Potent, selective, and orally bioavailable inhibitors of the mammalian target of rapamycin kinase domain exhibiting single agent antiproliferative activity

Article abstract of DOI:10.1021/jm301389h

Selective inhibitors of mammalian target of rapamycin (mTOR) kinase based upon saturated heterocycles fused to a pyrimidine core were designed and synthesized. Each series produced compounds with K_i < 10 nM for the mTOR kinase and >500-fold sel

Full text of DOI:10.1021/jm301389h

Subscriber access provided by FONDREN LIBRARY, RICE UNIVERSITY
Article
Potent, Selective and Orally Bioavailable Inhibitors of
the Mammalian Target of Rapamycin (mTOR) Kinase
Domain Exhibiting Single Agent Antiproliferative Activity
Michael Koehler, Philippe Bergeron, Elizabeth Blackwood, Krista K. Bowman, Yung-Hsiang
Chen, Gauri Deshmukh, Xiao Ding, Jennifer Epler, Kevin Lau, Leslie Lee, Lichuan Liu, Cuong
Q. Ly, Shiva Malek, Jim Nonomiya, Jason Oeh, Daniel Fred Ortwine, Deepak Sampath, Steve
Sideris, Lan Trinh, Tom Truong, Jiansheng Wu, Zhonghua Pei, and Joseph P. Lyssikatos
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm301389h • Publication Date (Web): 30 Nov 2012
Downloaded from http://pubs.acs.org on December 5, 2012
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Potent, Selective and Orally Bioavailable Inhibitors
of the Mammalian Target of Rapamycin (mTOR)
Kinase Domain Exhibiting Single Agent
Antiproliferative Activity
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Michael F.T. Koehler*†, Philippe Bergeron†, Elizabeth Blackwood‡, Krista K. Bowman§, Yungꢀ
Hsiang Chen#, Gauri Deshmukh#, Xiao Ding#, Jennifer Epler‡, Kevin Lau†, Leslie Lee‡, Lichuan
Liu#, Cuong Ly†, Shiva Malek‡, Jim Nonomiya , Jason Oeh‡, Daniel F. Ortwine†, Deepak
Sampath‡, Steve Sideris , Lan Trinh , Tom Truong‡, Jiansheng Wu◊, Zhonghua Pei† and Joseph
P. Lyssikatos†.
†Department of Discovery Chemistry, ‡Department of Translational Oncology, §Department of
Structural Biology, Department of Small Molecule Biochemical Pharmacology, #Department of
Drug Metabolism and Pharmacokinetics and ◊Department of Protein Chemistry, Genentech, Inc.,
1 DNA Way, South San Francisco, CA 94080, United States
ABSTRACT: Selective inhibitors of mTOR kinase based upon saturated heterocycles fused to a
pyrimidine core were designed and synthesized. Each series produced compounds with Ki <
10nM for the mTOR kinase and > 500ꢀfold selectivity over closely related PI3 kinases. This
potency translated into strong pathway inhibition as measured by phosphorylation of mTOR
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 47
1
2
3
4
5
6
7
8
9
substrate proteins and antiproliferative activity in cell lines with a constitutively active PI3K
pathway. Two compounds exhibiting suitable mouse PK were profiled in in vivo tumor models
and were shown to suppress mTORC1 and mTORC2 signaling for over 12 hours when dosed
orally. Both compounds were additionally shown to suppress tumor growth in vivo in a PC3
prostate cancer model over a 14 day study.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION:
As the role of misregulation of signaling pathways in various cancers has become better
understood, activating mutations in the growth factor receptor tyrosine kinase/phosphoinositide
3ꢀkinase (PI3K)/Akt pathway stand out as being implicated in a multitude of cancers. More
recent work has shown that the mammalian target of rapamycin (mTOR) serves as an essential
component of this signaling pathway. mTOR is a serineꢀthreonine protein kinase of the PI3K
superfamily and forms two distinct signaling complexes, mTOR complexes 1 and 2 (mTORC1 &
2).
The mTORC1 complex is known to phosphorylate 4EBP1 and S6K1, thereby stimulating CAP
dependent translation, protein synthesis and cell proliferation. mTORC2 has been found to
phosphorylate and activate Akt upstream of mTORC1. The natural product rapamycin 1a and its
semisynthetic analogs temsirolimus 1b and everolimus 1c (Figure 1) are selective allosteric
inhibitors of most of the functions of mTORC1 but do not inhibit mTORC2. While everolimus
has demonstrated efficacy when combined with exemestane in hormone receptor positive breast
cancer,1 both temsirolimus and everolimus have been shown to have only modest efficacy in
treating renal cell carcinoma and mantle cell lymphoma. This limited efficacy has been ascribed
to the existence of two feedback pathways, mediated by mTORC2,
which induce
hyperphosphorylation of Akt in the presence of rapamycin.2 These results prompted a great
2
ACS Paragon Plus Environment

