Discovery of a Parenteral Small Molecule Coagulation Factor XIa Inhibitor Clinical Candidate (BMS-962212)

Article abstract of DOI:10.1021/acs.jmedchem.7b01171

Factor XIa (FXIa) is a blood coagulation enzyme that is involved in the amplification of thrombin generation. Mounting evidence suggests that direct inhibition of FXIa can block pathologic thrombus formation while preserving normal hemostasis. Preclinical studies using a variety of approaches to reduce FXIa activity, including direct inhibitors of FXIa, have demonstrated good antithrombotic efficacy without increasing bleeding. On the basis of this potential, we targeted our efforts at identifying potent inhibitors of FXIa with a focus on discovering an acute antithrombotic agent for use in a hospital setting. Herein we describe the discovery of a potent FXIa clinical candidate, 55 (FXIa K_i = 0.7 nM), with excellent preclinical efficacy in thrombosis models and aqueous solubility suitable for intravenous administration. BMS-962212 is a reversible, direct, and highly selective small molecule inhibitor of FXIa.

Full text of DOI:10.1021/acs.jmedchem.7b01171

Subscriber access provided by READING UNIV
Article
Discovery of a parenteral small molecule coagulation
Factor XIa inhibitor clinical candidate (BMS-962212)
Donald J. P. Pinto, Michael J. Orwat, Leon M. Smith II, Mimi L Quan, Patrick Y. S. Lam, Karen A
Rossi, Atsu Apedo, Jeffrery M. Bozarth, Yiming Wu, Joanna J Zheng, Baomin Xin, Nathalie Toussaint,
Paul Stetsko, Olafur Gudmundsson, Brad D. Maxwell, Earl J Crain, Pancras C Wong, Zhen Lou,
Timothy W. Harper, Silvi A Chacko, Joseph E Myers, Steven Sheriff, Huiping Zhang, Xiaoping
Hou, Arvind Mathur, Dietmar A Seiffert, Ruth R. Wexler, Joseph M Luettgen, and William R Ewing
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01171 • Publication Date (Web): 27 Oct 2017
Downloaded from http://pubs.acs.org on October 29, 2017
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 73
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Discovery of a parenteral small molecule coagulation
Factor XIa inhibitor clinical candidate (BMS-962212)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Donald J. P. Pinto,* Michael J. Orwat, Leon M. Smith II, Mimi L. Quan, Patrick Y. S. Lam, Karen
A. Rossi, Atsu Apedo, Jeffrey M. Bozarth, Yiming Wu, Joanna J. Zheng, Baomin Xin, Nathalie
Toussaint, Paul Stetsko, Olafur Gudmundsson, Brad Maxwell, Earl J. Crain, Pancras C. Wong,
Zhen Lou, Timothy W. Harper, Silvi A. Chacko, Joseph E. Myers Jr., Steven Sheriff, Huiping
Zhang, Xiaoping Hou, Arvind Mathur, Dietmar A. Seiffert, Ruth R.Wexler, Joseph M. Luettgen,
William R. Ewing
Bristol-Myers Squibb Company, Research and Development, P.O. Box 5400, Princeton, New
Jersey 08543, United States
ABSTRACT
Factor XIa (FXIa) is a blood coagulation enzyme that is involved in the amplification of thrombin
generation. Mounting evidence suggests that direct inhibition of FXIa can block pathologic
thrombus formation while preserving normal hemostasis. Preclinical studies using a variety of
approaches to reduce FXIa activity, including direct inhibitors of FXIa, have demonstrated good
antithrombotic efficacy without increasing bleeding. Based on this potential, we targeted our
efforts at identifying potent inhibitors of FXIa with a focus on discovering an acute antithrombotic
agent for use in a hospital setting. Herein we describe the discovery of a potent FXIa clinical
candidate, 55 (FXIa Ki = 0.7 nM), with excellent preclinical efficacy in thrombosis models and
aqueous solubility suitable for intravenous administration. BMS-962212 is a reversible, direct, and
highly selective small molecule inhibitor of FXIa.
ACS Paragon Plus Environment

