Functionalized 3-benzylidene-indolin-2-ones: Inducers of NAD(P)H-quinone oxidoreductase 1 (NQO1) with antiproliferative activity

Article abstract of DOI:10.1016/j.bmc.2008.12.052

Functionalized benzylidene-indolin-2-ones are widely associated with antiproliferative activity. The scaffold is not normally associated with chemoprevention in spite of the presence of a nitrogen-linked Michael acceptor moiety that may predispose members to induction of NQO1, a widely used biomarker of chemopreventive potential. To investigate this possibility, we have synthesized and evaluated a series of functionalized 3-benzylidene-indolin-2-ones for induction of NQO1 in murine Hepa1c1c7 cells as well as antiproliferative activity against two human cancer cell lines (MCF-7, HCT116). The benzylideneindolinones were found to be good inducers of NQO1 activity, with 85% of test compounds able to increase basal NQO1 activity by more than twofold at concentrations of ≤10 μM. By contrast, fewer compounds (11%) tested at the same concentration were able to reduce cell viability by more than 50%. Structure activity relationships showed that the nitrogen linked Michael acceptor moiety was an essential requirement for both activities. This common feature notwithstanding, substitution of the 3-benzylidene-indolin-2-one core structure affected NQO1 induction and antiproliferative activities in dissimilar ways, underscoring different structural requirements for these two activities. Nonetheless, promising compounds (10, 42, 45-48) were identified that combine selective induction of NQO1 with potent antiproliferative activity. A potential advantage of such agents would be the ability to provide added protection to normal cells by the up-regulation of NQO1 and other phase II enzymes while simultaneously targeting neoplastic cells.

Full text of DOI:10.1016/j.bmc.2008.12.052

Bioorganic & Medicinal Chemistry 17 (2009) 2077–2090
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Functionalized 3-benzylidene-indolin-2-ones: Inducers of NAD(P)H-quinone
oxidoreductase 1 (NQO1) with antiproliferative activity
*
Wei Zhang, Mei-Lin Go
Department of Pharmacy, Faculty of Science, National University of Singapore, 18 Science Drive 4, Singapore 117543, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history:
Functionalized benzylidene-indolin-2-ones are widely associated with antiproliferative activity. The scaf-
fold is not normally associated with chemoprevention in spite of the presence of a nitrogen-linked
Michael acceptor moiety that may predispose members to induction of NQO1, a widely used biomarker
of chemopreventive potential. To investigate this possibility, we have synthesized and evaluated a series
of functionalized 3-benzylidene-indolin-2-ones for induction of NQO1 in murine Hepa1c1c7 cells as well
as antiproliferative activity against two human cancer cell lines (MCF-7, HCT116). The benzylideneindoli-
nones were found to be good inducers of NQO1 activity, with 85% of test compounds able to increase
Received 29 October 2008
Accepted 22 December 2008
Available online 30 December 2008
Keywords:
Benzylideneindolinones
Induction of NADP(H): quinone
oxidoreductase I, antiproliferative activity
Nitrogen-linked Michael acceptor
basal NQO1 activity by more than twofold at concentrations of 610 lM. By contrast, fewer compounds
(11%) tested at the same concentration were able to reduce cell viability by more than 50%. Structure
activity relationships showed that the nitrogen linked Michael acceptor moiety was an essential require-
ment for both activities. This common feature notwithstanding, substitution of the 3-benzylidene-indo-
lin-2-one core structure affected NQO1 induction and antiproliferative activities in dissimilar ways,
underscoring different structural requirements for these two activities. Nonetheless, promising com-
pounds (10, 42, 45–48) were identified that combine selective induction of NQO1 with potent antiprolif-
erative activity. A potential advantage of such agents would be the ability to provide added protection to
normal cells by the up-regulation of NQO1 and other phase II enzymes while simultaneously targeting
neoplastic cells.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
of reactive and toxic semiquinone intermediates7,8 and an indirect
antioxidant effect which maintains endogenous antioxidants in
Carcinogenesis is a complex and protracted multistep process
involving distinct yet closely linked events of tumour initiation,
promotion and progression.1,2 The multi-stage nature of carcino-
genesis has led to the proposal that timely intervention with die-
tary or pharmaceutical agents may result in the prevention, delay
or reversal of the process. This concept is termed cancer chemo-
prevention and was espoused by Sporn et al. as a strategy for
reducing cancer mortality.3 Chemopreventive agents may block
or reverse the premalignant stages of initiation and promotion.
They may also retard the transformation of precancerous cells
to malignant cells. The mechanisms of chemoprevention are com-
plex and involve multiple signaling transduction pathways.4–6
The enzyme NAD(P)H: quinone oxidoreductase 1 (NQO1) is
widely recognized for its role in protecting cells against the toxic
effects of pro-oxidant chemicals. In the presence of oxidative
stress, expression of NQO1 is increased to provide timely protec-
tion of the cell by various mechanisms such as the reduction of
exogenous quinones to quinols, thus by-passing the formation
their reduced/active states.9–11 The induction of NQO1 is associ-
ated with the up-regulation of other protective Phase II en-
zymes.12,13 Since the measurement of NQO1 induction is a well
established procedure with medium-throughput capabilities, it
is widely employed as a biomarker to screen compounds for che-
mopreventive potential against cancer initiation.14–16 NQO1 is
also up-regulated in many tumours and recent reports that it
can stabilize the tumour suppressor protein p53 either by inhib-
iting its degradation or by a direct protein–protein interaction has
heightened interest in the enzyme.17,18 Observations that NQO1-
deficient mice had lower levels of p53 induction, reduced apopto-
sis, increased susceptibility to chemically induced tumours and
impaired NF-jB function have further underscored its importance
in carcinogenesis.19–21 With this background in mind, the focus of
our investigations is to develop compounds with chemopreven-
tive potential based on their ability to induce NQO1 activity in
cell-based assays.
The seminal work of Talalay and co-workers has led to the rec-
ognition of the Michael acceptor motif as a recurrent structural
entity in compounds with chemopreventive activity.22–24 Cinna-
mates, chalcones, curcuminoids, coumarins, phenylbutenones,
* Corresponding author. Tel.: +65 6516 2654; fax: +65 6779 1554.
E-mail address: phagoml@nus.edu.sg (M.-L. Go).
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.12.052

2078
W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
bis(benzylidene) cycloalkanones and triterpenoids are relevant
examples.22,23,25 In the present study, we propose that functional-
ized 3-benzylidene-indolin-2-ones should likewise be included as
compounds with chemopreventive potential. The presence of the
Michael acceptor motif makes the benzylideneindolinone scaffold
a likely candidate for such activity. One difference is that the exo-
cyclic double bond in this template is linked to an amide carbonyl,
and not the normal carbon-linked carbonyl found in previously
reported compounds (Fig. 1a). Thus the motif is better described
as a ‘nitrogen-linked Michael acceptor’. Since little is known of
the chemopreventive properties of such an entity, investigations
on 3-benzylidene-indolin-2-ones would provide timely insight.
Arylideneindolinones are widely reputed to be anti-tumour
agents.26,27 A representative member of this class is sunitinib
which is currently used in the clinics as a multi-targeting tyrosine
kinase inhibitor with antiangiogenic activity (Fig. 1b). If antitu-
mour benzylideneindolinones have chemopreventive properties,
this may mean that when deployed for their effects on cancer
cells, the same agent would confer cytoprotection to normal/
non-metastasized cells, an interesting proposition with potential
clinical benefits. It is thus appropriate to determine if there is
any overlap in antiproliferative and chemopreventive properties
in the benzylideneindolinone scaffold, and if so, to establish
structure–activity relationships for future lead optimization.
Our interest in the Michael acceptor-like motif in 3-benzyl-
iden-indolin-2-ones is also motivated by the mechanistic aspects
of NQO1 induction. Briefly, induction of NQO1 gene expression is
controlled by the arylhydrocarbon (AhR)/xenobiotic response ele-
ment (XRE) and the Kelch-like ECH-associated protein 1 (Keap1)/
nuclear factor erythroid 2-related factor (Nrf2)/antioxidant re-
sponse element (ARE) signaling pathways. Compounds that in-
duce NQO1 by stimulating the AhR/XRE pathway are described
as bi-functional inducing agents.28 By contrast, compounds that
transcriptionally activate the NQO1 gene through the ARE path-
way are designated mono-functional inducers. Bi-functional
inducers are AhR agonists and induce both Phase I (CYP1A1,
1A2, 1B1) and Phase II (including NQO1, GSTA2, UGT1A6) en-
zymes.29,30 Mono-functional inducers are disruptors of the
Keap1/Nrf2 complex. The liberated Nrf2 migrates to the nucleus
where it binds to the ARE enhancer regions to bring about tran-
scription of Phase II enzymes.24 It has been argued that induction
of Phase I enzymes is potentially deleterious to the cell because
CYP450 can activate polycyclic aromatic hydrocarbons to form
ultimate carcinogens. Thus mono-functional inducers are deemed
more desirable than bi-functional inducers.2,16 But this may be an
oversimplification of a complex issue, in view of emerging evi-
dence of cross-talk between the AhR and Nrf2 gene batteries.6,31
Compounds with Michael acceptor motifs are associated with
mono-functional induction of NQO1.22–24 On the other hand, in-
dole phytochemicals like indole-3-carbinol and brassinin are bi-
functional inducers (Fig. 1c).32,33 Since the 3-benzylideneindolin-
2-one template embraces both structural elements—the indole
ring and the embedded Michael acceptor moiety, this raises the
question as to whether bi- or mono-functional induction would
prevail in these compounds. To address these issues, we have
synthesized a series of functionalized 3-benzylidene-indolin-2-
ones and related compounds and evaluated their chemopreven-
tive potential in terms of the induction of NQO1 activity in He-
pa1c1c7 cells. The compounds were also evaluated for induction
of CYP1A1 activity on the same cell line to determine if there is
concurrent induction of CYP1A1 activity and if so, the extent to
which CYP1A1 is induced compared to NQO1. Finally, the com-
pounds were screened for antiproliferative activity on two cancer
cell lines with the aim of assessing the extent to which both
properties are found in the benzylideneindolinone scaffold.
2. Results
2.1. Synthesis of target compounds
The synthesized compounds were divided into four classes
based on their structural features. Class 1 comprised 3-benzyl-
ideneindolin-2-ones with different substituents on the benzyli-
dene ring B. The choice of substituents was guided by the Craig
Plot to ensure that groups of different lipophilicites and electronic
effects were adequately represented.34 Ring B was also replaced
by heterocycles (pyridine, indole) or naphthalene. Class 2 com-
pounds were substituted on both the indole ring A and the ben-
zylidene ring B. The groups on ring B were 30-OCH3, 30-OH-40-
OCH3 or 30-CF3. For each ring B substituent, three different ring
A groups were explored (F, Cl, OCH3). Class 3 compounds were
structural variants of Class 1 compounds and consisted of the fol-
lowing modifications: (i) N-methylation of the indolinone nitro-
gen (49, 50); (ii) reduction of the exocyclic double bond (51,
52) and (iii) shift in the positions of the carbonyl and exocyclic
double bond to give the regioisomeric 2-benzylidene-indolin-3-
ones (53, 54). Class 4 consisted of isoindigos (3,30-bisindoles)
which may be viewed as two indolin-2-ones joined by an exocy-
clic double bond and with two nitrogen-linked Michael acceptor
moieties in the scaffold. The structures of Class 1–4 compounds
are given in Table 1. Of the 61 compounds listed in this Table,
29 members (1–6, 8, 9, 11, 12, 16, 17, 20, 24, 26–28, 33–36, 42,
47, 51–53, 55, 57, 58) have been reported in the literature27,35–
(a)
(b)
N
O
NH
O
N
H
N
H
F
O
Nitrogenlinked
MichaelAcceptor
Motif
N
H
Sunitinib
S
(c)
(c)
48
(see Supplementary data for exact citations). Except for com-
HN
pound 55,49 none have been evaluated for chemopreventive activ-
ity based on the induction of NQO1.
CH2OH
S
The Class 1 and 2 compounds (1–48) were synthesized by a
Knoevenagel reaction between the aldehyde and oxindole. The N-
methylated 3-benzylidieneindolin-2-ones 49, 50 were obtained
in a similar way starting from N-methyloxindole. The condensation
was carried out in ethanol with piperidine as base catalyst by re-
flux or in a microwave reactor (Scheme 1). The 2-benzylindolin-
2-ones 51 and 52 were synthesized by catalytic hydrogenation of
the exocyclic double bond of the benzylidineindolin-2-ones 41
N
H
N
H
Brassinin
Indole-3-carbinol
Figure 1. (a) Nitrogen linked Michael acceptor moiety in benzylideneindolinones;
(b) structure of sunitinib; (c) structures of indole-3-carbinol and brassinin.