Page 3 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
interest in developing inhibitors of the kinase domain of mTOR which would inhibit both
mTORC1 and mTORC2, a number of which have been reported recently. 3ꢀ28
RESULTS AND DISCUSSION:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Having previously developed both the PI3K selective29 compound 2 (GDCꢀ0941, Figure 1)
and the potent dual PI3K/mTOR inhibitor30 3 (GDCꢀ0980, Figure 1), we focused on producing
an inhibitor specific for mTOR that would enable us to assess the role mTOR signaling was
playing in various cell lines and cancer types. Our starting point was morpholino
thienopyrimidine 4, which had been found to be a potent inhibitor of the mTOR kinase with a
Ki=3 nM which was also 20ꢀfold selective for mTOR over PI3Kα.
O
O
H
OH
OMe
1a
1b
R=H
R=
O
N
O
OH
O
O
OH
H
OMe
OR
O
OH
1c
R= CH2CH2OH
O
OMe
O
N
HO
O
O
S
N
O
N
O
N
O
N
S
N
N
S
N
S
N
N
NH
N
HO
N
O
N
N
N
N
N
H
H
N
NH2
2
3
(GDC-0980)
4
(GDC-0941)
Figure 1. Rapamycin and its analogs 1aꢀc, selective PI3K inhibitor 2 (GDCꢀ0941), dual
PI3K/mTOR inhibitor 3 (GDCꢀ0980) and 4, which exhibits 20 fold mTOR selectivity.
Our initial efforts were directed at moving away from the thienopyrimidine scaffold used in
our previous work, which we knew from previous experience could present difficulties with
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 47
1
2
solubility, oral bioavailability and poor pharmacokinetics.30 In an effort to improve the predicted
solubility, clogP and predicted metabolic stability of our core, we targeted replacement of the
thiophene with a saturated ring. Pyrrolidine was an attractive replacement as it provided a
functional handle for further diversification after assembly of the core. Molecules of this type
could be rapidly assembled from the Bocꢀ2,4ꢀdichloroꢀ5ꢀpyrrolopyrimidine 5 through
nucleophilic displacement of the C4 chloride followed by a Suzuki coupling at C2. Potassium
acetate was included in the coupling to minimize the cleavage of the ethyl urea, which was a
major side product when sodium carbonate was the exclusive base. Final diversification was
accomplished by removal of the Boc group followed by treatment with an appropriate
electrophile, as shown in Scheme 1.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The in vitro potencies of these compounds were measured against recombinant mTOR kinase
domain using a LanthaScreen FRET assay and against PI3Kα using a fluorescence polarization
assay, as previously described.31 PI3Kα was chosen as our initial antiꢀtarget as a representative
class I PI3 kinase. We were pleased to discover that, with the exception of the unsubstituted
pyrrolidine, all of these compounds were potent inhibitors of the mTOR kinase domain.
Unexpectedly, the compounds also exhibited improved selectivity against PI3Kα relative to the
starting thienopyrimidine (Table 1). At the same time, we observed chemical instability in a
subset of these molecules upon standing in DMSO solution in air. We did not characterize this
instability extensively, but supposed it stemmed from oxidation of the doubly activated benzylic
methylenes. We therefore began investigating alternative saturated ring systems to identify those
which would be tolerated in place of the pyrrolidine and would not exhibit this instability.
Scheme 1. Synthesis of pyrrolopyrimidines 6aꢀea
4
ACS Paragon Plus Environment

Page 5 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (a) morpholine, DIPEA, 90%; (b) (4ꢀethylurido)phenylboronic acid
pinacol ester, Pd(PPh3)4, KOAc, Na2CO3, H2O/CH3CN, 120 °C, 96%; (c) TFA, CH2Cl2, then PSꢀ
carbonate resin; (d) electrophile, DIPEA, CH2Cl2, (e) 2ꢀchloropyrimidine, DIPEA, DMF, 120 ºC
Table 1. Inhibition data for pyrrolopyrimidines 6aꢀea
O
N
N
R N
N
O
N
N
H
H
mTOR
Ki (nM)
PI3K
(fold)
R
6a
6b
6c
6d
6e
Boc
H
10
21
82
122
2-pyrimidine 3.0
126
82
SO2Me
3.7
5.1
CONHEt
42
aAssays described in reference 30.
We began by investigating the possibility of expanding the pyrrolidine ring to a piperidine,
thereby eliminating one of the doubly activated methylenes. Of the two possibilities, we first
examined those in which the nitrogen on the saturated ring was in the 6ꢀposition. The synthesis
of these compounds proved straightforward (Scheme 2), beginning with the formation of the
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 47
1
2
3
4
5
6
7
8
9
pyrimidine dione from methyl 1ꢀbenzylꢀ4ꢀoxopiperidineꢀ3ꢀcarboxylate. This could be converted
to the dichloride 7 and the benzyl protecting group exchanged for Boc using ACEꢀCl32 to give 8.
Nucleophilic aromatic substitution with a morpholine followed by Suzuki coupling and Boc
removal gave compounds 9a and 9b, which could be functionalized as needed through acylation,
sulfonylation or nucleophilic aromatic substitution to give the desired 6ꢀazaꢀ
tetrahydroquinazolines 10.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Synthesis of 6ꢀazaꢀtetrahydroquinazolines 10a
Cl
N
Cl
N
Bn
CO2Me
O
Bn
Boc
a, b
c-e
f - h
N
N
N
N
N
Cl
Cl
7
8
O
N
O
10a R1=Boc, R2=H
10b R1=Boc, R2=Me
R
N
N
N
R2
10c R1=MeSO2, R2=Me
10d R1=Ac, R2=Me
10e R1=2-pyrimidine, R2=H
10f R1=2-pyrimidine, R2=Me
10g R1=2-pyrazine, R2=H
10h R1=2-pyrazine, R2=Me
i
R1
HN
N
N
N
O
O
9a R=H
9b R=Me
N
H
N
N
N
H
H
H
aReagents and conditions: (a) urea, MeOH, NaOMe, reflux, 64%; (b) POCl3, reflux, 75%; (c)
ACEꢀCl, CH2Cl2, 0 °C to RT, then reflux; (d) MeOH, reflux; (e) PSꢀcarbonate resin, Boc2O,
CH2Cl2, 88% for three steps; (f) morpholine or (S)ꢀmethylmorpholine, DIPEA, DMF; (g) (4ꢀ
ethylurido)phenylboronic acid pinacol ester, Pd(PPh3)4, KOAc, Na2CO3, H2O / CH3CN, 120 °C;
(h) TFA, CH2Cl2, then PSꢀcarbonate resin; (i) electrophile, DIPEA, CH2Cl2
The 6ꢀazaꢀtetrahydroquinazolines 10aꢀh proved to be less potent but more selective inhibitors
of mTOR than the equivalent pyrrolopyrimidines 6, as seen in Table 2. Additionally, all proved
stable under our typical handling conditions. To investigate the source of the diminished activity,
we examined compound 10f docked into a homology model of mTOR that was built using PI3Kγ
as a template (Figure 2). The model indicates that the 6ꢀNꢀsubstituents make a close approach to
the protein, but are not predicted to make unfavorable interactions. The observed loss of affinity
could be partially compensated for by incorporating an (S)ꢀ2ꢀmethylmorpholine, which has been
6
ACS Paragon Plus Environment

