Discovery of triazine-benzimidazoles as selective inhibitors of mTOR

Article abstract of DOI:10.1016/j.bmcl.2011.02.007

mTOR is part of the PI3K/AKT pathway and is a central regulator of cell growth and survival. Since many cancers display mutations linked to the mTOR signaling pathway, mTOR has emerged as an important target for oncology therapy. Herein, we report the discovery of triazine benzimidazole inhibitors that inhibit mTOR kinase activity with up to 200-fold selectivity over the structurally homologous kinase PI3Kα. When tested in a panel of cancer cell lines displaying various mutations, a selective inhibitor from this series inhibited cellular proliferation with a mean IC₅₀ of 0.41 μM. Lead compound 42 demonstrated up to 83% inhibition of mTOR substrate phosphorylation in a murine pharmacodynamic model.

Full text of DOI:10.1016/j.bmcl.2011.02.007

Bioorganic & Medicinal Chemistry Letters 21 (2011) 2064–2070
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of triazine-benzimidazoles as selective inhibitors of mTOR
Emily A. Peterson a, , Paul S. Andrews b, Xuhai Be c, Alessandro A. Boezio a, Tammy L. Bush d, Alan C. Cheng e,
⇑
James R. Coats a, Adria E. Colletti c, Katrina W. Copeland a, Michelle DuPont d, Russell Graceffa a,
Barbara Grubinska d, Jean-Christophe Harmange a, , Joseph L. Kim f, Erin L. Mullady b, Philip Olivieri a,
Laurie B. Schenkel a, Mary K. Stanton g, Yohannes Teffera e, Douglas A. Whittington f, Ti Cai d, Daniel S. La a,
⇑
a Medicinal Chemistry, Amgen Inc., 360 Binney St., Cambridge, MA 02142, USA
b Lead Discovery, Amgen Inc., 360 Binney St., Cambridge, MA 02142, USA
c Pharmacokinetics and Drug Metabolism, Amgen Inc., 360 Binney St. Cambridge, MA 02142, USA
d Oncology Research, Amgen Inc., 360 Binney St., Cambridge, MA 02142, USA
e Molecular Structure, Amgen Inc., 1120 Veterans Blvd., South San Francisco, CA 94080, USA
f Molecular Structure, Amgen Inc., 360 Binney St. Cambridge, MA 02142, USA
g Pharmaceutics R&D, Amgen Inc., 360 Binney St., Cambridge, MA 02142, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
mTOR is part of the PI3K/AKT pathway and is a central regulator of cell growth and survival. Since many
cancers display mutations linked to the mTOR signaling pathway, mTOR has emerged as an important
target for oncology therapy. Herein, we report the discovery of triazine benzimidazole inhibitors that
inhibit mTOR kinase activity with up to 200-fold selectivity over the structurally homologous kinase
Received 7 December 2010
Revised 28 January 2011
Accepted 2 February 2011
Available online 12 February 2011
PI3K
a
. When tested in a panel of cancer cell lines displaying various mutations, a selective inhibitor from
M. Lead compound 42 demonstrated
this series inhibited cellular proliferation with a mean IC50 of 0.41
l
Keywords:
mTOR
Kinase inhibitor
Oncology
up to 83% inhibition of mTOR substrate phosphorylation in a murine pharmacodynamic model.
Ó 2011 Elsevier Ltd. All rights reserved.
Benzimidazole
Triazine
The mammalian target of rapamycin (mTOR) kinase has been
identified as a critical signaling node in the promotion of cancer
cell growth and survival. mTOR senses the cellular availability of
nutrients and energy, serves as a downstream regulator of growth
factor signaling, and is ultimately responsible for mediating cell
functions such as apoptosis, autophagy, translation, cell cycle reg-
ulation, and cytoskeleton reorganization.1 mTOR functions through
two distinct multi-protein complexes, mTOR complex 1 (mTORC1)
and mTOR complex 2 (mTORC2). Each complex is defined by its
interaction with either Raptor (regulatory-associated protein of
mTOR) as in the case of mTORC1 or Rictor (rapamycin-insensitive
companion of mTOR) as in the case with mTORC2.
