Imidazo[1,5-a]pyrazines: Orally efficacious inhibitors of mTORC1 and mTORC2

Article abstract of DOI:10.1016/j.bmcl.2011.01.139

The discovery and optimization of a series of imidazo[1,5-a]pyrazine inhibitors of mTOR is described. HTS hits were optimized for potency, selectivity and metabolic stability to provide the orally bioavailable proof of concept compound 4c that demonstrate

Full text of DOI:10.1016/j.bmcl.2011.01.139

Bioorganic & Medicinal Chemistry Letters 21 (2011) 2092–2097
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Imidazo[1,5-a]pyrazines: Orally efficacious inhibitors of mTORC1 and mTORC2
Andrew P. Crew a, , Shripad V. Bhagwat a, Hanqing Dong a, Mark A. Bittner b, Anna Chan a, Xin Chen a,
⇑
Heather Coate a, Andrew Cooke b, Prafulla C. Gokhale b, Ayako Honda a, Meizhong Jin a, Jennifer Kahler a,
Christine Mantis b, Mark J. Mulvihill a, Paula A. Tavares-Greco a, Brian Volk a, Jing Wang a,
Douglas S. Werner a, Lee D. Arnold a, Jonathan A. Pachter a, Robert Wild b, Neil W. Gibson a
a (OSI) Oncology, OSI Pharmaceuticals, Inc., 1 Bioscience Park Drive, Farmingdale, NY 11735, USA
b (OSI) Oncology, OSI Pharmaceuticals, Inc., 2860 Wilderness Place, Boulder, CO 80301, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery and optimization of a series of imidazo[1,5-a]pyrazine inhibitors of mTOR is described. HTS
hits were optimized for potency, selectivity and metabolic stability to provide the orally bioavailable
proof of concept compound 4c that demonstrated target inhibition in vivo and concomitant inhibition
of tumor growth in an MDA-MB-231 xenograft model.
Received 6 January 2011
Revised 27 January 2011
Accepted 31 January 2011
Available online 3 February 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
mTOR
mTORC1
mTORC2
Imidazo[1,5-a]pyrazine
Oral inhibitor
The mammalian target of rapamycin (mTOR) is a 289 kDa
serine/threonine protein kinase and a member of the PI3K-related
kinase (PIKK) family that plays an important role in the PI3K/
mTOR/AKT pathway to integrate both extracellular and intracellu-
lar signals and serve as a central regulator of cell metabolism,
growth, proliferation and survival.1 In mammalian cells, mTOR re-
sides in at least two physically and functionally distinct signaling
complexes, mTORC1 and mTORC2. mTORC1 directly phosphory-
lates S6 kinase 1 (S6K1 T389) and the eIF4E binding protein 1
(4E-BP1 S65/T70) through which growth and protein synthesis
are respectively regulated.2 The prototypical mTOR inhibitors
based around the macrolide rapamycin (‘rapalogs’) have been
shown to inhibit phosphorylation of S6K1 and 4E-BP1 through
inhibition of mTORC1 and hence, mTORC1 is considered as a rapa-
mycin-sensitive complex.3 mTORC2 also phosphorylates 4E-BP1
(T37/46), and is additionally involved in phosphorylation of Akt
on S473, the activation of which drives cancer-related cellular
responses such as cell growth and proliferation.4 Rapamycin and
the rapalogs do not directly inhibit mTORC2.
consequences compared to those agents inhibiting mTORC1
only.5,6 The list of agents now identified as dual mTORC1/2 inhib-
itors includes PI-103,7a BEZ-235,7b AZD8055,7c Torin1,7d PP242,7e
WYE-1251327f, INK-1287g and our own dual inhibitor OSI-027, an
imidazotriazine-based small molecule which is currently in
phase-I clinical trials.8 Herein we describe our early medicinal
and computational chemistry efforts to optimize the high through-
put screening (HTS) hits, imidazo[1,5-a]pyrazin-8-amine 1a and
1b, focusing on the substitutions at C-1 and C-3 positions to drive
potency and DMPK properties. Proof of concept compound 4c was
discovered as a potent and orally bioavailable inhibitor with in vivo
efficacy in a mouse xenograft model, and became the starting point
for multivariant optimization and the discovery of OSI-027.
