Generation of Leads for γ-Secretase Modulation

Article abstract of DOI:10.1021/acs.jmedchem.0c00446

Herein, we disclose three structurally differentiated γ-secretase modulators (GSMs) based on an oxadiazine scaffold. The analogues from series I potently inhibit the generation of Aβ42 in vitro when the substituents at 3 and 4 positions of the oxadiazine moiety adopt an α orientation (cf. 11). To address the concern around potential reactivity of the exocyclic double bond present in series I toward nucleophilic attack, compounds containing either an endocyclic double bond, such as 20 (series II), or devoid of an olefinic moiety, such as 27 (series III), were designed and validated as novel GSMs. Compound 11 and azepine 20 exhibit robust lowering of CSF Aβ42 in rats treated with a 30 mg/kg oral dose.

Full text of DOI:10.1021/acs.jmedchem.0c00446

Subscriber access provided by CARLETON UNIVERSITY
Article
Generation of Leads for #-Secretase Modulation
Mihir Baran Mandal, Alexei V. Buevich, John P. Caldwell, Lynn A. Hyde, Xianhai
Huang, Xiaoxiang Liu, Brian A. McKittrick, Robert Mazzola, Dmitri A Pissarnitski,
Anandan Palani, Lili Zhang, Eric M. Parker, Li Xiao, Diane Rindgen, and John Zhu
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00446 • Publication Date (Web): 27 Jul 2020
Downloaded from pubs.acs.org on August 3, 2020
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 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Generation of Leads for γ-Secretase Modulation
‡
Mihirbaran Mandal,*‡ Alexei Buevich,† John P. Caldwell, Lynn Hyde,§ Xianhai Huang‡, Xiaoxiang
Liu,‡ Brian McKittrick,‡ Robert D. Mazzola,‡ Dmitri Pissarnitski,‡ Anandan Palani,‡ Lili Zhang,§ Eric
Parker,§ Li Xiao, Diane Rindgen,# and Zhaoning Zhu‡
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
‡Department of Medicinal Chemistry, #Department of Pharmacokinetics, Pharmacodynamics, and Drug
Metabolism, †Department of NMR Structure Elucidation, §Department of Neuroscience, and
Department of Modeling and Informatics
Merck & Co., Inc., 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA.
ABSTRACT
Herein we disclose three structurally differentiated γ-secretase modulators (GSMs) based on an
oxadiazine scaffold. The analogs from series I potently inhibits the generation of Aβ42 in vitro when the
substituents at 3 and 4 positions of the oxadiazine moiety adopts an α orientation (cf. 11). To address the
concern around potential reactivity of the exocyclic double bond present in series I towards nucleophilic
attack, compounds containing either an endocyclic double bond, such as 20 (series II), or devoid of an
olefinic moiety, such as 27 (series III), were designed and validated as novel GSMs. Compound 11 and
azepine 20 exhibit robust lowering of CSF Aβ42 in rats treated with a 30 mg/kg oral dose.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 49
1
2
3
4
5
6
F
F
F
F
7
8
9
O
N
O
N
3
4
O
N
N
N
F
MeO
N
O
F
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
MeO
N
N
N
N
F
N
N
20
27
11
(Series II)
(Series III)
(Series I)
Introduction
Despite enormous efforts both in industry and in academia, a disease-modifying therapeutic for
Alzheimer’s disease (AD) has not yet been approved.1 Current therapies for AD, which are only
palliative, include four acetylcholine esterase inhibitors such as donepezil, rivastigmine, tacrine and
galantamine, and one NMDA antagonist, memantine.2 Thus, the development of agents that will halt or
reverse the disease progression still represents an important unmet medical need. Toward this overall
goal, testing the amyloid hypothesis by inhibiting BACE and γ-secretase was recognized as a promising
strategy.3,4 Concurrent with the efforts to discover novel BACE15 and γ-secretase inhibitors (GSI),6
designing novel γ-secretase modulators (GSM) to lower the formation of highly amyloidogenic Aβ42 with
concomitant increase in the formation of Aβ37 from APP processing while sparing non-amyloid
substrates were undertaken.7 Our initial efforts resulted in the discovery of two novel chemotypes
represented by oxadiazoline 3 and oxadiazine 4 inspired by the classes of compounds represented as 18
and 2.9 Representative analogs from these series demonstrated robust lowering of CSF Aβ42 after oral
dosing.7h Subsequently, analog 57d generated by ring expansion from 4, and the analog 67i with
substitution at 3 position of oxadiazine were also validated as GSMs.
2
ACS Paragon Plus Environment

Page 3 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
N
OH
N
O
N
F
F
NEt2
R
S
R
N
N
N
N
N
H
F
F
Me
F
MeO
N
4
3
1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
(33 nM)
R =
(29 nM)
N
F
F
O
N
O
O
N
R
N
F
R
N
R
N
F
F
F
2
5
6
(64 nM)
(39 nM)
(48 nM)
Figure 1. Structures of first non-NSAID γ-secretase modulators 1, 2 and the discovery of
oxadiazoline 3, oxadiazine 4, 5 and 6, and in parenthesis Aβ42 IC50.
In our further efforts to identify structurally differentiate leads using oxadiazine scaffold, several
strategies were undertaken: 1) developing SAR at the C-3 position of oxadiazine moiety of 4 while
incorporating smaller groups at the 4 position (C-4) (Figure 2, Strategy I); 2) reconfiguration of the
exocyclic double bond into the more hindered β, β-disubstituted endocyclic olefin (Figure 3, Strategy II);
and 3) removal of olefinic bond without incorporation of an additional aromatic ring (Figure 5, Strategy
III). This article describes the guiding principles for each of these strategies, the synthesis, and
evaluation of the resulting molecules as GSMs.
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
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 49
1
2
3
4
5
6
7
8
9
Results and discussion
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Strategy I: SAR at 3 position (C-3) of oxadiazine moiety of 4 while incorporating smaller groups at
the 4 position (C-4).
Design considerations:
Analysis of reported GSMs revealed a novel chemical series exemplified by analog 7 with an IC50
of 87 nM for the inhibition of Aβ42 in vitro (Figure 2).10 Structural alignment of compound 7 with 4
suggested that the aryl moiety from the C-2 position of the imidazole 7 projected into a distinct binding
pocket previously unexplored using 4-substituated oxadiazine scaffold (cf. 4). Although direct follow up
of this finding by adding the second aryl at the C-3 position of the oxadiazine moiety of 4 was a potential
approach, we recognized that this strategy would generate analogs with poor physicochemical properties.
4
ACS Paragon Plus Environment

