Exploring the Anticancer Activity of Functionalized Isoindigos: Synthesis, Drug-like Potential, Mode of Action and Effect on Tumor-Induced Xenografts

Article abstract of DOI:10.1002/cmdc.201200018

Meisoindigo has been used as an indirubin substitute for the treatment of chronic myeloid leukemia (CML) for several years. In view of its poor solubility and erratic absorption, several investigations have focused on developing analogues with more desirable physicochemical profiles. Here, we investigated the structure-activity relationship (SAR) of meisoindigo with respect to its antiproliferative activity on leukemic K562 cells and found that appending a phenalkyl side chain onto the lactam NH resulted in analogues that retained good activity. Furthermore, analogues in which the phenyl ring was substituted with a basic heterocycle were significantly more soluble than meisoindigo while retaining acceptable antiproliferative profiles. The most promising analogue (E)-1-(2-(4-methylpiperazin-1-yl)ethyl)-[3,3′-biindolinylidene]-2,2′-dione (5-4) is more potent than meisoindigo across a panel of malignant cells, with at least 40 times greater solubility than meisoindigo, little or no tendency to aggregate in solution and capable of significantly extending the lifespans of animals with K562 induced xenografts. Mechanistically, it induced apoptotic cell death and disrupted the progression of K562 cells from the G₁ to G₂ phase. Taken together, our findings highlighted the feasibility of addressing the physicochemical deficits of the isoindigo scaffold by systematic modifications which was achieved without overt loss of growth inhibitory activity.

Full text of DOI:10.1002/cmdc.201200018

MED
DOI: 10.1002/cmdc.201200018
Exploring the Anticancer Activity of Functionalized
Isoindigos: Synthesis, Drug-like Potential, Mode of Action
and Effect on Tumor-Induced Xenografts
Xi Kai Wee, Tianming Yang, and Mei Lin Go*[a]
Meisoindigo has been used as an indirubin substitute for the
treatment of chronic myeloid leukemia (CML) for several years.
In view of its poor solubility and erratic absorption, several in-
vestigations have focused on developing analogues with more
desirable physicochemical profiles. Here, we investigated the
structure–activity relationship (SAR) of meisoindigo with re-
spect to its antiproliferative activity on leukemic K562 cells and
found that appending a phenalkyl side chain onto the lactam
NH resulted in analogues that retained good activity. Further-
more, analogues in which the phenyl ring was substituted
with a basic heterocycle were significantly more soluble than
meisoindigo while retaining acceptable antiproliferative pro-
files. The most promising analogue (E)-1-(2-(4-methylpiperazin-
1-yl)ethyl)-[3,3’-biindolinylidene]-2,2’-dione (5-4) is more potent
than meisoindigo across a panel of malignant cells, with at
least 40 times greater solubility than meisoindigo, little or no
tendency to aggregate in solution and capable of significantly
extending the lifespans of animals with K562 induced xeno-
grafts. Mechanistically, it induced apoptotic cell death and dis-
rupted the progression of K562 cells from the G1 to G2 phase.
Taken together, our findings highlighted the feasibility of ad-
dressing the physicochemical deficits of the isoindigo scaffold
by systematic modifications which was achieved without overt
loss of growth inhibitory activity.
Introduction
The isomeric bisindoles (indigo, indirubin and isoindigo), col-
lectively known as indigoids, are associated with varied phar-
maceutical and chemical applications (Figure 1). The indigos
are valuable textile dyes that have been used since antiquity.[1]
Functionalized indirubins are widely reported to be potent in-
hibitors of various kinases including CDK,[2–5] GSK3b[4,6] and
aurora kinases,[7] and thus are potential therapeutic leads for
cancer. In comparison, the isoindigo scaffold has received
scant attention, although an isoindigo, meisoindigo (1)
(Figure 1), has been used for the treatment of chronic myeloid
leukemia (CML) in the People’s Republic of China since
1992.[8–11] Reports on the mode of action of 1 suggests that it
has several targets. It inhibits DNA and RNA biosynthesis in rat
carcinosarcoma cells,[12] and in common with other DNA-tar-
geting agents, induces differentiation of cancer cells by de-
creasing the expression of the oncogene c-myb.[13] On the cell
cycle, 1 arrests the progression of leukemic cells through the
G0/G1 phase and increases levels of the cell cycle inhibitor pro-
teins p21 and p27.[14] There are conflicting views on whether
1 inhibits the phosphorylation mediated by cyclin-dependent
kinases,[15–17] possibly due to different assay protocols used to
obtain the results. Its effects on apoptosis are pronounced and
involve the intrinsic mitochondrial pathway, with up-regulation
of pro-apoptotic proteins and concurrent down-regulation of
antiapoptotic proteins.[14] Inhibition of angiogenesis through
diminished VEGF secretion has been observed.[18]
A significant drawback of functionalized indigoids is their
poor aqueous solubility, due in part to the “brick-dust” nature
of the scaffold, which promotes tight crystal packing. Not un-
expectedly, efforts in medicinal chemistry have thus far fo-
cused on modifications that would enhance solubility. One ap-
proach has been to attach sugar moieties to the bisindole,
[a] Dr. X. K. Wee, Dr. T. Yang, Dr. M. L. Go
Department of Pharmacy, Faculty of Science
National University of Singapore
18 Science Drive 4, 117543 (Singapore)
E-mail: phagoml@nus.edu.sg
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201200018.
Figure 1. Structures of bisindoles, meisoindigo, natura and 5-4.
ChemMedChem 2012, 7, 777 – 791
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
777

M. L. Go et al.
MED
giving rise to N-glycosides, and in the case of indiru-
bins, O-glycosides from glycosylation of the 3’-oxime
derivatives.[16,19–22] Unfortunately, greater solubility is
often achieved at the expense of biological activity.
For example, the water-soluble N-xylosyl analogue of
isoindigo has negligible antiproliferative activity
which was restored when the hydroxy groups on the
sugar moiety was acetylated.[16,19] The resulting com-
pound (natura, Figure 1) was apparently designed to
increase the bioavailability and bioactivity of 1, al-
though it has reportedly poor solubility in water.[16,19]
A large number of N-glycosylisoindigos have been
found to be kinase inhibitors but only when the hy-
droxy groups on the sugar moiety were O-benzylated
or O-acetylated.[21] It is clear that designing biologi-
cally active bisindoles that combine acceptable solu-
bility–permeability profiles with potent activity is
a daunting task.
Table 1. Series 1: IC50, K562 values, ClogP and estimated solubilities (pH 7.4).
R1
ClogP[a] Solubility
[mm][b]
IC50
[mm][c]
1
Me
2.04
673
7.75ꢀ0.71
5.30ꢀ0.97
1-1
4.37
7.6
1-2
1-3
4.73
4.73
7.6
7.6
4.55ꢀ0.06[d]
6.09ꢀ1.01
1-4
1-5
4.95
5.89
5.2
18.95ꢀ3.29
3.79ꢀ0.57[d]
11.8
In an earlier study, we have evaluated several isoin-
digos, including 1, for their antiproliferative activities
on a panel of human cancer cells.[17] The structure–
activity relationship (SAR) revealed key positions on
the scaffold that were sensitive to modification, nota-
bly, the N-methyl of 1, which could be replaced by
N-phenpropyl or N-(p-methoxy)phenethyl leading to
significant improvements in activity. In order to
expand the nascent SAR derived from that investiga-
tion, we have undertaken a systematic exploration of
the isoindigo scaffold with the aim of identifying ana-
logues that reconcile potency (assessed in terms of
antiproliferative activity on the CML cell line K562)
with acceptable aqueous solubility. In this regard, we
found that the isoindigo scaffold could be modified
to accommodate these dual requirements. The most
1-6
1-7
1-8
1-9
5.19
5.08
4.43
4.48
5.3
2.6
1.70ꢀ0.22[d]
2.34ꢀ0.33[d]
9.31ꢀ1.18
<1.0
10.5
3.84ꢀ0.53[d]
1-10
5.11
4.9
1.75ꢀ0.14[d]
[a] Determined with ChemDraw 7.0. [b] Determined with ACD/Labs 12.0 Solubility DB.
[c] Values represent the mean ꢀSD for nꢁ3 determinations. [d] Significantly lower
than meisoindigo 1 (p <0.05, one-way ANOVA, Dunnett’s post hoc).
promising compound 5-4 is more potent than 1 on K562 cells,
has improved solubility and shows no tendency to form aggre-
gates in solution. 5-4 increased the survival of K562 xenograft-
bearing nude mice on intratumoral administration. Mechanisti-
cally, it induces cell cycle arrest at G1 and promotes apoptosis
of K562 cells.
p-methoxy substituent was also introduced to the phenyl ring
of 1-6 to give 1-10.
Series 2 comprises ring substituted analogues of 1 (Table 2).
The substituents (fluoro, chloro, bromo and methoxy) were in-
troduced at positions 5, 6 or 7 of ring A (2-1 to 2-12) or the
corresponding positions on ring B (2-13 to 2-24). The choice of
substituents was determined largely by the commercial availa-
bility of reagents and, in the case of bromine, its striking asso-
ciation with antiproliferative activity, when attached to the in-
dirubin scaffold.[23] As each series was synthesized and evaluat-
ed iteratively, inputs from one series determined the modifica-
tions made to the next series. Thus, having noted the excep-
tional activity of analogue 2-20 with 6’-methoxy in series 2
(Table 2), this group was incorporated in series 3 (Table 3),
which also intended the replacement of the N1-methyl sub-
stituent with side chains present in some of the more potent
compounds of series 1. In a related manner, having deduced
that the optimal substituent at N1 is the p-methoxyphenethyl
side chain from series 3, this feature was kept constant in
series 4 (Table 4), which now saw other groups (methoxy,
fluoro, chloro or bromo) introduced to either ring A or B of the
scaffold. Thus, based on the activities of compounds from
Results
Design and synthesis of target compounds
The target compounds were broadly divided into five groups
(series 1–5) which are presented in Tables 1–5, respectively.
Series 1, shown in Table 1, comprises compounds with varying
phenalkyl side chains attached to the NH at position 1 (N1) and
is based on two hits that were identified earlier,[17] namely 1-(p-
methoxyphenethyl) analogue 1-3 and 1-(3-phenylpropyl) ana-
logue 1-6. Here, 1-3 was modified to give its o-methoxy and
m-methoxy regioisomers 1-1 and 1-2), respectively, the
p-fluoro analogue 1-4 and the cyclohexyl analogue 1-5. The
alkyl side chain of 1-6 was rigidified by unsaturation (1-7, 1-8)
or rendered less lipophilic by inserting an ether linkage (1-9). A
778
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2012, 7, 777 – 791

