Structure-Based Design of an Iminoheterocyclic β-Site Amyloid Precursor Protein Cleaving Enzyme (BACE) Inhibitor that Lowers Central Aβ in Nonhuman Primates

Article abstract of DOI:10.1021/acs.jmedchem.5b01995

We describe successful efforts to optimize the in vivo profile and address off-target liabilities of a series of BACE1 inhibitors represented by 6 that embodies the recently validated fused pyrrolidine iminopyrimidinone scaffold. Employing structure-based design, truncation of the cyanophenyl group of 6 that binds in the S3 pocket of BACE1 followed by modification of the thienyl group in S1 was pursued. Optimization of the pyrimidine substituent that binds in the S2′-S2″ pocket of BACE1 remediated time-dependent CYP3A4 inhibition of earlier analogues in this series and imparted high BACE1 affinity. These efforts resulted in the discovery of difluorophenyl analogue 9 (MBi-4), which robustly lowered CSF and cortex Aβ₄₀ in both rats and cynomolgus monkeys following a single oral dose. Compound 9 represents a unique molecular shape among BACE inhibitors reported to potently lower central Aβ in nonrodent preclinical species.

Full text of DOI:10.1021/acs.jmedchem.5b01995

Subscriber access provided by Flinders University Library
Article
Structure-Based Design of an Iminoheterocyclic beta-
Site Amyloid Precursor Protein Cleaving Enzyme (BACE)
Inhibitor that Lowers Central A# in non-Human Primates
Mihirbaran Mandal, Yusheng Wu, Jeffrey Misiaszek, Guoqing Li, Alexei V. Buevich, John P. Caldwell,
Xiaoxiang Liu, Robert D. Mazzola, Peter Orth, Corey Strickland, Johannes H. Voigt, Hongwu
Wang, Zhaoning Zhu, Xia Chen, Michael Grzelak, Lynn A. Hyde, Reshma Kuvelkar, Presscott
T. Leach, Giuseppe Terracina, Lili Zhang, Qi Zhang, Maria S. Michener, Brad Smith, Kathleen
Cox, Diane Grotz, Leonard Favreau, Kaushik Mitra, Irina Kazakevich, Brian A. McKittrick, William
J Greenlee, Matthew E. Kennedy, Eric M. Parker, Jared N Cumming, and Andrew W Stamford
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01995 • Publication Date (Web): 03 Mar 2016
Downloaded from http://pubs.acs.org on March 8, 2016
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.

Journal of Medicinal Chemistry
Structure-Based Design of an Iminoheterocyclic beta-Site
Amyloid Precursor Protein Cleaving Enzyme (BACE) Inhibitor
that Lowers Central Aβ in non-Human Primates
‡
‡
‡
‡
‡
Mihirbaran Mandal, ,Yusheng Wu, Jeffrey Misiaszek, Guoqing Li, Alexei Buevich, John P.
‡
‡
‡
‡
‡
‡
Caldwell, Xiaoxiang Liu, Robert D. Mazzola, Peter Orth, Corey Strickland, Johannes Voigt,
Hongwu Wang, Zhaoning Zhu, Xia Chen, Michael Grzelak, Lynn A. Hyde, Reshma Kuvelkar,
Prescott T. Leach, Giuseppe Terracina, Lili Zhang, Qi Zhang, Maria S. Michener, Brad Smith,
Kathleen Cox, Diane Grotz, Leonard Favreau, Kaushik Mitra, Irina Kazakevich, Brian A.
McKittrick, William Greenlee, Matthew E. Kennedy, Eric M. Parker, Jared N. Cumming, and
Andrew W. Stamford
‡
‡
§
§
§
§
§
§
§
§
∥
∥
#
#
#
#
⊥
‡
‡
§
§
*,‡
*,‡
‡
§
∥
Department of Global Chemistry, Department of Neuroscience, Department of Safety Assessment and
#
⊥
Laboratory Animal Research, Department of Discovery Pharmaceutical Sciences, and Department of
Pharmacokinetics, Pharmacodynamics, and Drug Metabolism
Merck Research Laboratories, 2015 Galloping Hill Road, Kenilworth, NJ 07033.
Abstract:
We describe successful efforts to optimize the in vivo profile and address off-target liabilities of
a series of BACE1 inhibitors represented by 6 that embodies the recently-validated fused pyrrolidine
iminopyrimidinone scaffold. Employing structure-based design, truncation of the cyanophenyl group of
6
that binds in the S3 pocket of BACE1 followed by modification of the thienyl group in S1 was
pursued. Optimization of the pyrimidine substituent that binds in the S2'-S2'' pocket of BACE1
remediated time-dependent CYP3A4 inhibition of earlier analogs in this series and imparted high
BACE1 affinity. These efforts resulted in the discovery of difluorophenyl analogue 9 (MBi-4) which
robustly lowered CSF and cortex Aβ in both rats and cynomolgus monkeys following a single oral
40
ACS Paragon Plus Environment

Page 1 of 58
Journal of Medicinal Chemistry
1
2
3
dose. Compound 9 represents a unique molecular shape among BACE inhibitors reported to potently
lower central Aβ in non-rodent preclinical species.
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
Introduction
Alzheimer’s disease (AD), is a progressive neurodegenerative disorder that affects an estimated
5
.3 million people in the United States and is responsible for 60 – 80% of all dementias, of which there
1
,2
were nearly 47 million worldwide in 2015.
In addition to the emotional and social burden that AD
exerts on patients and their families, the economic burden is enormous. In the US alone, the total cost of
1
care for patients with AD and other dementias was approximately US $226 billion in 2015. It is
expected that the number of people with AD will triple by 2050 resulting in a public health crisis unless
1
,2
disease modifying agents become available. Since current AD medications, the acetylcholine esterase
inhibitors and the N-methyl-D-aspartate receptor antagonist, memantine, only produce modest and
3
transient improvements in cognitive function, there is an urgent medical need to discover an agent that
slows the underlying neurodegenerative process. Major advances in the fields of genetics, cell biology,
biochemistry and neuroscience have provided some level of understanding of the pathophysiology of
4
AD. Two major neuropathological hallmarks of AD are extracellular amyloid plaques and intracellular
neurofibrillary tangles which are found predominantly in the cortex and hippocampus. Amyloid plaques
are composed of oligomerized and fibrillar β-amyloid (Aβ) peptides, primarily 42 amino acid residues in
length, which are derived from amyloid precursor protein (APP) through sequential cleavage of APP by
5
the beta-site APP cleaving enzyme 1 (BACE1) followed by γ-secretase. According to the amyloid
hypothesis, accumulation of Aβ peptides in the brain is causative of the underlying neurodegenerative
process of AD. Therefore, inhibition of Aβ peptide synthesis is a potentially disease-modifying
approach to the treatment of AD.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 58
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
Over recent years, BACE1 inhibition has emerged as the primary small molecule approach for
blocking Aβ peptide synthesis given the mechanism-based toxicity that attends γ-secretase inhibition,
6
resulting in dose-limiting side effects. Complementary approaches targeting central Aβ via anti-Aβ
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
monoclonal antibodies have also been tested in the clinic, but the effects on cognitive improvement
reported thus far either did not meet their target endpoints or, if they were significant, represented only a
7
minimal effect and require further trials. Taken together, the clinical data generated to-date with both
γ-secretase inhibitors and anti-Aβ antibodies do not reflect a definitive or even robust vetting of the
amyloid hypothesis.
In support of a focus on BACE1 inhibition to ultimately test this hypothesis, BACE1 knockout
8
(
KO) mice do not produce Aβ peptides, and a cross between these knockout mice and a transgenic
mouse model of early onset Alzheimer’s disease showed significant improvement in the overall disease
9
phenotype in regard to both reduction of amyloid levels and decreased memory deficits. BACE1 KO
8
mice were initially reported to possess no overt negative phenotype, but more recently loss of BACE1
function in these animals has been shown to result in subtle morphological and behavioral alterations,
10
and physiological non-APP substrates of BACE1 have been identified. Among the best characterized
of these substrates is neuregulin 1, both type III and type I. Phenotypes associated with reduced
proteolytic processing of these two substrates in BACE1 KO mice are a reduction of peripheral nerve
1
1
12
myelination and a reduction in muscle spindle formation, respectively. Importantly, these and many
phenotypes in homozygous BACE1 KO mice are not seen in BACE1 heterozygous mice and are not
replicated in adult wild type mice given selective BACE inhibitors. While our understanding of BACE1
biology continues to evolve, and careful monitoring of safety in BACE1 inhibitor clinical trials is
critical, the sum total of current data offers the potential for partial pharmacological inhibition of
BACE1 in adult patients to be a safe approach to disease modification in Alzheimer’s disease.
ACS Paragon Plus Environment

