Computer-aided rational drug design: A novel agent (SR13668) designed to mimic the unique anticancer mechanisms of dietary indole-3-carbinol to block Akt signaling

Article abstract of DOI:10.1021/jm070040e

Indole-3-carbinol (I3C) is a naturally occurring anticancer agent and has entered clinical trials for cancer prevention. However, the clinical development of I3C has been impeded by its poor metabolic profile. The active components of I3C were used to develop a novel class of indole analogs to optimize I3C's anticancer actions, including blocking growth factor-stimulated Akt activation. The most promising of these analogs, SRI3668, exhibited potent oral anticancer activity against various cancers and no significant toxicity.

Full text of DOI:10.1021/jm070040e

3412
J. Med. Chem. 2007, 50, 3412-3415
Letters
Computer-Aided Rational Drug Design: A Novel
Agent (SR13668) Designed to Mimic the Unique
Anticancer Mechanisms of Dietary
Indole-3-Carbinol to Block Akt Signaling
Wan-Ru Chao, Dawn Yean, Khalid Amin, Carol Green, and
Ling Jong*
SRI International, 333 RaVenswood AVenue, Menlo Park, California
ReceiVed January 11, 2007
Abstract: Indole-3-carbinol (I3C) is a naturally occurring anticancer
agent and has entered clinical trials for cancer prevention. However,
the clinical development of I3C has been impeded by its poor metabolic
profile. The active components of I3C were used to develop a novel
class of indole analogs to optimize I3C’s anticancer actions, including
blocking growth factor-stimulated Akt activation. The most promising
of these analogs, SR13668, exhibited potent oral anticancer activity
against various cancers and no significant toxicity.
Figure 1. I3C and its four active in vivo oligomers.
Indole-3-carbinol (I3C, 1), a dietary component found
exclusively in cruciferous vegetables, is known to suppress
proliferation and induce apoptosis of various cancer cells,
including breast, ovarian, lung, cervical, colon, prostate, and
liver, and was the subject of several early-phase clinical trials
for cancer prevention.1 Numerous anticancer mechanisms of I3C
have been proposed,2,3 but the most intriguing is that I3C was
able to selectively inhibit activation (phosphorylation) of Akt
in the tumor-derived MDA-MB-468 breast and PC-3 prostate
cancer cell lines, but not in nontumorigenic human HBL100
breast and CRL2221 prostate epithelial cells.4,5 I3C was also
reported to abrogate epidermal growth factor (EGF)-induced
activation of Akt in PC-3 cells and decrease expression of
downstream modulators of the Akt survival pathway.6 The
protein kinase Akt is a major therapeutic target for various
cancers known to promote cell survival, invasion, and resistance
to hormone-, chemo-, and radio-therapy-induced apoptosis.7
Several kinase inhibitors of the PI3K/Akt pathway, such as
wortmannin and LY294002, have suffered from the negative
effects of their broad-spectrum kinase inhibition and unaccept-
able toxicity. Thus, I3C might offer the promise of selectively
inhibiting Akt activation in cancer cells.
However, the clinical development of I3C has been impeded
by its poor metabolic profile and low potency. I3C (1) itself
does not possess anticancer activity, but must be acted on by
gastric acid in the stomach to form the active metabolites (acid
condensation products or oligomers) responsible for its in vivo
anticancer effects.8 Currently, only four of the over 20 oligomers
have been shown to have anticancer activity. These four active
oligomers, shown in Figure 1, are DIM (2, a linear dimer of
I3C), LT (3, a linear trimer), CT (4, a cyclic trimer), and ICZ
(5, a planar molecule),9 and DIM, the major I3C oligomer, has
been shown to reduce prostate tumor growth in mice.10 How-
ever, these active oligomers represent less than 20% of the total
oligomers formed in vivo, and their formation strongly depends
Figure 2. Superimposing the low-energy conformer of each oligomer.
To enhance clarity, double bonds and hydrogen atoms are not displayed.
on the pH of gastric acid in the stomach, thus, significant
interindividual variation was observed in plasma oligomer values
in a phase I clinical trial.11 Because of the unique anticancer
mechanism and apparent safety of I3C, we decided to use the
four active I3C oligomers as lead compounds to develop a novel
class of indole analogs, which are expected to act by I3C-like
mechanisms against several cancers and to offer substantially
improved potency and activity over I3C that would make them
applicable to cancer therapy as well.
Our lead-based rational drug design was founded on a
computer-aided structural analysis of the four chemically distinct
active oligomers (DIM, LT, CT, and ICZ) to identify common
structural factors responsible for their anticancer activity.