Page 7 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
shown to enhance potency and selectivity for mTOR when used in place of morpholine as a
hinge binder9,21 as this methyl group projects into a hydrophobic cavity created by Trp2239 that
PI3Kα lacks. In the examples below, the methyl group improved mTOR potency an average of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6.6ꢀfold, with an 8.0ꢀfold improvement in selectivity against PI3Kα. Encouraged by the
improved selectivity of these compounds for the mTOR kinase, we decided to investigate the
isomeric 7ꢀazaꢀtetrahydroquinazoline series as our modeling indicated that substituents on the
piperidine nitrogen would be more completely solvent exposed.
Table 2. Inhibition data for 6ꢀazaꢀtetrahydroquinazoline compounds 10aꢀha
O
N
N
R2
R1
N
N
O
N
N
H
H
mTOR
Ki (nM)
PI3K
(fold)
R1
R2
10a
Boc
Boc
H
48
12
103/48
214/224
10b
10c
10d
Me
Me
Me
H
MeSO2
Ac
6.8 >1000/>1000
11
17
2.3
20
>900/>900
120/61
10e 2-pyrimidine
10f 2-pyrimidine Me
10g 2-pyrazine
>900/>900
69/91
H
10h 2-pyrazine Me
2.4 >1000/>1000
aAssays described in ref 30.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 47
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Compound 10f docked into a homology model of mTOR that was built using PI3Kγ as
a template9. Residues differing from those in PI3Kγ are labeled. The molecular surfaces for both
ligand and protein are shown to illustrate the intimate interaction provided by the presence of the
(S)ꢀmethyl substituent on the morpholine. The protein surface is color coded by hydrophobicity
(tanꢀbrown is hydrophobic, cyanꢀblue is hydrophilic, green is in between). The ligand surface is
magenta.
We prepared the 7ꢀazaꢀtetrahydroquinolines using chemistry analogous to that developed for
the synthesis of the pyrrolopyrimidines (Scheme 3). When assayed (Table 3), we found that this
series of compounds had recovered the potency and selectivity for mTOR over both PI3Kα and
PI3Kδ we had observed in the pyrrolidine series, and all were again found to be stable under
typical storage conditions.
Scheme 3. Synthesis of 7ꢀazaꢀtetrahydroquinazolines 12aꢀga
8
ACS Paragon Plus Environment

Page 9 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
N
O
Cl
N
a
b
N
N
N
N
N
N
Boc
Cl
R
N
O
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Boc
N
Cl
N
N
12b R=Boc
12a R=H
11
c
H
H
e, f
d
12c R=SO2Me
12d R=CO2Me
12e R=Bz
or g, h
c, d
O
N
O
N
N
N
N
N
Boc
N
O
Y
Bz
N
O
Y
X
X
13a X= NH, Y= OEt
13b X= O, Y= NHEt
12f X= NH, Y= OEt
12g X= O, Y= NHEt
aReagents and conditions: (a) morpholine, DIPEA; (b) (4ꢀethylurido)phenylboronic acid pinacol
ester, Pd(PPh3)4, KOAc, Na2CO3, H2O/CH3CN, 120 °C; (c) TFA, CH2Cl2, then PSꢀcarbonate
resin; (d) electrophile, DIPEA, CH2Cl2; (e) 4ꢀhydroxyphenylboronic acid pinacol ester,
Pd(PPh3)4, KOAc, Na2CO3, H2O/CH3CN, 120 °C, 89%; (f) EtNCO, THF, 70 °C, 24h, 38%; (g)
4ꢀaminophenylboronic acid pinacol ester, Pd(PPh3)4, KOAc, Na2CO3, H2O/CH3CN, 120 °C,
86%; (h) ethyl chloroformate, THF, DIPEA, 70%
Our initial inhibition data indicated that both polar and nonpolar functionality was tolerated on
the piperidine nitrogen with little change in the biochemical potency. Docking of these molecules
into a homology model of mTOR shows Nꢀsubstituents are directed into a largely solvent
exposed region of the protein (Figure 3). The basic piperidine, left unmodified, reduced the
potency approximately ten fold, but reducing the basicity through formation of a carbamate,
sulfonamide, urea or ester resulted in compounds of approximately equivalent Ki. Modeling
shows this nitrogen is in a fairly hydrophobic region of the protein, flanked by Ile2163, Leu2185,
Trp2239, Met2345, and Ile2356, and thus poorly tolerant of a formal charge. Modifications of
the ethyl urea were much less well tolerated; replacement of the nitrogen adjacent to the phenyl
ring with an oxygen (12f) resulted in a 200ꢀfold reduction in mTOR binding and an almost
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 47
1
2
3
4
5
6
7
8
9
complete loss of selectivity against PI3Kα and δ. The second urea nitrogen was also replaced
with oxygen (12g), resulting in a 20ꢀfold reduction in binding, but no significant loss in
selectivity. The requirement for hydrogen bond donors in both positions is consistent with our
binding model, in which each of the urea nitrogens forms a hydrogen bond to the sidechain of
Asp2195.9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Having now developed a number of compounds potent in the biochemical assay, we examined
their cellular activity in two separate assays. For this, we used the pTENꢀnull prostate carcinoma
NCIꢀPC3 cell line, in which the loss of pTEN leads to phosphorylation of Akt at threonine 308,
activating the Akt kinase, which then phosphorylates and activates mTOR along with its other
substrates. In this cell line, we measured the inhibition of the phosphorylation of serine 473 of
Akt, a known mTORC2 substrate, and also assessed the antiproliferative activity of our
compounds more broadly through determination of an IC50 in a 3ꢀday viability assay (Table 3).
In each case, we found that cellular potency tracked roughly with the biochemical potency of the
compounds, with compound 12h being the most potent in both assays.
Table 3. In vitro data for 7ꢀazaꢀtetrahydroquinazoline compounds 12aꢀh.a
10
ACS Paragon Plus Environment