nases (RTKs) leading to association with the adaptor protein IRS1
(insulin receptor substrate 1). IRS1 recruits PI3K to the cell mem-
a
brane, which converts membrane bound phosphatidylinositol-4,
5-bisphosphate (PIP2) to phosphatidylinositol-3,4,5-trisphosphate
(PIP3). PIP3 causes translocation of AKT as well as PDK1 to the cell
membrane through association with the conserved pleckstrin
homology domain on both proteins. This event leads to the phos-
phorylation of AKT by PDK1 at the T308 site. Fully activated AKT
is generated upon the phosphorylation at S473 by mTORC2. AKT
phosphorylates several proximal substrates including PRAS40
and TSC2.2 Both of these events lead to activation of mTORC1,
which phosphorylates 4EBP1 and S6K, integral participants in pro-
tein translation and cell growth. mTOR activation also occurs in re-
sponse to elevated nutrient or amino acid levels through RAG
GTPase activation as well as the Ras/Raf/MAPK pathway.3 In-
creased activity through any of these mechanisms can lead to in-
creased signaling through mTOR and ultimately unregulated cell
growth. Recently, rapamycin and its analogs, which are allosteric
inhibitors of mTORC1 exclusively, have received much attention
as cancer therapeutics. Efforts in this arena have culminated in
the approval of AfinitorÒ (Everolimus)4 and ToriselÒ (Temsiroli-
mus)5 for renal cell carcinoma. As a result of mTORC1 inhibition,
Both complexes play critical roles in the PI3K/AKT cascade, a
pathway wherein mutations within its signaling components are
frequently present in cancer cells. Mechanistically, PI3K/AKT path-
way initiation occurs through activation of receptor tyrosine ki-
⇑
Corresponding authors. Tel.: +1 617 444 5027, fax: +1 617 621 3907 (E.A.P.);
tel.: +1 617 444 5189, fax: +1 617 621 3908 (D.S.L).
E-mail addresses: epeterso@amgen.com (E.A. Peterson), daniel.la@amgen.com
(D.S. La).
Currently at Constellation Pharmaceuticals, 215 First St., Suite 200, Cambridge
MA 02142, USA
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.02.007

E. A. Peterson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2064–2070
2065
rapamycin disrupts the regulating negative feedback loop, which
involves the destabilization of IRS1 protein by S6K phosphoryla-
tion. The suppression of this feedback mechanism, leading to in-
creased signaling through the PI3K/AKT pathway, could present a
liability as increased pAKT levels have been observed in patients
treated with rapamycin.6 Therefore, discovery of an ATP competi-
tive inhibitor, which would target both mTOR complexes through
inhibition of mTOR kinase activity, is an attractive approach to pre-
venting the increased AKT signaling produced by exclusive inhibi-
tion of mTORC1.7 Furthermore, it is expected that an mTOR
selective inhibitor would demonstrate increased tolerability over
a dual PI3K/mTOR inhibitor due to the effects of PI3K inhibition
on glucose homeostasis and insulin resistance. In the area of
ATP-competitive mTOR selective inhibitors, several compounds
have entered the clinic, including AZD80558 and OSI-027.9 Accord-
ingly, as part of our oncology portfolio, we endeavored to develop a
selective small molecule inhibitor of mTOR.
Our effort to develop a new class of selective mTOR kinase
inhibitors began with a high-throughput screen of our compound
collection. A promising hit that emerged from the screening cam-
paign was triazine benzimidazole 1 (Fig. 1A), which inhibited phos-
phorylation of mTOR substrate 4EBP1 in a Lantha-Screen enzyme
assay with an IC50 = 97 nM. It was also moderately potent in two
U-87 cellular assays that measure the phosphorylation of 4EBP1
(at T37/46, downstream of mTORC1) and AKT (at S473, which is
downstream of mTORC2). As mentioned above, a goal for our pro-
gram was to develop an mTOR selective inhibitor, which is chal-
lenging due to the significant homology between the active site
the affinity pocket.11 These interactions are likely similar to those
within the mTOR ATP binding domain, since the residues making
these interactions are conserved between mTOR and PI3K
model of the mTOR kinase domain based on sequence homology
with PI3K was developed to potentially gain a better understand-
c. A
c
ing of the mTOR ATP binding site. The residues labeled in blue are
those that are predicted to be different in the mTOR active site,
which we hoped to exploit in the development of a selective mTOR
inhibitor (Fig. 1B).