Computational analysis of the low-energy conformer of each
oligomer, using the SYBYL 7.0 software package (Tripos, St.
Louis, MO), revealed that the N-N′ distances were similar in
all four active oligomers (5.9 Å for DIM and 5.5 Å for LT,
CT, and ICZ). Superimposing their low-energy conformers
indeed showed excellent overlap of their indole rings (Figure
2). We speculated that the N-N′ distance might play an
important role in their anticancer effects. Our initial structure-
activity relationship (SAR) studies thus focused on the effects
of the N-N′ distance on the anticancer activity of the oligomers.
We chose to use the MCF-7 human breast cancer cell line for
our initial biological screen because I3C has entered phase I
and II clinical trials for breast cancer prevention.1 We later
expanded our studies to include prostate, ovarian, and lung
cancer.
* To whom correspondence should be addressed. Tel.: 650-859-6121.
Fax: 650-859-3153. E-mail: ling.jong@sri.com.
10.1021/jm070040e CCC: $37.00 © 2007 American Chemical Society
Published on Web 06/29/2007

Letters
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15 3413
Table 2. In Vivo Antitumor Effect of SR13650
Table 1. Modification of the N-N′ Distance and the Planarity of DIM
SR13650
(10 mg/kg)
tumor growth
inhibition (%)
intraperitoneal
oral
43
26
afforded 6, in which steric hindrance caused the N-N′ distance
to increase from 5.9 Å to 7.0 Å. Inserting one phenyl ring
between the two indole rings of DIM generated 7, with a
minimum -energy N-N′ distance of 9.0 Å. To reduce the N-N′
distance between the two indole rings, we connected the 2 and
2′ positions of indole rings to give 8, with an N-N′ distance of
4.7 Å. In vitro biological screens showed that both 6 and 7,
with N-N′ distances greater than 6 Å, lacked anticancer activity.
Analog 8, exhibiting improved anticancer activity, was then
chosen as the ideal template for further structural modifications.
Unlike the other three active oligomers, ICZ is a planar
molecule and was reported to be more potent than DIM in
inhibiting the growth of colon cancer cells.12 We thus designed
analog 9 to examine if introducing the planarity into compounds
will further enhance their anticancer activity. Changing the
hybridization of the carbon bridge atom of DIM from tetrahedral
sp3 to trigonal sp2 produced 9, which retained an N-N′ distance
similar to that of DIM. In vitro biological results indicate that
9 has retained or slightly improved growth inhibition activity.
This interesting result prompted us to design analog 10
(SR13650), in which we incorporated the planarity of 9 into 8,
having an optimal N-N′ distance. Our attempt produced the
first potent lead, 10, with an IC50 of 0.08 µM.
a IC50 is one-half the maximum inhibitory concentration and is the
average of quadruple experiments. The highest dose level tested was 10
µM.
Scheme 1a
To determine if the in vitro anticancer activity of our lead,
SR13650, would offer potent tumor growth inhibition in vivo,
we conducted a small preliminary tumor xenograft study using
MCF-7 cells. SR13650 at 10 mg/kg was administered daily by
oral gavage or ip injection for 30 days. We were pleased to
find that SR13650 indeed exhibited antitumor effects in animal
models at that dose level. We were also disappointed by the
low oral antitumor activity of SR13650, as compared to its ip
antitumor effect (Table 2). Although the in vivo responses could
be different for a given dose depending on the route of
administration, a reasonable doubt was that SR13650 might have
low oral bioavailability due to its highly lipophilic nature. We
thus decided to attempt chemical modifications to improve its
oral antitumor effect.
Increasing the polarity of compounds is known to influence
their membrane permeability and thus improve their oral
bioavailability. We thus decided to introduce electron-withdraw-
ing groups into SR13650 to increase its dipole moment
(polarizability) and possibly its oral bioavailability. To facilitate
our SAR study and compound synthesis, we chose to use DIM
as a template for selection of a suitable polar substituent that
would later be applied on SR13650.
Computational analysis showed that the presence of the polar
substituent at the C-5 position of the indole rings would have
the most significant effect on their polarizability. The electron-
withdrawing groups were thus introduced at the C-5 and C-5′
positions of DIM. Analogs 24 through 27 were designed to
evaluate the influence of various functionalities and steric
hindrance on the antiproliferative activity of DIM (Table 3).
DIM can easily be chemically modified via halogen-metal
exchange of 5,5′-dibromo-DIM 29 followed by nucleophilic
substitution to give the desired compounds (Scheme 2).