Page 11 of 47
Journal of Medicinal Chemistry
1
2
3
4
O
5
6
7
8
N
N
N
9
R
X
N
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
X
Y
mTOR
Ki (nM)
pAKT
IC50 (nM)
PI3K
(fold)
NCI-PC3
IC50 (nM)
R
Y
12a
12b
12c
12d
12e
12f
H
NH
NH
NH
NH
NH
O
NH
NH
NH
NH
NH
NH
O
140
12
36 / 17
48 / 79
10 / 50
22 / 92
38 / 86
6.5 / 6.5
79 / 79
60/130
nd
22
43
37
22
nd
115
25
nd
nd
Boc
MeSO2
CO2Me
Bz
21
430
220
340
nd
12
6.5
1500
130
Bz
NH
12g
12h
Bz
4800
290
2-pyrimidine NH
NH
5.5
aAssays described in ref 30. nd = not determined
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 47
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Compound 12i docked into a homology model of mTOR that was built using PI3Kγ as
a template9. Residues differing from those in PI3Kγ are labeled. Predicted hydrogen bonding
interactions with the protein are shown as dashed ovals, and the molecular surfaces for both
ligand and protein are shown to illustrate the intimate interaction provided by the presence of the
(S)ꢀmethyl substituent on the morpholine. The protein surface is color coded by hydrophobicity
(tanꢀbrown is hydrophobic, cyanꢀblue is hydrophilic, green is in between). The ligand surface is
yellow.
As it had been successful in the case of the 6ꢀazaꢀtetrahydroquinazolines, we again attempted
to improve the in vitro potency and selectivity of the 7ꢀazaꢀtetrahydroquinazoline compounds by
incorporating (S)ꢀ2ꢀmethylmorpholine as our hinge binder (Table 4). This proved successful, and
we sought to better quantify the effect by evaluating the potency and selectivity of all 51 pairs of
7ꢀazaꢀtetrahydroquinazolines in our database. Each pair contained the phenethyl urea and
differed only in the presence or absence of the 2ꢀS methyl group on the morpholine. Matched
pair analysis33 showed that the change in mTOR Ki was modest over these pairs, with those
12
ACS Paragon Plus Environment

Page 13 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
molecules containing the methyl group being 3.2ꢀfold more potent on average. The effect on
selectivity was much more pronounced, with the methyl group containing compounds producing
an average increase of 19.5ꢀfold in selectivity for PI3Kα. This is consistent with our modeling
results which predict an improved interaction with Trp2239 in mTOR as well as an unfavorable
interaction with the larger phenylalanine residue in place of Leu2185 in PI3Kα. We were unable
to observe any additional improvement in potency and or selectivity by incorporating a 2ꢀS ethyl
morpholine in compound 12j.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4. In vitro data for 7ꢀazaꢀtetrahydroquinazoline compounds 12iꢀna
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 47
1
2
3
4
O
5
6
7
8
N
N
R2
N
N
9
R1
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
N
H
H
potency/selectivity
target modulation
proliferation
mTOR
PI3K
pAKT
pp70S6K
NCI-PC3 MCF7neo/Her2
Ki (nM)
IC50 (nM) IC50 (nM)
R1
R2
Me
IC50 (nM)
IC50 (nM)
(fold)
N
1.5
1.3
3.3
1.5
9.1
nd
150
57
500/>1000
410/>1000
12i
12j
N
N
N
40
55
Et
38
N
N
N
1.7
4.3
7.8
110
320/470
12k
Me
N
1.4
3.5
13
35
19
140
230
290
69
94/180
12l
Me
Me
N
N
8.7
410/360
12m
H3C
S
N
4.4
340/570
22
62
310
130
12n
Me
N
O
H3C
aAssays described in ref 30. nd = not determined
Having determined that the 7ꢀazaꢀtetrahydroquinazolines with a 2ꢀS methyl morpholine as
hinge binding element provided superior potency and selectivity, we varied the heterocycles
appended to the 7ꢀposition of our scaffold. We examined the diazines and a few additional
electronꢀpoor heterocycles, as shown in Table 4, and found that they exhibited similar singleꢀ
digit nanomolar potency with selectivity varying, but almost always exceeding 100 fold against
14
ACS Paragon Plus Environment

Page 15 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
both PI3Kα and PI3Kδ. We observed the expected suppression of both mTORC1 and mTORC2
in a pTENꢀnull prostate carcinoma NCIꢀPC3 cell line34 as measured by phosphorylation of
p70S6K and Akt, respectively. We were pleased to observe potent antiproliferative activity for
these compounds on both the PTEN null NCIꢀPC3 prostate cell line as well as the MCFꢀ
7neo/Her2 breast cancer cell line, which has an activating mutation in PI3K. Additionally,
compounds 12hꢀn showed no inhibition of the cytochrome P450 (CYP) mediated metabolism of
standard substrates for the 3A4 (midazolam and testosterone), 2D6 (dextromethorphan), 2C9
(warfarin), 1A2 (phenacetin) or 2C19 (mephenytoin) isoforms35 indicating that they were at a
reduced risk for drugꢀdrug interactions when coꢀdosed with compounds metabolized principally
by only one of these isoforms.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As compounds 12hꢀn exhibited similar potency and selectivity in our enzymatic assays as well
as potent antiproliferative activity in two separate cell lines, we used a series of mouse PK
studies to determine which were most suitable for advancement to in vivo tumor models. We
included similar compounds 10f and 10h from our 6ꢀaza tetrahydroquinazoline series for
comparison as well. We were gratified to observe that a microsomal model of metabolic stability
enabled us to predict the in vivo clearance of compounds 12 in mice fairly well (Table 5). The
intravenous clearance for these compounds ranged from 24ꢀ49% of mouse liver blood flow while
the bioavailability of these compounds ranged from 22ꢀ70%. Compound 12h appeared most
attractive overall, combining the lowest clearance with the highest oral bioavailability, resulting
in an AUC after oral dosing which was twoꢀfold higher than the other 7ꢀazaꢀ
tetrahydroquinazolines examined.
Table 5. In vitro and in vivo PK data for selected 7ꢀazaꢀtetrahydroquinazolines
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 47
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aCompound dosed iv as a solution in 80% PEG/10% EtOH and dosed PO as a suspension in
MCT. Clhep predicted from incubation with microsomes. h/r/m = human/rat/mouse. nd = not
determined
b
c
d
In addition to PK analysis, compound 12h was examined in a panel of 250 kinases at a
concentration of 1 ꢀM. It proved exquisitely selective with no inhibition greater than 25%
observed in kinases other than mTOR and the three class I PI3Kα, γ and δ, which were
approximately 50% inhibited at 1ꢀM. Compounds 10f and 12i were spot checked against a
smaller panel of 50 kinases and exhibited still greater selectivity, with no inhibition greater than
25% for any kinase assayed at 1ꢀM, including the PI3K isoforms. This illustrates the general
nature of the selectivity derived from this scaffold and the still greater selectivity possible with 2ꢀ
S methyl morpholine used as the hinge binder.
At this point, we had a number of potent and selective compounds in hand that exhibited
sufficiently long half lives in vivo suitable for further study of the effects suppression of mTOR
kinase activity has on antiproliferative potency. We began by investigating the ability of these
compounds to suppress signaling through the PI3K/Akt/mTOR pathway using NCIꢀPC3 prostate
16
ACS Paragon Plus Environment