Anticancer Activity of Functionalized Isoindigos
MED
series 3 and 4, the optimal combination of groups on rings A/B
and N1 could be deduced.
acceptable activity in several variants of the phenethyl side
chain (series 1, Table 1), we speculated that a viable approach
would be, to replace the lipophilic aromatic ring of the side
chain with a solubilizing basic heterocyclic ring, such as, mor-
pholine, thiomorpholine, piperidine or 1-methylpi-
perazine. Inserting an azomethine nitrogen in the
isoindigo scaffold to give azaisoindigo was also con-
sidered, as estimated values point to a potential de-
IC50
Series 5 shown in Table 5 comprises compounds that were
designed to have greater aqueous solubility. Having observed
Table 2. Series 2: IC50, K562values, ClogP and estimated solubilities (pH 7.4).
R2/R3
ClogP[a]
Solubility
[mm][b]
[mm][c]
crease in lipophilicity (logD7.4 value of 1.03 and 0.6
for isoindigo and azaisoindigo, respectively) and an
increase in solubility (aqueous solubility of 62 mm
and 330 mm for isoindigo and azaisoindigo, respec-
tively) with this modification.
2-1
2-2
2-3
2-4
2-5
2-6
2-7
2-8
5-F
5-Cl
5-Br
5-OMe
6-F
6-Cl
6-Br
6-OMe
7-F
7-Cl
2.86
3.43
3.58
2.63
2.86
3.43
3.58
2.63
2.86
3.43
3.58
2.63
3.11
3.68
3.83
2.79
3.11
3.68
3.83
2.79
3.11
3.68
3.83
2.79
642.2
209.2
259.0
692.1
339.8
157.7
166.1
297.1
727.2
234.9
284.4
639.9
713.6
231.7
287.2
656.2
951.5
238.1
191.4
470.1
744.2
241.4
292.8
669.2
19.24ꢀ1.30
32.23ꢀ2.46
>30
5.18ꢀ0.91
10.87ꢀ0.93
11.08ꢀ1.27
9.84ꢀ0.69
9.74ꢀ1.45
13.45ꢀ1.38
17.12ꢀ0.96
20.82ꢀ0.65
5.78ꢀ0.96
18.20ꢀ0.89
21.85ꢀ3.55
14.41ꢀ1.88
5.93ꢀ1.13
14.82ꢀ1.65
18.45ꢀ2.40
6.71ꢀ1.03
3.18[d] ꢀ0.34
32.89ꢀ2.67
Three synthetic routes were adopted for the syn-
thesis of series 1–5. The first approach involves the
N-alkylation of isatin (or ring-substituted isatin)
under basic conditions, followed by aldol condensa-
tion with oxindole (or ring-substituted oxindole)
under acidic conditions (Scheme 1). This route was
adopted for the synthesis of series 3 and 4, and se-
lected compounds in series 1 (1-1 to 1-5, 1-7, 1-8, 1-
10) and series 2 (2-1 to 2-7, 2-9 to 2-11, 2-20).
In the second approach, which is essentially a re-
versal of the first approach, commercially available 1-
methyloxindole is reacted with a ring-substituted
isatin in an acid-catalyzed aldol condensation reac-
tion (Scheme 2). Most of the series 2 compounds
(2-8, 2-12 to 2-19, 2-21 to 2-24) were synthesized by
this route. In the case of 2-8 and 2-12, the substitut-
ed 1-methyloxindole was synthesized by a methyla-
tion reaction with dimethylsulfate.
2-9
2-10
2-11
2-12
2-13
2-14
2-15
2-16
2-17
2-18
2-19
2-20
2-21
2-22
2-23
2-24
7-Br
7-OMe
5’-F
5’-Cl
5’-Br
5’-OMe
6’-F
6’-Cl
6’-Br
6’-OMe
7’-F
7’-Cl
7’-Br
7’-OMe
>30[e]
>30[e]
>30[e]
[a] Determined with ChemDraw 7.0. [b] Determined with ACD/Labs 12.0 Solubility DB.
[c] Values represent the mean ꢀSD for nꢁ3 determinations. [d] Significantly lower
than 1 (p<0.05, one-way ANOVA, Dunnett’s post hoc). [e] Could not be determined
due to limited solubility. >50% viability was observed at the stated concentration.
The third approach involves condensing isatin and
oxindole to give isoindigo, followed by alkylation of
the latter under basic conditions (Scheme 3).
Direct alkylation of isoindigo is a short and
convenient approach, especially as isoindigo
is obtained in excellent yields (91%).[17] How-
ever, the separation of the target compound
from unreacted isoindigo by column chroma-
tography was cumbersome and, consequent-
Table 3. Series 3: IC50, K562 values, ClogP and estimated solubilities (pH 7.4).
R1
ClogP[a] Solubility
[mm][b]
IC50
[mm][c]
3-1
3-2
5.26
5.18
4.9
4.5
2.75ꢀ0.33[d,e]
4.67ꢀ0.77[d]
ly, this route was adopted only when other
approaches had failed to give the target
compounds in good yields, as in the case of
1-6, 1-9, 5-2 and 5-3.
>20[f]
Unlike in series 1–4, the alkyl halides used
for series 5 (5-1 to 5-3, 5-5) had to be synthe-
sized. As shown in Scheme 4, 2-bromoetha-
nol was reacted with the basic heterocycle
(morpholine, thiomorpholine, or piperidine)
to attach the hydroxyethyl side chain to the
basic nitrogen, followed by displacement of
the hydroxy group by chlorine from thionyl
chloride. 1-(2-Chloroethyl)morpholine was
then reacted with isatin to give compound
41, which subsequently was condensed with
oxindole (or 6-methoxyoxindiole) to give 5-
1 and 5-5. In the case of 1-(2-chloroethyl)th-
3-3
4.80
4.7
3-4
3-5
4.80
4.80
4.7
7.0
11.43ꢀ0.66
0.50ꢀ0.06[d,e,g]
[a] Determined with ChemDraw 7.0. [b] Determined with ACD/Labs 12.0 Solubility DB. [c] Values
represent the mean ꢀSD for nꢁ3 determinations. [d] Significantly lower than 1 (p<0.05, one-
way ANOVA, Dunnett’s post hoc). [e] Significantly lower than 2-20 (p<0.05, one-way ANOVA,
Dunnett’s post hoc). [f] Could not be determined due to limited solubility. >50% viability was
observed at the stated concentration. [g] Significantly lower than corresponding series 1 ana-
logue 1-3 (p<0.05, one-way ANOVA, Dunnett’s post hoc).
ChemMedChem 2012, 7, 777 – 791
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
779

M. L. Go et al.
MED
iomorpholine and 1-(2-chloroethyl)piperidine, these were di-
give 1-(2-bromoethyl)indolin-2,3-dione 42 and then reacted
with 2-oxindole to give the corresponding isoindigo, after
which the bromine atom in the side chain was displaced by 1-
methylpiperazine. This approach has the advantage of facilitat-
ing the separation of the less polar intermediate from the final
product by column chromatography.
rectly reacted with isoindigo to give 5-2 and 5-3.
A
different approach was followed for 5-4 and 5-6
(Scheme 5), which both have a 4-methyl-1-piperazinylethyl
side chain. Here, isatin was alkylated with dibromoethane to
The azaisoindigos 5-7 and 5-8 were prepared from
the common intermediate 7-aza-2-oxindole 46
(Scheme 6) which was synthesized by a reported
method.[24] The intermediate was reacted with 1-
Table 4. Series 4: IC50, K562 values, ClogP and estimated solubilities (pH 7.4).
R2 or R3
ClogP[a] Solubility
[mm][b]
IC50
[mm][c]
methylisatin or 1-(p-methoxyphenethyl)isatin to give
the desired products 5-7 and 5-8.
4-1 5-F
4.87
5.44
5.59
4.65
4.87
5.44
5.59
4.65
4.87
5.44
5.59
7.2
2.3
4.2
9.4
4.8
2.3
2.1
4.7
9.7
2.3
4.2
18.58ꢀ0.82
4-2 5-Cl
4-3 5-Br
4-4 5-OMe
4-5 6-F
4-6 6-Cl
4-7 6-Br
4-8 6-OMe
4-9 7-F
>30[d]
>30[d]
>30[d]
>30[d]
>30[d]
Antiproliferative activity on K562 cells
10.75ꢀ1.62
5.15ꢀ0.90[e]
22.24ꢀ2.51
The antiproliferative activities of the target com-
pounds were determined on K562 cells and ex-
pressed in terms of half maximal growth inhibitory
concentrations (IC50). Tables 1–5 list the IC50 values for
series 1–5 compounds as well as their estimated lipo-
philicities (ClogP) and aqueous solubilities.
4-10 7-Cl
4-11 7-Br
>30[d]
>30[d]
4-12 5’-F
4-13 5’-Cl
4-14 5’-Br
4-15 6’-F
4-16 6’-Cl
4-17 6’-Br
3-5 6’-OMe
4-18 7’-F
4-19 7’-Cl
5.13
5.70
5.85
5.13
5.70
5.85
4.80
5.13
5.70
5.85
4.89
6.06
4.76
3.73
3.75
9.7
2.3
4.2
12.0
2.3
2.1
7.0
9.7
2.3
4.2
4.6
<1.0
2.4
12.6
6.6
14.03ꢀ1.53
12.01ꢀ0.51
12.04ꢀ0.60
2.42ꢀ0.28[e,f]
1.05ꢀ0.10[e,f]
1.06ꢀ0.17[e,f]
0.50ꢀ0.06[e,f]
Initially, we had identified 1-3 and 1-6 as the most
promising compounds, with IC50, K562 values of 1.4 mm
and 1.6 mm, respectively, relative to 6.7 mm for 1.[17]
A
>10[d]
re-determination of the IC50, K562 values gave compara-
ble values for 1 (7.75 mm) and 1-6 (1.70 mm), but not
for 1-3 (6.09 mm). Interestingly, 1-6 was again the
most potent compound in series 1, followed closely
by its p-methoxyphenpropyl analogue (1-10,
IC50, K562 =1.75 mm). The SAR for 1-6 showed that the
activity was significantly decreased only when a triple
bond was inserted into the phenpropyl side chain (1-
8). Other modifications, such as introducing a double
bond (1-7) or an ether linkage (1-9) in the side chain
of 1-6 were well tolerated, with only incremental
losses to activity. In the case of 1-3, which was initial-
ly thought to be a potent hit, the replacement of the
p-methoxyphenethyl side chain with a non-aromatic
cyclohexylethyl side chain (1-5) significantly im-
>30[d]
>10[d]
>30[d]
4-20 7’-Br
4-21 6’-NMe2
4-22 6’-CF3
4-23 6’-CN
4-24 6’-SO2Me
4-25 6’-NH(C=O)Me
9.81ꢀ1.21
13.82ꢀ0.14
10.30ꢀ0.54
9.79ꢀ1.54
[a] Determined with ChemDraw 7.0. [b] Determined with ACD/Labs 12.0 Solubility DB.
[c] Values represent the mean ꢀSD for nꢁ3 determinations. [d] Could not be deter-
mined due to limited solubility. >50% viability was observed at the stated concentra-
tion. [e] Significantly lower than 1 (p<0.05, one-way ANOVA, Dunnett’s post hoc).
[f] Significantly lower than 1-3 (p<0.05, one-way ANOVA, Dunnett’s post hoc).
proved activity, while a relocation of the p-methoxy substitu-
ent to the ortho or meta position of the ring (1-1, 1-2) did not
adversely affect activity. On the other hand, inserting a p-fluoro
group on the phenyl ring of 1-3 was detrimental, which was in
line with earlier observations, where a ring substitution with
p-methyl, p-hydroxy or p-nitro groups resulted in sharp losses
in activity.[17]
Scheme 1. Reagents and conditions: a) Alkyl halide R1ꢂX, DMF, CaH2, 808C, 2–
4 h; b) Alkyl halide R1ꢂX, DMF, K2CO3, MW, 1508C, 15–40 min; c) AcOH, HCl,
reflux, 16 h; d) AcOH, HCl, MW, 2008C, 30 min.
Scheme 2. Reagents and conditions: a) NaH, Me2SO4, xylene, 508C, 3 h;
b) AcOH, HCl, reflux, 16 h.
Scheme 3. Reagents and conditions: a) Alkyl halide R1ꢂX, DMF, K2CO3, 808C,
3 h; b) Alkyl halide R1ꢂX, DMF, K2CO3, MW, 1508C, 15–40 min
780
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2012, 7, 777 – 791