Page 5 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
R1
R2
O
N
O
N
3
N
N
F
F
Ar
4
Ar
4
MeO
N
F
O
N
R1
R2
3
9
N
Ar
=
F
N
Ar
4
O
R1
R2
2
N
Br
R1 = aryl
N
8
Ar
N
15
Ar
N
R2 = small alkyl
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
moiety
7
10
(a)
Cpd
ΔE (Kcal/mol)
(9ea - 9ae)
-4.34
Cpd
ΔE (Kcal/mol)
(10ee - 10aa)
9
10
-0.04
(b)
(c)
Figure 2. The rationale for designing disubstituted oxadiazines. (a) Structure of 4 and 7, and
conceptualization of 3,4-disubstituted oxadiazine 8. Compounds 9 and 10 are the diastereomeric pairs
differing in orientation of the substitution at their 3 position; (b) Overlay of 9ea (yellow)and 7 (magenta)
with unsubstituted phenyl at its 2 position; (c) Conformational analysis by DFT at the b3lyp/6-31+G(d,p)
level of theory. 9ea (R1 = equatorial Ph, R2 = axial Me), 9ae (R1 = axial Ph, R2 = equatorial Me), 10ee
(R1 = equatorial Ph, R2 = equatorial Me), and 10aa (R1 = axial Ph, R2 = axial Me).
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
Thus, the computational study was initiated with C-3 aryl and smaller alkyl groups at C-4 (cf. 8,
Figure 2a). Relative to our earlier SAR studies at C-3 with no substitution at C-4, where the Aβ42 IC50
was independent of the stereochemistry at C-3, the disusbtitution both at C-3 and C-4 would provide the
opportunity to probe the effect of conformational restriction. Additionally, we wanted to explore the
effect of substitution at C-4 in the context of C-3 only substituted oxadiazine (cf. 6, Figure 2a) reported
earlier.7i
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 49
1
2
3
4
5
6
7
8
9
Based on the conformational analysis by DFT at the b3lyp/6-31+G(d,p) level of theory we prioritized the
synthesis of diastereomer 9 over 10. Conformer 9ae (C-3 pseudo axial and C-4 pseudo equatorial
orientation) of diastereomer 9 was stabilized by 4.3 kcal/mmol relative to its conformer 9ea (C-3 pseudo
equatorial and C-4 pseudo axial orientation) (Figure 2c). Additionally, conformer 9ae adopted the
desired binding mode with its C-3 aryl projecting into the unexplored binding pocket occupied by the C-2
aryl group of 7, while positioning the C-4 alkyl substituent in the binding pocket previously explored
using C-4-aryl oxadiazine (cf. 3,4,5-trifluorophenyl in 4) (Figure 2b). By comparison, diastereomer 10 is
conformationally more flexible, and the substituents at C-3 and C-4 position of these conformers either
did not project into the targeted binding pockets concurrently (10aa) or only with significant distortion of
the oxadiazine moiety (10ee) (Supporting Information, Figure S1).
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 vitro Characterization:
Consistent with the conformational analysis, analog 11 with C-3 3,5-difluorophenyl and C-4
methyl, exhibited Aβ42 IC50 of 52 nM, while the conformationally flexible analog 12 (obtained as the
undesired isomer during synthesis), was significantly less potent (Aβ42 IC50 of 9289 nM). Increasing bulk
at C4 (analog 13), addition of a fluorine into the difluorophenyl at C-3 (analog 14), and replacement of
the methyl with a polar hydroxyl methyl moiety (analog 15) were tolerated. Compounds 11 and 14 were
selected for further profiling based of their potencies, permeabilities, and lack of major off target
liabilities (Table 2). Furthermore, the ability of 11 and 14 to lower Aβ42 in plasma and CSF was assessed
using in vivo studies.
6
ACS Paragon Plus Environment

Page 7 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
Table 1. GSM Profile of 3,4-Disubstituted Oxadiazine
6
7
8
9
O
N
R1
R2
3
4
N
MeO
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
R1
R2
IC50 (nM)
IC50 (µM)
IC50 Aβ total/ IC50
Aβ42
Cpd
a,b
a
(orientation)
(orientation)
Aβ42
Aβ total
3,5-difluorophenyl
Me
(α)
11
52
12
248
NA
174
(α)
3,5-difluorophenyl
Me
(α)
12
13
9289
113
>20
19
(β)
)
3,5-difluorophenyl
i-Pr
(α)
(α)
3,4,5-difluorophenyl
Me
(α)
14
15
55
69
>20
15
>367
221
(α)
3,4,5-difluorophenyl
CH2OH
(α)
(α)
aProtocols for determination of Aβ42 IC50 (nM) and Aβ total IC50 have been previously described.7h All the
measurements are geometric mean of at least n=2 measurements.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 49
1
2
3
4
5
Strategy II - Reconfiguration of the exocyclic double bond into the more hindered β, β
disubstituted endocyclic olefin.
6
7
8
9
Analysis of previous SAR:
Along with the aforementioned efforts to introduce structural diversity, we undertook several strategies11
to mitigate the concern related to the exocyclic double bond towards nucleophilic attack that was present
in our previous GSMs. We generated multiple scaffolds following a cyclization strategy where
heterocycles were installed by tethering the distal methylene of the exocyclic double bond and the amide
carbonyl as exemplified by compounds 16 and 17 (Figure 3).11a, b While successful in masking the
double bond with minimal loss of in vitro potency, this strategy introduced an additional aromatic ring
into the scaffold. In addition, none of these analogs exhibited statistically significant CSF Aβ42 reduction
in in vivo at 30 mg/kg oral dose except for 17, which demonstrated only modest CSF Aβ42 lowering at 30
mg/kg with 45% reduction due to poor PK and lack of adequate brain exposure. In another approach,11c
the exocyclic methylene was replaced by an “NH” (cf. 18) to create an intramolecular hydrogen bond
with the adjacent amide carbonyl to mimic the 5-membered heterocycles present in 16 and 17.
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
R1
=
N
N
X
Y
distal
methylene
R2
N
R2
R1
N
N
R1
Z
N
O
F
F
R2
=
Cl
R3
R1
N
16
17
H
O
F
N
R3
2
R1
N
R3
=
F
F
F
18
Figure 3: Previous strategies to mask the exocyclic double bond.
8
ACS Paragon Plus Environment

