Conformationally-restricted cyclic sulfones as potent and selective mTOR kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2012.05.104

Novel conformationally-restricted mTOR kinase inhibitors with cyclic sulfone scaffold were designed. Synthesis and structure-activity relationship (SAR) studies are described with emphasis on optimization of the mTOR potency and selectivity against class I PI3Kα kinase. PF-05139962 was identified with excellent mTOR biochemical inhibition, cellular potency, kinase selectivity and in vitro ADME properties.

Full text of DOI:10.1016/j.bmcl.2012.05.104

Bioorganic & Medicinal Chemistry Letters 22 (2012) 5114–5117
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Conformationally-restricted cyclic sulfones as potent and selective mTOR
kinase inhibitors
⇑
Kevin K.-C. Liu , Simon Bailey, Dac M. Dinh, Hieu Lam, Chunze Li, Peter A. Wells, Min-Jean Yin,
Aihua Zou
Pfizer Global Research and Development, Chemistry Department, 10770 Science Center Drive, La Jolla, CA 92120, USA
Pfizer Global Research and Development, Oncology, 10770 Science Center Drive, La Jolla, CA 92120, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Novel conformationally-restricted mTOR kinase inhibitors with cyclic sulfone scaffold were designed.
Synthesis and structure–activity relationship (SAR) studies are described with emphasis on optimization
of the mTOR potency and selectivity against class I PI3Ka kinase. PF-05139962 was identified with excel-
Received 5 March 2012
Revised 18 May 2012
Accepted 29 May 2012
Available online 6 June 2012
lent mTOR biochemical inhibition, cellular potency, kinase selectivity and in vitro ADME properties.
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
mTOR
Kinase inhibitors
Oncology
Conformationally-restricted
Mammalian target of rapamycin (mTOR) signaling pathway
plays a central role in driving tumor cell proliferation, survival,
angiogenesis and metastasis by responding to nutrients, growth
factors, and cellular energy level.1 mTOR is a mammalian serine/
threonine kinase discovered in 1994 and a member of PI3K like ki-
nase (PIKK) family of proteins. mTOR exists in two different com-
plexes,2 mTORC1, a rapamycin sensitive complex signaling to
S6K1 and 4E-BP1 to promote translation and cell growth; and
mTORC2, an rapamycin insensitive complex signals to AKT, an
important oncoprotein that activates a broad anti-apoptotic mech-
anism for cell survival. These two distinct mTOR components func-
tion through two different sets cell regulations. Rapamycin analogs
(rapalogues) are protein–protein inhibitors and allosteric inhibi-
tors of mTOR through mTORC1 but not mTORC2.3
Rapalogues such as RAD001 (everolimus, Novartis), CCI-779
(temsirolimus, Pfizer), and AP23573 (ridaforolimus, deforolimus,
Merck) entered anti-tumor clinical trials recently.4 Despite the
high expectation for their application in oncology based on sound
rationale related to the presumed mechanism-of-action, the rapa-
logues have only met with modest success. Most notable is the
utility of these agents as monotherapy in renal cell cancer (RCC)
and mantle cell lymphoma. Existence of this rapamycin insensitive
component of the mTOR signaling pathway thus provides new
opportunities to further inhibit mTOR related signaling. A mTOR
kinase inhibitor should prevent signaling through both mTORC1
and mTORC2 to have a broad and advantageous spectrum of phar-
macology over rapalogues to exploit the full therapeutic potential
of targeting mTOR. In addition, there are reports that rapalogues
led to PI3K/AKT activation and attenuated the agents’ antitumor
activities.5 It remains to be seen if the same activation is observed
in selective mTOR kinase inhibitors.
To find high quality leads of mTOR kinase inhibitors, we started
our own HTS and literature search in parallel. We were drawn to a
series of low molecular weight mTOR inhibitors published by
Astrazeneca colleagues6 as general structure 1 (Scheme 1). The
simple morpholinylpyrimidine core in these compounds provides
us a great opportunity to find other novel templates by scaffold-
hopping. After analyzing literature data on this series of com-
pounds, we found that most of these analogues have either small
alkyl groups or cyclic propyl group (as R2 and R3 in general struc-
O
N
O
N
conformational restriction
O
S
N
N
O
O2S
R4
N
N
R3 R2
N
N
O
N
O
N
R1
R1
2
1
⇑
Corresponding author.
E-mail address: Liu_Kang_Zhi_Kevin@Lilly.com (K.K.-C. Liu).
Scheme 1. Design of conformationally-restricted cyclic sulfones 2.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.05.104