Although 24-29 all have increased dipole moment, only 24
(ester) exhibited improved antiproliferation activity (Table 3).
The lack of activity in 25 (acid), 26 (amide), and 27 (sulfone)
a Reagents and conditions: (a) CH3COCH3, Dy(OTf)3, EtOH, H2O, 55%;
(b) (CH3CO)2O, 2-methyl-2-oxazoline, 65%; (c) NaOH, EtOH, H2O, 80%;
(d) PhSO2Cl, NaOH, (n-Bu)4NHSO4, CH2Cl2, 99%; (e) Br2, CCl4, 95%;
(f) t-BuLi, B(Oi-Pr)3, THF, H3O+, 85%; (g) Pd(PPh3)4, 1,3-dibromobenzene,
aq Na2CO3, DME, 75%; (h) K2CO3, MeOH, H2O, 90%; (i) LAH, THF,
98%; (j) NCS, (CH3)2S, CH2Cl2; (k) xylenes, reflux, 63%; (l) Sc(OTf)3,
CH2Cl2; (m) Raney Ni, EtOH, 65%; (n) 10% aq NaOH, 77%; (o)
(t-BuOCO)2O, DMAP, THF (95%); (p) 2,2,6,6-tetramethylpiperidine (TMP),
n-BuLi, THF, -78 °C; ClCO2C2H5, warmed to 0 °C; (q) CF3CO2H, CH2Cl2,
72%.
In our lead-based drug design, we tried to minimize structural
change in the compounds while optimizing their potency to
ensure that our novel indole analogs would retain the safety
and unique biological activity of I3C. Analogs 6 to 9 were
designed and synthesized to determine the optimal N-N′
distance for anticancer activity (Table 1 and Scheme 1). DIM
(2) was used as the initial scaffold for chemical modification.
Adding a gem-dimethyl group on the one-carbon bridge of DIM

3414 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15
Table 3. Introduction of Polar Substituents into DIM
Letters
Table 4. Modification Dipole Moment of SR13650
MCF-7
IC50 (µM)
dipole
moment (D)
MCF-7
IC50 (µM)
dipole
moment (D)
cmpd
R
cmpd
R1
R2
2 (DIM)
H
5.5
1.75
>10
>10
>10
0.5
2.4
2.5
7.9
5.9
10 (SR13650)
35 (SR13661)
36 (SR13667)
37 (SR13668)
H
OCO2C2H5
OCO2C2H5
C3F7
0.08
0.11
0.09
0.19
1.1
2.4
1.4
4.5
24
CO2C2H5
CO2H
CON(CH3)2
SO2CH3
CO2C2H5
CO2C2H5
CO2C2H5
25
26
27
OCH3
Scheme 2a
a Reagents and conditions: (a) 37% HCHO, CF3CO2H, THF, 60%; (b)
(t-BuOCO)2O, DMAP, THF, rt, 94%; (c) t-BuLi, THF, -78 °C,
ClCO2CH2CH3 (31; 50%), or ClCON(CH3)2, THF, -78 °C to 0 °C (32;
81%), or CH3SSCH3, THF, -78 °C to 0 °C (33; 56%); (d) 160 °C, neat,
5 min, 80-90%; (e) aq NaOH, C2H5OH, 60 °C, 86%; (f) mCPBA, CH2Cl2,
rt, 82%.
Scheme 3a
a Reagentsandconditions: (a)TMP,n-BuLi,THF-78°C,ClCO2CH2CH3,
warm to -10 °C (38; 94%), or ClCO2CH2CH3, -78 °C (40; 92%), or
(C3F7CO)2O, warm to 0 °C (39; 79%); (b) CH3I, K2CO3, DMF, THF, 91%;
(c) t-BuLi, ClCO2CH2CH3, THF, -78 °C, 80-90%; (d) CF3CO2H, CH2Cl2,
rt, 90-98%.
indicated that, in addition to the dipole moment, the nature of
substituents also plays an important role in determining the
antiproliferative activity of DIM. The ester functionality was
then selected as the best choice for enhancing the dipole moment
of indole analogs without diminishing their anticancer activity.
Introducing ester groups into 10 at the C-5 and C-5′ positions
on the indole rings created 35 (SR13661), which has an
improved dipole moment compared to 10 (Scheme 3 and Table
4). Because fluorine groups are commonly used to increase in
vivo drug absorption, we next replaced the carbonate group of
35 with a heptafluoropropyl group to create 36 (SR13667),
which had dipole moment comparable with that of 10. Finally,
to maximize the dipole moment of 10, we introduced both
electron-withdrawing ester groups and an electron-donating
methoxyl group into the same molecule to create 37 (SR13668),
which had greater dipole moment than any other analog.
