Small Molecule Proprotein Convertase Subtilisin/Kexin Type 9 (PCSK9) Inhibitors: Hit to Lead Optimization of Systemic Agents

Article abstract of DOI:10.1021/acs.jmedchem.8b00650

The optimization of a new class of small molecule PCSK9 mRNA translation inhibitors is described. The potency, physicochemical properties, and off-target pharmacology associated with the hit compound (1) were improved by changes to two regions of the molecule. The last step in the synthesis of the congested amide center was enabled by three different routes. Subtle structural changes yielded significant changes in pharmacology and off-target margins. These efforts led to the identification of 7l and 7n with overall profiles suitable for in vivo evaluation. In a 14-day toxicology study, 7l demonstrated an improved safety profile vs lead 7f. We hypothesize that the improved safety profile is related to diminished binding of 7l to nontranslating ribosomes and an apparent improvement in transcript selectivity due to the lower strength of 7l stalling of off-target proteins.

Full text of DOI:10.1021/acs.jmedchem.8b00650

Subscriber access provided by Kaohsiung Medical University
Article
Small Molecule Proprotein Convertase Subtilisin/Kexin type 9
(PCSK9) Inhibitors: Hit to Lead Optimization of Systemic Agents
Allyn Timothy Londregan, Liuqing Wei, Jun Xiao, Nathanael G Lintner, Donna N. Petersen,
Robert Dullea, Kim F. Mcclure, Michael W Bolt, Joseph S Warmus, Steven B. Coffey, Chris
Limberakis, Julien Genovino, Benjamin A. Thuma, Kevin D. Hesp, Gary E Aspnes, Benjamin Reidich,
Christopher T. Salatto, Jeffrey R Chabot, Jamie H.D. Cate, Spiros Liras, and David W. Piotrowski
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00650 • Publication Date (Web): 07 Jun 2018
Downloaded from http://pubs.acs.org on June 7, 2018
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 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Small Molecule Proprotein Convertase
Subtilisin/Kexin type 9 (PCSK9) Inhibitors: Hit to
Lead Optimization of Systemic Agents
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
Allyn T. Londregan,*, † Liuqing Wei, † Jun Xiao, † Nathanael G. Lintner, ‡ Donna Petersen,#
Robert G. Dullea, § Kim F. McClure, ǁ Michael W. Bolt, ǂ Joseph S. Warmus, † Steven B. Coffey, †
Chris Limberakis, † Julien Genovino, † Benjamin A. Thuma, † Kevin D. Hesp, † Gary E. Aspnes, ǁ
Benjamin Reidich, § Christopher T. Salatto, § Jeffrey R. Chabot,^ Jamie H. D. Cate,‡, ┴,◊,= Spiros
Lirasǁ and David W. Piotrowski*,†
†Pfizer Medicinal Chemistry, Groton, Connecticut 06340, United States
‡Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley,
California, 94720, United States
ǁPfizer Medicinal Chemistry, Internal Medicine Research Unit, Pfizer Worldwide Research and
Development, Cambridge, Massachusetts, 02139, United States
#Primary Pharmacology Group, Pharmacokinetics, Dynamics and Metabolism, Pfizer Worldwide
Research and Development, Groton, Connecticut, 06340, United States
ǂDrug Safety Research & Development, Pfizer Worldwide Research & Development,
Cambridge, Massachusetts, 02139, United States
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 53
1
2
3
4
5
^Pfizer Pharmacokinetics, Dynamics and Metabolism Modeling and Simulation, Pfizer
Worldwide Research and Development, Cambridge, Massachusetts, 02139, United States
6
7
8
9
§Internal Medicine Research Unit, Pfizer Worldwide Research and Development, Cambridge,
Massachusetts, 02139, United States,
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
┴QB3 Institute, University of California, Berkeley, Berkeley, California, 94720, United States
=Department of Chemistry, University of California, Berkeley, Berkeley, California, 94720,
United States
◊Physical Biosciences Division, Lawrence Berkeley National Laboratory, Berkeley, California,
94720, United States
KEYWORDS: PCSK9, proprotein, convertase, lipoprotein, LDLR, protein synthesis, translation,
LDLꢀC
ABSTRACT: The optimization of a new class of small molecule PCSK9 mRNA translation
inhibitors is described. The potency, physicochemical properties and the offꢀtarget pharmacology
associated with the hit compound (1) were improved by changes to two regions of the molecule.
The last step in the synthesis of the congested amide center was enabled by three different routes.
Subtle structural changes yielded significant changes in pharmacology and offꢀtarget margins.
These efforts led to the identification of 7l and 7n with overall profiles suitable for in vivo
evaluation. In a 14ꢀday toxicology study, 7l demonstrated an improved safety profile vs. lead 7f.
We hypothesize that the improved safety profile is related to diminished binding of 7l to nonꢀ
translating ribosomes and an apparent improvement in transcript selectivity due to the lower
strength of 7l stalling of offꢀtarget proteins.
ACS Paragon Plus Environment
2

Page 3 of 53
Journal of Medicinal Chemistry
1
2
3
4
INTRODUCTION
5
6
7
8
Proprotein Convertase Subtilisin/Kexin type 9 (PCSK9) is a secreted 71 kDa serine proteinase
consisting of an Nꢀterminal prodomain, a catalytic domain, and a Cꢀterminal domain. The protein
is primarily expressed in the liver and is exported to plasma where it down regulates cell surface
levels of the low density lipoprotein receptor (LDLR) via endocytosis and lysosomal
degradation.1 The affinity of PCSK9 for LDLR becomes progressively tighter at lower pH with
the equilibrium dissociation Kd reaching low nanomolar concentrations at lysosomal pH.2 After
initial endocytosis, the PCSK9ꢀLDLR complex is believed to undergo a sorting process within
the endosome. By a mechanism potentially involving accessory proteins, LDLR is either
recycled3 to the cell surface or carried to its destruction in the lysosome with PCSK9.4 As such
PCSK9 has a significant role in the maintenance of plasma low density lipoprotein cholesterol
(LDLꢀC), a primary risk factor for cardiovascular disease. Human genetics and the PCSK9
neutralizing antibodies confirm the important role of PCSK9 in LDLꢀC regulation. Individuals
with PCSK9 loss of function mutations have both lower LDLꢀC and reduced cardiovascular
risk.5 Equally, the PCSK9 antibodies alirocumab, evolocumab and bococizumab have
demonstrated 35ꢀ69% reduction in plasma after 12 weeks of biweekly parenteral administration.6
Although antibodies targeting mature PCSK9 are the most advanced clinically, other points of
intervention and modalities have been reported. In all cases, small molecule progress has lagged
behind. Unlike the majority of the proprotein convertase family members, the prodomain of
PCSK9 remains tightly bound to the catalytic domain in its mature form.7 This is attributed to the
lack of a secondary cleavage event within the prodomain thought to destabilize its association
with catalytic subunit of the convertase. In spite of this, attempts have been made to target the
hydrolase center (Ser386, His226, Asp186) of mature PCSK9 based on computed models
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
3

