Novel benzofuran-3-one indole inhibitors of PI3 kinase-α and the mammalian target of rapamycin: Hit to lead studies

Article abstract of DOI:10.1016/j.bmcl.2010.02.082

A series of benzofuran-3-one indole phosphatidylinositol-3-kinases (PI3K) inhibitors identified via HTS has been prepared. The optimized inhibitors possess single digit nanomolar activity against p110α (PI3K-α), good pharmaceutical properties, selectivity versus p110γ (PI3K-γ), and tunable selectivity versus the mammalian target of rapamycin (mTOR). Modeling of compounds 9 and 32 in homology models of PI3K-α and mTOR supports the proposed rationale for selectivity. Compounds show activity in multiple cellular proliferation assays with signaling through the PI3K pathway confirmed via phospho-Akt inhibition in PC-3 cells.

Full text of DOI:10.1016/j.bmcl.2010.02.082

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2586–2590
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel benzofuran-3-one indole inhibitors of PI3 kinase-
mammalian target of rapamycin: Hit to lead studies
a and the
Matthew G. Bursavich a, Natasja Brooijmans b, Lawrence Feldberg c, Irwin Hollander c, Stephen Kim c,
Sabrina Lombardi a, Kaapjoo Park a, Robert Mallon c, Adam M. Gilbert a,
*
a Medicinal Chemistry, Chemical Sciences, Wyeth Research, Pearl River, NY 10965, USA
b Computational Chemistry, Chemical Sciences, Wyeth Research, Pearl River, NY 10965, USA
c Discovery Oncology, Wyeth Research, Pearl River, NY 10965, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of benzofuran-3-one indole phosphatidylinositol-3-kinases (PI3K) inhibitors identified via HTS
Received 17 January 2010
Revised 18 February 2010
Accepted 19 February 2010
Available online 23 February 2010
has been prepared. The optimized inhibitors possess single digit nanomolar activity against p110
(PI3K- ), good pharmaceutical properties, selectivity versus p110 (PI3K- ), and tunable selectivity ver-
sus the mammalian target of rapamycin (mTOR). Modeling of compounds 9 and 32 in homology models
of PI3K- and mTOR supports the proposed rationale for selectivity. Compounds show activity in multiple
a
a
c
c
a
cellular proliferation assays with signaling through the PI3K pathway confirmed via phospho-Akt inhibi-
tion in PC-3 cells.
Keywords:
Kinase
Phosphatidylinositol-3-kinases
Mammalian target of rapamycin
Ó 2010 Elsevier Ltd. All rights reserved.
Phosphatidylinositol-3-kinases (PI3K) are lipid kinases that
phosphorylate the 30-hydroxyl position of phosphatidylinositol
4,5-diphosphate (PIP2) to phosphatidylinositol 3,4,5-triphosphate
(PIP3). Eight PI3Ks have thus far been identified and categorized
into four classes (IA, IB, II and III) based on substrate specificity
and sequence homology. Deregulation of class IA PI3Ks, namely
We have recently published our efforts developing purines and
pyrazolo[3,4-d]pyrimidines as selective inhibitors of both PI3K-
a
and mTOR.8,9 Herein we report the hit to lead optimization and
structure–activity relationships (SAR) of benzofuran-3-one indoles,
another class of novel PI3K-a inhibitors with good selectivity ver-
sus PI3K- and mTOR.
c
p110
a
(PI3K-
a
), leads to elevated PIP3 levels and downstream acti-
The synthesis of the benzofuran-3-one indole inhibitors is
shown in Scheme 1.10 The synthesis begins with commercially
available indoles 2 or their construction from 1 via standard Suzuki
cross-coupling conditions.
vation of Akt.1,2 This ultimately promotes growth, survival, prolif-
eration, enhanced migration, and adhesion in cancer cells.
