Discovery of novel morpholino-quinoxalines as PI3Kα inhibitors by pharmacophore-based screening

Article abstract of DOI:10.1039/c2md00255h

A pharmacophore model of PI3Kα inhibitors was built using the DiscoveryStudio 2.0 package. Pharmacophore-based screening (PBS) retrieved a series of novel morpholino-quinoxalines as PI3Kα inhibitors, as exemplified by 1a (PI3Kα IC₅₀: 0.44 μM).

Full text of DOI:10.1039/c2md00255h

View Online / Journal Homepage
Dynamic Article Links <
C
MedChemComm
Cite this: DOI: 10.1039/c2md00255h
www.rsc.org/medchemcomm
CONCISE ARTICLE
Discovery of novel morpholino–quinoxalines as PI3Ka inhibitors by
pharmacophore-based screening†‡
Peng Wu,a Yi Su,b Xiaowen Liu,b Jingying Yan,a Yong Ye,a Lei Zhang,b Jianchao Xu,a Shaoyu Weng,a Yani Li,a
a
Tao Liu, Xiaowu Dong, Maotang Sun, Bo Yang, Qiaojun He and Yongzhou Hu
a
a
b
b
a
*
Received 9th October 2011, Accepted 3rd April 2012
DOI: 10.1039/c2md00255h
A pharmacophore model of PI3Ka inhibitors was built using the DiscoveryStudio 2.0 package.
Pharmacophore-based screening (PBS) retrieved a series of novel morpholino–quinoxalines as PI3Ka
inhibitors, as exemplified by 1a (PI3Ka IC50: 0.44 mM). All target compounds showed good in vitro
cytotoxicity against tested human cell lines. A pharmacophore mapping analysis and docking study
indicated that both the morpholino group and the sulfonyl group contributed significantly to the potent
PI3Ka inhibitory activity and cytotoxicity of the compounds.
The phosphoinositide 3-kinase (PI3K) pathway is one of the
most highly activated and mutated signaling cascades for a wide
spectrum of cancers.1,2 The PI3Ks are lipid kinases which differ
in substrate preference and regulation mechanism.3 Among
these, PI3Ka has emerged as a promising target for cancer
therapy due to prevalent gain-of-function mutations which have
been observed in Pik3ca—the gene which encodes the p110a
catalytic subunit of PI3Ka.4 In addition, loss-of-function muta-
tions in the phosphotase and tensin homologue (PTEN)
contribute indirectly to the activation of the PI3K pathway.5
Thus, considerable interests are currently being focused on
developing inhibitors targeting PI3K, especially PI3Ka, for the
treatment of cancer.6–9
The pharmacophore model was built based on 75 reported
PI3K inhibitors with varied potency and chemical structures,
among which 25 were selected to be a training set based on the
principle of structural diversity and wide coverage of activity
range (ESI, Fig. S1†), the remaining 50 molecules with maximal
3D diversity and continuous bioactivity magnitude constituted
a test set (ESI, Fig. S2†).11 Ten pharmacophore models of PI3Ka
inhibitors were generated by using the Catalyst package 4.11
(ESI, Table S1 and S2†). The best pharmacophore model
(Hypo1), evaluated in terms of cost functions and other statis-
tical parameters such as correlation coefficient and RMSD value,
featured four pharmacophore elements: two hydrogen bond
acceptors (HBA), one hydrophobic element (HY) and one ring
aromatic element (RA) (Fig. 1). Hypo1 was then validated using
the test set containing 50 reported PI3K inhibitors (see ESI,
Fig. S3 and Table S3†) and CatScarmble randomization method
(see ESI, Table S4†).
As a part of our group’s continued work on developing small-
molecule inhibitors targeting the PI3K signaling pathway,10
a pharmacophore model was built in an attempt to retrieve novel
PI3Ka inhibitors. Pharmacophore-based screening against an
in-house database revealed a series of morpholino–quinoxalines
as potential PI3Ka inhibitors. Synthesized target compounds
showed good to excellent in vitro cytotoxicity against tested
human cell lines. Further pharmacophore mapping and docking
studies indicated that the morpholino group and the sulfonyl
group were important pharmacophores for PI3K inhibition. The
identified leads, as exemplified by 1a, provided starting points for
further modificaiton and optimization.
DiscoveryStudio 2.0/Ligand Pharmacophore mapping
protocol was used with Hypo1 as the pharmacophore model to
screen an in-house database consisting of approximately three-
thousand compounds that were synthetically available in our
laboratory to retrieve molecules as novel PI3K inhibitors.
aZJU-ENS Joint Laboratory of Medicinal Chemistry, College of
Pharmaceutical Sciences, Zhejiang University, Hangzhou 310058, China
bInstitute of Pharmacology and Toxicology, College of Pharmaceutical
Sciences, Zhejiang University, Hangzhou 310058, China. E-mail: huyz@
zju.edu.cn; Fax: +86 571 8820 8460; Tel: +086 571 8820 8460
† Electronic supplementary information (ESI) available: Pharmacophore
establishment and validation, experimental data, materials and methods
for cytotoxicity and PI3Ka enzymatic assay. See DOI:
10.1039/c2md00255h
Fig. 1 (a) The best pharmacophore model (Hypo1) with four elements;
‡ The authors have declared no conflict of interest.
and (b) 3D spatial relationship and distance constrains of Hypo1.
This journal is ª The Royal Society of Chemistry 2012
Med. Chem. Commun.

