Discovery, Optimization, and Evaluation of Potent and Highly Selective PI3Kγ-PI3KδDual Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.8b02014

An electronic density model was developed and used to identify a novel pyrrolotriazinone replacement for a quinazolinone, a commonly used moiety to impart selectivity in inhibitors for PI3Kγand PI3Kδ. Guided by molecular docking, this new specificity piece was then linked to the hinge-binding region of the inhibitor using a novel cyclic moiety. Further structure-activity relationship optimization around the hinge region led to the discovery of candidate 26, a highly potent and selective PI3Kγ-PI3Kδdual inhibitor with favorable drug metabolism and pharmacokinetic properties in preclinical species.

Full text of DOI:10.1021/acs.jmedchem.8b02014

Subscriber access provided by Stockholm University Library
Article
Discovery, Optimization and Evaluation of Potent
and Highly Selective PI3K#-PI3K# Dual Inhibitors
Hong Jia, Guangxiu Dai, Weiguo Su, Kun Xiao, Jianyang Weng, Zhulin Zhang, Qing Wang,
Tianhai Yuan, Fuying Shi, Zheng Zhang, Wei Chen, Yang Sai, Jian Wang, Xiong Li, Yu Cai, Jun
Yu, Ping Ren, Jennifer Venable, Tadimeti Rao, James Patrick Edwards, and Scott D. Bembenek
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b02014 • Publication Date (Web): 29 Apr 2019
Downloaded from http://pubs.acs.org on April 30, 2019
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 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 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.
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 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery, Optimization and Evaluation of Potent
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
and Highly Selective PI3Kγ−PI3Kδ Dual Inhibitors
Hong Jiaǂ, Guangxiu Daiǂ*, Weiguo Suǂ, Kun Xiaoǂ, Jianyang Wengǂ, Zhulin Zhangǂ, Qing Wangǂ,
Tianhai Yuanǂ, Fuying Shiǂ, Zheng Zhangǂ, Wei Chenǂ, Yang Saiǂ, Jian Wangǂ, Xiong Liǂ, Yu Caiǂ,
Jun Yuǂ, Ping Renǂ, Jennifer Venable, Tadimeti Rao, James P. Edwards, Scott D. Bembenek.
ǂHutchison MediPharma Limited, Building 4, 720 Cai Lun Road, Zhangjiang Hi-Tech Park,
Shanghai 201203, China;
Janssen Pharmaceuticals Research & Development, 3210 Merryfield Row, San Diego, CA
92121, United States
##Janssen Pharmaceutical Research & Development, 1400 McKean Road, Spring House, PA
19477, United States
*Correspondence author: guangxiud@hmplglobal.com
KEYWORDS: PI3Kγ, PI3Kδ, Phosphoinositide 3-kinases, kinase inhibitors, structure-based
drug design, SBDD, RA, rheumatoid arthritis.
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 45
1
2
3
4
5
6
7
8
9
ABSTRACT: An electronic density model was developed and used to identify a novel
pyrrolotriazinone replacement for a quinazolinone, a commonly used moiety to impart selectivity
in inhibitors for PI3K and PI3K. Guided by molecular docking, this new specificity piece was
then linked to the hinge–binding region of the inhibitor using a novel cyclic moiety. Further SAR
optimization around the hinge region led to the discovery of candidate 26, a highly potent and
selective PI3Kγ−PI3Kδ dual inhibitor with favorable DMPK properties in preclinical species.
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
Phosphoinositide 3-kinases (PI3-kinases or PI3Ks) are a family of enzymes involved in
cellular functions such as cell growth, proliferation, differentiation, motility, survival and
intracellular trafficking. Based on the sequence homology and lipid substrate specificity, the
PI3K family is divided into class I, class II, and class III. The class I PI3Ks are the most
extensively studied and the focus of numerous drug discovery efforts. All class I PI3Ks are
heterodimeric proteins. Each contains a small regulatory domain and a large 110 kDa catalytic
domain which occur in four isoforms differentiated as p110α, p110β, p110γ and p110δ.1-3
All four class I catalytic PI3K isoforms show a characteristic expression pattern in vivo.
PI3Kα and PI3Kβ are ubiquitously expressed and play key roles in cell growth, survival, and
proliferation; inhibition of PI3Kα and PI3Kβ mainly targets cancer therapy.4-5 PI3Kγ is widely
expressed in granulocytes, monocytes, and macrophages, while the PI3Kδ isoform is also found
in B and T cells. Mice containing knockouts for the genes encoding either PI3Kδ or PI3Kγ are
viable and fertile but exhibit significant defects in innate immunity as well as adaptive responses.
6-8
Thus isoform-specific PI3Kδ or PI3Kγ inhibitors may have therapeutic benefits on
autoimmune diseases without interfering with the general PI3K signaling critical to the normal
functions of other cellular systems.
ACS Paragon Plus Environment
2

Page 3 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Rheumatoid arthritis (RA) is an immune-mediated disease causing debilitating joint
inflammation and tissue damage. The activities of T cells, B cells, mast cells, macrophages, and
neutrophils have all been shown to affect at least one or several aspects of the disease.
Considering their functions in these cells, their activities and the abundant expression in RA
synovium and fibroblast-like synoviocytes (FLSs), PI3Kγ and PI3Kδ represent possible
therapeutic targets for rheumatoid arthritis.6, 9-11
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
Several PI3Kδ inhibitors have entered clinical development. Idelalisib, also known as GS‑
1101 or CAL‑101 and branded Zydelig™, (1 in Figure 1), is the first FDA-approved PI3Kδ
inhibitor to treat follicular lymphoma (FL), and small lymphocytic lymphoma (SLL) as a single
agent, and to treat relapsed chronic lymphocytic leukemia (CLL) in combination with anti-CD20
Rituximab.12 Umbralisib (TGR-1202) (2 in Figure 1) is a small molecule PI3Kδ inhibitor
developed by TG Therapeutics and is currently in Phase 3 clinical trials in combination with
anti-CD20 ublituximab (TG-1101) for patients with hematologic malignancies.13 In recent years,
more selective PI3Kδ inhibitors have been advanced into clinical trials.14
Discovery of PI3Kγ specific inhibitors have faced the challenge of isoform selectivity
due to high homology between PI3Kγ and PI3Kα. Infinity Pharmaceuticals reported a highly
potent and selective PI3Kγ inhibitor IPI-549 (3 in Figure 1), which is being evaluated in a Phase
1 study as a monotherapy and in combination with nivolumab in patients with advanced solid
tumors.15
Although inhibition of either PI3Kγ or PI3Kδ alone presents a unique therapeutic
opportunity in autoimmune and oncology diseases, simultaneous inhibition of both these
isoforms may yield superior clinical efficacy in the treatment of a variety of complex immune-
mediated inflammatory diseases.16-17 Both enzymes are known to help support the growth and
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 45
1
2
3
4
5
6
7
8
9
survival of malignant B-cells and T-cells. They play a role in the formation and maintenance of
the supportive tumor microenvironment.Targeting both PI3Kγ and PI3Kδ for oncology
indications, such as chronic lymphocytic leukemia, acute myeloid leukemia, non-Hodgkin’s
lymphoma, multiple myeloma also has been studied.18-19
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
Duvelisib (4 in Figure 1, IPI-145) and branded COPIKTRA™, the first reported dual
inhibitor of PI3Kγ and PI3Kδ by Infinity Pharmaceuticals, demonstrated that the combined
inhibition of adaptive and innate immune function can be achieved through dual blockade,
leading to significant therapeutic advantage in multiple inflammatory and autoimmune diseases
in animal models.20 However, duvelisib failed to demonstrate a positive clinical response in RA
patients in a double-blind, placebo-controlled Phase 2 study with background methotrexate.21
Based on our own PKPD modeling, we speculated that neither clinical dose (1 mg and 5 mg,
both twice daily) would provide sufficient drug exposure to suppress both targets, (especially
PI3Kγ), to produce a clinical response. In 2018, duvelisib was approved for the treatment of
patients with relapsed or refractory chronic lymphocytic leukemia/small lymphocytic lymphoma
and follicular lymphoma. Rhizen Pharmaceuticals S.A. have also completed a Phase 1 clinical
study with the PI3Kγ−PI3Kδ dual inhibitor tenalisib (RP-6530) (5 in Figure 1) to treat patients
with relapsed or refractory hematologic malignancies, including T-cell Lymphoma. 22-23
Figure 1. Structures of PI3Kδ and PI3Kγ inhibitors
ACS Paragon Plus Environment
4