Page 3 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1: Discovery of bicyclic iminopyrimidinones 6, 7 and 8 from fragment 1 by rational design and
SAR development through iminohydantoins 2 and 3 and iminopyrimidinone 4.
Because BACE1 is an aspartyl protease, the early phase of BACE1 inhibitor design was
1
3
dominated by substrate-based peptidic and peptidomimetic approaches that had been successful in the
design of inhibitors of related aspartyl proteases, such as HIV protease and renin. These efforts resulted
in many potent inhibitors, but these generally lacked suitable molecular properties to achieve significant
14
brain penetration and in vivo CNS Aβ lowering activity at appropriate pharmacological doses.
Therefore, focus shifted to non-peptidic BACE1 inhibitors that would be more likely to possess
properties conducive to CNS activity.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 58
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
As shown in Figure 1, our work in this area began with the identification of isothiourea 1 through
1
5
a fragment based approach, and we have disclosed the progression of this fragment via lead compound
1
6
2
to high affinity brain penetrant and bioavailable inhibitors belonging to two distinct chemical series,
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
17
18
19
iminohydantoins (e.g. 3) and iminopyrimidinones (e.g. 4 and 5).
These efforts, independent of
2
0
others achieved through HTS and fragment based approaches, involved significant de novo design
elements with extensive application of X-ray crystallography and molecular modeling. In the
iminopyrimidinone series, additional SAR development that focused on structural diversity and
conformational constraint led to design of a novel bicyclic iminopyrimidinone scaffold represented by
21
compounds 6-7 and 8 (Figure 1).
Compound 6 reduced cortex Aβ in Sprague Dawley rats (-42% compared to vehicle) after
40
2
1
administration of a single dose (30 mg/kg, PO). Unfortunately, it had relatively high PSA that likely
contributed to modest brain penetration (brain/plasma 0.1), was a potent time dependent CYP3A4
inhibitor and showed significant hERG inhibition (70% @ 1 ꢀM). To overcome these issues, we
investigated truncation of the cyanophenyl moiety of 6 that binds in the BACE1 S3 pocket (shown
schematically in Figure 1). We concurrently explored SAR of the fluoropyrimidine group of 6 that
binds in the S2'-S2" pockets to rescue the associated loss of affinity and to address the time-dependent
CYP3A4 inhibition. Subsequent modification of the aryl group in S1 was pursued to further improve
the affinity and selectivity of the series. These efforts culminated in discovery of inhibitor 9 (MBi-4;
Figure 2) which showed robust pharmacodynamic A
β lowering activity in rats and monkeys. It is worth
noting that the vast majority of published small molecule BACE inhibitors that lower Aβ in monkey and
/
or dog, including all that have entered clinical trials, have substituents that occupy the S3 subsite of
22
BACE, and none of these structures occupy the S2'' subsite (Figure 2). In this way, our present work
validates a novel molecular shape, a fused pyrrolidine iminopyrimidinone, that only occupies S1 and
ACS Paragon Plus Environment

Page 5 of 58
Journal of Medicinal Chemistry
1
2
3
S2'-S2'', and thus achieves a similar overall profile with unique structural diversity. The discovery and
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
characterization of compound 9 will be the central focus of this article.
F
HN
HN
HN
HN
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
S
N
N
CHF2
N
HN
HN
N
F
O
N
F
Eli Lilly
(LY-2811376)22a
Astra-Zeneca
(AZD-3839)22b
N
F
F
N
N
NC
CF3
F
HN
HN
HN
S
N
H
N
O
O
H
F
HN
N
N
9
(MBi-4)
O
O
O
F
F
Eli Lilly
LY-2886721)22c
Novartis
(NB-360)22d
(
HN
HN
S
HN
O
NC
HN
N
H
F
O
O
N
HN
HN
N
N
F
F
O
O
F
O
F
F
F
Janssen22e
Amgen22f
Pfizer22g
Figure 2: Structures of representative published BACE inhibitors that lower central Aβ in monkeys
and/or dogs.
Chemistry
Synthetic routes to P1 thiophene analogs 10-17 are shown in Scheme 1. Bromothiophene core
21
2
6 was prepared following our previously described synthetic route and was debrominated under
standard hydrogenation conditions to provide thiophene 27 as a flexible intermediate. Re-bromination
of compound 27 with N-bromosuccinimide (NBS) afforded isomeric bromothiophene 28 that was
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 58
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
subjected to palladium-mediated cyanation using Zn/ZnCN followed by removal of the 2,2,2-
2
trichloroethoxycarbonyl (Teoc) group with tetrabutylammonium fluoride (TBAF) to provide pyrrolidine
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
a
Scheme 1. Synthesis of P1 thiophene analogs 10-17
BocN
HN
BocN
HN
BocN
HN
BocN
HN
N
N
N
N
O
O
O
O
a
b
c, d
N
N
N
N
Br
S
S
S
S
H
Teoc
Teoc
Teoc
Br
NC
26
27
28
29
g
e, f
O
1
2
BocN
HN
HN
HN
10 R = OMe, R = H
N
N
O
1
2
1
1
1
1 R = OMe, R = Me
1
2
5 R = OEt, R = Me
1
2
6 R = SMe, R = Me
N
N
1
2
S
S
17 R = NHMe, R = Me
Teoc
N
F
Cl
NC
N
R2
30
R1
d
HN
HN
N
O
BocN
HN
N
O
1
1
2 R = Me
e, f
1
1
3 R = Et
N
1
S
14 R = nPr
N
N
F
S
H
N
Cl
R1
Cl
MeO
31
a
Reagents and conditions: (a) 10% Pd/C, H , methanol, rt, 92%; (b) NBS, DMF, 60%; (c) Zn(CN) , Zn,
2
2
PdCl (dppf)•CH Cl , DMA, then Et N, (Boc) O, CH Cl , 71%; (d) TBAF, THF, 80%; (e) heteroaryl
2
2
2
3
2
2
2
halide, tris(dibenzylideneacetone)dipalladium(0), Na-OtBu, (2-biphenylyl)di-tert-butylphosphine
JohnPhos), toluene, 65 °C, 30-80%; (f) TFA, CH Cl ; (g) NCS, DMF, 75-80%.
(
2
2
ACS Paragon Plus Environment