Further insight was obtained by inspection of a surface repre-
sentation of PI3Kc bound to 1, which revealed a small pocket adja-
cent to the triazine (Fig. 1C, highlighted in yellow). The presence of
this extra space and the adjacent tyrosine residue differentiates the
mTOR and PIKK kinases from the majority of other protein kinases,
which generally contain a carbonyl group that fills this pocket and
often interacts with kinase inhibitors through a hydrogen bonding
interaction.12 We therefore investigated substitution of the R1 po-
sition of the triazine linker-binder as a means to provide selectivity
over kinases outside the PIKK family (Table 1). Lead compound 1
was modified to analog 10 (R1 = methyl) and evaluated in a panel
of 48 kinases revealing that this small modification removed off-
target kinase activity and improved the potency on mTOR (Table
1A). Selectivity over PI3Ka, however, remained a challenge. Expan-
sion of the R1 position to groups larger than methyl (12, 13, and 16)
led to a significant decrease in potency, which is consistent with
the limited volume of the pocket adjacent to Tyr867.
A variety of triazine analogs were synthesized and screened
with the aim of improving potency and selectivity (Table 1B). At-
tempts to replace the methyl acetonitrile group with aromatic,
benzylic or small alkyl groups lead to decreased potency (2–5). Re-
moval or replacement of the R2 amine also led to inactive com-
pounds (7 and 8). The only modification that improved potency
was removal of the methylacetonitrile group to give triazine amine
6. Changing the triazine to a pyrimidine (11, X = N ? 14, X = CH)
was also investigated but provided no additional improvement in
and kinase domains of mTOR and the class I PI3Ks (a, b, d, and
c
). Inhibitor 1 demonstrated modest (3Â) selectivity over PI3K
a,
and when profiled in a binding assay (Ambit) against a panel of
402 kinases, was also found to bind kinases in the Raf pathway
as well as KDR, TGFBR2 and DCAMKL3 at [1 lM]. Herein we de-
scribe the design, structure–activity relationships (SAR), and
in vivo profile of selective mTOR inhibitors arising from the initial
discovery of benzimidazole triazine 1.
potency or selectivity over PI3Ka. Consideration of the SAR devel-
Due in part to the large size of the mTOR protein (289 kDa),
solving a crystal structure of the mTOR kinase domain has been a
challenge to the field. However, due to the high homology between
oped in Table 1B revealed that methyl aminotriazine 11 was the
optimal linker-binder element. Inhibitor 11 was evaluated against
a custom panel of 100 kinases to understand its broader selectivity
profile. This effort revealed that 11 also demonstrated selectivity
for mTOR and the PIKK kinases with the exception of FLT3
mTOR and PI3K
of the kinase domain of PI3K
inhibitor 1 binds to the mTOR active site. A co-crystal structure
of 1 in the kinase domain of PI3K was obtained and showed that
there are several important contacts that compound 1 makes with-
in the ATP binding site of PI3K
(Fig. 1B).10 Of particular impor-
tance is dual hydrogen bonding interaction between the
c
within the ATP binding site, the crystal structure
c
could provide insight into how
(POC = 14) and CSNK2A1 (POC = 15) at [1 lM].
c
The representative route used to synthesize the triazine-benz-
imidazole compounds is illustrated in Scheme 1, as exemplified
by the synthesis of 11. Methylation of cyanuric chloride (17) with
methyl magnesium bromide provided methyl dichlorotriazine 18,
which underwent subsequent SNAr displacement with 2-chloro-
benzimidazole (19). Amination of the remaining chloride on the
triazine installed the linker-binding moiety and provided the
c
a
triazine amine and Val882 (also known as the linker or hinge res-
idue) and the displacement by the phenol of a structural water
molecule that bridges Asp841 and Tyr867 in the area known as
Figure 1. (A) HTS hit 1; (B) co-crystal structure of 1 with PI3Kc; (C) surface representation of PI3Kc bound to 1.

2066
E. A. Peterson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2064–2070
Table 1
contains the catalytic lysine (Fig. 1B). We therefore anticipated that
a polar substituent would be favored, as demonstrated by the
diminished mTOR potency for phenyl compound 24 (Table 2).
The most potent inhibitor of mTOR in this series was bis-phenol
25, which also had sub-nanomolar potency on three of the PI3K
isoforms. In seeking to replace the phenol group, which could
potentially pose a pharmacokinetic liability, several phenol bio-
isosteres were tested with varying results. While the more tradi-
tional carboxamide phenol bioisostere13 26 proved to be less
potent than the parent phenol, N-linked amides and ureas in the
para position were potent although non-selective (27 and 28). In
general, it was found that arenes attached through a six-membered
ring demonstrated excellent potency, but were not selective
Modification of the linker-binder moiety provides broader kinase selectivity and
improved potency
Compd
C-RAF
KDR
CK1d
FGFR1
A. Broader kinase selectivity (POC)a
1
10
3
92
40
100
40
90
43
100
against PI3Ka.