Anticancer Activity of Functionalized Isoindigos
MED
The results for series 2 (Table 2) showed that there was gen-
erally little benefit in introducing substituents on ring A or B of
meisoindigo 1. Only compound 2-20 (6’-methoxy) was more
potent than 1 (IC50, K562 =3.18 mm). Curiously, the methoxy sub-
stitutent fared better than the halogens when introduced at
every position on rings A and B, except position 7’ on ring B,
where both halogens and methoxy were
Table 5. Series 5: IC50, K562 values, ClogD and estimated solubilities (pH 7.4).
poorly tolerated. Notionally, this might
be related to the proximity of position 7’
[a]
R1
R3 ClogD7.4 Solubility
IC50
to the NH of the isoindigo scaffold. It
could be that substitution at this position
sterically hinders hydrogen bonding at
this site. Indeed, methylation of this NH
group in 1 to give dimethylisoindigo was
found to abolish antiproliferative activity
on K562 cells,[17] supporting a critical hy-
drogen bonding role for the NH moiety.
As mentioned earlier, the intent of
series 3 is to determine, whether the ac-
tivity would be improved in analogues
that carry both the optimal 6’-methoxy
group and the favored sidechain(s) iden-
tified in series 1. Disappointingly, this was
not found to be so. In fact, only analogue
3-5 (IC50, K562 =0.50 mm), which bears the
6’-methoxy group and the p-methoxy-
phenethyl side chain, was more active
[mm][b]
[mm][c]
5-1
5-2
5-3
H
H
H
1.13
1.87
1.47
258.4
6.81ꢀ0.14
81.7
7.91ꢀ0.51
7.70ꢀ0.74
720.3
5-4
5-5
5-6
5-7
H
OMe
OMe
–
0.88
1.15
0.90
0.85
602.4
192.7
456.4
3.97ꢀ0.21[d]
3.78ꢀ0.51[d]
5.89ꢀ0.67[d]
Me
1644.5 17.25ꢀ0.51
than its series 1 analogue 1-3 (IC50, K562
=
6.09 mm). The remaining compounds in
series 3 showed sharp losses in activity
when the 6’-methoxy group was intro-
duced.
5-8
–
3.03
20.1
>20[e]
The SAR of 3-5 was further explored in
series 4 by varying the position and type
of substitution on rings A and B (Table 4).
As in series 2, there was no added bene-
fit arising from substitution of either
ring A or B. However, unlike series 2, sub-
stitution on ring B of series 4 resulted in
more active analogues than substitution
[a] Determined with ACD/Labs 12.0 LogD. [b] Determined with ACD/Labs 12.0 Solubility DB.
[c] Values represent the mean ꢀSD for nꢁ3 determinations. [d] Significantly lower than 1 (p<
0.05, one-way ANOVA, Dunnett’s post hoc). [e] Could not be determined due to limited solubility.
>50% viability was observed at the stated concentration.
Scheme 4. Reagents and conditions: a) 2-Bromoethanol, toluene, triethyl-
amine, reflux, 4 h; b) Thionyl chloride, 0–808C, 2 h; c) Isoindigo, K2CO3, DMF,
808C, 4 h; d) Isatin, DMF, 808C, 4 h; e) 2-Oxindole or 6-methoxy-2-oxindole,
AcOH, HCl, reflux, 16 h.
Scheme 5. Reagents and conditions: a) 1,2-Dibromoethane, K2CO3, DMF, RT,
4 h; b) 2-Oxindole or 6-methoxy-2-oxindole, AcOH, HCl, reflux, 16 h; c) N-
Methylpiperazine, triethylamine, DMF, 808C.
ChemMedChem 2012, 7, 777 – 791
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
781

M. L. Go et al.
MED
namely NB4 (an acute promyelocytic leukemic cell line), HuH7
(a hepatocellular carcinoma cell line), RCC786 (a renal carcino-
ma cell line) and HCT116 (a colon carcinoma cell line). They
were also screened against non-malignant IMR90 (human lung
fibroblasts) to establish selective activity. Seven analogues, in-
cluding 1, were shortlisted for screening, namely those that
were most active on K562 cells in their respective series: 1-6,
1-10 of series 1, 2-20 of series 2, 3-5 of series 3 and 5-4 of
series 5. Compound 1-3 was included to represent a compound
with moderate potency. The results are given in Table 6.
Two general trends are noted. First, the isoindigos were dis-
tinctly more potent on leukemic cells (K562, NB4) than on ma-
lignant cells derived from solid tumors (HCT116, HuH7,
RCC786). Notably, 1-3, 1-6, 1-10, and 3-5 had submicromolar
Scheme 6. Reagent and conditions: a) PBPB, tBuOH, RT, 6 h; b) Pd/C, H2,
50 psi, RT, 4 h; c) 1-Methylisatin or 4, EtOH, HCl, reflux, 4 h.
on ring A. Position 6’ was again
Table 6. IC50 values [mm] of selected compounds on K562, NB4, HuH7, RCC786, HCT116 and IMR90 cells.[a]
the most favored position in
series 4, with both 6’-methoxy
and 6’-halogen modifications
giving rise to exceptionally
active analogues. However, the
poor activity of analogues with
other substituents at position 6’,
such as, dimethylamino (4-21),
trifluoromethyl, (4-22), cyano (4-
23), methylsulfonyl (4-24) and
acetamide (4-25) bears consider-
ation, as it suggests that the
type of group at position 6’, and
not just its occupancy, plays
Compd
K562
NB4
HuH7
RCC786
HCT116
IMR90
1
7.75ꢀ0.71
6.09ꢀ1.01
2.38ꢀ0.27
20.43ꢀ0.95
7.51ꢀ0.31[b]
7.87ꢀ0.33[b]
6.62ꢀ0.32[b]
7.39ꢀ0.61[b]
5.16ꢀ0.94[b]
9.49ꢀ0.51[b]
5.53ꢀ0.33
5.22ꢀ0.39
4.46ꢀ0.54
3.99ꢀ0.44
31.96ꢀ1.61
12.76ꢀ0.44[b]
6.76ꢀ0.83[b]
7.03ꢀ0.46[b]
17.57ꢀ1.60[b]
4.96ꢀ0.32[b]
12.01ꢀ0.27
7.94ꢀ0.34[b]
5.76ꢀ0.16[b]
5.83ꢀ0.63[b]
1-3
1-6
1-10
2-20
3-5
5-4
0.68ꢀ0.05[b]
1.70ꢀ0.22[b] 0.89ꢀ0.07[b]
1.75ꢀ0.14[b] 0.98ꢀ0.11[b]
3.18ꢀ0.34[b] 2.52ꢀ0.23
0.50ꢀ0.06[b] 0.18ꢀ0.01[b]
3.97ꢀ0.21[b] 4.32ꢀ0.16
>20[c]
>20[c]
7.13ꢀ1.30
4.80ꢀ0.83[b]
3.52ꢀ0.21[b] 15.69ꢀ1.14[b]
9.42ꢀ0.42[b]
[a] Values represent the mean ꢀSD for nꢁ3 determinations. [b] Significantly lower than 1 on the specified cell
line (p<0.05, one-way ANOVA, Dunnett’s post hoc). [c] Could not be determined due to limited solubility.
>50% viability was observed at the stated concentration.
a decisive role in influencing the activity.
IC50 values on leukemic NB4, but not on solid tumor cell lines.
Second, many of the isoindigos maintained significantly higher
activity than 1 across the panel of malignant cells (K562, NB4,
HuH7, HCT116). Exceptions were 2-20 and 5-4 on leukemic
NB4 cells, which were not more active than 1, and compound
5-4, which was the only compound that was more potent than
1 on RCC786 cells. When the compounds were tested on the
non-malignant cell line IMR90, selective targeting against the
leukemic cells (K562, NB4) was evident, with selectivity for the
malignant cells ranging from 1.6–9.6-fold (K562) and 2.2–27-
fold (NB4). Compound 3-5 exhibited impressive selective activi-
ty on the leukemic cells relative to IMR90.
Of the compounds of series 5, only the three analogues 5-4,
5-5, and 5-6 were significantly more potent than 1 (Table 5). A
unique role might be inferred for the 4-methyl-1-piperazinyl-
ethyl side chain, as it is present in two of the three active com-
pounds (5-4, 5-5). However, both compounds did not achieve
the submicromolar growth inhibitory activity of 3-5, which re-
mains the most active analogue identified thus far. The azai-
soindigos 5-7 and 5-8 fared poorly in terms of activity
(IC50, K562 >15 um). In both these compounds, the azomethine
nitrogen is at position 7’, which has been found to be general-
ly intolerant of substitution. Wang and co-workers[25] have re-
ported more promising activity for their series of functionalized
azaindigos, in which the azomethine nitrogen was placed at
position 7. Their best compound had an IC50 value of 10.5 mm
on the prostate cancer cell line DU145, compared to 17.4 mm
for meisoindigo 1. Possibly, more active analogues could be
obtained, if the azomethine nitrogen is introduced at positions
other than 7’ of ring B. Even so, there are doubts whether the
azaisoindigo scaffold would indeed improve water solubility, as
5-8 was too insoluble to permit determination of its IC50 value.
Aqueous solubilities of 3-5 and 5-4
Thus far, we have identified 3-5 as the most potent analogue
and 5-4 as the most promising water soluble analogue against
K562 cells. Notwithstanding its potency, 3-5 is predicted to
have poor solubility at pH 7.4 (7.0 mm), while estimates suggest
that 5-4 is likely to be as soluble as 1 (5-4: 602 mm; 1: 673 mm,
Tables 1 and 5). Hence, it was of interest to determine the
aqueous solubilities (pH 7.4) of 3-5, 5-4, and 1. Determinations
were made on multiscreen solubility filter plates by a method
which involved measuring the amount of compound dissolved
in phosphate buffer (pH 7.4) after 24 h of agitation. As seen in
Table 7, both 3-5 and 5-4 were more soluble than 1 by almost
50-fold. The inclusion of a solubilizing piperazine ring in 5-4
Antiproliferative activity on other malignant and non-malig-
nant cell lines
Next, selected analogues from each series were screened for
growth inhibitory effects on other human malignant cell lines,
782
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2012, 7, 777 – 791

Anticancer Activity of Functionalized Isoindigos
MED
similar conditions. K562 cells were injected into the flank of im-
munocompromised mice to induce tumor formation. Once
a palpable tumor was observed (day 0), the compound (10 mm)
was administered directly into the tumor on days 0, 3 and 10.
The animals were monitored up to day 28, barring the absence
of censored events that would require euthanizing the ani-
mals.
Table 7. Solubility and light scattering properties of meisoindigo 1, 3-5
and 5-4.
Compd
Exptl Solubility
Calcd Solubility
Light Scattering Count
[kcps][c]
[mm][a]
[mm][b]
1[d]
8.34ꢀ1.48
22.36ꢀ3.65
673
7
480ꢀ15 (10 mm)
11 ꢀ1 (1 mm)
455ꢀ14 (10 mm)
4ꢀ1 (0.5 mm)
11 (10 mm)
3-5[d]
5-4[e]
Here, we chose to test the three compounds at the same
dose of 10 mm, in spite of differences in their growth inhibitory
properties (IC50, K562 value of 1>5-4>3-5). The original intent
was to arbitrarily test at concentrations approximately twice
the IC50, K562 value, which would be 16 mm for 1, 10 mm for 5-4,
and 1 mm for 3-5. However, this was not possible, due to the
poor solubility of 1, which did not allow dosing at the pro-
posed concentration. As 1 was readily administered at 10 mm,
it was decided that all three compounds would be evaluated
at this dose, although it would mean that 3-5 would be given
at almost 20 times its IC50 value, exceeding by far that of 1 and
5-4.
348.38ꢀ3.23[f]
602
1 (1 mm)
[a] Determined in universal buffer (pH 7.4) containing 1% v/v DMSO, 24 h
agitation, 288C. Values represent the mean ꢀSD for nꢁ3 determinations.
[b] Calculated by ACD/Labs 12.0 Solubility DB, pH 7.4. [c] Determined in
phosphate buffer (5 mm, pH 7.4) containing 1% v/v DMSO. Counts repre-
sent the mean ꢀSD for nꢁ3 independent determinations. [d] Concentra-
tion of initial solution from which filtrate was derived=100 mm. [e] Con-
centration of initial solution from which filtrate was derived=400 mm.
[f] Solubility can exceed 348 mm because no precipitate was observed in
the starting solutions (400 mm).
has indeed dramatically escalated solubility to the millimolar
range. In the case of meisoindigo (1), its solubility has been
grossly overestimated, possibly because the algorithms used in
the software did not take into account the brick-dust character
of the molecules.
Table 8 summarizes the mean and median survival times of
the xenograft bearing mice in the control and treatment
groups. Mice in the control group (treated with saline-DMSO)
Table 8. Kaplan–Meier analysis for comparison between treatment
groups in xenograft-bearing mice.
Aggregation potential of 3-5 and 5-4
Control
1
3-5
5-4
There is growing interest in the ability of bioactive substances
to form colloid-like aggregates in aqueous solutions.[26–32]
These “aggregators” might inhibit proteins in a nonspecific
manner and lead to the misinterpretation of biological data.
Common characteristics shared by aggregators are low solubil-
ity, high lipophilicity and extended conjugation.[29] Noting that
indirubin and indigo have been reported to form aggre-
gates,[28] there was concern that this could be a shared proper-
ty of the present series of functionalized isoindigos. The aggre-
gation phenomenon can be monitored by measuring the light
scattering capacity of a known concentration of the test com-
pound in phosphate buffer. In the presence of aggregates, sig-
nificant scattering is observed, giving rise to a high count rate.
Table 7 gives the results for 3-5, 5-4 and 1. Benzyl benzoate,
a known aggregator, gave a count rate of 680 kilocounts per
second (kcps) at 250 mm. The scattering capacities of the test
compounds were monitored at a lower concentration of
10 mm. At this concentration, 1 and 3-5 registered counts of
480 kcps and 455 kcps, respectively. The scattering effect is
concentration dependent and was negligible at the lowest test
concentrations (1 mm for 1 and 0.5 mm for 3-5). In the case of
5-4, negligible scattering was observed at 10 mm, indicating
that its improved solubility has served to negate the aggregat-
ing potential of the isoindigo scaffold.
No. of animals
8
2
19
5
1
26
5
0
16
6
5
No. survived after 28 d
Median survival time [d]
Mean survival time [d]
No. of censored events[b]
ILS [%][c]
>28[a]
19.6ꢀ1.9 23.8ꢀ1.8 19.6ꢀ2.5 26.7ꢀ1.2
2
–
–
1
0
5
+21.3
0.631
ꢂ0.1
0.658
+35.9
0.028
Log-rank test (p)[d]
[a] Median could not be determined because 5 of 6 animals in this group
survived beyond 28 d. [b] The only censored event was the survival of an-
imals beyond 28 d. [c] Increase in mean life span (ILS)=[(mean survival
time of treated group ꢂ mean survival time of control group)/mean sur-
vival time of control group]ꢁ100%. [d] p value gives the statistical signifi-
cance between control and treated animals.
survived for 19.6 days, while those treated with 1 and 5-4 lived
longer at 23.8 days and 26.7 days, respectively. Animals treated
with 3-5 survived as long as the control animals. To determine
if the increases in mean lifespans observed with 1 and 5-4
were significant, the data were analyzed by the log-rank test
(pair-wise) comparison statistic p in the Kaplan–Meier analysis.
The results show that only mice treated with 5-4 had signifi-
cantly longer mean life spans (p=0.028). While animals treated
with 3-5 or 1 were not significantly long-lived, neither were
their life spans prematurely curtailed, indicating that these
compounds were relatively nontoxic at the given doses. The
survival curves of animals treated with the three test com-
pounds are given in Supporting Information.
In vivo effects of 3-5 and 5-4 on mice bearing K562 xeno-
grafts
The body weights of the treated animals were monitored
daily and were found to remain constant throughout the test
period (not shown). Changes in the tumor volumes were also
tracked over 28 days (Figure 2). It is seen that tumor volumes
3-5 and 5-4 were administered to xenograft-bearing animals
to assess their in vivo effects. Meisoindigo 1 was tested under
ChemMedChem 2012, 7, 777 – 791
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
783