W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
2079
and 20, respectively. Reduction of 41 was accompanied by the loss
Table 1 (continued)
of the 6-chloro group to give 51.
(D) Class 4 compounds (55–61)
7'
H
N
6'
R'
5'
Table 1
O
B
Structures of compounds in Class 1–4
3'
4
(A) Class 1 compounds (1–38)
4'
3
5
A
O
R
6
7
N
H
H
N
6
7
O
A
5
Compound
R
R0
Compound
R
R0
4
55
56
57
58
H
H
H
H
H
59
60
61
6-CF3
6-CF3
6-CF3
H
5-F
5-Cl
6-F
5-Cl
6-Cl
R'
2'
B
4'
3'
a
Ring B is replaced by pyridine, indole or naphthalene.
Compound
R0
Compound
R0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
H
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34a
35a
36a
37a
38a
40-CH3
20-OCH3
40-C 2H5
20-F
30-OCH3
40-OCH3
R1
30-F
40-F
20,40-(OCH3)2
20,50-(OCH3)2
20,60-(OCH3)2
30,50-(OCH3)2
30,40,50-(OCH3)3
30-O-C6H5
20-OH
20,40-(F)2
30-Cl
CHO
40-Cl
a
R
O
20,40-(Cl)2
20-Cl, 40-F
20-CF3
+
O
R
N
R1
N
R2
R2
30-OH
30-CF3
20,50-(OH)2
30,50-(OH)2
20-OH, 40-OCH3
30-OH, 40-OCH3
30-OCH3, 40-OH
20-CH3
40-CF3
40-CN
Scheme 1. Synthetic pathway for Class 1 and 2 compounds (1–48). Reagents and
conditions: (a) piperidine (cat.), ethanol, microwave, 140 °C, 15 min or reflux, 18 h.
R2 = Methyl for 49 and 50. Substituted benzaldehydes replaced by appropriate
carbaldehydes for 34–38.
20-Pyridinyl
40-Pyridinyl
30-Indolyl
10-Napthalenyl
20-Napthalenyl
30-CH3
(B) Class 2 compounds (39–48)
O
O
H
OH
7
H
N
6
5
a
+
O
A
4
R"
R
N
H
R
N
Ac
R'
2'
Scheme 2. Synthetic pathway for substituted 2-benzylidene indolin-3-ones (53,
54). Reagents and conditions: (a) 50% aqueous ethanol, concentrated hydrochloric
acid, reflux 2 h.
B
4'
3'
Compound R0
R00
Compound R0
R00
39
40
41
42
43
30-OCH3
6-F
5-Cl
6-Cl
6-OCH3 47
48
44
45
46
30-OH, 40-OCH3 6-Cl
The 2-benzylidineindolin-3-ones (53, 54) were obtained by
an acid catalyzed aldol condensation (Scheme 2). The presence
of acid promoted the loss of the N-acetyl group from 1-(3-hy-
droxy-1H-indol-1-yl) ethanone and generated a carbocation in
30-OCH3
30-OCH3
30-OCH3
30-OH, 40-OCH3 6-OCH3
30-CF3
30-CF3
30-CF3
6-F
6-Cl
6-OCH3
30-OH, 40-OCH3 6-F
the reacting aldehyde. Reaction between the
a carbon of 1H-
(C) Class 3 compounds: 3-Benzylidene-1-methylindolin-2-ones (49,50), 3-
benzylindolin-2-ones (51,52) and 2-benzylidene-indolin-3-ones (53,54)
indol-3-ol and the carbocation proceeded with loss of water
to yield the desired product. The isoindigos (55–61) were ob-
tained by an acid catalyzed reaction between an oxindole and
isatin (Scheme 3). The mechanism is also an acid catalyzed al-
dol condensation and involved enolization of the oxindole, pro-
tonation of the non-amide carbonyl oxygen of the isatin,
followed by condensation and dehydration to give the desired
isoindigo.
CH3
B
H
N
B
R'
N
O
A
R'
A
A
O
N
H
O
R'
B
2'
2.2. Assignment of configuration of synthesized compounds
4'
3'
The synthesized compounds (except 51, 52) could exist as
either E or Z isomers due to the presence of the exocyclic double
bond. Thus it was necessary to determine if synthesis had given
49:
50:
R0 = 30-OCH3
R0 = 30-OH, 40-OCH3
51:
52:
R0 = 30-OCH3
R0 = 40-CH3
53:
54:
R0 = 30-OCH3
R0 = 30-OH, 40-OCH3

2080
W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
H
N
O
O
R2
a
O
+ R2
O
R1
N
H
N
H
R1
O
N
H
Scheme 3. Synthetic pathway for isoindigos (55–61). Reagents and conditions: (a) glacial acetic acid, concentrated hydrochloric acid, reflux 16 h.
rise to a single (predominant) isomer or to a mixture of isomers.
Analyses of the 1H and 13C NMR spectra of the Class 1–3 com-
pounds (except 51, 52) showed that they were obtained as a single
(or predominant) isomer.
Hence, the predominant isomer was assigned the
configuration.
E
1H chemical shifts could not be used to assign the configura-
tion of 7 as it had methoxy groups on both the 20 and 60 positions
of ring B. Thus, a 2D NOESY analysis was carried out. If 7 was
present as the Z isomer, NOESY analysis would reveal interactions
between the vinylic proton and H-4 on the indolinone ring. Since
the latter interaction was not observed in the spectrum (Supple-
mentary data), it was concluded that 7 was present as the E
isomer.
To determine if the Class 1–2 compounds were E or Z isomers,
the chemical shifts of the aromatic protons on the benzylidene ring
B (H-20 and H-60) were analyzed. In the Z but not E isomer, the H-20
or H-60 protons would be shifted downfield due to deshielding by
the carbonyl group. In the literature, chemical shifts of 7.85–
8.53 ppm were assigned to the Z isomer, and 7.45–7.84 ppm to
the E isomer.26,27 For compounds already reported to be E isomers
and whose chemical shifts were available from the literature, com-
parisons of our experimental values with reported values provided
the needed confirmation.27,36,37,39,40,44–46
For compounds whose stereochemistries have not been previ-
ously assigned, the chemical shifts of the H-20 and H-60 protons
at 7.45–7.84 ppm were used to conclude that the synthesized
compounds were E isomers. Compounds 23, 29, 32, 46, 47 were
obtained as a mixture of E and Z isomers but with one predom-
inant isomer. When the 1H spectra of the isomers (after separa-
tion by column chromatography) were examined, the chemical
shifts of H-20 and H-60 were found at 7.90–8.01 ppm for the pre-
dominant isomer, and 8.40–8.50 ppm for the minor isomer.
In the case of 21 and 37, X-ray analyses showed that they were
present as E isomers (Fig. 2) and this led us to assign the E config-
uration to other 3-(benzylidene) indolin-2-ones with 40 substitu-
ents (4, 24, 27, 33), as well as 38 which is a regioisomer of 37. A
check on the 1H chemical shifts of these compounds provided fur-
ther confirmation.
The Class 3 compounds 53 and 54 are 2-benzylideneindolin-3-
ones. 53 was reported to be a Z isomer38 and a comparison of the
reported 13C chemical shifts with the experimentally obtained
values for 53 showed good agreement. By the same token, 54
was assigned a Z configuration based on the similarity of its spec-
tral characteristics (13C and 1H chemical shifts, aromatic region)
to 53.
Figure 2. X-ray structures of 21 and 37.