Further investigation revealed that substituting five-membered
ring heterocycles for the affinity pocket moiety provided potent
inhibitors of mTOR that were moderately selective for mTOR over
Compd R1
R2
X
mTORb p4EBP1c pAKTd P13Kae
(IC50 M)
B. Potency on mTOR and selectivity over P13K
a
, l
1
H
NHCH2CN
NHPh
NHMe
NH2CH2CF3
NHCH2-2-pyr
NH2
H
OMe
NH2
NHCH2CN
NH2
N
N
N
N
N
N
N
N
0.097
2.3
0.25
>50
21
1.7
>25
2.9
>25
>25
0.70
0.87
0.18
3.2
0.32
0.72
0.19
3.6
PI3K
potency and selectivity was pyrazole 32, which had an IC50 of
27 nM for the inhibition of mTOR, was 30-fold selective over PI3K
a. The compound that demonstrated the best combination of
2
H
3
H
a,
4
H
5
H
2.4
3.6
and demonstrated up to 32Â selectivity over the other PI3K iso-
forms. Substitution on the carbons of the pyrazole ring led to a sig-
nificant decrease in mTOR potency, with the exception of indazole
36, which demonstrated diminished potency in the p4EBP1 cellular
assay. Alkylation of the pyrazole nitrogen was also detrimental (33,
34, and 38) as was substitution with other related heterocycles
such as isoxazole 35 and isothiazole (not shown). In addition,
migration of the pyrazole nitrogens to give 4-pyrazolyl isomer 37
or five-membered heterocycles bearing three heteroatoms such
as triazole or oxadiazole reduced the potency, as exemplified by
40.
6
H
0.021
3.0
>50
0.71
>25
>25
3.5
0.38
0.15
6.3
>25
1.2
8.9
0.034
1.6
1.9
0.13
2.0
1.7
7
H
8
H
9
H
CH 0.11
0.42
0.017
0.009
0.41
2.64
0.043
0.60
2.1
1.0
10
11
12
13
14
15
16
Me
Me
Et
n-Pr
Me
Me
N
N
N
N
0.008
0.007
0.39
4.7
0.032
0.004
0.34
3.5
0.042
0.37
0.52
NH2
NH2
NH2
H
CH 0.087
CH 0.37
N
OMe NHCH2CN
>50
21.5
a
b
c
Kinase binding assay, POC = % of control at [1
Inhibition of kinase activity (Lantha-Screen).
Cell assay measuring p4EBP1 in U87 cells.
Cell assay measuring pAKT (S473) in U87 cells.
Inhibition of kinase activity (Alpha-screen).
lM].
More drastic changes to the affinity pocket region of the mole-
cule were not tolerated. Replacing the NH linkage between the
benzimidazole ring and the affinity pocket moiety with an oxygen
d
e
or sulfur resulted in IC50’s >25 lM. An intramolecular hydrogen
bond exists between the NH at the 2-benzimidazole position and
the proximal nitrogen on the triazine ring. This interaction is
essential for maintaining co-planarity of the triazine and benz-
imidazole rings, which appears to be a requirement for potency
on mTOR. It was also found that only planar, aromatic groups in
the affinity pocket resulted in potent inhibition of mTOR, as attach-
ment of alkyl rings, chains, benzyl groups, amides or ureas gave
significantly less potent compounds, as exemplified by 41.
benzimidazole chloride intermediate 21 (R = H), from which diver-
gent SAR exploration of the affinity pocket could be easily achieved
by SNAr displacement. For compound 11, 3-aminophenol reacted
exclusively at the aniline amine upon microwave heating at
140 °C in 2-butanol.