Page 3 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
1a
Cardiovascular (CV) disease continues to be the leading cause of death worldwide. Thrombus
formation in the arterial circulation can cause myocardial infarction or ischemic stroke. In the
venous circulation, blood clots can cause local pain and swelling. A venous blood clot can break
free and travel to the lungs causing a pulmonary embolism (PE) leading to sudden death. Together,
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
myocardial infarction, ischemic stroke, and venous thromboembolism (VTE) account for a large
portion of deaths due to cardiovascular disease.1a,b
Oral anticoagulant and antiplatelet medicines are highly effective for the long-term prevention of
thrombotic events.2a-d However, in the acute phase, it is desirable to have an antithrombotic agent
which can treat the event and prevent an early recurrence when the risk is known to be the highest.
During this acute period, oral medicines are less suited because they require hours to achieve their
maximum plasma concentration and therefore their maximum antithrombotic effect. Further, a
long-acting oral antithrombotic is not preferred if the potential for an urgent surgical intervention
is foreseeable or if the risk of hemorrhage outweighs the benefits. For decades, heparin has been
3
the most widely used acute antithrombotic agent. Heparin accelerates the antithrombin III
mediated inhibition of multiple blood coagulation proteases. Although heparin has a rapid onset
of action and, relative to oral agents, a rapid offset of action, it has some properties which limit its
use and effectiveness. Heparin requires dose titration which is often accomplished using the
activated partial thromboplastin time (aPTT) targeted to achieve 1.5- to 2.5-times the normal
aPTT. Overdose of heparin can cause serious bleeding complications. Repeated heparin
administration in rare cases can cause heparin-induced thrombocytopenia (platelet consumption)
which increases the risk of bleeding and other complications.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Factor XIa (FXIa) inhibition has emerged as a means to achieve anticoagulation without significant
4
a
effects on hemostasis. Human FXI deficiency (hemophilia C) was first described as a mild to
4
b
moderate bleeding disorder. However, these subjects rarely suffer from spontaneous bleeding
episodes, and when bleeding does occur it most often is associated with trauma or surgery,
especially in tissues with high fibrinolytic activity. Further, epidemiologic studies show an
increased risk of thrombosis in subjects with elevated FXI levels and some protection from
thrombosis in subjects with reduced FXI levels.4c,d Preclinically, inhibition of FXIa as an effective
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
a
5b
antithrombotic approach came from studies in rats and rabbits with a FXIa active site directed
irreversible inhibitor, and from studies in genetically modified FXI-deficient mice.5c,d At Bristol-
Myers Squibb we have developed a strategy to identify small molecule parenteral FXIa inhibitors
that have the potential to address the high unmet medical need in the acute care setting. This
strategy was strengthened with additional preclinical studies with other direct FXIa inhibitors,6a-n
antibodies which blocked FXI activation or FXIa activity,7a-c and antisense oligonucleotides
(
ASO)8a,b which blocked FXI synthesis. Furthermore our hypothesis to target FXIa inhibition was
supported by a phase II clinical study with a FXI-directed ASO in subjects undergoing elective
knee arthroplasty which showed that reducing FXI levels was an effective means to reduce
8c
thrombosis and also appeared to be safe with respect to bleeding. Based on our preclinical data
with small molecule FXIa inhibitors, and the recent clinical data with the ASO, the potential for
FXIa inhibition to produce a superior antithrombotic profile with a low additional bleeding risk is
a very appealing strategy to address the unmet medical need in the acute care setting.
6a
We previously reported on a FXIa tool molecule 1 (FXIa Ki = 0.2 nM) which was found to be
highly efficacious in a rabbit electrically-mediated carotid arterial thrombosis (ECAT) model, with
9
minimal increase in bleeding time at a dose that provided 80% inhibition of thrombosis. Recently,
ACS Paragon Plus Environment

Page 5 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
we also reported on the discovery of a series of phenylalanine diamide FXIa inhibitors (e.g.
compound 2) having a p-chlorophenyltetrazole P1 group and a p-aminobenzoic acid P2' moiety.6d
Extensive SAR within this series culminated in the identification of an extended piperazine P1'
urea analog, 3 (FXIa Ki = 2 nM), with excellent anticoagulant activity (aPTT IC1.5x = 0.50 µM)
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
h
and antithrombotic efficacy (EC50 = 2.8 µM) in a rabbit ECAT thrombosis model. Compound 3
was a highly polar zwitterionic molecule with poor aqueous solubility and not suitable for further
evaluation. Herein, we describe our continuing efforts directed at identifying a first small molecule
parenteral FXIa inhibitor which culminated in a Phase I clinical candidate.
Figure-1: Novel FXIa inhibitors identified at Bristol-Myers Squibb
N
NH
H
CH3
N
N
O
O
H N
2
O
NH
N
N
N N
N
H
H
H
N
N
N
N
N
HN
N
HN
O
OH
O
OH
CO H
2
O
O
O
Cl
Cl
CONH2
16a
26d
Compound
Compound 36h
FXIa Ki = 2.0 nM
aPTT EC1.5x = 0.5 µM
Compound
FXIa Ki = 0.2 nM
FXIa Ki = 1.0 nM
aPTT EC1.5x = 0.5 µM
aPTT EC1.5x = 4.2 µM
RESULTS AND DISCUSSION
Our initial attempts at improving solubility and in vitro anticoagulant activity within the
phenylalanine chemotype proved to be challenging. However, based on the FXIa bound x-ray
6d
structure for compound 2, we reasoned that the potential existed for discovering an alternative
scaffold, which would allow us to incorporate functionalities to provide adequate solubility
without compromising on the FXIa potency. We began this effort by closely examining the FXIa
bound x-ray structure for compound 2. Some of the key attributes for inhibitor 2 are (1) the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
chlorophenyltetrazole acrylamide P1 that enables the chlorine and the tetrazole moieties to occupy
the S1 pocket making key interactions with Tyr 228 at the bottom of the pocket and the disulfide
at the top of the S1 pocket, (2) the acrylamide linker that places the amide carbonyl such that it
directly interacts with enzyme residues in the oxyanion region, while the carboxamide NH (which
is attached to the P2' benzoic acid group) interacts with a conserved water, (3) the benzoic acid P2'
group that forms a strong interaction with Tyr 143 and His 40 and (4) the phenyl moiety of the
phenylalanine that is within contact distance of the disulfide bridge in the S1' region.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure-2: FXIa bound x-ray structure of phenylalanine analog 2 (3.3 Å, PDB 5E2O)6h
'
P1
Strategy-1
N
N
H
N
N
N
HN
'
O
OH
P1
P2
O
O
Cl
2 FXIa Ki = 1.0 nM
aPTT EC1.5x = 4.2 µM
'
P1
N
N
H
N
N
N
N
'
O
OH
P1
P2
O
O
Cl
4
FXIa Ki = 23 nM
aPTT EC1.5x = 36 µM
The benzylic carbon bearing the P1' phenyl group and the acrylamide NH appeared to be in the
same plane. This observation prompted us to determine whether we could connect these two
atoms in the form of a proline. A careful look at the proline scaffold model for compound 4
showed that relative to compound 2, this connectivity would be a viable approach without
significantly altering the overall conformation and binding to FXIa. Indeed, compound 4,
ACS Paragon Plus Environment