Page 17 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
tumors implanted subcutaneously into Taconic NCR.nude mice. Compounds 12h and 12i were
selected for this study due to their potency, selectivity and favorable mouse PK profile.
After administration at three separate doses, tumor and plasma samples were obtained after six
hours (Figure 4). The plasma levels observed for the two compounds were as expected based on
our previous oral administration in mice. The pharmacodynamic (PD) effect of our inhibitors
was assessed by measuring the fold reduction of phosphoꢀserine 437, normalized to total Akt, as
well as the fold reduction of phosphoꢀserines 235 and 236, normalized to total S6, relative to
their timeꢀmatched vehicle control. Inhibition of S6 phosphorylation is indicative of inhibition of
the mTORC1 complex while inhibition of Akt phosphorylation is indicative of inhibition of the
mTORC2 complex. Both molecules potently inhibited phosphorylation of both substrates after
six hours. The performance of each drug was comparable to that of 10 mg/kg GDCꢀ0980, which
was run as a contemporaneous positive control, and can be considered a benchmark of maximum
knockdown for this model and assay method.36 12h showed a dose response between 12.5 mg/kg
and 100 mg/kg, while all doses of 12i tested performed similarly to GDCꢀ0980. We did not
repeat the study of 12i at lower doses to determine the lowest dose capable of achieving full
knockdown of mTOR substrates.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 47
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Plasma levels of 12h and 12i 6 hours following oral administration in PC3 tumorꢀ
bearing mice along with the fold decreases of phosphorylated mTORC1 and 2 substrates relative
to timeꢀmatched vehicle controls. Error bars represent one standard deviation from the depicted
mean value. Four mice were dosed for each compound, and phosphoprotein levels were
measured in duplicate samples for each mouse.
We next sought to evaluate the difference in inhibition profile between the present mTOR
kinase inhibitors and rapamycin, which would be expected to reduce only phosphorylation of
mTORC1 substrates such as S6. We therefore repeated our PK/PD experiment, sampling the
tumor and blood over a 24h period, comparing a 100 mg/kg dose of 12i to a 10mg/kg dose of
rapamycin (Figure 5). In this instance, we used the U87 glioma model, which is also PTEN
18
ACS Paragon Plus Environment

Page 19 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
deficient. By extending sampling to 24 hours, we were able to observe the clearance of 12i and
the recovery of phosphoꢀAkt and phosphoꢀS6 levels to vehicle levels by the end of the
experiment. As anticipated, we found that administration of rapamycin inhibited phosphorylation
of S6, but did not reduce phosphoꢀAkt levels. Instead, we observed the previously noted
hyperphosphorylation of Akt on serine 473 at all time points beyond one hour.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Plasma levels of 12i and rapamycin following administration in U87 tumorꢀbearing
mice along with the fraction of phosphorylated mTORC1 and 2 substrates relative to time
matched vehicle controls. Error bars represent one standard deviation from the depicted mean
value. Four mice were dosed for each compound and time point, and phosphoprotein levels were
measured in duplicate samples for each mouse.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 47
1
2
3
4
5
6
7
8
9
Having obtained an understanding of the extent of PI3K/Akt/mTOR pathway inhibition we
could obtain from our compounds at a given dose and time point, we decided to probe the extent
to which these compounds could suppress tumor growth in vivo. A 14ꢀday efficacy study was
run in the NCIꢀPC3 model, with daily dosing. 12h was administered at doses from 6.25 mg/kg to
100 mg/kg QD in an effort to span the range of pathway suppression observed in our PK/PD
experiments. Rapamycin was dosed at 10 mg/kg, a dose previously determined to be efficacious
in this model.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
After 14 days of dosing, all animals remained on study, but the 100 mg/kg dose group had lost
slightly more than 10% of their starting body weight, which triggers a dosing holiday in our
protocol, and indicates that we are approaching the maximum tolerated dose (MTD). By way of
comparison, the vehicle group lost 3.7% of their body weight, as this model is mildly cachexic.
The tumor growth curves (Figure 6a) were calculated by linear mixed effects (LME) modeling,
and clearly illustrate that stronger pathway inhibition results in steadily increasing inhibition of
growth and that tumor stasis has not been reached at the highest dose examined. The 10 mg/kg
QDx14 course of rapamycin performed well in this model, approaching tumor stasis. While
tolerated, this dose achieves a greater rapamycin exposure level in mice than can be analogously
achieved in humans37ꢀ39, which may account for its efficacy in this model.
We subsequently examined the growth inhibitory properties of the more potent 12i in the same
model, again at a range of doses. At the highest dose of 200 mg/kg dosed QD, we again observed
body weight loss exceeding 10% at the end of the study. We observed stronger tumor growth
inhibition at all doses of 12i than was observed for 12h at the identical dose. We again noted that
maximum inhibition of tumor growth was only achieved at doses greater than that which could
be tolerated by the mice without a dosing holiday over the course of a two week study (Figure
20
ACS Paragon Plus Environment

Page 21 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6b). In this study, we noted that the 50 mg/kg dose appeared more efficacious than the 100
mg/kg dose, but attribute this to experimental variability resulting from our small cohort size of 5
animals per group. This was substantiated by the results of a second study performed
simultaneously with material from the same lot of 12i, in which we examined the effect of BID
dosing at the same total dose per day. In this study, the 50 mg/kg/day dose resulted in slightly
lower efficacy than the 100 mg/kg/day dose as measured by the average tumor volume (data not
shown).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 47
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Growth inhibition curves for compounds 12h (a) and 12i (b) used to treat
subcutaneously implanted PC3 tumors in mice.
These studies clearly demonstrated that we had met our goal of producing a potent, selective,
orally available inhibitor of the mTOR kinase. Our compounds robustly inhibited tumor growth
in cell lines in which the PI3K/Akt/mTOR signaling pathway is constitutively active due to the
loss of PTEN. 12i achieved tumor stasis at the highest 200 mg/kg/day dose examined, which
appears to also be approaching the limit of tolerability for this molecule.
CONCLUSIONS:
We have developed a novel class of potent, highly selective selective mTOR kinase domain
antagonists. The addition of saturated heterocycles fused to the central pyrimidine ring, along
with the previously described phenyl ethyl urea and Sꢀmethyl morpholine hinge binder provided
compounds with nanomolar in vitro potency and high selectivity against PI3K and other kinases.
This biochemical potency translated into potent growth inhibition in cell lines where the
PI3K/Akt/mTOR signaling pathway is constitutively active due to the loss of PTEN or due to an
22
ACS Paragon Plus Environment