W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
2081
As for the Class 4 compounds, we found recurrent similarities in
the 13C and 1H spectra of all the members (55–61). Since isoindigo
(55) was reported to have an E configuration,42 the other isoindigos
(56–61) in Class 4 were likewise assigned as E isomers.
A chiral centre at C3 is present in the reduced compounds 51
and 52, but no attempt was made to determine if they were R, S
or a racemate. The compounds were purified in the usual way
and used as such for biological testing.
Table 2
CD and IC50 values of Class 1–4 compounds on Hepa1c1c7 cells
Number
Ring A (R)
Ring B (R0)
QR Induction of Hepa1c1c cellsaCD (
lM)
IC50 (lM) on Hepa1c1c7 cells
Class 1b
1
2
3
4
5
6
7
8
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
0.92 0.09
0.25 0.06
0.27 0.06
0.87 0.11
>11
0.013 0.003
>14
>16
>25
>25
>25
>25
11.1
>25
13.9
16.0
>25
10.8
>25
>25
>25
>25
>25
11.9
14.6
>25
>25
>25
>25
>25
>25
>25
>25
5.9
20-OCH3
30-OCH3
40-OCH3
20,40 -(OCH3)2
20,50-(OCH3)2
20,60-(OCH3)2
30,50-(OCH3)2
30,40,50-(OCH3)3
30 -OC6H5
20-OH
9
>25
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
1.7 0.4
9.0 1.0
1.1 0.2
2.4 0.6
5.3 0.6
0.49 0.13
0.12 0.03
0.86 0.08
0.85 0.07
0.27 0.05
1.6 0.2
2.4 0.4
0.85 0.07
1.73 0.4
5.1 0.7
3.3 0.6
0.12 0.03
0.45 0.1
2.3 0.6
0.33 0.08
4.5 0.6
0.027 0.006
0.031 0.009
20.7 2.1
2.7 0.5
14.4 1.3
0.17 0.05
>19
30-OH
20,50-(OH)2
30,50-(OH)2
20-OH, 40-OCH3
30-OH, 40-OCH3
30-OCH3, 40-OH
20CH3
30-CH3
40-CH3
40 -C2H5
20-F
30-F
40-F
20,40-(F)2
30-Cl
40-Cl
6.7
20, 40-(Cl)2
20-Cl, 40-F
20-CF3
10.6
>25
>25
3.9
12.0
>25
19.9
>25
>25
18.6
>25
30-CF3
40-CF3
40-CN
Ring B = 20-pyridinyl
Ring B = 40-pyridinyl
Ring B = 30-indolyl
Ring B = 10-napthalenyl
Ring B = 20-napthalenyl
0.42 0.06
Class 2b
39
40
41
42
43
44
45
46
6-F
5-Cl
6-Cl
6-OCH3
6-F
6-Cl
6-OCH3
6-F
6-Cl
6-OCH3
30-OCH3
0.17 0.05
0.36 0.06
0.039 0.01
0.13 0.05
0.11 0.05
0.076 0.01
0.24 0.08
0.016 0.007
0.0041 0.0009
0.085 0.014
9.8
30-OCH3
17.5
>25
9.8
15.6
>25
9.8
2.2
1.1
2.6
30-OCH3
30-OCH3
30-OH, 40-OCH3
30-OH, 40-OCH3
30-OH, 40-OCH3
30-CF3
47
48
30-CF3
30-CF3
Class 3f,g,h
49c
—
—
—
—
—
—
30OCH3
1.07 0.15
1.2 0.3
10.9 1.1
>25
0.49 0.07
1.9 0.2
>25
15.5
>25
>25
21.2
>25
50c
30-OH, 40-OCH3
30-OCH3
51d
52d
40-CH3
53e
30-OCH3
54e
30-OH, 40-OCH3
Class 4f
55
56
57
58
H
H
H
H
H
0.025 0.008
>25
>25
20.5
>25
6-F
5-Cl
6-Cl
0.0083 0.0016
0.0015 0.0007
0.0017 0.00097
(continued on next page)

2082
W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
Table 2 (continued)
Number
Ring A (R)
Ring B (R0)
QR Induction of Hepa1c1c cellsaCD (
lM)
IC50
(l
M) on Hepa1c1c7 cells
59
60
61
6-CF3
6-CF3
6-CF3
H
5-F
6-Cl
0.00056 0.00021
0.00096 0.00016
0.00069 0.0002
>25
>25
9.8
Structures of Class 1–4 are as follows:
H
N
B
R'
B
R'
O
O
B
R'
B
R'
A
A
R
O
A
A
O
A
O
N
R
N
H
N
H
N
O
R'
B
N
H
H
b Class 1 and Class 2
c Class 3 (49,50)
d Class 3 (51,52)
e Class 3 (53,54)
f Class 4
a
Concentration required to bring about a twofold increase in the NQO1 activity of the treated cells compared to the untreated Hepa1c1c7 cells. Mean and SD (n = 3).
2.3. Induction of NQO1 in Hepa1c1c7 cells
(27, 0.45
1.6
M) > 40-C2H5 (21, 2.4
20.7
M). For 30-monosubstituted compounds, it was 30-CF3 (31,
0.027
M) > 30-Cl (26, 0.12 M) > 30-OCH3 (3, 0.27 M), 30-CH3
(19, 0.27 M) > H (1, 0.92
M), 30-OH (12, 1.1 M) > 30-OC6H5 (10,
1.70 lM). When the Hansch p and Hammett r
M), 30-F (23, 1.73
l
M) > 40-OCH3 (4, 0.87
M) > 40-F (24, 5.1
l
M), H (1, 0.92
l
M) > 40-CH3 (20,
l
M) > 40-CN (33,
l
l
l
The compounds were screened for induction of NQO1 activity in
the murine hepatoma (Hepa1c1c7) cells.14 Briefly, the assay is
based on the generation of NADPH when glucose-6-phosphate
(G6P) is reduced by G6P dehydrogenase in the presence of its
cofactor NADP. NADPH serves as an electron donor for the NQO1
mediated reduction of menadione (a quinone) to menadiol (a
diphenol). The latter reduces MTT to formazan (a coloured prod-
uct) and its formation is monitored in the assay. A compound that
induces NQO1 increases the rate at which menadiol is formed, and
hence the generation of formazan.
l
l
l
l
l
l
l
values of these substituents50 were analyzed for correlation to
CD (expressed as ꢁlog10 CD or pCD), induction activity was signif-
icantly correlated to
r (Spearman’s q = 0.42, p = 0.045, 2 tailed) but
not
sity of ring B and to a lesser extent by its lipophilic character. The
positive value indicated a preference for electron withdrawing
(+ ) groups like CF3 and halogens.
When ring B was di- or tri-substituted, no clear trend was ob-
p. Thus, activity was determined primarily by the electron den-
q
The compounds were first screened for their growth inhibi-
tory IC50 values on the Hepa1c1c7 cells to ensure that concentra-
tions used in the induction assay were not cytotoxic to the cells.
r
served. By contrast to the preference for electron withdrawing
groups observed for mono-substituted ring B, di-substitution of
ring B with two halogens did not improve activity, as seen from
Compounds that did not affect cell viability at 25
concentration tested) were listed as having IC50 values exceeding
25
M. It can be seen from Table 2 shows that ꢀ80% of the com-
pounds had IC50 values P10 M. Once this was established, the
lM (highest
l
a comparison of 28 (20,40-Cl, CD 2.3
0.12
M) and 27 (40-Cl, CD 0.45
3.3
M) with 22 (20-F, CD 0.85
l
M) with 26 (30-Cl, CD
M), as well as 25 (20,40-F, CD
M) and 23 (30-F, CD 1.73
M).
l
l
l
l
l
compounds were tested over a 10–100-fold concentration range
for induction activity, which was expressed in terms of the CD,
defined as the concentration required to bring about a twofold
increase in NQO1 activity of treated cells compared to untreated
Hepa1c1c7 cells. The results are given in Table 2. Positive con-
trols for this assay were the known NQO1 inducers sulforaphane
and b-napthoflavone (BNF). Their CD values were 0.26 ( 0.04)
l
In the case of OCH3, the introduction of two or three such
groups on ring B (5, 7, 8, 9) caused a marked decrease in induc-
tion activity. A notable exception is the di-methoxylated deriva-
tive 6 (20,50-OCH3, CD 0.013
lM) which was extremely potent
and in fact, the most potent inducer among the Class 1 com-
pounds. Its potency may be related to conversion to an active
metabolite (perhaps a quinone) that mediates the induction
effect.
l
M and 0.028 ( 0.003)
lM, which agreed well with reported
values.25
As seen from Table 2, not all compounds had CD values. Nota-
bly, 5, 7, 8 and 37 had no effect on NQO1 induction at concentra-
tions lower than their IC50 Hepa1c1c7, and could not be tested at
concentrations greater than IC50 Hepa1c1c7 because of cytotoxic ef-
fects on the cells. Their CD values were recorded as >IC50 Hepa1c1c7
rounded up to the nearest unit. No induction was observed for 9
When a non-phenyl ring was present as ring B, variable out-
comes were observed. An interesting observation was the different
induction potencies of the isomeric napthalenyl derivatives (37,
38). The 10 napthalenyl isomer (37, CD > 19
l
M) was significantly
M). The
less active than the 20-napthalenyl isomer (38, CD 0.42
l
20 and 40-pyridinyl isomers (34, 35) also showed regioselectivity
in their induction properties.
and 52 when they were tested at 25 lM which was the highest
concentration used in the assay. Their CDs were listed as >25
in Table 2.
l
M
Class 2 comprised of a small group of ten 3-benzylideneindo-
lin-2-ones that were substituted on both rings A and B. They
were generally more potent NQO1 inducers than the Class 1
The 3-substituted indolin-2-ones of Class 1 (n = 38) showed an
impressive 1000-fold variation in induction activity. Since only
ring B was modified, it was clear that the nature of this ring was
a critical determinant of activity. Ring B was either mono-, di- or
tri-substituted or replaced by a non-phenyl ring B. These changes
had different effects on induction activity as discussed in the fol-
lowing paragraph.
compounds, with CD values ranging from 0.36 lM to 4 nM. To
determine if substitution on ring A was instrumental in the im-
proved induction profiles, the following comparisons (analyzed
by one-way ANOVA) were made. In the case of 3 (30-OCH3 on
ring B, no ring A substituent) and 39–42 (30-OCH3 on ring B, var-
ious substituents on ring A), greater induction was observed for
41 (6-Cl) and 42 (6-OCH3). For 31 (30-CF3 on ring B, no ring A
substituent) and 46–47 (30-CF3 on ring B, various ring A substit-
uents), only 47 (6-Cl) improved activity. But in the case of 16
(30-OH, 40-OCH3 on ring B, no ring A substituent) and 43–45
Among the Class 1 compounds with monosubstituted ring B,
good activity was associated with the CF3, Cl and OCH3 groups at
either at 30 or 40 position of ring B. The sequence of activity for
40-monosubstituted compounds was 40-CF3 (32, 0.031 M) > 40-Cl
l