Using the synthesis described for 11 in Scheme 1, which incor-
porates the affinity pocket moiety as the last step, a wide variety of
phenol replacements could be investigated in a rapid fashion with
the aim of improving potency against mTOR and selectivity over
Analysis of the co-crystal structure of pyrazole 32 in the PI3K
c
active site revealed that the pyrazole resided within hydrogen
bonding distance of the catalytic Asp836 and Lys833 residues in
PI3Kc (Fig. 2). A key difference in the mTOR active site based on
the PI3K
a, b, d, and c isoforms (Table 2). The affinity pocket region
within the ATP binding site is a carboxylate-rich region that also
Scheme 1. Synthesis of inhibitor 11. Reagents and conditions: (a) MeMgBr, Et2O/THF, À78 °C to rt (85%); (b) iPr2EtN, 1,4-dioxane (55%); (c) 7 N NH3 in MeOH, rt (90%);
(d) 2-butanol, microwave, 140 °C (70%).

E. A. Peterson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2064–2070
2067
Table 2
Modification of the affinity pocket
Compd
R
IC50
pAKTc
(lm)
mTORa
0.007
4EBP1b
0.15
P13Kad
P13Kb/d/c
23
24
0.009
0.19
0.004
0.009/0.003/0.008
1.0/0.36/0.47
0.42
>25
0.26
25
26
0.0002
0.097
0.005
0.19
0.0006
0.0002/0.0004/0.005
1.03/0.040/0.48
0.175
0.002
3.9
0.29
27
0.34
0.011
0.006
0.017/0.011/0.013
28
29
0.013
0.29
0.34
1.5
0.18
0.12
0.023
0.63
0.41/0.071/0.02
1.4/0.89/0.11
30
31
0.028
0.027
0.96
1.5
0.14
0.12
0.23
0.20
1.0/0.24/0.34
0.63/0.32/0.20
32
33
34
35
0.027
0.27
0.34
0.15
0.39
5.3
0.055
0.58
0.77
0.26
0.80
6.2
0.87/0.51/0.32
28.8/4.6/4.1
22.3/0.53/0.86
3.1/1.3/2.3
6.5
10.4
0.84
3.3
36
0.06
4.2
0.25
0.96
0.42/0.27/0.19
37
38
0.19
>50
5.3
0.18
0.54
0.46
1.9
1.0/0.69/0.61
6.9/3.3/2.3
>25
39
40
41
0.15
33.1
21.4
>25
>25
>25
0.68
4.6
0.52
NDe
1.6
1.8/0.14/0.026
ND
0.68
2.4/0.71/0.03
a
b
c
Inhibition of kinase activity (Lantha-Screen).
Cell assay measuring p4EBP1 in U87 cells.
Cell assay measuring pAKT (S473) in U87 cells.
Inhibition of kinase activity (Alpha-Screen).
ND = not determined.
d
e

2068
E. A. Peterson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2064–2070
Pharmacokinetic (PK) studies in male Sprague–Dawley rats re-
vealed that 32 demonstrated relatively high clearance (CL = 1.9 L/
h/kg), short half-life (T1/2 = 0.72 h), poor exposure (IV
AUC = 1680 nM h) and poor bioavailability (F = 6.3%). The high
clearance observed in vivo was inconsistent with the in vitro
intrinsic clearance rates, which were relatively low (79 and
52
Similarly, the in vitro predicted clearance obtained from rat hepa-
tocytes was comparatively low (28 L/(min mg)). To further
lL/(min mg) in human and rat liver microsomes, respectively).
l
understand the high in vivo clearance, we conducted in vivo
metabolism studies by dosing bile duct-cannulated rats with 32.
Using LC/MS analysis, multiple glucuronidation products of both
the parent (32) and of oxidative metabolites were identified in
rat bile (Scheme 2).
To mitigate the oxidation of the benzimidazole ring, bis-fluori-
nated analog 46 was synthesized and evaluated in vivo in rats. This
modification had minimal impact on the potency or selectivity for
mTOR, but resulted in an inferior PK profile (Table 3). As an alter-
native strategy, potential sites of glucuronidation were targeted as
a means of reducing the in vivo clearance. The pyrazole in the affin-
ity pocket was addressed first using the isoxazole congener 35.
Although 35 was a much less potent inhibitor of mTOR and exhib-
Figure 2. Co-crystal structure of 32 in the kinase active site of PI3Kc.
our homology model is the presence of Glu724 in place of the
Asp836 present in PI3K . We reasoned that the additional side
c
ited no selectivity over PI3Ka, the in vivo PK data was obtained to
chain length of the glutamate, when compared to Asp, could lead
to a tighter binding interaction with the pyrazole and also move
the lysine closer to the benzimidazole and pyrazole nitrogens. This
residue difference could lead to a more effective binding interac-
tion for 32 within mTOR and in turn, impact selectivity over PI3K.
determine if the pyrazole was the primary site of glucuronidation.