Page 7 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
showed modest FXIa binding affinity (FXIa Ki = 23 nM). Figure-3 shows an overlay of the FXIa
bound x-ray structures for compound 2 and proline analog 4. While the orientation of the P1
groups are similar in the S1 pocket for 2 and 4, some subtle binding changes were seen
elsewhere which may account for the decrease in binding affinity. For example, the acrylamide
linker carbonyl oxygen atom interaction with oxyanion hole residue Gly 193 differs slightly
between 2 and 4 (~3.0 Å for compound 4 and ~2.7 Å for compound 2). The incorporation of the
proline scaffold in 4 positions the 3(R)-phenyl P1' group to be in close proximity to His 57,
eliciting an edge to face interaction and also appears to be slightly extended over the disulfide
bridge. In contrast, for compound 2, a slight twist of the phenylalanine P1' group is observed
relative to compound 4. In the S2' region, the benzoic acid P2' group in 4 appears to be slightly
extended relative to compound 2. Various substitutions were made on the proline linked P1'
phenyl ring, but none afforded potencies that were better than compound 4.
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: Overlay of x-ray structures of compound 2 (3.3 Å; PDB 5E2O) (blue) and 4 (2.9 Å;
PDB 5QCK) (brown) in the active site of FXIa.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
To address the potency issue, we embarked on another strategy (Figure-4). This strategy relied on
a connectivity that involved the phenyl P1' group and the acrylamide nitrogen as shown in Figure-
4. This concept generated isomeric tetrahydroisoquinoline (THIQ) scaffolds (Figure-4; strategy 2
and 3, compounds 5 and 6). The computational and modeling experiments suggested that the THIQ
scaffold 6 would be best suited for this purpose and could make several key FXIa
Figure-4: Evolution of the tetrahydroisoquinoline (THIQ) FXIa scaffold.
ACS Paragon Plus Environment

Page 9 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
P1
N N
N N
H
H
N
N
Strategy-2
N
N
N
HN
N
N
S
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
O
OH
'
O
5
P1
P2
COOH
O
O
Cl
2
Cl
FXIa Ki = 1700 nM
Strategy-3
N N
H
N
N
N
N
S
O
O
COOH
6
Cl
FXIa Ki = 11 nM
aPTT EC1.5x > 40 µM
interactions previously seen with 2 and 4. To test this hypothesis, homochiral compounds 5 and 6
were prepared. We were delighted that compound 6 (FXIa Ki = 11 nM) performed as predicted by
our computational study and was significantly more potent than 5 (FXIa Ki = 1700 nM). Figure-5
shows an overlay of the FXIa bound x-ray structure of 6 (2.11 Å) with 2, from which it can be
seen that the location of the chiral center in 6 is shifted relative to compound 2. One plausible
explanation for this observation was that, in order to accommodate the rigid THIQ scaffold, a slight
shift in the orientation of 6 in the FXIa active site was necessary. This change also resulted in the
benzoic acid P2' group in 6 being extended slightly into the S2' region. Importantly, the change
notwithstanding, the highly constrained THIQ scaffold maintained all the binding interactions seen
previously with compounds 2 and 4 in the FXIa binding site, and additionally the scaffold formed
an edge to face interaction with His 57 and also appears to engage the Cys 42 - Cys 58 disulfide
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
residues in the S1' region. Although the potency changes were not significant at this juncture, the
overall strategy of incorporating a novel THIQ scaffold proved invaluable for further optimization.
Figure-5: Overlay of FXIa bound x-ray structures 2 (3.3 Å; PDB 5E2O) (blue) and 6 (2.11 Å;
PDB 5QCL) (brown)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
S1’
S2’
S1
The encouraging FXIa data obtained for 6 prompted us to also evaluate a select set of 5 and 6-
membered heterobicyclic THIQ analogs (Table-1). The 5,6,7,8-tetrahydro-1,6-naphthyridine 7
(FXIa Ki = 56 nM) and 1,2,3,4-tetrahydro-2,7-naphthyridine 8 (FXIa Ki = 160 nM) were weaker
relative to 6, but were within 3-fold of each other in terms of potency. From the 5,6-heterobicyclic
chemotype, the unsubstituted 4,5,6,7-tetrahydro-2H-pyrazolo[4,3-c]pyridine 9 (FXIa Ki = 640
nM) was significantly less potent. Although the NH of 9 provided us with an opportunity for
ACS Paragon Plus Environment