Our efforts towards identifying dual mTORC1/2 inhibitors
began with a high throughput screening campaign of the OSI com-
pound library using human mTOR isolated from HeLa cell lysates in
a chemiluminescence-based ELISA assay,9 monitoring phosphory-
lation of full-length 4E-BP1 at T37/46, the so-called rapamycin
insensitive sites.6c Compounds were screened at a concentration
In the past several years, significant effort has been devoted to
the discovery of small molecule ATP-competitive mTOR inhibitors,
which by virtue of the identity of the respective kinase domains,
can interact with both mTOR complexes, with enhanced
of 10
l
M with 2-fold compression, and in order to identify robust
starting points, a final ATP concentration of 100
lM was used.
Among the multiple chemotypes identified, imidazo[1,5-a]-
pyrazine hits 1a and 1b (Fig. 1) were selected for follow-up based
on our previous and positive experiences with this bicyclic
skeleton that culminated in the identification of the clinical dual
IGF-1R/IR agent OSI-906.10 Cognizant of the IGF-1R activity
⇑
Corresponding author. Tel.: +1 631 962 0600; fax: +1 631 845 5671.
E-mail address: acrew@osip.com (A.P. Crew).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.01.139

A. P. Crew et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2092–2097
2093
M)
Table 2
NH2
Ar
SAR of different bicyclic aryl derivatives
1
1a: Ar = 3-OH-phenyl
1b: Ar = 7-quinolinyl
N
8
a
a
N
Compd Ar
IC50
(l
M) Compd Ar
IC50 (l
N
3
S
1d
1.36
2.35
2.60
7.15
>10
1k
1.30
>10
>10
0.46
0.40
Figure 1. Representative imidazo[1,5-a]pyrazine HTS hits.
O
associated with this chemotype, we evaluated hits 1a and 1b, along
with the known potent IGF-1R inhibitor 1c,11 in both mTOR and
IGF-1R biochemical assays. As shown in Table 1, the phenol deriv-
ative 1a exhibited activities for both targets. However, the 7-quin-
olinyl derivative 1b demonstrated a preference for mTOR and in
contrast, 2-phenyl-7-quinolinyl analog 1c showed the a reverse
preference suggesting that the hydrophobic pocket that produc-
tively accommodates the additional phenyl group in IGF-1R,11
was either absent or topologically different in mTOR. This SAR
observation indicated that the mTOR and IGF-1R activities of this
series could be controlled through incorporation of appropriate
substitution at the 1-position.
N
1e
1l
H
N
N
N
N
1f
1m
1n
1o
1g
1h
NH
A variety of analogs at the C-1 position were prepared, including
differently substituted phenyls and other bicyclic aryl groups,
using the general synthetic routes described in our previous com-
munication.12 It was found that replacement of 3-hydroxyl group
in compound 1a with –NHAc, –NHSO2Me, –NH2, –CONH2, –CH2OH
or its movement to the 4-position resulted in total loss of mTOR
N
NH
O
H
potency (IC50 >10 lM). In contrast, the SAR of close analogs of
N
quinoline 1b was more productive (Table 2). Removal of the quin-
oline nitrogen (1d) afforded no appreciable change in potency, sig-
nifying that it played no significant part in binding to the protein.
Its reposition to the 3-position (1e) or the addition of a second
nitrogen to form a quinoxaline (1f) similarly had no effect on po-
tency, neither did replacement of the distal ring with 5-membered
nitrogen and sulphur heterocycles (1i, 1j and 1k). The 6-quinolinyl
analog 1g however, was approximately 8-fold less potent than the
parent 1b, and benzofuran 1l was inactive, suggestive of an unfa-
vorable electrostatic interaction between the heteroatom and the
protein. Such effect in the case of the quinoxaline (1f) could be
somewhat ameliorated by the second nitrogen in the bicyclic ring,
possibly through electron-withdrawing effects. Repositioning of
1i
1j
2.48
3.12
1p
1q
0.26
0.13
NH
S
a
Biochemical ELISA assay in the presence of 100
P2 experiments
lM ATP; values are the mean of
the naphthalene and indole attachment points (a-naphthenoid
1h and 3-indole 1m), resulted in a complete loss of activity.
