Discovery of 2-methylpyridine-based biaryl amides as γ-secretase modulators for the treatment of Alzheimer's disease

Article abstract of DOI:10.1016/j.bmcl.2013.09.041

γ-Secretase modulators (GSMs) are potentially disease-modifying treatments for Alzheimer's disease. They selectively lower pathogenic Aβ42 levels by shifting the enzyme cleavage sites without inhibiting γ-secretase activity, possibly avoiding known adverse effects observed with complete inhibition of the enzyme complex. A cell-based HTS effort identified the sulfonamide 1 as a GSM lead. Lead optimization studies identified compound 25 with improved cell potency, PKDM properties, and it lowered Aβ42 levels in the cerebrospinal fluid (CSF) of Sprague-Dawley rats following oral administration. Further optimization of 25 to improve cellular potency is described.

Full text of DOI:10.1016/j.bmcl.2013.09.041

Bioorganic & Medicinal Chemistry Letters 23 (2013) 6447–6454
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of 2-methylpyridine-based biaryl amides as
modulators for the treatment of Alzheimer’s disease
c-secretase
a
a
a
a
a
Jian Jeffrey Chen a, , Wenyuan Qian , Kaustav Biswas , Chester Yuan , Albert Amegadzie , Qingyian Liu ,
Thomas Nixey a, Joe Zhu a, Mqhele Ncube a, Robert M. Rzasa a, Frank Chavez Jr. a, Ning Chen a,
Frenel DeMorin a, Shannon Rumfelt a, Christopher M. Tegley a, Jennifer R. Allen a, Stephen Hitchcock a,
Randy Hungate a, Michael D. Bartberger a, Leeanne Zalameda a, Yichin Liu a, John D. McCarter a,
Jianhua Zhang b, Li Zhu b, Safura Babu-Khan b, Yi Luo b, Jodi Bradley b, Paul H. Wen b, Darren L. Reid c,
Frank Koegler d, Charles Dean Jr. d, Dean Hickman e, Tiffany L. Correll f, Toni Williamson b,d, Stephen Wood b
⇑
a Department of Therapeutic Discovery, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
b Department of Neuroscience, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
c Pharmaceutical R&D, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
d Department of Comparative Biology and Safety Sciences, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
e Department of Pharmacokinetics and Drug Metabolism, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
f Analytical Research & Development, Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
c
-Secretase modulators (GSMs) are potentially disease-modifying treatments for Alzheimer’s disease.
They selectively lower pathogenic Ab42 levels by shifting the enzyme cleavage sites without inhibiting
-secretase activity, possibly avoiding known adverse effects observed with complete inhibition of the
Received 9 August 2013
Revised 12 September 2013
Accepted 13 September 2013
Available online 29 September 2013
c
enzyme complex. A cell-based HTS effort identified the sulfonamide 1 as a GSM lead. Lead optimization
studies identified compound 25 with improved cell potency, PKDM properties, and it lowered Ab42 levels
in the cerebrospinal fluid (CSF) of Sprague-Dawley rats following oral administration. Further optimiza-
tion of 25 to improve cellular potency is described.
Keywords:
c
c
-Secretase modulator (GSM)
-Secretase inhibitor (GSI)
Ó 2013 Elsevier Ltd. All rights reserved.
Alzheimer’s disease (AD)
Amyloid b-protein (Ab)
Methylpyridine
Amide as sulfonamide replacement
Alzheimer’s disease (AD) is the most common dementia affecting
more than 5 million patients in the US. According to the amyloid
hypothesis,1–3 aggregation of the amyloid b-protein (Ab), particu-
larly its 42-residue variant (Ab42), plays a direct role in the
pathogenesis of AD, and therefore agents that could lower Ab levels
in the brain should be a useful treatment for AD. Ab is generated
through sequential cleavage of amyloid precursor protein (APP) by
lowering pathogenic Ab42 levels by shifting the enzymatic cleavage
sites to generate shorter and less aggregation-prone forms of the Ab
peptide such as Ab37 and Ab38.10–19 The processing of other
c-
secretase substrates such as Notch does not appear to be affected
by the -secretase modulators (GSMs), thus avoiding side effects
c
associated with GSIs. Herein, we describe the discovery of an orally
efficacious 2-methylpyridine-based amide GSM 2520 through lead
optimization of a high throughput screen (HTS) lead sulfonamide 1.