Page 7 of 58
Journal of Medicinal Chemistry
1
2
3
2
9. To generate the analogous pyrrolidine with a P1 chlorothiophene 31, intermediate 27 was
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
chlorinated using N-chlorosuccinimide (NCS) to afford compound 30 that was subsequently treated with
TBAF. Compounds 29 and 31 were used for preparation of compounds 10-17 in two steps – palladium-
mediated N-heteroarylation with a variety of heteroaryl halides followed by treatment of the resulting
products with TFA to remove the tert-butoxycarbonyl (Boc) protecting group.
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
Synthesis of compound 9 is depicted in Scheme 2 as a representative example for the approach to
P1 phenyl analogs 18-22. Commercially available nitrile 32 was condensed with aqueous glyoxylic acid
using potassium carbonate, and product 33 spontaneously precipitated from the reaction mixture.
Hydrolysis of the nitrile moiety of 33 using a mixture of sulfuric and formic acid produced, after
2
1
dehydrative cyclization, anhydride 34. In contrast to our previous efforts, anhydride 34 was not
hydrolyzed to the corresponding diacid but rather was used directly in the dipolar cycloaddition with
commercially available N-benzyl-1-methoxy-N-((trimethylsilyl)methyl)methanamine 35. After the
cycloaddition reaction was complete, removal of the solvent and treatment of the crude residue with
methanol resulted in precipitation of methyl ester 36 as a single regioisomer from solvolysis of the initial
cycloaddition product. The carboxylic acid moiety of 36 was converted to an amine via
diphenylphosphoryl azide (DPPA)-mediated Curtius rearrangement using 2-(trimethylsilyl)ethanol as
the isocyanate trapping agent followed by unmasking the free amine by treatment with HCl to give
intermediate 37. Boc-protected iminopyrimidinone core 38 was formed by coupling between 37 and t-
1
8
butyl N-[(methylamino)thioxomethyl]carbamate.
using chiral HPLC to provide desired isomer 39 and its enantiomer 40. Removal of the benzyl group of
9 followed by palladium mediated coupling of pyrrolidine 41 with 2-chloro-5-fluoro-4-methoxy-6-
methylpyrimidine and subsequent treatment of the resulting intermediate with TFA provided compound
The enantiomers of compound 38 were separated
3
9
.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 58
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
a
Scheme 2. Synthesis of 2,4-difluorophenyl analog 9 and related analogs 18-22.
O
CO Me
2
CO H
2
MeO
NC
Ar
O
TMS
HO C
NC
Ar
33
b
c
2
a
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
O
+
Ar
N
N
Bn
Ar
Bn
32
34
35
36
BocN
BocN
BocN
CO Me
2
N
N
N
O
O
O
H N
2
d, e
f
g
HN
Ar
HN
Ar
HN
Ar
+
Ar
N
Bn
N
N
N
Bn
Bn
Bn
37
38
39
40
HN
HN
N
N
O
O
BocN
N
HN
Ar
HN
Ar'
O
i, j
h
HN
Ar
N
N
N
H
N
F
N
F
N
N
41
MeO
MeO
9 Ar = 2,4-F -Ph
18 Ar'= Ph
19 Ar' = 2-F-Ph
2
2
2
2
0 Ar' = 4-F-Ph
1 Ar' = 2,5-F -Ph
2 Ar' = 2,4,5-F -Ph
2
3
a
Reagents and conditions: (a) glyoxylic acid (50% aqueous), K CO , MeOH, rt, 89%; (b) 7%
2
3
concentrated H SO in formic acid, 100 °C, 94%; (c) cat. TFA in THF, 0 °C to rt; concentration then
2
4
dilution with MeOH, rt, 57%; (d) DPPA, Et N, toluene, 66 °C then acetic acid, TMS-ethanol, reflux,
3
4
8%; (e) 4N HCl in dioxane, rt; (f) t-butyl N-[(methylamino)thioxomethyl]carbamate, EDCI, DIEA,
DMF, rt, 63%; (g) HPLC, AD column; (h) H , 20% Pd(OH) /C, MeOH, rt, 91%; (i) 2-chloro-5-fluoro-4-
2
2
methoxy-6-methylpyrimidine, Pd(OAc) , Na-OtBu, toluene, 2'-(di-tert-butylphosphino)-N,N-dimethyl-
2
[
1,1'-biphenyl]-2-amine, 120 °C; (j) TFA, DCM, 58%.
Preparation of analogs 23-25 required changes to the synthetic route due to perturbation of the
stereoelectronics of the neighboring carbons imposed by 2,6-difluoro substitution. Scheme 3 shows this
ACS Paragon Plus Environment