M. L. Go et al.
MED
value and up to approximately four times this con-
centration. The results are listed in Table 9.
The control K562 cells were not synchronized in
this analysis and showed a high cell density at the G1
phase and about equal amounts of cells in the S and
G2 phases. As seen from Table 9, meisoindigo
1 (IC50 =7.75 mm) caused minimal disruption to the
cell cycle, notwithstanding a small but significant de-
crease in the proportion of cells in the G2 phase at
the highest concentration tested (20 mm). However,
a concurrent increase of cells in the G1 phase, typical
of a G1 arrest, was not observed.
3-5 was investigated at lower concentrations (0.1–
10 mm) in view of its potent growth inhibitory effects
(IC50 =0.5 mm). The results showed that it induced
a concentration dependent decrease in the propor-
tion of cells in G1 while concurrently increasing those
in the G2 phase, a profile which attests to G2 arrest.
There was also a significant increase in the propor-
tion of cells in the sub-G1 phase which is associated
with cellular debris arising from necrotic or apoptotic
cells.
Figure 2. Changes in tumor size (mm3) of xenograft-bearing mice treated with the vehi-
!
&
*
^
cle control ( ) and test compounds, meisoindigo ( ), 3-5 ( ), 5-4 ( ).
of control animals and those treated with 1 and 3-5 increased
steadily from day 0 to day 14, but showed greater variability
and less defined trends thereafter, due to the progressive
demise of test animals after day 14. In the case of animals
treated with 5-4, tumor volumes actually decreased from day 0
to day 10 before resuming a gradual increase. When changes
in tumor volumes were analyzed for statistical difference from
control animals by one-way ANOVA (Dunnett’s post hoc), only
animals treated with 5-4 between day 6 (p=0.026) to day 14
(p=0.036) had tumors that were statistically smaller than
those of control animals over the same period. It was good to
note that of the six mice treated with 5-4, the tumors of two
mice completely regressed and remained as such until day 28.
In a single case, the tumor shrunk initially for the first 10 days,
but increased thereafter. The tumors of the remain-
In the case of 5-4, which was investigated at 1–15 mm, dis-
ruption of the cell cycle at the G1 phase (increase in G1, de-
crease in G2) was evident only at the highest concentration
(15 mm). There was a concurrent increase in the proportion of
cells in the sub-G1 phase. Thus, in spite of their structural simi-
larity, all three compounds had very different effects on the
cell cycle, namely G2 arrest for 3-5, G1 arrest in the case of 5-4
and no clear cut effects for 1.
Effects of 3-5 and 5-4 on apoptosis of K562 cells
Meisoindigo 1 is known to induce apoptosis in leukemic
cells,[14] and it was of interest to determine if 3-5 and 5-4 have
similar effects. Hence, apoptosis was investigated by double
ing animals grew steadily up to day 28, but did not
Table 9. FACS analyses of the various phases of the cell cycle of leukemic K562 cells
on exposure to meisoindigo, 3-5 and 5-4.
exceed the tumor diameter of 1.5 cm, which was the
threshold for euthanasia.
c [mm]
Sub-G1 [%][a]
G1 [%][a]
S [%][a]
G2 [%][a]
Effects of 3-5 and 5-4 on the cell cycle of K562
cells
Control[b]
1
–
1
5
10
15
20
0.1
0.5
1
5
10
1
5
10
15
1.80ꢀ0.46
1.42ꢀ0.25
1.91ꢀ0.71
2.56ꢀ0.39
1.89ꢀ0.15
2.23ꢀ0.20
1.38ꢀ0.11
2.15ꢀ0.24
4.97ꢀ0.20[c]
14.11ꢀ0.31[c]
9.87ꢀ0.44[c]
1.26ꢀ0.02
1.24ꢀ0.10
2.18ꢀ0.11
9.76ꢀ0.46[c]
40.02ꢀ4.49
43.92ꢀ1.37
45.71ꢀ2.29
40.16ꢀ0.54
41.64ꢀ0.82
39.09ꢀ0.61
41.90ꢀ0.81
32.44ꢀ0.90[c]
19.88ꢀ0.71[c]
9.99ꢀ0.11[c]
4.17ꢀ0.20[c]
44.34ꢀ1.06
44.07ꢀ1.30
42.41ꢀ0.55
45.30ꢀ0.23[c]
29.81ꢀ1.09
29.37ꢀ0.22
29.63ꢀ0.92
30.27ꢀ0.20
30.82ꢀ0.16
31.59ꢀ0.18[c]
29.61ꢀ0.50
25.90ꢀ0.62[c]
35.76ꢀ0.82[c]
9.03ꢀ0.19[c]
6.18ꢀ0.23[c]
29.37ꢀ0.68
28.79ꢀ0.57
28.69ꢀ0.18
22.40ꢀ0.29[c]
26.76ꢀ2.29
26.08ꢀ1.29
23.46ꢀ2.58
27.52ꢀ0.62
22.44ꢀ0.65[c]
23.70ꢀ0.44[c]
24.11ꢀ0.91
35.68ꢀ1.61[c]
34.57ꢀ1.31[c]
57.24ꢀ0.22[c]
71.31ꢀ0.82[c]
25.41ꢀ1.37
26.43ꢀ1.62
27.16ꢀ0.77
18.19ꢀ0.59[c]
Having established that 3-5 and 5-4 are promising
hits with good potencies and aqueous solubilities,
we proceeded to examine their effects on the cell
cycle of K562 cells. The intent is to derive a better
understanding of the mechanistic basis underlying
their antiproliferative activities. Cell cycle arrest was
investigated by fluorescence activated cell sorter
analysis (FACS) using flow cytometry. K562 cells were
incubated with the test compound for 24 h, after
which cellular DNA content was analyzed to deter-
mine the proportion of cells in the different phases
of the cell cycle. Each compound was tested over
a range of concentrations that spanned its IC50, K562
3-5
5-4
[a] Values represent the mean ꢀSD for nꢁ3 determinations. [b] Control cells treated
with media (0.4% v/v DMSO) in the absence of test compound. [c] Significantly differ-
ent (p <0.05) from control (one-way ANOVA, Dunnett’s post hoc).
784
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2012, 7, 777 – 791