Additionally, PI3K-
a deregulation is involved a quarter of breast
cancers in all stages. Also, PTEN (the phosphatase that regulates
PIP3 levels) deregulation is associated with resistance to chemo-
therapeutic agents that target the EGFR/Her2 signaling pathway.3,4
Indoles 2 were formylated with POCl3 in DMF to give 3.11 In
some instances, the indole nitrogens of 3 were methylated with
MeI using NaH in DMF.12 Condensation of 4,6-dihydroxybenzofu-
ran-3(2H)-one 4 with indoles 3 under catalytic acid conditions pro-
vided the desired benzofuran-3-one indoles 5.13
Thus PI3K-a has emerged as a premier target for cancer treatment
and the focus of many drug discovery efforts.5
A high-throughput screen of our compound library identified a
series of benzofuran-3-one indoles (Fig. 1) with activity against
Compounds were assayed using PI3K-a and PI3K-c fluorescence
polarization assays to monitor PIP3 production. Compounds were
PI3K-a, p110c (PIK3-c), a related PI3K subtype (class IB) targeted
further assayed for mTOR potency using the DELFIA format
for autoimmune and inflammation diseases,6 and the mammalian
target of rapamycin (mTOR), a related kinase targeted in cancer
chemotherapy that signals downstream of Akt.7 Since many of
the effects of Akt are mediated via mTOR, a dual PI3K/mTOR inhib-
itor has the potential to prevent cancer cell proliferation and in-
duce apoptosis by fully suppressing Akt activation.
OH
O
R2
O
HO
R1
N
R3
* Corresponding author. Tel.: +1 845 602 4865; fax: +1 845 602 5561.
E-mail address: gilbera@wyeth.com (A.M. Gilbert).
Figure 1. Benzofuran-3-one indole PI3K inhibitor scaffold identified via HTS.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.02.082

M. G. Bursavich et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2586–2590
2587
OH
O
OHC
R2
R2
R2
a
b
d
Br
R2
R1
R1
N
H
O
N
H
N
HO
R1
R
N
R
1
2
3
R = H
R = Me
c
5
Scheme 1. Reagents and conditions: (a) R1B(OH)2, Na2CO3, cat. Pd(PPh3)4, dioxane, reflux; (b) POCl3, DMF, 0–23 °C; (c) MeI, NaH, DMF; (d) 4,6-dihydroxybenzofuran-3(2H)-
one, cat. HCl, EtOH, reflux.
monitoring phosphorylated His6-SK6. Cellular proliferation activ-
ity was determined by monitoring cell growth densities in Caco2,
LoVo, and PC3MM2 cell cultures.14 Mechanism based cellular as-
says of PI3K signaling in PC-3 cells was determined by immuno-
blotting protein lysates for phosphorylated-Akt (p-Akt) and
phosphorylated 4E binding protein 1 (p-4EBP1).
The 5-indole (R2) substituent was initially varied while main-
taining the R1 substituent as a proton, methyl, or aryl group (Table
1). The importance of the R2 substituent was evident in the early
analogs where R1 = H. Changing R2 to a halogen (7 and 8) gives a
significant PI3K-a potency improvement over 6 with at least five-
fold selectivity versus PI3K-
c
and mTOR. Incorporation of an
Table 1
Structure–activity relationships of benzofuran-3-one indoles
OH
O
R2
O
HO
R1
N
H
R1
R2
PI3K-
a
IC50 (nM)
PI3K-
c
IC50 (nM)
mTOR IC50 (nM)
Sol.b
a
a
a
Compd
6
7
8
H
H
H
H
H
Me
Me
Me
Phenyl
2-Pyridyl
(4-F)Phenyl
(4-Cl)Phenyl
2-Naphthyl
Phenyl
3-Pyridyl
H
Cl
Br
OMe
OBn
H
Br
OMe
H
H
H
H
H
OMe
OMe
6754
492
422
30
1889
3197
179
3
1034
3650
1580
542
800
2
nt
9200
11,000
3700
265
5000
500
190
3
130
22
123
190
280
10
14
1
0
12
1
0
1
1
1
3
3628
1625
269
nt
nt
2220
25
2781
nt
nt
1458
1820
11
9
10
11
12
13
14
15
16
17
18
19
20
1
1
1
0
1
10
1
1
nt = not tested.
a
Values are means of two experiments, standard deviations are 10%.