Replacement of the proximal ring with 5-membered heterocycles
Table 1
mTOR and IGF-1R potencies of compounds 1a–c
a
Compd
Ar
IC50
(
lM)
mTOR
1.97
IGF-1R
0.52
OH
N
1a
1b
1c
0.90
>9
>10
N
0.08
a
Biochemical ELISA assay in the presence of 100
P2 experiments.
lM ATP; values are the mean of
Figure 2. Binding of 1k to the mTOR homology model.

2094
A. P. Crew et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2092–2097
Table 3
Table 5
Deviations from co-planarity for the representative 6,6-, 6,5- and
5,6-bicyclic systems attaching to the 1-position
SAR of imidazopyrazine-3-cyclohexyl derivatives
Compd
Ring system
aa (°)
1d
1h
1i
6,6
6,6
6,5
6,5
5,6
5,6
5,6
ꢀ39.7
ꢀ51.8
ꢀ43.6
ꢀ42.6
ꢀ24.2
ꢀ23.6
0.3
NH
NH2
N
1j
N
1n
1o
1p
N
a
Torsional angle between the imidazopyrazine core and a 1-
substituent after geometrical optimization of whole compound
using ab initio method B3LYP and basis set 6-31G⁄⁄ (ref. Schro-
dinger Inc.).
R
Compd
R
IC50
(l
M)
ERb
a
provided modest/good improvements in potency (e.g. 1n, 1o, 1p
and 1q). By analogy with examples 1b and 1d, the additional nitro-
gen of azaindole 1o offered no benefit over indole 1n.
A homology model of the mTOR kinase was built based on the
PI3K-c and PI3K-c crystal structures 1E8X and 2WXF, respectively.
Mouse
Human
3a
3b
3c
3d
3e
3f
OH
NH2
0.22
1.0
0.67
0.28
0.35
0.51
0.79
0.76
ND
0.74
0.34
0.50
0.64
0.74
0.78
ND
NHAc
NHCHO
NHCO-cyclobutyl
NHCO-3-furyl
NHSO2Me
0.21
0.15
0.17
0.09
0.41
The complex modeling with compounds indicated that the
imidazopyrazine derivatives fit the ATP-binding site with the
8-amino group and 7-nitrogen atom forming two hydrogen bonds
with the backbone groups of Gly2238 and Val2240 considered to
be the hinge residues of mTOR (Fig. 2), mimicking the binding
mode of the purine ring in ATP. In addition, the model identified
a number of pharmacophoric elements important for the binding
of this chemotype to mTOR that helped the development and
understanding of the associated SAR. (1) A narrow channel accom-
modates the C-1 aromatic substituent, and forces it to adopt an
approximate coplanar conformation with respect to the imidazo-
pyrazine core. This suggests why the 5,6-bicyclic systems
(1n, 1o, 1p and 1q) are generally more potent than the 6,6-bicyclic
systems or 6,5-bicyclic systems, since the former systems have
larger preferences toward the approximate co-planarity than the
latter. Table 3 shows the results of ab initio geometrical calcula-
tions that support the increased torsional angle hypothesis for
the 6,6 and 6,5 systems versus the 5,6 systems; (2) Asp2195 which
is located at the end of the narrow channel destabilizes the binding
3g
ND = not determined.
a
Biochemical ELISA assay in the presence of 100 lM ATP; values are the mean of
P2 experiments.
b
ERs determined with mouse and human liver microsomes.
of compounds 1g and 1l by presenting repulsive electrostatic inter-
actions to the nitrogen and oxygen atoms in the distal ring of the
bicycle. In the case of 1g, the increased Van der Waals contacts
associated with the larger quinoline moiety partially offset the det-
rimental electronic effects to recover some potency unlike in the
case of the benzofuran in 1l. By comparison, Asp2195 stabilizes
compound 1i by forming a hydrogen bond with the indole NH moi-
ety, and compound 1k via a favorable oxygen–sulfur contact.