Page 23 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activating mutation in PI3K. Compounds 12h and 12i suppressed mTORC1 and mTORC2 in
cellular and in vivo PK/PD experiments, and did not induce the hyperphosphorylation of Akt as
rapamycin did at time points after 3 hours. This potential advantage did not, however, translate
into greater efficacy in our PC3 xenograft model when rapamycin and mTOR inhibitors were
compared at doses approaching their MTD in mice. This may be due in part to the fact that the
rapamycin levels achieved in mice cannot be reproduced in man, limiting its efficacy. What
remains to be understood is whether the tolerability in man of these different modes of pathway
inhibition differs and whether mTOR can be effectively combined with other targeted
chemotherapeutics. Compounds 12h and 12i were found to exhibit time dependent inhibition of
CYP 3A4, which prevented their progression into the clinic, however further refinement of this
scaffold enabled us to select a clinical candidate which joined the PI3K selective and dual
PI3K/mTOR inhibitors undergoing human clinical trials in an effort to address these questions
directly.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION:
All chemicals were purchased from commercial suppliers and used as received. Flash
chromatography was carried out with prepacked silica cartridges from either ISCO or SiliCycle
on an ISCO Companion chromatography system using gradient elution. NMR spectra were
recorded on a Bruker AV III 400 or 500 NMR spectrometer and referenced to tetramethylsilane.
Preparative HPLC was performed on a Polaris C18 5 ꢁm column (50 mm x 21 mm), eluting with
mixtures of water/acetonitrile. All final compounds were purified to >95% chemical and optical
purity, as assayed by LC/MS. This analysis was performed on an Agilent 1100 HPLC coupled
with Agilent MSD mass spectrometer using ESI as the ionization source. The LC separation was
performed using an Agilent ZORBAX SBꢀC18, 1.8mm, 30 x 2.1 mm column with a flow rate of
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 47
1
2
3
4
5
6
7
8
9
0.4 ml/minute. The gradient consisted of 3 ꢀ 97% acetonitrile in water, each containing 0.05%
TFA, over 7 min. 97% acetonitrile was then maintained for 1.5 min, followed by reequilibration
at 3% acetonitrile for 1.5 min. The LC column temperature was maintained at 40 ˚C. UV
absorbance was measured at 220 and 254nm along with a full scan mass spectrum. For select
compounds, a longer gradient LC was used (method B). In these cases, the gradient consisted of
2 ꢀ 98% acetonitrile in water, each containing 0.05% TFA, over 25.5 min. 98% acetonitrile was
then maintained for 2.5 min, followed by reequilibration at 2% acetonitrile for 1.5 min. UV and
MS detection was performed as before.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
InꢀLife PK−PD Study. Human prostate cancer NCIꢀPC3 cells (National Cancer Institute,
Frederick, MD) were implanted subcutaneously into the right hind flanks of female NCR.nude
mice (5 × 106 cells in 100 ꢁL of Hank’s balanced salt solution). Tumors were monitored until
they reached a mean tumor volume of approximately 500 mm3. Then similarly sized tumors were
randomly assigned to groups (n = 4). Compounds were formulated as suspensions in 0.5%
methylcellulose/0.2% Tween 80 (MCT) and dosed orally at 25, 50, and 100 mg/kg (100 ꢁL
dose/25 g animal). Tumor and plasma samples were harvested at 1, 6, and 10 h postdose.
Blood samples were harvested by terminal cardiac puncture and collected into tubes containing
K2EDTA as an anticoagulant. Samples were kept chilled on ice until centrifugation at
(1500−2000)g for 5 min; plasma was transferred to a new vial and snap frozen. Tumor samples
were transferred to a 1.5 mL snapꢀcap polypropylene tube, weighed, and then flash frozen on dry
ice. Samples were stored at −80 °C prior to PD biomarker evaluation.
Xenograft Tumor Lysates. Frozen tumors were lysed in iceꢀcold extraction buffer by
mechanical disruption and agitation with garnet/ceramic sphere matrix in a FastPrepꢀ24 (MP
Biomedicals, Solon, OH; five 20 s cycles with rest periods on wet ice). Extraction buffer
24
ACS Paragon Plus Environment