Page 5 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Here, we describe our efforts toward identifying a novel series of potent and highly selective
PI3Kγ−PI3Kδ dual inhibitors, which led to the discovery of the efficacious and orally
bioavailable 26 as a drug candidate for treating RA and other autoimmune diseases.
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
RESULTS AND DISCUSSIONS
Molecular docking was used to identify the key interactions of 1 in the ATP binding pocket
of both PI3Kγ and PI3Kδ. Figure 2 show these interactions for 1 docked in PI3K. Here we see
N3 and N9 of the purine hinge-binding region of the molecule form hydrogen bonds with the
backbones residues V882 and E880 respectively (In PI3K, the corresponding residues are V828
and E826). The quinazolinone motif, or specificity piece, makes a hydrophobic interaction with
M804 (M752 in PI3K) and an edge-on- interaction with W812 (W760 in PI3K). The
interaction with W812 is hypothesized to be the stronger of the two interactions, as it is a specific
interaction involving the proximal proton of W812 directly interacting with the system of the
quinazolinone (seen in Figure 2). We note that this proposed docking mode was in excellent
agreement with the later reported X-ray crystal structure of 1 bound in the PI3Kδ (4XE0 in PDB)
12. Moreover, based on our modeling and known literature crystal structures, compounds 2, 4 and
5 are expected to have binding poses similar to 1 in both PI3Kγ and PI3Kδ.
The aforementioned interactions of 1 and overall low-strain fit into the binding pocket
mostly ensure its potency in both PI3Kγ and PI3Kδ. 25 Further, ligands engaging the P-Loop with
a specificity piece tend to have excellent overall kinome selectivity while showing a varying
degree of activity over the PI3K isoforms.13 Given the variations in conformational plasticity of
the P-Loop (note that the similarity in the P-Loop among the PI3K isoforms ranges between 65–
68%) and that distinction in its overall interaction with the specificity piece of a ligand exists, 24
we expected that this region could provide an opportunity to achieve PI3K isoform selectivity. 25
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 45
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. Proposed docking pose of 1 in PI3K. The hinge region makes hydrogen bond
interactions with the backbones of V882 and E880. The specificity piece (shown in the box),
makes a hydrophobic interaction with M804, and an edge-on- interaction with W812, which is
the stronger of the two interactions. 2CHW in PDB was used for the modeling.
The first goal of the program was to identify a replacement for the quinazolinone in 1 that
would present more favorable DMPK properties. Ideally this novel specificity piece would
achieve both similar fit and interactions (especially with W812) in the P-Loop region.
As the first step in identifying viable replacement specificity piece motifs, a virtual screen
was performed on that portion of the molecule using the bioisostere replacement tool BROOD.26
We found the results from BROOD to be of more qualitative, rather than quantitative, value. In
other words, while BROOD provides a viable means of lead hopping and brainstorming, we have
found the actual score to be of little value in prioritizing molecules for synthesis.
Seeking a scoring tool that would be more useful for prioritization of medicinal chemistry
efforts, motifs of interest from the BROOD virtual screen were subsequently run through the
ACS Paragon Plus Environment
6

Page 7 of 45
Journal of Medicinal Chemistry
1
2
3
27
electrostatic similarity tool EON, whereby we obtained an electrostatic surface representation
for each novel motif along with the original reference motif (quinazolinone) of 1. We then
hypothesized the major interaction of the specificity piece was on the side interacting with
W812. Finally, through visual inspection of the electrostatic surfaces (Figure 3), we sought out
those new motifs that were comparable to 1 on that side. (In the supporting information, we
showed examples of poor affinity specificity pieces and their electrostatic surfaces. See Figures
S1–S4.) Although EON does provide an overall score for each piece against the chosen reference
(e.g., 1), we found it to be in poor agreement with our SAR. We therefore pursued the described
visual inspection approach as a feasible alternative.
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. The electrostatic surfaces of the (parent) specificity pieces for compound 1
(quinazolinone) and the (electronically similar) novel pyrrolotriazinone motif identified from
modeling.
In this way, several novel replacement motifs for the specificity piece were identified.
However, it was the pyrrolotriazinone that showed the most electronic similarity (according to
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 45
1
2
3
4
5
6
7
8
9
the described paradigm) to the quinazolinone motif of 1, and therefore was chosen for successive
chemistry. We describe our lead optimization of the pyrrolotriazinone series of PI3K/
inhibitors in detail below.
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 docking mode of 1 in PI3Kγ (Figure 2) showed the open-chain-amine linker placed the
specificity piece and the hinge perpendicular to each other resulting in a propeller-like shape.
Given the overall orientation of the molecule along with its interactions within the binding
pocket, we realized that cyclization of the linker into a 5-membered pyrrolidine would be a
feasible approach for structural modification to reinforce parameters by locking the conformation
(Figure 4). This proposal was further substantiated upon docking of the de novo compound 6
(and subsequently 7) in PI3Kγ, which displayed key interactions (in the hinge and specificity
pieces regions) and overall fit in the binding pocket (Figure 5) to be comparable to 1. The
docking model also revealed that the new cyclic-amine linker would more fully occupy (than the
open-chain-amine linker) a small hydrophobic pocket formed by M804, I963 and M953.
Figure 4. Rational design of PI3Kγ and PI3Kδ inhibitors.
ACS Paragon Plus Environment
8