Page 9 of 58
Journal of Medicinal Chemistry
1
2
3
modified synthetic route for compound 23. Note that the pyrrolidine intermediates analogous to
compound 46 en route to analogs 24 and 25 were made by a slightly different sequence as shown in
Scheme 4 for intermediate 56.
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
a
Scheme 3. Synthesis of 2,6-difluorophenyl analog 23
a
Reagents and conditions: (a) Pd(PPh ) , AcOH, dioxane, 80 °C, 34%; (b) LiOH•H O, THF, 40 °C; (c)
3 4
2
acetic anhydride, 90 °C, 99%; (d) cat. TFA in THF, 0 °C to rt; concentration then dilution with pre-
mixed MeOH/Et N, -60 °C to rt, 56%; (e) 20% Pd(OH) on carbon, H , MeOH; (f) CbzCl, Et N, MeOH,
3
2
2
3
THF, 64% from 45; (g) DPPA, Et N, toluene, then TMS-ethanol, 120 °C, 67%; (h) 4N HCl in dioxane,
3
rt; (i) t-butyl N-[(methylamino)thioxomethyl]carbamate, EDCI, DIEA, DMF, rt, 57%; (j) HPLC, AD
column; (k) 20% Pd(OH) on carbon, H , MeOH, rt, 98%; (l) 2-chloro-5-fluoro-4-methoxy-6-
2
2
methylpyrimidine,
tris(dibenzylideneacetone)dipalladium(0),
Na-OtBu,
(2-biphenylyl)di-tert-
butylphosphine (JohnPhos), toluene, 65 °C, 46%; (m) TFA, DCM, 63%.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 58
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
a
Scheme 4. Synthesis of pyrrolidine intermediate en route to 2,3,6-trifluorophenyl analog 24
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
a
Reagents and conditions: (a) NaCN, EtOH, H O, reflux, 79%; (b) glyoxalic acid monohydrate, K CO ,
MeOH; (c) 10% conc. H SO in formic acid, reflux; (d) 5% conc. H SO in MeOH, reflux, 41%; (e) N-
benzyl-1-methoxy-N-((trimethylsilyl)methyl)methanamine, cat. TFA, THF, rt, 100%; (f) 15% conc.
H SO in water, reflux; (g) Ac O, 90 °C, then, MeOH and Et N, -78 °C; (h) Pd(OH) , MeOH, rt; (i)
2
2
3
2
4
2
4
2
4
2
3
2
CbzCl, Et N, THF, rt, 79%.
3
Results and Discussion
2
1
Our previous SAR studies generated three optimized lead compounds 6-7, and 8 from
iminopyrimidinone series with BACE1 K values of 3.5, 5.6 and 0.9 nM respectively (Figure 3). Despite
i
their similar structures and affinities, X-ray co-crystal structures with BACE1 showed that these
molecules each exhibit unique binding features in S2'-S2'' (Figure 3(a)). The fluoropyrimidine moiety
of compound 6 binds in an open hydrophobic space near S2' and establishes a hydrogen bonding
7
6
network with Trp through a highly conserved water molecule in a sub-pocket we have termed S2''.
While the pyridine substituent of analog 7 overlays with the pyrimidine of compound 6 very well, it
projects the cyano substituent into S2'' to displace the conserved water molecule and forms a direct
7
6
hydrogen bond with Trp . In sharp contrast to the two modes above, the methoxypyridine moiety of 8
ACS Paragon Plus Environment

Page 11 of 58
Journal of Medicinal Chemistry
1
2
3
shifted roughly 2 Å relative to the other analogs to displace two conserved waters and displaying the
methoxy group deep in the S2" binding pocket.
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
(a)
(
b)
Figure 3: (a) Superimposed X-ray conformation of iminopyrimidinone 6 (yellow, PDB ID code: 5hd0),
(green, PDB ID code: 4h3g), 8 (magenta, PDB ID code: 5hdu) in the active site of BACE1 with red
7
surface showing the location of catalytic aspartic acids. (b) Structures of 6, 7 and 8 with their respective
BACE1 K , CYP3A4 IC (pre), CatD/BACE1 selectivity and calculated properties.
i
50
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 58
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
The different arrangements of the pyridine and pyrimidine moieties offered several potential
strategies for further optimization of the overall inhibitor profiles. At the outset, however, we were
cognizant that optimization of 6, 7 and 8 (Figure 3(b)) should avoid the incorporation of additional
molecular weight and lipophilicity since this could result in profiles unsuitable for a CNS agent and
would likely only exacerbate critical off target liabilities (e.g. hERG). Therefore our strategy was to
remove the P3 cyanophenyl moiety to afford lower molecular weight scaffolds that occupy two binding
sites (S1 and S2''), from which we could pursue SAR to rescue any intrinsic potency loss while fine
tuning physicochemical properties (cLogP, PSA) into CNS drug-like ranges. Such inhibitors would be
structurally differentiated from the majority of optimized amidine-based analogs that occupy the S1-S3
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
2
2a-f,23
pockets.
Additionally we envisaged that optimization based on 6 and 8 would necessitate
modification of their fluoropyrimidine moieties. It was hypothesized that the time-dependent CYP3A4
inhibition observed with 6 and 8 (Figure 3(b)) was a result of bioactivation of the fluoropyrimidine
moiety to generate a reactive intermediate. Evidence for this was provided by incubation of 6 with
recombinant CYP3A4, supplemented with glutathione, which led to the formation of a product of m/z
751.2104 corresponding to a glutathione adduct of 6 with concomitant loss of fluoride.
Implementation of this strategy using 6 and 7 as starting points failed to provide an inhibitor with
an acceptable overall profile. On the other hand, compound 8 presented some potential advantages as a
starting point including higher BACE1 affinity and similar or better selectivity for BACE1 over
cathepsin D (CatD) compared to 6 and 7. The importance of achieving substantial selectivity for
BACE1 over CatD has been highlighted by the recent report of ocular toxicity and neurodegeneration in
2
2a
rats that was attributed to CatD inhibition with the modestly selective BACE1 inhibitor LY-2811376.
ACS Paragon Plus Environment

Page 13 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
Truncation of the P3 cyanophenyl group of compound 8 to P1 cyanothiophene analog 10 led to a
4
0-fold drop in BACE1 affinity (Table 1). Introduction of a methyl group at the 6-position of the
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
pyrimidine gave compound 11 resulting in a 4-fold gain of BACE1 affinity. Replacement of the P1 5-
cyano substituent of 11 with chloro (12) resulted in similar affinity and cellular potency for Aβ40
lowering. On the other hand, further increase in lipophilicity resulting from replacement of the
pyrimidine methyl substituent of 12 with ethyl (13) or n-propyl (14) was detrimental to cellular potency.
The X-ray crystal structure of 11 complexed with BACE1 showed that truncation and subsequent
pyrimidine substitution did not affect the overall binding mode of this series. The methoxy group of the
pyrimidine remains deep in the S2" binding pocket while the methyl group projects toward the S2'
binding site as designed (Figure 4). Selectivity for BACE1 over CatD for the truncated analogues was
generally maintained or eroded relative to 8, with the exception of 15 which appeared to be highly
BACE1 selective (Table 1). In the absence of a crystal structure of 15 and its analogues in complex with
CatD, a convincing explanation for the observed high selectivity of 15 is not readily apparent. A recent
publication disclosing inhibitors that bind a region of BACE1 similar to that for the inhibitors described
herein attributed the observed CatD selectivity to be a result of interactions between the inhibitors and
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
7
8
40
22g
Tyr and Trp of CatD.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 58
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
a
Table 1. SAR investigation of truncated thiophene analogs.
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
1
R
2
R
3
R
MW
ClogP)
BACE1
K (nM)
HEK293 Aβ40
IC50 (nM)
CatD /
BACE1
Cpd
8
LE
(
i
468 (2.8)
1.4
18
0.35
95
1
1
1
1
0
1
2
3
CN
CN
Cl
H
OMe
OMe
OMe
OMe
OMe
OEt
401 (1.0)
416 (1.5)
435 (2.7)
439 (3.2)
453 (3.8)
430 (2.0)
432 (2.0)
414 (1.6)
57
15
10
11
15
48
45
336
238
40
0.35
0.46
0.39
0.39
0.35
0.33
0.34
0.30
93
39
Me
Me
Et
51
70
Cl
203
190
361
36
ND
18
14
15
Cl
nPr
Me
Me
Me
CN
CN
CN
>2000
5.2
1
1
6
7
SMe
NHMe
157
66
a
Protocols for determination of inhibitor K values against purified human BACE1 and Cathepsin D, as well as for cellular
i
16
BACE1 IC values, have been previously described.
50
ACS Paragon Plus Environment