Page 25 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(Biosource; Carlsbad, CA) was supplemented with protease inhibitors (F. HoffmanꢀLa Roche;
Mannheim, Germany), 1 mM phenylmethysufonyl fluoride, and phosphatase inhibitor cocktails
1 and 2 (SigmaꢀAldrich; St. Louis, MO). After thorough disruption, tumor lysates were clarified
by centrifugation at 4 °C. Protein concentrations were determined using the BCA protein assay
kit (Pierce; Rockford, IL) and normalized to a standard concentration of 5 mg/mL.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
PD Biomarker Assays. The Meso Scale Discovery MultiꢀSpot biomarker detection system
(Meso Scale Discovery; Gaithersburg, MD) was used to determine the levels of Akt
phosphorylated at serine 473 (pAkt) and S6RP phosphorylated at serine 235/236 (pS6RP). These
doubleꢀdeterminant immunoassays quantify protein levels on the basis of measurements of
electrochemoluminescence intensity. In the case of pS473 and total Akt, both of these were
measured in the same well using a duplex assay plate. All other markers required separate plates
to measure phosphorylated and total protein. Protein was loaded at 20 ꢁg/well for Akt and 10
ꢁg/well for S6RP.
Relative levels of phosphoprotein were determined by comparison of treated tumor lysates
with an average of the timeꢀmatched vehicle controls.
InꢀLife Efficacy Study. NCIꢀPC3 cells were subcutaneously implanted as above. Tumors
were grown to an average volume of 200mm3 before dosing was started. Extra mice were
implanted, to insure that each treatment group of 5 mice had similarlyꢀsized tumors. Tumor
volumes were determined using digital calipers (Fred V. Fowler Company, Inc.) using the
formula (L x W x W)/2. Average tumor volume was calculated by applying curve fitting to the
Log2 transformed individual tumor volume data using a linear mixedꢀeffects model using the R
package nlme, version 3.1ꢀ97 in R v2.12.0. Percent tumor growth inhibition, %TGI, was
calculated as the percentage of the area under the fitted curve (AUC) for the respective dose
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 47
1
2
3
4
group per day in relation to the vehicle, such that %TGI
=
100
x
1ꢀ
5
6
7
8
(AUCtreatment/day)/(AUCvehicle/day). Tumors were measured, and mice weighed twice per
week. If weight loss was >15%, the mice were weighed daily and euthanized if weight loss
reached 20%. 10% body weight loss was the limit for tolerated doses. Mice whose tumors
reached 2000mm3 were euthanized.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
tertꢀbutyl 2ꢀchloroꢀ4ꢀmorpholinoꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidineꢀ6(7H)ꢀcarboxylate. To a
mixture of tertꢀbutyl 2,4ꢀdichloroꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidineꢀ6(7H)ꢀcarboxylate (1.2 g, 4.1
mmol) and N,Nꢀdiisopropylamine (1.4 mL, 8.3 mmol) in isopropanol (8 mL) was added
morpholine (0.430 mL, 5.0 mmol). The mixture was stirred at room temperature for 1hr and was
then concentrated in vacuo. The resulting residue was purified by flash column chromatography
(40% EA/Hex) to give tertꢀbutyl 2ꢀchloroꢀ4ꢀmorpholinoꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidineꢀ6(7H)ꢀ
carboxylate(1.378 g, 98%).
tertꢀbutyl 2ꢀ(4ꢀ(3ꢀethylureido)phenyl)ꢀ4ꢀmorpholinoꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidineꢀ6(7H)ꢀ
carboxylate
(6a).
2ꢀchloroꢀ4ꢀmorpholinoꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidineꢀ6(7H)ꢀcarboxylate
(0.366 mmol), (4ꢀethylureido)phenylboronic acid pinacol ester (0.439 mmol),
tetrakis(triphenylphosphine)palladium (0.022 mmol), 1 M aq. KOAc (0.55 mL), 1 M aq. Na2CO3
(0. 55 mL) and MeCN (2 mL) were mixed in a microwave reaction tube. The mixture was heated
in a microwave reactor at 120 °C for 25 min. The mixture was extracted with EtOAc (3x) and the
combined organic extracts were concentrated. The resulting residue was purified by flash column
chromatography (40% EA/Hex) to give 6a (149 mg, 83%). LC/MS: 4.43 min, M+=469.2,
MW=468.6 1H NMR (500 MHz, CDCl3) δ 8.28 (d, J = 8.7, 2H), 7.38 (d, J = 8.8, 2H), 7.10 (d, J
= 8.6, 1H), 6.81 (d, J = 8.8, 1H), 4.78 (s, 2H), 4.57 (s, 2H), 3.93 – 3.59 (m, 8H), 3.36 – 3.29 (m,
2H), 1.53 (s, 9H), 1.17 (t, J = 7.2, 3H).
26
ACS Paragon Plus Environment