Page 9 of 45
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
Figure 5. Docking models of 6 (left) and 7 (right) in PI3Kγ. Hydrophobic interactions occur
within a small pocket formed by M804, I963 and M953. 2CHW in PDB was used for the
modeling.
As predicted from the docking model, 6 showed good inhibition of both PI3Kγ and PI3Kδ
with an IC50 of 0.012 μM and 0.002 μM, respectively. Compound 6 retained equivalent activity
for PI3Kδ and improved activity for PI3Kγ compared to compound 1 (1: IC50 = 0.104 μM and
0.002 μM in PI3Kγ and PI3Kδ in biochemical assays). With an IC50 of 0.004 μM and 0.001 μM
in PI3Kγ and PI3Kδ respectively, 7 demonstrates improved activity at PI3Kγ and is comparable
to 4 (IC50 = 0.002 μM and 0.001 μM in PI3Kγ and PI3Kδ respectively). These data helped verify
the docking model as well as confirmed the success of replacing quinazolinone with
pyrrolotriazinone and cyclic linker. However, both 6 and 7 both displayed poor metabolic
stability in rat liver microsomes and further improvement in isoform selectivity over PI3Kβ was
still needed. Initial SAR investigations were primarily focused on improving the selectivity
profile of 6 and 7 while monitoring the impact on metabolic stability in rat and human liver
microsome assays. Modifications were evaluated at key locations on the chemotype: C-5 of the
pyrrolotriazinone (R1), C-3 of the pyrrolidine (R2), and the hinge-binding region of the molecule
(W). As shown in Table 1, when C5-Cl of 6 was changed to F or H, the resulting analogs 8 and 9
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 45
1
2
3
4
5
6
7
8
9
respectively demonstrated 5.6-fold and 9-fold lower activities against PI3Kγ. (In the supporting
information, we compared the electrostatic surfaces of all three. See Figures S5–S6.).
Introduction of a F in either the R2 or R3 positions (10 and 11) increased the isoform selectivity
over PI3Kβ (IC50 ratio of PI3Kβ/γ is 26 for 10; ~ 90 for 11), while retaining potent IC50 values
respectively (PI3Kγ: 0.016 μM, 0.011 μM; PI3Kδ: 0.003 μM). Removal of Cl at R1 in 7 gave the
unsubstituted compound 12, which was 50-fold more selective over PI3Kβ comparatively with
only a slight loss in potency against PI3Kγ, although significant instability in liver microsomal
assays was noted. Finally, the m-F phenyl compound 13 displayed comparable on-target
activities and isoform selectivity as 7.
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 (R)-enantiomers proved to be much less active. The (R)-enantiomer of 9 showed only
14.4% inhibition of PI3Kγ and 17.7% inhibition of PI3Kδ at 1 µM. Similarly the corresponding
(R) isomer of 12 inhibited PI3Kγ and PI3Kδ by 21.2% and 7.7% at 1 µM, respectively. Changes
to the hinge W were evaluated next. Altering the C-3 nitrile of the pyrimidinyl indole 7 to methyl
ketone in compound 14 resulted in the improved selectivity of PI3Kγ and PI3Kδ over PI3Kβ by
3 fold with similar on-target activities as 7. Combination of modifications to R1 and R2 with this
new hinge resulted in compounds 15-17. The addition of a F to the pendant phenyl or the
pyrrolidine linker, and exchanging Cl with F, maintained equivalent potencies and similar
selectivity. We also evaluated a 2-amino pyrimidine-5-carbonitrile hinge motif as in 18. Potency
and selectivity of this molecule was comparable to 14 (IC50 = 0.007 μM, 0.001 and 0.298
respectively for PI3Kγ, PI3Kδ and PI3Kβ). In terms of PI3Kα, all compounds showed less than
50% inhibition at 1 µM, except 4 (81%).
As in the enzymatic assays, compounds 10-12 and 16-18 were consistently potent in C5a-
induced and anti-IgM induced phosphorylation of Akt kinase, with single-digit nanomolar IC50’s
ACS Paragon Plus Environment
10

Page 11 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
in PI3Kδ and <0.050 μM in PI3Kγ. Unfortunately, these compounds displayed poor to moderate
stabilities in either human or rat liver microsomes after incubation for 30 min (Table 1), and
were predicted to be rapidly cleared in rodents.
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
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 45
1
2
3
4
Table 1. SAR of pyrrolidine-linker compounds.a
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
Microsomal
stability, remaining
(%)
PI3K cell IC50
(µM)
PI3K enzyme IC50 (µM)c
Cpd
R1
R2
R3
W
γ
δ
β
γ
δ
RLM
HLM
1 (CAL-101)
4 (IPI-145)
0.104
0.002
0.293
0.337
0.004
--
--
0.002
0.012
0.001
0.002
0.008
0.091
0.025
--
0.001
--
--
--
6
7
Cl
Cl
F
H
H
H
H
F
H
H
H
H
H
F
38.8
72.6
0.004
0.067
0.108
0.016
0.011
0.001
0.003
0.010
0.003
0.003
0.034
52.9% b
17.4%b
0.420
--
--
--
--
38.7
51.8
44.3
43.8
34.6
--
--
8
9
H
--
--
--
10
11
Cl
Cl
0.020
0.008
0.001
0.003
--
H
54.8% b
64.3
ACS Paragon Plus Environment
12

Page 13 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
12
13
H
H
F
H
H
0.019
0.003
0.002
0.001
47.7% b
0.025
0.040
--
0.003
--
19.1
--
2.8
--
Cl
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
14
15
16
Cl
Cl
Cl
H
F
H
H
F
0.002
0.013
0.006
0.0005
0.003
0.002
0.051
0.170
0.191
--
--
--
41.4
60.4
--
17.9
20.1
11.7
--
H
0.007
0.042
0.003
17
18
F
H
H
H
H
0.004
0.007
0.002
0.001
0.165
0.298
0.005
0.003
26.1
23.5
25.1
--
Cl
0.022
a
b
c
--: Not tested; : IC50 or percentage of inhibition is based on a single run of experiments except 4 (n=3); : Percentage of inhibition at 1 μM; : all compounds
showed less than 50% inhibition at 1 µM, except 4 (81%), in the PI3Kα enzymatic assay.
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 45
1
2
3
4
5
6
7
8
9
Further PK & metabolite ID studies showed that the oxidation of the pyrrolidine linker was
one of the major metabolism pathways of these compounds (Figure 6). To increase the
metabolic stability by mitigating the oxidation of the hydrophobic pyrrolidine, a 4-membered
ring azetidine linker was introduced, which reduced clogP and maintained the propeller
conformation essential for potency and selectivity. As expected, the metabolic stability of the
azetidine compounds was improved compared to the corresponding pyrrolidine compounds. For
example, percent remaining of the azetidines 20, 21, and 22 in rat microsomes was 63.9%,
77.2% and 77.5% respectively, and is much improved when compared to the corresponding
pyrrolidine analogs 7, 14 and 18. This trend towards increased stability is also evident in the
human liver microsomes (Table 2).
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． In vitro metabolite identification of 12 in human and mouse liver microsomes.
ACS Paragon Plus Environment
14