Journal of Medicinal Chemistry
Page 4 of 53
1
2
3
4
5
6
7
8
9
wherein the prodomain has been removed.8 The fact that dissociation of the prodomain can be
observed in the context of domain exchange experiments supports potential access to the
catalytic triad.9 In principle, an active site binder could displace the prodomain from the catalytic
domain leading to improper folding and secretion of PCSK9. Nevertheless, the demonstration
that the proteolytic activity is not required for PCSK9 functional regulation of LDLR10 puts into
question the effectiveness of molecules designed to bind the catalytic triad or neighboring
pockets of secreted PCSK9.2a The functional action of PCSK9 is therefore believed to be more a
result of pH sensitive proteinꢀprotein interactions over diverse regions of its surface.11 To obtain
competitive inhibitors would then require inhibition of PCSK9’s interaction with the LDLR.
PCSK9 is the preferred target between the two proteins as acting directly at the LDLR presents
the risk of interfering with its role in LDLꢀC regulation. One of the primary contact patches with
PCSK9 at neutral pH is the epidermal growth factor precursor homology domain A (EGFꢀA) of
the LDLR. While parenteral antiꢀPCSK9 antibodies and adnexins are capable of gaining affinity
by extending beyond this small (530 Å) and relatively featureless EGFꢀA interacting region on
PCSK9, peptide approaches have struggled to achieve the size and properties sufficient to
facilitate oral administration.12 For small molecules, gaining sufficient perch on the flat EGFꢀA
interacting region of PCSK9 is understandably even more challenging but a recent disclosure
suggests that the adjacent P’ꢀhelix site offers some hope for disruption of the PCSK9/LDLR
interaction with small molecules.13 A review of other small molecule inhibitors has been
disclosed.14
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
For these reasons attempts have been made to target PCSK9 prior to its maturation. The recent
demonstration that the PCSK9 catalytic domain is capable of intermolecular (in trans) substrate
cleavage has led to the hope of identifying small molecules that can inhibit the intramolecular (in
ACS Paragon Plus Environment
4

Page 5 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
cis) autoꢀprocessing of proPCSK9.15 The inhibition of PCSK9 mRNA translation has also been
explored. Both antisense locked nucleic acids and short interfering RNA (siRNA) of PCSK9
messenger RNA have progressed to human clinical trials with significant PCSK9 and LDLꢀC
lowering being reported for the siRNA.16
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
Previously, we described the identification of small molecule inhibitors of PCSK9 mRNA
translation by a cellꢀbased HTS designed to find molecules that decreased PCSK9 secretion to
the cell media.17 From this effort, PFꢀ00932239 (1) was identified as a hit (Table 1) and
subsequently shown to decrease PCSK9 secretion and increase cell surface LDLR and LDLꢀC
uptake in recombinant and native cell lines via a mechanism attributed to the inhibition of
PCSK9 early in its translation. PFꢀ06446846 (7f) was later identified as an advanced prototype
suitable for biological, safety and mechanistic evaluation, the details of which were recently
described.18 In parallel, efforts to achieve a liverꢀselective PCSK9 inhibitor via a proꢀdrug
strategy were also explored.19 Herein, we disclose our medicinal chemistry efforts to transform 1
into molecules with high systemic bioavailability and an improved balance of safety and efficacy
suitable for further exploration.
RESULTS AND DISCUSSION
As discussed above, compound 1 was identified from a cellꢀbased HTS. Subsequent
localization of the inhibition point in protein translation informed the development of a high
capacity luciferase reporter assay (PCSK9 ProLuc) to monitor the translation of a truncated
human PCSK9 (1ꢀ152) construct within transfected Huh7 cells as a surrogate for the native
protein. This was the primary assay initially used to screen for compounds with improved
pharmacology. The molecular weight and LogD of 1 were acceptable, as were permeability
(RRCK) and safety (dofetilide binding, CYP inhibition, cell toxicity) assessments (see Table 1).
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 53
1
2
3
4
5
6
7
8
9
However, the intrinsic clearance of 1 in rat liver microsomes (RLM) was high, when compared
to human, and represented an initial hurdle in the development of a viable rodent tool compound.
The initial optimization process was therefore designed to improve potency and reduce clearance
of compounds through the careful modulation of lipophilicity, while maintaining the favorable
properties of 1. A screening cascade was implemented to facilitate identification of an orally
bioavailable tool compound suitable for in vivo studies. Parallel assessments of activity (PCSK9
ProLuc), rat and human liver microsomal (HLM) stability, and in vitro safety were used to triage
compounds.
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 1. Pharmacological, Absorption, Metabolism, and Safety Properties of 1
MW: 389.1, LogDa: 1.1, LipEb: 4.5
PCSK9 ProLuc IC50 (ꢁM)
Cell Toxicity (ꢁM)c
2.63
> 20
RRCK (Papp, 10ꢀ6 cm/sec)d
RLM Clint (ꢁL/min/mg)
HLM Clint (ꢁL/min/mg)
CYP (% inh at 3 ꢁM)
DOF Ki (ꢁM)e
5.4
> 510
20.6
1A2, 2D6, 2C9, 3A4 < 20 %
13
ACS Paragon Plus Environment
6