Page 9 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
While this approach was effective in maintaining the in vitro potency without incorporating an
additional aromatic ring, it introduced unsaturation in the piperidinone moiety, and analogs from this
series generally lacked robust in vivo efficacy. For example, analog 18 with in vitro potency of 40 nM
lowered CSF Aβ42 by 40% at 100 mg/kg oral dose in rodents.
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
Design Considerations:
F
F
O
N
O
O
N
O
F
F
F
N
N
R1
N
N
R1
H
H
R1
R1
F
F
19c
19t
21
20
A IC50 = 462 nM
F
F
MeO
N
O
O
N
N
R1
=
N
OH
N
CO2Et
N
R1
R1
O
HO
IC50 = 512 nM
23
22
Figure 4: Conformational analysis and building hypothesis for the synthesis of 20 and 21. The
calculated energy difference between 19t and 19c is ~ 2.3 kcal/mmol (energy calculations were done at
the b3lyp/6-31+g(d,p) level of theory).
Therefore, in subsequent efforts, we focused on designing scaffolds without the introduction of
additional aromatic rings or unsaturation while mitigating the concern around the exocyclic double bond.
Detailed SAR analysis of close analogs of compound 2 from the literature12 led to the identification of the
acyclic analog 19 with Aβ42 IC50 of 462 nM (Figure 4). Since it is well known that the conformers with
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 49
1
2
3
4
5
6
7
8
9
global minimum energy or even local minimum energy from the potential energy surface often are not
the bioactive conformers,13 we wanted to probe the effect of constraining the cisoid conformer of 19 (cf.
19c) by linking the amide NH with the methylene distal to amide (bottom red double headed arrow on
19c) on the in vitro potency. This arrangement forms a sterically encumbered endocyclic β, β-
disubstituted double bond with reduced susceptibility for nucleophilic addition.14 While such strategy
could be perceived as high risk due to presence transoid conformation among all the previously reported
GSMs, it was inspired by the serendipitous discovery of 23 with an IC50 of 512 nM during our SAR
studies in the oxadiazoline series (Figure 4). Among several approaches to constrain the cisoid
conformation 19c, azepines 20 and 21 were prepared.
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 vitro Characterization:
Azepines 20 and 21 displayed cell Aβ42 IC50s of 107 nM and 143 nM respectively while
maintaining a good modulator profile as shown by selectivity vs. Aβtotal (Table 2). These analogs were
highly permeable and exhibited reduced binding to the hERG K+ channel relative to analog 2 (IC50 = 990
nM) in a radiolabeled MK-499 binding15 assay and possessed no major off target liability (Table 2).
These compounds were selected for characterization in a rat pharmacodynamic model to assess their
ability to reduce CSF Aβ42 in vivo.
10
ACS Paragon Plus Environment

Page 11 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
Strategy III: Removal of the olefinic bond without incorporation of additional aromaticity
Design considerations and validation as GSM:
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
Our third strategy to mitigate the concern around the reactivity of the exocyclic double bond
began with careful consideration of the structural features of 2 and 4 (Figure 5). This analysis revealed
that the lone pair on the oxygen (represented as red) of the amide carbonyl of 2 projected toward the
fluorobenzyl moiety was not necessary to maintain potency as it was encapsulated by the formation of
oxadiazine ring (2→4). The importance of the other lone pair (represented as blue) oriented toward the
exocyclic double bond was unknown. Generation of analog 2411a with IC50 of 116 nM (racemic)
eliminating the nitrogen lone pair oriented toward the exocyclic double bond via formation of the 5-
membered heterocycles underscored its lack of importance for binding affinity (comparison of IC50s
between 24 and 25). This SAR suggests that the amide carbonyl lone pairs of 2 or the of N-1 nitrogen
lone pair of triazole 24 may not be playing a significant role on the binding affinity via hydrogen bonding
or ionic interactions. Recent reports by Petterssen et al. from Pfizer,7g where the exocyclic double bond
was replaced by the isosteric amide, further suggested that the amide carbonyl lone pair of 2 was either
not important or the distance between the H-bond acceptor and the methoxyaryl-imidazole right-hand-
side was flexible.
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 49
1
2
3
4
5
6
7
8
9
a
O
N
1
N
2
N
b
N
O
a
F
F
O
N
b
O
N
F
F
O
N
N
N
N
4
N
F
N
N
F
F
24
2
IC50 = 64 nM
IC50 = 116 nM (racemic)
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
4
, Ar = 4-fluoro-phenyl, IC50 = 33 nM
25
, Ar = 3,4,5-fluorophenyl, IC50 = 37 nM
O
N
O
O
N
N
CF3
O
N
O
F
O
N
O
N
N
N
O
N
N
F
Cl
N
N
F
28
IC50 = 15 nM
26
27
IC50 = 400 nM
IC50 = 1554 nM
Figure 5. Analysis on the importance of the lone pairs of the amide carbonyl of 2 for potency and
building hypothesis for the synthesis 26 and 27 via deconstruction of 24.
To probe this hypothesis, prototype compound 26 with IC50 of 1,554 nM (racemic) was generated
by replacing the triazole in 24 with an amide moiety. Importantly, 26 lacked the lone pairs that were
initially present at the N1 position in 24. Furthermore, replacement of the amide of 26 by an oxadiazine
along with the introduction of conformational constraint provided analog 27 with an IC50 of 400 nM
(racemic). The increased sp3 character and reduced aromatic ring count relative to 24 together with the
lack of the exocyclic double bond, compound 27 represented a promising novel chemotype for GSM.
Further improvement in potency may be necessary before in vivo characterization of this series.
Moreover, independently a recent report revealed that enhancement of potency was feasible via SAR
studies related to the chemotype exemplified by compound 27. For example, analog 28 exhibits
improved potency and in vivo efficacy.7a
12
ACS Paragon Plus Environment