At the start of our program only a few nonsteroidal anti-inflam-
matory drugs (NSAID)-derived carboxylic acids and imidazole-
based heterocycles represented by compounds 2 and 3 were
b-secretase (BACE) and
become popular AD drug discovery targets over the last decade
c
-secretase (GS).4 These enzymes have
and several inhibitors of each enzyme have been advanced into clin-
ical trials.5–7 However, development of
c
-secretase inhibitors (GSIs)
are challenging, largely due to mechanism-based toxicity associated
with inhibition of other
-secretase substrates such as Notch.8,9 An
disclosed chemotypes of c
-secretase modulators (GSMs).11 Both 2
c
and 3 have the same imidazole phenyl moiety that was subse-
alternative approach, called
c-secretase modulation, involves
quently used in a wide variety of GSMs.11
We used a high-throughput homogeneous time-resolved fluo-
rescence (HTRF) assay to screen for inhibitors of Ab42 production
in Human Embryonic Kidney 293 cells (HEK293 cells) with stable
expression of guinea pig Swedish mutant SFV-APP695sw. A cell
⇑
Corresponding author. Tel.: +1 805 447 8498; fax: +1 805 480 1337.
E-mail address: jianc@amgen.com (J.J. Chen).
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.09.041

6448
J. J. Chen et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6447–6454
Cl
S
H
N
O
S
N
N
H
N
Cl
O
O
N
N
N
2
1
Torrey Pines
HTS lead
IC50(Aβ42): 1.54 μM
IC50(Aß40): > 10 μM
IC50(Aß42): 0.13 μM
IC50(Aß40): 0.56 μM
(our data)
H
O
N
N
S
F
N
O
N
O
Cl
N
N
3
4
Eisai E-2012
Roche's HTS lead
IC50(Aß42): 0.82 μM
(reported data)
IC50(Aß42): 0.16 μM
IC50(Aß40: 1.95 μM
(our data)
Figure 1. Non-NSAID based GSM and the HTS lead 1.
viability assay (Cell Titer-Blue from Promega) was used to
eliminate cytotoxic compounds. A Notch luciferase reporter gene
assay (in which a luciferase construct was inserted under the con-
trol of the Notch promoter, in TE-671 cells and Luciferase lumines-
cence was measured following three days of compound treatment)
generation GSMs 2–3 and Roche’s recently reported oxazole-based
GSM lead, 4 (see Fig. 1).21
Lead 1 selectively inhibited the formation of Ab42 (IC50 = 1.54 -
lM) over Ab40 and Notch (both IC50 > 10 lM). In our initial
optimization, we explored the replacement of the oxazole group
by other five and six-membered heteroaryls (Table 1).
Replacement of the nitrogen in the oxazole by CH resulted in over
was used to eliminate nonselective GSIs. Compound
1 was
identified as a novel GSM which is structurally related to the early
Table 1
Sulfonamide SAR
X
H
O
S
N
Ar
X
O
a
c
c
Compd
Ar
X
Ab42 cell IC50
(l
M)
RLMb
166
(
lL/min/mg)
HLMb
32
(lL/min/mg)
CYP3A4 IC50
1.6
(lM)
CYP2D6 IC50
(
lM)
Mouse B/Pd
0.03
O
1
Cl
1.54 0.0.36
2.9
N
O
O
5
6
Cl
Cl
>10
>399
287
>399
94
13
27
>27
2.44
15.1
0.16
0.07
N
N
O
N
7
Cl
>10
>399
>399
>27
>27
N
8
9
Cl
Cl
>10
>399
<14
384
<14
15.3
0.1
11.8
0.1
HN
N
N
2.10 0.23
N
N
N
10
11
Cl
Cl
9.4
113
NA
22
>27
NA
>27
NA
MeO
N
1.06 0.08
NA
N
N
12
13
Cl
Cl
2.75 0.46
>10
85
35
0.6
0.2
0.09
N
>399
>399
16.2
>27
N

J. J. Chen et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6447–6454
6449
Table 1 (continued)
a
c
c
Compd
Ar
X
Ab42 cell IC50
(l
M)
RLMb
>399
(
lL/min/mg)
HLMb
45
(lL/min/mg)
CYP3A4 IC50
3.5
(lM)
CYP2D6 IC50
7.2
(lM)
Mouse B/Pd
N
14
Cl
>10
N
N
15
16
Cl
0.72 0.91
0.28 0.19
>399
290
497
399
4.7
9.8
2.0
0.10
N
N
CF3
>27
MeO
17
18
CF3
CF3
0.53 0.04
1.83 1.33
NA
NA
NA
2.2
NA
1.9
OMe
N
>399
306
N
19
20
CF3
CF3
1.26 0.96
0.21 0.11
169
324
193
371
4.5
0.70
0.7
N
N
12.2
21
CF3
4.06
122
270
>27
1.80
a
b
c
Values with standard deviations shown were the average of at least three independent runs.
Compounds were incubated with microsomes for 30 min at a concentration of 1 M.