W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
2083
(30-OH-40-OCH3 on ring B, various ring A groups), substitution on
ring A did not improve activity. Thus, compounds with 6-Cl on
ring A fared better than their unsubstituted ring A analogues
only when coupled with certain ring B groups, namely 30-OCH3
and 30-CF3. This may be true for other substituents as well but
with the limited examples available, it is not possible to deter-
mine whether ring A or B plays a dominant role.
The activities of the Class 3 compounds were particularly inter-
esting because they shed light on the role of the nitrogen linked
Michael acceptor for NQO1 induction. The results showed that it
was important to retain an unsubstituted ring nitrogen (N-methyl
derivatives 49, 50 were less active) and to keep the carbon–carbon
double bond intact. Reduction of the double bond gave compounds
without the Michael acceptor moiety (51, 52) and this was accom-
Table 3
CYP1A1 and NQO1 induction ratios of test compounds
Numbera
Ring A (R)
Ring B (R0)
CYP1A1 induction ratiob
NQO1 induction ratioc
NQO1/CYP1A1 Quotientd
Class 1e
1 (5
2 (1
3 (1
4 (5
5
6 (0.05
7
8
9
10 (7.5
11 (20
12 (5
13 (10
14 (20
15 (2.5
16 (0.5
17 (5
18 (5
19 (1
l
l
l
l
M)
M)
M)
M)
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
1.01 0.04
2.48 0.09
1.26 0.08
1.04 0.01
Not done
3.08 0.31
Not done
Not done
5.13
4.03
4.21
4.50
—
3.32
—
—
5.08
1.63
3.34
4.3
—
1.08
—
20-OCH3
30-OCH3
40-OCH3
20,40 -(OCH3)2
20,50-(OCH3)2
20,60-(OCH3)2
30,50-(OCH3)2
30,40,50-(OCH3)3
30 -OC6H5
20-OH
l
M)
M)
—
—
Not done
—
l
1.08 0.02
1.02 0.02
1.11 0.03
2.87 0.25
1.04 0.03
2.35 0.23
2.66 0.16
1.32 0.10
1.66 0.10
1.08 0.05
1.07 0.04
1.08 0.01
1.26 0.10
1.11 0.02
1.64 0.09
1.89 0.18
1.21 0.01
1.06 0.03
1.21 0.06
1.20 0.04
1.20 0.02
1.40 0.09
1.22 0.03
1.03 0.04
1.08 0.04
1.01 0.01
1.28 0.07
Not done
3.34
3.35
3.94
7.69
3.82
3.59
3.13
4.00
4.45
4.04
3.05
4.96
6.80
4.19
5.69
5.06
4.47
3.67
4.24
3.52
2.96
3.85
2.74
2.20
3.92
2.08
4.33
—
3.09
3.28
3.55
2.68
3.67
1.53
1.17
3.03
2.68
3.74
2.85
4.59
5.39
3.77
3.47
2.68
3.69
3.46
3.50
2.93
2.47
2.75
2.24
2.13
3.63
2.06
3.38
—
lM)
l
M)
30-OH
l
l
M)
M)
20,50-(OH)2
30,50-(OH)2
20-OH, 40-OCH3
30-OH, 40-OCH3
30-OCH3, 40-OH
20CH3
l
l
M)
M)
M)
l
l
M)
l
M)
M)
l
l
l
30-CH3
20 (7.5
21 (7.5
22 (5
M)
M)
40-CH3
40 -C2H5
l
20-F
23 (7.5
24 (20
25 (15
26 (0.5
27 (2.5
28 (10
29 (2.5
30 (20
31 (0.1
32 (0.1
33 (25
34 (10
35 (20
M)
30-F
lM)
40-F
l
M)
20,40-(F)2
l
M)
M)
M)
30-Cl
l
l
40-Cl
20,40-(Cl)2
20-Cl, 40-F
20-CF3
l
M)
lM)
l
M)
M)
M)
M)
M)
30-CF3
l
40-CF3
l
l
l
40-CN
Ring B = 20pyridinyl
Ring B = 40pyridinyl
Ring B = 30indolyl
Ring B = 10 napthalenyl
Ring B = 20 napthalenyl
36 (1
37
38 (2.5
lM)
l
M)
1.06 0.01
4.61
4.35
Class 2e
39 (1
lM)
6-F
5-Cl
6-Cl
6-OCH3
6-F
6-Cl
6-OCH3
6-F
6-Cl
6-OCH3
30-OCH3
1.25 0.06
1.32 0.08
1.51 0.13
1.13 0.01
2.05 0.07
2.33 0.13
1.05 0.01
1.23 0.08
2.40 0.27
1.36 0.07
3.21
4.34
2.56
3.08
4.10
5.09
4.34
3.31
5.20
4.22
2.57
3.29
1.70
2.72
2.00
2.18
4.13
2.69
2.17
3.10
40 (2.5
41 (0.1
42 (0.5
43 (0.5
l
l
l
l
M)
30-OCH3
M)
M)
M)
30-OCH3
30-OCH3
30-OH, 40-OCH3
30-OH, 40-OCH3
30-OH, 40-OCH3
30-CF3
44 (1
45 (1
l
l
M)
M)
46 (0.05
47 (0.05
48 (0.5
l
l
M)
M)
30-CF3
30-CF3
lM)
Class 3f,g,h
49f (5
50f (5
l
l
M)
M)
—
—
—
—
—
—
30OCH3
1.18 0.05
1.08 0.03
1.78 0.10
Not done
4.78
2.64
2.23
—
4.05
2.44
1.25
—
30-OH, 40-OCH3
30-OCH3
51g (20
52g
l
M)
40-CH3
53h (5
l
M)
l
30-OCH3
1.13 0.07
1.38 0.11
3.88
4.21
3.43
3.05
54h (7.5
M)
30-OH, 40-OCH3
Class 4i
55 (0.1
l
M)
H
6-F
5-Cl
6-Cl
6-CF3
6-CF3
H
H
H
H
H
5-F
1.44 0.11
2.75 0.14
2.31 0.15
2.09 0.06
3.01 0.26
3.72 0.20
2.80
2.56
2.42
2.88
2.99
3.09
1.94
0.93
1.05
1.38
0.99
0.83
56 (5 nM)
57 (5 nM)
58 (50 nM)
59 (5 nM)
60 (5 nM)
(continued on next page)

2084
W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
Table 3 (continued)
Numbera
Ring A (R)
Ring B (R0)
CYP1A1 induction ratiob
NQO1 induction ratioc
NQO1/CYP1A1 Quotientd
61 (5 nM)
Sulforaphane (1
BNF (0.01 M)
6-CF3
—
—
6-Cl
—
—
3.23 0.21
0.92 0.04
1.13j
2.96
3.89
1.83
0.92
4.22
1.62
lM)
l
Structures of Class1–4 are as follows:
a
Compound number and concentration used in CYP1A1 induction assay (ꢀ4 ꢂ CD for NQO1 activity) is given in parentheses.
b
Determined on Hepa1c1c7 cells at concentrations given in column 1. Mean SD, n = 3. CYP1A1 induction ratio = (FluorescenceCells+Test Compound ꢁ FBlank)/Fluores-
cenceControl Cells ꢁ FBlank).
c
Determined on Hepa1c1c7 cells at the same concentration as CYP1A1 induction ratio. Values were read off from plot of degree of induction versus concentration for each
compound.
d
Quotient = NQO1 induction activity/CYP1A1 induction activity.
H
N
B
R'
B
R'
O
O
B
R'
B
R'
A
A
R
O
A
A
O
A
O
N
R
N
H
N
H
O
N
R'
B
N
H
H
e Class 1 and Class 2
f Class 3 (49,50)
g Class 3 (51,52)
h Class 3 (53,54)
i Class 4
j
This value was read off from a curve constructed with concentrations 0.005, 0.05, 0.5 and 5 lM BNF. Therefore no SD is assigned to this value. At 0.05 lM BNF, CYP1A1
ratio was 1.27 0.12.
panied by a significant drop in NQO1 induction. As mentioned ear-
lier, 53 and 54 differ from 3 and 16 in the positions of the carbonyl
and exocylic double bond. In 53 and 54, the double bond (and not
the carbonyl group) is adjacent to the ring nitrogen, while the
opposite is true for 3 and 16. This modification had mixed effects
on induction activity, with slightly diminished activities observed
tion of NQO1 relative to CYP1A1. The results are tabulated in
Table 3.
TCDD, BNF and sulforaphane were used as controls in this exper-
iment. TCDD is a known inducer of CYP1A149 and in our hands, it
gave a ratio of 2.99 in the EROD assay at 1 nM. BNF and sulforaphane
are bifunctional and monofunctional NQO1 inducers, respectively.25
Hence, BNF should induce both CYP1A1 and NQO1 activities in He-
pa1c1c7 cells, whereas sulforaphane should induce NQO1 activity
to a greater degree than CYP1A1 activity. We found that BNF at
for 53 (0.46
lM) and 3 (0.27
lM) but a magnitude difference for
54 (1.9 M) and 16 (0.12
l
l
M). Thus, of the three modifications
investigated in Class 3, reduction of the exocyclic double bond
which abolished the Michael acceptor motif had the greatest im-
pact on activity.
0.01 lM induced NQO1 activity by 1.83 times (NQO1 induction ra-
tio)andCYP1A1activityby1.13times(CYP1A1inductionratio),thus
giving an NQO1/CYP1A1 quotient of 1.62. Others have shown that
BNF induced CYP1A1 activity to a greater extent than NQO1 activity
buttheseexperimentswerebasedon analyzingthemicrosomalliver
extracts or mRNA levels of these proteins in rats treated with
BNF.52,53 These may have been more sensitive assays than the cell-
based determinations used here. On the other hand, the results ob-
tained with sulforaphane were in keeping with its monofunctional
The isoindigos of Class 4 were outstanding for their potent induc-
tion of NQO1. The increase in size, lipophilicity and rigidity of these
compounds compared to the Class 1 and 2 compounds may have
contributed to their improved profiles. Isoindigo (55) was earlier re-
ported to be a potent NQO1 inducer in Hepa1c1c7 cells (ED50
320 nM).49 In our hands, it had a CD of 25 nM, but we found greater
induction potencies in other members of Class 4 that were substi-
tuted with electron withdrawing groups. Subnanomolar CD values
were found for 59, 60 and 61 which had 6-CF3 on one indole ring
and H, 5-F or 6-Cl on the other ring, respectively. They were the most
potent NQO1 inducers identified in the entire series.
character. At 1 lM, it barely induced CYP1A1 activity (ratio = 0.92)
but increased NQO1 activity by nearly fourfold (ratio = 3.89).
In general, the Class 1 and 2 compounds were weak inducers of
CYP1A1, with only eight compounds (2, 6, 13, 15, 16, 43, 44, 47)
increasing CYP1A1 activity by more than twofold. The Class 3 com-
pounds showed that N-methylation (49, 50) had negligible effect
but the reduction of the exocyclic double bond (51) caused an al-
most twofold increase in CYP1A1 induction. The 2-benzylidene-
indolin-3-ones (53, 54) had mixed effects compared to their iso-
mers 3 and 16, respectively. The isoindigos (56–61) in Class 4 were
strong inducers of CYP1A1, except for the isoindigo 55 which had
moderate CYP1A1 induction (1.44). The disubstituted analogues
(60, 61) were particularly potent inducers.
More important than the CYP1A1 induction ratio was the quo-
tient of NQO1 versus CYP1A1 induction ratios as this value gave
an indication of whether the compound preferentially induced
NQO1 to CYP1A1. Strong induction of CYP1A1 may not be unfa-
vourable if it is counter-balanced by even stronger induction of
NQO1. Here, we observed an interesting trend when the quotients
were compared to the NQO1 induction activity, namely that com-
pounds with larger quotients were actually weaker NQO1 inducers.
For example, it was noted that the isoindigo 55 was the least po-
tent NQO1 inducer among the Class 4 compounds (CD 25 nM)
2.4. Effect of compounds on 7-ethoxyresorufin O-deethylase
(EROD) activity in Hepa1c1c7 cells
Since induction of NQO1 activity was observed for many Class
1–4 compounds, it was important to determine if phase 1 enzyme
activity was induced as well. The EROD assay was used for this pur-
pose.51 EROD activity describes the rate at which CYP1A1 caused
the O-deethylation of 7-ethoxyresorufin to resorufin. If a com-
pound induced CYP1A1 activity, more resorufin would be formed
compared to control untreated Hepa1c1c7 cells and the ratio of
resorufin fluorescence in treated versus control Hepa1c1c7 cells
would increase. This ratio was determined for compounds at con-
centrations that were approximately 4 times their CD values. A
fixed concentration was not used here because it would be difficult
to select this concentration in view of the large variation in NQO1
induction activities. The quotient of NQO1 and CYP1A1 induction
ratios (ratio determined at the same concentration) was also ob-
tained for each compound. Values >1 would imply greater induc-