No significant improvement in the in vivo clearance, exposure or
half life was observed as a result of this substitution.14 Finally,
we examined alkylation of the triazine amine as a means to inhibit
Table 3
Efforts to improve the PK properties by substitution of the triazine linker-binder
Compd
R
IC50
pAKT
(l
m)
Rat PK (1 mg/kg IV,a 2 mg/kg POb)
mTOR
4EBP1
P13K
a
P13Kb/d/
c
CL (L/h/kg)
Vss(L/kg)
AUC (nM h, IV)
T1/2 (h, IV)
F(%)
PPB(%)rat
32
42
43
44
45
35
46
H
Me
CH2CN
(CH2)2CN
CH2CCH
––
0.027
0.081
0.031
0.060
0.080
0.15
0.40
0.60
0.60
1.3
1.4
0.33
0.34
0.055
0.10
0.073
0.082
0.15
0.80
2.2
6.2
7.7
4.0
0.87/0.51/0.32
4.5/0.89/3.2
8.5/4.0/3.6
15.4/5.8/6.2
3.7/4.6/3.7
1.9
1.1
1680
3600
559c
1020d
ND
0.72
0.56
0.45c
0.70d
ND
6.3
20.7
0
ND
ND
ND
6.0
97.2
99.2
98.4
ND
ND
97.7
99.0
0.93
2.5c
ND
ND
1.8
0.76
1.5c
ND
ND
1.4
0.26
0.045
0.84
0.49
3.1/1.3/2.3
1.0/0.11/0.25
1930
964
0.50
0.70
––
0.024
3.0
3.0
ND = not determined.
a
Formulation: DMSO.
Formulation: 1% Tween 80, 2% HPMC in water adjusted to pH = 2.2 wth MSA.
IV 0.5 mg/kg (DMSO).
Rat PO, 5 mg/kg.
b
c
d
Scheme 2. Metabolites produced from treatment of rats with 32.

E. A. Peterson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2064–2070
2069
Table 4
binding to the 402 kinases tested in an Ambit binding assay at
[1 M], with the exception of TGFBR2 (Table 4, POC of 17). Inde-
pendent testing of 43 against TGFb signaling in a pSMAD cellular
assay revealed no significant inhibition of TGFb up to a 10 M con-
Inhibition of proliferation in cancer cell lines by 43
l
l
centration of 43. The clean kinase profile exhibited by 43 allowed
its use as a tool to understand the applicability of an mTOR inhib-
itor toward various cancer models.
Compound 43 was used to investigate the effect of a selective
mTOR inhibitor on a panel of tumor cell lines as a model to predict
whether selective inhibition of mTOR kinase activity would have
broad efficacy over a variety of tumors. Focusing on melanoma,
lung and breast cancer cell lines, compound 43 was tested in a pa-
nel of 18 different cancer cell lines of varying mutation status. Sub-
micromolar proliferation inhibition was observed upon treatment
with 43 in all of the cell lines tested with a mean IC50 = 0.41 lM.
The greatest inhibition was seen in BT474 and UACC-893 breast
cancer cells that harbor HER2, PI3K and p53 mutations (Table 4).
The efficacy across a variety of cell lines further supports inhibition
of mTOR kinase activity as a worthwhile endeavor toward cancer
treatment for a broad patient population.
a
Cell line
Gene mutation/amplification
IC50 (lm)
Although compound 43 demonstrated the best potency/selec-
tivity combination, the high rate of clearance and poor exposure
observed for this compound precluded further testing in vivo.
The analog that provided the most improvement in the PK proper-
ties was the triazine methylamine analog 42. Compound 42
showed slightly diminished potency against mTOR, but maintained
the selectivity over the PI3K isoforms when compared with lead
compound 32. Importantly, the installation of the methyl group re-
duced the in vivo clearance from 1.9 L/h/kg (32) to 0.93 L/h/kg (42),
increased the oral bioavailability by 3Â to give F = 21% and im-
A375
A549
BT474
Calu-6
BRAF
0.38
0.43
0.20
0.59
0.32
0.34
0.53
0.22
0.54
0.27
0.69
0.42
0.60
0.46
0.43
0.47
0.35
0.14
KRAS, LKB1
HER2, P13K, p53
KRAS
P13K, PTEN
BRAF, PTEN
BRAF, APC, RB
P13K, p53
CAL51
COLO829
DU4475
HCC-1954
MCF7
MDA-MB-231
MDA-MB-468
NCI-H2009
NCI-H1975
NCI-H460
NCI-H82
SK-MEL-28
U87MG
P13K
KRAS, BRAF, p53
PTEN, EGFR, p53
KRAS, p53,RB
P13K, EGFR, p53
KRAS, P13K, LKBI
CMYC, RB
BRAF
PTEN
HER2, P13K, p53
proved the solubility (solubility in 0.01 N HCl = 156
37 g/mL for 32).