Page 7 of 53
Journal of Medicinal Chemistry
1
2
3
4
aLogD determined by shake flask method. bLipE = ꢀLogIC50PCSK9 ProLuc – LogD. cCellTiterꢀ
5
6
d
Glo® ATP viability assay. RRCK = Modified MadinꢀDarby permeability assay.20 eDOF =
7
8
dofetilide fluorescence polarization binding assay.
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
Although the chemical series represented by 1 appeared highly amenable to rapid analogue
generation, synthetic modifications were initially found to be quite challenging. The common
obstacle to all analogues was the formation of the central tertiary amide bond. Compound 1 was
divided into three fragments: the heterocyclic head group (magenta), the acyl linker (blue), and
the amine (black), for the purpose of exploring the SAR. Mindful of the desire to rapidly explore
the chemical space around this chemotype, we developed three complementary synthetic
methods to expand the SAR via singleton and parallel synthetic methods (see Scheme 1). In
Route A, an Abramovitch rearrangement21 of amide 4 and Nꢀoxide 5 was performed to afford 6,
which was subsequently deprotected under acidic conditions to give 7. This route was useful for
initial SAR generation where the amine and amide fragments could be readily explored by
parallel chemistry starting from isoquinoline Nꢀoxide and diverse sets of monoꢀBoc diamines and
carboxylic acids. No compounds more potent than 1 were found (data not shown) so further
analogue work was restricted to use of amine 2. It was soon determined that isoquinolineꢀNꢀ
oxide was a uniquely competent reaction partner in the Abramovitch rearrangement, and that
diversification to monoꢀheterocyclic Nꢀoxides was complicated by low conversions and
regiochemical mixtures of products. As such, a threeꢀstep sequence (Route B) was developed
that would allow for more significant variation. Palladiumꢀmediated coupling of 2 with 2ꢀ
bromopyridine 8 afforded amine 9b, which was then coupled with various carboxylic acids to
give analogues 7, after Bocꢀdeprotection. This sequence, however, was less tolerant to diverse
substrates than anticipated. The initial BuchwaldꢀHartwig amination required significant
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 53
1
2
3
4
5
6
7
8
9
optimization and was found to be highly substrateꢀspecific. Furthermore, amines 9a, 9b were
found to be a poor amidation partners, presumably due to steric hindrance and their electronꢀ
deficient nature. Ultimately, Route B proved to be most useful for gram scale preparation, where
varied substrates were not utilized. Route C was the most successful diversityꢀoriented approach
for analogue generation. In this sequence, Bocꢀprotected amine 2 underwent an SNAr reaction
with 2ꢀhaloꢀazineꢀNꢀoxide 10 to give 11, which was directly coupled with 3 to yield 12. The Nꢀ
oxide and Bocꢀprotecting group were then readily removed under mild conditions22 to afford the
desired analogues 7. In this case, the SNAr reaction is electronically enhanced by the Nꢀoxide,
which further enables the amide coupling of 11 and 3 via an intraꢀmolecular activation
mechanism.23
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Synthetic Approaches to Tertiary Benzamidesa
NH2
BocN
Route A:
H
N
5
N
O
O
O
2
BocN
R1
R1 OH
a
b
R2 =
N
or
N
Cl
3
4
Route B:
R2
R2
N
R2
N
2
3
NH
O
O
N
c
BocN
BocN
HN
e
Cl
d
R1
R1
Br
8
9a R2 = 1-isoquinolinyl
9b R2 = 3-chloro-2-pyridinyl
6
7
Route C:
N
N
O
Cl
O
Cl
O
N
O
2
3
NH
N
Cl
BocN
BocN
f
g
h, e
Cl
R1
11
12
10
aReagents and conditions: (a) 1,1’ꢀcarbonyldiimidazole, DMF, 30 oC. (b)
trifluoromethanesulfonic anhydride, 2,6ꢀlutidine, DCM, ꢀ30 to 25 oC. (c) chloro(diꢀ2ꢀ
norbornylphosphino)(2ꢀdimethylaminoferrocenꢀ1ꢀyl) palladium (II), KOtBu or NaOtAm,
o
o
o
toluene, 105ꢀ110 C. (d) 3, oxalyl chloride, DCM, 0ꢀ25 C, then LHMDS, THF, 25 C or
ACS Paragon Plus Environment
8

Page 9 of 53
Journal of Medicinal Chemistry
1
2
3
4
o
o
iPr2EtN, DMAP, THF, 65 C. (e) HCl in 1,4ꢀdioxane or TFA in DCM, 25 C. (f) iPr2EtN,
5
6
o
DMAP, CsF, nBuOH, 100 C. (g) bromotripyrrolidinophosphonium hexafluorophosphate
7
8
(PyBroP), iPr2EtN, THF. (h) Et3SiH, MeOH, 10% PdꢀAl2O3.
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
Using a parallelꢀenabled version of the chemistry developed for route A, the amide vector of 1
was optimized using a diverse set of readily available carboxylic acids. Of note was the
discovery of the weakly active biphenyl analogue 7a (Table 2). Most significantly, this
modification conferred improved intrinsic clearance, but at the expense of potency. A concise
library protocol (Scheme 2) was therefore developed to quickly assess the SAR of more diverse
paraꢀheterocyclic benzoyl amides, with the goal to mitigate this deficiency. A straightforward
Suzuki coupling of heterocyclic halides with boronate ester 14a followed by Bocꢀdeprotection,
afforded several new analogues. Compounds 7b, 7c and 7d had significantly improved profiles,
with the latter being the first analogue with both subꢀmicromolar activity (PCSK9 ProꢀLuc
assay) and favorable ADME properties. Further potency improvements were achieved with Nꢀ
triazolopyridine variant 7e, which was synthesized from the appropriate benzoic acid 20 obtained
via the sequence found in Scheme 3.
Scheme 2: Library-Expansion of the Benzamide Vectora
aReagents and conditions: (a) Bis(pinacolato)diboron, KOAc, PdCl2(dppf), 1,4ꢀdioxane, 85 oC.
(b) ArX, Cs2CO3, PdCl2(dppf), 1,4ꢀdioxane, 85 oC. (c) HCl, 1,4ꢀdioxane, 25 oC.
Scheme 3. Synthesis of varied N-triazolo-biaryl acidsa
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 53
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (a) Pd(OAc)2, K2CO3, (±)ꢀ2,2′ꢀbis(diphenylphosphino)ꢀ1,1′ꢀ
binaphthalene, NaI, toluene, 110 oC. (b) RaneyꢀNi, H2, EtOH, 25 oC. (c) NaNO2, AcOH, H2O, 0ꢀ
25 oC, (d) NaOH, MeOH, 60 oC.
Initial attempts to further improve the potency and ADME properties of the compounds
through modification of the isoquinoline moiety were not fruitful. Briefly, insertion of nitrogen
atoms into the bicyclic ring (i.e. isomeric naphthyridines) or monoꢀsubstitution of the 5, 6, 7 or
8ꢀposition with small substituents (e.g. F, Me, OMe) provided no potency or ADME advantage
(data not shown). Thus, our attention was directed towards an isosteric replacement for the
isoquinoline moiety. As shown in Table 2, 3ꢀchloropyridine served this purpose well. The direct
comparator 7f to isoquinoline 7e was equipotent, sufficiently stable in liver microsomes, and
exhibited modestly improved LipE (7f = 6.0 vs 7e = 5.4) suggesting that future analogues would
benefit from this modification. Distomer 7g was also prepared, but it was significantly less
potent. The greater potency for Rꢀenantiomers is consistent with previous findings associated
with 1. Therefore, all subsequent analogues used (R)ꢀ3ꢀaminopiperidine 2 as their starting points.
ACS Paragon Plus Environment
10