Page 13 of 49
Journal of Medicinal Chemistry
1
2
3
Table 2. Collective Profile of 11, 14, 20, and 21
4
5
6
7
8
9
Measured parameters
Compounds
20
11
14
21
143
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
IC50 (nM) Aβ42
IC50 (µM) Aβtotal
Papp (10-6cm/s)
hERG K+ MK-499
Binding IC50 (nM)
PXR (μM)
52
55
107
12
17
>20
14
14
16
23
22
2007
1607
3352
2608
15
16; 33; 49
73
>10
>30
>15
30; 8.0; 8.0
NS
3A4; 2C9; 2D6 (μM)
29; 37; 49 34; 18; 48
In vivo rat CSF Aβ42
reduction (30 mg/kg)
CP (µM); CB (µM)a
80
52
8.8; 7.5
13.1; 11.1
14.2; 2.1
1.4; 1.3
aMeasured total concentrations in plasma (CP), and brain (CB) of rats at 3h post oral dose of 30
mg/kg. Papp, apparent permeability; hERG K+, Human ether-a-go-go-related gene (hERG) potassium
channels; PXR, Pregnane X receptor.
In vivo characterization:
Concurrent with our efforts to identify novel structural motifs described above, analogs 11 and 14 from
the disubstituted oxadiazine series (Table 1) as well as azepines 20 and 21 (Figure 4) were evaluated to
measure their ability to lower CSF Aβ42 in vivo at a 30 mg/kg oral dose after 3 h drug administration
using male CD rats (Table 2). Analogs 11 and 14 exhibited robust CSF Aβ42 lowering (73% and 80%
respectively) at 3 h post dose. Azepine 20 showed 52% lowering of CSF Aβ42 while no significant
reduction was observed with azepine 21. In a dose response study, compound 11 showed a dose
dependent effect on Aβ42 lowering from 3-30 mg/kg oral dose while no significant reduction was
observed with 1 mg/kg dose relative to control levels (Figure 6a). Approximated free drug concentration
in the brain at 3 h post dose calculated using plasma free fraction showed that the concentrations of 11
are within 2-3 fold relative to its cellular IC50 for 30 and 10 mg/kg oral dose (37 nM and 21 nM for 30
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 49
1
2
3
4
and 10 mg/kg dose respectively). But, the free drug concentration of 11 is much lower at 3 mg/kg dose
5
6
(4 nM).16
7
8
9
11: Dose Response
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
15
10
5
100
80
60
40
20
0
Dose
(mg/kg)
30
CSF Aβ42
reduction
73
CP
CB
Plasma Concentration
Brain concentration
CSF
(µM) (µM)
8.88
4.83
0.91
7.51
4.2
10
3
52
24
0.81
0
100
1
10
3
30
(a)
(b)
Figure 6. (a) Dose dependent reduction of Aβ42 by compound 11 in male CD rats. The value of
CSF Aβ42 is the geometric mean of 29 (for 3 mg/kg and 10 mg/kg dosing group) and 10 animals (30
mg/kg dosing group) respectively, and the value of CP and CB is the geometric mean of 10 (3 mg/kg
dosing group), 15 (10 mg/kg dosing group) and 5 (30 mg/kg dosing group) animals. (b) Measured total
concentration of 11 in plasma (CP), and brain (CB) of rats at doses ranging from 3 mg/kg to 30 mg/kg.
All animal procedures were performed according to the protocol reviewed and approved by the
institutional Animal Care and Use Committee of Merck & Co., Inc., Kenilworth, NJ, USA.
As shown in Table 3, compound 11 displayed oral bioavailability of 79% with a Cmax and total AUC of
1.3 μM and 8 μM·h respectively with 2 mg/kg oral dose. Additionally, it exhibited low in vivo clearance
and moderate volume of distribution that resulted in a MRT of 1.7 h.
14
ACS Paragon Plus Environment

Page 15 of 49
Journal of Medicinal Chemistry
1
2
3
Table 3. Rat pharmacokinetic parameters of compound 11a
4
5
6
7
8
MRT
(h)
Pgp-Efflux
(rat)d
AUC(0-∞)
(µM·h)b
Cmax
PPB
(% bound)
Cl
Vd,ss
F%b
79
(µM)b
(mL/min/kg)c
(L/kg)c
2.46
8
1.3
99.5
10
1.1
1.7
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
b
c
aValues are average of two experiments. Data from 2 mg/ kg oral dose Data from 1 mg/ kg iv dose.
dRat Pgp Effux was measured in LLC-PK1 cells transfected with rat MDR1A and reported as a ratio of
(B to A)/(A to B). AUC, area under the plasma drug concentration-time curve; Cmax, highest
concentration; F%, oral bioavailability; PPB, plasma protein binding; Cl, clearance; Vd, volume of
distribution; MRT, mean residence time; Pgp, p-glycoprotein.
Conclusions
Modulation of γ-secretase to probe the amyloid cascade hypothesis was one of the multipronged
approaches that we had undertaken, and our previous work resulted in the identification of two distinct
GSM series represented by compound 3 and 4. To enhance structural diversity and to mitigate concern
around the exocyclic double bond present in oxadiazine series, we pursued a number of strategies to
identify new leads. Three distinct strategies were inspired by existing chemical matter and executed upon
using a “hypothesis driven drug discovery” approach, which is complementary to lead identification via
traditional approaches such as uHTS, ALIS and FBLD screens. The conceptualization of the 3,4-
disubstituted oxadiazine series was inspired by compound 7 reported in the patent literature.
Representative compounds from this series demonstrated robust reduction of CSF Aβ42 as exemplified by
compounds 11 and 14. In a further attempt to mask the exocyclic double bonds present in these
molecules, analogs 20 and 21 with a β, β- disubstituted endocyclic double bond were designed,
synthesized and validated as novel GSMs, and analog 20 exhibited 52% lowering of CSF Aβ42 with 30
mg/kg oral dose. Based on the dose responsive study with 11, we surmised that the free concentration of
compounds in the brain is required to be near their IC50 for robust lowering of CSF Aβ42. Finally,
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 49
1
2
3
4
5
6
7
8
9
discovery of 27, lacking both the exocyclic double bond and the additional aromatic ring led to a novel
GSM chemotype. A recent report from the literature confirmed that such molecular motifs embodied the
potential for the generation of orally bioavailable molecules that might ultimately translate into novel
therapeutics for the treatment of Alzheimer’s disease.7a Additional investigation of these GSMs in the
context of newly developed cryo-electron microscopy structures of human γ-secretase in complex with a
Notch fragment and transmembrane APP fragment would provide the opportunity to further optimize
these molecules.17
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
Chemistry
Our initial route for the synthesis of 11, 12 and related analogs commenced with the three-step
synthesis of aldehyde 31 from the known intermediate 29 (Scheme 1).7h The reaction of 31 with (3,5-
difluorophenyl)magnesium bromide led to alcohol 32 as a diastereomeric mixture. Subjection of 32 to
three-step sequence, namely Mitsunobu reaction with N-hydroxyphthalimide, phthalimide removal and
H3PO4 mediated dehydrative cyclization resulted in analogs 11 and 12. A detailed investigation revealed
that the addition of (3,5-difluorophenyl)magnesium bromide generated the syn isomer as the major
product via chelation control, albeit with low diastereoselectivity (step d). Furthermore, use of the pure
diastereomer separated from the mixture of diastereomers represented as 32 in the Mitsunobu reaction
with N-hydroxyphthalimide resulted in complete retention of stereochemistry at the alcohol bearing
center. We surmised that this retention of stereochemistry resulted from double inversion due to the
neighboring group participation by the proximal amide moiety.
Scheme 1. Initial approach to 3,4- disubstituted oxadiazines 11 and 13-15 based on Mitsunobu
reaction.
16
ACS Paragon Plus Environment