Page 15 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Surprisingly, while stability was improved in the azetidine analogues, both enzymatic and
cellular activities decreased significantly. The enzymatic inhibition of 19 was much weaker
against PI3Kγ and PI3Kδ (IC50 = 0.271 μM and 0.084 μM respectively) than the corresponding
pyrrolidine 6. Improved selectivity over PI3Kβ was seen for 20 when compared to 7 (PI3Kβ/γ =
28, PI3Kβ/δ= 73), but 8-fold less potent at PI3Kγ and 13-fold weaker at PI3Kδ with an IC50 of
0.034 μM and 0.013 μM respectively. Also, the inhibition of Akt phosphorylation was
significantly impacted (IC50 = 0.152 μM and 0.021 μM, respectively). Consistent with the
pyrrolidine analogue 14, 21 was still active in the enzymatic assays (IC50 = 0.005 μM for PI3Kγ
and 0.006 μM for PI3Kδ), and it showed much improved selectivity over PI3Kβ (~ 200 fold).
Although 21 was still active and selective in the enzymatic assays, it exhibited reduced potency
in C5a-induced phosphorylation of Akt with an IC50 of 0.142 μM. While the 2-aminopyrimidine
hinge binding motif in 22 displayed diminished activity, it was still potent at PI3Kγ (IC50 = 0.027
μM) and PI3Kδ (IC50 = 0.012 μM) and selective (37.0% inhibition against PI3Kβ and < 50%
against PI3Kα at 1 μM in enzymatic assay) (Table 2).
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 2. SAR of azetidine-linker compounds a
PI3K enzyme IC50 (µM)
PI3K cell IC50 (µM)
Stability (%)
Cpd
W
γ
δ
β
γ
δ
RLM HLM
α
19
0.271
0.084
18.5% b -1.5% b
--
--
78.5
63.9
95.4
89.8
0.034±
0.009
(n=3)
0.013±
0.001 (n=2)
0.152±0.090
(n=5)
0.021±0.001
(n=2)
20
0.945
-0.3% b
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 45
1
2
3
4
5
6
7
8
9
0.142±0.056
(n=2)
0.021±0.013
(n=2)
21
22
0.005
0.027
0.006
0.012
45.9% b
8.4% b
77.2
77.5
83
37.0% b <50% b
0.241
0.016
71.2
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
b
--: Not tested; a. IC50 or percentage of inhibition is based on a single run of experiments; : Percentage of inhibition
at 1 μM
The balanced profile of compound 22 along with the ability to readily modulate the C-5
substituent led us to explore the SAR of the 2-aminopyrimidine as an effective hinge motif. As
shown in Table 3, 23 demonstrated reduced biochemical potency than 22 towards PI3Kγ and
PI3Kδ with an IC50 of 0.182 μM and 0.059 μM, respectively. Replacing the nitrile of compound
22 with an alkyne, resulted in 24, which maintained isoform PI3K activities and selectivity in
both enzymatic and cellular assays; however, the rat liver microsomal stability was only 8.9%.
Changing nitrile of compound 22 to a methyl sulfone gave 25, which had a complete loss of
activities. The introduction of a methyl ketone into the C-5 position of the pyrimidine yielded 26,
which was both active and selective at PI3Kγ and PI3Kδ (IC50 at PI3Kα = 3.7 μM in enzyme and
>10 μM in cell), as well as very potent on the inhibition of C5a and anti-IgM induced Akt
phosphorylation (IC50 = 0.011 ±0.005 μM for PI3Kγ and 0.025 ±0.009 μM for PI3Kδ). In vitro
ADME screening demonstrated that 26 was more stable in RLM and HLM (63.5% and 81.5%).
Further modification of 26 to give the cyclopropyl (27) and CF3 (28) did not produce the desired
potency, indicating larger substitutes may not be well tolerated in this region.
Table 3. SAR of pyrimidine-hinge compounds.a
ACS Paragon Plus Environment
16

Page 17 of 45
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
PI3K enzyme IC50 (µM)
Cpd
R
γ
δ
β
23
24
Cl
0.182
0.059
0.060
22.9% b
0.052
20.4% b
22.7% b
--
25
26
18.1% b
0.004 ±0.002
0.005 ±0.003
0.590 ±0.254
27
28
50.7% b
7.1% b
--
0.043
0.524
>1
a
b
--: Not tested; : IC50 and percentage of inhibition is based on a single run of experiments except 26 (n=3); :
percentage of inhibition at 1 μM.
Compound 26 demonstrated high specificity towards the PI3K family in a panel of 458
kinases (395 non-mutated kinases and 63 mutated human kinases) at a test concentration of 10
μM. Furthermore, 26 showed nearly no activity against a panel of 50 GPCRs, ion channels, and
transporters.
We further selected 26 for additional in vitro cellular assays to understand the impact on the
functional activities of PI3Kδ and PI3Kγ. Assays for PI3Kδ functions include anti-IgM
stimulated CD69 expression on B cells in human whole blood; anti-IgD stimulated CD86
expression on B cells in rat whole blood; and fMLP induced CD63 expression on basophils in
human whole blood. Functional assays to evaluate PI3Kγ activity include anti-IgE induced CD63
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 45
1
2
3
4
5
6
7
8
9
expression on basophils in human whole blood. As shown in Table 4, 26 exhibited potent
inhibitory effects on basophil and B cell activation. For example, IC50 was 0.042 μM and 0.337
μM on the inhibition of basophil activation in the human whole blood, and 0.044 μM on B cell
activation in the rat whole blood.
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. Additional cellular assays of compound 26 (n ≥ 3)
HWB basophil CD63 IC50 (µM)
HWB B cell CD69 IC50
(µM)
RWB B cell CD86 IC50
(µM)
Cpd
Anti-IgE
fLMP
4
0.014±0.006
0.063±0.037
0.002±0.001
0.044±0.015
0.015±0.012
0.042±0.022
2.206±0.335
0.337±0.089
26
Experiments were performed in female Wistar rats to further evaluate the inhibitory effect of
26 on anti-IgD induced B cell activation ex vivo. Oral administration of 26 effectively inhibited
anti-IgD induced B cell activation at 2h, with an inhibition of 88.3% at a dose of 1 mg/kg.
Instillation of KC/GRO (also called CXCL1, Gro-α and neutrophil-activating protein 3 (NAP-3)
ect, a CXC chemokine) into the rat subcutaneous air pouch induced robust neutrophil chemotaxis
7. In an air pouch study in rats, 26 blocked
KC/GRO induced neutrophil migration at 4 h, with an inhibition of 87.9% at a dose of 10 mg/kg,
confirming the potent inhibitory activity of 26 on PI3Kγ in vivo.
When administered in vivo for pharmacokinetic profiling, 26 demonstrated moderate
clearance in rats and low clearance in dogs, moderate volumes of distribution, and high oral
bioavailability (Table 5). No glutathione (GSH) adducts were identified in human liver
microsomes when GSH was co- incubated as a cofactor.
Table 5. PK profiles of compound 26 in SD rats and in Beagle dogs.
ACS Paragon Plus Environment
18