Page 11 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
As previously reported,18 compound 7f was advanced into in vivo safety and efficacy studies,
which are briefly summarized here. During the 14 d dosing period, 7f was tolerated when orally
administered at doses of 5, 15, and 50 mg/kg daily to male Sprague Dawley rats. A reduction in
total cholesterol (30%) and LDL (58%) was observed at 50 mg/kg/day. However, decreased
cellularity of the bone marrow also occurred at this dose that was associated with decreases in
hematocrit and hemoglobin (~9ꢀ10%), decreased total nucleated cell counts (30%), absolute
maturing (48%), proliferating (32.9%), and total (46%) erythroid cells, decreased white blood
cells (52%), lymphoyctes (54%), monocytes (63%), eosinophils (54%) and basophils (100%), as
well as decreased total T cells (54%), helper T cells (54%), cytotoxic T cells (54.5%), and B
cells (58%). In addition, minimal single cell necrosis of the crypt cells in the ileum was observed
at 50 mg/kg/day. No doseꢀlimiting changes were noted at the 5 and 15 mg/kg dose level when
compared to vehicle.
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
In consideration of these toxicology findings, which were not evident in the upfront CellTiterꢀ
Glo® ATP viability assay, in vitro bone marrow cellular toxicity assays for rat (initial toxicology
species, rat lineage (ꢀ) cells) and human (CD34+ cells)19 were developed to screen for future
compounds devoid of bone marrow effects. Indeed, when 7f was assessed in the rat lineage (ꢀ)
assay, it reduced cell viability (IC50 = 1.2 ꢀM) within threeꢀfold margin of its functional activity
(PCSK9 ProꢀLuc IC50 = 0.38 ꢀM). As such, all new compounds would be assessed for rat bone
marrow toxicity alongside assays in the existing screening funnel.
Without a clear understanding of the SAR driving bone marrow toxicity, we initiated point
mutations based on structure 7f. A methyl group position scan of the triazolopyridine of 7f
resulted in the equipotent 5ꢀmethyl 7h and 6ꢀmethyl 7i analogues. Other isomeric methylated
variants (data not shown) were significantly less potent. Of note, 7i showed a significant
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 53
1
2
3
4
5
6
7
8
9
improvement in bone marrow toxicity (rat lineage (ꢀ) IC50 > 19.7 ꢀM) while 7h only had a twoꢀ
fold margin. This suggested that seemingly minor modifications to our lead series could be
fruitful in yielding improved molecules. The benzoyl moiety was also modified to give
picolinamide 7j and nicotinamide 7k analogues. Diazine variants of this ring led to compounds
with intrinsic microsomal clearances that were essentially reduced to the limit of assay detection,
presumably due to lowered LogD, but these changes generally resulted in lower potencies (data
not shown). The combination of picolinamide and the beneficial 6ꢀmethyl group addition led to
7l, which gratifyingly maintained favorable ADME and potency while exhibiting a >30ꢀfold
window between bone marrow toxicity and PCSK9 ProꢀLuc activity. Analogues containing
heterocycles other than triazolopyridine with sufficient potency and ADME properties were also
reꢀexamined using a similar single point mutation strategy (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. Potency and Properties of Selected Analogues
PCSK9
ProLuc
PCSK9
ELISA
DOF Ki
Cell Toxicity
IC50 (ꢀM)
HLM
Clint
RLM
Clint
Compound
R1
R2
LogDb LipEc
d
d
(ꢀM)e
IC50 (ꢀM)a IC50 (ꢀM)a
7a
54%h
4.30
39%h
9.32
2.5
0.7
2.4
4.7
10.7
11.0
57.7
123
2.2
ꢀ
7b
7c
A
A
>40.0
>20f
1.06
2.38
0.9
5.1
13.3
ꢀ
16.0
2.9f
ACS Paragon Plus Environment
12

Page 13 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7d
7e
A
A
0.69
0.39
1.29
0.90
1.1
5.4
14.5
31.2
5.4
1.4f
1.4
0.8
5.4
6.0
21.8
9.7
ꢀ
15.8
18.1
ꢀ
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
7f
0.38
0.82
18.2
1.2f
7g (S)
61%h
24%h
>20f
7h
7i
B
B
B
B
B
0.45
0.90
1.2
5.5
9.0
38.5
34.0
1.0f
0.41
0.71
0.78
1.44
1.2
0.3
5.6
6.0
8.8
16.5
15.3
3.3
>19.7f
>12.7f
<8.0
28.8
7j
7k
7l
4.10
0.85
9.24
1.37
0.1
0.7
5.6
5.7
<8.0
9.3
14.7
15.1
12.7
5.9
ꢀ
>20f/27.1g
7m
7n
B
B
ꢀ
ꢀ
2.97
2.82
1.37
1.37
ꢀ
ꢀ
18.2
<8.0
ꢀ
ꢀ
14.1
38.2
30.1g
43.7g
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 53
1
2
3
4
5
6
ꢀ
3.08
1.07
8.7
ꢀ
38.4
19.6g
7o
B
7
8
9
aValues are reported as geometric means (n > 3). Please see supporting information for the
number of experiments and associated error. bLogD determined by shake flask method. cLipE = ꢀ
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
d
e
LogIC50PCSK9 ProLuc–LogD. Intrinsic clearance ꢀL/min/mg. DOF = dofetilide fluorescence
polarization binding assay. CellTiterꢀGlo® rat bone marrow cell toxicity with puromycin as high
f
point effect (HPE) standard. CellTiterꢀGlo® rat bone marrow cell toxicity with 7f as the HPE
g
h
standard. Percent effect at 20 ꢀM.
In order to establish an in vitro/in vivo toxicological correlation to help differentiate the
analogues of interest, a group of compounds were selected for a sideꢀbyꢀside safety assessment
in rats at a 50 mg/kg dose. The 50 mg/kg dose is meaningful because this dose of 7f elicited the
in vivo effects on rapidly proliferating cells. Thus, 7f, 7g, 7i, 7j, and 7l were selected as they
represent analogues that span the ranges of potencies in both the PCSK9 ProꢀLuc and rat lineage
(ꢀ) assays (Figure 1).
puromycin
Rat Lineage (-) IC (ꢀM)
50
ACS Paragon Plus Environment
14