In line with previous observations on simple 3-cyclobutylimi-
dazo[1,5-a]pyrazine derivatives,11 the examples described in Table 2
Table 4
Table 6
SAR of imidazopyrazine-3-cyclobutyl derivatives
SAR of imidazopyrazine-3-piperidin-4-yl derivatives
NH
NH2
N
NH
NH2
N
N
N
N
N
3'
R
N
R
a
Compd
R
IC50
(lM)
ERb
ERb
a
Mouse
Human
Compd
R
IC50
(
lM)
1n
2a
2b
H
OH
CH2OH
0.46
0.37
0.32
0.94
0.60
0.79
0.92
0.74
0.70
Mouse
Human
4a
4b
4c
H
>10
0.69
0.12
0.64
0.43
0.67
0.18
0.44
0.68
COMe
CONMe2
2c
4.0
0.20
0.59
0.44
0.58
N
N
N
4d
1.37
ND
ND
N
O
2d
0.35
N Ac
N
a
a
Biochemical ELISA assay in the presence of 100
P2 experiments.
lM ATP; values are the mean of
Biochemical ELISA assay in the presence of 100
P2 experiments.
lM ATP; values are the mean of
b
b
ERs determined with mouse and human liver microsomes.
ERs determined with mouse and human liver microsomes.

A. P. Crew et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2092–2097
2095
hydrogen-bonding opportunity for some compounds, however,
due to the desolvation penalties incurred in the binding processes,
these hydrogen bonds appeared to contribute little gain to the
binding energies (Fig. 3). The model also indicated that the imi-
dazopyrazine 3-substituent was embraced by a large space with
open access to the solvent that could accommodate both cyclobu-
tyl and cyclohexyl systems. As shown in Tables 5 and 6, cyclohexyl
and piperidine groups tethered to distal polar groups were well
tolerated at the imidazopyrazine 3-position. Generally these polar
groups could be used to modulate the overall physico-chemical
properties and metabolic stability of the series without gross
detriment to potency. In particular acetamide 3c and urea 4c of-
fered a promising combination of sub-micromolar potency and
moderate ERs, so these two agents were selected for further
in vitro and in vivo profiling.
The syntheses of compounds 2a–d, 3a–g and 4a–d shown in
Tables 4–6 are described in Schemes 1 and 2. The bicyclic
intermediate 6 was prepared by the amide coupling of 1-(3-chloro-
pyrazin-2-yl)methylamine (5) with 3-methylidenecyclobutane-
carboxylic acid followed by POCl3-mediated cyclization.11 Olefin
6 was converted to its diol derivative with NMO, and then
iodinated at the 1-position followed by cleavage of the diol moiety
with sodium periodate to afford ketone 7. Reduction of 7 was
achieved with sodium borohydride to stereo-specifically provide
the cis-derivative, which was converted to compound 2a through
ammonolysis and a Suzuki coupling. Compounds 2c and 2d were
prepared through the reductive amination of ketone 7 with the
appropriate piperazine derivatives, again stereo-specifically afford-
ing the cis-products. The hydroboration of 6 was less selective
however, producing a 5:1 mixture of cis:trans isomers that was
inseparable by silica gel chromatography, as was the iodinated
mixture 8. Formation of a p-nitrobenzoate derivative allowed chro-
matographic isolation of pure cis-isomer. Upon ammonolysis at the
8-position, simultaneous deprotection of the benzoate ester was
realized to give 9, which was converted to compound 2b through
Suzuki chemistry. The syntheses of analogs 3a–g and 4a–d from
intermediates 10, 13 and 15 incorporated many of the same chem-
istries outlined for the cyclobutyl compounds in Scheme 1. In addi-
tion, the synthesis of 3a involved a lithium aluminum hydride
reduction of ester 10 to alcohol 11 that did not compromise the
8-chloro functionality and in fact was a quantitative conversion.
The formation of the amine-derived compounds 3b–g and 4a–d
also required extra steps for removal of the Cbz groups and subse-
quent final acylations.
Figure 3. Binding of 3b to the mTOR homology model.
exhibited poor in vitro metabolic stability with extraction ratios
(ER) greater than 0.9, that is, predicting for >90% first-pass
metabolism in vivo. Metabolic identification studies indicated that
oxidation of the cyclobutyl ring at its unsubstituted distal position
(30-position in Table 4) was the primary liability of this system,
therefore various modifications at this solvent-exposed region
were pursued while maintaining the C-1 substituent as 2-indolyl
(Tables 4–6).