Page 19 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
PO
Cmax
IV
T1/2
Species
Dose
(mg/kg)
F
(%)
AUC
(h∙ng/mL)
Dose
Cl
Vss
AUC
(ng/mL)
(mg/kg) (mL/min/kg) (h) (L/kg) (h∙ng/mL)
Rata
10
5
107.9
82.4
2730
14342
13125
2.5
1
12.7
5.4
1.5
6.3
1.2
1.7
3274
3261
Dogb
2532
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
b
a: 20% HP-beta-CD, pH2.5; : 20% HP-beta-CD, pH4.3. Values were the mean values of male for rats, male and
female for dogs.
We dosed 26 in an established rat collagen-induced arthritis model at 0.1, 1 and 10 mg/kg
twice daily for 7 days from day 10-16. It significantly reduced the average paw volumes in a
dose-dependent manner, indicating significant anti-inflammatory effects; the estimated ED50 was
0.25 mg/kg, BID (Figure 7). Given at 1.0 mg/kg BID, 26 produced a paw volume reduction
comparable to the positive control YiSaiPu (a rhTNFR: Fc) from day 11 to day 16; and a 10
mg/kg BID administration of 26 demonstrated strong efficacy from day 11 and reversed the
disease progression nearly returning the animal back to normal on day 16.
2 .4
2 .2
# #
V e h ic le (2 0 % H P C D , p H 2 .5 ) P O B ID
C o m p o u n d 2 6 0 .1 M P K P O B ID
2 .0
1 .8
1 .6
1 .4
1 .2
1 .0
*
C o m p o u n d 2 6 1 .0 M P K P O B ID
Y iS a iP u 1 0 M P K IP Q O D
**
**
C o m p o u n d 2 6 1 0 M P K P O B ID
N a iv e
**
1 0
1 1
1 2
1 3
1 4
1 5
1 6
D ays post im m u nization
Figure 7. Hind paw swelling reduction of compound 26 in rat CIA model. The paw volume was
expressed as mean ± SEM. n= 6 in Naive groups. n=9 in the other groups. *p<0.05, **p<0.01 vs
vehicle group, #p<0.05, ##p<0.01 vs naive group. 26 was dosed twice daily for 7 days.
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 45
1
2
3
4
CHEMISTRY
5
6
7
8
Compounds 30a-e were prepared starting with a coupling reaction between amide 29a-c and
various substituted proline or azetidine carboxylic acids, followed by intramolecular dehydration
to produce 31a-e.28 The desired compounds were generated by two routes, depending on the
nature of the R2 substituent. When R2 is not a hydroxyl, 31a-b and 31d-e were converted to 32a-
e.29 After NBoc de-protection , 32a-e were treated with the halides of the selected hinge motif W
in SnAr reactions to give the final compounds 6-9, 11-12, 14, 16-24 and 26, as well as 25a and
36a-b as intermediates. If R2 is hydroxyl, the compound was treated with DAST to introduce
fluorine onto the pyrrolidine. The resulting compound 33 was used to make compounds 10, 13,
and 15 in a similar fashion as 6 (Scheme 1).
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
When
W
is 4-(2-amino-5-bromo)pyrimidinyl, 25a was coupled with sodium
methanesulfinate to give 25. 36a-b could be converted to compounds 27-28 by oxidation,
followed by the displacement of sulfonyl with NH2 (Scheme 2).
Scheme 1. Synthesis of compounds 6-26 and intermediates 25a, 36a-b a
ACS Paragon Plus Environment
20

Page 21 of 45
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
o
Reagents and conditions: (i) EDC∙HCl, THF, rt; (ii) KOH, EtOH, H2O, 100 C; (iii) substituted or unsubstituted
phenylboronic acid, 4Å molecular sieves, Cu(OAc)2, pyridine, DCM, rt; (iv) aq. HCl or conc. HCl, MeOH; (v) n-
o
BuOH, Et3N, reflux; (vi) DAST, DCM, 0 C to rt; (viii) sodium methanesulfinate, Pd2(dba)3, Xantphos, K3PO4,
120oC, N2.
Scheme 2. Synthesis of compounds 27-28 a
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 45
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
Reagents and conditions: (i) m-CPBA, DCM, 0 oC to rt.; (ii) DCM, NH3, rt.
CONCLUSION
In conclusion a combination of an electronic density model approach and molecular docking
led to design of a novel series of potent and isoform-selective PI3Kγ – PI3Kδ dual inhibitors
characterized by a substituted pyrrolotriazinone specificity piece connected to the hinge-binding
region of the molecule through a cyclic linker. After optimization of the hinge and linker, the
azetidine 26 was identified as a promising preclinical candidate of PI3Kγ and PI3Kδ inhibitors
with potent efficacy in a rat CIA model and desirable pharmacokinetic properties.
EXPERIMENTAL SECTION
All reagents and solvents employed were purchased commercially and used without further
purification. Temperature and pressure is at or near atmospheric. NMR data were recorded on a
Varian 400-MR machine. Chemical shifts are expressed as δ units using tetramethylsilane as the
external standard (in NMR description, s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet;
and br, broad peak). All coupling constants (J) are reported in Hertz. MS data were checked by
Agilent 6120 or Agilent 1100. Mass spectra were measured with an Agilent quadrupole 6110
spectrometer using an ESI+APCI source coupled to an Agilent 1200 HPLC system or an 1100
series LC/MSD trap spectrometer using an ESI source coupled to an Agilent 1100 HPLC system.
The IC50 values were calculated using XL-Fit 2.0 software. Compound purity was determined by
ACS Paragon Plus Environment
22