W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
2085
5
4.5
4
3.5
3
2.5
2
1.5
1
potent inducers were found in B (CD = 5–10
(CD = 0.1–2.5
in D (CD = 5 and 50 nM, n = 9). The mean NQO1/CYP1A1 quotients
for each group were compared to other groups by one-way ANO-
VA with Tukey as post hoc test. Figure 3 shows the quotients for
each concentration grouping. Only the quotients obtained at con-
centration ranges A (15 and 20 lM) and C (0.1–2.5 lM) did not
differ significantly from each other at p < 0.05. The other compar-
isons (A vs B or D; B vs C or D; C vs D) showed significant differ-
ences. Thus compounds tested at the 5–10 lM range (B) had
l
M, n = 17) and C
l
M, n = 21). The strongest inducers were grouped
0.5
0
significantly larger quotients than compounds tested at lower
concentration ranges (C and D). The compounds in B were mod-
erately potent NQO1 inducers but preferentially induced NQO1
to CYP1A1. In contrast, the more potent inducers found in cate-
gory D induced both NQO1 and CYP1A1 to nearly the same
extent.
A
B
C
D
Concentration Range
Figure 3. NQO1/CYP1A1 quotients for compounds tested at the following concen-
tration range: (A) 20 and 15 M; (B) 5–10 M; (C) 0.1–2.5 M; (D) 5 and 50 nM.
Quotients at all concentration ranges were significantly different (p < 0.05, one-way
l
l
l
ANOVA, Tukey post-hoc) except those at A and C.
2.5. Antiproliferative activity of compounds on HCT116, MCF7
and CCL186 cells
but it had the largest NQO1/CYP1A1 quotient (1.94) in its Class. In
Class 1 and 2, the most potent NQO1 inducer was 47 (NQO1 induc-
tion ratio 5.20, Table 3) and it had a modest NQO1/CYP1A1 quo-
tient of 2.17. Similar correlations were noted for 6, 21 and 45.
To quantify this trend, the compounds were categorized accord-
ing to the concentrations used to determine the CYP1A1 and NQO1
induction activities. Since the compounds were used at ꢀ4 ꢂ CD
for NQO1 activity, weak NQO1 inducers were tested at higher con-
centrations while strong NQO1 inducers were tested at lower con-
Class 1–4 compounds were initially screened at a fixed concen-
tration of 10 lM on the human cancer cell lines HCT116 and MCF7
cells. Only eleven compounds (5–7, 9, 10, 37, 42, 45–48) caused
cell death by more than 50% at this concentration and these com-
pounds were further evaluated for their IC50 on the two cancer cell
lines (Table 4). The small number (11, 18%) of compounds found to
affect cell viability may imply that more stringent requirements
exist for antiproliferative activity, as compared to chemopreven-
tion where 85% of members induced NQO1 activity with CD values
centrations. In all, 10 concentrations ranging from 20
lM to 5 nM
were used and these were grouped into four concentration ranges.
<10 lM (Table 2).
Weak inducers (CD 15 and 20 lM, n = 8) were found in A and more
Table 4
IC50 and selectivity ratios of 3-substituted indoline-2-ones on human breast cancer (MCF-7), human colon cancer (HCT116) and normal human diploid embryonic lung fibroblast
(CCL-186) cell lines
R'
B
H
R
O
A
N
H
a
Compound
Ring A (R)
Ring B (R0)
IC50 (lM)
MCF-7
HCT-116
CCL-186
5.8 0.8
5
H
20,40(OCH3)2
20,50(OCH3)2
20,60(OCH3)2
8.8 0.5
0.7
3.3 0.5
1.8
6
H
5.1 0.1
3.5
4.8 0.7
3.7
5.8 0.9
4.1
3.7 0.8
3.4
3.9 0.6
6.4
1.2 0.4
1.5
3.7 0.3
3.1
8.4 0.3
0.9
7.7 0.2
2.4
6.2 0.3
2.9
10.5 0.7
2.2
8.2 0.7
1.5
19.6 0.9
1.3
1.3 0.4
1.4
7.8 0.9
1.4
8.9 0.3
0.8
18.1 1.1
17.8 0.9
23.6 1.5
12.7 0.6
25.0 1.8
1.8 0.2
11.3 0.4
7.5 0.7
6.6 0.5
5.4 0.3
7
H
9
H
30,40,50- (OCH3)3
30-OC6H5
10
37
42
45
46
47
48
H
H
Ring B = 10 napthalenyl
30OCH3
6-OCH3
6-OCH3
6-F
30OH-40-OCH3
30-CF3
6-Cl
6-OCH3
30-CF3
3.8 0.3
1.7
1.9 0.5
2.8
7.5 0.4
0.9
3.8 0.6
1.4
30-CF3
a
Concentration required to reduce cell survival by 50% after 72 h incubation with test compound. Mean SD, n = 3. Values in bold and italics represent the selectivity ratio:
or IC50 CCL186/IC50
IC50 CCL186/IC50
.
HCT116
MCF7

2086
W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
Only Class 1 and 2 compounds are represented in Table 4. Class
stituents at position 6, namely 6-OCH3, 6-Cl and 6-F. Of these, 6-
OCH3 appears to be preferred because all three compounds (42,
45, 48) with 6-OCH3 are listed in Table 4, irrespective of the group
on ring B. Hence, it may be concluded that 6-OCH3 (ring A) and 30-
CF3 (ring B) are preferred groups for antiproliferative activity, not-
withstanding that the most active compound 42 (6-OCH3, 30-OCH3)
did not have both features in the same molecule. The substituent
preferences for antiproliferative activity are comparatively well
defined, unlike NQO1 induction activity as described in Section 2.3.
3 compounds which had modifications made to the nitrogen-
linked Michael acceptor moiety had negligible antiproliferative
activity. This was also true for the isoindigos of Class 4, in spite
of their strong induction of NQO1 activity.
The compounds in Table 4 had comparable antiproliferative
activities on both cancer cell lines, with IC50 values ranging from
1.2 to 19.6 lM. When these activities were compared to those ob-
tained on a normal cell line (human lung fibroblasts), modest
selectivities (approximately twofold on average) were observed.
Some compounds like 5, 46 and 47 discriminated poorly between
cancerous and normal cells and had selectivity ratios that were
close to 1.
2.6. Effect of selected Class 1 and 2 compounds on cell cycle of
HCT116 cells
For compounds in Class 1, the presence of two or more methoxy
groups on ring B strongly favoured activity. Compounds 5, 6 and 7
are regioisomers with dimethoxy groups on ring B. Their good
activity stands in contrast to another regioisomer 8 (30,50-OCH3
on ring B) which had negligible activity and was not short-listed
for IC50 determination. This may mean that there are preferred
positions for dimethoxy groups on ring B. The other Class 1 com-
pounds (9, 10, 37) in Table 4 have in common bulky substituents
on ring B or a bulky ring B. Compound 9 has three methoxy groups
on ring B (30,40,50-OCH3). A 40-phenoxy group is present in 10. In 37,
ring B was replaced by 10-napthalenyl. Interestingly the 20-napt-
halenyl isomer 38 was not found to have significant antiprolifera-
tive activity. 37 and 38 were also observed to have very different
To have a better understanding of the mechanisms underlying
the antiproliferative activities of the compounds listed in Table 4,
their effects on the cell cycle was investigated by fluorescence-acti-
vated cell sorter (FACS) analysis using flow cytometry. The exper-
iments were designed to examine the effects of the test
compounds on synchronized cell populations at the G1 or G2/M
phase. G1 synchronized cells were obtained by growing cells to
confluence for up to 5 days and then stimulating their re-entry into
the cell cycle by sub-culturing at lower densities. After 24 h of
growth, these cells would be aligned at the G2/M phase.
Cells released from the G1-block were exposed to a fixed con-
centration (ꢀ1.5 ꢂ IC50 on HCT116 cells) of test compound for
24 h, after which the distribution of cells at the different phases
were determined, and compared to control untreated cells. Cells
aligned at the G2/M phase were similarly treated. Representative
FACS diagrams of control cells released from G1 block and G2 block
are give in Figure 4. The proportion of cells in each phase after re-
lease from G1- or G2-arrest are given in Table 5.
NQO1 induction activities, but with 38 (CD 0.42
lM) more potent
than 37 (CD >19 M). These examples underscore the different
l
structural requirements among Class 1 compounds for these two
activities.
The Class 2 compounds are strongly represented for antiprolif-
erative activity. Of the 10 Class 2 compounds available, half of
them were found to have antiproliferative activity. As stated ear-
lier, the Class 2 compounds had only three different ring B substit-
uents (30-OCH3, 30-OH-40- OCH3, 30-CF3). Interestingly, all members
with 30-CF3 (46, 47, 48) were listed in Table 4, indicating the impor-
tant role played by this substituent, quite independently of the ring
A substituent. Similarly, ring A of Class 2 had only 3 different sub-
Analysis of the results revealed that the compounds caused
either G1 arrest (9, 10, 46, 47) or G2 arrest (5–7, 42, 45, 48). The
exception was 37 which did not interfere with cell cycle progres-
sion. Its antiproliferative activity may arise from other mechanisms
like autophagy or necrosis. Taking 46 as an example of a compound
associated with G1 arrest, the FACS diagram showed that when
cells released from G1 block were treated with 46, the progression
Figure 4. FACS diagrams showing effects of representative compounds 46, 48 on HCT116 cells released from (A) G1-block and (B) G2-block. R6, R3, R4, R5 refer to sub-G1, G1,
S and G2/M phases, respectively. 46 and 48 were tested at 15 M and 5
M, respectively (ꢀ1.5 ꢂ IC50).
l
l

