Ho Han Kiat of AꢀSTAR, Singapore. K562 and HL60 cells were
sub-cultured when they reached a density of 106 cells/ml and cells
in their 6–12 passages were used for the experiments. HCT116 and
IMR-90 cells were sub-cultured when they reached 90% confluency
and used within 4–10 passages for the assay.
4.14. 1-(p-Acetylamino-phenethyl)-isoindigo (11)
Isoindigo (262 mg, 1 mmol) and K2CO3 (166 mg, 1.2 mmol)
were dissolved in 2 ml of anhydrous DMF in a 5 ml microwave ves-
sel and allowed to stir at room temperature for 15 min. N-(4-(2-
Bromoethyl)phenyl)acetamide (266 mg, 1.1 mmol) was dissolved
in anhydrous DMF (1 ml) and injected into the vessel. The mixture
was subjected to microwave irradiation for 30 min at 150 °C. The
mixture was then extracted with ethyl acetate/brine, the organic
fraction was dried (anhydrous Na2SO4), concentrated in vacuo
and the residue was purified by column chromatography with
1:1 hexane/ethyl acetate as eluting solvent. Compound 11 was
recrystallized in ethanol to give red crystals. Yield 21%; mp: 278–
279 °C. 1H NMR (300 MHz, DMSO-d6): d = 10.90 (s, 1H), 9.85 (s,
1H), 9.08 (d, J = 8.1 Hz, 1H), 8.95 (d, J = 7.8 Hz, 1H), 7.46 (d,
J = 8.4 Hz, 2H), 7.41 (t, J = 8.1 Hz, 1H), 7.35 (t, J = 7.8 Hz, 1H), 7.18
(d, J = 8.4 Hz, 2H), 7.07 (d, J = 8.1 Hz, 1H), 7.02 (t, J = 7.8 Hz, 1H),
6.95 (t, J = 7.5 Hz, 1H), 6.84 (d, J = 7.8 Hz, 1H), 3.96 (t, J = 6.9 Hz,
2H), 2.87 (t, J = 6.6 Hz, 2H), 2.00 (s, 3H); 13C NMR (75 MHz,
DMSO-d6): d = 168.9 (CO–NH), 168.3 (CO–NH), 167.0 (C@O),
144.2, 137.7, 133.9, 133.0, 132.9, 132.7, 132.2, 129.3, 129.1,
121.7, 121.6, 121.2, 120.8, 119.0, 109.7, 108.7, 41.0, 32.3; MS(APCI)
m/z 424.4 [(M+H+) calculated for C26H22N3O3, 424.5]. Anal. Calcd
for C26H21N3O3 (423.45): C, 73.74; H, 5.00. Found: C, 73.56; H, 5.20.
An aliquot (99
l
l, 1.0 ꢁ 105 cell/ml) of K562 cells in media was
added to each well in a microtitre plate, followed by an aliquot
(1 ll) of the test compound that was previously prepared at a
100-fold concentration. The final concentration of DMSO in each
well was 0.4% v/v. The same procedure was followed for the
HL60 cells except that a higher cell density (2.0 ꢁ 105 cells/ml)
was used. The test compounds were incubated with K562 cells
for 72 h at 37 °C, 5% CO2, while the HL60 cells were incubated for
five days due to their slower growth rate. After the incubation per-
iod, an aliquot (25 ll) of MTT in media was added to give a final
concentration of 0.5 mg/ml per well. The plate was gently agitated,
incubated for 3 h at 37 °C, after which it was centrifuged (1200 g,
10 min), the supernatant removed by aspiration and DMSO
(100 ll) added to dissolve the formazan crystals. The plate was
read at 550 nm. Controls were wells of similar composition except
for the exclusion of test compound (‘vehicle’) and wells containing
only 0.4% DMSO in media (‘blank’). A slightly different protocol was
followed for the adherent HCT116, HuH7 and IMR-90 cells.