Compounds were incubated with microsomes for 5 (CYP3A4) or 10 (CYP2D6) min at a concentration of 3
l
lM.
d
Measured at 2 h after 10 mg/kg oral dosing.
a sevenfold loss of potency (5). The 1,3,4-oxadiazole analog 6 was
almost equipotent. The isomeric 1,2,4-oxadiazole 7 had over a sev-
enfold loss of potency compared to 1. Of the three five-membered
diazole isomers 8–11, only the 1-imidazole 9 maintained the po-
tency. This compound was metabolically stable but was also a po-
tent inhibitor of CYP3A4 and CYP2D6. Introduction of a methyl
group at the 2-position of the imidazole ring resulted in four fold
loss of potency (10). Using (3-methoxy-4-(4-methyl-1H-imida-
zol-1-yl)phenyl present in E2012 (3) to replace 4-(1H-imidazol-
1-yl)phenyl only improved the potency by twofold over 9.22
Of the six-membered heterocycles examined (12–14), the 4-
pyridine analog 12 was slightly less potent than oxazole 1 but
had better metabolic stability in rat liver microsomes. Introduction
of a 2-methyl group improved the potency by fourfold and reduced
CYP inhibitory activity (15). Modification of the sulfonamide
portion identified the 3,5-bis(trifluorophenyl)sulfonamide 16 with
a twofold improvement in potency.
Comparing the structures of 16, 3, and 11, it is reasonable to
assume that 2-methyl pyridine in 16 mimics the methyl imidazole
group in 3 and 11. In the GSM literature the methoxy group is ubiq-
uitous. In recent publications where the effect of the methoxy
group replacement was studied, the potency was reduced in all
cases with the magnitude of the drop dependent upon other
portions of the molecule.12,18,21 We investigated the effect of the
methoxy on the phenyl group of 16. Interestingly the opposite ef-
fect was observed (a twofold decrease in potency) (17). Moving the
methoxy group next to the amine group led to a threefold decrease
in potency (18). Of the several methyl group replacements pre-
pared (19–21), only the ethyl group retained potency.
responsible for the low brain exposure. The corresponding amide
of 15 was prepared and had a much higher B/P ratio (1.46) than
previous compounds under the same conditions (see compound
22, Table 2); however, it showed only partial inhibition of Ab42
production in the cellular assay. At higher concentrations (up
to10
(POC) reached plateau at 58%. Our curve-fitting algorithm auto-
matically assigned an IC50 of >10 M if the minimum POC was
lM top concentration) the Ab42 percent of DMSO control
l
>50%. Variation of the substituents on the carboxamide phenyl ring
offered little improvement in cellular potency (data not shown).
Most compounds showed incomplete inhibition of Ab42 produc-
tion. The best compound identified from this SAR effort was 3,5-
bis(trifluoromethyl) analog 23. It exhibited a minimum POC of
52% at 10 lM. The observation of partial inhibition in this class
of GSM is consistent with proposed allosteric mechanism (vide
infra).
As our next step in establishing the SAR of this GSM chemotype,
we examined the effect of introducing a methylene spacer between
the amide carbonyl function and the disubstituted aromatic ring.
The amide 24 was found to be only slightly less potent than the
sulfonamide 16.
Flexible alignment23 of both 24 and 23 to the lowest energy
conformations of the analogous sulfonamide 16 (predicted from
quantum mechanics) provided a qualitative rationale for the in-
creased potency of 24 compared to 23. The methylene spacer of
24 (Fig. 2, yellow) allowed the pendant, disubstituted aryl ring to
better adopt the approximately perpendicular orientation pre-
dicted for the sulfonamide species (green) than does an amide
lacking the spacing group (pink).
Next we investigated the mouse brain exposure of 15, two
hours after oral dosing at 10 mg/kg. Although the brain/plasma
ratio (B/P = 0.10) was slightly better than that of 1 (0.03), it was
still considered to be low for an efficacious CNS drug. Three other
sulfonamides 6, 9, and 12 also showed low B/P ratios (Table 1). This
led us to postulate that the polar aryl sulfonamide group might be
Introduction of a methyl group at the a-carbon of the amide
gave two enantiomers 25 and 26, separated by chiral supercritical
fluid chromatography (SFC). The absolute stereochemistry of 25
and 26 were determined via comparison of their predicted and
experimental vibrational circular dichroism (VCD) spectra and
optical rotations.24–26 The more active (R)-isomer 25 was fivefold

6450
J. J. Chen et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6447–6454
Table 2
Lead generation
H
N
Y
O
N
RLMb
19
(
lL/min/mg)
HLMb
(l
L/min/mg)
CYP3A4 IC50
>27
(
lM)
CYP2D6 IC50
>27
(lM)
Mouse B/Pd
1.46
a
c
c
Compd
Y
Ab42 cell IC50
(lM)
Cl
22
>10e
91
Cl
CF3
23
24
25
>10f
14
106
18
>27
>27
>27
>27
13
CF3
CF3
0.46 0.11
0.26 0.10
22
CF3
CF3
<14
<14
4.1
2.0g
2.0g
CF3
CF3
26
>10
<14
<14
>27
2.6
CF3
a
b
c
d
e
f
Values with standard deviations shown were the average of at least three independent runs.