Since compound 42 exhibited the most promising combination
lg/mL vs
l
of potency, selectivity and desirable pharmacokinetic properties,
we sought to determine whether it would inhibit mTOR activity
in vivo. Before assessing the in vivo pharmacological activity of
42, it was evaluated against a panel of 100 kinases to determine
its selectivity profile. Inhibitor 42 demonstrated excellent selectiv-
ity for mTOR and did not significantly inhibit any of the other
kinases in the panel, showing only moderate inhibition of YSK4
UACC-893
a
Inhibition of proliferation was measured by ATP-lite assay after cells treated
with compound for 72 h in 10% serum media; also pictured: Ambit Kinome Scan for
43 at [1 M]: the red circle on the kinome tree represents TGFBR2 binding.
Although tested in the Ambit assay, mTOR and PI3KK are not pictured on the
kinome tree.
l
(POC = 35) and CSNK2A1 (POC = 46) at [1 l
M].15 In order to
demonstrate that 42 would inhibit mTOR kinase activity in vivo,
a pharmacodynamic assay was conducted to measure the change
in HGF-induced phosphorylation of downstream mTOR substrates
pAKT and pS6RP (Fig. 3). Inhibitor 42 was administered to mice by
intraperitoneal (IP) injection at 10, 30 and 100 mg/kg (N = 3 ani-
mals/group).16 Six hours post-dose, human recombinant HGF was
injected intravenously and the livers were harvested. The phos-
phorylation of AKT (S473, downstream of mTORC2) and S6K (phos-
phorylation site S240/244, downstream of mTORC1) was measured
N-glucuronidation through modification of the electronics or steric
hindrance of the triazine amine. Consideration of the SAR devel-
oped in Table 1 allowed us to quickly determine which groups
would be compatible with maintaining mTOR potency. With re-
gard to selectivity, compound 43 was superior to compounds pre-
viously synthesized, showing 200-fold selectivity over PI3Ka, and
greater than 100-fold selectivity over the other PI3K isoforms.
Additionally, 43 was exquisitely selective for mTOR, showing no
Figure 3. Pharmacodynamic assay measuring the changes in of HGF-induced phosphorylation of pAKT and pS6RP in the mouse liver as a result of treatment with compound
42.

2070
E. A. Peterson et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2064–2070
using MSD analysis. Treatment with 42 at 100 mg/kg resulted in
significant inhibition of both AKT (83% inhibition, p <0.0001) and
S6RP (78%, p = 0.0005) phosphorylation in liver tissue at the six
hour time point. The mean unbound plasma concentration at the
References and notes
1. Ciuffreda, L.; Di Danza, C.; Incani, U. C.; Milella, M. Curr. Cancer Drug Targets
2010, 10, 484.
2. Li, Y.; Corradetti, M. N.; Inoki, K.; Guan, K.-L. Trends Biochem. Sci. 2004, 29, 32.
3. Long, X.; Ortiz-Vegas, S.; Lin, Y.; Avruch, J. J. Biol. Chem. 2005, 280, 23433.
4. Motzer, R. J.; Escudier, B.; Oudard, S.; Hutson, T. E.; Porta, C.; Bracarda, S.;
Grünwald, V.; Thompson, J. A.; Figlin, R. A.; Hollaender, N.; Kay, A.; Ravaud, A.
Cancer 2010, 116, 4256.
5. (a) Rini, B. I. Clinical Cancer Res. 2008, 14, 1286; (b) Rini, B.; Kar, S.; Kirkpatrick,
P. Nat. Rev. Drug Disc. 2007, 6, 599.
6. García-Echeverría, C. Bioorg. Med. Chem. Lett. 2010, 20, 4308.
7. Sabatini, D. M. Nat. Rev. Cancer 2006, 6, 729; Faivre, S.; Kroemer, G.; Raymond,
E. Nat. Rev. Drug Disc. 2006, 5, 671.
100 mg/kg dose was 0.65
the AKT cellular IC50 and 1.6-fold higher than the pS6 cellular
IC50 of 0.39 M. Treatment with 30 mg/kg resulted in a mean un-
bound plasma concentration of 0.30 M providing significant inhi-
lM, which was greater than 6-fold over
l
l
bition of AKT phosphorylation only (71%, p <0.0001).