Page 17 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
H
O
O
Ar
O
O
O
OH
Me
d
a, b, c
R1
+
R1
N
Me
OH
Cl
R1
N
Me
H2N
32
30
NPhth
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
29
31
O
N
Ar
O
N
Ar
O
Ar
N
N
O
e
f, g
+
R1
Me
R1
R1
Me
N
Me
11
33
12
MeO
N
F
R1
=
=
Ar
N
F
Reagents and conditions: (a) EDCI, HOBT, DIEA, DMF; (b) NaOMe (25 wt.% in MeOH), THF, 25%;
(c) (COCl)2, DMSO, DCM, –78 °C, 100%; (d) 3,5-difluorophenyl magnesium bromide, THF, 48%; (e)
N-hydroxyphthalimide, n-Bu3P, ADDP, 2MeTHF, 18%; (f) NH2NH2, EtOH; (g) H3PO4, n-BuOH, 100
°C, 29% yield for 11, and 8% yield for 12.
Based on these findings, modified synthetic sequences were envisioned to access 11 and close
analogs rapidly (Scheme 2). The amino alcohol 35 obtained in three-step from 34 was condensed with
acid 29. Subsequent reaction of the resulting amide with NaOMe led to the cyclized compound 36.
Mitsunobu reaction of 36 with N-hydroxyphthalimide followed by hydrazine mediated phthalimide
removal and subsequent H3PO4 mediated dehydrative cyclization led to the formation of 11 with overall
retention in configuration at the alcohol bearing center. Analogs 14 were prepared following this general
synthetic sequence.
Scheme 2. Stereoselective approach to 3,4-disubstituted oxadiazines based on the observation of
retention of configuration Mitsunobu reaction in Scheme 1.
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 49
1
2
3
4
5
6
7
8
9
H
O
Ar
R
O
d, e
Bn2N
R
H2N
OH
OH
Ar
MeO
a, b, c
N
+
29
R
N
N
35
Ar
R
36
34
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
N
N
3,5-difluorophenyl
3,5-difluorophenyl
3,4,5-trifluorophe
=
Ar
Me,
=
=
=
11, R
f, g, h
MeO
isopropyl,
13 R
,
nyl
=
Ar
Me,
14, R
N
N
Reagents and conditions: (a) (COCl)2, DMSO, DCM, –78 °C; (b) 3,5-difluorophenyl magnesium
bromide, THF, 100%; (c) Pd/C, H2, MeOH, 84%; (d) EDCI, HOBT, DIEA, DMF; (e) NaOMe (25 wt.%
in MeOH), THF, 81%; (f) N-hydroxyphthalimide, n-Bu3P, DIAD, THF, 76%; (g) NH2NH2, EtOH; (h)
H3PO4, n-BuOH, 100 °C, 51%.
The synthesis of hydroxymethylated analog 15 is described in (Scheme 3). The key intermediate
38 was prepared in three steps from alcohol 37 via Swern oxidation, non-chelation controlled 3,4,5-
trifluorophenyl Grignard addition and selective removal of benzyl groups from the amino functionality.
Amino alcohol 38 was progressed to compound 40 following the reaction sequence described in Scheme
2. Treatment of 40 with BCl3 to remove the benzyl group led to 15.
Scheme 3. Synthesis of analog 15 based on the observation of retention of configuration Mitsunobu
reaction.
18
ACS Paragon Plus Environment

Page 19 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
F
F
F
F
F
O
f, g, h
HO
a, b, c
d,e
R1
HO
HO
R1
+
OH
Cl
O
F
8
9
Bn2N
N
H2N
OBn
OBn
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
OBn
38
37
29
39
F
F
F
F
F
MeO
N
i
O
O
N
N
R1
=
N
F
N
R1
R1
N
OBn
OH
15
40
Reagents and conditions: (a) (COCl)2, DMSO, DCM, –78 °C; (b) 3,5-difluorophenyl magnesium
bromide, THF, 94%; (c) Pd/C, H2, MeOH, 100%; (d) EDCI, HOBT, DIEA, DMF; (e) NaOMe, MeOH,
66%; (f) N-hydroxyphthalimide, PPh3, DIAD, THF, 0 °C; (g) NH2NH2, EtOH, 48%; (h) H3PO4, n-BuOH,
100 °C, 86%; (i) BCl3, DCM, 0 °C, 23%.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 49
1
2
3
4
Scheme 4. Synthesis of key intermediate 47 for synthesis of azepine analogs 20 and 21
5
6
7
8
O
O
CO2t-Bu
MeO
N
Br
d
c
Me3Sn
Me3Sn
a
b
CO2t-Bu
Cl
OtBu
+
Cl
9
N
Me3Sn
Cl
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
Cl
44
43
45
41
42
O
O
OH
CO2t-Bu
Cl
e
MeO
Cl
MeO
N
N
N
N
46
47
Reagents and conditions: (a). n-BuLi, Et2O, THF, –30 °C, (Boc)2O, 34%; (b). Me6Sn2, Pd(Ph3)4, 80 °C,
THF, 77%; (c). CuCl, DMF, H2O, rt, 72%; (d). CuCl, Pd(dppf)Cl2, DMF, 100 C; (e) TFA, DCM, rt,
o
68%.
The synthesis of acid 47 for the preparation of analogs 20 and 21 was illustrated in (Scheme 4).
Alkyne 42 prepared in a single step from 41 was treated with hexamethylditin to provide intermediate 43.
Treatment of 43 with CuCl resulted in selective removal of trimethyltin α to the ester group. Stille
coupling between 44 and 45 followed by treatment of the resulting intermediate with TFA provided the
key acid 47, which was used for the synthesis of both core structure 20 and 21.
20
ACS Paragon Plus Environment