Page 11 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substitution at this position, we hypothesized that the shallow S1' pocket would not allow for bulky
substitutions. To test this hypothesis, we prepared the 2-methyl analog 10, which shows a modest
3
-fold improvement in FXIa binding affinity relative to 9; whereas the 2-phenyl (11) and 2-benzyl
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
(
12) analogs were much weaker in line with our modeling predictions. The 2,3-dimethyl analog
13 (FXIa Ki = 29 nM) showed comparable binding affinity to 6, although a loss in affinity was
seen for the regioisomeric analogs 14 (FXIa Ki = 80 nM) and 15 (FXIa Ki = 300 nM). This limited
SAR for alternative THIQ scaffolds appeared to favor compound 6 for further optimization studies.
Table-1: 5, and 6-Membered heterobicyclic THIQ replacements
R
N N
H
N
N
N
N
O
O
COOH
Racemic
Cl
aFXIa Ki nM
Compound
6
R
1
1
(chiral)
N
7
51
N
8
9
160
NH
N
640
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
N
N
1
1
0
1
270
Ph
N
N
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2000
Ph
N
N
1
400
29
1
1
2
3
N
N
N
N
N
14
15
80
N
300
aK
values were obtained from purified human enzyme at 25 °C and were averaged from multiple determinations (n=2), as described
i
in Ref. 6c.
We next focused on exploring alternative P2' groups (Table-2). It was interesting to note that the
meta-benzoic acid compound 16 was ~26-fold less active when compared to 6. We also explored
the methylcarbamate and indazole P2' groups previously identified in the phenylimidazole series
6
c
of compounds. The carbamate P2' containing compounds 17 and 18 showed a significant loss in
FXIa potency. A similar finding was also found for indazole compounds 19 and 20, suggesting the
ACS Paragon Plus Environment

Page 13 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
location of the carbamate and indazole P2' moieties in the S2' pocket are slightly altered resulting
in weaker interactions with the enzyme residues in this region.
Table-2: P2' THIQ structure activity relationships
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
N
N
N
N
R
O
O
X
Homochiral
Cl
Compound
6
X
H
R
aFXIa Ki nM
COOH
1
1
290
1
1
1
6
7
8
H
H
H
COOH
NHCO Me
2
870
890
NHCO Me
2
N
N
1
2
9
0
H
H
57
83
H
NH2
N
N
H
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
NHCO Me
2
2
1
F
F
110
COOH
22
3.7
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
aK
in Ref. 6c.
values were obtained from purified human enzyme at 25 °C and were averaged from multiple determinations (n=2), as described
i
The proline compound 4 and the THIQ 6 show that the 4-chlorophenyltetrazole P1 group, which
was carefully optimized for the diamide 2, is effective in other chemotypes such as in the
phenylalanine series.6d,6h Additionally in the phenylalanine series, we had observed that subtle
changes on the P1 phenyl group such as incorporation of a fluoro group adjacent to the 4-Cl moiety
6
h
impacted the FXIa potency. The introduction of a 3-fluoro-4-chlorophenyltetrazole P1 in 21
FXIa Ki = 110 nM, Table-2) afforded an 8-fold improvement in FXIa binding compared to 17
(
lacking the 3-fluoro substituent. A similar trend was also observed for 22 (FXIa Ki = 3.7 nM)
relative to 6. To understand this improvement in potency, we obtained the FXIa bound structure
for compound 21 and compared it with the bound conformation of 6 (Figure-6). Not too
surprisingly, the P1 4-chloro group in 21 engages the Tyr 228 residue, while the adjacent 3-fluoro
moiety appears to be in close contact with Thr 213 (3.0 Å). This interaction could in part result in
the slight potency gain seen with the 3-fluoro-4-chlorophenyltetrazole P1 in 21. For the S2'
interactions, the carbamate P2' group in 21 appears to orient above the benzoic acid group in 6.
The slight shift in the orientation of the carbamate group potentially weakens the interactions with
the enzyme residues in the S2' region, even though a close contact of the carbamate carbonyl and
Tyr 143 (2.8 Å), and a corresponding network water interaction with His 40 is observed.
Figure-6: An overlay of the FXIa bound x-ray structures of compound 6 (2.1 Å; PDB 5QCL)
(
red) and 21 (2.2 Å; PDB 5QCM) (purple).
ACS Paragon Plus Environment