Aqueous solubility determined at pH 7.4. Values in lg/mL.
b
Table 2
Structure–activity relationships of benzofuran-3-one N-methyl indoles
OH
O
R2
O
HO
R1
N
Me
R1
R2
PI3K-
a
IC50 (nM)
PI3 K-
c
IC50 (nM)
mTOR IC50 (nM)
Sol.b
a
a
a
Compd
21
22
23
24
H
H
H
1106
106
7
6949
587
54
800
62
4
2
1
1
2
OMe
OMe
H
Me
Phenyl
2665
nt
96
nt = not tested.
a
Values are means of two experiments, standard deviations are 10%.
Aqueous solubility determined at pH 7.4. Values in lg/mL.
b

2588
M. G. Bursavich et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2586–2590
Several analogs were synthesized to investigate the role of an N-
OHC
R1
OHC
methyl indole substituent (Table 2). The R1 = R2 = H indole 21
shows modest kinase potency. Addition of an R2 = OMe substituent
(22) improves potency against all three enzymes by one log unit.
OMe
OMe
a, b
R1
N
N
H
The R1 = Me, R2 = OMe analog 23 shows excellent PI3K-
a
and
. Com-
pound 24 incorporating an R1 phenyl group interestingly shows
good mTOR potency with 26-fold selectivity over PI3K-
3
R2 n
mTOR potency with nearly 10-fold selectivity over PI3K-
c
25
a.
n = 1, 2
A quick look at the pharmaceutical properties of these benzo-
furan-3-one indoles reveals potential solubility and CYP 3A4 inhi-
c
bition liabilities. For the compounds with PI3K-
(i.e., 9, 13, 19, 20, and 23) the respective CYP 3A4% inhibition at
M is 58%, 79%, 77%, 57%, and 75%. To address these potential
c IC50 <100 nM
OH
O
3
l
OMe
issues and further investigate the selectivity potential, a series
of indoles containing water-solubilizing groups were designed
and prepared.
O
HO
R1
N
Construction of these benzofuran-3-one N-alkyl amino indoles
is shown in Scheme 2.10 The synthesis began with NaH-mediated
alkylation15 of indole 3 to provide the alkyl chloride intermediates
which were then reacted with various amines to provide N-alkyl
amino indoles 25. Condensation of 4,6-dihydroxybenzofuran-
3(2H)-one 4 with indole 25 under catalytic acidic conditions pro-
vided the desired benzofuran-3-one indoles 26.13
R2n
26
n = 1, 2
Scheme 2. Reagents and conditions: (a) BrCH2CH2Cl or BrCH2CH2CH2Cl, NaH, DMF;
(b) amine, K2CO3, KI, ACN, reflux; (c) 4,6-dihydroxybenzofuran-3(2H)-one, cat. HCl,
EtOH, reflux.
Examination of the resulting benzofuran-3-one N-alkyl amino
indoles reveals excellent PI3K-
mTOR selectivity (Table 3). Among the analogs containing a two
carbon linker (n = 1), the R1 = H analog 27 shows good PI3K-
po-
tency and the greatest selectivity over mTOR. The R1 = Me analogs
(28–31) afford potent single digit nM PI3 K- and similar mTOR
potency with 10–80-fold selectivity over of PI3K- . All of the ana-
logs with a three carbon linker (n = 2) and R1 = H show good PI3K-
potency and greater than 10-fold selectivity over both PI3K- and
mTOR. As seen previously, the corresponding R2 methyl analogs
(35–38) display more potent PI3K- activity with 10–80-fold
selectivity versus PI3K- and/or mTOR. Addition of the N-alkyl
a potency and tunable PI3K-c and
R2 = OMe substituent in 9 further improves PI3K-
maintaining 10-fold selectivity over PI3K- and mTOR. The
R2 = OBn analog 10 shows a substantial potency reduction com-
pared to the corresponding OMe analog 9. PI3K- potency can be
improved by adding an R1 = Me substituent. While the R2 = H ana-
log 11 possesses low- M PI3K- activity, the halogenated analog
12 and the OMe-substituted 13 show significant potency increases.