Page 15 of 58
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
(
a)
(b)
Figure 4: (a) Superimposition of X-ray crystal structures of 11 (yellow, PDB ID code: 5hdv) and 8
green, PDB ID code: 5hdu) complexed with BACE1. Red surface represents the catalytic dyad Asp
and Asp . (b) Overlay of X-ray crystal structures of 11 (yellow, PDB ID code: 5hdv), 15 (blue, PDB
ID code: 5hdx), 16 (purple, PDB ID code: 5hdz) and 17 (green, PDB ID code: 5he5). The red ball
represents the bridging water molecule between the Trp and the nitrogen of the pyrimidine of 15, 16
3
2
(
2
28
76
and 17. Pyrimidine moieties of 15, 16 and 17 have shifted almost 1.9 Å towards S2' binding site.
To probe if methoxy was optimal for binding in S2'', the methoxy group of 11 was homologated
to the ethoxy analog 15 which resulted in a three-fold loss in BACE1 affinity and a similar loss of
cellular potency. We next examined the isosteric aminomethyl and thiomethyl replacements.
Thiomethyl substitution was tolerated with affinity similar to that of the ethoxy analog (16, BACE1 K =
i
4
5 nM). In contrast, aminomethyl analog 17 displayed a significant drop in affinity (BACE1 K = 336
i
nM). Interestingly, the X-ray structures of 15, 16 and 17 complexed with BACE1 showed that the
pyrimidine moiety had flipped to orient the ethoxy, thiomethyl and aminomethyl moieties toward the S2'
binding pocket and that the pyrimidine ring had shifted from the S2" binding site by approximately 1.9
7
6
Å, engaging in an interaction with Trp through conserved water molecules mimicking the binding
modes of compound 6 (Figure 4(b)). Whereas the in-plane anti-oriented methoxy group of the
pyrimidine ring of 11 is positioned optimally in the S2" pocket, larger groups such as ethoxy and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 58
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
thiomethyl (e.g. 15 and 16), or groups that lack the energetics to adopt preferentially the anti-orientation
such as the aminomethyl group (e.g. 17) occupy the relatively larger S2' binding site.
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 SAR with optimized S2'-S2'' moiety
The SAR explorations described above demonstrated that truncation from S3 and optimization of
the substituents in S2'' and S2' could generate potent, ligand-efficient analogs relative to their higher
molecular weight precursors. In fact, the optimal S2'-S2'' substituent, 4-methoxy-5-fluoro-6-
methylpyrimidine present in analogs 11 and 12 imparted excellent affinity with minimal (methyl)
occupancy of S2', and both of these analogs gratifyingly showed reduced time-dependent inhibition of
CYP3A4 compared to starting point 8 (IC50 2 µM for 8 vs. 24 µM and 18 µM for 11 and 12
respectively). We attribute the reduced time-dependent CYP3A4 inhibition of 11 and 12 to the
increased steric demand of the newly introduced methyl substituent reducing the rate oxidation of the
pyrimidine moiety and/or redirecting metabolism to the methyl group itself.
Given that the S1 thienyl moiety was originally installed to provide an optimal vector for the
now absent aryl substituent into the S3 pocket, we next chose to replace this group with phenyl in
combination with the optimized S2'-S2" pyrimidine substituent. In this regard, compound 18 was well
tolerated with a BACE1 K of 36 nM and whole cell potency of 58 nM (Table 3). A series of mono- and
i
polyfluorinated phenyl analogs was also prepared to assess SAR related to in vitro potency, selectivity,
and rat PK parameters. As shown in Table 3, introduction of fluorine to the P1 phenyl was generally
beneficial, imparting improved K values (e.g. 9, 19-22). The improvement in potency may be due to
i
general hydrophobic contacts within the lipophilic S1 site, or may be driven by electrostatic interactions
1
15
71
of the fluorine with the phenyl rings of Trp and with the hydroxyl group of flap Tyr as suggested by
24
the X-ray co-crystal structure of the 2,4-difluorophenyl analogue 9 (Figure 5(a)). The
ACS Paragon Plus Environment

Page 17 of 58
Journal of Medicinal Chemistry
1
2
3
Table 3. SAR investigation of fluorophenyl substitution in the presence of the 4-methoxy-5-fluoro-6-
methylpyrimidinyl substituent.
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
a
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
HEK293
Aβ40 IC50
CYP3A4
(preincubation,
µM)
1
MW
ClogP)
BACE1
K (nM)
CatD/
BACE1
b
Cpd
R
C (µM)
B
B/P
(
i
(
nM)
3
84
1
8
36
58
349
ND
ND
>30
(
2.3)
4
02
1
2
9
0
6
7
17
16
143
354
0.074
0.32
0.7
0.9
>30
(
2.5)
402
(2.5)
34.1
4
20
9
5
9
14
28
458
158
0.072
0.056
1.1
0.7
>30
>30
(
2.6)
F
4
20
2
2
1
2
3
(
2.6)
F
438
2.7)
8
35
49
446
636
151
99
ND
ND
0.088
0
NA
ND
0.4
>20
>20
>20
~30
(
F
F
4
20
2
20
53
30
(2.6)
4
38
24
234
93
(
2.7)
F
F
F
4
56
25
F
NA
(
2.8)
a
Protocols for determination of inhibitor K values against purified human BACE1 and cathepsin D, as well as for cellular
i
16 b
BACE1 IC values, have been previously described. Concentration in brain homogenate 6h post 10 mpk PO dosing.
50
exception to the potency improvements with fluorine addition are the 2,6-difluoro-containing analogs
2
3-25. These analogs are all slightly less active than their 2-fluoro-6-proteophenyl counterparts (analogs
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 58
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
230
9
, 19-22), possibly due to a weakly-negative interaction with carbonyl of Gly and carbonyl of the side
3
2
chain carboxylic acid of Asp (Figure 5(b)).
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
(a)
(b)
7
1
Figure 5: (a) X-ray crystal structure of 9 complexed with BACE1. The hydroxyl moiety of Tyr of the
flap is 2.9 Å away from the 2-fluorine of 2,4-difluorophenyl group of compound 9. (b) Superimposition
of X-ray crystal structures of 9 (yellow, PDB ID code: 5he7) and a racemic sample of 23 (green, PDB
ID code: 5he4) complexed with BACE1. The 6-fluorine of the bound enantiomer of 23 is 3.1 Å away
230
from the backbone carbonyl of Gly and is 3.4 Å away from the carbonyl of the side chain carboxylic
3
2
acid of Asp .
Importantly, in all cases, the addition of fluorine did not significantly erode the cell Aβ40 IC50
values relative to the K values of these analogs, and compounds 9, 19 and 20 showed cell IC values of
i
50
<20 nM. In addition, all of these analogs showed at least 100-fold selectivity over CatD and did not
show significant inhibition of CYP 3A4 under preincubation conditions. Of these, the 2,4-
difluorophenyl analogue 9 was selected for further profiling based on its favorable overall profile with
respect to BACE1 affinity, cell potency and 460-fold selectivity over CatD.
Compound 9 displayed good selectivity for BACE1 over the human aspartic proteases pepsin,
(>20000x), renin (6820x) and cathepsin E (648x), but was not selective versus the closely related
ACS Paragon Plus Environment