In conclusion, by using a combination of targeted SAR with
information derived from the X-ray co-crystal structures with
8. Chresta, C. M.; Davies, B. R.; Hickson, I.; Harding, T.; Cosulich, S.; Critchlow, S.
E.; Vincent, J. P.; Ellson, R.; Jones, D.; Sini, P.; James, D.; Howard, Z.; Dudley, P.;
Hughes, G.; Smith, L.; Maguire, S.; Hummersone, M.; Malagu, K.; Menear, K.;
Jenkins, R.; Jacobsen, M.; Smith, G. C. M.; Guichard, S.; Pass, M. Cancer Res. 2010,
70, 288.
PI3Kc, novel triazine-benzimidazole mTOR inhibitors were discov-
ered that demonstrated moderate selectivity over the PIKK family
of kinases. Although this class of inhibitors suffered from relatively
high in vivo clearance and low solubility, compound 42, which
demonstrated the best combination of mTOR potency, selectivity
over other kinases and pharmacokinetic properties was tested
in vivo in our mouse pharmacodynamic assay. Inhibitor 42 showed
dose-dependent inhibition of mTOR kinase activity, showing up to
83% inhibition of downstream substrate pAKT. Additionally, 43 was
used as a tool to understand the efficacy profile against 18 cell lines
of varying mutation status. Improvements upon this series with re-
gard to potency, selectivity and desirable pharmacokinetic proper-
ties will be reported in due course.
9. Barr, S.; Russo, S.; Buck, E.; Epstein, D.; Miglarese, M. AACR 101st Annual
Meeting, Washington, DC, 2010, p 1632.
10. Atomic coordinates for the co-crystal structures of
1 and 32 have been
deposited in the RCSB with PDB codes 3QAQ and 3QAR, respectively.
11. Knight, Z. A.; Gonzalez, B.; Feldman, M. E.; Zunder, E. R.; Goldenberg, D. D.;
Williams, O.; Loewith, R.; Stokoe, D.; Balla, A.; Toth, B.; Balla, T.; Weiss, W. A.;
Williams, R. L.; Shokat, K. M. Cell 2006, 125, 733.
12. (a) Liu, Y.; Gray, N. S. Nat. Chem. Biol. 2006, 2, 358; (b) Zhang, J.; Yang, P. L.;
Gray, N. S. Nat. Rev. Cancer 2009, 9, 28.
13. Carboxamides as Phenol Bioisostere papers: (a) Breslin, H. J.; Cai, C.;
Miskowski, T. A.; Coutinho, S. V.; Zhang, S.-P.; Hornby, P.; He, W. Bioorg. Med.
Chem. Lett. 2006, 16, 2505; (b) Wentland, M. P.; Lou, R.; Ye, Y.; Cohen, D. J.;
Richardson, G. P.; Bidlack, J. M. Bioorg. Med. Chem. Lett. 2001, 11, 623.
14. The in vitro intrinsic clearance was also low for 35 (39 and 46 lL/(min mg), in
Acknowledgments
HLM and RLM respectively).
15. The custom Ambit 100 kinase binding assay contains a diverse array of kinases
from different families and is intended be representative of the wider range of
kinases.
16. Although the bioavailibility for 42 was improved over 32 and 43, the oral
exposure for 42 did not increase proportional to dose (2 mg/kg PO,
AUC = 1490 nM h versus 50 mg/kg PO, AUC = 2030 nM h). From this result it
was determined that IP dosing might be the best method to achieve desired
target coverage.
The authors would like to acknowledge Wei Hu, Tisha San
Miguel, and Leeanne Zalameda for enzyme and cell assay support.
We are also grateful to Loren Berry, Liyue Huang, Meghan Langley,
and Jonathan Roberts for PKDM support. Thanks to Jingzhou Liu for
metabolite identification studies, Pete Yakowec and Jin Tang for
PI3Kc expression and purification as well as Huilin Zhao and Linda
Epstein for mTOR expression and purification.