Compound 13 is a single digit nM PI3K- /mTOR inhibitor with
nearly 10-fold selectivity over PI3K-
. The series of R1 aryl substi-
tuted analogs with R2 = H substituents (14–18) showed good
mTOR potency with selectivity over both PI3K- and PI3K- . As
seen with the R1 = H and Me series, R2 = OMe analogs 19 and 20
show excellent PI3K- potency and demonstrate selectivity versus
PI3K- and mTOR.
a potency while
a
c
a
a
c
a
l
a
c
a
a
c
c
amino indoles also improves the solubility of this scaffold and
moderates the CYP 3A4 liability. Among the N-alkyl amino indoles
a
c
27–38 only 29, 33 and 36 possess a CYP 3A4% inhibition at 3
l
M
a
>30% (64%, 42%, and 68%, respectively).
c
Table 3
Structure–activity relationships of benzofuran-3-one N-alkylamine indoles
OH
O
OMe
O
HO
R1
N
nn = 1,2
R2
R1
R2
PI3K-a
IC50 (nM)
PI3K-c IC50 (nM)
mTOR IC50 (nM)
Sol.b
a
a
a
Compd
n
27
28
29
30
31
32
33
34
35
36
37
38
1
1
1
1
1
2
2
2
2
2
2
2
H
N-(CH2)2OH piperazine
NMe2
Morpholine
N-Me piperazine
N-(CH2)2OH piperazine
NMe2
Morpholine
N-(CH2)2OH piperazine
NMe2
Morpholine
N-Me piperazine
N-(CH2)2OH piperazine
36
9
11
5
976
283
155
473
322
2419
646
1358
404
76
215
18
9
11
7
2125
430
1650
118
18
nd
1
6
30
30
32
3
25
32
1
Me
Me
Me
Me
H
H
H
Me
Me
Me
Me
4
40
42
37
5
8
15
6
402
109
153
108
10
36
nd = not determined.
a
Values are means of two experiments, standard deviations are 10%.
Aqueous solubility determined at pH 7.4. Values in lg/mL.
b

M. G. Bursavich et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2586–2590
2589
Figure 2. (A) Compound 9 docked in PI3K-
are shown in blue dashed lines to the hinge region (Val851), the catalytic lysine (Lys802), the DFG-motif aspartic acid (Asp933), and Tyr836. The indole nitrogen points
towards one of the residues in the specificity surface (His855) and solvent. (B) Overlay of docked binding modes of compound 32 in complex with PI3K- and mTOR
homology models. Overlay of the predicted binding mode of 32 with PI3K (green and dark green carbons) and mTOR (magenta and purple carbons). Hydrogen bonds are
shown in green (PI3K- ) and pink (mTOR) dashed lines. The solubilizing tail cannot form hydrogen bonding interactions with the PI3K- histidine (His855), but in mTOR can
potentially form a hydrogen bond to Asp2244.
a homology model. Close-up of the PI3K-a binding site in complex with 9 based on docking studies. Hydrogen bonding interactions
a
a
a
Docking models of 9 in a PI3K-
a
homology model (based on a
and mTOR homol-
the substituent/inhibitor. It is also possible that the quaternary
nitrogen has an electrostatic clash with Lys2171 in mTOR, which
published structure of PI3K- ) and 32 in PI3K-
c
a
ogy models are shown inFigure 2A and B, respectively. The docking
model of 9 reveals the 5-OMe substituent binds to the hinge region
(Val851), and the OH groups bind to the catalytic lysine (Lys802),
the DFG-motif aspartic acid (Asp933), and Tyr836. Docking studies
of 32 show the N-alkylamine substituent points towards the selec-
tivity surface and solvent. R2 Substituents are not predicted to form
hydrogen bonding interactions with the protein and these substit-
uents thus function as solubilizing tails, with relatively little im-
pact on potency.