Page 19 of 58
Journal of Medicinal Chemistry
1
2
3
BACE1 homolog BACE2 (K = 2.7 nM). Expressed predominantly in the periphery, BACE2 processes
i
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
2
5
the melanocyte structural protein PMEL17 required for proper pigmentation in rodents. BACE2 is
also expressed in pancreatic β-cells and loss of BACE2 function has been associated with improved
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
2
6
glucose homeostasis. While knowledge of BACE2 mediated biological processes continues to evolve,
current understanding mitigates concerns related to BACE2 inhibition for development of a therapeutic
agent. In a counterscreen of a panel of human receptors, enzymes and ion channels, 9 did not show
significant off target activity with the exception of histamine H2 (K = 4.9 ꢀM), melatonin MT1 (K =
i
i
2
.9 ꢀM), and motilin (K = 19 ꢀM) receptors.
i
Table 4. Rat, dog and monkey pharmacokinetic parameters of 9
Hep. Clint
In vivo
clearance b
mL/min/kg
92
t1/2
Species
Dose a AUC(
Cmaxa
(µM)
V ,ss
(L/kg)
0-∞)
6
d
a
%F ꢀL/min/10
b
b
(mg/kg)
(µM.hr)
(hr)
cells
Rat
Dog
3
0.30
0.1
19
8.2
33
53
73
48
9.0
1.1
2.2
7.3
3
3
0.39
2.6
0.19
0.49
35
15
3.6
5.7
Monkey
a
b
Oral. Intravenous (3 mg/kg for rat and 1 mg/kg for dog and monkey).
Compound 9 did not significantly inhibit human CYP’s 1A2, 2C8, 2C9, 2C19, 2D6 and 3A4 at
concentrations below 20 µM and thus has minimal risk of interaction with drugs that are metabolized by
the major human CYPs, demonstrating a marked improvement over its progenitor 6. Inhibition of the
hERG channel by 9 (IC = 1.9 ꢀM) was also reduced relative to 6.
5
0
The pharmacokinetic parameters of 9 are shown in Table 4. Oral bioavailability of 9 was low in
rats and dogs and moderate in monkeys, consistent with the high Cl observed in hepatocytes of the
int
respective species. In vivo total clearance exceeded liver blood flow in rats and dogs, indicating that
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 58
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
clearance pathways other than hepatic extraction contribute to total clearance. In this regard, 9 was
unstable in rat and dog plasma suggesting that its plasma instability may constitute a significant extra-
hepatic clearance pathway in these species. In cynomolgus monkeys, total in vivo clearance of 9 was
moderate and in concert with the moderate volume of distribution resulted in higher plasma exposure
and a longer half-life in vivo. The compound achieves substantial total brain concentrations in rats and
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
Dose
mg/kg, PO)
CP
(nM)
5.6
17.53
121.2
770.4
2981
CB
(nM)
3.0
13.5
56.3
346.9
1781.1
4043
b
C
CSF(nM)
c
(
a
0.3
<bloq
<bloq
<bloq
a
1
3
10
a
5.4
26.1
80.7
119.8
30
1
3
00
00
8805
14186
5018
a
b
bloq; below limit of quantitation (0.6 nM).
Plasma concentration. Brain concentration
c
(a)
(b)
Figure 6. (a) Measured total concentration of 9 in plasma (C ), CSF (C ) and brain (C ) of rats at
P
CSF
B
doses ranging from 0.3 mg/kg to 300 mg/kg. (b) Effect of acute oral administration of compound 9 on
the reduction of plasma, CSF and cortical Aβ levels in rats.
4
0
monkeys (brain/plasma ratio of 0.8 and 3.2, respectively). The improved brain penetration of 9
21
compared to 6 (brain / plasma ratio of 1.1 versus 0.1, respectively) is consistent with the reduction in
2
2
calculated PSA achieved during optimization (PSA 94 Å vs. 109 Å , respectively). While instability of
9 in plasma from rat, dog and monkey precluded accurate determination of plasma protein binding
values, the compound is more stable in human plasma (95% bound). In human hepatocytes 9 exhibited
6
low intrinsic clearance (1.8 ꢀL/min/10 cells), in contrast with the high hepatocyte clearance observed in
rat, dog and monkey hepatocytes (Table 4).
ACS Paragon Plus Environment

Page 21 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
Compound 9 was evaluated to assess its ability to lower Aβ in the plasma, brain and CSF of
4
0
rats (Figure 6). Dose related increases in exposure of 9 in plasma, brain and CSF was seen across the
dose range of 0.3 – 300 mg/kg, and dose responsive lowering of Aβ40 was observed in all three
compartments. A dose of 0.3 mg/kg of 9 lowered Aβ40 in plasma to 40% of the control levels
corresponding to an ED for Aβ lowering of 0.5 mg/kg in plasma. On the other hand, potency for
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
5
0
40
CSF and cortex Aβ lowering was considerably shifted rightward with ED of 5.2 mg/kg and 12.7
4
0
50s
mg/kg respectively. Notably, levels of 9 in the CSF were below the limit of quantitation at the 3 mg/kg
dose, a dose at which significant CSF Aβ lowering was observed. Consistent with this observation, 9
4
0
was determined to be a rat Pgp substrate in an MDCK cell line overexpressing MDR1 (efflux ratio >30;
−
6
Papp (41 × 10 cm/s) indicating that CSF levels of 9 cannot be considered a reliable predictor of its free
2
7
brain concentration.
Dose (mg/kg)
10
30
Plasma
4.11
0.34
7.0
100
AUC0-48h (µM.h)
Cmax (µM)
Tmax (h)
0.608
0.067
5.0
22.9
1.14
20.0
CSF
AUC0-48h (µM.h)
Cmax (µM)
0.0046 0.035 0.304
0.0009 0.0046 0.0234
Tmax (h)
6.0
6.0
34.0
(a)
(b)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 58
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
(
c)
(d)
(e)
(f)
Figure 7: Acute PK/PD profile of compound 9 in cisterna magna ported rhesus monkeys. Plasma and
CSF samples were collected at 2, 4, 6, 8, 12, 24 and 48 h after oral administration of 9 (10, 30 or 100
mg/kg) and analyzed for both compound concentrations and Aβ40 levels. (a) Plasma and CSF
concentrations of 9 over time. (b) Pharmacokinetic parameters of 9. (c) Plasma Aβ levels. (d) CSF
40
Aβ levels. (e) CSF Aβ levels. (f) CSF sAPPβ levels.
4
0
42
The pharmacodynamic profile of 9 was also determined in non-human primates. Single doses of
compound 9 (10, 30, and 100 mg/kg) were orally administered to cisterna magna ported rhesus monkeys
enabling evaluation of its effect on the temporal dynamics of BACE1-related biomarkers in CSF and
ACS Paragon Plus Environment