Page 15 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 1: Rat Lineage (ꢀ) IC50 vs. Human ProꢀLuc IC50. Summary of in vitro toxicity/potency
correlations. Markers are colored by RRCK (AB flux, Papp, 10ꢀ6 cm/sec) permeability, where red
< 2, yellow 2ꢀ10, and green > 10. Plusꢀshaped markers indicate compounds also assessed in vivo.
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
A 14ꢀday oral gavage toxicology study in male rats (5/group) administered vehicle or one of
five different compounds at a dose of 50 mg/kg was conducted. Mortality, clinical signs and
hematological endpoints were monitored throughout the course of the study. Toxicokinetic
analysis was performed at 1, 3, 6, and 24 h postꢀdose on day 14. Importantly, after necropsy on
day 15, key microscopic evaluations of bone marrow, ileum, duodenum, jejunum, and liver were
conducted. Based on AUC (ng h/mL), the exposure was within fourꢀfold for all compounds
•
dosed. Although decreased bone marrow cellularity and crypt cell necrosis previously observed
in rats administered 7f were also observed in this study, no other compounds tested in this study
had these findings. A discolored/mottled liver was observed with all compounds tested.
Hepatocyte vacuolation was observed in animals given 7i, which was not considered dose
limiting. These results confirmed that compounds with rat lineage (ꢀ) assay IC50 values greater
than 7f had no evidence of bone marrow or crypt cell toxicity. The results are depicted
graphically in Figure 2.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 53
1
2
3
4
5
7g
g
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
7l
7i
7j
7f
Rat Lineage (-) IC50 (ꢀM)
Figure 2: Rat Lineage (ꢀ) IC50 vs. Human ProꢀLuc IC50. Summary of in vitro toxicity/potency
correlations. Markers are colored by white blood cell toxicity where red is white blood cell toxic
and green is not white blood cell toxic. Shapes are sized by area under the curve (AUC) for the
total drug exposure vs. time.
After having established an in vitro/in vivo toxicology correlation, changes to the screening
funnel were instituted to further refine compound selection for additional in vivo work. Two
changes were made aimed at defining more stringent selection criteria: (1) use of an enzymeꢀ
linked immunosorbent assay (ELISA) (PCSK9 ELISA) to assess potency and (2) use of
compound 7f as the high point effect (HPE) standard in the rat lineage (ꢀ) assay. It was
anticipated that the former change would provide a more realistic assessment of PCSK9
inhibition since it employed the native, human fullꢀlength PCSK9 construct. In general, the two
primary pharmacology assays correlated well, especially for the more potent compounds. For
example, compare 1 (PSCK9 ProLuc IC50 = 2.63 ꢀM vs. PCSK9 ELISA IC50 = 8.37 ꢀM) and 7f
PSCK9 ProLuc IC50 = 0.38 ꢀΜ vs. PCSK9 ELISA IC50 = 0.82 ꢀΜ). The latter change was
ACS Paragon Plus Environment
16

Page 17 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
instituted to reflect the correlation between in vitro and in vivo toxicity for 7f. For example,
compare 7l, rat lineage (ꢀ) IC50 (puromycin HPE) > 20 ꢀM vs. rat lineage (ꢀ) IC50 (7f HPE) =
7
8
9
27.1 ꢀM.
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
On the basis of these newly delineated criteria, 7l emerged as a compound with the best
balance of potency in the PCSK9 ELISA assay, minimal toxicity in the rat lineage (ꢀ) assay and
tolerability in vivo. However, it still had an undesirable ion channel profile as judged from low
micromolar activity in the dofetilide binding assay. This was confirmed with hERG patch clamp
data (hERG IC50 = 3.56 ꢁM). Therefore, we evaluated a wider range of compounds that could
lead to compounds that maintained the desirable qualities, have an increased window versus
bone marrow toxicity and an improved hERG profile. On the basis of the favorable dofetilide
profile of 7b, pendent pyrazines were further explored that resulted in identification of
compounds 7m and 7n. Both of these compounds showed lower bone marrow toxicity and were
less potent in the hERG patch clamp assay (7m, hERG IC50 = 43.2 ꢁM; 7n, hERG IC50 = 59.8
ꢁM). In anticipation of in vivo safety studies, rat and nonꢀhuman primate PK studies were
performed. Of note, the presence of an active metabolite 7o was detected following
administration of 7m to rats. Despite negligible levels of 7o detected in nonꢀhuman primate (in
vitro and in vivo) and human (in vitro), the increased potency in the rat lineage (ꢀ) cell assay and
the demonstrated correlation of this assay to rat in vivo toxicity precluded 7m from further
consideration.
On the basis of the promising profiles of 7l and 7n, extensive ADME and safety analyses were
performed in anticipation of advancing an in vivo tool. The data in comparison to former lead
compound 7f is shown in Table 3.
Table 3. Pharmacological, ADME and in vitro Safety Properties of 7f, 7l and 7n
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 53
1
2
3
4
5
6
7
8
9
7f
7l
7n
407.9
2.0
0.8
ꢀ
MW
433.9
2.1
448.9
1.5
clogP
LogDa
0.8
0.7
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
Cell Toxicity (ꢀM)b
RRCK (Papp, 10ꢀ6 cm/sec)c
MDR1ꢀMDCK AB
(10ꢀ6 cm/sec)
> 20
15.0
1.5
> 20
6.2
1.3
10.5
0.7
MDR1ꢀMDCK BA
(10ꢀ6 cm/sec)
20.6
19.9
20.0
CYP IC50 (ꢁM) or
% inhbition at 3 ꢁM
Fraction unbound
(rat plasma)
1A2, 2D6, 2C9, 3A4 1A2, 2D6, 2C9, 3A4 1A2, 2D6, 2C9, 3A4
> 30 ꢁM
20%
all < 13%
24%
> 30 ꢁM
41%
hERG IC50 (ꢁM)d
Cerep panel of 72 assays
screened at 10 ꢀM
8.5
3.6
59.8
3 targets with
2 targets with
Ki/IC50 < 10 ꢀM
1 target with
Ki/IC50 < 10 ꢀM
Ki/IC50 < 10 ꢀM
aLogD determined by shake flask method. CellTiterꢀGlo® ATP viability assay. RRCK =
Modified MadinꢀDarby permeability assay. dhERG = human EtherꢀaꢀgoꢀgoꢀRelated Gene patchꢀ
clamp assay.
b
c
Compounds 7l and 7n were assessed in advanced in vivo safety studies. Of note, 7l was
assessed in a 14ꢀday rat toxicology study at doses of 50, 150 and 500 mg/kg/day. Although
mortality and doseꢀlimiting toxicity was observed at ≥ 150 mg/kg/day, 50 mg/kg/day was
ACS Paragon Plus Environment
18