In mTOR on the other hand, an aspartic acid (Asp2244) is pres-
ent in the selectivity surface allowing the possible formation of a
hydrogen bond with the quaternary nitrogen of the R2 substituent
(i.e., NMe2 in 32 and 35, morpholine in 33 and 36). Despite the pos-
sible formation of a hydrogen bond with the enzyme, some R2 sub-
stituents decrease potency against mTOR significantly as seen in
32. One possible explanation for the potency decrease is that the
hydrogen bond formation results in a decrease of the entropy of
is a tryptophan in PI3K-
a.
Several of the most potent PI3K-
a
analogs were tested for cellu-
lar proliferation activity against Caco-2 epithelial colorectal adeno-
carcinoma cells, LoVo human colon carcinoma cells, and PC3
human prostate cancer cells (Table 4). Inspection of the cellular
data reveals the biochemical single digit nanomolar, dual PI3K-a/
mTOR inhibitors trend to more promising cellular potency (com-
pare 19–20 with 35–38). It also appears that for the LoVo and
PC3 data, the better compounds tend to have higher clog P values
(ꢀ>4.3).
To ensure that compounds were inhibiting PI3K signaling in
cells, compound 19 was tested for p-Akt (at Thr308) inhibition in
PC-3 human prostate cancer cells. Immunostaining for phosphory-
lated 4E binding protein 1 (p4EBP1) was also conducted as a mar-
ker of pathway inhibition downstream of mTOR ( Fig. 3). p-Akt
inhibition is seen starting at 0.3
sitometer assessment of the p-Akt inhibition blots indicates an
lM and is complete at 3 lM. Den-
Table 4
Cellular proliferation inhibition of selected benzofuran-3-one indoles
a
a
a
a
a
Compd
PI3K-a
IC50 (nM)
mTOR IC50 (nM)
Caco IC50
(l
M)
LoVo IC50
(lM)
PC3 IC50
(lM)
clog P
13
19
20
23
28
29
30
31
35
36
37
38
3
2
1
7
9
11
5
4
5
8
3
9
1
4
18
9
11
7
118
18
153
108
9.0
2.1
7.5
7.2
2.8
2.0
4.7
9.0
9.3
4.7
10.0
10.0
3.6
1.2
4.7
2.5
1.7
2.3
2.2
5.9
5.7
4.6
7.5
10.0
4.0
0.8
1.9
1.3
1.8
1.5
2.4
6.9
2.8
2.4
9.5
10.0
4.19
5.80
4.38
4.70
4.82
4.67
3.74
3.20
5.07
5.03
4.10
3.57
15
7
a
Values are means of two experiments; standard deviation is 10%.

2590
M. G. Bursavich et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2586–2590
tion in PC-3 cells. This lead series provides the basis for
developing more potent, novel inhibitors of the PI3K signaling
pathway with the desired selectivity profile.
References and notes
1. Vanhaesebroeck, B.; Leevers, S. J.; Ahmadi, K.; Timms, J.; Katso, R.; Driscoll, P.
C.; Woscholski, R.; Parker, P. J.; Waterfield, M. D. Annu. Rev. Biochem. 2001, 70,
535.
2. Liu, P.; Cheng, H.; Roberts, T. M.; Zhao, J. J. Nat. Rev. Drug Disc. 2009, 8, 627.
3. Saal, L. H.; Holm, K.; Maurer, M.; Memeo, L.; Su, T.; Wang, X.; Yu, J. S.;
Malmstrom, P. O.; Mansukhani, M.; Enoksson, J.; Hibshoosh, H.; Borg, A.;
Parsons, R. Cancer Res. 2005, 65, 2554.