Page 15 of 73
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
The orientation of the THIQ scaffold in the FXIa enzyme S1' pocket provided us with a unique
opportunity to explore substitutions on this scaffold. Figure-6 shows the THIQ in 21 to be in close
proximity to the disulfide in the S1' region, such that substitution at either the 7 or 8 positions of
the THIQ ring would result in a steric clash. On the other hand, substitutions at either the 5 or 6
positions of the THIQ seemed feasible, though some steric clash with the S1' residues with
substitution at the 6 position of the THIQ could not be ruled out. Substitution at the 5-position of
the THIQ appeared to project into the S2 pocket of the FXIa active site. To test this hypothesis,
we prepared the select set of compounds listed in Table-3. From this initial set, the compounds
that showed promise were indeed the 5-substituted compounds, such as the dimethylamino
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compound 25 (FXIa Ki = 6 nM), methoxycarbonyl compound 23 (FXIa Ki = 17 nM) and cyano
compound 24 (FXIa Ki = 39 nM). The FXIa affinity of the dimethylamino compound 25 was the
most encouraging. Based on the availability of the THIQ scaffolds, we explored a relatively small
set of compounds with substitutions at positions 6, 7 and 8 on the THIQ ring. From this set it was
clear that a small substituent at the 6-position of the THIQ, such as the 6-methyl 27 was comparable
to compound 6 in terms of potency, but larger groups as in the 6-methylester analog 26 resulted in
a loss of FXIa affinity.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table-3: Exploratory THIQ scaffold substitutions
6
5
7
8
N N
H
N
N
N
N
O
O
COOH
Racemic
Cl
a
Compound
(chiral)
R
H
FXIa Ki nM
6
11
17
39
6
140
27
2
2
2
2
3
4
5
6
5-CO Me
2
5-CN
5-NMe
2
6-CO
2
Me
27
6-Me
2
2
3
3
8
9
0
1
7-CO
7-CN
7-Me
2
Me
230
400
28
8-CO
2
Me
2200
aK
values were obtained from purified human enzyme at 25 °C and were averaged from multiple determinations (n=2), as described
i
in Ref. 6c.
Other substitutions at either the 7 or 8 positions of the THIQ ring also showed a loss in FXIa
affinity relative to compound 6. We were delighted with these results as it validated our hypothesis
ACS Paragon Plus Environment

Page 17 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for substitution on the THIQ scaffold. Importantly, compound 25 opened up the possibility for an
extensive SAR at the previously unexplored S2 pocket of FXIa via substitution at the 5-position
of the THIQ to probe for additional binding interactions.
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: Effect of electron withdrawing groups at the 5-position of THIQ scaffold – S2 SAR
S2 SAR
R
N N
H
N
N
N
N
O
O
COOH
Racemic
Cl
Compound
R
aFXIa Ki nM
2
3
3
3
2
3
CO
COOH
CONMe
2
Me
17
74
25
2
O
OMe
3
4
N
Me
O
27
17
18
12
6
35
N
O
3
3
6
7
N
O
O
O
CO Et
2
N
CO Et
2
38
N
aK
in Ref. 6c.
values were obtained from purified human enzyme at 25 °C and were averaged from multiple determinations (n=2), as described
i
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure-7: FXIa bound x-ray structure of compound 38 (2.3 Å; PDB 5QCN)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
At first we explored the effectiveness of a select set of carboxamido groups (compounds 33-38)
at the 5-position of the scaffold (Table-4). In general, the FXIa affinities were comparable to that
observed for compound 23; however, the carboxamide 38 (FXIa Ki = 6 nM) was approximately
2
-3 fold better. The 5-carboxylic acid 32 had ~4-fold less FXIa affinity when compared to the
corresponding ester compound 23. An FXIa bound x-ray structure for compound 38 (2.30 Å)
was obtained (Figure-7). The S2 region of the FXIa enzyme shows an array of water molecules.
The P2 carboxamide compound 38 appears to derive its potency by engaging these water
molecules. The carbonyl oxygen of the piperidinyl carboxylate interacts with the S2 enzyme
residues through a network of waters. While this is an intriguing find, it is also interesting to
ACS Paragon Plus Environment