Anticancer Activity of Functionalized Isoindigos
MED
staining the K562 cells with annexin V and propidium iodide at
different concentrations of the test compounds. Cells that
were positively stained by annexin V, but not propidium
iodide, were deemed to be apoptotic, while those that were
positively stained by both annexin V and propidium iodide
were likely to be necrotic. Representative plots are given in the
Supporting Information. The distribution of treated cells in
each quadrant of the plot was monitored for significant varia-
tions from the control untreated cells (Table 10). The results
showed that both 3-5 and 5-4 increased the proportion of
apoptotic cells, and interestingly, at concentrations that closely
corresponded to those required to disrupt the cell cycle,
namely at the IC50 value of 3-5 and at a higher concentration
of 15 mm for 5-4. 1 also increased the proportion of apoptotic
cells but only at higher concentrations (15 mm and 20 mm).
tribute significantly to the activity on K562 cells. An intact 3,3’-
double bond to ensure planarity and a methyl substitution at
only one of the lactam NH (N1) groups are key features de-
duced from our earlier report.[17] Here we show that replacing
the methyl group of 1 with phenalkyl side chains, notably,
phenpropyl, is a viable means of improving activity. Moreover,
with the exception of a selected ring substitution (1-4) and re-
stricted flexibility of the alkyl side chain (1-8), most modifica-
tions of the phenalkyl moiety were well tolerated, with many
of the resulting analogues (series 1) displaying greater activity
than 1. In order to negate the adverse increases in lipophilicity
associated with these modifications, the phenyl ring was re-
placed by several basic heterocycles. This strategy did indeed
address the physicochemical deficits of the phenalkyl ana-
logues, in spite of modest antiproliferative activities (5-4 to 5-
6), which did not match the submicromolar potencies of 3-5.
The other approach of inserting an azomethine nitrogen to im-
prove solubility was less successful, as the resulting 7’-azaisoin-
digos showed dramatic losses in activity and were possibly less
soluble than 1.
Table 10. Distribution of normal, apoptototic and necrotic K562 cells
treated with different concentrations of 1, 3-5 and 5-4.
c [mm]
Normal [%][a]
Apoptotic [%][a]
Necrotic [%][a]
We also examined the effect of ring substitutions on rings A/
B of the scaffold and found that substitutions on rings A and B
produced differential effects on activity, even with the same
substituent. There is a general preference for a substitution on
ring B, particularly at position 6’. At this position, the associa-
tion of 6’-methoxy with good activity is striking, but contin-
gent on the side chain present at the lactam nitrogen (N1).
Thus, activity was optimal when 6’-methoxy was paired with 1-
methyl (2-20) and 1-(p-methoxyphenethyl) (3-5), but not with
side chains present in 3-1 to 3-4 or the water soluble analogue
5-6.
Control[b]
1
93.77ꢀ1.67
93.25
93.36
3.42ꢀ0.73
3.70
3.48
1.92ꢀ0.62
2.42
2.51
1
5
10
15
20
0.1
0.5
1
5
10
1
93.58
3.77
2.07
83.50ꢀ0.88
82.90ꢀ0.47
95.56ꢀ0.03
87.27ꢀ1.31
90.26ꢀ0.12
82.34ꢀ0.14
78.18ꢀ0.05
92.72
10.15ꢀ0.32[c]
11.27ꢀ0.31[c]
2.74ꢀ0.10
8.91ꢀ0.79[c]
7.72ꢀ0.19[c]
14.03ꢀ0.27[c]
16.76ꢀ0.08[c]
3.74
5.11ꢀ0.42[c]
5.04ꢀ0.33[c]
1.28ꢀ0.62
2.77ꢀ1.06
1.62ꢀ0.08
3.05ꢀ0.26[c]
4.57ꢀ0.14[c]
2.63
3-5
5-4
5
93.29
3.43
2.42
10
15
94.02
79.84ꢀ0.87
3.73
1.59
It is encouraging to note that the antiproliferative activities
of the potent functionalized isoindigos (1-6, 1-10, 2-20, 3-5,
5-4) were not limited to K562 alone, but were also observed at
even lower IC50 values (except for 2-20, 5-4) on another leuke-
mic cell line, NB4. Another positive finding is the selective ac-
tivity on leukemic cells vis-ꢂ-vis the non-malignant human
lung fibroblast cells, IMR90. More modest antiproliferative ac-
tivities were observed against malignant cells derived from
solid tumors, but even then, most of the synthesized com-
pounds maintained significantly greater potencies than 1. No-
tably, 5-4 was the only compound that was consistently more
potent than 1 on the solid tumor cell lines HuH7, HCT116 and
RCC786.
14.06ꢀ1.24[c]
5.41ꢀ0.14[c]
[a] Proportion of normal, apoptotic and necrotic cells were deduced from
the FACS analysis of cell populations in the lower left, lower right and
upper right quandrants, respectively. [b] Control cells treated with media
(0.4% v/v DMSO) in the absence of test compound, n=5 determinations.
[c] Significantly different (p <0.05) from control (one-way ANOVA, Dun-
nett’s post hoc). Values represent the mean ꢀSD for nꢁ3 determina-
tions. For n=2 determinations, no SD was calculated.
Discussion
Meisoindigo 1 and its glycosylated analogue, natura, are argua-
bly the most widely reported isoindigos with anticancer prop-
erties. 1 is used in China as an efficacious and well-tolerated
substitute for indirubin in the treatment of CML,[9] while natura
was designed to improve the bioavailability and potency of
1.[16,19] Both compounds reportedly induce apoptosis, inhibit
cyclin-dependent kinases and arrest tumor growth in mice
transplanted with Walker 256 cancer cells.[16,19] The aim of the
present study is to investigate the SAR of 1 with reference to
its antiproliferative activity, and to use these findings to identi-
fy key features on the scaffold that could be modified to en-
hance the physicochemical profile without overt loss of activi-
ty. To this end, we have obtained a reasonably comprehensive
understanding of the critical features of the scaffold that con-
Previous approaches at improving the water solubility of the
isoindigo scaffold have focused almost exclusively on substitut-
ing the lactam NH with sugar moieties. Here we showed that
significant improvements in aqueous solubility and a concur-
rent decrease in aggregation potential can be achieved by ap-
pending solubilizing basic heterocycles to the NH moiety.
Through this approach, we were able to identify analogues
(5-4, 5-5, 5-6) that achieve drug-like profiles, as well as antipro-
liferative activities exceeding that of 1. In fact, it is tempting to
attribute the in vivo efficacy of 5-4 to its improved physico-
chemical profile and to further propose that the modest po-
tency of 5-4 has been compensated or even overcome by its
favorable solubility. In comparison, 3-5 with greater potency
ChemMedChem 2012, 7, 777 – 791
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
785

M. L. Go et al.
MED
tography was carried out using Silica Gel 60 (230–400 mesh,
Merck). H NMR spectra were collected on a Spectrospin 300 Ultra-
but poorer solubility failed to demonstrate activity, in spite of
being administered at a dose (10 mm) that was almost 20 times
its IC50 value. We were not able to demonstrate the in vivo effi-
cacy of 1 on the K562 xenografts. Such a study has not been
reported, although 1 is reputed to be effective on xenografts
induced by HT29 colon cancer cells[33] and HL60 acute myeloid
leukemic cells,[13,14] but at higher doses and at greater frequen-
cies. Criticism might be levied at the intratumoral route em-
ployed, but this was necessitated by the absence of pharmaco-
kinetic data of the test compounds. Nonetheless, care was
taken to ensure that the injection was made directly to the
palpable tumor and that no damage was afflicted to neighbor-
ing tissues.
1
shield (Bruker, Billerca, MA, USA) and referenced to residual solvent
peaks (CDCl3 at d=7.26, [D6]DMSO at d=2.50) as internal stand-
ards. 13C NMR spectra (75 MHz) were determined on the same in-
strument and are reported in ppm (d) relative to residual CDCl3
(d=77.0) and [D6]DMSO (d=39.5). Coupling constants (J) are re-
ported in Hertz (Hz). Nominal mass spectra were captured on
a LcQ mass spectrometer (Thermo Finnigan MAT GmbH, Bremen,
Germany) with chemical ionization (APCI) and m/z values for the
molecular ion are reported. High resolution accurate mass spectra
were analyzed on a micrOTOF-QII mass spectrometer (Bruker, Biller-
ca, MA, USA) by electrospray ionization (ESI). Microwave-assisted
reactions were carried out on a Biotageꢃ Initiator microwave syn-
thesizer (Biotage AB, Uppsala, Sweden). Purity of final compounds
were verified by either combustion analysis (C, H) (Vario Micro
Cube, Elementar, Hanau, Germany) or by reverse phase HPLC on
two different eluting systems (isocratic mode). Analytical and
purity data of final compounds are provided in the Supporting In-
formation.
Meisoindigo 1 has been variously reported to arrest the cell
[13,14]
cycle at either the G1
or G2 phase,[33] depending on the
type of cells investigated. It has also been noted to induce
apoptosis in multiple leukemic cell lines, possibly by down reg-
ulating antiapoptotic Bcl-2, and upregulating pro-apoptotic
(Bak, Bax) and cell cycle-related proteins (p21,p27).[14] Here, we
found that 1 (20 mm) decreased the proportion of cells in the
G2 phase but did not increase those in the G1 phase, which is
characteristic of G1 arrest. Apoptosis was observed at a similar
concentration range (10–20 mm). Strikingly different profiles
were observed for 3-5 (G2 arrest at IC50) and 5-4 (G1 arrest at
concentrations>IC50), indicating that relatively minor modifica-
tions can significantly alter the mechanistic pathways involved
in growth inhibition. The potent G2 arrest associated with 3-5
is interesting, as it has been shown not to inhibit CDK2, which
is involved in progression through the G2 phase.[17] Thus, the
disruption of G2 progression by 3-5 could be caused by other
factors, such as an increase in the production of endogenous
cell cycle inhibitors, down regulation of cyclins required for
CDK activation or induction of arylhydrocarbon receptor (AhR)
signaling pathways leading to cell cycle arrest. The latter is
a plausible alternative, as isoindigo has been reported to be an
AhR agonist,[34] and the antiproliferative activity of 1-methylat-
ed indirubins has been linked to the activation of AhR signal-
ing pathways.[35] At this stage of our investigations, the mecha-
nistic basis underlying the ability of 5-4 to induce apoptosis
and cause G1 arrest in K562 cells remains to be resolved.
General method for the preparation of 1-alkyl isatin intermedi-
ates, 2–4, 6, 7, and 11. Isatin (1 equiv), anhyd K2CO3 (1.2 equiv)
and anhyd DMF (2–4 mL) were added to a sealed microwave
vessel (5 mL) and stirred at RT for 15 min, followed by injection of
the alkyl halide (1.1 equiv). Stirring was continued for another
5 min before irradiation in a microwave reactor at 1508C for 15–
30 min. After H2O was added to the mixture, it was either filtered
to remove the crude product, which was then rinsed with deion-
ized H2O, or extracted with EtOAc/CH2Cl2. For the latter, the organic
layer was dried with anhyd Na2SO4 and filtered, and the organic
solvent was removed in vacuo. Purification by column chromatog-
raphy followed by recrystallization with EtOH gave the desired
product.
General method for the preparation of 1-alkyl isatin intermedi-
ates, 5, 8, and 9). Isatin (1 equiv), anhyd CaH2 (1.2 equiv) and
anhyd DMF (2–4 mL) were stirred in a flask for 15 min at RT. The
alkyl halide (1.1 equiv) was added and the mixture heated at 808C
for 2–4 h. Extraction was performed using EtOAc or CH2Cl2 with
brine and purified as described for compound 2.
(2-Chloroethoxy)benzene (10). Phenol (10 mmol), anhyd K2CO3
(15 mmol) and MeOH (8 mL) were stirred in a sealed vessel (20 mL)
at RT for 15 min and then heated in a microwave reactor to 1408C
for 30 min. The mixture was extracted with CH2Cl2 and the organic
layer was washed with an aq K2CO3 solution, dried (anhyd Na2SO4)
and filtered. The solvent was removed in vacuo and the residue
purified by column chromatography (isocratic, 100% n-hexane) to
In conclusion, our findings have provided a better under-
standing of the SAR of the isoindigo scaffold with regards to
its antiproliferative activity and have shown that the physico-
chemical deficits of the scaffold are surmountable by systemat-
ic modifications of structure and could be achieved without
overt loss of growth inhibitory activity.
1
give 10 as a colorless oil (674 mg, 43%): H NMR (300 MHz, CDCl3):
d=7.21 (t, J=8.1 Hz, 2H), 6.89 (t, J=7.5 Hz, 2H), 6.83 (d, J=8.4 Hz,
2H), 4.13 (t, J=5.7 Hz, 2H), 3.71 ppm (t, J=5.7 Hz, 2H); 13C NMR
(75 MHz, [D6]DMSO): d=158.1, 129.5, 121.4, 114.7, 77.4, 77.0, 76.6,
67.9, 41.9 ppm.
General method for the preparation of substituted 1-methylisa-
tins, 12–18 and 20–22. The 1-methylated isatins were prepared by
reacting the corresponding substituted isatin with methyl iodide
as described earlier.[17]
Experimental Section
Chemistry
General details for chemical synthesis. Reagents were obtained
from commercial suppliers and used without further purification.
Melting points were determined on a Gallenkamp melting point
apparatus (Weiss–Gallenkamp, Loughborough, UK) and reported as
uncorrected values. Reactions were routinely monitored by thin
layer chromatography (TLC) using pre-coated plates (Silica Gel 60,
F254, Merck) and visualized with UV light. Flash column chroma-
General method for the preparation of substituted 1-methyl-2-
oxindoles, 19 and 23. To a solution of substituted 2-oxindole
(2 mmol) in xylene (4 mL) was added NaH (2 mmol), followed by
dimethylsulfate (2 mmol). The solution was allowed to stir at RT for
15 min, and then at 508C for 3 h. The mixture was concentrated in
vacuo and extracted with EtOAc/brine. The organic layer was dried
786
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2012, 7, 777 – 791