As shown in Table 4, substitutions at the cyclobutyl 30-position
had the desired effect on reducing ERs in both mouse and human
liver microsomes. The hydroxyl (2a), hydroxymethyl (2b) and N-
Ac piperazinyl (2d) derivatives were tolerated, however the basic
N-Me piperazine derivative 2c was significantly less active. Inspec-
tion of the mTOR homology model indicated that the cationic
Lys2171 residue located in the proximity of the 30-substituents,
constituted a destabilizing factor for the compounds bearing a
basic amino group. This unfavorable electronic effect was also
observed for other compounds bearing a basic moiety (3b, 4a
and 4d). On the other hand, Lys2171 offered
a
potential
Cl
Cl
I
Cl
2a
d, e, f
g, e, f
N
N
N
a, b, c
N
NH2
N
N
N
N
2c-d
6
5
7
O
h, b
Cl
I
NH2
N
I
N
N
N
i, e
f
N
N
2b
OH
OH
8
9
Scheme 1. Preparation of compounds 2a–d. Reagents and conditions: (a) NMO (50% w/w in water), cat. potassium osmate dihydrate, THF; (b) N-iodosuccinimide, DMF, 60 °C;
(c) NaIO4, THF–water (3:1, v/v), rt; (d) NaBH4, MeOH–DCM (1:1, v/v), rt; (e) ammonia, 2-propanol, 110 °C, 24 h; (f) 2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-indole,
PdCl2(dppf), K2CO3, DME–water (4:1, v/v), 100 °C, 12 h; (g) NaBH(OAc)3, amine, DCE, rt; (h) 9-BBN, THF, then NaBO3ꢁH2O in water; (i) 4-nitrobenzoyl chloride, DIEA, DCM.

2096
A. P. Crew et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2092–2097
Cl
Cl
NH2
I
N
N
N
O
N
a
N
b, c
d
N
N
N
3a
N
O
HO
11
HO
12
10
NH2
N
Cl
I
NH2
N
Ar
N
N
N
N
N
b, c
d, e
N
f
N
3c-g
HN
Cbz
HN
Cbz
H2N
14
3b
13
NH2
N
Cl
I
NH2
N
Ar
N
N
N
d, e
g
N
b, c
N
N
4b-d
N
N
N
N
Cbz
Cbz
H
16
4a
15
Scheme 2. Preparation of compounds 3a–g and 4a–d. Reagents and conditions: (Ar = 1H-indol-2-yl): (a) lithium aluminum hydride, THF, ꢀ78 °C to rt; (b) N-iodosuccinimide,
DMF, 60 °C; (c) ammonia, 2-propanol, 110 °C, 24 h; (d) 2-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-indole, PdCl2(dppf), K2CO3, DME–water (4:1, v/v), 100 °C, 12 h; (e)
concd HCl, rt, 12 h; (f) carboxylic acid, EDC, DIEA, DCM for 3c–f; MeSO2Cl, DIEA, DCM for 3 g; (g) HOAc, EDC, DIEA, DCM for 4b; carbamoyl chloride, DIEA, DMF for 4c–d.
especially against other members of the mTOR pathway (IGF-1R,
Table 7
PDK1, PI3Kb) although overall 4c appeared slightly better in its
profile with P5-fold selectivity for mTOR over the kinases
evaluated.
Compounds 3c and 4c were also profiled in cells for mechanistic
effects on T37/46 p4E-BP1 and S473 pAkt.13 Despite their modest
biochemical potency, both 3c and 4c completely inhibited T37/46
4E-BP1 phosphorylation in MDA-MB-231 cells with IC50 values of
In vitro kinase selectivity of compounds 3c and 4c
Kinasea
3c
4c
mTOR
Abl
EGFR
cRaf
IGF-1R
IR
KDR
0.21
1.96
3.58
0.18
2.72
22.6
4.21
0.96
11.2
>30
0.12
>10
>10
1.7
>30
>30
>10
>10
>30
>30
0.77
7.4 and 7.8
strated a maximal inhibition of 40% at 20
agents inhibited S473 pAkt in BT-20 cells with IC50 values of 1.26
and 1.38 M, respectively, indicative of inhibition of mTORC2.