4.15. 1,10-Dimethyl-isoindigo (12)
Four hundred milligrams of dry K2CO3 (excess) were added to
isoindigo (265 mg,1 mmol) dissolved in 2.5 ml anhydrous DMF in
a 5 ml microwave vessel and allowed to stir at room temperature
for 15 min. Methyl iodide (624 ll, 10 mmol) was injected and the
mixture was stirred at room temperature for 5 min before micro-
HCT116, HUH7 and IMR-90 cells in media (100 ll) were seeded
at 10,000 cells/well, 6000 cells/well, and 4000 cells/well, respec-
tively. The cells were allowed to adhere to the well over 24 h after
which media containing the test compound (prepared in 0.5% v/v
DMSO-media) was added and the plates incubated for 72 h at
37 °C, 5% CO2. After incubation, the supernatant in each well was
wave irradiation at 150 °C for 10 min. Five milliliters of water were

7570
X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
removed by aspiration and the cell layer gently rinsed with PBS. An
aliquot of MTT (100 l, 0.5 mg/ml in phosphate buffer) was added
to each well, incubated for 4 h, 37 °C, after which the MTT solution
was removed and DMSO (100 l) added to dissolve the formazan
crystals. Absorbances were measured within 30 min at 590 nm
on a microplate reader. Control wells were maintained as de-
scribed earlier. The viability of cells at a given concentration of test
compound was determined from the following expression:
at 300 K and pH 7.4. The non-heavy atoms were minimized using
the AMBER force field and unbound water molecules at the active
site were deleted. No force field minimization was carried out on
the heavy atoms. The co-crystallized ATP was extracted from the
crystal structure (1FIN) to give an apo-CDK2/CyclinA structure.
The extracted ATP structure was then re-docked onto the apo-CDK2/
cyclin A structure with the software GOLD 4.0 (CCDC, Cambridge,
UK). The variation (measured in terms of the heavy atom root mean
square deviation RMSD) between the most favored re-docked pose of
ATP and the original ATP in 1FIN was 1.84 Å, which was an acceptable
variation considering that ATP has a large number of rotatable bonds.
This process served as a means of validating the subsequent docking of
the test compounds as it showed that the software parameters selected
for the process were appropriate.
The test compounds used in the docking exercise were drawn
with MOE. As the bisindole ring of isoindigo was reported to be
planar31, semi-empirical PM3 method was used to energy mini-
mize (gradient 0.00001 kcal/mol) the isoindigo ring. Thereafter,
the potential of the heavy atoms of the isoindigo ring were con-
strained using fixed potential and the substituents groups were
added and minimized with MMFF94ꢁ forcefield (gradient
0.00001 kcal/mol) and saved as mol2 files. This procedure was
found to be necessary to ensure planarity of the isoindigo ring
as the usual forcefield minimization was found to cause the ring
l
l
Percentage viability ¼ ðAdrug ꢂ Ablank=Avehicle ꢂ AblankÞ ꢁ 100;
where Adrug = absorbance of wells with cells grown in vehicle
(DMSO/media) and treated with test compound, Ablank = absorbance
of wells with only 100 ll DMSO and Avehicle = absorbance of wells
with cells grown in vehicle (DMSO/media). Each concentration of
test compound was evaluated at least four times on separate occa-
sions and different cell passages. The IC50 (concentration that inhib-
ited 50% of cell growth) was determined from the sigmoidal curve
obtained by plotting percentage viability versus logarithmic con-
centration of test compound using Graphpad Prism 4 (San Diego,
CA).
4.18. Determination of CDK2 inhibitory activity using the IMAP
assay
The IMAP 800-tp kit (Molecular Devices R8155) was purchased
for the assay. The EC70 value of the CDK2/cyclinA (Millipore #14–
448) enzyme was first determined using the protocol described
to assume a non-planar conformation. Partial charges were
added using the Gasteiger and Hückel charge calculation method
using Sybyl8.0 (Tripos). Docking of the test compounds was con-
ducted on the apo-CDK2/CyclinA protein using GOLD 4.0. The
geometric center of the co-crystallized ATP was determined
using MOE. All CDK2 atoms within 12 Å spherical radius of the
geometric center were defined as the docking pocket and it in-
cluded the ‘hinge’ region amino acids of the kinase active site.
All ligands were added as a single input in one GOLD run. The
number of genetic operations performed were determined by
using the automatic (ligand-dependent) genetic algorithm
parameter settings. Search efficiency was set at the default
100%. The number of poses returned by GOLD was determined
by the default settings. Atom and bond types were assigned
automatically for the ligands. No constraints were set for the
GOLD run.
The CDK2ꢂligand interactions of all docked poses were ana-
lyzed using the Protein Ligand Interaction Fingerprint (PLIF) mod-
ule in MOE. PLIF uses a fingerprint scheme comprising of various
interactions. Hydrogen bonds, ionic interactions and surface con-
tacts were considered according to the residue of origin. Since
known inhibitors of CDK2 formed hydrogen bonds with Glu81
and/or Leu83 in the hinge region of the active site, the BitSelector
function of PLIF was used to filter out those ligands that had no
probability of forming interactions with Glu81 and Leu83 in any
of their docked poses. Among those ligands that were likely to
form interactions with Glu81 and Leu83, the probability of each
ligand forming such interactions were scored using PLIF for each
ligand. PLIF calculated hydrogen bonds between polar atoms
using a method based on protein contact statistics,32 whereby a
pair of atoms were scored by their distances and orientation.