Figure 3. Inhibition of PI3K signaling in PC-3 cells. PC-3 cells were treated with
DMSO, 10 lM LY 294022 (positive control) or the indicated doses of 19 after 6 h.
4. She, Q. B.; Solit, D.; Basso, A.; Moasser, M. M. Clin. Cancer Res. 2003, 9, 4340.
5. Wee, S.; Lengauer, C.; Wiederschain, D. Curr. Opin. Oncol. 2008, 20, 77.
6. Ameriks, M. K.; Venable, J. D. Curr. Top. Med. Chem. 2009, 9, 738.
7. Tsang, C. K.; Qi, H.; Liu, L. F.; Zheng, X. F. S. Drug Discovery Today 2007, 12, 112.
8. Gilbert, A. M.; Nowak, P.; Brooijmans, N.; Bursavich, M. G.; Dehnhardt, C.;
Santos, E. D.; Feldberg, L. R.; Hollander, I.; Kim, S.; Lombardi, S.; Park, K.;
Venkatesan, A. M.; Mallon, R. Bioorg. Med. Chem. Lett. 2010, 20, 636.
9. Venkatesan, A. M.; Dehnhardt, C. M.; Chen, Z.; Santos, E. D.; Dos Santos, O.;
Bursavich, M.; Gilbert, A. M.; Ellingboe, J. W.; Ayral-Kaloustian, S.; Khafizova,
G.; Brooijmans, N.; Mallon, R.; Hollander, I.; Feldberg, L.; Lucas, J.; Yu, K.;
Gibbons, J.; Abraham, R.; Mansour, T. S. Bioorg. Med. Chem. Lett. 2010, 20, 653.
10. All newly prepared compounds were characterized by 1H NMR and reversed
phases-HPLC/MS spectroscopy.
11. Gastpar, R.; Goldbrunner, M.; Marko, D.; von Angerer, E. J. Med. Chem. 1998, 41,
4965.
12. Tanaka, M.; Kaneko, T.; Akamatsu, H.; Okita, M.; Chiba, K.; Obaishi, H.; Sakurai,
H.; Yamatsu, I. Eur. J. Med. Chem. 1995, 30, 179.
13. Teo, C. C.; Kon, O. L.; Sim, K. Y.; Ng, S. C. J. Med. Chem. 1992, 35, 1330.
14. Yu, K.; Toral-Barza, L.; Discafani, C.; Zhang, W. G.; Skotnicki, J.; Frost, P.;
Gibbons, J. J. Endocr. Relat. Cancer 2001, 8, 249.
15. de la Mora, M. A.; Cuevas, E.; Muchowski, J. M.; Cruz-Almanza, R. Tetrahedron
Lett. 2001, 42, 5351.
Protein lysates were prepared and immunoblotted for phospho-Akt T308 (p-Akt
(T308)) and phosphorylated 4E binding protein 1 (p4EBP1).
IC50 = 385 nM. Suppression of p-Akt is indicative of inhibition of
cellular PI3Ks which are upstream of KT in the PI3K signaling
pathway.
In conclusion, we have presented a series of the benzofuran-3-
one indole PI3K inhibitors identified via HTS efforts. Structure
based design with a PI3K-
PI3K- X-ray structure led to compounds with improved selectiv-
ity, solubility and CYP inhibition. Our hit to lead optimization has
identified potent PI3K- inhibitors with good selectivity versus
PI3K- and mTOR. This scaffold has also provided dual PI3K-
mTOR inhibitors with good selectivity over PI3K- . Compounds
a homology model derived from a
c
a
c
a/
c
showed activity in various cellular proliferation assays with signal-
ing through the PI3K pathway confirmed via phospho-Akt inhibi-