Compounds are incubated with microsomes for 30 min at a concentration of 1 M.
Compounds were incubated with microsomes for 5 (CYP3A4) or 10 (CYP2D6) min at a concentration of 3
l
lM.
Measured at 2 h after 10 mg/kg oral dosing.
Maximum 53% inhibition at 10
Maximum 58% inhibition at 10
Obtained from rat PD studies.
l
l
M.
M.
g
more potent (Ab42 IC50 = 0.26
was less active against Ab40 (IC50 = 3.1
about 45-fold selective against Notch (IC50 = 11.8
Compound 25 exhibited low-to-moderate passive permeability
(human LLC-PK1 Avg Papp: 7.8 cm/s) and was not an efflux
substrate (efflux ratio: 1.2). Pharmacokinetic analysis of 25 showed
that in vivo clearance in rats was 0.58 L/h/kg, Vss was 4.3 L/kg, and
t1/2 was 6 h following an iv dose (2 mg/kg). The oral bioavailability
was 100% based on a 5 mg/kg dose. In cynomolgous monkeys,
l
M) than the original HTS lead 1. It
M) than against Ab42 and
M).
in vivo clearance was 0.16 L/h/kg, Vss was 3.1 L/kg, and t1/2 was
15.9 h following an iv dose (2 mg/kg). Oral bioavailability was
56% based on a 5 mg/kg dose.
l
l
Due to its low in vivo clearance and good oral bioavailability in
rat, we tested 25 in a pharmacodynamic (PD) model of Ab reduc-
tion. It was administered in a single oral dose to naïve Sprague-
Dawley rats and Ab levels were measured in CSF and brain tissue.
CSF Ab42 levels were reduced by 41% versus vehicle control
(100 mg/kg po) after 4 h whereas no significant inhibition was
observed with 26 (opposite enantiomer of 25) even though both
have similar plasma and brain exposure. In brain homogenate,
Ab42 levels were reduced by 33%. Consistent with in vitro data,
no significant lowering of Ab40 was observed with 25 (data not
shown).
l
To further investigate the time-course of inhibition by 25, rats
were dosed at 100 mg/kg and samples were collected at the indi-
cated time points and analyzed for both drug and Ab42 levels.
There was a clear relationship between 25 plasma concentrations
and Ab lowering in the brain and CSF (Fig. 3). Sustained CSFAb low-
ering (>36 hours) was observed (Fig. 3) and the plasma and brain
concentrations of 25 were quite stable from 4 h to 36 h post dose.
The data were not consistent with t1/2 of 6 h obtained with 2 mg/kg
iv dose. This might be caused by saturated absorption of 25 at
100 mg/kg. The brain/plasma ratios measured in the PD studies
for 25 were 2.0. This represented a significant improvement over
the original HTS lead 1.
Figure 2. Alignment of 24 (yellow) and 23 (pink) to parent sulfonamide 16 (green).

J. J. Chen et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6447–6454
6451
M)
Table 3
SAR of compound 25 core modification
a
Compd Structures
Ab42 cell IC50 (l
H
N
CF3
O
24
27
0.46 0.11
N
CF3
H
O
N
N
CF3
1.45 0.44
0.16 0.06
O
N
CF3
N
O
F
O
3
N
N
O
N
28
29
1.53 0.9
Figure 3. Time course of in vivo brain and CSF Ab42 reduction with 25 in rats
following single, oral dose of 100 mg/kg and correlation with the plasma
F
a
N
concentration of 25.
H
N
CF3
CF3
CF3
O
O
O
0.50 0.15
Having identified 25 as a novel GSM with moderate in vitro and
in vivo effects on Ab42 reduction we were interested to explore
whether compounds of this type would cause Notch-related toxic-
ities similar to those observed with earlier generations of GSIs. To
assess general and Notch-related toxicity, the compound was
dosed orally at 30, 100, and 300 mg/kg in rat for four days. For
comparison, a known nonNotch selective GSI, LY411575 (our data
Notch IC50: 1.2 nM),9 was also dosed to a separate group of animals
in the same study at 1 and 3 mg/kg. In contrast to LY411575, no
Notch-associated effects were identified for 25 (see Supplemental
data section for more details).