Page 23 of 58
Journal of Medicinal Chemistry
1
2
3
plasma. Plasma and CSF samples were collected at 2, 4, 6, 8, 12, 24 and 48 h post dose to analyze
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 and Aβ40 levels. In addition, Aβ and sAPPβ were analyzed at each time point in the
42
samples from CSF. A dose dependent increase in exposure of 9 was observed in plasma and CSF
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 7(a)). Correspondingly, dose and time dependent reduction of CSF Aβ40 was observed with
profound peak reduction (-77% relative to baseline) observed at 24 h following the high dose (Figure
(d)). The maximum lowering of plasma Aβ40 at each dose was not differentiated, with apparent
7
plateauing at 50-60% reduction as a result of the limited window of quantification by the plasma Aβ40
assay. However, a dose-dependent effect on the duration of plasma Aβ inhibition was observed which
4
0
correlated with plasma exposures of 9 (Figure 7(c)). Peak reductions of CSF Aβ occurred at time
4
0
points that were significantly delayed relative to the timing (Tmax) of the plasma and CSF Cmax
concentrations of 9, with the delay (~6 h) more pronounced at the lower doses (cf. Figures 7(a) and
7
(d)). Reductions of Aβ and sAPPβ in the CSF followed a temporal pattern very similar to that of Aβ
42 40
(Figures 7(e) and 7(f)). This hysteresis effect is likely related to the kinetics of central Aβ turnover and
28
transit to the CSF from within the brain.
Figure 8. Acute effect of compound 9 (10 mg/kg) on CSF and cortical Aβ levels in cynomolgus
40
monkeys (*p = 0.0014, **p < 0.0007).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 58
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
The acute effect of 9 on the reduction of CSF and cortex Aβ was investigated in cynomolgus
4
0
monkeys (Figure 8). Acute oral administration of 9 at 10 mg/kg resulted in 58% reduction of Aβ in
40
cortex compared with the vehicle control group when evaluated at 4 h post dose. Interestingly, CSF
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
A
β40 reduction (35% relative to control) was not as pronounced as that observed in cortex at the same
time point.
The greater reduction of cortex Aβ40 compared to CSF Aβ40 at the 4 h time point is thought to be
a manifestation of the dynamics of Aβ40 transit and turnover under the conditions of acute
administration of 9. A close temporal relationship between BACE1 inhibition of Aβ40 synthesis and the
magnitude of Aβ40 inhibition would be expected in cortex where Aβ40 is directly produced. In contrast,
the degree of reduction of Aβ40 in the CSF under non-steady state conditions of acute inhibitor
administration is dependent on the kinetics of Aβ40 clearance from CSF and of transit of newly
2
8
synthesized A
β
into the CSF compartment from the primary sites of synthesis in the brain.
This
hypothesis is consistent with the observed delay to maximum CSF Aβ40 lowering of 12-24 h post dosing
in rhesus monkeys (Figure 7(d)).
Conclusions
Design and development of potent BACE1 inhibitors for advanced clinical studies has been
challenging due to the intrinsic nature of the BACE1 active site and the localization of the enzyme
within intracellular compartments of the brain. To achieve high affinity, inhibitors require polar
functionality to interact with the active site catalytic aspartic acids complemented by lipophilic groups
that bind in the adjacent specificity pockets. Successful inhibitor development requires an appropriate
balance of properties to achieve cell and brain penetration while minimizing off target liabilities. Our
earlier work employing the novel bicyclic iminopyrimidinone scaffold clearly illustrated these
ACS Paragon Plus Environment

Page 25 of 58
Journal of Medicinal Chemistry
1
2
3
challenges, resulting in the potent BACE1 inhibitor 6 which suffered from significant CYP3A4 and
hERG inhibition, and sub-optimal brain exposure. Using 6 as a starting point, improvement of the
ancillary profile and CNS properties was achieved by truncation of the P3 moiety and optimization of
the pyrimidine moiety binding in the S2'-S2'' region, delivering 11. Subsequent optimization of the P1
aryl group of 11 to improve affinity and selectivity delivered 2,4-difluorophenyl analogue 9 which is
structurally diverse from our previously reported S1-S3 binding inhibitors and afforded robust central
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
Aβ reduction in rats and monkeys.
EXPERIMENTAL SECTION
Procedures for X-ray crystallography, aspartyl protease K determination, cell Aβ IC50
i
40
determination and determination of Aβ levels in rat brain and CSF samples were conducted as
1
5,16,18
previously described.
Synthesis: Unless otherwise mentioned, all the reagents and solvents were obtained from
commercial sources and used without further purification. Air sensitive chemistries were performed
under an atmosphere of nitrogen or argon. Purification of the final compounds to >95% purity were
carried out either using a prepacked silica gel cartridge (Analogix or Biotage or ISCO) or a reverse
phase C18 column (mobile phase, A = 0.05% TFA in water and B = 0.05% TFA in acetonitrile, gradient
= 10% to 95% B over 10 min). All NMR data unless otherwise specified were collected at 400 MHz on
Varian or Bruker instruments. Chemical shifts are reported in ppm relative to the residual solvent peak
1
in the indicated solvent, and for H NMR, multiplicities, coupling constants in Hertz, and numbers of
protons are indicated parenthetically. Microwave assisted reactions were performed in a Smith
synthesizer from Personal Chemistry. Purity and MS information was obtained via LC-electrospray-
mass spectroscopy with a C18 column using a gradient of 5% to 95% MeCN in water with 0.05% TFA
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 58
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
as the mobile phase. The purity of the samples was assessed using a UV detector at 254 nm. An
additional analytical reverse phase HPLC system was used to assess the purity of final compounds using
an ELSD detector and a UV detector monitoring both 220 nm and 254 nm.
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
tert-Butyl ((4aR,7aR)-7a-(5-cyanothiophen-2-yl)-3-methyl-4-oxohexahydro-1H-pyrrolo[3,4-
d]pyrimidin-2(3H)-ylidene)carbamate (29). To a solution of 26 (1.8 g, 3.1 mmol) in methanol (40
mL) was added 10% palladium on carbon (0.9 g), and the resulting mixture was stirred under the
atmospheric of H for 1.5 h. The reaction mixture was filtered, the filtrate was concentrated and the
2
1
residue was purified using 0 to 100% ethyl acetate in hexanes to provide 27 (1.41 g, 92%). H NMR
(
400 MHz, CDCl ) δ 10.4 (m, 1H), 7.31 (m, 1H), 6.97-6.99 (m, 2H), 4.20 (m, 2H), 3.68-4.06 (m, 4H),
3
3.50 (m, 1H), 3.30 (s, 3H), 1.53 (s, 9H), 1.01 (m, 2H), 0.03 (s, 9H).
To a solution of 27 (1.41 g, 2.8 mmol) in DMF (15 mL) was added NBS (0.6 g, 3.4 mmol), and
the resulting mixture was stirred for 12 h. The reaction mixture was diluted with ethyl acetate (100 mL),
washed with brine, and then with water. The organic layer was dried with Na SO and concentrated.
2
4
The crude product was purified by silica gel chromatography using 0 to 100% ethyl acetate in hexanes
1
to provide 28 (0.96 g, 60%). H NMR (400 MHz, CDCl ) δ 10.39 (m, 1H), 6.94 (d, J = 4.0 Hz, 1H),
3
6
.76 (m, 1H), 4.20 (m, 2H), 3.68-4.04 (m, 4H), 3.43 (m, 1H), 3.29 (s, 3H), 1.53 (s, 9H), 1.01 (m, 2H),
.03 (s, 9H).
0
To a solution of 28 (1.0 g, 1.7 mmol) in DMA (3 mL) were added PdCl (dppf) (0.2 g, 0.17
2
mmol), Zn(CN) (0.2 g, 1.7 mmol) and Zn powder (0.1 g, 1.7 mmol). The resulting solution was
2
degassed and heated at 80 °C for 4 h. The reaction mixture was then diluted with ethyl acetate and
passed through a silica gel plug. The filtrate was evaporated, and to a solution of the resulting material
in dichloromethane (5 mL) was added Et N (0.5 mL, 3.5 mmol) and (Boc) O (0.572 g, 2.6 mmol). The
3
2
ACS Paragon Plus Environment