W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
2087
Table 5
that left this feature intact, like N-methylation and switching the
positions of the carbonyl and double bond, decreased activity to
a lesser degree. The isoindigos in Class 4 were by far the most po-
tent NQO1 inducers in this study. It is tempting to attribute their
strong activities to the presence of two Michael reaction moieties
embedded in the isoindigo template but this would require verifi-
cation. The unique physicochemical profiles of isoindigos in terms
of size, lipophilicity and conformational rigidity may play a role as
well.
Having established a structural role for the Michael acceptor
moiety, we proceeded to investigate its functional contribution
to NQO1 induction. Compounds with Michael acceptor moieties
are generally associated with the induction of NQO1 without con-
current induction of CYP1A1 activity (mono-functional inducers).
Our results showed that the compounds also induced CYP1A1
activity in the Hepa1c1c7 cells, albeit to varying degrees. Class
4 isoindigos were particularly strong inducers of CYP1A1 (quo-
tients 6 1) and there were others (2, 6, 15, 16, 41, 51) that in-
duced CYP1A1 and NQO1 to almost the same extent
(quotients < 2). It is tempting to attribute this profile to the pres-
ence of the nitrogen in the Michael acceptor moiety because in-
doles like indole-3-carbinol and brassinin are known to be
bifunctional inducers and should induce both NQO1 and CYP1A1.2
Nonetheless, it was heartening to note that there were as many
compounds (1, 4, 21, 38, 45, 49) that induced NQO1 to a greater
extent than CYP1A1 and with quotients that were comparable or
Effect of test compounds on various phases of cell cycle
Compounda
HCT 116 cells released from G1 blockb
Control (0 h)
Control (24 h)
Sub-G1
G1
S
G2/M
0.46
1.94
2.64
2.64
3.52
3.78
3.62
2.96
4.22
5.26
12.68
12.39
4.39
80.19
34.03
26.14
24.24
20.86
56.85
55.93
32.69
10.56
17.49
64.93
65.55
17.50
6.8
11.57
49.47
58.82
46.51
59.63
18.54
18.98
38.42
71.54
56.49
14.58
15.13
59.16
14.90
12.37
26.68
16.00
21.42
21.83
26.91
13.81
21.13
8.04
5 (5 lM)
6 (10
7 (10
9 (15
l
l
l
M)
M)
M)
10 (15
37 (30
42 (2.5
45 (15
46 (15
47 (15
l
l
M)
M)
l
M)
M)
M)
M)
l
l
l
7.10
48 (5
lM)
19.12
HCT 116 cells released from G2 blockb
Control (0 h)
Control (24 h)
1.17
1.24
12.49
9.83
9.51
4.13
31.12
55.88
34.66
39.33
32.97
42.06
63.51
51.32
26.01
26.01
64.15
53.47
30.55
12.11
14.93
13.43
16.68
12.42
16.84
13.10
14.61
15.36
15.36
8.82
55.97
27.90
39.43
34.26
44.99
36.45
21.09
28.26
43.57
43.57
25.00
28.39
31.58
5 (5
lM)
6 (10
7 (10
9 (15
l
l
l
M)
M)
M)
10 (15
37 (30
42 (2.5
45 (15
46 (15
47 (15
l
l
M)
M)
2.16
5.68
l
M)
M)
M)
M)
30.54
15.54
2.49
10.62
20.69
l
l
l
8.05
48 (5
lM)
17.96
exceeded that of sulforaphane (1 lM, quotient = 4.22), a known
a
mono-functional NQO1 inducer. Analysis of the NQO1/CYP1A1
quotients also revealed an interesting trend in that compounds
with favourable quotients (>3) were generally those with moder-
Concentration used is indicated in brackets. ꢀ1.5 ꢂ IC50
.
HCT116
b
Proportion of cells in each phase was the mean of 3 separate determinations.
ate NQO1 induction activity (CD ꢀ 1–2
lM). In contrast, com-
pounds that were very strong or weak NQO1 inducers tend to
from G1 to G2/M was interrupted (Fig. 4A). Notable increases in
the sub-G1 phase were observed (Table 5). As anticipated, 46 had
no effect on the transition of cells from G2/M to G1 (Fig. 4B). In
the case of 48 which caused G2-arrest, it did not interfere with
the progression of cells released from G1 (Fig. 4A) but prevented
cell transit from G2/M to G1 (Fig. 4B). Thus, the proportion of cells
in G1 did not increase after exposure (24 h) to 48 (30.55% in G1,
compared to 55.88% in untreated cells) (Table 5). There were also
significant increases in apoptotic cells in the sub-G1 phase
(Fig. 4B, Table 5).
discriminate poorly between the two proteins.
Recent investigations revealed evidence of cross-talk between
the AhR/XRE (associated with bi-functional induction) and
Keap1/Nrf2/ARE (associated with mono-functional induction) sig-
naling pathways and the coordinate regulation of the AhR and
Nrf gene batteries.54–56 The expression of NQO1 would then be reg-
ulated by these interactions.6,31 If so, the distinction between
mono- and bi-functional inducers may not be as clear-cut and
the perceived advantages of mono-functional over bi-functional
inducers may be oversimplified.
A structural trend was observed among the compounds that
caused G1 or G2 arrest. Compounds (5–7, 42, 45, 48) that caused
G2 arrest had in common the presence of one or two methoxy
groups on ring A or B. Compounds 5–7 are regioisomers with
two methoxy groups on ring B. 42, 45 and 48 had 6-OCH3 on ring
A. Rings B of 42 and 45 were also methoxylated. By contrast, com-
pounds (9, 10, 46, 47) that caused G1 arrest were either halogen-
tated (46, 47) or had bulky ring B substitutents (9, 10). The
outlier status of 37 which caused neither G1 nor G2 arrest is not
unexpected since it is structurally different from the other
compounds.
Turning to the antiproliferative activities of the Class 1–4 com-
pounds, only a small number of compounds were found to cause at
least 50% cell death at 10 lM. It would seem that the structural
templates of Class 1–4 are better suited for chemoprevention than
antiproliferative activity. Nonetheless, the presence of an intact
nitrogen-linked Michael acceptor moiety remains a requirement
for both activities. Within each class, we noted differences in struc-
tural requirements. For example, Class 1 compounds with alkoxy
groups (dimethoxy, trimethoxy) groups on ring B had good anti-
proliferative activities but were not particularly outstanding for
NQO1 induction. Among Class 2 compounds, 30-CF3 on ring B and
6-OCH3 on ring A were important substitutents for antiprolifera-
tive activity but preferences for NQO1 induction were less clear
cut. However, we did identify a small number of compounds that
combined good antiproliferative activity with chemopreventive
potential (as evaluated from their NQO1/CYP1A1 quotient). These
were mostly Class 2 compounds (42, 45–48) and one compound
from Class 1 (10). These compounds had NQO1/CYP1A1 quotients
that exceeded 2. The Class 2 compounds (42, 45–48) are particu-
larly promising because they have CD values in the nanomolar
range but yet show at least twofold selectivity for NQO1 induction.
These compounds could potentially have the dual effects of cyto-
protection towards normal cells while being cytotoxic to cancer
cells.
3. Discussion
Our results show that functionalized 3-benzylideneindolin-2-
ones were inducers of NQO1, a Phase II enzyme that is widely used
as a biomarker for cancer chemoprevention. The Class 1 and 2 com-
pounds (n = 48) yielded an impressive 5000-fold variation in
induction activity, with several members reporting CD values in
the nanomolar range. An analysis of structure–activity relation-
ships highlighted the critical role of the nitrogen-linked Michael
acceptor motif for activity. The reduction of the exocyclic double
bond effectively abolished the Michael acceptor feature and re-
sulted in a significant lowering of activity whereas modifications

2088
W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
An unexpected finding was how structural features in the com-
7–9, 14, 18, 22, 20, 24, 27, 33–36 were synthesized by conventional
heating (reflux, 18 h) using the same quantities of reactants dis-
solved in a larger volume (20 ml) of ethanol. The reaction mixture
was then cooled to room temperature and the resulting precipitate
removed by filtration, carefully washed with cold ethanol and
recrystallized at least once from ethanol to give the desired prod-
uct. Some compounds (2, 5, 6, 9, 12, 13, 22) were not sufficiently
soluble in ethanol to permit recrystallization. These compounds
were washed many times with ethanol until TLC showed a single
spot. In the case of 14, it did not precipitate from the reaction mix-
ture on cooling. The solvent was then removed under vacuum and
the oil purified by column chromatography (hexane/ethyl ace-
tate = 4:1) until TLC gave a single spot.
pounds with antiproliferative activity were linked to their effects
on cell cycle progression. Compounds with at least two methoxy
groups on rings A and B (5–7, 42, 45) were found to cause G2 ar-
rest, possibly by interfering with the formation of the mitotic spin-
dle. By contrast, compounds with bulky substituents (trimethoxy,
phenoxy) on ring B (9, 10) and halogenated groups on both rings
(46–48) caused G1 arrest which may involve inhibition of cyclin
dependent kinase activity. Evidence of apoptosis was observed
for these compounds. Compound 37, a structural outlier, did not
disrupt the cell cycle at either phase. These structural trends serve
to highlight how seemingly small variations on the ring structure
could affect biological responses.
In conclusion, we have demonstrated the chemopreventive po-
tential of benzylideneindolinones as seen from their ability to in-
duce NQO1 activity. 85% of the compounds evaluated had CD
4.3. General procedure for synthesis of substituted 3-
benzylindolin-2-ones (51, 52)
values that were less than 10 lM, highlighting the potential of this
scaffold for chemoprevention. As anticipated, SAR pointed to the
importance of an intact nitrogen linked Michael acceptor moiety
for this activity. On the other hand, the presence of the nitrogen
may have predisposed the compounds to the induction of both
CYP1A1 and NQO1 activities although there were candidates that
selectively induced NQO1 by fourfold or more in this series. In
the case of the antiproliferative activity, only a handful of com-
pounds were able to successfully combine selective induction of
Compound 20 or 41 (100 mg) was dissolved in ethanol (15 ml),
10 mg Pd/C (10%) was added and the mixture was reacted with H2
in a shaking Parr Hydrogenator at 50 psi, 16 h, room temperature
(30 °C). The catalyst was removed by filtration. The solvent was re-
moved under reduced pressure on a rotary evaporator to give the
crude product which was purified by column chromatography
(hexane/ethyl acetate = 3:1) or recrystallization from ethanol.
NQO1 with good antiproliferative activity (IC50 < 10
l
M on two
4.4. General procedure for synthesis of substituted 2-
benzylidene indolin-3-ones 53 and 5457
cancer cell lines). Most of these compounds were from Class 2
and even among this small number (42, 45–48), interesting differ-
ences in the way they affected the cell cycle were noted. They are
potentially useful leads for compounds that combine antiprolifera-
tive activity against cancer cells with a degree of cytoprotection
mediated through NQO1 induction in normal cells.
1-(3-Hydroxy-1H-indol-1-yl)ethanone (5 mmol) was dissolved
in 50% aqueous ethanol (25 ml) containing concentrated HCl
(0.5 ml) and heated to reflux for 1 h under protection of argon.
The aldehyde (5 mmol) in ethanol was then added and heating
was continued for another 2 h. The reaction mixture was cooled
to room temperature, after which the solvent was removed under
reduced pressure and the residue purified by column chromatogra-
phy (hexane/ethyl acetate = 4:1).
4. Experimental
4.1. Chemistry
4.5. General procedure for synthesis of isoindigos 55–6158
Melting points (uncorrected) were determined on a melting
point apparatus in open glass capillary tubes. 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 residual CHCl3 (d
7.25) and DMSO (d 2.49) as internal standards. 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). Reactions were routinely monitored by thin layer
chromatography (TLC) on pre-coated plates (Silica Gel 60 F254,
Merck), with ultraviolet light as visualizing agent. Column chroma-
tography was carried out with Silica Gel 60 (0.04–0.063 mm) with
hexane/ethyl acetate as eluting solvents. Nominal mass spectra
were collected on LcQ Finnigan MAT mass spectrometer with
chemical ionization (APCI) as probe. The purity of final compounds
was verified by reverse phase HPLC on two different solvent sys-
tems (isocratic mode) and details are given in Supplementary data,
which also contains information on the melting points, nominal
mass and NMR data of the final compounds (1–61).
A
solution of the substituted indolin-2-one (oxindole)
(0.5 mmol) and indoline-2, 3-dione (isatin, 0.5 mmol) in glacial
acetic acid (20 ml) and concentrated hydrochloric acid (0.5 ml)
was heated to reflux for 16 h under argon protection. The reaction
mixture was cooled to room temperature and the precipitate re-
moved by vacuum filtration and washed liberally with ethyl ace-
tate until a single spot was detected on TLC.
4.6. X-ray crystallography
Crystals of 21 or 37 were grown in ethanol and mounted on
glass fibres. X-ray data were collected with a Bruker AXS SMART
APEX diffractometer, using Mo K
a radiation at 223 K, with the
SMART suite of Programs (SMART version 5.628 (200), Bruker AXS
Inc., Madison, WI). Data was processed and corrected for Lorentz
and polarisation effects with SAINT
ker AXS Inc., Madison, WI), and for absorption effect with SADABS
SADABS, version 2.10, 2001, University of Göttingen). Structural
(SAINT+ version 6.22a (2001) Bru-
(
solution and refinement were carried out with the SHELXTL, suite
of programs (SHELXTL, Version 6.14 (2000), Bruker AXS Inc., Madison,
WI). The structure was solved by direct methods to locate the hea-
vy atoms, followed by difference maps for the light, non-hydrogen
atoms. All non-hydrogen atoms were generally given anisotropic
displacement parameters in the final model whereas H-atoms
were placed at calculated positions. X-ray crystal data of 21 and
37 are given in Supplementary data.
4.2. General procedure for synthesis of 3-substituted indolin-2-
ones (1–48)
Equimolar amounts (0.5 mmol) of the oxindole and aldehyde
were dissolved in ethanol (10 ml), a drop of piperidine (20 ll)
was added and the mixture heated in a sealed vessel (10 ml)
flushed with argon. The vessel was heated to 140 °C for 15 min
in a microwave synthesizer (Biotage InitatorÒ). Compounds 1–5,