Page 19 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
tolerated. This provided a 10ꢀfold improvement in exposure (total, Cave) over 7f. For 7n, bile duct
inflammation (at ≥ 50 mg/kg/day), degeneration/regeneration of duodenal or jejunal crypt cell
epithelium (at ≥ 150 mg/kg/day) and mortality (at 500 mg/kg/day) were observed in a 14ꢀday
toxicity study in rats, and prolongation of PR and QRS intervals in a cardiovascular study in
telemetered cynomolgus monkeys (at 75 mg/kg/day) excluded 7n from further study. Despite the
unfavorable hERG profile of 7l, it emerged as the compound with the best pharmacological
profile. Therefore, we sought to understand the in vitro selectivity profile of 7l versus the
previously profiled 7f and assess its in vivo activity. Of note, using the previously described
ribosome binding assay that employs [3H]ꢀ7f as the radioligand,18,19 7l was found to bind to the
80S isolated ribosomes more weakly than 7f (Table 4) in spite of comparable activity for the
inhibition of PCSK9.
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: Toxicology Summaries for 7f vs 7l
PCSK9
ELISA
Rat Lin(ꢀ) IC50/
Cmpd Human CD34+
IC50
80S
Hematopoietic Total Cmax Total Cave NOAEL Total
Ki (ꢀM)
cell findings (ng/mL)
(nM)
Cave exposure
IC50 (ꢀM)
direct
1640
50 mg/kg
478 nM
(15 mg/kg)
4231 nMa
(50 mg/kg)
2.9 ꢁM/2.7 ꢁM
7 ꢁM
0.82
1.37
2419
7f
7l
indirect
9660
150 mg/kg
*65 ꢁM/>20 ꢁM 63 ꢁM
15869
*puromycin (HPE). afree Cave = 1015 nM based on 24% rat plasma protein binding.
In an attempt to understand the difference in overall in vivo toxicity and ribosome binding
profiles, ribosome profiling in rat intestinal epithelial cells (IECꢀ6) was performed. We chose
this cell line because of the histopathological intestinal findings with 7f. IECꢀ6 cells were found
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 53
1
2
3
4
5
6
7
8
9
to have a differential toxicity response to 7l and 7f based on a CellTiterꢀGlo® assay with a high
divergence at 4 µM (SI Figure 2). IECꢀ6 cells were treated with 4 µM 7l or 7f for 1 h and direct
offꢀtargets were identified using ribosome profiling and our previously described analysis
method.18 In this approach, the cumulative fraction of reads along the annotated protein regions
are plotted for the compound treatment and vehicle control. If a compound induced stall occurs,
these two plots diverge (SI Figure 2). We define the magnitude of the maximum divergence as
Dmax and the codon at which the maximum divergence occurs as the Dmax position. If a given
protein coding region has a stall, the Dmax position will be 3’ to the stall. For genes with a Dmax
value greater than three standard deviations from the mean, reads 3’ to the Dmax position are
extracted and used for differential expression analysis with DESeq. Our previous ribosome
profiling study with 7f demonstrated that the direct effects of compound on translation are
observable using ribosome footprint data alone with no changes in transcript levels as measured
by mRNA seq18 so we did not include mRNAꢀseq in this current study.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Fiftyꢀtwo IECꢀ6ꢀexpressed nascent chains were found to be sensitive to 7f with a DESeq false
discovery rate of 0.1 and a twoꢀfold decrease in reads mapping 3’ to the Dmax position. Eleven
7lꢀsensitive nascent chains were detected with the same criteria (Figure 3, SI Table 1). Eight of
the 7lꢀsensitive nascent chains were also sensitive to 7f while Ndc80 and Cox15 were only
detected for 7l. All of the offꢀtargets that were identified using one compound were also
downregulated by the other compound, albeit not in all cases to the level of a twoꢀfold change
and/or an individual false discovery rate of 0.1. With the exceptions of Ndc80 and Cox15, all of
the sensitive proteins were more affected by 7f than 7l (Figure 3 CꢀD). Additionally, 49 genes
were differentially regulated (false discovery rate < 0.1) upon treatment with 7f but did not
display a buildup of reads indicative of a stall. Most of these displayed an expression level
ACS Paragon Plus Environment
20

Page 21 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
change in the same direction but with a lower magnitude with 7l treatment (SI Figure 2). These
indirectlyꢀaffected proteins likely represent a downstream cellular response. These data indicate
that 7f and 7l target the same group of direct offꢀtarget proteins with 7l having a much lower
effect on all sensitive nascent chains except Ndc80 and Cox15 (Figure 3).
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
21

Journal of Medicinal Chemistry
Page 22 of 53
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. 7l affects a smaller portion of the translatome than 7f. Volcano plots of ribosome
profiling data from (3A) 4 µM 7f and (3B) 4 µM 7l. The points in red indicate genes with a Zꢀ
scoreꢀtransformed Dmax > 2 in either the 7f or 7l datasets. Direct offꢀtargets (large red circles)
are defined as having a potential buildup of reads to indicate a stall as indicated by a Zꢀscore
transformed Dmax > 2, a false discovery rate < 0.1 and at least a twoꢀfold change. (3C)
Heatmap displaying the levels of downregulation of compoundꢀsensitive nascent chains. (3D)
Scatterplot of protein translation levels in reads per kilobase per million (RPKM) total reads
ACS Paragon Plus Environment
22

Page 23 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
following treatment with 4 µM 7f or 4 µM 7l. The directly affected proteins listed in Figure 3C
are represented by a larger red spot. The majority of the directly affected proteins display a lower
translation level in the 7f than the 7l condition. RPKM values are calculated based on reads
aligning 3’ to the Dmax position of the Dmax Zꢀscore is greater than 2.0 or 3’ to codon 50
otherwise.
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 improved rat safety profile for 7l prompted its assessment of in vivo efficacy in the
humanized PCSK9 mouse model. Figure 4 shows relative plasma PCSK9 levels, normalized to t
= 0 min, for vehicle, 30 mg/kg and 100 mg/kg doses over 0ꢀ24 h. The entire time course was
used for determining the concentrationꢀresponse relationship using an indirect effect PK/PD
model. The free Ceff for PCSK9 lowering by 7l was determined to be 175 nM. The ribosome
profiling data combined with the improved in vivo safety profile suggested that 7l was not only
more selective, but also a more tolerated variant of 7f.
2.0
Vehicle
30 mg/kg
1.5
1.0
0.5
0.0
100 mg/kg
0
10
20
30
Time (h)
Figure 4: Plasma PCSK9 Levels in hPCSK9 Mice Following Single Oral Dose Administration
of 7l. Data for the 300 mg/kg dose not shown (see SI).
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 53
1
2
3
4
CONCLUSION
5
6
7
8
We have described an orally active small molecule inhibitor of PCSK9 synthesis designed
for systemic exposure that demonstrates an improved in vivo safety profile. Crucial to the
expansion of the SAR around progenitor 1 was the development of synthetic methodologies to
address the encumbered tertiary amide at the center of the molecule. These chemistries made it
possible to efficiently explore alternatives to the isoquinoline and the dihydrocinnamate groups.
Libraryꢀenabled methods allowed for exploration of each vector. Minimal changes to the head
group and amine moiety were tolerated but small changes to the tail fragment resulted in
compounds with similar potencies and improved in vivo toxicology profiles. Interestingly,
ribosome profiling indicates a change in the degree of stalling of offꢀtarget transcripts rather than
a significantly altered selectivity profile. While these data are derived from one cell type and it
remains unclear whether this pattern is repeated globally to account for the improved overall
safety profile. Based on the mode of action for these compounds, a clear challenge for future
work will be ensuring sufficient selectivity for inhibition of PCSK9 mRNA translation. A greater
understanding of the determinants of translational selectivity will likely be necessary.
EXPERIMENTAL SECTION
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
All procedures performed on animals in this study were in accordance with established
guidelines and regulations, and were reviewed and approved by Pfizer Institutional Animal Care
and Use Committee. Pfizer animal care facilities that supported this work are fully accredited by
AAALAC International.
Materials and Methods: All chemicals, reagents and solvents were purchased from
commercial sources and used without further purification. Compound 2 (CAS # 188111ꢀ79ꢀ7)
was purchased from CNH Technologies (catalog # Cꢀ3103R), Woburn, MA. Compound 7f was
ACS Paragon Plus Environment
24