Having established 25 as a viable lead for further optimization,
we studied the SAR of the core modification of the molecule. For
ease of synthesis we used its achiral analog 24 as the starting point
for comparison (Table 3). In the aniline portion (left hand side),
replacement of methylpyridine with commonly used imidazole
group resulted in a threefold loss of potency (27). This trend was
the same as that observed in the sulfonamide series. Replacement
of the imidazole group in 3 by methylpyridine resulted in a tenfold
drop in potency (28). These results clearly demonstrated the differ-
ent SAR between the amide and E-2012 (3) series.
Similar to the sulfonamide SAR, introduction of the methoxy
group in the aniline phenyl resulted in slight loss of potency
(29). Changing the phenyl group to a pyridyl led to threefold loss
of potency for one regio-isomer and complete loss of potency for
the other (30 and 31). The meta-isomer of 24 (32) was completely
inactive. Methylation of the amide NH abolished the GSM activity
(33). Replacement of the amide NHCOCH2 group by the urea
NHCONH group decreased the potency by twofold (34). Replace-
ment of the amide NHCO moiety in 25 by the reverse amide CONH
resulted in loss of potency (35). A small library of 80 reverse
amides with different amines was synthesized and the most potent
compounds had IC50 values in the micro molar range (e.g., com-
pound 36). Due to the divergent SAR observed between methyl
pyridine and imidazole series, we made 55 imidazole phenyl based
reverse amides through parallel chemistry. In general, compounds
in the series were more potent than the corresponding methyl
pyridine analogs (compare 36 and 37). Lipophilic groups were pre-
ferred for improved potency. One of the most potent compounds
CF3
CF3
CF3
N
N
OMe
N
H
N
30
1.35
H
N
N
31
32
33
>10
>10
>10
N
N
CF3
O
N
H
CF3
CF3
N
O
O
CF3
CF3
N
N
H
N
H
N
CF3
34
35
1.11 0.46
O
O
CF3
N
H
>10
N
N
N
CF3
Cl
N
H
36
37
2.12
0.42
Cl
O
N
H
N
O
identified (38), had a cell IC50 of 0.32
unstable (RLM >399 L/min/mg). At that time we decided not to
pursue the series further and instead focused on methylpyridine
lM but was metabolically
l
(continued on next page)

6452
J. J. Chen et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6447–6454
Table 4
Table 3 (continued)
SAR of the carboxamides
a
Compd
Structures
Ab42 cell IC50
(
lM)
R
H
N
Ph
X
O
O
Ph
N
H
38
0.32
N
N
N
N
N
O
a
H
Compd
R
X
Ab42 cell IC50 (lM)
N
N
CF3
45
46
47
48
49
50
51
52
53
54
55
56
Me
H
Et
nPr
nBu
iBu
Me
Me
Me
Me
Me
Me
Me
Me
Me
H
0.54 0.35
2.78 0.29
1.05
0.54 0.25
0.49 0.08
0.17 0.01
0.18 0.08
1.38
>10
6.0
4.80
>10
O
39
40
>10
CF3
OH
H
N
CF3
O
iBu single isomer 1
iBu single isomer 2
HO(CH2)3
Me
Me
2.13
CF3
1-Morpholine
N-Me-piperazine
H
N
Me
CF3
CF3
CN
41
0.82
O
O
O
O
O
N
N
57
58
59
Me
Me
Me
0.16 0.05
0.66
N
CF3
CF3
CF3
N
N
N
N
N
H
N
O
0.14 0.03
42
43
44
45
>10
O
60
Me
0.26 0.08
NMe2
H
N
CO2H
CN
61
62
Me
Me
>10
0.25 0.01
7.0
63
64
Me
Me
0.22 0.03
0.57
CN
H
N
0.57 0.07
0.54 0.35
NC
OMe
65
66
67
Me
Me
Me
0.15 0.05
0.12 0.02
0.070 0.014
H
N
NMe2
N
N
N
N
O
(R) Meb
(S) Meb
0.031 0.008
0.070 0.022
0.020 0.006
O
O
O
68
69
70
a
Values with standard deviations shown were the average of at least three
independent runs.
(R) Bub
i
71
72
73
Me
Me
Me
2-Pyr
3-Pyr
4-Pyr
>10
1.29 0.09
1.20 0.09
based compounds. Independently a group from Pfizer recently
reported similar imidazole-based amides as GSMs.16
Subtle modification of the carboxylic acid (right hand side)
portion of 25 had dramatic effects on the cell potency. For example,
the aza modification of the phenyl group resulted in complete loss
O
N
74
75
Me
Me
0.15 0.10
0.17 0.12
N
N
N
N
of potency (39). Replacement of the
droxy group decreased the potency by eightfold (40). The
tyl group was threefold less active (41). The -dimethyl group
a
-methyl group by an
a-hy-
a-isobu-
a
NH
76
77
Me
Me
0.93
completely abolished the cell activity. Replacement of one of the
trifluoromethyl groups by cyano group resulted in 27-fold loss of
potency (43). The replacement of both trifluoromethyl groups by
one or two methyl groups only had minor effect on the potency
(44 and 45). Compound 45 was subsequently used as a lead for fur-
ther optimization of the carboxylic acid (right hand side) portion
(Table 4).