W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
2089
4.7. Materials for biological assays
tration of DMSO was 0.5% v/v. The cells were incubated with test
compound for 48 h, after which the medium was removed, the well
Menadione, digitonin, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphe-
nyltetrazolium bromide (MTT), b-naphthoflavone (BNF), dicuoma-
rol, salicylamide, propidium iodide, RNase A, 7-ethoxyresorufin
and sulforaphane were purchased from Sigma–Aldrich (St. Louis,
MO). 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) was purchased
from AccuStandard (New Haven, USA). Other reagents were of ana-
lytical grade.
washed with 200
and then incubated with 5
cylamide in 200 l of medium at 37 °C, 40 min. Readings were ta-
l
l of 1ꢂphosphate-buffered saline solution (PBS)
l
M 7-ethoxyresorufin and 2 mM sali-
l
ken at kexcitation of 530 nm and kemission of 590 nm on a fluorometer.
Fluorescence (if any) of the test compound was determined at the
same wavelengths to take into account its contribution to the ob-
served readings. Readings of empty wells (no cells, ‘blank’) and
wells with Hepa1c1c7 cells in medium containing 0.5% DMSO
but without test compound (‘control’) were also determined.
TCDD, a strong inducer of CYP1A1 activity, as well as BNF and sul-
foraphane were used as controls.
4.8. Determination of NQO1 activity14
Hepa1c1c7 were purchased from American Type Culture Collec-
tion (Rockville, MD). and cultured in alpha-minimum essential
medium (a-MEM) without nucleosides, and containing 10% (v/v)
heat- and charcoal-treated fetal calf serum (1 g of charcoal per
100 ml of serum; 90 min at 5 °C), 0.15% sodium bicarbonate, 0.01%
penicillin G, 0.01% streptomycin sulfate in an humidified atmo-
sphere of 5% CO2 at 37 °C. Cells were sub cultured when they
reached 80–90% confluency and used within 6–17 passages for
determinations. For the assay, about 10,000 cells were grown in
CYP1A1 induction activity was given by the expression:
Degree of induction ¼ FCellsþtest compound ꢁ FBlank=FControl ꢁ FBlank
4.10. Determination of antiproliferative activity by the
microculture tetrazolium (MTT) assay60
each well of a 96-well plate for 24 h in a-MEM. Total protein content
The antiproliferative activity of the test compounds were deter-
mined by the MTT assay on the following cell lines: MCF7 (human
breast cancer cell line), HCT116 (human colon cancer cell line) and
CCL186 (normal human diploid embryonic lung fibroblast). The
cell lines were purchased from American Type Culture Collection
(Rockville, MD). MCF7 and CCL186 cells were cultured in Eagle’s
Minimum Essential Medium (EMEM) with 1 mM sodium pyruvate,
per was found to be 0.036 mg per well based on the Bradford assay.
Stock solutions of test compounds were prepared in DMSO and ali-
quots were added to each well to give the desired concentration. The
final concentration of DMSO in each well was kept at 0.5% v/v or low-
er. After incubation for 48 h, the media was decanted and the cells
lyzed by a solution containing 0.8% w/v digitonin and 2 mM EDTA
with incubation at 37 °C (10 min). The plates were gently agitated
1.5 g/l sodium bicarbonate, 2 mM
L-glutamine supplemented
on an orbital shaker for 10 min at 23 °C. An aliquot (200 ll) of a solu-
with10% FBS and 0.01% antibiotics. HCT116 cells were cultured in
McCoy’s 5A Medium with 2.2 g/l sodium bicarbonate supple-
mented with 10% FBS and 0.01% antibiotics. Cells were plated at
the following densities: 10,000 cells/well (MCF7), 4000 cells/well
(HCT116), 5000 cells/well (CCL186). They were grown for 24 h in
tion (‘complete reaction mixture’) was then added to each well. This
solution was freshly prepared just prior to use and consisted of
7.5 ml 0.5 M Tris–Cl (pH = 7.4), 100 mg bovine serum, 1 ml 1.5%
Tween-20, 0.1 ml 7.5 mM FAD, 1 ml 150 mM glucose 6-phosphate,
90
genase, 45 mg MTT and deionized water made up to a final volume
of 150 ml. Menadione (1 l of 50 mM menadione dissolved in aceto-
ll 50 mM NADP, 300 units yeast glucose-6-phosphate dehydro-
a 96-well plate, with 100 ll of corresponding medium in each well.
The test compounds were prepared in DMSO and diluted to a series
of concentrations with medium. Not more than 1% DMSO (final
concentration) was present in each well. The test compounds were
l
nitrile for every milliliter of reaction mixture was added just before
the mixture was dispensed into the wells. After addition, the plate
was gently agitated for 5 min and the reaction was quenched by
incubated with the cells for 72 h, after which 100 ll MTT solution
(0.5 mg/ml in 1 ꢂ PBS) was added for 3 h and the cells lysed to re-
addition of a solution (50 ll) of 0.3 mM dicoumarol in 0.5% DMSO
lease the formazan product. The latter was dissolved in DMSO
and 5 mM potassium phosphate, pH = 7.4. A blue colour due to the
formation of formazan was observed in each well, the absorbance
of which was measured at 590 nm on a plate reader. Blank wells
contained no cells and control wells contained Hepa1c1c 7 cells
treated with medium containing 0.5% DMSO but without the test
compound. NQO1 induction activity of a compound tested at a given
concentration was determined from the equation:
(150 ll) and absorbance determined within 30 min at 590 nm on
a microtitre plate reader. Cell survival was given by the expression:
Cell survival ð%Þ ¼ ½ðAcellsþtest compound ꢁ AblankÞ=ðAuntreated cells
ꢁ AblankÞꢃ ꢂ 100
where A is the absorbance of formazan measured at 590 nm in the
Degree of induction ¼ Atest compound ꢁ Ablank=Acontrol ꢁ Ablank
test (Acells
compound), control (Auntreated cells) or blank (Ablank)
+
test
wells. Each concentration of test compound was evaluated on 3 sep-
arate occasions. The concentration (IC50) that inhibited 50% of cell
growth was determined from the sigmoidal curve obtained by plot-
ting % surviving cells versus concentration using OriginPro 7.5 SR1
(Version V7.5776 B776), OriginLab Corporation, MA.
The MTT assay was also carried out on Hepa1c1c7 cells for the
purpose of determining cytotoxicity of test compounds. The same
procedure was adopted except that the incubation period of com-
pound with cells was 48 h, instead of 72 h.
where A is absorbance of formazan measured at 590 nm.
Sulforaphane and BNF were determined under similar condi-
tions as positive controls. The results were plotted (degree of
induction vs concentration) with OriginPro 7.5 SR1 (Version
V7.5776 B776), OriginLab Corporation, MA and the concentration
of test compound required to double the basal NQO1 activity
(CD) was determined. Three separate determinations were made
and CD was reported as mean SD for n = 3.
4.9. Determination of 7-ethoxyresorufin O-deethylase (EROD)
activity in Hepa1c1c7 cells59
4.11. Determination of the effects of test compounds on the cell
cycle of HCT116 cells by cell cytometry
Hepa1c1c7 cells were plated at a density of 10,000 cells per well
in a 96-well plate and cultured for 24 h. A stock solution of test
compound was prepared in DMSO and serially diluted with med-
ium to give the desired concentration in the well. The final concen-
HCT116 cells were grown to confluence and maintained in this
state for at least 5 days without change of media. The serum
starved cells were then trypsinized and subcultured at densities