Page 25 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
synthesized as previously reported.18 All reactions were performed under an atmosphere of
nitrogen unless otherwise noted. Nuclear magnetic resonance spectra (1H, 13C NMR) were
recorded with 400 MHz, 500 MHz or 600 MHz Varian or Bruker spectrometers. Chemical shifts
are expressed in parts per million downfield from tetramethylsilane. The peak shapes are denoted
as follows: s, singlet; d, doublet; t, triplet; q, quartet; m, multiplet; br s, broad singlet. Due to
differences in solvents used in sample preparation and, in some cases amount of water present,
not all exchangeable protons were observable. Mass spectrometry (MS) was performed via
atmospheric pressure chemical ionization (APCI) or electron scatter (ESI) ionization sources.
Silica gel chromatography was performed using a medium pressure Biotage or ISCO system
using columns preꢀpackaged by various commercial vendors including Biotage and ISCO.
Whatman preꢀcoated silica gel plates (250 ꢁm) were used for analytical TLC. Microanalyses
were performed by Intertek Pharmaceutical Services and were within 0.4% of the calculated
values. All library compounds were > 95% purity as judged by LCMS (UV 254 nm or ELSD
detection).
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
Synthesis
General Procedure A: (Corresponding to Scheme 1, Route A, parallel format). Step 1: To a
solution of carboxylic acid (3, 125 ꢁmol, 1.0 equiv) in 250 ꢁL DMF was added a solution of 1,
o
1’ꢀcarbonyldiimidazole (250 ꢁL, 0.5 M in DMF). The mixture was shaken at 30 C for 1.5 h,
then a solution of tertꢀbutyl (3R)ꢀ3ꢀaminopiperidineꢀ1ꢀcarboxylate (2, 250 ꢁL, 0.5 M in DMF)
o
was added. The reaction was shaken at 30 C until complete by LCMS (typically 16 h). The
solvents from the crude reaction mixture were evaporated to dryness by Speedvac, the residue
was dissolved in DCM (3.0 mL) and washed with water (2 × 3 mL) and brine (3 mL). The
organic phase was dried over Na2SO4, filtered and concentrated by Speedvac to give the amide
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 53
1
2
3
4
5
6
7
8
intermediates that were used directly in the next step. Step 2: In a glove box, 2,6ꢀlutidine (32 ꢁL,
276 ꢁmol, 2.2 equiv) was added to a solution of amide (4) from Step 1 in 500 ꢁL DCM. The
reaction was cooled to ꢀ30 oC, followed by dropwise addition of a solution of
9
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
trifluoromethanesulfonic anhydride (275 ꢁL, 0.5 M in DCM). The reaction was stirred at ꢀ20 C
for 1.5 h, followed by the addition of a solution of isoquinoline Nꢀoxide 5 (200 ꢁL, 0.69M in
o
DCM). The reaction was removed from the glove box. The reaction was shaken at 30 C until
complete by LCMS (typically 16 h). The solvents from the crude reaction mixture were
evaporated by Speedvac and the residue was purified by preparative HPLC to give the NꢀBoc
intermediates (6). Step 3: A solution of TFA/DCM (2.0 mL, v/v = 1/4) was dispensed into the
o
purified NꢀBoc intermediates from Step 2. After shaking for 1 h at 30 C, the solvents were
evaporated by Speedvac. The crude material was purified by reversed phase HPLC to afford the
desired products (7).
General Procedure B: (Corresponding to Scheme 1, Route B, parallel format). Step 1: To
the carboxylic acid 3 (0.21 mmol, 1.0 equiv) in DCM (2 mL) was added two drops of DMF
followed by oxalyl chloride (21 ꢁL, 0.25 mmol). The reaction was stirred at rt for 2 h and then
concentrated. Step 2: The crude acid chloride from Step 1 (~0.21 mmol) was added to a solution
of amine 9 (0.11 mmol) in THF (2 mL), iPr2EtN (96 ꢁL, 0.54 mmol) and DMAP (2 mg). After
heating overnight at 65 oC, the reaction was cooled and poured into saturated NaHCO3, extracted
with EtOAc, washed with brine, dried (Na2SO4) and concentrated to afford 6. Step 3: The
residue was dissolved in DCM (3 mL) and treated with TFA or HCl (1 mL). After 1 h, the
mixture was concentrated. The crude material was either purified by reversed phase HPLC or
triturated with diethyl ether (if sufficiently pure after the previous step) to afford the desired
ACS Paragon Plus Environment
26