The score was expressed as a percentage probability of forming
a H bond. The individual score for each interaction (Glu81 as do-
nor, Leu83 as donor, Leu83 as acceptor) was obtained and
summed up to produce a new score called HingeSum which re-
flected the total probability of a ligand establishing either one of
the three possible interactions with Glu8a/Leu83 at the hinge re-
gion of CDK2. Only the docked pose of each compound with the
highest HingeSum was retained while the other poses were dis-
carded. This method of selecting poses placed emphasis on poses
that formed significant interactions with the hinge region of
CDK2.
in the kit. Aliquots (5
buffer, (ii) CDK2/cyclinA enzyme (3.96, 1.32, 0.44, 0.15, 0.05,
0.02, 0.005 unit/ml), (iii) 120 M ATP and (iv) 400nM lyophilized
ll) of the following: (i) complete reaction
l
FAM-Histone H1-derived peptide (5FAM-GGGPATPKKAKKL-COOH)
(Molecular Devices, R7252) were added in sequence to into a well
(384-well black microplate, Sigma, CLS3571) such that the final
concentration of ATP and peptide substrate were 30
100 nM, respectively. The mixture was incubated for 1 h for the
phosphorlyation reaction to take place, after which 60 l of the
lM and
l
progressive binding solution (1/400 dilution of binding reagent)
was added and incubated for another hour. Fluorescent polariza-
tion (FP) readings were taken on a microplate reader (MDC Spec-
tramax M5) at kExcitation 485 nm and kEmission 530 nm. The EC70 of
CDK2/cyclin A was determined from the sigmoidal curve obtained
by plotting FP versus logarithmic enzyme concentration (GraphPad
Prism 4, San Diego, CA). The assay was repeated in the presence of
10
(40
120
l
l
M test compound. Briefly, 5
M in 4% v/v DMSO), CDK2/cyclin A at 4ꢁ EC70 concentration,
lM ATP and 400nM lyophilized FAM-Histone H1-derived pep-
ll aliquots of the test compound
tide was added to each well. The final concentration of DMSO in
the well was maintained at 1% v/v. Control wells contained com-
plete reaction buffer (5 ll) in place of the same volume of enzyme
solution and the ‘vehicle’ control were wells that did not contain
test compound but with other reagents unchanged. The enzyme
inhibition activity of test compound at 10
the following equation:
lM was calculated using
Percentage inhibition ¼ ðFPvehicle ꢂ FPdrug=FPvehicle ꢂ FPcontrol
ꢁ 100;
Þ
where FP is the fluorescent polarization units (mP) measured for the
test compound, vehicle control and control. Each test compound
was evaluated at least twice on separate occasions.
4.19. Docking with CDK2
The CDK2 protein complex (1FIN) was protonated in the pres-
ence of its ligand ATP using the Protonate3D process in Molecular
Operating Environment (MOE, version 2007.09, Montreal, Canada)

X. K. Wee et al. / Bioorg. Med. Chem. 17 (2009) 7562–7571
7571
11. Liu, X. M.; Wang, L. G.; Li, H. Y.; Ji, X. J. Biochem. Pharmacol. 1996, 51,
1545.
Acknowledgments
12. Xiao, Z.; Wang, Y.; Lu, L.; Li, Z.; Peng, Z.; Han, Z.; Hao, Y. Leukocyte Res. 2006, 30,
54.
13. Wang, L.; Liu, X.; Chen R. WO Patent 03/051900, 2003.
14. Sassatelli, M.; Bouchikhi, F.; Aboab, B.; Anizon, F.; Fabbro, D.; Prudhomme, M.;
Moreau, P. Anti-cancer Drug 2007, 18, 1069.
15. Sassatelli, M.; Bouchikhi, F.; Messaoudi, S.; Anizon, F.; Debiton, E.; Barthomeuf,
C.; Prudhomme, M.; Moreau, P. Eur. J. Med. Chem. 2006, 41, 88.
16. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Adv. Drug Delivery Rev.
1997, 23, 3.
This work was supported by National University of Singapore
Academic Research Fund (R-148-000-084-112) to GML. The re-
search scholarship to Wee Xi Kai from Ministry of Education,
Republic of Singapore and National University of Singapore is
gratefully acknowledged.
Supplementary data
17. Veber, D. F.; Johnson, S. R.; Cheng, H.-Y.; Smith, B. R.; Ward, K. W.; Kopple, K. D.
J. Med. Chem. 2002, 45, 2615.
18. Jeffery, P. D.; Russo, A. A.; Polyak, K.; Gibbs, E.; Hurwitz, J.; Massague, J.;
Pavletich, N. P. Nature 1995, 376, 313.