Rapamycin had no effect on inhibition of S473 pAkt at 20 M.
Functionally, both compounds demonstrated antiproliferative
lM, respectively. In contrast, rapamycin only demon-
lM. Additionally, both
MEK1
PDK1
l
PI3K
Src
c
5.30
l
a
Run in presence of 100
lM ATP.
effects in MDA-MB-231 cells with IC50 values of 7.4 and 9.3 lM,
Table 8
Mouse pharmacokinetic parameters for compounds 3c and 4c
160
Compound 3c
Compound 4c
Route of administration and parameters
3c
4c
140
iv
Dose (mg/kg)
CL (mL/min/kg)
1
2
5
7
120
100
80
60
40
20
0
t
(h)
1.6
6876
0.2
20
30.6
73824
54
0.9
12092
0.4
½
AUC0–1 (ng h/mL)
Vss (L/kg)
Dose (mg/kg)
p.o.
20
Cmax
(
l
M)
35.9
49421
102
AUC0–1 (ng h/mL)
F (%)
Control
4h
8h
Further in vitro characterization of compounds 3c and 4c in-
cluded an assessment of selectivity versus a small panel of kinases
(Table 7). Both agents offered a reasonable degree of selectivity
Figure 4. Pharmacodynamic inhibition of T37/46 4E-BP1 phosphorylation in MDA-
MB-231 xenograft following oral dosing.

A. P. Crew et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2092–2097
2097
and Ms. Viorica Lazarescu and Dr. Minghui Wang for analytical
support.
Additionally we would like to recognize Dr. Arno G. Steinig,
Dr. An-Hu Li and Mr. Anthony Nigro from the OSI IGF-1R team
by whom some of the chemical seeds and intermediates used in
this work were generated.11,12
350
300
250
200
150
100
50
Vehicle Control
100 mg/kg qd 1-14
References and notes
1. (a) Faivre, S.; Kroemer, G.; Raymond, E. Nat. Rev. Drug Disc. 2006, 5, 671; (b)
Guertin, D. A.; Sabatini, D. M. Cancer Cell 2007, 12, 9; (c) Wullschleger, S.;
Loewith, R.; Hall, M. N. Cell 2006, 124, 471; (d) Chiang, G. G.; Abraham, R. T.
Trends Mol. Med. 2007, 13, 433; (e) Meric-Bernstam, F.; Gonzalez-Angulo, A. M.
J. Clin. Oncol. 2009, 27, 2278.
0
1
6
11
Days
16
21
2. (a) Laplante, M.; Sabatini, D. M. J. Cell Sci. 2009, 122, 3589; (b) Proud, C. G.
Biochem. Soc. Trans. 2009, 37, 227.
3. (a) Loewith, R.; Jacinto, E.; Wullschleger, S.; Lorberg, A.; Crespo, J. L.; Bonenfant,
D.; Oppliger, W.; Jenoe, P.; Hall, M. N. Mol. Cell 2002, 10, 457; (b) Hay, N.;
Sonenberg, N. Genes Dev. 2004, 18, 1926.
Figure 5. Tumor growth inhibition of compound 4c in MDA-MB-231 xenograft.
4. (a) Sarbassov, D. D.; Guertin, D. A.; Ali, S. M.; Sabatini, D. M. Science 2005, 307,
1098; (b) Engelman, J. A. Nat. Rev. Cancer 2009, 9, 550.
5. O’Reilly, K. E.; Rojo, F.; She, Q.-B.; Solit, D.; Mills, G. B.; Smith, D.; Lane, H.;
Hofmann, F.; Hicklin, D. J.; Ludwig, D. L.; Baselga, J.; Rosen, N. Cancer Res. 2006,
66, 1500.
respectively, commensurate with the mechanistic data and sug-
gesting that these phenotypic effects were target mediated. Rapa-
mycin was unable to inhibit any proliferation in this cell line
6. (a) Bhagwat, S. V.; Crew, A. P.; Gokhale, P. C.; Yao, Y.; Kahler, J.; Arnold, L. D.;
Epstein, D. M.; Wild, R.; Pachter, J. A. Abstr. Am. Assoc. Cancer Res. 2009, 100,
3711; (b) Bhagwat, S. V.; Crew, A. P.; Gokhale, P. C.; Yao, Y.; Kahler, J.; Epstein,
D. M.; Wild, R.; Pachter, J. A. Abstr. Am. Assoc. Cancer Res. 2010, 101, 4487; (c)
Choo, A. Y.; Yoon, S.-O.; Kim, S. G.; Roux, P. P.; Blenis, J. PNAS 2008, 105, 17414.