Anticancer Activity of Functionalized Isoindigos
MED
(anhyd Na2SO4), filtered, concentrated, and purified by column
chromatography (isocratic, hexane/EtOAc 4:1) to give the desired
product.
the mixture was extracted with CH2Cl2/brine. The organic fraction
was dried (anhyd Na2SO4), filtered, concentrated and purified by
column chromatography (n-hexane/EtOAc 3:1) to give the desired
product 41 as orange crystals (192 mg, 74%): 1H NMR (300 MHz,
CDCl3): d=7.63–7.59 (m, 2H), 7.12 (t, J=7.8 Hz, 1H), 6.92 (d, J=
8.1 Hz, 1H), 3.86 (t, J=6.6 Hz, 2H), 3.66 (t, J=4.5 Hz, 4H), 2.64 (t,
J=6.6 Hz, 2H), 2.53 ppm (t, J=4.5 Hz, 4H); 13C NMR (75 MHz,
[D6]DMSO): d=183.6, 158.2, 150.8, 138.4, 124.5, 123.3, 117.4, 111.0,
66.2, 54.7, 53.2, 37.0 ppm; MS(APCI+): m/z [M+H]+ calcd for
C14H16N2O3: 261.12, found: 260.9.
6-Aminoindolin-2-one (24). The synthesis was performed as re-
ported by Khanwelkar et al.[36] Briefly, a solution of 2,4-dinitrophe-
nylacetic acid (10 mmol) in MeOH (100 mL) was hydrogenated in
a Parr hydrogenator with 10% Pd/C (500 mg) at RT for 2.5 h. After
MeOH (100 mL) was added, the mixture was sonicated at 508C for
5 min and filtered through celite under reduced pressure. The fil-
trate was concentrated to ca. 30 mL, followed by the addition of
2.5n HCl (6 mL) and held at reflux for 16 h. The solution was ad-
justed to pH 10 with K2CO3 and then extracted with EtOAc/brine.
The organic fraction was dried (anhyd Na2SO4), filtered and concen-
trated in vacuo to give desired product 24 as a pale brown
powder (53%, lit. 65%), which darkened rapidly on exposure to air
and light.
1-(2-Bromoethyl)indoline-2,3-dione (42). Isatin (10 mmol) and
anhyd K2CO3 (20 mmol) were dissolved in anhyd DMF (12 mL) and
stirred at RT for 15 min. 1,2-Dibromoethane (12 mmol) was added
to the solution, and stirring was continued at RT for 16 h. The sus-
pension was filtered under reduced pressure and the residue
rinsed with DMF to give the desired compound. The filtrate was
concentrated in vacuo and H2O (20 mL) was added. A precipitate
was obtained, which was removed by filtration under reduced
pressure, washed with H2O, combined with the earlier obtained
material and dried in an oven at 1008C. 42 was obtained as a red
solid (5.44 g, 67%): 1H NMR (300 MHz, CDCl3): d=7.64–7.53 (m,
2H), 7.14 (t, J=7.5 Hz, 1H), 7.00 (d, J=8.1 Hz, 1H), 4.14 (t, J=
6.6 Hz, 2H), 3.61 ppm (t, J=6.6 Hz, 2H); 13C NMR (75 MHz, CDCl3):
d=182.6, 158.2, 150.4, 138.4, 125.7, 124.0, 117.6, 110.2, 41.9,
27.0 ppm.
6-(Dimethylamino)indolin-2-one (25). Sodium cyanoborohydride
(4.5 mmol) was added to a solution of 24 (2 mmol) in glacial acetic
acid (3 mL) followed by formaldehyde (4.5 mmol), and the mixture
was stirred at RT for 24 h. The suspension was concentrated in
vacuo and extracted with EtOAc/brine. The organic layer was dried
(anhyd Na2SO4), filtered, concentrated in vacuo and purified by
column chromatography (CH2Cl2/EtOH 100:0 to 98:2) to give the
1
desired product 25 as an off-white powder (296 mg, 84%): H NMR
(300 MHz, [D6]DMSO): d=10.17 (s, 1H), 6.97 (d, J=8.1 Hz, 1H), 6.26
1
2
(dd, J=8.1 Hz, J=2.1 Hz, 1H), 6.19 (d, J=2.1 Hz, 1H), 3.31 (s, 2H),
2.85 ppm (s, 6H); 13C NMR (75 MHz, CDCl3): d=177.1, 150.4, 144.5,
124.5, 112.7, 105.2, 94.3, 40.3, 35.0 ppm; MS(APCI+): m/z [M+H]+
calcd for C10H12N2O: 177.09, found: 177.1
(E)-1-(2-Bromoethyl)-[3,3’-biindolinylidene]-2,2’-dione (43) and
(E)-1-(2-bromoethyl)-6’-methoxy-[3,3’-biindolinylidene]-2,2’-
dione (44). Compounds 43 and 44 were prepared by reacting 42
(2 mmol) with 2-oxindole (2 mmol) or 6-methoxy-2-oxindole
(2 mmol), respectively, in acetic acid at reflux as described for com-
N-(2-Oxoindolin-6-yl)acetamide (26). Concd H2SO4 (2 drops) was
added to a solution of 24 (2 mmol) in acetic anhydride (4 mmol)
and stirred at RT for 2 h, followed by the removal of the solvent in
vacuo. The residue was extracted with EtOAc/brine, and the organ-
ic fraction was dried (anhyd Na2SO4), filtered and concentrated to
give the desired product 26 as light brown solid (262 mg, 69%):
1H NMR (300 MHz, [D6]DMSO): d=10.33 (s, 1H), 9.88 (s, 1H), 7.35 (s,
1H), 7.08 (d, J=7.8 Hz, 1H), 6.97 (d, J=8.1 Hz, 1H), 3.23 (s, 2H),
2.02 ppm (s, 3H); MS(APCI+): m/z [M+H]+ calcd for C10H10N2O2:
191.07, found: 191.1.
1
pound 2-1. 43 (254 mg, 35%): H NMR (300 MHz, CDCl3): d=9.13
(d, J=7.8 Hz, 1H), 7.67 (s, 1H), 7.38–7.30 (m, 2H), 7.11–7.03 (m,
2H), 6.87 (d, J=7.8 Hz, 1H), 6.81 (d, J=7.8 Hz, 1H), 4.20 (t, J=
7.2 Hz, 2H), 3.60 ppm (t, J=7.2 Hz, 2H). 44 (55 mg, 7%): 1H NMR
(300 MHz, CDCl3): d=9.12 (d, J=9.0 Hz, 1H), 9.07 (d, J=8.1 Hz,
1H), 7.98 (s, 1H), 7.34 (t, J=7.8 Hz, 1H), 7.06 (t, J=7.8 Hz, 1H), 6.86
1
2
(d, J=7.8 Hz, 1H), 6.54 (dd, J=8.7 Hz, J=1.8 Hz, 1H), 6.85 (d, J=
1.8 Hz, 1H), 4.20 (t, J=7.2 Hz, 2H), 3.86 (s, 3H), 3.59 ppm (t, J=
6.9 Hz, 2H).
3,3-Dibromo-1H-pyrrolo[2,3-b]pyridin-2(3H)-one (45). The syn-
thesis was performed as reported by Marfat et al.[24] Briefly, pyridini-
um bromide perbromide (PBPB, 4 equiv) was added to a stirred so-
lution of 7-azaindole (10 mmol) dissolved in tert-butanol (70 mL) in
small portions over 6 h at RT. The solvent was removed in vacuo,
and the residue extracted with EtOAc/brine. The organic fraction
was dried (anhyd Na2SO4), filtered and concentrated to give the
crude residue, which was thrn dissolved in CH2Cl2, sonicated for
2 min, and filtered under reduced pressure. Recrystallization in tol-
uene gave product 45 as a light brown solid which darkens on
General method for the preparation of substituted 1-(4-meth-
oxyphenethyl)isatins for series 4 (27–37). The method described
for 5 was followed. Briefly, the substituted isatin was reacted with
p-methoxyphenethyl bromide in the presence of CaH2 in DMF by
holding at reflux (808C) for 2–4 h.
General method for the preparation of chloroethylmorpholine
(38), chloroethylthiomorpholine (39) and chloroethylpiperidine
(40). The basic heterocycle (10 mmol morpholine, thiomorpholine
or piperidine, Et3N (12 mmol) and 2-bromoethanol (12 mmol) were
dissolved in toluene (5 mL) and stirred at 808C for 4 h. The reaction
mixture was concentrated in vacuo, and CH2Cl2 (5 mL) was added.
The solution was cooled to 08C and SOCl2 (30 mmol) was added
dropwise. After effervescence had subsided, the mixture was
heated at 808C for 2 h. The crude mixture was adjusted to pH 10
with K2CO3 and extracted using CH2Cl2/brine. The product was ob-
tained by column chromatography as an oil which solidified on
standing.
1
standing (2.34 g, 80%, lit. 86%): H NMR (300 MHz, [D6]DMSO): d=
12.0 (brs, 1H), 8.21 (d, J=5.1 Hz, 1H), 8.00 (d, J=7.5 Hz, 1H),
1
7.18 ppm (dd, J=7.5 Hz, 2J=5.1 Hz, 1H).
1H-Pyrrolo[2,3-b]pyridin-2(3H)-one (46). The synthesis was per-
formed as reported by Marfat et al.[24] 10% Pd/C (200 mg) was
added to a solution of 45 (1 mmol) in EtOH. Hydrogenation was
performed at 50 psi H2 and RT for 4 h in a Parr hydrogenator. The
mixture was then filtered over celite, the solvent was removed in
vacuo and the residue was used for the proceeding reaction with-
1-(2-Morpholinoethyl)indoline-2,3-dione (41). Isatin (1 equiv) and
anhyd K2CO3 (1.1 equiv) were dissolved in anhyd DMF and stirred
at RT for 15 min. 38 (1.1 equiv) was added, and after the reaction
was continued for 4 h at 808C, the pH was adjusted to 8–9 and
1
out further purification (97 mg, 72%, lit. 75%): H NMR (300 MHz,
[D6]DMSO): d=11.42 (brs, 1H), 8.06 (d, J=5.4 Hz, 1H), 7.71 (d, J=
7.2 Hz, 1H), 7.05 (t, J=6.0 Hz, 1H), 3.60 ppm (s, 2H); 13C NMR
ChemMedChem 2012, 7, 777 – 791
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
787