Page 19 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
note that 38 and the dimethylamino analog 25 (FXIa Ki = 6 nM, Table-3) display similar FXIa
affinities.
Table-5: A selection of 5-amino substituted THIQ analogs.
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
R
N N
H
N
N
N
N
O
O
COOH
Racemic
Cl
Compound
R
aFXIa Ki nM aPTT IC 1.5x
µM
Me
N
2
5
Me
6
>40
Me
3
9
N
O
21
23
26
H
NHCO
2
Me
40
>40
H
N
H
N
CO Et
2
4
4
1
2
O
25
17
<5
<5
22
>40
0.65
11
H
H
N
N
O
N
NH
43
CO Me
2
N
N
4
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
N
4
5
5.8
39
aK
values were obtained from purified human enzyme at 25 °C and were averaged from multiple determinations (n=2), as described
i
in Ref. 6c. aPTT (activated partial thromboplastin time) in vitro clotting assay was performed in human plasma as described in
reference 6c.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Encouraged by similar results for compounds containing either a carboxy or an amino P2
substituent as reflected by the FXIa potencies for compounds 25 and 38, we explored additional
5-substituted amino P2 containing compounds (Table-5). The less basic compounds such as the
acetamide and urea P2 compounds 39-42 showed a 3-4-fold loss in FXIa binding affinity relative
to the dimethylamino compound 25, while the more polar 5-piperazine containing compound 43
(
FXIa Ki < 5 nM) was significantly better. Compound 43 also showed excellent anticoagulant
activity in the clotting assay (aPTT IC1.5x = 0.65 µM). A similar finding was also observed with
the piperazine carbamate compound 44, although its clotting activity was much weaker. The effect
of basicity on FXIa affinity and clotting activity is further accentuated by the observation that the
lactam P2 analog 45 (FXIa Ki 5.8 nM, aPTT IC1.5x = 39 µM) displayed good FXIa affinity, but
also showed weak anticoagulant activity. Compound 43 at this juncture, exhibited the
characteristics suitable for a parenteral agent in terms of potency but to advance into further studies
acceptable solubility was required. The solubility was measured (< 5 µg/ml) and determined to be
low for its amorphous form. The poor solubility trends were a characteristic feature for a number
of compounds in Table-5. The racemic THIQ P2 compounds as in Table-5 allowed us to rapidly
identify key functional groups and trends that could be further exploited with homochiral
compounds.
The THIQ analogs that were substituted at the 5-position allowed us to reach our FXIa potency
and the in vitro anticoagulant criteria. Moreover, we used the information shown in Table-5 to
further our understanding of the SAR with the goal of achieving a highly potent compound with
ACS Paragon Plus Environment

Page 21 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
good aqueous solubility. To do this, we incorporated the highly optimized 3-fluoro-4-
chlorophenyltetrazole P1 group with homochiral THIQ scaffolds (Table-6). This resulted in the
identification of highly potent FXIa compounds. Importantly, the THIQ scaffold allowed for the
incorporation of a diverse set of P2 groups that were well tolerated in the active site of FXIa. For
example, compounds that contained morpholine (46), thiomorpholine (47), sulfonyl (48) and
piperidine (49) P2 groups had sub-nanomolar FXIa affinity, with the morpholine analog (46) also
demonstrating good anticoagulant activity (aPTT1.5x = 0.9 µM). A similar finding was also
observed with the zwitterionic compounds, such as 4-dimethylamino piperidine compound 50
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
(
FXIa Ki = 0.1 nM, aPTT1.5x = 0.1 µM), piperazinone analogs 54 and 55 and compounds such as
spirolactam compound 51, lactam 52, and morpholinone 53. The lactam containing compounds
2, 53 and 55 all showed acceptable solubility at pH 6.4. Among these, compound 55 had the most
balanced profile in terms of FXIa affinity, anticoagulant activity and amorphous aqueous solubility
pH 6.4 Sol = 0.88 mg/mL). Compound 55 represented a milestone compound for the parenteral
5
(
program and had all the physical chemical properties we sought for advancing the compound into
development.
A model of compound 55 (Figure-8) in the FXIa active site shows the P2 N-methylpiperazinone
group making a liphophilic interaction with Tyr 94 and Tyr 58B. Additionally, the benzoic acid
P2' and the chlorophenyltetrazole P1 groups make the requisite interactions at the S1 and S2'
regions as previously described for similar compounds.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table-6: The SAR 5-substituted heterocyclic P2 homochiral THIQ analogs.
R
N N
H
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
N
N
N
N
O
O
COOH
F
Homochiral
Cl
Compound
R
aFXIa
Ki nM
aPTT IC50
Solubility
@pH 6.5
mg/ml
1.5x
µM
O
S
4
4
6
7
N
0
.38
0.91
2.5
-
-
N
0.49
SO2
N
48
2.0
2.2
29
24
-
-
4
9
N
Me
N
Me
50
0.098
0.11
NT
N
N Me
51
52
N
0.07
1.4
0.45
0.84
0.094
0.92
O
O
N
ACS Paragon Plus Environment