Page 21 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
Scheme 5. Failed attempt to transform the hydroxyl functionality to hydroxylamine for the
synthesis of 20.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
MeO
H
O
N
O
Ar
O
Ar
a
R =
O
O
47
+
t-BuMe2SiO
H2N
Ar
N
N
X
N
N
R
R
Ar = 3,4-5-trifluorophenyl
49
48
50
Reagents and conditions: (a) 1. N-hydroxyphthalimide, nBu3P, DIAD, THF; 2. N-hydroxyphthalimide,
nBu3P, ADDP, THF.
With the acid 47 in hand, the initial route for the synthesis of 20 involved the synthetic sequences
described in (Scheme 5). And the 7-membered lactam 49 was prepared for the subsequent conversion of
its hydroxyl to hydroxyl amine moiety functionality. However, the conversion of 49 to 50 under a
variety of conditions was met with failure.
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 49
1
2
3
4
Scheme 6. Synthesis of 20 via development of new synthetic strategy.
5
6
7
8
NPhth
BocHNO
Ar
O
Ar
HO
Ar
HO
Ar
b
e
a
c, d
9
(Allyl)2N
(Allyl)2N
(Allyl)2N
H2N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
54
53
51
52
BocHN
O
BocHN
O
O
Ar
O
N
Ar
O
Ar
N
BocHNO
Ar
g
f
h, i
N
N
H
47
+
R
R
Cl
H2N
R
55
58
56
57
MeO
O
N
Ar
O
N
Ar
N
N
j
N
N
+
R =
R
R
Ar = 3,4,5-trifluorophenyl
20
59
Reagents and conditions: (a). Allyl bromide, K2CO3, n-butyronitrile, 120 °C, 50%; (b) N-
hydroxyphthalimide, n-Bu3P, DIAD, THF, 0 °C; (c) NH2NH2, DCM, EtOH; (d) (Boc)2O, rt, 65%; (e)
1,3-Dimethylbarbituric acid, Pd(PPh3)4, DCM, rt, 84%; (f) EDCI, HOBt, DIEA, DMF, 0 °C, 85%; (g)
KHMDS, THF/DMF, –50 °C, 56%; (h) TFA, DCM, rt; (i) H3PO4, 105 °C; (j) Chiralpak AD column,
22% for 20, 17% for 59.
The synthetic strategy was revised to preinstall the hydroxylamino moiety into the amino alcohol
51 (c.f. 55) prior to formation of 7-membered lactam (Scheme 6). The amino moiety of 51 was protected
with the bis-allyl group for facile generation of free amine 55 without cleavage of the N-O bond
(54→55). Condensation of amine 55 with acid 47 followed by cyclization using KHMDS led to the 7-
membered lactam 57. Removal of the Boc group followed by treatment of the resulting free
hydroxylamine with H3PO4 afforded oxadiazine 58. The two enantiomers of 58 were separated using the
AD column with 30% i-PrOH in hexanes to afford 20.
22
ACS Paragon Plus Environment

Page 23 of 49
Journal of Medicinal Chemistry
1
2
3
Scheme 7. Unanticipated transformation of 61 to 62 leading to stereocontrolled synthesis of 20.
4
5
6
7
8
9
As in
scheme 6
HO
F
HO
O
F
F
b, c, d
a
F
F
F
F
N
O
20
H2N
(Allyl)2N
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
F
(Allyl)2N
F
60
62
61
Reagents and conditions: (a). Allyl bromide, K2CO3, MeCN, 80 °C, 38%; (b) N-hydroxyphthalimide, n-
Bu3P, DIAD, THF, 0 °C; (c) NH2-NH2, EtOH, DCM, 0 °C; (d) (BOC)2O, 37%.
We recognized that enantiomerically pure amino alcohol 60 may be useful for the synthesis of 21
with the 3,4,5-trifluorophenyl moiety at the 4-position of the oxadiazine moiety (Scheme 8). The route
began with the transformation of amino alcohol 60 to the bis-allyl protected amino alcohol 61.
Interestingly, under Mitsunobu reaction conditions, the bis-allyl protected amino alcohol 61 was
converted to 62 with apparent migration of the 3,4,5-trifluorophenyl from the nitrogen to the oxygen
bearing carbon of the allylated amino alcohol 61. This transformation occurred presumably via in situ
formation of aziridinium ion followed by regio and stereoselective opening of the aziridinium ring by the
N-hydroxyphthalimide. Although intermediate 62 was not useful for the synthesis of 21, it could be used
for the synthesis of enantiopure 20 following the exact sequences of reactions described in Scheme 6
(details in the supporting information).
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 49
1
2
3
4
Scheme 8. Stereocontrolled synthesis of 21
5
6
7
8
NPhth
O
BocHNO
H2N
HO
BocHNO
b, c
a
d
47
+
9
Ar
CbzHN
Ar
CbzHN
Ar
CbzHN
Ar
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
66
64
65
63
BocHN
ONHBoc
O
O
O
f
e
g
Ar
21
Ar
Cl
N
N
H
MeO
N
MeO
N
Ar = 3,4,5-trifluorophenyl
N
N
68
67
Reagents and conditions: (a) N-hydroxyphthalimide, Ph3P, DIAD, THF, –30 °C, 98%; (b) NH2NH2,
DCM, EtOH, 88%; (c) DCM, (Boc)2O, rt, 80%; (d) Pd/C, H2, EtOH, rt, 100%; (e) EDCI, HOBt, DIEA,
DMF, 0 °C, 65%; (f) KHMDS, DMF, –30 °C, 64%; (g) POCl3, 80 °C, 20%.
For the synthesis of 21, the amino moiety of amino alcohol 60 was protected with Cbz to
attenuate its nucleophilicity, thereby preventing the formation of the aziridinium ion in the following
Mitsunobu reaction (Scheme 8). The Cbz protected alcohol 63 underwent Mitsunobu reaction in a
desired fashion without rearrangement to generate 64. Protecting group manipulation followed by
removal of Cbz afforded amine 66, which was progressed to 21 following the similar sequences of
reaction as described in Scheme 6.
24
ACS Paragon Plus Environment