7. (a) Raynaud, F. I.; Eccles, S.; Clarke, P. A.; Hayes, A.; Nutley, B.; Alix, S.; Henley,
A.; Di-Stefano, F.; Ahmad, Z.; Guillard, S.; Bjerke, L. M.; Kelland, L.; Valenti, M.;
Patterson, L.; Gowan, S.; Brandon, A. H.; Hayakawa, M.; Kaizawa, H.; Koizumi,
T.; Ohishi, T.; Patel, S.; Saghir, N.; Parker, P.; Waterfield, M.; Workman, P. Cancer
Res. 2007, 67, 5840; (b) Maira, S.-M.; Stauffer, F.; Brueggen, J.; Furet, P.; Schnell,
C.; Fritsch, C.; Brachmann, S.; Chène, P.; Pover, A. D.; Schoemaker, K.; Fabbro,
D.; Gabriel, D.; Simonen, M.; Murphy, L.; Finan, P.; Sellers, W.; García-
Echeverría, C. Mol. Cancer Ther. 2008, 7, 1851; (c) Chresta, C. M.; Davies, B. R.;
Hickson, I.; Harding, T.; Cosulich, S.; Critchlow, S. E.; Vincent, J. P.; Ellston, 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; (d) Thoreen, C.
C.; Kang, S. A.; Chang, J. W.; Liu, Q.; Zhang, J.; Gao, Y.; Reichling, L. J.; Sim, T.;
Sabatini, D. M.; Gray, N. S. J. Biol. Chem. 2009, 284, 8023; (e) Apsel, B.; Blair, J. A.;
Gonzalez, B.; Nazif, T. M.; Feldman, M. E.; Aizenstein, B.; Hoffman, R.; Williams,
R. L.; Shokat, K. M.; Knight, Z. A. Nat. Chem. Biol. 2008, 4, 691; (f) Yu, K.; Shi, C.;
Toral-Barza, L.; Lucas, J.; Shor, B.; Kim, J. E.; Zhang, W.-G.; Mahoney, R.; Gaydos,
C.; Tardio, L.; Kim, S. K.; Conant, R.; Curran, K.; Kaplan, J.; Verheijen, J.; Ayral-
Kaloustian, S.; Mansour, T. S.; Abraham, R. T.; Zask, A.; Gibbons, J. J. Cancer Res.
2010, 70, 621; (g) Jessen, K.; Wang, S.; Kessler, L.; Guo, X.; Kucharski, J.;
Staunton, J.; Lan, L.; Elia, M.; Stewart, J.; Bown, J.; Li, L. Mol. Cancer Ther. 2009, 8,
B148.
8. Bhagwat, S. V.; Crew, A. P. Curr. Opin. Invest. Drugs 2010, 11, 638.
9. Bhagwat, S. V.; Kahler, J.; Yao, Y.; Maresca, P.; Brooks, M.; Crew, A.; Boisclair,
M.; Pachter, J. A. Assay Drug Dev. Technol. 2009, 7, 471.
10. Mulvihill, M. J.; Cooke, A.; Rosenfeld-Franklin, M.; Buck, E.; Foreman, K.;
Landfair, D.; O’Connor, M.; Pirritt, C.; Sun, Y.; Yao, Y.; Arnold, L. D.; Gibson, N.
W.; Ji, Q. S. Future Med. Chem. 2009, 1, 1153.
11. Mulvihill, M. J.; Ji, Q.-S.; Coate, H. R.; Cooke, A.; Dong, H.; Feng, L.; Foreman, K.;
Rosenfeld-Franklin, M.; Honda, A.; Mak, G.; Mulvihill, K. M.; Nigro, A. I.;
O’Connor, M.; Pirrit, C.; Steinig, A. G.; Siu, K.; Stolz, K. M.; Sun, Y.; Tavares, P. A.