Page 23 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
O
N
5
3
2.0
0.55
0.37
O
NH
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
54
1.4
0.3
0.027
O
Me
N
5
5
0.71
0.3
0.88
N
aK
values were obtained from purified human enzyme at 25 °C and were averaged from multiple determinations (n=2), as described
i
in Ref. 6c. aPTT (activated partial thromboplastin time) in vitro clotting assay was performed in human plasma as described in
reference 6c. NT = Not tested. All solubilities are from amorphous compounds.
Figure-8: A docked model of compound 55 in FXIa
PHARMACOKINETIC PROFILE FOR THIQ FXIa INHIBITORS
The compounds that were chosen for evaluation in rat PK studies (cassette dosing format, Table-
7
) satisfied the criteria we had established, i.e., potent FXIa binding, good anticoagulant activity
as measured in the aPTT assay and adequate aqueous solubility. From this study we were seeking
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compounds with short half-life (< 1 h) to instill a rapid onset and offset of pharmacology.
Compound 50 showed low clearance values and also had a short half-life. Compounds 51-53 and
5
5, had moderate clearance values and likewise demonstrated short half-lives in the desirable
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
range. Compound 55 best met the pre-determined criteria of low to moderate clearance and a
pharmacokinetic half-life of less than 1 h, and thus emerged as a potential leading candidate. The
pharmacokinetic profile of 55 was subsequently evaluated in rats, dogs and cynomologus monkeys
(Table-8). The clearance in rats and cynos was higher (~2-fold) than that observed with dogs. The
volume of distribution was the lowest in rat and moderate in dogs and cynos. The mean residence
time and half-life were less than 1 h in all three species.
Table-7: Intravenous (IV) cassette dosing pharmacokinetics (PK) in rats
Compd.
Microsome
CL
Vss
AUCtotal
t
1/2
T
1/2 min
ml/min/kg
L/kg
nM●h
h
(h/r)
5
0
90, 70
53, 67
4.0
10
23
16
64
16
0.24
0.22
0.18
0.68
0.81
0.75
4796
2156
624
1.18
0.74
0.35
1.39
0.25
0.88
51
5
5
2
3
31, 7.5
40, 9.7
248
54
>120, 45
>120, 59
183
5
5
309
intravenous dosing solution 70%PEG400; 20%H2O; 10%EtOH. CL = clearance, Vss = steady-state volume of
distribution, AUC = area under the curve, t1/2 = half life
Table-8: Multi species pharmacokinetic data for compound 55.
ACS Paragon Plus Environment

Page 25 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Animal
Species
Rat
Dose
mg/kg
0.5
CL
Vss
L/kg
0.32
0.71
1.02
1.16
AUCtotal
nM●h
441
t
1/2
ml/min/kg
h
31
43
15
31
0.22
0.6
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
Rabbit
Dog
0.5
270
4.75
2.03
7757
1590
1.05
0.78
Cyno
CL = clearance, Vss = steady-state volume of distribution, AUC = area under the curve, t1/2 = half life
IN VIVO EFFICACY IN THE RABBIT ARTERIOVENOUS SHUNT MODEL
Compound 55 was next advanced into the rabbit arteriovenous shunt (AV-shunt) thrombosis
model11a for efficacy determination. When evaluated in this model, compound 55 (rabbit FXIa Ki
=
3 nM; rabbit activated plasma kallikrein Ki 1,100 nM) demonstrated a dose-response
relationship with respect to inhibiting thrombus formation (Figure-9). At the highest dose
evaluated (0.23 + 1.55 mg/kg/h IV), the compound showed an 83 % inhibition of thrombus
formation and increased aPTT up to 1.8-fold over baseline, with no prothrombin time (PT)
changes noted. Changes in aPTT but not PT are consistent with the mechanism of inhibiting
FXIa. Compound 55 was evaluated in a rabbit cuticle bleeding model,11b at twice the dose used
in the AV-shunt model (0.46 mg/kg+3.1 mg/kg/h) and did not increase the bleeding time as
compared to vehicle. The combination of aspirin (4 mg/kg/h IV) and 55 (0.46 mg/kg+3.1
mg/kg/h) resulted in bleeding times that were not greater than for aspirin treatment alone.12a
Figure-9. Antithrombotic and clotting effects (aPTT and PT) of Compound 55 in the rabbit AV-
Shunt model. Data are means ± SEM (n = 3 per group).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
C l o t t i n g T i m e s i n R a b b i t A V - S h u n t
R a b b i t A V - S h u n t
8 3 %
a P T T
P T
2
1
1
1
1
. 0
. 8
. 6
. 4
. 2
1
0 0
8
6
4
2
0
0
0
0
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4 6 %
1 . 0
5
1
5
5
. 5
1
0
.
+
3
.
2
0
0
2 3 +
0
.
C o m p o u n d 6 1 ( m g /k g
+
m g /k g /h I V )
C o m p o u n d 6 1 ( m g /k g
+
m g /k g /h I V )
ADDITIONAL PROFILE FOR COMPOUND 55
The protease selectivity profile for compound 55 is shown in Table-11. The compound is a
highly potent FXIa inhibitor (FXIa Ki = 0.7 nM) with an excellent protease selectivity profile.
The aqueous solubility for compound 55 was measured in buffer systems that would be clinically
acceptable, and its long term solution stability was also evaluated. The crystalline form of 55
Table-11. The human serine protease selectivity profile for 55.
Human Enzyme Ki
55
0.71
a
(
nM)
Factor XIa
Factor VIIab
Factor IXa
>13,000
>27,000
>9,000
>3,000
310
Factor Xab
Factor XIIa
Plasma kallikrein
Tissue kallikrein-1
Thrombinb
>10,000
>13,000
ACS Paragon Plus Environment