View Online
Among the molecules with high fit-values and potent predictive
activity, the interesting lead, 2-morpholino-3-phenyl-
sulfonylquinoxaline (1a, predictive PI3Ka IC50 of 1.0 mM) which
contains a morpholinoaryl scaffold—as shown in the widely
studied PI3K inhibitor LY294002, which was believed to be one
of the privileged structures for PI3K inhibition12—was selected
for further synthesis and biological evaluation. 2-Morpholino-3-
phenylthioquinoxaline (2, predictive PI3Ka IC50 of 38.0 mM),
the predictively inactive PI3K compound that shares the mor-
pholinoaryl scaffold with 1a but lacks the sulfonyl group, was
also synthesized to test the importance of the sulfonyl group for
PI3K inhibition. In addition, 2-piperidinyl-3-phenyl-
sulfonylquinoxaline 3, which has a piperidinyl group rather than
a morpholino group at the 2-position of the quinoxaline ring,
was synthesized to evaluate the impact of the morpholino group
on PI3K inhibition (Fig. 2).
The 2-chloro-3-arylsulfonylquinoxaline intermediates 7a–e
were synthesized via the method reported previously by our
group starting from the arylsulfonylchlorides 4a–e to give the
arylsulfonylhydrazides 5a–e, whose reaction with 2,3-dichlor-
oquinoxaline 6 affored 7a–e.10 A final step of 7a–e with mor-
pholine in isopropyl alcohol under microwave irradiation yielded
the target compounds 1a–e. Substitution of 2,3-dichlor-
oquinoxaline 6 with morpholine gave 2-morpholino-3-chlor-
oquinoxaline 8, which was further reacted by thiophenol to give
compound 2. Reaction of 2-chloro-3-phenylsulfonylquinoxaline
7a with piperidine in isopropyl alcohol under microwave irradi-
ation yielded compound 3 (Scheme 1).
Scheme 1 Synthesis of morpholino–quinoxaline derivatives 1a–e, 2, and
piperidinyl–quinoxaline 3. Conditions and reagents: (a) hydrazine
hydrate, H2O, 0 ꢀC–r.t., 90–99%; (b) ethanol, reflux, 14–32%; (c) mor-
pholine, microwave irradiation, isopropyl alcohol, 80 ꢀC, 72–97%; (d)
morpholine, isopropyl alcohol, reflux, 93%; (e) thiophenol, DMF, NaH,
r.t., 70%; and (f) piperidine, microwave irradiation, isopropyl alcohol, 80
ꢀC, 92%.
Compounds 1a–e, 2, and 3 were tested for their PI3Ka enzy-
matic inhibitory activity using a competitive fluorescence polar-
ization kinase activity assay.13 LY294002 was used as the positive
control. The results are presented in Table 1. Compound 1a
showed a PI3Ka IC50 value of 0.44 mM, which is comparable to
that of LY294002 and close to the predictive value of 1.0 mM.
Derivatives of 1a with electro-donating methyl (1b) and methoxy
(1c) groups and electro-withdrawing fluoro (1d) and bromo (1e)
groups at the 4-position of the sulfonylphenyl ring also exhibited
favorable PI3Ka inhibitory activities ranging from 1.58 to
2.89 mM. While compound 2 with a thio group at the 3-position
of the quinoxaline ring showed an IC50 value bigger than 50 mM,
and compound 3 with a piperidinyl group at the 2-position of the
quinoxlaine ring showed an IC50 value of 21.63 mM. This result
suggested that both the sulfonyl group and the morpholino
group are important for the PI3K inhibitory activity in this series
of compounds.
Compounds 1a–e, 2, and 3 were then tested for their in vitro
cytotoxicity against four human cancer cell lines, prostate cancer
cells PC3, lung adenocarcinoma epithelial cells A549, colon
cancer cells HCT116, and promyelocytic leukemia cells HL60.
LY294002 was used as the reference compound. The results are
summarized in Table 2. All morpholino–quinoxaline derivatives
1a–e showed an obviously improved cytotoxicity compared with
that of LY294002, and 1a–e were generally more active than 2
Table 1 Inhibition of PI3Ka by target compounds
IC50 (mM)a
Cpd.
Tested value
Predictive value
1a
1b
1c
1d
1e
2
3
0.44
1.58
1.40
2.89
2.49
>50
21.63
0.63
1.0
1.0
0.7
0.5
1.9
38.0
16.0
2.6
LY294002b
a
b
The mean value of at least two separate determinations. Reported
value, ref. 12.
Fig. 2 Structures of 1a, 2, 3 and LY294002.
Med. Chem. Commun.
This journal is ª The Royal Society of Chemistry 2012