Page 23 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
HPLC , which confirmed the purity of ≥ 95% for all final biological testing compounds. HPLC
method: Aglilent 1200 HPLC, Column: SB-C18 5u 150 x 4.6 mm; Mobile phase: H2O (0.1%
formic acid) and MeOH (0.1% formic acid); Gradient: 5% MeOH in 0-10 min, 95% MeOH in
10−13 min, 5% MeOH in 13-16 min; Flow rate: 1 mL/min; Detector: 254 nm.
Preparation of (S)-2-(1-(9H-purin-6-yl)pyrrolidin-2-yl)-5-chloro-3-phenylpyrrolo[2,1-
f][1,2,4]triazin-4(3H)-one (6)
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
tert-butyl
(S)-2-((2-carbamoyl-3-chloro-1H-pyrrol-1-yl)carbamoyl)pyrrolidine-1-
carboxylate (30a) : To a solution of 1-amino-3-chloro-1H-pyrrole-2-carboxamide (700 mg, 4.40
mmol) and (S)-1-(tert-butoxycarbonyl)pyrrolidine-2-carboxylic acid (1040 mg, 4.84 mmol) in
THF was added EDC∙HCl (950 mg, 4.95 mmol). The reaction mixture was stirred at room
temperature for 3.5 h, then was diluted with water and extracted with EtOAc. The combined
organic layers were dried over anhydrous Na2SO4, and concentrated to yield the 30a as a white
solid, which was used for the next step without further purification (1.25 g, 80%). 1H NMR (400
MHz, DMSO-d6) δ 11.28 (s, 0.6H), 11.27 (s, 0.4H), 7.49 (s, 1H), 7.09 (s, 0.6H), 7.03 (s, 0.4H),
6.82 (d, J = 3.2 Hz, 0.4H), 6.71 (d, J = 2.8 Hz, 0.6H), 6.17 (d, J = 2.8 Hz, 0.6H), 6.14 (d, J = 3.2
Hz, 0.4H), 4.19 – 4.12 (m, 1H), 3.39−3.32 (m, 1H), 3.28−3.20 (m, 1H), 2.17−2.02 (m, 2H),
1.88−1.71 (m, 2H), 1.37 (s, 3H), 1.33 (s, 6H). MS (m/z): 256.9 (M-Boc+H)+.
tert-butyl (S)-2-(5-chloro-4-oxo-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-yl)pyrrolidine-1-
carboxylate (31a) : To a solution of 30a (1.25 g, 3.51 mmol) in EtOH (40 mL) was added a
solution of KOH (1.00 g, 17.56 mmol) in water (40 mL). The reaction mixture was heated to 100
⁰C for 2 days. After cooling to room temperature, the mixture was diluted with water and
adjusted to pH4.5 with aq.HCl (1.0 N). The resulting precipitate was filtered off and dried to give
31a as a white solid (850 mg, 72%).28 1H NMR (400 MHz, DMSO-d6) δ 7.54 (d, J = 2.8 Hz,
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 45
1
2
3
4
5
6
7
8
9
0.6H), 7.51 (d, J = 2.8 Hz, 0.4H), 6.54-6.51 (m, 1H), 4.49−4.46 (m, 0.4H), 4.43 - 4.39 (m, 0.6H),
3.54−3.44 (m, 1H), 3.33−3.26 (m, 1H), 2.26−2.10 (m, 1H), 1.99 – 1.71 (m, 3H), 1.34 (s, 3H),
1.13 (s, 6H). MS (m/z): 338.7 (M+H)+.
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
tert-butyl
(S)-2-(5-chloro-4-oxo-3-phenyl-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-
yl)pyrrolidine-1-carboxylate (32a): A mixture of 31a (850 mg, 2.51 mmol), phenylboronic acid
(613 mg, 5.02 mmol), molecular sieves (4Å, 2 g), Cu(OAc)2, (915 mg, 5.02 mmol), and pyridine
(1.00 mL, 12.55 mmol) in dry DCM (30 mL) was stirred overnight at room temperature under
dry air. The mixture was concentrated in vacuo and purified by flash chromatography, eluting
with 0−100% gradient of MeOH in H2O, to yield 32a as a white solid (250 mg, 24%).29 1H NMR
(400 MHz, DMSO-d6) δ 7.67 (d, J = 2.8 Hz, 0.6H), 7.65 (d, J = 2.8 Hz, 0.4H), 7.62−7.46 (m,
4H), 7.33 (d, J = 7.2 Hz, 0.4H), 7.25 (d, J = 7.6 Hz, 0.6H), 6.66 - 6.63 (m, 1H), 4.21−4.17 (m,
0.4H), 4.16−4.13 (m, 0.6H), 3.46−3.36 (m, 1H), 3.25−3.17 (m, 1H), 2.12−1.64 (m, 4H), 1.34 (s,
3H), 1.25 (s, 6H). MS (m/z): 414.7(M+H)+.
(S)-2-(1-(9H-purin-6-yl)pyrrolidin-2-yl)-5-chloro-3-phenylpyrrolo[2,1-f][1,2,4]triazin-
4(3H)-one (6): A mixture of 32a (250 mg, 0.603 mmol) in a solution of HCl in MeOH (4 N, 25
mL) was stirred at room temperature for 2 h. The mixture was concentrated in vacuo. The
residue (40 mg, 0.13 mmol) was dissolved in n-BuOH (3 mL), and then 6-chloro-9H-purine (22
mg, 0.14 mmol) and Et3N (0.10 mL, 0.65 mmol) were added. The reaction mixture was refluxed
for 2 h. The mixture was concentrated to remove n-BuOH. The resulting residue was purified by
flash chromatography, eluted with 0−100% gradient of MeOH in H2O, to yield the target
compound 6 as a white solid (35 mg, 63%). Purity: 96.0%. MS (m/z): 432.7 (M+H)+. H NMR
1
(400 MHz, DMSO-d6) δ: 12.95 (s, 1H), 8.65−8.05 (m, 3H), 7.72−7.40 (m, 5H), 6.57−6.50 (m,
ACS Paragon Plus Environment
24