K. K.-C. Liu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5114–5117
5115
Scheme 2. Both the (R) and (S) methyl morpholines have about the same potency and selectivity.
Scheme 3. PF-05139962, a potent and selective mTOR inhibitor.
ture 1) on the carbon in between pyrimidine core and the sulfonyl
shown in Scheme 2. Compound 3 has mTOR Ki = 25 nM and PI3K
a
side chain. Possible reasons for this chemical modification are (1)
to increase metabolic and chemical stability of the sulfonyl side
chain, (2) to restrict movement of the sulfonyl side chain, (3) to
stabilize a favorable binding conformation for better potency.
Based on these assumptions, we decided to make a series of con-
formationally-restricted cyclic sulfones (2, in Scheme 1) to test
our hypothesis.
Ki = 514 nM. It has about 20-fold selectivity over PI3K
a
. The com-
pound has low in vitro clearance as indicated by the low human li-
ver microsomes (HLM) and human hepatocyte (HHEP) clearance.
In addition, the compound should have good permeability based
on its good RRCK7 (Scheme 2). It is important for kinase inhibitors
to have good cell permeability to be biochemical efficient and
show good cellular potency.
5-Membered ring cyclic sulfone (3) with plain morpholine was
made first for testing, and encouraging result was obtained as
Methyl group was introduced to the morpholine moiety to see if
we can get better potency of mTOR and selectivity against PI3Ka by

5116
K. K.-C. Liu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5114–5117
Figure 1. PF-05139962 purposed binding in mTOR.
Figure 2. PF-05139962 kinase selectivity at 1 lM.
O
N
O
N
Table 1
PF-05139962 rat PK, no in vitro in vivo correlation
O
O
O
S
S
7
N
N
O
Chiral
N
O
O
N
O
N
N
N
N
8
CL
mL/min/kg
66.2
2.56
1.05
30
N
mTOR Ki = 50 nM
PI3Kα Ki > 4550 nM
HLM = 58 ul/min/mg
mTOR Ki = 120 nM
Vdss
T1/2
F
mL/kg
Hours
%
O
PI3Kα 7% @ 100 nM
HLM = 25 ul/min/mg
HHEP = 11 ul/min/million
S
N
O
HHEP = 28 ul/min/million
N
O
fe
%
5.5
Scheme 4. Different region-isomers.
N
N
increasing steric bulkiness and further restricting conformation of
the molecule. We did see selectivity against PI3K improvement
a
from 20Â to 80–100Â between 3 and 4R or 3 and 4S. However,
the mTOR biochemical activity of 4R or 4S was only slight improved
over compound 3. Interestingly, compounds 4R and 4S are about
equal potent for both mTOR and PI3Ka regardless the different ste-
ducing methyl group to the morpholine moiety, both compounds
6S (PF-05139962) and 6R are also inactive to PI3K Ki more than
4.5 M. The 6-membered ring cyclic sulfone appears to be a better
template for selectivity because of lacking PI3K activity. We are
pleased to see that these compounds still keep good ADME profile
(Scheme 3). Most interestingly and contrary from what was ob-
reochemistry of the methyl group (Scheme 2). With these interest-
ing data, we decided to expand the SAR to 6-membered ring cyclic
sulfone and walk the sulfone group to different positions in the ring.
Compound 5, the 6-membered ring cyclic sulfone as the homo-
analog of compound 3, is 2Â less potent than compound 3 on
a
l
a
mTOR kinases but it has PI3K
a
Ki more than 4.5
lM. After intro-
O
N
O
N
O
O
S
N
S
N
O
O
N
N
O
O
N
N
N
N
9
10
mTOR Ki = 5 nM, PI3Ka Ki = 3 uM
S473 IC50 = 55 nM
MW = 457
mTOR Ki = 7 nM, PI3Ka 18% @100 nM
S473 IC50 = 84 nM
MW = 443
O
N
O
S
N
O
N
O
N
N
11
mTOR Ki = 62 nM, PI3Ka 3% @100 nM
S473 IC50 = 1290 nM
MW = 429
Scheme 5. Other morpholine replacements.