M. L. Go et al.
MED
(75 MHz, CDCl3): d=175.7, 156.8, 142.7, 133.8, 122.0, 117.6,
35.1 ppm; MS(APCI+): m/z [M+H]+ calcd for C7H6N2O: 135.05,
found: 134.9.
satin (27–37) was reacted with 2-oxindole in acetic acid at reflux to
give the desired product on workup and purification.
General method for the preparation of series 4 compounds
(4-12 to 4-25). Compounds 4-12 to 4-25 were prepared following
the method described for compound 2-1. Briefly, 1-(p-methoxyphe-
nethyl)isatin (4) was reacted with the appropriate substituted 2-ox-
indole in acetic acid at reflux.
General method for the preparation of 1-alkylated isoindigos
1-1, 1-2, 1-5. The alkylated isatin (1 equiv 2, 3 or 6) and 2-oxindole
(1.1 equiv) were dissolved in acetic acid (1–3 mL) containing concd
HCl (2 drops) in a sealed vessel and stirred at RT for 5 min. The
mixture was heated to 2008C, 20–30 min in a microwave reactor-
and after cooling, concentrated in vacuo and extracted with
CH2Cl2/brine or EtOAc/brine. The organic fraction was dried (anhyd
Na2SO4), filtered, concentrated and purified by column chromatog-
raphy to give the desired product.
General method for the preparation of 5-1 and 5-5. Compound
41 (1 equiv) and 2-oxindole (1.1 equiv) were dissolved in acetic
acid (2–5mL) containing concd HCl (2 drops). The mixture was
held at reflux at 1408C for 12–16 h, then concentrated in vacuo,
adjusted to pH 10 and extracted with CH2Cl2/brine. The organic
fraction was dried (anhyd Na2SO4), filtered, concentrated in vacuo
and purified by column chromatography to give 5-1. 5-5 was pre-
pared in a similar manner by reacting 41 with 6’-methoxy-2-oxin-
dole instead of 2-oxindole.
General method for the preparation of 1-alkylated isoindigos
(1-3, 1-4, 1-7, 1-8, 1-10). The alkylated isatin (1 equiv 4, 5, 8, 9 or
11) and 2-oxindole (1.1 equiv) were dissolved in acetic acid (2–
5 mL) containing concd HCl (2 drops) and held at reflux for 12–
14 h. The mixture was purified as described for compound 1-1.
General method for the preparation of 5-2 and 5-3. Isoindigo
(1 equiv) and anhyd K2CO3 (1.1 equiv) were dissolved in anhyd
DMF (4–8 mL) and stirred at RT for 15 min. 39 or 40 (1.1 equiv) was
added to the mixture, which was then heated to 808C for 3 h. The
mixture was extracted with CH2Cl2/brine and adjusted to pH 14.
The organic fraction was dried (anhyd Na2SO4), filtered, concentrat-
ed and purified by column chromatography to give the desired
product.
General method for the preparation of 1-alkylated isoindigos
(1-6, 1-9). Isoindigo (1 equiv) and anhyd K2CO3 (1.1 equiv) were dis-
solved in anhyd DMF (2–4 mL) in a sealed microwave vessel and
stirred at RT for 15 min. The appropriate alkyl halide (1.1 equiv)
was added and the mixture heated in a microwave reactor to
1508C for 30 min. Extraction was performed using either CH2Cl2/
brine or EtAc/brine and the organic fraction was dried (anhyd
Na2SO4), filtered, concentrated in vacuo and purified by column
chromatography to give the desired product.
General procedure for synthesis of 5-4 and 5-6. A solution of N-
methylpiperazine (1 equiv) and Et3N (2 equiv) was stirred in anhyd
DMF (4 mL) at RT for 15 min. 43 or 44 (1 equiv) was added and the
mixture was heated to 808C for 16 h. The reaction mixture was ex-
tracted with CH2Cl2/brine at pH 10, and the organic layer dried
(anhyd Na2SO4), filtered and concentrated in vacuo. The product
was purifed by column chromatography.
General method for the preparation of series 2 compounds
(2-1 to 2-7 and 2-9 to 2-11). The appropriate substituted 1-methyl
isatin (1 equiv 12-18, 20-22) and 2-oxindole (1 equiv) were dis-
solved in acetic acid (2–5 mL) containing concd HCl (2 drops). The
mixture was held at reflux for 16 h. If the desired product precipi-
tated from the reaction mixture on cooling, it was removed by fil-
tration under reduced pressure, rinsed with cold acetic acid and re-
crystallized in EtOH. If no precipitate was observed, the solution
was concentrated in vacuo and extracted with CH2Cl2/brine or
EtOAc/brine. The organic fraction was dried (anhyd Na2SO4), fil-
tered, concentrated and purified by column chromatography to
give the desired product.
General method for the preparation of 5-7 and 5-8. 46 (1 equiv)
was reacted with 1-methylisatin or 1-(p-methoxyphenethyl)isatin
(4) (1 equiv) as described for compound 2-1. The mixture was held
at reflux in EtOH for 16 h. The precipitate was removed by filtration
under reduced pressure and rinsed with ice-cold EtOH to give the
desired product.
General method for the preparation series 2 compounds (2-8,
2-12). The appropriate substituted 1-methyl-2-oxindole (1 equiv
19, 23) and isatin (1 equiv) were reacted in acetic acid and worked
up as described for compound 2-1.
Biology
Cell lines. Human chronic myelogenous leukemic cells K562, human
normal lung fibroblast cells IMR90 and human colon carcinoma
cells HCT116 were purchased from American Type Culture Collec-
tion (Rockville, MD, USA). Human acute promyelocytic leukemia
cells NB4, human liver carcinoma cells HuH7 and human renal car-
cinoma cells RCC786 were gifts from Dr. Matiullah Khan (National
University of Singapore), Dr. Ho Han-Kiat (National University of
Singapore) and Dr. John Yuen (Singapore General Hospital) respec-
tively. Iscove’s modified Dulbecco’s medium (IMDM, catalog no.
I7633), high glucose Dulbecco modified eagle’s medium (DMEM,
catalog no. D1152) and Eagle’s minimum essential medium (EMEM,
catalog no. M0268) were purchased from Sigma–Aldrich in powder
form and reconstituted according to manufacturer’s instructions.
RPMI-1640 (Hyclone) and phosphate buffered saline (PBS) were
purchased from the National University Medical Institutes (NUMI,
Laboratory Supplies, National University of Singapore). Fetal bovine
serum (FBS) was obtained from Gibco (South America). 3-(4,5-Di-
methylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) was
purchased from Duchefa Biochemie (Haarlem, Netherlands). All
other chemicals were purchased from local suppliers and were of
molecular biology or cell culture grade.
General method for the preparation of series 2 compounds
(2-20, 2-24). The appropriate commercially available R3-substituted
2-oxindole (1 equiv) and 1-methylisatin (1 equiv) were reacted in
acetic acid and worked up as described for compound 2-1.
General method for the preparation of series 2 compounds
(2-13 to 2-19, 2-21 to 2-23). The appropriate commercially avail-
able R3-substituted isatin (1 equiv) and 1-methyl-2-oxindole
(1 equiv) were reacted in acetic acid and purified as described for
compound 2-1.
General method for synthesis of series 3 compounds (3-1 to
3-5). The method described for compound 1-1 was used. Briefly,
the alkylated isatin (7, 11, 2, 3 or 4) was reacted with 6-methoxy-2-
oxindole in acetic acid at reflux to give the desired product on
workup and purification.
General method for the preparation of series 4 compounds
(4-1 to 4-11). The method described for compound 1-3 was ap-
plied. Briefly, the appropriate substituted 1-(p-methoxyphenethyl)i-
788
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2012, 7, 777 – 791

Anticancer Activity of Functionalized Isoindigos
MED
Propagation of cells. K562 cells were cultured in IMDM supplement-
ed with FBS (10% v/v), penicillin G (100 mgLꢂ1) and streptomycin
(100 mgLꢂ1). The cells were sub-cultured by a ratio of 1:10 when
a cell density of 106 cellsmLꢂ1 was attained and passages 4–12
were used for viability experiments. NB4 and HCT116 cells were
grown in RPMI-1640 supplemented with FBS (10% v/v), penicillin G
(100 mgLꢂ1) and streptomycin (100 mgLꢂ1). NB4 cells were sub-cul-
tured at a split ratio of 1:8 when the cell density reached 106
cellsmLꢂ1. Only cells of passages 8–14 were used for experiments.
HCT116 cells were subcultured (ratio of 1:5) when 90% confluent
and used within passages 9–18 for experiments. HuH7 cells were
propagated in high glucose DMEM supplemented with FBS (10%
v/v), sodium pyruvate (1 mm), penicillin G (100 mgLꢂ1) and strepto-
mycin (100 mgLꢂ1). The cells were subcultured (ratio of 1:4) when
90% confluent and used within passages 8–16 for experiments.
RCC786 cells were cultured in high glucose DMEM supplemented
with FBS (10% v/v), penicillin G (100 mgLꢂ1) and streptomycin
(100 mgLꢂ1). The cells were subcultured (ratio of 1:4) when 90%
confluent and used within passages 9–16 for experiments. IMR90
cells were grown in EMEM supplemented with FBS (10% v/v), peni-
cillin G (100 mgLꢂ1) and streptomycin (100 mgLꢂ1). The cells were
subcultured (ratio of 1:3) when 90% confluent and used within
passages 6–13 for viability experiments.
a pre-determined wavelength and used to construct a calibration
curve for the test compound. Next, a stock solution of the test
compound in DMSO was prepared at a known concentration, dilut-
ed with universal buffer (pH 7.4), dispensed into wells in the Multi-
screenꢃ Solubility filter plate, and agitated for a period of time.
The suspension was filtered, the filtrate collected and diluted with
acetonitrile to give the same solvent composition used to prepare
the calibration solutions. The absorbance of the diluted filtrate was
read at the predetermined wavelength and the concentration of
the filtrate (equivalent to the solubility of the test compound) was
determined from the calibration curve.
Assessment of aggregation tendency by dynamic light scattering.
Stock solutions (10 mm) of test compounds were prepared in
DMSO and serially diluted with potassium phosphate buffer (5 mm,
pH 7.4, pre-filtered before use) to give final concentrations of 0.5,
1, 5 and 10 mm. The concentration of DMSO was maintained at 1%
v/v in the final solutions. Measurements were carried out using
a Zetasizer Nano ZS system (Malvern Instruments, Worcestershire,
UK) equipped with a 4 mW He-Ne laser at 633 nm. The detector
angle was 908. Three or more independent measurements of de-
rived count rates (in kilocounts per second, kcps) were made for
each concentration of test compound, using at least two separate-
ly prepared solutions. Results are represented as mean values ꢀ
standard deviation (SD). Data collection and quality assessment
were carried out using the software supplied with the instrument.
MTT assay for determination of cell viability. The MTT assay for K562
cells was described earlier.[17] The same method was applied to
NB4 cells except that cells were seeded at a lower density of 4000
cells per well in 96-well plates. For the non-adherent cells (HuH7,
HCT116, RCC786, IMR90), the MTT assay was modified as follows:
The cells in their respective media were seeded at densities of
6000 (HuH7), 2500 (HCT116), 10000 (RCC-786) and 4000 (IMR90)
cells per well. Cells adhered to the base of the wells after a 16–
18 h incubation period at 378C and 5% CO2. The media was re-
moved by aspiration and replaced by fresh media containing the
test compound at varying concentrations. The final concentration
of DMSO vehicle was kept at 0.4% v/v per well for all cell lines.
After the cells were incubated for 72 h at 378C and 5% CO2, the
media was removed by aspiration, and the cells were gently rinsed
with PBS to remove residual compound. MTT (100 mL, 0.5 mgmLꢂ1
in PBS) was added to each well and incubated for 4 h at 378C. The
MTT solution was removed gently by aspiration and the formazan
crystals dissolved in DMSO (100 mL). Absorbances were measured
within 30 min at 595 nm on a microplate reader as described earli-
er. Cell viability at a given concentration is determined from Equa-
tion (1), where Atest compound =absorbance of wells with cells exposed
to test compound in media; Avehicle =absorbance of wells with cells
in media; and Ablank =absorbance of wells with DMSO only.
Cell cycle analysis. K562 cells were seeded at a density of 2.0ꢁ105
cellsmLꢂ1 in each well of a 6-well plate and treated with a known
concentration of test compound for 24 h at 378C and 5% CO2. The
cells were then harvested, centrifuged at 200 g for 10 min and
rinsed once with chilled PBS. 106 cells were suspended in 300 mL of
PBS and ice cold abs EtOH (700 uL) was added dropwise to the cell
suspension with intermittent vortexing between additions. The cell
sample was kept at 48C overnight and subsequently stored at
ꢂ208C prior to analysis. For analysis, the sample was centrifuged
at 200 g for 10 min to remove the fixing solution and the cell
pellet was rinsed with chilled PBS and resuspended in staining so-
lution (1 mL), which contained 0.2 mgmLꢂ1 RNase A (MP Biomedi-
cals, Illkirch Cedex, France) and 20 mgmLꢂ1 propidium iodide
(Sigma–Aldrich, catalog no. P4170) in PBS. The cell suspension was
then kept in the dark at RT for 15 min, after which it was immedi-
ately analyzed for cell cycle distribution on a Cytomation Cyan LX
instrument (Dako, Fort Collins, CO, USA) equipped with an Ar solid
state laser (488 nm) using Summit (Version 4.3) software. 20000
cells were read for each determination. Each test compound was
evaluated at four or five concentrations with at least three repeti-
tions per concentration.
Atest compoundꢂAblank
Apoptosis. The Annexin V-FITC Apoptosis Detection Kit (Sigma–
Aldrich, catalog no. APOAF) was used. K562 cells were seeded at
a density of 2.0ꢁ105 cellsmLꢂ1 in each well of a 6-well plate and
treated with a known concentration of test compound for 24 h (or
48 h) at 378C and 5% CO2. Cells were harvested, centrifuged
(200 g, 10 min), the cell pellet rinsed with chilled PBS and resus-
pended in 1ꢁ binding buffer at a density of 106 cellsmLꢂ1. Annex-
in V-FITC conjugate protein (5 mL) and propidium iodide solution
(10 mL) were added to an aliquot (500 mL) of cell suspension and
incubated at RT for 10 min in the dark, after which flow cytometric
analysis was immediately carried out on the Cytomation Cyan LX
instrument (Dako, Fort Collins, CO, USA) using the Summit (Version
4.8) software. Unstained treated cells, treated cells stained with an-
nexin V-FITC only and treated cells stained with propidium iodide
only were used for calibration and compensation. 20000 cells were
read for each sample determination. Each test compound was eval-
ð1Þ
Viability ½%ꢃ ¼ ð
Þ ꢄ 100
AvehicleꢂAblank
Each concentration was tested at least three times on separate oc-
casions, using two different stock solutions, and on cells of differ-
ent passage numbers. The IC50 value was determined from the sig-
moidal curve obtained by plotting the percentage viability versus
logarithmic concentration of the test compound using GraphPad
Prism 5.0 (San Diego, CA, USA).
Determination of aqueous solubility: Determination of aqueous solu-
bility was carried out on Multiscreenꢃ Solubility filter plates
(MSSLBPC10, Millipore Billerca, MA, USA). The manufacturer’s pro-
tocol was followed. Briefly, various concentrations of the test com-
pound were prepared in universal buffer (pH 7.4)/acetonitrile/
DMSO. The UV absorbances of these solutions were obtained at
ChemMedChem 2012, 7, 777 – 791
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
789