R.; Yao, Y.; Gibson, N. W. Bioorg. Med. Chem. 2008, 16, 1359.
12. Mulvihill, M. J.; Ji, Q.-S.; Werner, D.; Beck, P.; Cesario, C.; Cooke, A.; Cox, M.;
Crew, A.; Dong, H.; Feng, L.; Foreman, K. W.; Mak, G.; Nigro, A.; O’Connor, M.;
Saroglou, L.; Stolz, K. M.; Sujka, I.; Volk, B.; Weng, Q.; Wilkes, R. Bioorg. Med.
Chem. Lett. 2007, 17, 1091.
(IC50 >50
In addition to moderate ERs, both agents exhibited good perme-
ability (PAMPA >250 nm/s) and solubility (>100 M at pH 7.4) and
so were progressed to mouse PK studies. As shown in Table 8, both
agents demonstrated high plasma exposures, as well as low clear-
ance and good bioavailability on oral dosing at 20 mg/kg.
Despite the modest cell potencies of these compounds, the
exposures obtained on oral dosing were sufficient to evaluate the
potential of these agents to exert pharmacodynamic effects in vivo.
To compensate for plasma protein binding effects (90.3% and 88.1%
for 3c and 4c, respectively) these agents were dosed at 100 mg/kg
to CD-1 nude mice bearing MDA-MB-231 xenografts. As shown in
Figure 4, both agents were able to effect significant inhibition of
4E-BP1 phosphorylation at 4 and 8 h.
In order to assess whether such mechanistic target inhibition
in vivo was sufficient to drive an associated functional effect, com-
pound 4c was dosed orally for 14 days at 100 mg/kg qd to CD-1
nude mice bearing subcutaneous MDA-MB-231 xenografts. The
agent was well tolerated (body weight loss 62%) and this dose
indeed resulted in 94% tumor growth inhibition (Fig. 5).14
In summary, our efforts towards the discovery of dual mTORC1/
2 inhibitors led to the identification of a series of 1,3-disubstituted
imidazo[1,5-a]pyrazin-8-amines. Optimization of the 1- and 3-
substituents resulted in compounds 3c and 4c that combined
sub-micromolar mTOR biochemical potency, mechanistic and
phenotypic effects in a rapamycin resistant cell line, and in vitro
metabolic stability. Both agents also exhibited emerging kinase
selectivity and excellent pharmacokinetics on oral dosing. In
particular, proof-of-concept compound 4c was identified that dem-
onstrated in vivo target inhibition in xenografts, with commensu-
rate and significant inhibition of tumor growth in this tumor line
and thereby establishing the viability of this series as orally effica-
cious dual mTORC1/2 inhibitors. Further optimization of com-
pound 4c in terms of potency, selectivity and general in vivo
properties, leading to the discovery of the clinical mTORC1/2 agent
OSI-027 will be described in future communications.
lM).
l
13. Falcon, B. L.; Barr, S.; Gokhale, P. C.; Chou, J.; Fogarty, J.; Depeille, P.; Miglarese,
M.; Epstein, D. M.; McDonald, D. M. Cancer Res. 2011, 71, 1.
14. Female nu/nu CD-1 mice were used in the xenograft studies. To assess anti-
tumor efficacy, MDA-MB-231 cells were implanted into the mammary fat pad.
Tumors were allowed to establish to 200 50mm3 before randomization into
treatment groups. Tumor volumes were determined twice weekly from caliper
measurements by V = (length ꢂ width2)/2. Tumor growth inhibition (TGI) was
determined by %TGI = {1 ꢀ [(Tt/T0)/(Ct/C0)]/1 ꢀ [C0/Ct]} ꢂ 100, where Tt = tumor
volume of treated animal ꢂ at time t, T0 = tumor volume of treated animal ꢂ at
time 0, Ct = median tumor volume of control group at time t, and C0 = median
tumor volume of control group at time 0. Mean %TGI was calculated for the
entire dosing period for each group. Significant anti-tumor activity is defined
as mean %TGI >50%.
Acknowledgements
We would like to thank Mr. Paul Maresca and Ms. Maureen
Brooks for developing and running the HTS, Mr. Peter Meyn,
Ms. Jo Dunseath and Mr. Roy Turton for in vitro ADMET support,