Although all three new indole analogs (35-37) exhibited in
vitro anticancer activity comparable with that of SR13650, in
vivo evaluation of their antitumor activity against MCF-7
showed that 35-37 indeed exhibited improved tumor growth
inhibition (45-60%), as compared to SR13650 (26%), at 10
mg/kg/day after 30 days of treatment. SR13668, with the greatest
Figure 3. Tumor growth inhibition effects of SR136668 in (A)
estrogen-independent MDA-MB-231 breast, (B) androgen-independent
PC-3 prostate, and (C) highly drug-resistant SKOV-3 ovarian tumor
xenografts. For each study, animals were dosed when tumor sizes
reached 70-100 mm3. The control and test groups consisted of 7-10
mice/group. SR13668 was orally administered daily alone (A) or in
combination with taxol (B and C). SR13668 and Taxol alone inhibited
tumor growth, but a combination of SR13668 and taxol significantly
reduced tumor burden by 70-90% (B and C). qd × 4 means continuous
treatment for 4 days (day 1 to day 4); bars, SE; all treatment vs control
P < 0.01 (ANOVA), except those marked with * (P < 0.05).
dipole moment, exhibited the most potent oral antitumor effect
in MCF-7 tumor models, so SR13668 was selected for further
evaluation in estrogen-independent MDA-MB-231 breast, PC-3
prostate, and SKOV-3 ovarian tumor xenograft models
(Figure 3).
In prostate and ovarian tumor models, we also assessed the
combination effect of SR13668 and Taxol, a commonly used
chemotherapeutic agent. SR13668 exhibited potent oral antitu-
mor activity in all tumor models tested. A pharmacokinetic (PK)
study using [14C]-labeled SR13668 in tumor-bearing athymic
mice showed that SR13668 was distributed extensively into the

Letters
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15 3415
inherit I3C’s unique anticancer action to inhibit various cancers
while sparing normal, nonneoplastic tissues. These results are
exciting because a diet-based compound may be safer for use
in long-term cancer treatment to maximize disease control while
minimizing undesirable side effects.
Acknowledgment. This work was supported by California
Breast Cancer Research Program Grants 4KB-0040, 6JB-0072,
and 9WB-0110 and Department of Defense Prostate Research
Cancer Program Grants DAMD 17-03-1-0021 and CCRP
3PF0126. The preclinical development of SR13668 has been
supported by the NCI/RAPID program.
Figure 4. Western blot analysis of pAkt and p-GSK3â in SR13668-
treated PC-3 cells. PC-3 cells were starved in serum-free medium. After
24 h, the cells were pretreated with SR13668 at various concentrations
for 10 min before stimulation with EGF (10 ng). Cell lysates were
prepared 2 h after EGF stimulation and used for Western blot analysis.
tissues and could be detected in tumors 72 h after a single oral
dose. However, increasing the oral dose of SR13668 did not
proportionally increase its exposure level, which may explained
why SR13668 seemed to lack dose-dependent tumor inhibition
in the MDA-MB-231 tumor xenograft studies (Figure 3A). We
reasoned that this result probably reflected the poor solubility
of SR13668 in our oral vehicle (0.5% aqueous hydroxypropyl-
cellulose), where SR13668 formed a suspension that might limit
in vivo absorption. As a result of an effort to improve the oral
formulation of SR13668, its blood concentration was markedly
increased (∼30-fold) and its oral antitumor effect could be
observed at dose levels as low as 1 mg/kg. To ensure that our
novel indole analog also inherited I3C’s unique anticancer action
to inhibit the Akt signaling pathway, we conducted similar
biological studies using SR13668. SR13668 blocked EGF-
stimulated Akt activation and inactivated its downstream effector
GSK3â in PC-3 prostate cancer cells in a dose-dependent
manner (Figure 4). SR13668 was also able to inhibit serum-
stimulated Akt phosphorylation in PC-3 and high pAkt ex-
pressed MDA-MB-468 breast cancer cells.