View Online
Table 2 In vitro cytotoxicities of target compounds
IC50 (mM)a
a surrogate due to its amenability and the fact that PI3Ka-small
molecule inhibitor has not been resolved yet.14 As shown in
Fig. 4, both 1a and 2 form a hydrogen bond through the mor-
pholino oxygen with the hinge residue of Val 882. Disparities in
docking modes were in two facets: an additional hydrogen bond
was formed involving one of the sulfonyl oxygens with Thr887,
and the quinoxaline scaffold was oriented toward the affinity
pocket for the PI3Kg-1a complex and close to the ribose binding
pocket for the PI3Kg-2 complex.
Cpd.
PC3
A549
HCT116
HL60
1a
1b
1c
1d
1e
2
18.88
9.34
0.88
25.44
3.40
>50
12.55
3.3
5.35
10.48
>50
35.02
3.61
45.58
6.48
56.01
4.47
0.95
4.80
3.41
1.74
5.18
27.08
9.94
19.35
6.65
5.64
7.25
>50
In summary, a series of novel morpholino–quinoxalines 1a–e
were discovered to be good PI3Ka inhibitors via pharmaco-
phore-based screening (PBS). All target compounds showed
good in vitro cytotoxicity against four tested human cell lines.
Mapping analysis and docking study indicated that both the
morpholino group and the sulfonyl group were important
pharmacophores for PI3Ka inhibition in this series of
compounds. The morpholino–quinoxaline lead compounds
described in this study provide promising starting points for
further development of novel quinoxalines as PI3K inhibitors.
3
LY294002
26.94
61.35
89.65
a
The mean value of at least two separate determinations.
and 3, especially against PC3, the PI3K activated cell line. For
example, compounds 1c, 2, 3 and LY294002 showed PC3 IC50
values of 0.88, bigger than 50, 26.94 and 61.35 mM, respectively.
Further pharmacophore mapping and docking analyses were
performed to better explain the potency gap between 1a and 2,
and understand the binding mode of the target compounds.
Catalyst software package 4.11 was used with Hypo1 as the
pharmacophore model to compare the mapping mode between
1a and 2. As shown in Fig. 3, 1a was mapped well with three of
the four pharmacophore elements with the oxygen atom of the
morpholin ring and one of the sulfonyl oxygens functioning as
hydrogen bond acceptors (Fig. 3a and c). Although the mor-
pholino oxygen was also indicated as a potential hydrogen bond
acceptor in 2, the other hydrogen bond acceptor element was
absent due to the lack of the sulfonyl group in 2. The overall
mapping result also showed that the quinoxaline scaffold of 2
was outside of the pharmacophore elements due to a change in
conformation compared with that of 1a (Fig. 3b and d).
Acknowledgements
We thank Dr Yi Wang, Miss Yun Jiang and Miss Wan Gong for
assistance in performing the fluorescence polarization assay. Dr
Luke R. Odell, Dr Krzysztof Wie˛ckowski, Dr Anna Wie˛ck-
owska, Dr Francesco Russo, Mr Marc Stevens, and Mr Andreas
Blomgren are acknowledged for useful discussion and
continuous support. Financial support provided by the
Erasmus Mundus EMECW-LiSUM project is gratefully
acknowledged.
Docking analysis using the C-DOCKER program within
DiscoveryStudio 2.0 package was performed to further explore
the binding mode of the target compounds. The co-crystal
structure of PI3Kg with small-molecule inhibitor was used as
Fig. 4 Docking analysis using PI3Kg co-crystal structure as a template
(PDB ID: 3L54). (a) Ribbon show of 1a docked into the ATP binding
pocket of PI3Kg; and (b) ribbon show of 2 docked into the ATP binding
pocket of PI3Kg. Selected residues are shown in yellow backbones, 1a in
pink backbone, 2 in wheat backbone, orange dashed lines indicate
hydrogen bonds, prepared using PyMOL.
Fig. 3 Pharmacophore mapping using Hypo1: (a) Hypo1 mapped with
1a in transparent show; (b) Hypo1 mapped with 2 in transparent show;
(c) Hypo1 mapped with 1a in solid show; and (d) Hypo1 mapped with 2 in
solid show. Hydrogen bond acceptor (HBA) in green, hydrophobic (HY)
in blue, and ring aromatic (RA) in orange.
This journal is ª The Royal Society of Chemistry 2012
Med. Chem. Commun.