K. K.-C. Liu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5114–5117
5117
O
O
OH
Cl
N
POCl3
NH2
N2H
O
O
NaH,
S
O
S
N
S
N
O
S
O
reflu
NaOEt,
15.3%
PhNMe
70%
O
N
OH
Cl
100%
D
A
B
C
O
O
B
O
O
O
O
O
N
N
N
Cl
N
N
H
N
H
O
N
H
N
N
G
oxone
N
N
N
O
S
S
S
N
O
PdCl2(dppf), DME,
H2O
Cl
K2CO3
,
S
THF, H2O
O
O
Cl
N
41.6%
Cl
N
H
N
H
30.5%
96.3%
D
E
F
PF-05139962
Scheme 6. Synthesis of PF-05139962.
served in the corresponding 5-membered ring sulfones (4R and
4S), compound 6S with (S)-methyl-morpholine moiety has mTOR
Ki = 8 nM which is superior to its stereoisomer 6R (mTOR
Ki = 52 nM).
Compound 5 was docked to our mTOR homogenous model as
shown in Figure 1. According to this model, the morpholinyl oxy-
gen interacts with Val-81 in the hinge region, oxygen of the sulfone
interacts with Trp-80 and the urea side chain play a critical role in
interacting with Lys-88 and Asp-36. Several attempts to replace
the urea group with other bioisosteres failed to deliver compounds
with decent mTOR inhibition since the flexible and free urea side
chain can maximize the H-bond interactions with Lys-88 and
Asp-36.
Two region-isomers, compounds 7 and 8 were synthesized
(Scheme 4). However, they are less potent than their correspond-
ing counter partners compounds 6S and 4R respectively. Com-
pound 7 also has higher clearance in HLM and HHEP as well.
We then moved our attention to modify the morpholine portion
of the molecule since data suggests this portion of molecule is re-
lated to compounds potency and selectivity. Among these modified
morpholine analogs, compounds 9–11 are highlighted and they
have desired potency and selectivity profiles as shown in Scheme
5. Especially, compound 9 and 10 that also have single digit mTOR
disconnection remains unknown and we decided to halt this series
compounds for more evaluation because of the uncertainty on hu-
man PK and dose predictions.
The synthesis of PF-05139962 is outlined in Scheme 6. Com-
pound A in anhydrous THF was treated with NaH at room temper-
ature then the mixture was heated to reflux to give crude
compound B, a thiopyranone, as a yellow oil which was used di-
rectly for next step without further purification. The crude thiopyr-
anone B was treated with urea and NaOEt in EtOH from rt. to 70 °C
to yield crude bicyclic compound C as a white solid. The crude
compound C was used directly and chlorinated with POCl3 in the
presence of N,N-dimethylaniline at 100 °C to give compound D in
70% yield as a yellow solid. 2-Methylmorpholine was coupled with
compound D in hot DMF under basic conditions (K2CO3) to give
compound E in 42% yield as a yellow solid. The thioether com-
pound E was oxidized with oxone in THF–H2O 1:1 mixed solvent
to generate the corresponding sulfone, compound F, as a yellow so-
lid in 93% crude yield. The crude compound F was coupled with
compound G, a phenylurea boronic ester, in the presence of Cs2CO3
and PdCl2(dppf) in DME–H2O mixed solvent at 100 °C under
microwave conditions to give PF-05139962 in 30.5% yield after
purification as a white solid.
In summary, a series of novel cyclic sulfones were designed
based on the concept of conformational restriction to generate po-
tent and selective mTOR inhibitors. Among these inhibitors, PF-
Kis, great selectivity against PI3K
a and have pS473 cellular IC50
<100 nM.
After further evaluating these sulfones, we decided to focus on
PF-05139962 which gives us most balanced profile that we are
looking for. As mentioned earlier, PF-05139962 is a potent mTOR
05139962 has more than 500-fold selectivity against PI3Ka and
good in vitro ADME profile. However, no in vitro in vivo PK corre-
lation was observed in rats and this disconnection confounds our
human PK predictions. Compounds in this series were halted for
further evaluation. Learning from this series of compounds was ap-
plied to develop better selective mTOR inhibitors and will be dis-
closed in due course.
inhibitor with great selectivity against PI3K
a and great in vitro
ADME profile. It has pS473 and pS6 cellular IC50 = 48 and 6 nM
respectively. It has great selectivity against other receptors and ki-
nases (Fig. 2 for its kinase selectivity). No genotoxicity was ob-
served on this compound and no more than 25% inhibiton was
observed for major CYP enzymes (3A4, 1A2, 2C9, 2D6) at 3
l
M.
References and notes
This compound has LE = 0.35 and LipE up to 6.8 which is in a very
desirable range for a kinase inhibitor.8
1. Engelman, J. A.; Luo, J.; Cantley, L. C. Nat. Rev. Genet. 2006, 7, 606.
2. Sarbassov, D. D.; Ali, S. M.; Sabatini, D. M. Curr. Opin. Cell Biol. 2005, 17, 596.
3. Guertin, D. A.; Sabatini, D. M. Sci. Signaling 2009, 2, No. Pe24.
4. Huang, S.; Houghton, P. J. Curr. Opin. Pharmacol. 2003, 3, 371.
5. Sun, S.-Y.; Rosenberg, L. M.; Wang, X.; Zhou, Z.; Yue, P.; Fu, H.; Khuri, F. R. Cancer
Res. 2005, 65, 7052.
6. Morris, Jeffrey James; Pike, Kurt Gordon 2009, WO 2009007748.
7. RRCK cell line (isolated from Madine–Darby Canine Kidney cells) was used to
assess compound permeability.
PF-05139962 was subjected to in vivo PK studies in rats to fur-
ther understand the PK-PD for this compound. Although, PF-
05139962 has good permeability and low in vitro clearance; high
clearance, short t1/2 and 30% bioavailability (F) were observed in
rats. The predicted in vivo clearance based on in vitro data is at
lease 10Â less than what was observed in in vivo in rats after pro-
tein binding correction (Table 1). The reason for this in vitro/in vivo
8. Edwards, M. P.; Price, D. A. Ann. Reports Med. Chem. 2010, 45, 381.