H
N
0.18 0.08
N
N
N
O
78
Me
0.65 0.24
N
a
Values with standard deviations shown were the average of at least three
independent runs.
Stereochemistry was randomly assigned for peaks separated by chiral SFC, see
Supplemental material for details.
We first investigated the effect of the
a-substitution. Only the
b
isobutyl group provided threefold improvement in potency (50).
Chiral separation of the racemate showed one enantiomer was
about 10-fold more potent than the other. Introduction of polar hy-
droxyl group into the
(53).
a-alkyl chain abolished the cellular potency
The effect of replacing the methyl group on the phenyl ring
was also investigated. Removal of the methyl group resulted in

J. J. Chen et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6447–6454
6453
about tenfold drop in potency (54). Replacement with polar
groups such as morpholino and N-methylpiperazinyl led to over
tenfold loss in potency (55 and 56). However, N-(2-pyridyl) pip-
erazinyl gave a threefold increase in potency (57). A variety of
substituted phenyls were good replacements of the methyl
group. Polar functional groups such as acetyl, cyano, carboxam-
ide, methoxy and substituted amino groups on the terminal
phenyl rings were not only tolerated but actually provided
potency enhancement (58–67). The most potent compound iden-
tified was the 4-morpholinophenyl analog 68 with IC50 of 31 nM.
tion of morpholinophenyl containing compounds, such as 70, with
a tenfold in vitro potency improvement. These compounds lacked
the in vivo PD efficacy achieved by 25, likely a result of poor
permeability and high cLogP. We also showed the SAR trend of
methylpyridine and the widely used methyl imidazole based com-
pounds was different in several cores.
Supplementary data
Supplementary data (synthesis of representative examples, pro-
cedures for chiral separation and stereochemistry confirmation,
in vitro and in vivo assays, and rat four-day Notch toxicity study)
associated with this article can be found, in the online version, at
http://dx.doi.org/10.1016/j.bmcl.2013.09.041.
As expected, replacement of the
a-methyl by a-isobutyl
improved the potency slightly (70, IC50 = 20 nM).
To reduce the lipohilicity of these more potent GSMs we studied
the effect of replacing the terminal phenyl group with a pyridyl.
Both 3- and 4-pyridyl analogs were only twofold less potent but
the 2-pyridyl was inactive (71–73). Unfortunately the 3-pyridine
analog of 68 was fivefold less potent. The morpholino group could
be replaced by the dimethyl amino group (75). The monomethyla-
mino group was fivefold less potent than dimethyl amino group
(76). The potency was restored with the more lipophilic N-isopro-
pyl amino group (77). Finally the pyrimidine analog of 68 (i.e. 78)
was twenty-fold less potent than 68. These data were consistent
with our findings and literature reports that GSMs tended to be
more lipophilic.
Through the modification of the carboxylic acid portion we
were able to improve the cell potency of the original lead 25 by
over tenfold. We tested some of these more potent GSMs (Table 4)
in the in vivo PD model. Unfortunately all showed <30% inhibition
of CSF Ab42 at 100 mg/kg when dosed orally. For example, 70 pro-
duced 27% Ab42 reduction four hours after a 100 mg/kg oral dose.
The unbound plasma concentration (290 nM) was significantly
higher than the measured CSF concentration (2 nM) (the brain con-
centration was not measured in the study). The large shift may
have been caused by the high lipophilicity (cLogP 6.86 calculated
using ChemDraw) and poor passive permeability (LLC-PK Avg Papp
References and notes
1. Hardy, J. A.; Higgins, G. A. Science 1992, 256, 184.
2. Goate, A.; Hardy, J. J. Neurochem. 2012, 120, 3.
3. Kung, H. F. ACS Med. Chem. Lett. 2012, 3, 265.
4. Zhang, H.; Ma, Q.; Zhang, Y.-W.; Xu, H. J. Neurochem. 2012, 120, 9.
5. Ghosh, A. K.; Brindisi, M.; Tang, J. J. Neurochem. 2012, 120, 71.
6. Kreft, A. F.; Martone, R.; Porte, A. J. Med. Chem. 2009, 52, 6169.
7. Imbimbo, B. P.; Panza, F.; Frisardi, V.; Solfrizzi, V.; D’Onofrio, G.; Logroscino, G.;
Seripa, D.; Pilotto, A. Expert Opin. Invest. Drugs 2011, 20, 325.