Page 25 of 49
Journal of Medicinal Chemistry
1
2
3
Scheme 9. Racemic synthesis of 27.
4
5
6
7
8
EtO
O
EtO
O
EtO
O
ref 20
O
a
b, c
O
O
F
OH
OH
BocN
Br
BocN
BocN
+
EtO
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
71
OH
F
69
70
72
OH
OH
O
O
d
e
O
F
O
F
HN
+
O
F
R
N
R
N
R
OH
F
F
74
76
F
75
73
NH2
O
O
N
MeO
N
O
O
F
h
f, g
R =
O
F
R
N
R
N
N
F
27
F
77
Reagents and conditions: (a) 3,5-Difluorophenol, DEAD, PPh3, THF, 0 °C; (b) LiAlH4, , THF, 42%; (c)
4N HCl in dioxane, rt; (d) EDCI, HOBt, DIEA, DMF, 50%; (e) Pd/C, H2, MeOH, 57%; (f) MeSO2Cl,
Et3N, N-hydroxyl phthalimide, DCM, 43%; (g) NH2NH2, EtOH, DCM, 63%; (h) POCl3, 90 °C, 22%.
Synthesis of 27 began with the ethyl bromoacetate 69, which was progressed to racemic 70 as the
major diastereomer following the literature procedure (Scheme 9).18 Mitsunobu reaction of 70 with 3,5-
difluorophenol resulted in 72 via inversion of stereochemistry at the carbon bearing alcohol. Reduction
of ester functionality in 72 to the corresponding alcohol followed by removal of the Boc group provided
73. Condensation of 73 with acid 74 followed by hydrogenation led to 76. Subjection of intermediate 76
to Mitsunobu reaction conditions with N-hydroxyphthalimide followed by hydrazine mediated
phthalimide removal generated 77, which on treatment with POCl3 resulted in 27.
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 49
1
2
3
4
EXPERIMENTAL SECTION
5
6
7
8
General. Measurement of Aβ42 IC50, Aβtotal IC50 and BACE1 Ki and determination of rat CSF
Aβ42 lowering in vivo were carried out following our previously described protocols.7g
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
Unless otherwise mentioned, all reagents and solvents purchased from commercial source were
used directly without further purification. Air sensitive reactions were performed under an atmosphere of
N2. Purification of all the final compounds to >95% purity was carried out either by prepacked silica gel
cartridge (Analogix, Biotage, or ISCO) or on a reverse phase C18 column. Water and acetonitrile used as
mobile phase A and B respectively for the C18 column were supplemented either with 0.05% TFA or
0.1% formic acid. All the NMR data were collected either at 400 or 600 MHz on a Varian or Bruker
instrument. Chemical shifts are reported in ppm relative to the residual solvent peak in the indicated
1
solvent, and for H NMR, multiplicities, coupling constants in Hertz, and numbers of protons are
indicated parenthetically. 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 supplemented with
0.05% TFA or 0.1% formic acid as the mobile phase. The purity of the samples was assessed using a UV
detector at 254 nm.
(3R,4S)-3-(3,5-Difluorophenyl)-9-((E)-3-methoxy-4-(4-methyl-1H-imidazol-1-yl)benzylidene)-4-
methyl-3,4,6,7,8,9-hexahydropyrido[2,1-c][1,2,4]oxadiazine (11) and (3S,4S)-3-(3,5-difluorophenyl)-
9-((E)-3-Methoxy-4-(4-methyl-1H-imidazol-1-yl)benzylidene)-4-methyl-3,4,6,7,8,9-
hexahydropyrido[2,1-c][1,2,4]oxadiazine (12)
To a solution of TFA salt of 29 (5 g, 11 mmol), HOBt (2.04 g, 13.3 mmol) and 30 (1 g, 13.3
mmol) in DMF (30 mL) was added EDCI (2.56 g, 13.3 mmol) followed by DIEA (4.28 mL, 24.5 mmol)
at 0°C. The resulting mixture was slowly warmed to rt, and stirred for 2 h. Then the reaction mixture
was diluted with EtOAc, transferred to a separatory funnel, washed with saturated aqueous solution of
26
ACS Paragon Plus Environment