Page 27 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Trypsinb
Activated Protein C
Plasmin
>6,200
>21,000
>25,000
>6,200
TPA
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
Urokinase
>15,000
11,000
Chymotrypsin
a
o
b
o
i i
K values in nM were obtained using human purified enzymes at 37 C. K values in nM were obtained using human purified enzymes at 25 C.
(
hemi-hydrate) showed good aqueous solubility (Table-9), and it had good stability for over 90
days and at varying temperatures (Table-10). The measured pKa values for 55 were 4.7 and 4.1
for the N-Me-piperazine and carboxylic acid groups, respectively. The free fraction for 55 in
human, rat, dog and cynomolgus monkey serum was 33.2 %, 44.8 %, 67.5 % and 51.8 %,
respectively. The compound also showed no CYP activity (IC50 values > 40 µM) and was stable
in human, rat and dog liver microsomes (T1/2 = 108, 59, 120 min, respectively). Compound 55 was
AMES negative.
Table-9: Pharmaceutical properties for compound 55
pKa Acid (a) Base (b)
HPLC Log D
4.1 (a); 4.7 (b)
1.40
pH
.7
7.4
Solubility (mg/mL)
6
0.26
1.5
11
>25
>100
Aqueous Solubility
8
8
8
.0
.3
.5
Table-10: Buffer-pH-temperature solution stability for 55 after 90 days
Buffers
pH
6.5
7.4
25 °C
99.9 %
100 %
37 °C
99.8 %
99.0 %
50 °C
95.4 %
94.6 %
0
0
.1 M Phosphate
.1 M Phosphate
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 73
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0.1 M Tris
8.0
99 %
99.7 %
94 %
CHEMISTRY
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compounds 4-6 were prepared in a three step sequence from commercially available materials
7a-c (Scheme-1). Standard amide couplings of 57a-c with BOP reagent and Hunig’s base,
5
followed by removal of the N-Boc protecting group and coupling with (E)-3-(5-chloro-2-(1H-
tetrazol-1-yl)phenyl)acrylic acid 58 afforded compounds 4-6 in good overall yields. The highly
elaborated THIQ scaffolds shown in Table-2 were accessed from commercially available starting
materials which allowed rapid preparation of the final products via a three component Ugi
reaction.10a,b For example, compound 38 (Scheme-2) was prepared from commercially available
2-(tert-butoxycarbonyl)-1,2,3,4-tetrahydroisoquinoline-5-carboxylic acid. Coupling of this
intermediate with (S)-ethyl-piperidine-3-carboxylate afforded compound 59 in good yield. To
perform the Ugi reaction it was necessary to access the imine 60. This was accomplished in a three
step sequence by first removal of the Boc protecting group followed by neutralization and
oxidation with excess activated MnO in dichloromethane to provide 60 in 74% overall yield.
2
Scheme-1: Method for the preparation of compounds 4-6.a
ACS Paragon Plus Environment

Page 29 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
H N
2
O
O
Compounds 4, 5,
6
COOH
N
N
COOH
N
COOH
Fmoc
Boc
1. POCl , pyridine or BOP or T3P,
3
CBz
DIEA
quant.)
(quant.,98%,
5
7c
5
7a
57b
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
2. TFA, DCM or H , Pd/C,
morpholine
2
7
7%)
(quant.,75%,
. 58a, DIEA or 58, T3P, DIEA
1%, quant.)
3
4
9
(quant.,
. TFA/ DCM
21%, 14%)
(
29%,
N
N
O
N
N
O
HO N
O
N
N
N
O
N
O
N
O
OH
DIC
O
R
R
5
8
R = H, 58a, quant.
R = 58b,
Cl
Cl
F,
quant.
Isonitrile intermediate 61 was prepared in two steps by heating tert-butyl-4-aminobenzoate in neat
ethyl formate. The ensuing formamide intermediate was subsequently dehydrated to the isonitrile
6
1 by treatment with POCl
assembled by the use of the three component Ugi reaction by combining intermediates 58, 60 and
1 in EtOH. Treatment with TFA afforded compound 38 in low yield. The preparation of
3
in THF containing TEA. The highly complex compounds were then
6
compound 55 required the availability of the imine intermediate 76. This was easily obtained from
commercially available 5-bromoisoquinoline. Ullmann coupling of this material with N-
methylpiperazinone in the presence of catalytic amount of CuI, 1,10-phenanthroline and K
2
CO at
3
o
130 C in DMSO afforded the coupled intermediate 74 (31% yield). This was easily converted to
the requisite imine intermediate 76 by hydrogenation with PtO
with excess MnO in good overall yields. Ugi coupling of 76 with chlorophenyltetrazole
Scheme-2: The synthesis of compounds 38 and 55.a
2
catalyst, followed by oxidation
2
ACS Paragon Plus Environment