19. Vulpetti, A.; Pevarello, P. Curr. Med. Chem. Anti-Cancer Agents 2005, 5,
561.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2009.09.008.
20. Papageorgiou, C.; Borer, X. Helv. Chim. Acta 1988, 71, 1079.
21. Elgazwy, A. S. H.; Atta-Allah, S. R. Afinidad 2008, 534, 148.
22. Azizan, J.; Fallah-Bagher-Shaidaei, H. H. Synth. Commun. 2003, 33, 789.
23. Edmonson, J. M.; Armstrong, L. S.; Martinez, A. O. J. Tissue Cult. Methods 1988,
11, 15.
References and notes
1. Tang, W.; Eisenbrand, G. In Chinese Drugs of Plant Origin: Chemistry,
Pharmacology and Use in Traditional and Modern Medicine; Springer:
Heidelberg, 1992.
2. Xiao, Z.; Hao, Y.; Liu, B.; Qian, L. Leukocyte Lymph. 2002, 43, 1763.
3. Hoessel, R.; Leclerc, S.; Endicott, J. A.; Nobel, M. E. M.; Lawrie, A.; Tunnah, P.;
Leost, M.; Damiens, E.; Marie, D.; Marko, D.; Niederberger, E.; Tang, W.;
Eisenbrand, G.; Meijer, L. Nat. Cell Biol. 1999, 1, 60.
4. Polychronopoulos, P.; Magiatis, P.; Skaltsounis, A.; Myrianthopoulos, V.;
Mikros, E.; Tarricone, A.; Musacchio, A.; Roe, S. M.; Pearl, L.; Leost, M.;
Greengard, P.; Meijer, L. J. Med. Chem. 2004, 47, 935.
5. Zhen, Y.; SØrensen, V.; Jin, Y.; Wiedlocha, A. Oncogene 2007, 26, 6372.
6. Kim, S.-A.; Kim, S.-W.; Chang, S.; Yoon, J.-H.; Ahn, S.-G. Cancer Lett. 2009, 274,
72.
7. Ribas, J.; Bettayeb, K.; Ferandin, Y.; Knockaert, M.; Garrofé-Ochoa, X.; Totzke, F.;
Schächtele, C.; Mester, J.; Polychronopoulos, P.; Magiatis, P.; Skaltsounis, A.-L.;
Boix, J.; Meijer, L. Oncogene 2006, 25, 6304.
8. Ji, X.-J.; Zhang, F.-R. Acta Pharm. Sin. 1985, 20, 137.
9. Clinical Therapeutic Coordinating Group of Meisoindigo Acta Haematol. Sin.
1988, 9, 135.
24. Marks, D. C.; Belov, L.; Davey, M. W.; Davey, R. A.; Kidman, A. D. Leukocyte Res.
1992, 16, 1165.
25. Wolff, M. E. In Burger’s Medicinal Chemistry and Drug Discovery, 5th ed.. In
Principles and Practise; John Wiley & Sons, 1995; Vol. 1, pp 507–509.
26. Sportsman, J. R.; Gaudet, E. A.; Boge, A. Assay Drug Dev. Tech. 2004, 2, 205.
27. Fabian, M. A.; Biggs, W. H.; Treiber, D. K.; Atteridge, C. E.; Azimioara, M. D.;
Benedetti, M. G.; Carter, T. A.; Ciceri, P.; Edeen, P. T.; Floyd, M.; Ford, J. M.;
Wodicka, L. M.; Patel, H. K.; Zarrinkar, P. P.; Lockhart, D. J. Nat. Biotechnol. 2005,
23, 329.
28. Vendonk, M. L.; Mortenson, P. N.; Hall, R. J.; Hartshorn, M. J.; Murray, C. W. J.
Chem. Inf. Model. 2008, 48, 2214.
29. Poon, R. Y. C.; Yamashita, K.; Adamczewski, J. P.; Hunt, T.; Shuttleworth, J.
EMBO J. 1993, 12, 3123.
30. Abbas, T.; Jha, S.; Sherman, N. E.; Dutta, A. Cell Cycle 2007, 6, 843.
31. Yuan, M. S.; Fang, Q.; Ji, L.; Yu, W. T. Acta Crystallogr., Sect.
E 2007, 63,
o4342.
32. Labute, P. J. Chem. Comput. Group 2001 http://www.chemcomp.com/journal/
10. Ji, X.; Liu, X.; Li, K.; Chen, R.; Wang, L. Biomed. Environ. Sci. 1991, 4, 332.
cstat.htm.