View Online
10 P. Wu, Y. Su, X. Liu, L. Zhang, Y. Ye, J. Xu, S. Weng, Y. Li, T. Liu,
S. Huang, B. Yang, Q. He and Y. Hu, Eur. J. Med. Chem., 2011, 46,
5540–5548.
11 P. Wu, W. H. Huang, X. W. Dong, M. T. Sun, J. C. Xu, S. Y. Weng
and Y. Z. Hu, Abstract of papers no. 105, 7th AFMC-IMCC, Cairns,
Australia, August 23–27, 2009.
References
1 P. K. Vogt, S. Kang, M.-A. Elsliger and M. Gymnopoulos, Trends
Biochem. Sci., 2007, 32, 342–349.
2 J. A. Engelman, J. Luo and L. C. Cantley, Nat. Rev. Genet., 2006, 7,
606–619.
12 M. Hayakawa, H. Kaizawa, H. Moritomo, T. Koizumi, T. Ohishi,
M. Okada, M. Ohta, S.-I. Tsukamoto, P. Parker, P. Workman and
M. Waterfield, Bioorg. Med. Chem., 2006, 14, 6847–6858.
13 B. E. Drees, A. Weipert, H. Hudson, C. G. Ferguson,
L. Chakaravarty and G. D. Prestwich, Comb. Chem. High
Throughput Screening, 2003, 6, 321–330.
3 B. Vanhaesebroeck, J. Guillermet-Guibert, M. Graupera and
B. Bilanges, Nat. Rev. Mol. Cell Biol., 2010, 11, 329–341.
4 Y. Samuels, Z. Wang, A. Bardelli, N. Silliman, J. Ptak, S. Szabo,
H. Yan, A. Gazdar, S. M. Powell, G. J. Riggins, J. K. V. Willson,
S. Markowitz, K. W. Kinzler, B. Vogelstein and V. E. Velculescu,
Science, 2004, 304, 554.
14 S. D. Knight, N. D. Adams, J. L. Burgess, A. M. Chaudhari,
M. G. Darcy, C. A. Donatelli, J. I. Luengo, K. A. Newlander,
C. A. Parrish, L. H. Ridgers, M. A. Sarpong, S. J. Schmidt,
G. S. Van Aller, J. D. Carson, M. A. Diamond, P. A. Elkins,
C. M. Gardiner, E. Garver, S. A. Gilbert, R. R. Gontarek,
J. R. Jackson, K. L. Kershner, L. Luo, K. Raha, C. S. Sherk,
C.-M. Sung, D. Sutton, P. J. Tummino, R. J. Wegrzyn,
K. R. Auger and D. Dhanak, ACS Med. Chem. Lett., 2010, 1, 39–43.
5 M. C. Hollander, G. M. Blumenthal and P. A. Dennis, Nat. Rev.
Cancer, 2011, 11, 289–301.
6 R. Marone, V. Cmiljanovic, B. Giese and M. P. Wymann, Biochim.
Biophys. Acta, Proteins Proteomics, 2008, 1784, 159–185.
7 P. Wu, T. Liu and Y. z. Hu, Curr. Med. Chem., 2009, 16, 916–930.
8 P. Wu and Y. Z. Hu, Curr. Med. Chem., 2010, 17, 4326–4341.
9 K. Kawauchi, T. Ogasawara, M. Yasuyama, K. Otsuka and
O. Yamada, Anti-Cancer Agents Med. Chem., 2009, 9, 550–559.
Med. Chem. Commun.
This journal is ª The Royal Society of Chemistry 2012