8. Searfoss, G. H.; Jordan, W. H.; Calligaro, D. O.; Galbreath, E. J.; Schirtzinger, L.
M.; Berridge, B. R.; Gao, H.; Higgins, M. A.; May, P. C.; Ryan, T. P. J. Biol. Chem.
2003, 278, 46107.
9. Wong, G. T.; Manfra, D.; Poulet, F. M.; Zhang, Q.; Josien, H.; Bara, T.; Engstrom,
L.; Pinzon-Ortiz, M.; Fine, J. S.; Lee, H. J. J.; Zhang, L.; Higgins, G. A.; Parker, E. M.
J. Biol. Chem. 2004, 279, 12876.
10. Pettersson, M.; Kauffman, G. W.; am Ende, C. W.; Patel, N. C.; Stiff, C.; Tran, T.
P.; Johnson, D. S. Expert Opin. Ther. Pat. 2011, 21, 205.
11. Oehlrich, D.; Berthelot, D. J. C.; Gijsen, H. J. M. J. Med. Chem. 2011, 54, 669.
12. Wan, Z.; Hall, A.; Sang, Y.; Xiang, J. N.; Yang, E.; Smith, B.; Harrison, D. C.; Yang,
G.; Yu, H.; Price, H. S.; Wang, J.; Hawkins, J.; Lau, L. F.; Johnson, M. R.; Li, T.;
Zhao, W.; Mitchell, W. L.; Su, X.; Zhang, X.; Zhou, Y.; Jin, Y.; Tong, Z.; Cheng, Z.;
Hussain, I.; Elliott, J. D.; Matsuoka, Y. Bioorg. Med. Chem. Lett. 2011, 21, 4832.
13. Qin, J.; Zhou, W.; Huang, X.; Dhondi, P.; Palani, A.; Aslanian, R.; Zhu, Z.;
Greenlee, W.; Cohen-Williams, M.; Jones, N.; Hyde, L.; Zhang, L. ACS Med. Chem.
Lett. 2011, 2, 471.
14. Sun, Z.-Y.; Asberom, T.; Bara, T.; Bennett, C.; Burnett, D.; Chu, I.; Clader, J.;
Cohen-Williams, M.; Cole, D.; Czarniecki, M.; Durkin, J.; Gallo, G.; Greenlee, W.;
Josien, H.; Huang, X.; Hyde, L.; Jones, N.; Kazakevich, I.; Li, H.; Liu, X.; Lee, J.;
MacCoss, M.; Mandal, M. B.; McCracken, T.; Nomeir, A.; Mazzola, R.; Palani, A.;
Parker, E. M.; Pissarnitski, D. A.; Qin, J.; Song, L.; Terracina, G.; Vicarel, M.;
Voigt, J.; Xu, R.; Zhang, L.; Zhang, Q.; Zhao, Z.; Zhu, X.; Zhu, Z. J. Med. Chem.
2012, 55, 489.
15. Peng, H.; Talreja, T.; Xin, Z.; Cuervo, J. H.; Kumaravel, G.; Humora, M. J.; Xu, L.;
Rohde, E.; Gan, L.; Jung, M. Y.; Shackett, M. N.; Chollate, S.; Dunah, A. W.;
Snodgrass-belt, P. A.; Arnold, H. M.; Taveras, A. G.; Rhodes, K. J.; Scannevin, R.
H. ACS Med. Chem. Lett. 2011, 2, 786.
16. Pettersson, M.; Johnson, D. S.; Subramanyam, C.; Bales, K. R.; am Ende, C. W.;
Fish, B. A.; Green, M. E.; Kauffman, G. W.; Lira, R.; Mullins, P. B.; Navaratnam, T.;
Sakya, S. M.; Stiff, C. M.; Tran, T. P.; Vetelino, B. C.; Xie, L.; Zhang, L.; Pustilnik, L.
R.; Wood, K. M.; O’Donnell, C. J. Bioorg. Med. Chem. Lett. 2012, 22, 2906.
17. Yngve, U.; Paulsen, K.; Macsari, I.; Sundstrom, M.; Santangelo, E.; Linde, C.;
Bogar, K.; Lake, F.; Besidski, Y.; Malmborg, J.; Stromberg, K.; Appelkvist, P.;
Radesater, A.-C.; Olsson, F.; Bergstrom, D.; Klintenberg, R.; Arvidsson, P. I.
MedChemComm 2013, 4, 422.
<1
lcm/s).
A major difficulty encountered in our lead optimization of the
methylpyridine based GSMs was the so-called ‘flat’ or ‘shallow’
SAR, a phenomenon widely described in the literature of alloste-
ric modulators of GPCRs.27–29 This is characterized by complete
loss of activity with minor structural changes and/or very shallow
and narrow SAR patterns. Another observation was the presence
of ‘nonadditive SAR’ whereby the substituent effect in one series
or molecule was not translatable into another, similar to the
observation reported for allosteric modulators of GPCRs.