Page 25 of 45
Journal of Medicinal Chemistry
1
2
3
4
1H), 5.34−5.26 (m, 0.5H), 4.67−4.59 (m, 0.5H), 4.33−4.25 (m, 0.5H), 4.11−4.03 (m, 0.5H),
5
6
7
8
3.89−3.83 (m, 0.5H), 3.62−3.58 (m, 0.5H), 2.35−2.15 (m, 2H), 1.98−1.81 (m, 2H).
Compounds 7-9 were prepared by a procedure similar to that described for the synthesis of
6.
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
(S)-4-(2-(5-chloro-4-oxo-3-phenyl-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-
yl)pyrrolidin-1-yl)-7H-pyrrolo[2,3-d]pyrimidine-5-carbonitrile (7). Purity: 99.4%. MS (m/z):
456.8 (M+H)+. 1H NMR (400 MHz, DMSO-d6) δ 12.84 (s, 1H), 8.29 (s, 1H), 8.26 (s, 1H),
7.78−7.71 (m, 1H), 7.68−7.43 (m, 5H), 6.55 (d, J = 3.0 Hz, 1H), 4.68−4.60 (m, 1H), 4.12−4.03
(m, 1H), 3.97−3.86 (m, 1H), 2.36−2.15 (m, 2H), 2.04−1.86 (m, 2H).
(S)-2-(1-(9H-purin-6-yl)pyrrolidin-2-yl)-5-fluoro-3-phenylpyrrolo[2,1-f][1,2,4]triazin-
1
4(3H)-one (8) Purity: 100%. MS (m/z): 416.8 (M+H)+. H NMR (400 MHz, CD3OD) 8.24 (s,
1H), 8.03 (s, 1H), 7.84−7.41 (m, 5H), 7.15−7.09 (m, 1H), 6.30−6.15 (m, 1H), 5.65−5.50 (m,
0.5H), 4.91−4.85 (m, 0.5H), 4.42−4.37 (m, 0.5H), 4.23−4.13 (m, 0.5H), 4.05−3.95 (m, 0.5H),
3.85−3.78 (m, 0.5H), 2.37−1.97 (m, 4H).
(S)-2-(1-(9H-purin-6-yl)pyrrolidin-2-yl)-3-phenylpyrrolo[2,1-f][1,2,4]triazin-4(3H)-one
1
(9) Purity: 96.1%. MS (m/z): 398.7 (M+H)+. H NMR (400 MHz, DMSO-d6) δ 12.94 (s, 1H),
8.28-8.01 (m, 2H), 7.79-7.32 (m, 6H), 6.97-6.75 (m, 1H), 6.55-6.36 (m, 1H), 5.43-5.25 (m,
0.5H), 4.73-4.60 (m, 0.5H), 4.40-4.25 (m, 0.5H), 4.14-4.01 (m, 0.5H), 3.94-3.81 (m, 0.5H), 3.72-
3.60 (m, 0.5H), 2.31-2.10 (m, 2H), 2.01-1.91 (m, 1H), 1.90-1.77 (m, 1H).
Preparation
of
5-chloro-2-((2S,4S)-4-fluoro-1-(9H-purin-6-yl)pyrrolidin-2-yl)-3-
phenylpyrrolo[2,1-f][1,2,4]triazin-4(3H)-one (10)
tert-butyl
(2S,4S)-2-(5-chloro-4-oxo-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-yl)-4-
fluoropyrrolidine-1-carboxylate (33): To a solution of 31c (prepared by a procedure similar to
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 45
1
2
3
4
that described for the synthesis of compound 31a using (2S,4R)-1-(tert-butoxycarbonyl)-4-
5
6
hydroxypyrrolidine-2-carboxylic
acid
as
the
material
instead
of
(S)-1-(tert-
7
8
9
butoxycarbonyl)pyrrolidine-2-carboxylic acid) (400 mg, 1.73 mmol) in DCM (50 mL) was
added DAST (726 mg, 4.52 mmol) at 0 ⁰C. The resulting mixture was stirred at 0 ⁰C for 1 h,
then allowed to be warmed up to room temperature and stirred for another 1 h. After the addition
of aq. NaHCO3 (10 mL), the mixture was extracted with DCM. The organic layers were washed
with brine, dried over anhydrous Na2SO4 and concentrated to give compound 33, which was
used in the next step without further purification.30 MS (m/z): 257.0 (M-Boc+H)+.
Compounds 10 was prepared by a procedure similar to that described for the synthesis of
compound 6 using compound 33 as the starting material instead of 31a. Purity: 95.1%. MS
(m/z): 451.1 (M+H)+. 1H NMR (400 MHz, DMSO-d6) δ: 8.38-8.10 (m, 3H), 7.71−7.52 (m, 4H),
7.46 (s, 1H), 6.59−6.49 (m, 1H), 5.39−5.29 (m, 1H), 4.88−4.34 (m, 1H), 4.24−3.93 (m, 2H),
2.31−2.17 (m, 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
Compounds 11-12 were prepared by a procedure similar to that described for the synthesis
of
6.(S)-2-(1-(9H-purin-6-yl)pyrrolidin-2-yl)-5-chloro-3-(3-fluorophenyl)pyrrolo[2,1-
1
f][1,2,4]triazin-4(3H)-one (11) Purity: 97.5%. MS (m/z): 450.8 (M+H)+. H NMR (400 MHz,
DMSO-d6) δ: 12.96 (s, 1H), 8.24−8.20 (m, 1H), 8.10 (s, 1H), 7.66−7.39 (m, 5H), 6.60−6.52 (m,
1H), 5.36−5.30 (m, 0.5H), 4.68−4.62 (m, 0.5H), 4.35−4.29 (m, 0.5H), 4.12−4.06 (m, 0.5H),
3.92−3.86 (m, 0.5H), 3.73−3.67 (m, 0.5H), 2.28−2.22 (m, 1H), 2.05−1.86 (m, 3H).
(S)-4-(2-(4-oxo-3-phenyl-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-yl)pyrrolidin-1-yl)-
1
7H-pyrrolo[2,3-d]pyrimidine-5-carbonitrile (12) Purity: 97.0%. MS (m/z): 422.6 (M+H)+. H
NMR (400 MHz, DMSO-d6) δ: 12.81 (s, 1H), 8.27−8.26 (m, 2H), 7.72−7.68 (m, 1H), 7.64−7.41
(m, 5H), 6.88 (dd, J = 4.3, 1.7 Hz, 1H), 6.47 (dd, J = 4.3, 2.7 Hz, 1H), 4.72−4.65 (m, 1H),
ACS Paragon Plus Environment
26