Page 27 of 53
Journal of Medicinal Chemistry
1
2
3
4
products (7).
5
6
7
8
9
General Procedure C: (Corresponding to Scheme 1, Route C, parallel format). Step 1:
Bromotripyrrolidinophosphonium hexafluorophosphate (PyBroP, 112 mg, 240 ꢁmol, 1.6 equiv),
Nꢀoxide 11 (150 ꢁmol, 1.0 equiv) and carboxylic acid 3 (180 ꢁmol, 1.2 equiv) were dissolved in
1 mL THF. After addition of iPr2EtN (70 ꢁL, 450 ꢁmol, 3 equiv), the mixture was shaken at 30
oC. After 16 h, the solvent was removed by Speedvac and the residue was submitted to purify by
preparative TLC to afford the intermediate (12). Step 2: Intermediate 12 (~150 ꢁmol, 1.0 equiv)
was dissolved in 1 mL MeOH. PdꢀAl2O3 (12.5 mg, 10% Pd basis, 12.5 ꢁmol, 0.1 equiv) and
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
triethylsilane (158 ꢁL, 1200 ꢁmol, 8.0 equiv) were then added. The mixture was shaken at 30 C.
After 5 h, the reaction mixture was filtered and the filtrate was concentrated by Speedvac to give
crude intermediate, which was used in the next step. Step 3: A solution of TFA/DCM (1.0 mL,
v/v = 1/4) was dispensed into the crude reaction mixture from Step 2. After shaking for 2 h at 30
oC, the solvents were evaporated by Speedvac. The crude material was purified by reversed
phase HPLC to afford the desired products (7).
General Procedure D: (Corresponding to Scheme 2, parallel format). To a solution of the
pinacol boronate 14 (1.0 equiv) in 1,4ꢀdioxane (0.10 M) was added the aryl/heteroaryl halide
(100 ꢁmol, 1.0 equiv) followed by Cs2CO3 solution (200 ꢁL, 2.0 equiv, 1.0 M) and Pd(dppf)Cl2
(2 ꢁmol, 0.02 equiv) to each vial under N2 atmosphere. The reaction was heated under nitrogen
o
at 100 C until complete by LCMS (typically 8ꢀ16 h). The solvents from the crude reaction
mixture were evaporated by Speedvac. Step 2: A solution of TFA/DCM (2.0 mL, v/v = 1/10)
o
was dispensed into the crude reaction mixture. After shaking for 2 h at 30 C, the solvents were
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 53
1
2
3
4
5
6
evaporated by Speedvac. The crude material was purified by reversed phase HPLC to afford the
desired products (7).
7
8
9
(R)-N-(isoquinolin-1-yl)-N-(piperidin-3-yl)-[1,1'-biphenyl]-4-carboxamide (7a). Prepared
according to General Procedure A with [1,1'ꢀbiphenyl]ꢀ4ꢀcarboxylic acid. Purified by preparative
HPLC (Kromasil Eternityꢀ5ꢀC18 150 × 30 mm × 5 ꢁm; 45ꢀ75% acetonitrile: water (0.2% formic
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
acid)) to afford 7a (13.4 mg, 21%). tR = 2.80 min, m/z = 408 [M+H]+
.
(R)-N-(isoquinolin-1-yl)-N-(piperidin-3-yl)-4-(pyrazin-2-yl)benzamide
(7b).
Prepared
according to General Procedure B with 4ꢀ(pyrazinꢀ2ꢀyl)benzoic acid and tertꢀbutyl (R)ꢀ3ꢀ
(isoquinolinꢀ1ꢀylamino)piperidineꢀ1ꢀcarboxylate (9a). Purified by trituration from diethyl ether
1
to afford 7b (36 mg, 35%). H NMR (400 MHz, MeOHꢀd4) δ 8.89 (br s, 1H), 8.56 (br s, 1H),
8.34−8.52 (m, 2H), 7.94−8.19 (m, 2H), 7.57−7.87 (m, 5H), 7.36 (br s, 2H), 4.52−5.05 (m, 1H),
3.45−3.58 (m, 1H), 3.12−3.26 (m, 1H), 2.91 (d, J = 10.5 Hz, 2H), 2.26−2.64 (m, 1H), 1.77−1.91
(m, 1H), 1.64 (br s, 2H). LCMS (Method 4) tR = 0.876 min, m/z = 410.0 [M+H]+
.
(R)-N-(isoquinolin-1-yl)-4-(1-methyl-1H-pyrazol-4-yl)-N-(piperidin-3-yl)benzamide
(7c).
Prepared according to General Procedure D with (R)ꢀ3ꢀ(Nꢀ(isoquinolinꢀ1ꢀyl)ꢀ4ꢀ(4,4,5,5ꢀ
tetramethylꢀ1,3,2ꢀdioxaborolanꢀ2ꢀyl)benzamido)piperidineꢀ1ꢀcarboxylate (14a) and 4ꢀbromoꢀ1ꢀ
methylꢀ1Hꢀpyrazole. Purified by preparative HPLC (Grace Vydac C18 150 × 20 mm × 5 ꢁm; 45ꢀ
85% acetonitrile: water (0.2% formic acid)) to afford 7c (7.5 mg, 14%). tR = 2.34 min; m/z = 412
[M+H]+ Analytical HPLC Method 1: tR = 2.91 min, 99.1%.
.
ACS Paragon Plus Environment
28

Page 29 of 53
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(R)-N-(isoquinolin-1-yl)-N-(piperidin-3-yl)-4-(pyrazolo[1,5-a]pyrimidin-3-yl)benzamide
(7d). Prepared according to General Procedure D with 14a and 3ꢀbromopyrazolo[1,5ꢀ
a]pyrimidine. Purified by preparative HPLC (DIKMA Diamonsil(2) C18 200 × 20 mm × 5 ꢁm;
45ꢀ75% acetonitrile: water (0.1% formic acid)) to afford 7d (9.5 mg, 34%). tR = 2.45 min; m/z =
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
449 [M+H]+. H NMR (400 MHz, CD3OD) δ 8.86 (dd, J = 1.8, 7.0 Hz, 1H), 8.56 (dd, J = 1.8,
4.1 Hz, 1H), 8.48−8.53 (m, 1H), 8.42 (s, 1H), 7.99−8.14 (m, 1H), 7.77−7.91 (m, 4H), 7.68 (br s,
2H), 7.30 (d, J = 7.4 Hz, 2H), 7.01 (dd, J = 4.1, 7.0 Hz, 1H), 5.27 (br s, 1H), 3.91 (d, J = 14.9
Hz, 1H), 3.70−3.81 (m, 1H), 2.88 (br s, 2H), 1.74−2.16 (m, 3H), 1.32 (d, J = 17.2 Hz, 1H).
Analytical HPLC Method 2: tR = 3.04 min, 96.8%.
(R)-4-(3H-[1,2,3]triazolo[4,5-b]pyridin-3-yl)-N-(isoquinolin-1-yl)-N-(piperidin-3-
yl)benzamide (7e). Prepared according to General Procedure A with 4ꢀ(3Hꢀ[1,2,3]triazolo[4,5ꢀ
b]pyridinꢀ3ꢀyl)benzoic acid. Purified by preparative HPLC (Phenomenex Gemini C18 250 × 21
mm × 10 ꢁm; 45ꢀ85% acetonitrile: water (0.2% formic acid)) to afford 7e (2.5 mg, 4%). tR =
2.48 min; m/z = 450 [M+H]+
.
N-(3-chloropyridin-2-yl)-N-[(3S)-piperidin-3-yl]-4-(3H-[1,2,3]triazolo[4,5-b]pyridin-3-
yl)benzamide (7g). Prepared in analogous fashion to 7f (see SI) with (S)ꢀ3ꢀ[(3ꢀchloropyridinꢀ2ꢀ
yl)amino]piperidineꢀ1ꢀcarboxylate (13.7 g, 86%). [α] = +55.9o (c = 1.965, MeOH). Analytical
HPLC Method 3: tR = 2.47 min, 99.4%.
N-(3-chloropyridin-2-yl)-4-(5-methyl-3H-[1,2,3]triazolo[4,5-b]pyridin-3-yl)-N-[(3R)-
piperidin-3-yl]benzamide (7h). Prepared in analogous fashion to 7f starting with 9b and 4ꢀ(5ꢀ
ACS Paragon Plus Environment
29