Page 27 of 58
Journal of Medicinal Chemistry
1
2
3
resulting solution was stirred at rt for 12 h and then evaporated to dryness. The crude product was
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
purified via silica gel chromatography eluting with 25% ethyl acetate in hexanes to provide a Teoc-
1
protected precursor to cyanothiophene core 29 (0.628 g, 71%). H NMR (400 MHz, CDCl ) δ 10.47 (m,
3
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
1
H), 7.51 (d, J = 4.5 Hz, 1H), 7.04 (m, 1H), 4.16-4.20 (m, 2H), 3.71-4.04 (m, 4H), 3.46-3.49 (m, 1H),
3.26 (s, 3H), 1.50 (s, 9H), 0.98 (m, 2H), 0.01 (s, 9H). To a solution of this intermediate (1.0 g, 1.9
mmol) in THF (2 mL) was added a solution of 1M TBAF in THF (5.7 mL, 5.7 mmol) at 0 °C, and the
resulting solution was warmed to rt over period of 30 minutes and then stirred for additional 3 h. The
reaction mixture was evaporated to dryness, and was directly loaded into silica gel column and purified
using 100% ethyl acetate to provide 29 (0.56 g, 80%).
General Procedure A for N-arylation and Subsequent Boc Removal. To a solution of the
pyrrolidine intermediate (1 equiv.) and heteroaryl halide (1.2 equiv.) in toluene (5 mL/mmol of
pyrroldine) were added tris(dibenzylidene-acetone)dipalladium(0) (0.1 equiv.), (2-biphenylyl)di-tert-
butylphosphine (JohnPhos, 0.15 equiv.) and Na-OtBu (2.2 equiv.). The resulting solution was degassed
and heated at 70 °C for 3 h, then cooled to room temperature, filtered through a pad of Celite, and the
crude material was purified via silica gel chromatography using ethyl acetate in hexanes as the eluant to
give an N-arylated product. The Boc protecting group was removed through treament of this material
with 20% TFA/CH Cl (2.5 mL/mmol). The deprotected compounds were purified by reverse phase
2
2
HPLC (C18) using 0 to 90% water / acetonitrile with 0.05% TFA.
5
-((4aR,7aR)-6-(5-Fluoro-4-methoxypyrimidin-2-yl)-2-imino-3-methyl-4-oxooctahydro-1H-
pyrrolo[3,4-d]pyrimidin-7a-yl)thiophene-2-carbonitrile (10). Compound 10 was prepared from
compound 29 (38 mg, 0.101 mmol) and 2-chloro-5-fluoro-4-methoxypyrimidine following general
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 58
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
1
procedure A (2.2 mg, 5%). H NMR (400 MHz, CD OD) δ 8.05 (d, J = 3.3 Hz, 1H), 7.76 (d, J = 4.0
3
Hz, 1H), 7.33 (d, J = 4.0 Hz, 1H), 4.41 (d, J = 12.4 Hz, 1H), 4.10 (m, 4H), 4.02 (s, 3H), 3.33 (s, 3H).
m/z: 402.2.
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-((4aR,7aR)-6-(5-Fluoro-4-methoxy-6-methylpyrimidin-2-yl)-2-imino-3-methyl-4-
oxooctahydro-1H-pyrrolo[3,4-d]pyrimidin-7a-yl)thiophene-2-carbonitrile (11). Compound 11 was
prepared from compound 29 (40 mg, 0.107 mmol) and 2-chloro-5-fluoro-4-methoxy-6-
1
methylpyrimidine following general procedure A (26 mg, 60%). H NMR (400 MHz, CD OD) δ 7.75
3
(
d, J = 4.0 Hz, 1H), 7.32 (d, J = 4.0 Hz, 1H), 4.88 (s, 3H), 4.40 (d, J = 12.9 Hz, 1H), 4.08 (m, 4H), 3.98
s, 3H), 3.33 (s, 3H), 2.28 (d, J = 2.9 Hz, 3H). m/z: 416.2.
(
tert-Butyl
pyrrolo[3,4-d]pyrimidin-2(3H)-ylidene)carbamate (31). To a solution of 27 (1.0 g, 2 mmol) in DMF
15 mL) was added NCS (0.265 g, 4 mmol), and the resulting mixture was stirred at 50 °C for 2 h, and
((4aR,7aR)-7a-(5-chlorothiophen-2-yl)-3-methyl-4-oxohexahydro-1H-
(
then cooled to room temperature. The reaction mixture was diluted with water and extracted with ethyl
acetate (3x50 mL). The organic layer was dried with Na SO , concentrated and the residue was purified
2
4
1
using 25% ethyl acetate in hexanes to provide 5-chlorothiophene 30 (0.846 g, 80%). H NMR (400
MHz, CDCl ) δ 10.39 (m, 1H), 6.78 (m, 2H), 4.19-4.21 (m, 2H), 3.71-4.04 (m, 4H), 3.43 (m, 1H), 3.29
3
(s, 3H), 1.52 (s, 9H), 1.0 (m, 2H), 0.03 (s, 9H).
To a solution of 30 (0.846 g, 1.6 mmol) in THF (2 mL) was added a solution of 1M TBAF in
THF (4.8 mL, 2.4 mmol) at 0 °C, and the resulting solution was warmed to rt over period of 30 minutes
and then stirred for additional 3 h. The reaction mixture was evaporated to dryness, and was loaded onto
silica gel column and purified using 100% ethyl acetate to provide 31 (0.417 g, 75%).
ACS Paragon Plus Environment