Page 27 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1ꢀethylꢀ3ꢀ(4ꢀ(4ꢀmorpholinoꢀ6,7ꢀdihydroꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidinꢀ2ꢀyl)phenyl)urea
(6b). tertꢀbutyl 2ꢀ(4ꢀ(3ꢀethylureido)phenyl)ꢀ4ꢀmorpholinoꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidineꢀ6(7H)ꢀ
carboxylate (c, 1.74 mmol) was dissolved in CH2Cl2 (5 mL), then TFA (3 mL) was added. The
mixture was stirred at room temperature for 2 hours. Isopropanol and toluene were added and the
resulting mixture was concentrated in vacuo. The resulting residue was azeotroped with toluene
two more times. Methanol and methylene chloride (4 mL each) were added to the resulting
residue, then PSꢀcarbonate (2.5ꢀ3.5 mmol N/gram resin, 1.45 g) was added. The mixture was
stirred at room temperature until the pH reached >7 (~ 1 hr). The mixture was filtered, and the
resin was washed with methanol and CH2Cl2. The filtrate was concentrated to provide 6b, which
was used without further purification in subsequent reactions. A portion of the material was
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
purified by reverse phase HPLC for assay. LC/MS: 2.74 min, M+=369.2, MW=368.4 H NMR
(400 MHz, DMSO) δ 9.57 (s, 1H), 8.75 (s, 1H), 8.19 (d, J = 8.9, 2H), 7.50 (d, J = 8.9, 2H), 6.38
– 6.08 (m, 1H), 4.71 (s, 2H), 4.38 (s, 2H), 3.72 (s, 8H), 3.24 – 3.03 (m, 2H), 1.06 (t, J = 7.2, 3H).
1ꢀethylꢀ3ꢀ(4ꢀ(4ꢀmorpholinoꢀ6ꢀ(pyrimidinꢀ2ꢀyl)ꢀ6,7ꢀdihydroꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidinꢀ2ꢀ
yl)phenyl)urea (6c). 6b (0.15 mmol), 2ꢀchloropyrimidine (0.21 mmol) and iꢀPrN2Et (0.6 mmol)
were mixed in DMF (0.6 mL) in a microwave reaction tube. The mixture was heated to 120 °C
and stirred for 15 min. After cooling to room temperature, the mixture was diluted with DMF.
The resulting mixture was purified by reverseꢀphase HPLC to give 6c. LC/MS: 3.80 min,
1
M+=447.3, MW=446.5 H NMR (500 MHz, DMSO) δ 8.71 (s, 1H), 8.47 (d, J = 4.8, 2H), 8.21
(d, J = 8.8, 2H), 7.51 (d, J = 8.8, 2H), 6.77 (t, J = 2.5, 1H), 6.18 (m, 1H), 4.99 (s, 2H), 4.68 (s,
2H), 3.78 (m, 8H), 3.20 – 3.15 (m, 2H), 1.26 (t, J = 6.9, 3H); LCꢀMS: m/z = + 447 (M+H)+.
1ꢀethylꢀ3ꢀ(4ꢀ(6ꢀ(methylsulfonyl)ꢀ4ꢀmorpholinoꢀ6,7ꢀdihydroꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidinꢀ
1
2ꢀyl)phenyl)urea (6d). LC/MS: 3.59 min, M+=447.2, MW=446.5 H NMR (400 MHz, DMSO)
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 47
1
2
3
4
5
6
δ 8.69 (s, 1H), 8.18 (d, J = 8.7, 2H), 7.49 (d, J = 8.8, 2H), 6.19 (br s, 1H), 4.84 (s, 2H), 4.51 (s,
2H), 4.04 – 3.76 (m, 8H), 3.20 – 3.08 (m, 2H), 3.06 (s, 3H), 1.06 (t, J = 7.2, 3H);
7
8
9
ethyl
2ꢀ(4ꢀ(3ꢀethylureido)phenyl)ꢀ4ꢀmorpholinoꢀ5Hꢀpyrrolo[3,4ꢀd]pyrimidineꢀ6(7H)ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
carboxylate (6e). LC/MS: 3.53 min, M+=440.3, MW=439.5 H NMR (500 MHz, DMSO) δ
8.73 (s, 1H), 8.18 (d, J = 6.7, 2H), 7.49 (d, J = 8.8, 2H), 6.26 – 6.15 (m, 1H), 4.82 (d, J = 13.9,
2H), 4.50 (d, J = 15.2, 2H), 4.21 – 4.07 (m, 2H), 3.80 – 3.69 (m, 8H), 3.19 – 3.06 (m, 2H), 1.25
(t, J = 7.0, 3H), 1.06 (t, J = 7.2, 3H).
6ꢀbenzylꢀ5,6,7,8ꢀtetrahydropyrido[4,3ꢀd]pyrimidineꢀ2,4(1H,3H)ꢀdione. Methyl 1ꢀbenzylꢀ4ꢀ
oxoꢀ3ꢀpiperidinecarboxylate (20.194 g, 71.168 mmol) and urea (9.031 g, 150.4 mmol) were
dissolved in methanol (150 mL, 3700 mmol). 46 mL of a 1M solution of sodium methoxide in
methanol was added dropwise, and then the reaction was heated to reflux under N2 for 96h. The
reaction was cooled to 0 °C and filtered to give a white solid. This was stirred vigorously for 30
minutes with 50 mL water, then cooled to 0 °C and filtered to give 6ꢀbenzylꢀ5,6,7,8ꢀ
tetrahydropyrido[4,3ꢀd]pyrimidineꢀ2,4(1H,3H)ꢀdione as a white solid (11.77g, 45.7 mmol) which
was dried under high vacuum overnight and then used without further purification. 1H NMR (D6ꢀ
DMSO, 400 MHz) δ 7.47 – 7.07 (m, 5H), 3.56 (s, 2H), 2.96 (s, 2H), 2.51 (t, 4H), 2.27 (t, J = 5.6,
2H).
6ꢀbenzylꢀ2,4ꢀdichloroꢀ5,6,7,8ꢀtetrahydropyrido[4,3ꢀd]pyrimidine (7). 6ꢀbenzylꢀ5,6,7,8ꢀ
tetrahydropyrido[4,3ꢀd]pyrimidineꢀ2,4(1H,3H)ꢀdione (16.95 g, 65.88 mmol) was added to
phosphoryl chloride (100 mL, 1070 mmol) in a 500 mL roundbottom flask equipped with a
stirbar and the solution was refluxed 3h under N2. The volatiles were removed under reduced
pressure, and the concentrate was poured over 250 mL ice. 3M NaOH was added to a final pH of
28
ACS Paragon Plus Environment

Page 29 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10, and then the aqueous phase was extracted with CH2Cl2 (3x150 mL). The combined organics
were dried over MgSO4, filtered and concentrated to a tan oil. This crude material was dissolved
in dichloromethane and concentrated onto silica gel and subjected to column chromatography
using a 120g column, with a gradient of 0% to 50% ethyl acetate in hexanes. The product
containing fractions were combined and evaporated under reduced pressure to give 7 as a pale
solid (14.63g, 49.7 mmol). 1H NMR (CDCl3, 400 MHz) δ 7.41 – 7.27 (m, 5H), 3.77 (s, 2H), 3.62
(s, 2H), 2.99 (t, J = 5.8, 2H), 2.81 (t, J = 5.8, 2H).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
tertꢀbutyl
2,4ꢀdichloroꢀ7,8ꢀdihydropyrido[4,3ꢀd]pyrimidineꢀ6(5H)ꢀcarboxylate
(8).
Compound 7 (1.84 g, 6.25 mmol) was dissolved in methylene chloride (40 mL, 700 mmol) in a
roundbottom flask equipped with a stirbar and the reaction was cooled to 0 °C. αꢀchloroethyl
chloroformate (0.810 mL, 7.50 mmol) was added slowly and the reaction was stirred at 0 °C for
15 minutes. The reaction was allowed to warm to RT, and then was refluxed for 1h. The volatiles
were removed under reduced pressure and the residue was redissolved in 20 mL MeOH and
heated to reflux for 30 min. The volatiles were removed under reduced pressure to give the
intermediate amine hydrochloride salt. This material was dissolved in 50 mL dry CH2Cl2, to
which was added 3.1 g tetraalkylammonium carbonate polymerꢀbound (2.5ꢀ3.5 mmol N/g) resin
and diꢀtertꢀbutyldicarbonate (2.45 g, 11.2 mmol). The mixture was stirred at RT for 1h and was
then filtered to remove the resin. The crude mixture was concentrated onto silica gel and
subjected to column chromatography using a 12g column, eluting with a gradient of 0ꢀ30%
EtOAc in hexanes. The product containing fractions were combined and evaporated under
reduced pressure to give 8 as a clear oil which slowly crystallized (1.68g, 5.52 mmol). 1H NMR
(CDCl3, 400 MHz) δ 4.56 (s, 2H), 3.75 (t, J = 5.9, 2H), 2.97 (t, J = 5.8, 2H), 1.50 (s, 9H).
29
ACS Paragon Plus Environment