2090
W. Zhang, M.-L. Go / Bioorg. Med. Chem. 17 (2009) 2077–2090
of 5 ꢂ 105 cells/well in six-well plates. Growth media contained
10% fetal calf serum. The test compound was added either imme-
diately to cells synchronized at G1 phase, or 24 h later when cells
were synchronized at G2 phase. The cells were incubated with test
compound for 24 h from the time of addition. After this time, the
cells were harvested, trypsinized and fixed in 70% ice-cold ethanol
for a minimum of 24 h. After centrifugation, the supernatant was
15. Talalay, P. Chemical Protection Against Cancer by Induction of Electrophile
Detoxication (Phase 2) Enzymes. In Cellular Targets for Chemoprevention; Steele,
V., Ed.; Boca CRC Press: Raton, 1992; pp 193–205.
16. Cuendet, M.; Oteham, C. P.; Moon, R. C.; Pezzuto, J. M. J. Nat. Prod. 2006, 69, 460.
17. Asher, G.; Lotem, J.; Cohen, B.; Sachs, L.; Shaul, Y. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 1188.
18. Gong, X.; Kole, L.; Iskander, K.; Jaiswal, A. K. Cancer Res. 2007, 67, 5380.
19. Long, D. J., 2nd; Waikel, R. L.; Wang, X. J.; Perlaky, L.; Roop, D. R.; Jaiswal, A. K.
Cancer Res. 2000, 60, 5913.
20. Iskander, K.; Gaikwad, A.; Paquet, M.; Long, D. J., 2nd; Brayton, C.; Barrios, R.;
Jaiswal, A. K. Cancer Res. 2005, 65, 2054.
21. Ahn, K. S.; Sethi, G.; Jain, A. K.; Jaiswal, A. K.; Aggarwal, B. B. J. Biol. Chem. 2006,
281, 19798.
22. Dinkova-Kostova, A. T.; Abeygunawardana, C.; Talalay, P. J. Med. Chem. 1998,
41, 5287.
23. Dinkova-Kostova, A. T.; Massiah, M. A.; Bozak, R. E.; Hicks, R. J.; Talalay, P. Proc.
Natl. Acad. Sci. U.S.A. 2001, 98, 3404.
discarded and the pellet was treated with RNase A (200
for 30 min at room temperature, followed by cell staining using
propidium iodide at a final concentration of 20 g/ml. The stained
lg/ml)
l
cells were then analyzed for cell cycle distribution in the sub-G1,
G1, S and G2/M phases on a Dako Cytomation Cyan LX (Dako Col-
orado, Fort Collins, CO, USA) equipped with an argon solid state la-
ser (488 nm) using the Summit (version 4.3) software. Compounds
were tested at a concentration of ꢀ1.5 ꢂ IC50, with IC50 determined
by the MTT assay on HCT116 cells.
24. Dinkova-Kostova, A. T.; Holtzclaw, W. D.; Cole, R. N.; Itoh, K.; Wakabayashi, N.;
Katoh, Y.; Yamamoto, M.; Talalay, P. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 11908.
25. Dinkova-Kostova, A. T.; Liby, K. T.; Stephenson, K. K.; Holtzclaw, W. D.; Gao, X.;
Suh, N.; Williams, C.; Risingsong, R.; Honda, T.; Gribble, G. W.; Sporn, M. B.;
Talalay, P. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 4584.
26. Sun, L.; Tran, N.; Tang, F.; App, H.; Hirth, P.; McMahon, G.; Tang, C. J. Med. Chem.
1998, 41, 2588.
4.12. Statistical analysis
27. Andreani, A.; Burnelli, S.; Granaiola, M.; Leoni, A.; Locatelli, A.; Morigi, R.;
Rambaldi, M.; Varoli, L.; Kunkel, M. W. J. Med. Chem. 2006, 49, 6922.
28. Prochaska, H. J.; Talalay, P. Cancer Res. 1988, 48, 4776.
29. Nebert, D. W.; Petersen, D. D.; Fornace, A. J., Jr. Environ. Health Perspect. 1990,
88, 13.
30. Sogawa, K.; Fujii-Kuriyama, Y. J. Biochem. 1997, 122, 1075.
31. Nioi, P.; Hayes, J. D. Mutat. Res. 2004, 555, 149.
32. Mehta, R. G.; Liu, J.; Constantinou, A.; Thomas, C. F.; Hawthorne, M.; You, M.;
Gerhuser, C.; Pezzuto, J. M.; Moon, R. C.; Moriarty, R. M. Carcinogenesis 1995, 16,
399.
Data was analyzed for statistical significance by one-way ANO-
VA followed by Tukey HSD as post-hoc test on SPSS 15.0 for Win-
dows, Chicago, IL. Spearman’s correlation analysis was carried out
on the same software. A level of probability of 0.05 was used as the
criterion for significance.
33. Hayes, J. D.; Kelleher, M. O.; Eggleston, I. M. Eur. J. Nutr. 2008, 47, 73.
34. Craig, P. N. J. Med. Chem. 1971, 14, 680.
Acknowlegements
35. Wahl, A.; Fericean, G. Annali di Chimica Applicata 1928, 9, 277.
36. Blake, K. W.; Jaques, B. J. Chem. Soc., Perkin Trans. 2 (1972–1999) 1973, 12, 1660.
37. Corsico Coda, A.; Gamba Invernizzi, A.; Righetti, P. P.; Tacconi, G.; Gatti, G. J.
Chem. Soc., Perkin Trans. 2 (1972–1999) 1984, 4, 615.
38. Burger, U.; Bringhen, A. O. Helv. Chim. Acta 1989, 72, 93.
39. Andreani, A.; Rambaldi, M.; Locatelli, A.; Bossa, R.; Galatulas, I.; Ninci, M. Eur. J.
Med. Chem. 1990, 25, 187.
This work was supported by National University of Singapore
Academic Research Fund R 148 000 084 112. The financial support
(research scholarship) to Zhang Wei from Ministry of Education,
Republic of Singapore and National University of Singapore is
gratefully acknowledged.
40. Andreani, A.; Rambaldi, M.; Locatelli, A.; Bongini, A.; Bossa, R.; Galatulas, I.;
Ninci, M. Eur. J. Med. Chem. 1992, 27, 167.
41. Hirao, K.; Morii, N.; John, T.; Takahashi, S. Tetrahedron Lett. 1995, 36, 6243.
42. Leclerc, S.; Garnier, M.; Hoessel, R.; Marko, D.; Bibb, J. A.; Snyder, G. L.;
Greengard, P.; Biernat, J.; Wu, Y. Z.; Mandelkow, E. M.; Eisenbrand, G.; Meijer, L.
J. Biol. Chem. 2001, 276, 251.
43. Volk, B.; Simig, G. Eur. J. Org. Chem. 2003, 20, 3991.
44. Bouerat, L.; Fensholdt, J.; Liang, X.; Havez, S.; Nielsen, S. F.; Hansen, J. R.; Bolvig,
S.; Andersson, C. J. Med. Chem. 2005, 48, 5412.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2008.12.052.
45. Bouerat, L. M. E.; Fensholdt, J.; Nielsen, S. F.; Liang, X.; Havez, S. E.; Andersson,
E. C.; Jensen, L.; Hansen, J. R. WO Patent 2004-DK8752005058309, 20041216.,
2005.
References and notes
1. Surh, Y. J. Nat. Rev. Cancer 2003, 3, 768.
46. Olgen, S.; Akaho, E.; Nebioglu, D. Farmaco 2005, 60, 497.
47. Pandit, B.; Sun, Y.; Chen, P.; Sackett, D. L.; Hu, Z.; Rich, W.; Li, C.; Lewis, A.;
Schaefer, K.; Li, P. K. Bioorg. Med. Chem. 2006, 14, 6492.
48. Liu, J. J., Z., Z. M., WO Patent WO 2008/055812 A1, 2008.
49. Nishiumi, S.; Yamamoto, N.; Kodoi, R.; Fukuda, I.; Yoshida, K.; Ashida, H. Arch.
Biochem. Biophys. 2008, 470, 187.
50. Kubinyi, H. The Quantitative Analysis of Structure–Activity Relationships. In
Burger’s Medicinal Chemistry and Drug Discovery; Wolff, M. E., Ed., 5th ed.;
Principles and Practice; John Wiley & Sons: New York, 1995; Vol. 1, pp 507–
509.
51. Burke, M. D.; Mayer, R. T. Drug Metab. Dispos. 1974, 2, 583.
52. Brauze, D. Toxicol. Lett. 2004, 152, 111.
2. Pezzuto, J. M.; Kosmeder, J. W., II; Park, E. J.; Lee, S. K.; Cuendet, M.; Gills, J.;
Bhat, K.; Grubjesic, S.; Park, H. S.; Mata-Greenwood, M.; Tan, Y. M.; Yu, R.;
Lantvit, D. D.; Kinghorn, A. D. Characterization of Natural Product
Chemopreventive Agents. In Strategies for Cancer Chemoprevention; Kelloff, G.
J., Hawk, E. T., Sigman, C. C., Eds.; Cancer Chemoprevention; Humana Press:
Totowa, NJ, 2005; Vol. 2, pp 3–37.
3. Sporn, M. B.; Dunlop, N. M.; Newton, D. L.; Smith, J. M. Fed. Proc. 1976, 35, 1332.
4. Chen, C.; Kong, A. N. Trends Pharmacol. Sci. 2005, 26, 318.
5. Kohle, C.; Bock, K. W. Biochem. Pharmacol. 2006, 72, 795.
6. Kohle, C.; Bock, K. W. Biochem. Pharmacol. 2007, 73, 1853.
7. Ross, D.; Siegel, D. Methods Enzymol. 2004, 382, 115.
8. Talalay, P.; Dinkova-Kostova, A. T. Methods Enzymol. 2004, 382, 355.
9. Landi, L.; Fiorentini, D.; Galli, M. C.; Segura-Aguilar, J.; Beyer, R. E. Free Radic.
Biol. Med. 1997, 22, 329.
53. Brauze, D.; Widerak, M.; Cwykiel, J.; Szyfter, K.; Baer-Dubowska, W. Toxicol.
Lett. 2006, 167, 212.
54. Jaiswal, A. K. Biochem. Pharmacol. 1994, 48, 439.
10. Siegel, D.; Bolton, E. M.; Burr, J. A.; Liebler, D. C.; Ross, D. Mol. Pharmacol. 1997,
52, 300.
55. Miao, W.; Hu, L.; Scrivens, P. J.; Batist, G. J. Biol. Chem. 2005, 280, 20340.
56. Ma, Q.; Kinneer, K.; Bi, Y.; Chan, J. Y.; Kan, Y. W. Biochem. J. 2004, 377, 205.
57. Daisley, R. W.; Elagbar, Z. A.; Walker, J. J. Heterocycl. Chem. 1982, 19, 1013.
58. Papageorgiou, C.; Borer, X. HeIv. Chim. Acta 1988, 71, 1079.
59. Marchand, A.; Barouki, R.; Garlatti, M. Mol. Pharmacol. 2004, 65, 1029.
60. Alley, M. C.; Scudiero, D. A.; Monks, A.; Hursey, M. L.; Czerwinski, M. J.; Fine, D.
L.; Abbott, B. J.; Mayo, J. G.; Shoemaker, R. H.; Boyd, M. R. Cancer Res. 1988, 48,
589.
11. Chan, T. S.; Wilson, J. X.; O’Brien, P. J. Methods Enzymol. 2004, 382, 89.
12. Kelly, V. P.; Ellis, E. M.; Manson, M. M.; Chanas, S. A.; Moffat, G. J.; McLeod, R.;
Judah, D. J.; Neal, G. E.; Hayes, J. D. Cancer Res. 2000, 60, 957.
13. McMahon, M.; Itoh, K.; Yamamoto, M.; Chanas, S. A.; Henderson, C. J.; McLellan,
L. I.; Wolf, C. R.; Cavin, C.; Hayes, J. D. Cancer Res. 2001, 61, 3299.
14. Prochaska, H. J.; Santamaria, A. B. Anal. Biochem. 1988, 169, 328.