M. L. Go et al.
MED
uated at five concentrations with at least three repetitions for
those concentrations that were associated with apoptosis.
same software. p-values <0.05 were considered significant for
both tests. Graphical representations were plotted with GraphPad
Prism 5.0 (San Diego, CA, USA).
In vivo studies. Experimental protocols for the in vivo study were
approved by the National University of Singapore (NUS) Institution-
al Animal Care and Use Committee (IACUC, 112/10) and were in
compliance with guidelines for the care and use of animals for sci-
entific purposes issued by the National Advisory Committee for
Laboratory Animal Research, Republic of Singapore. Balb/c female
nude athymic nude mice (16–20 g body weight, 5–8 weeks old)
were obtained from the Biological Resource Center (Singapore).
Animals were kept under controlled environmental conditions (19–
268C, relative humidity <70%, 12 h dark-light cycle, Center for Life
Science vivarium, NUS) and given free access to H2O and standard
feed. They were randomly assigned to one of four groups, namely
control animals to be treated with vehicle (1% DMSO v/v saline),
and treated animals earmarked to receive meisoindigo, 3-5 or 5-4.
Each group comprised 5–8 mice. To establish the xenograft, K562
cells (2.0ꢁ107 in 150 mL in saline) were injected subcutaneously in
the right dorsal flank of the animal previously sedated with gas-
eous isoflurane (5% v/v in O2 for induction of anaesthesia and 2–
3% v/v in O2 for maintenance). Tumor formation at the site was
monitored carefully and treatment was initiated when a palpable
tumor of approximately 150 mm3 was evident (day 0). The test
compound (10 mm in 100 mL saline containing 1% v/v DMSO) was
injected into the tumor on three separate occasions—day 0, day 3
and day 10. Animals in the control group were similarly injected
with saline containing 1% v/v DMSO on days 0, 3 and 10. Solutions
of test compounds were freshly prepared prior to injection and
sterilized by filtration using a 0.22 mm DMSO-safe Acrodiscꢃ syringe
filter (Pall Life Sciences, Ann Arbor, USA). The animals were moni-
tored daily for changes in body weight, tumor volume and signs of
distress. Tumor volume was calculated using the formula for the
volume of an ellipsoid [Equation (2)].
Acknowledgements
This work was made possible by funding to M.L.G. from the Na-
tional University of Singapore (RP 148000032112) and a National
University of Singapore research scholarship to X.K.W.
Keywords: antiproliferation · cancer · drug-like properties ·
isoindigos · structure–activity relationships
[1] J. Balfour-Paul in Indirubin, the Red Shade of Indigo, Life in Progress Edi-
tion, (Eds.: L. Meijer, N. Guyard, L. A. Skaltsounis, G. Eisenbrand), Roscoff,
France, 2006, pp. 1–6.
[2] R. Hoessel, S. Leclerc, J. A. Endicott, M. E. M. Nobel, A. Lawrie, P. Tunnah,
M. Leost, E. Damiens, D. Marie, D. Marko, E. Niederberger, W. C. Tang, G.
Eisenbrand, L. Meijer, Nat. Cell Biol. 1999, 1, 60–67.
[3] D. Marko, S. Schatzle, A. Friedel, A. Genzlinger, H. Zankl, L. Meijer, G. Ei-
senbrand, Br. J. Cancer 2001, 84, 283–289.
[4] S. Leclerc, M. Garnier, R. Hoessel, D. Marko, J. A. Bibb, G. L. Snyder, P.
Greengard, J. Biernat, Y. Z. Wu, E.-M. Mandelkow, G. Eisenbrand, L.
Meijer, J. Biol. Chem. 2001, 276, 251–260.
[5] M. J. Moon, S. K. Lee, J. W. Lee, W. K. Song, S. W. Kim, J. I. Kim, C. Cho,
S. J. Choi, Y. C. Kim, Bioorg. Med. Chem. 2006, 14, 237–246.
[6] E. Damiens, B. Baratte, D. Marie, G. Eisenbrand, L. Meijer, Oncogene
2001, 20, 3786–3797.
[7] V. Myrianthopoulos, P. Magiatis, Y. Ferandin, A. L. Skaltsounis, L. Meijer,
E. Mikros, J. Med. Chem. 2007, 50, 4027–4037.
[8] Z. J. Xiao, Y. S. Hao, B. C. Liu, L. S. Qian, Leukemia Lymphoma 2002, 43,
1763–1768.
[9] Z. J. Xiao, Y. S. Haoin Indirubin, the Red Shade of Indigo, Life in Progress
Edition, (Eds.: L. Meijer, N. Guyard, L. A. Skaltsounis, G. Eisenbrand),
Roscoff, France, 2006, pp. 203–208.
ðp ꢄ L ꢄ W2Þ
[10] Z. J. Xiao, L. S. Qian, B. C. Liu, Y. S. Hao, Br. J. Haematol. 2000, 111, 711–
712.
ð2Þ
V ¼
6
[11] F. Chen, L. Z. Li, D. X. Ma, S. X. Yan, J. Z. Sun, M. H. Zhang, C. Y. Ji, M.
Hou, Leukemia Res. 2010, 34, e75–e77.
[12] X. M. Liu, L. G. Wang, X. J. Ji, Chin. J. Pharmacol. Toxicol. 1991, 5, 204–
208.
[13] X. M. Liu, L. G. Wang, H. Y. Li, X. J. Ji, Biochem. Pharmacol. 1996, 51,
1545–1551.
[14] C. C. Lee, C. P. Lin, Y. L. Lee, G. C. Wang, Y. C. Cheng, H. E. Liu, Leukemia
Lymphoma 2010, 51, 897–905.
[15] M. Huang, Y. S. Lee, P. C. Ho, Rapid Commun. Mass Spectrom. 2010, 24,
729–741.
[16] L. Wang, X. Liu, R. Chen, (Natrogen Therapeutics, Inc., Whitefish Bay,
USA), WO 03/051900, 2003; [Chem. Abstr. 2003, 139, 47135].
[17] X. K. Wee, W. K. Yeo, B. Zhang, V. B. C. Tan, K. M. Lim, T. E. Tay, M. L. Go,
Bioorg. Med. Chem. 2009, 17, 7562–7571.
[18] Z. J. Xiao, Y. Wang, L. Lu, Z. J. Li, Z. Peng, Z. C. Han, Y. S. Hao, Leuk. Res.
2006, 30, 54–59.
[19] L. Wang (Flushing, USA), X. Liu (Flushing, USA), R. Chen (Foster City,
USA), US 6933315, 2005.
[20] M. Sassatelli, F. Bouchikhi, S. Messaoudi, F. Anizon, E. Debiton, C. Bartho-
meuf, M. Prudhomme, P. Moreau, Eur. J. Med. Chem. 2006, 41, 88–100.
[21] M. Sassatelli, F. Bouchikhi, B. Aboab, F. Anizon, D. Fabbro, M. Prud-
homme, P. Moreau, Anti-Cancer Drugs 2007, 18, 1069–1074.
[22] G. Karapetyan, K. Chakrabarty, M. Hein, P. Langer, ChemMedChem 2011,
6, 25–37.
[23] J. Ribas, K. Bettayeb, Y. Ferandin, M. Knockaert, X. Garrofe-Ochoa, F.
Totzke, C. Schachtele, J. Mester, P. Polychronopoulos, P. Magiatis, A. L.
Skaltsounis, J. Boix, L. Meijer, Oncogene 2006, 25, 6304–6318.
[24] A. Marfat, M. P. Carta, Tetrahedron Lett. 1987, 28, 4027–4030.
[25] Z. H. Wang, T. Wang, S. N. Yao, J. C. Chen, W. Y. Hua, Q. Z. Yao, Arch.
Pharm. 2010, 343, 160–166.
Length (L) and width (W) of the tumor were measured with elec-
tronic vernier calipers and reported to two decimal places. The
mice were monitored for 28 d after the first injection of test com-
pound or vehicle (day 0), after which they were euthanized with
CO2. Animals were also euthanized (before day 28) if they complied
with the following criteria: (i) tumor diameter >1.5 cm, (ii) tumor
ulceration, infection or inflammation, (iii) ruffled fur, hunched back
appearance or inappetent state, (iv) 10% or more body weight loss
over 24 h or 20% body weight loss relative to control group, or
(v) moribund or pre-moribund state. For the purpose of analyzing
the results using the Kaplan–Meier survival analysis, censored
events were recorded. These are events in which the period of sur-
vival of the animal could not be determined due to the following:
(i) animal remained alive after the study period (28 d), (ii) death of
animal due to unrelated causes such as sudden death or infection
and (iii) conditions which warrants euthanasia due to unrelated dis-
ease pathology or infection (e.g., pin-worm infestation in animal).
Statistical Analysis. Data was analyzed for statistical significance by
one-way ANOVA followed by Dunnett’s post-hoc test on SPSS 18.0
for Windows (SPPS Inc., Chicago, IL, USA). p <0.05 was taken as
the criterion for significance. Survival data were analyzed by
Kaplan–Meier (with log-rank test, pairwise comparison) on IBM
SPSS Statistics 19.0 (SPPS Inc., Chicago, IL, USA) following the pro-
tocol described by Chan and co-workers.[37] Tumor size data were
analyzed by One-way ANOVA with Dunnet post-hoc test on the
790
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2012, 7, 777 – 791

Anticancer Activity of Functionalized Isoindigos
MED
[26] S. L. McGovern, E. Caselli, N. Grigorieff, B. K. Shoichet, J. Med. Chem.
2002, 45, 1712–1722.
[27] S. L. McGovern, B. T. Helfand, B. Feng, B. K. Shoichet, J. Med. Chem.
2003, 46, 4265–4272.
[35] M. Knockaert, M. Blondel, S. Bach, M. Leost, C. Elbi, G. L. Hager, S. R.
Nagy, D. Han, M. Denison, M. Ffrench, X. Z. P. Ryan, P. Magiatis, P. Poly-
chronopoulos, P. Greengard, L. Skaltsounis, L. Meijer, Oncogene 2004,
23, 4400–4412.
[28] S. L. McGovern, B. K. Shoichet, J. Med. Chem. 2003, 46, 1478–1483.
[29] J. Seidler, S. L. McGovern, T. N. Doman, B. K. Shoichet, J. Med. Chem.
2003, 46, 4477–4486.
[30] K. E. D. Coan, B. K. Shoichet, J. Am. Chem. Soc. 2008, 130, 9606–9612.
[31] A. K. Doak, H. Wille, S. B. Prusiner, B. K. Shoichet, J. Med. Chem. 2010, 53,
4259–4265.
[36] R. R. Khanwelkar, G. S. Chen, H. C. Wang, C. W. Yu, C. H. Huang, O. Lee,
C. H. Chen, C. S. Hwang, C. H. Ko, N. T. Chou, M. W. Lin, L. M. Wang, Y. C.
Chen, T. H. Hseu, C. N. Chang, H. C. Hsu, H. C. Lin, Y. C. Shih, S. H. Chou,
H. W. Tseng, C. P. Liu, C. M. Tu, T. L. Hu, Y. J. Tsai, J. W. Chern, Bioorg. Med.
Chem. 2010, 18, 4674–4686.
[37] Y. H. Chan, Singapore Med. J. 2004, 45, 249–256.
[32] B. K. Shoichet, J. Med. Chem. 2006, 49, 7274–7277.
[33] M. X. Zuo, Y. Li, H. B. Wang, J. H. Zhou, H. Y. Li, H. Liu, H. Q. Xin, S.
Zhang, X. G. Chen, J. Chemother. 2008, 20, 728–733.
[34] S. Nishiumi, N. Yamamoto, R. Kodoi, I. Fukuda, K. I. Yoshida, H. Ashida,
Arch. Biochem. Biophys. 2008, 470, 187–199.
Received: January 10, 2012
Published online on March 13, 2012
ChemMedChem 2012, 7, 777 – 791
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
791