Inhibition of the PI3K/Akt signaling pathway by PI3K
inhibitor has been shown to induce glucose intolerance and
hyperinsulinemia. For example, wortmannin, an irreversible
PI3K inhibitor, is known to cause an exaggerated increase in
blood glucose concentration. Assessing the effect of SR13668
on glucose metabolism in mice showed that SR13668 has no
adverse effects on the fasting glucose levels or body weights
after 14 days of oral treatment with SR13668 at 500 mg/kg/
day, a dose more than 50 times higher than the dose needed for
antitumor activity. Screening a broad selection of 32 kinase
targets, including Akt(1,2,3), PI3K, and PDK1, indicated that
SR13668 is not a kinase inhibitor. SR13668 appears to inhibit
Akt activation by blocking growth factor-stimulated Akt phos-
phorylation, and it does not target the ATP substrate binding
site. Preclinical safety studies showed that SR13668 is not
genotoxic in Ames mutagenicity tests or in the mouse micro-
nucleus test. In a 14-day toxicity study in which Sprague-
Dawley rats were orally dosed with SR13668 at 25, 75, 200, or
600 mg/kg/day, no drug-related mortality, changes in body or
organ weights, or other signs of significant toxicity were seen
at any dose level. In summary, SR13668, developed and
improved from dietary I3C’s active oligomers, appeared to
Supporting Information Available: Experimental details,
characterization data, and details for in vitro and in vivo assays.
This material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) NCI, DCPC Chemoprevention Branch and Agent Development
Committee. Clinical development plan: Indole-3-carbinol. J. Cell
Biochem. Suppl. 1996, 26, 127-136.
(2) Broadbent, T. A.; Broadbent, H. S. 1s1. The chemistry and
pharmacology of indole-3-carbinol (indole-3-methanol) and 3-(meth-
oxymethyl)indole [Part I]. Curr. Med. Chem. 1998, 5, 337-352.
(3) Broadbent, T. A.; Broadbent, H. S. 1. The chemistry and pharmacol-
ogy of indole-3-carbinol (indole-3-methanol) and 3-(methoxymethyl)-
indole [Part II]. Curr. Med. Chem. 1998, 5, 469-491.
(4) Howells, L. M.; Gallacher-Horley, B.; Houghton, C. E.; Manson,
M. M.; Hudson, E. A. Indole-3-carbinol inhibits protein kinase B/Akt
and induces apoptosis in the human breast tumor cell line MDA
MB468 but not in the nontumorigenic HBL100 line. Mol. Cancer
Ther. 2002, 1, 1161-1172.
(5) Li, Y.; Chinni, S. R.; Sarkar, F. H. Selective growth regulatory and
pro-apoptotic effects of DIM is mediated by AKT and NF-kappaB
pathways in prostate cancer cells. Front Biosci. 2005, 10, 236-243.
(6) Chinni, S. R.; Sarkar, F. H. Akt inactivation is a key event in indole-
3-carbinol-induced apoptosis in PC-3 cells. Clin. Cancer Res. 2002,
8, 1228-1236.
(7) West, K. A.; Castillo, S. S.; Dennis, P. A. Activation of the PI3K/
Akt pathway and chemotherapeutic resistance. Drug Resist. Updates
2002, 5, 234-248.
(8) Anderton, M. J.; Manson, M. M.; Verschoyle, R. D.; Gescher, A.;
Lamb, J. H.; Farmer, P. B.; Steward, W. P.; Williams, M. L.
Pharmacokinetics and tissue disposition of indole-3-carbinol and its
acid condensation products after oral administration to mice. Clin.
Cancer Res. 2004, 10, 5233-5241.
(9) De Kruif, C. A.; Marsman, J. W.; Venekamp, J. C.; Falke, H. E.;
Noordhoek, J.; Blaauboer, B. J.; Wortelboer, H. M. Structure
elucidation of acid reaction products of indole- 3-carbinol: detection
in vivo and enzyme induction in vitro. Chem.-Biol. Interact. 1991,
80, 303-315.
(10) Nachshon-Kedmi, M.; Fares, F. A.; Yannai, S. Therapeutic activity
of 3,3′-diindolylmethane on prostate cancer in an in vivo model.
Prostate 2004, 61, 153-160.
(11) Reed, G. A.; Arneson, D. W.; Putnam, W. C.; Smith, H. J.; Gray, J.
C.; Sullivan, D. K.; Mayo, M. S.; Crowell, J. A.; Hurwitz, A. Single-
dose and multiple-dose administration of indole-3-carbinol to
women: Pharmacokinetics based on 3,3′-diindolylmethane. Cancer
Epidemiol., Biomarkers PreV. 2006, 15, 2477-2481.
(12) Bonnesen, C.; Eggleston, I. M.; Hayes, J. D. Dietary indoles and
isothiocyanates that are generated from cruciferous vegetables can
both stimulate apoptosis and confer protection against DNA damage
in human colon cell lines. Cancer Res. 2001, 61, 6120-6130.
JM070040E