Page 27 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4.12−4.06 (m, 1H), 3.96−3.89 (m, 1H), 2.35−2.15 (m, 2H), 2.06−1.83 (m, 2H). 4-((2S,4S)-2-(5-
chloro-4-oxo-3-phenyl-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-yl)-4-fluoropyrrolidin-1-
yl)-7H-pyrrolo[2,3-d]pyrimidine-5-carbonitrile (13). Compound was prepared by a procedure
similar to that described for the synthesis of compound 6 using compound 33 as the starting
material instead of 31a. Purity: 95.3 %. MS (m/z): 475.1 (M+H)+. 1H NMR (400 MHz, CD3OD)
δ: 8.23 (s, 1H), 7.98 (s, 1H), 7.75 (d, J = 8.0 Hz, 1H), 7.65−7.55 (m, 3H), 7.48 (d, J = 7.6 Hz,
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
1H), 7.30 (d, J = 3.0 Hz, 1H), 6.46 (d, J = 3.0 Hz, 1H), 5.39−5.31 (m, 1H), 5.22−5.16 (m, 1H),
4.56−4.41 (m, 2H), 2.51−2.41 (m, 1H), 2.22−2.16 (m, 1H).
(S)-2-(1-(5-acetyl-7H-pyrrolo[2,3-d]pyrimidin-4-yl)pyrrolidin-2-yl)-5-chloro-3-
phenylpyrrolo[2,1-f][1,2,4]triazin-4(3H)-one (14). Compound was prepared by a procedure
1
similar to that described for the synthesis of 6. Purity: 98.9%. MS (m/z): 474.1 (M+H)+. H
NMR (400 MHz, DMSO-d6) δ: 12.29 (br, 1H), 8.16 (s, 1H), 8.03 (s, 1H), 7.80−7.35 (m, 6H),
6.57 (s, 1H), 4.47−4.44 (m, 1H), 3.81−3.64 (m, 2H), 2.48 (s, 3H), 2.11−1.93 (m, 3H), 1.65−1.56
(m, 1H).
2-((2S,4S)-1-(5-acetyl-7H-pyrrolo[2,3-d]pyrimidin-4-yl)-4-fluoropyrrolidin-2-yl)-5-
chloro-3-phenylpyrrolo[2,1-f][1,2,4]triazin-4(3H)-one (15). Compound was prepared by a
procedure similar to that described for the synthesis of compound 6 using compound 33 as the
1
starting material instead of 31a. Purity: 97.0%. MS (m/z): 492.1 (M+H)+. H NMR (400 MHz,
DMSO-d6) δ: 12.42 (br, 1H), 8.23 (s, 1H), 8.13 (s, 1H), 7.64−7.55 (m, 1H), 7.54−7.45 (m, 5H),
6.59 (d, J = 3.0 Hz, 1H), 5.24−5.02 (m, 1H), 4.74−4.63 (m, 1H), 4.19−3.97 (m, 1H), 3.92−3.83
(m, 1H), 2.51 (s, 3H), 2.44-2.21 (m, 2H).
Compounds 16-24 were prepared by a procedure similar to that described for the synthesis
of 6.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 45
1
2
3
4
5
6
7
8
9
(S)-2-(1-(5-acetyl-7H-pyrrolo[2,3-d]pyrimidin-4-yl)pyrrolidin-2-yl)-5-chloro-3-(3-
fluorophenyl)pyrrolo[2,1-f][1,2,4]triazin-4(3H)-one (16) Purity: 98.3%. MS (m/z): 492.6
(M+H)+. 1H NMR (400 MHz, DMSO- d6) δ: 12.28 (br, 1H), 8.17 (s, 1H), 8.04 (s, 1H), 7.66−7.31
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
(m, 5H), 6.58 (s, 1H), 4.59−4.38 (m, 1H), 3.94−3.62 (m, 2H), 2.48 (s, 3H), 2.15−1.89 (m, 3H),
1.67−1.64 (m, 1H).
(S)-2-(1-(5-acetyl-7H-pyrrolo[2,3-d]pyrimidin-4-yl)pyrrolidin-2-yl)-5-fluoro-
3phenylpyrrolo[2,1-f][1,2,4]triazin-4(3H)-one (17) Purity: 98.1%. MS (m/z): 458.1 (M+H)+.
1H NMR (400 MHz, CDCl3) δ 11.60 (s, 1H), 8.23 (s, 1H), 7.76 (d, J = 7.9 Hz, 1H), 7.64 (s, 1H),
7.62−7.46 (m, 3H), 7.30−7.26 (m, 1H), 7.08−7.00 (m, 1H), 6.16 (d, J = 3.1 Hz, 1H), 4.81 (t, J =
6.8 Hz, 1H), 4.04−3.92 (m, 1H), 3.88-3.72 (m, 1H), 2.59 (s, 3H), 2.18−2.09 (m, 1H), 2.08−1.93
(m, 3H).
(S)-2-amino-4-(2-(5-chloro-4-oxo-3-phenyl-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-
1
yl)pyrrolidin-1-yl)pyrimidine-5-carbonitrile (18) Purity: 98.9%. MS (m/z): 433.1 (M+H)+. H
NMR (400 MHz, DMSO-d6) δ 8.18 (s, 1H), 7.85 (d, J = 7.7 Hz, 1H), 7.60−7.46 (m, 5H), 6.97
(br, 2H), 6.59 (d, J = 3.0 Hz, 1H), 4.57−4.56 (m, 1H), 3.93 (br, 1H), 3.77−3.73 (m, 1H),
2.07−2.04 (m, 2H), 1.89 (br, 1H), 1.70−1.60 (m, 1H).
(S)-2-(1-(9H-purin-6-yl)azetidin-2-yl)-5-chloro-3-phenylpyrrolo[2,1-f][1,2,4]triazin-
1
4(3H)-one (19) Purity: 97.4%. MS (m/z): 418.7 (M+H)+. H NMR (400 MHz, DMSO-d6) δ:
12.93 (s, 1H), 8.18 (s, 1H), 8.09 (s, 1H), 7.72−7.41 (m, 6H), 6.59 (s, 1H), 5.18−5.04 (m, 1H),
4.19−4.03 (m, 2H), 2.68−2.60 (m, 1H), 2.24−2.16 (m, 1H).
(S)-4-(2-(5-chloro-4-oxo-3-phenyl-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-yl)azetidin-
1-yl)-7H-pyrrolo[2,3-d]pyrimidine-5-carbonitrile (20) Purity: 96.3%. MS (m/z): 442.8
(M+H)+. 1H NMR (400 MHz, DMSO-d6) δ: 8.24−8.23 (m, 2H), 7.70−7.41 (m, 6H), 6.61 (s, 1H),
ACS Paragon Plus Environment
28

Page 29 of 45
Journal of Medicinal Chemistry
1
2
3
4
5
6
5.13−5.05 (m, 1H), 4.38−4.28 (m, 1H), 4.15−4.09 (m, 1H), 2.66−2.58 (m, 1H), 2.10−1.98 (m,
1H).
7
8
9
(S)-2-(1-(5-acetyl-7H-pyrrolo[2,3-d]pyrimidin-4-yl)azetidin-2-yl)-5-chloro-3-
phenylpyrrolo[2,1-f][1,2,4]triazin-4(3H)-one (21). Purity: 96.3%. MS (m/z): 460.2 (M+H)+. 1H
NMR (400 MHz, DMSO-d6 ) δ: 12.41 (s, 1H), 8.21 (s, 1H), 8.14 (s, 1H), 7.67−7.52 (m, 5H),
7.49−7.43 (m, 1H), 6.66−6.62 (m, 1H), 5.05−4.95 (br, 1H), 4.33−4.23 (m, 1H), 3.78−3.72 (m,
1H), 2.49−2.44 (m, 1H), 2.40 (s, 3H), 1.89−1.79 (m, 1H).
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
(S)-2-amino-4-(2-(5-chloro-4-oxo-3-phenyl-3,4-dihydropyrrolo[2,1-f][1,2,4]triazin-2-
1
yl)azetidin-1-yl)pyrimidine-5-carbonitrile (22) Purity: 98.4%. MS (m/z): 419.1 (M+H)+. H
NMR (400 MHz, DMSO-d6) δ 8.15 (s, 1H), 7.74 (s, 1H), 7.58−7.52 (m, 4H), 7.45−7.38 (m, 1H),
7.14 (br, 2H), 6.66 (d, J = 3.0 Hz, 1H), 5.04−4.80 (m, 1H), 4.10−4.00 (m, 2H), 2.58−2.51 (m,
1H), 2.09−2.00 (m, 1H).
(S)-2-(1-(2-amino-5-chloropyrimidin-4-yl)azetidin-2-yl)-5-chloro-3-phenylpyrrolo[2,1-
1
f][1,2,4]triazin-4(3H)-one (23) Purity: 100%. MS (m/z): 428.0 (M+H)+. H NMR (400 MHz,
DMSO- d6) δ 7.71 (d, J = 2.8 Hz, 1H), 7.70 (s, 1H), 7.62-7.49 (m, 4H), 7.39−7.37 (m, 1H), 6.64
(d, J = 3.0 Hz, 1H), 6.28 (s, 2H), 4.81−4.77 (m, 1H), 4.18−4.12 (m, 1H), 4.02−3.96 (m, 1H),
2.46−2.39 (m, 1H), 2.01−1.95 (m, 1H).
(S)-2-(1-(2-amino-5-ethynylpyrimidin-4-yl)azetidin-2-yl)-5-chloro-3-phenylpyrrolo[2,1-
1
f][1,2,4]triazin-4(3H)-one (24) Purity: 98.5%. MS (m/z): 418.0 (M+H)+. H NMR (400 MHz,
CDCl3) δ 8.00 (s, 1H), 7.60-7.47 (m, 3H), 7.40 (d, J = 7.5 Hz, 1H), 7.30 (d, J = 2.9 Hz, 1H),
7.15−7.10 (m, 1H), 6.49 (d, J = 3.0 Hz, 1H), 5.13−5.03 (m, 1H), 4.85 (s, 2H), 4.39−4.34 (m, 1H),
4.16−4.07 (m, 1H), 3.14 (s, 1H), 2.38−2.18 (m, 2H).
ACS Paragon Plus Environment
29