Page 27 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
NaHCO3 and then with brine. The organic layer was dried with MgSO4, filtered, concentrated, and the
crude product was taken to the next step without further purification. To this crude product in THF (30
mL) was added NaOMe (25 wt. % in MeOH, 3.61 g, 16.71 mmol) at 0 °C. The resulting mixture was
slowly warmed to rt, and stirred for 14 h. Then the reaction mixture was diluted with EtOAc, neutralized
with 1N HCl and washed with brine. The organic layer was dried with MgSO4, filtered, evaporated, and
the crude product was purified by silica gel chromatography using 0–100% EtOAc in hexanes to afford
alcohol precursor to aldehyde 31 (1 g, 25 % yield).
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
To a solution of oxalyl chloride (0.49 mL, 5.63 mmol) in DCM (4 mL) was added a solution of
DMSO (0.79 mL, 11.3 mmol) in DCM (2 mL) at −78 °C. The reaction mixture was stirred at –78 °C for
10 min, then to this mixture was added a solution of alcohol precursor to aldehyde 31 (1 g, 2.8 mmol) in
DCM (4 mL) and stirred at –78 °C for 30 min. Then was added Et3N (1.96 mL, 14 mmol), and the
resulting solution was slowly warmed to rt over 1 h. The reaction mixture was diluted with EtOAc,
transferred to a separatory funnel, washed first with saturated aqueous solution of NaHCO3 and then with
brine. The organic layer was dried with MgSO4, filtered, evaporated, and the crude aldehyde 31 was
taken directly to the next step without further purification.
To a solution of this crude 31 (0.99 g, 2.8 mmol) in THF (10 mL) was added a solution of (3,5-
difluorophenyl)magnesium bromide (2.81 mL, 5.6 mmol, 2M in diethyl ether) at 0 °C, and the resulting
mixture was slowly warmed to rt and stirred for 2 h. Then the reaction mixture was diluted with EtOAc,
transferred to a separatory funnel, neutralized with 1N HCl and washed with brine. The organic layer
was separated, dried with MgSO4, filtered, evaporated and the crude product was purified by silica gel
chromatography using 0–10 % methanol in DCM to provide 32 as the mixture of diastereomers (0. 63 g,
48 % yield).
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 49
1
2
3
4
5
6
7
8
9
To a solution of 32 (0.17 g, 0.36 mmol) in 2-MeTHF (5 mL) was added N-hydroxyphthalimide
(0.10 g, 0.73 mmol), ADDP (0.12 g, 0.47 mmol) and n-Bu3P (0.11 mL, 0.47 mmol). The resulting
mixture was degassed and heated at 80 °C for 18 h. The reaction mixture was diluted with EtOAc,
transferred to a separatory funnel, washed with saturated aqueous solution of NaHCO3 and then with
brine. The organic layer was dried with MgSO4, filtered, concentrated, and the crude product was
purified by silica gel chromatography using 1% Et3N in EtOAc to provide corresponding Mitsunobu
adduct 33 (0.04 g, 18 % yield).
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
To a solution of 33 (0.04 g, 0.06 mmol) in EtOH (0.2 mL) was added hydrazine (0.07 mL, 2.23
mmol) at 0 °C, and the resulting solution was slowly warmed to rt and stirred for 30 min. The reaction
mixture was diluted with EtOAc, transferred to a separatory funnel, washed with saturated aqueous
solution of NaHCO3 and brine, and the crude product was taken to the next step without further
purification.
To a solution of this crude product (0.03 g, 0.06 mmol) in butan-1-ol was added H3PO4 (0.01 mL,
0.17 mmol) and heated at 100 °C for 12 h. The reaction mixture was concentrated, and the crude reaction
mixture was purified by reverse phase HPLC to provide 11 (0.01 g, 29 % yield) and 12 (0.002 g, 7.83 %
1
yield). H NMR and m/z of compound 11 (400 MHz, CD3OD) δ 9.19 (d, J = 1.6 Hz, 1H), 7.67 – 7.58
(m, 2H), 7.47 (s, 1H), 7.36 (d, J = 1.5 Hz, 1H), 7.27 (dd, J = 8.2, 1.2 Hz, 1H), 7.10 – 6.98 (m, 3H), 5.34
(d, J = 2.2 Hz, 1H), 4.13 (qd, J = 6.6, 2.4 Hz, 1H), 3.97 (s, 3H), 3.78 (m, 1H), 3.58 (m, 1H), 3.04 – 2.79
(m, 2H), 2.44 (d, J = 1.1 Hz, 3H), 2.21 – 1.95 (m, 2H), 1.15 (d, J = 6.6 Hz, 3H). m/z: 465. 1H NMR and
m/z of compound 12 (400 MHz, CD3OD) δ 9.16 (d, J = 1.6 Hz, 1H), 7.63 – 7.51 (m, 2H), 7.35 – 7.26
(m, 2H), 7.22 – 7.12 (m, 3H), 7.04 (tt, J = 8.9, 2.2 Hz, 1H), 5.16 (d, J = 3.6 Hz, 1H), 4.16 (td, J = 6.5, 3.7
Hz, 1H), 3.93 (s, 3H), 3.68 (m, 1H), 3.46 (m 1H), 2.79 (m, 2H), 2.42 (d, J = 1.1 Hz, 3H), 2.07 – 1.78 (m,
2H), 1.47 (d, J = 6.5 Hz, 3H). m/z: 465.
28
ACS Paragon Plus Environment

Page 29 of 49
Journal of Medicinal Chemistry
1
2
3
4
5
(3R,4S)-3-(3,5-Difluorophenyl)-9-((E)-3-methoxy-4-(4-methyl-1H-imidazol-1-yl)benzylidene)-4-
methyl-3,4,6,7,8,9-hexahydropyrido[2,1-c][1,2,4]oxadiazine (11) according to the synthetic route
described in Scheme 2.
6
7
8
9
To a solution of oxalyl chloride (2 mL, 23.5 mmol) in DCM (100 mL) was added a solution of
DMSO (2.78 mL, 39.2 mmol) in DCM (10 mL) at –78 °C and stirred for 10 min. Then was added a
solution of 34 (5 g, 19.5 mmol) in DCM (10 mL) and stirred at –78 °C for 0.5 h before the addition of
Et3N (10.9 mL, 78 mmol). The reaction mixture was slowly warmed to 0 °C, diluted with EtOAc,
transferred to a separatory funnel, washed with saturated aqueous solution of NaHCO3 and then with
brine. The organic layer was dried with MgSO4, filtered, concentrated, and the crude aldehyde was taken
to the next step without further purification.
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
To a solution of (3,5-difluorophenyl)magnesium bromide (100 mL, 50 mmol, 0.5M in diethyl
ether) was added a solution of the crude aldehyde from the previous step (4.81 g, 19 mmol) in THF (30
mL) at –78 °C, and the reaction mixture was slowly warmed to rt and stirred for 30 min. The reaction
mixture was diluted with EtOAc, transferred to a separatory funnel, washed with saturated aqueous
solution of NaHCO3 and then with brine. The organic layer was dried with MgSO4, filtered,
concentrated, and the crude product was purified by silica gel chromatography using 0–10% EtOAc in
hexanes to afford corresponding alcohol (6.98 g, 100% yield).
To a solution of this alcohol (6.98 g, 19 mmol) in MeOH (100 mL) was added 10% Pd/C (2 g, 1.8
mmol), and the resulting solution was degassed and stirred under the atmosphere of H2 for 12 h. The
reaction mixture was filtered, and the filtrate was evaporated to dryness to afford 33 (3 g, 84 % yield)
which was taken to the next step without further purification.
To a mixture of 35 (3 g, 16 mmol), TFA salt of 29 (7.19 g, 16 mmol) and HOBt (2.95 g, 19.2
mmol) in DMF (32 mL) was added EDCI (3.69 g, 19.2 mmol) followed by DIEA (6.16 mL, 35.3 mmol)
at 0 °C. The resulting mixture was slowly warmed to rt and stirred for 2 h. The reaction mixture was
29
ACS Paragon Plus Environment