Although we have not carried out any mechanistic studies of
our GSMs, several reports suggested an allosteric mechanism of
GSM action on
c
-secretase.30–32 This mechanism could also ex-
plain the SAR observation mentioned above including the pres-
ence of partial enzyme inhibition for certain amides such as 22
and 23 (similar to partial antagonists observed in allosteric mod-
ulators of GPCRs).29
18. Bischoff, F.; Berthelot, D.; De Cleyn, M.; Macdonald, G.; Minne, G.; Oehlrich, D.;
Pieters, S.; Surkyn, M.; Trabanco, A. A.; Tresadern, G.; Van Brandt, S.; Velter, I.;
Zaja, M.; Borghys, H.; Masungi, C.; Mercken, M.; Gijsen, H. J. M. J. Med. Chem.
2012, 55, 9089.
19. Huang, X.; Aslanian, R.; Zhou, W.; Zhu, X.; Qin, J.; Greenlee, W.; Zhu, Z.; Zhang,
L.; Hyde, L.; Chu, I.; Cohen-Williams, M.; Palani, A. ACS Med. Chem. Lett. 2010, 1,
184.
20. Hitchcock, S. A.; Chen, J. J.; Ncube, M.; Nixey, T.; Amegadzie, A.; Kunz, R.; Qian,
W.; Chen, N.; Tegley, C. M.; Demorin, F.; Yuan, C. C.; Liu, Q.; Zhu, J.; Peterkin, T.;
Adams, J. A.; Hu, E.; Chavez, F. WO 2009075874, 2009. Chem. Abstr. 2009,
739301.
21. Lübbers, T.; Flohr, A.; Jolidon, S.; David-Pierson, P.; Jacobsen, H.; Ozmen, L.;
Baumann, K. Bioorg. Med. Chem. Lett. 2011, 21, 6554.
In summary, starting from an initial oxazole based HTS lead 1
we have identified 2-methylpyridine as a good replacement for
oxazole with improved PKDM properties. To improve CNS expo-
sure, the sulfonamide group was replaced by the corresponding
carboxamide functionality. Further homologation of the amide
group resulted in the lead 25 with significant improvement in
the in vitro potency, which could be rationalized by the observed
alignment of low energy conformations of the homologated amide
to the sulfonamide, unlike the benzamide analog. This lead optimi-
zation effort therefore led to a successful replacement of a
sulfonamide moiety to improve CNS penetration and in vivo po-
tency. In a rat PD model, 25 demonstrated statistically significant
lowering of Ab42 when administrated orally at 100 mg/kg. Further
optimization of the carboxamide portion in 25 led to the identifica-
22. A recent patent application from Shionogi disclosed a close analog of 11 (3,4-
dichlorophenyl sulfonamide instead of the 3,5-dichlorophenyl sulfonamide in
11) with similar cell IC50 (0.9
Chem. Abstr. 2010, 71724.
lM), see Kato, I.; Kank, K. EP2455370, 2012.
23. Cho, S. J.; Sun, Y. J. Chem. Inf. Model. 2005, 46, 298.
24. Freedman, T. B.; Cao, X.; Dukor, R. K.; Nafie, L. A. Chirality 2003, 15, 743.

6454
J. J. Chen et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6447–6454
25. Stephens, P. J.; Devlin, F. J.; Pan, J.-J. Chirality 2008, 20, 643.
26. Stephens, P. J.; Devlin, F. J.; Cheeseman, J. R.; Frisch, M. J. J. Phys. Chem. A 2001,
105, 5356.
27. Conn, P. J.; Kuduk, S. D.; Doller, D. In Annu. Rep. Med. Chem.; Manoj, C. D., Ed.;
Academic Press, 2012; Vol. 47, p. 441.
28. Zhao, Z.; Wisnoski, D. D.; O’Brien, J. A.; Lemaire, W.; Williams, D. L., Jr.;
Jacobson, M. A.; Wittman, M.; Ha, S. N.; Schaffhauser, H.; Sur, C.; Pettibone, D.
J.; Duggan, M. E.; Conn, P. J.; Hartman, G. D.; Lindsley, C. W. Bioorg. Med. Chem.
Lett. 2007, 17, 1386.
29. Melancon, B. J.; Hopkins, C. R.; Wood, M. R.; Emmitte, K. A.; Niswender, C. M.;
Christopoulos, A.; Conn, P. J.; Lindsley, C. W. J. Med. Chem. 2012, 55, 1445.
30. Rennhack, A.; Jumpertz, T.; Ness, J.; Baches, S.; Pietrzik, C. U.; Weggen, S.; Bulic,
B. Bioorg. Med. Chem. 2012, 20, 6523.
31. Ohki, Y.; Higo, T.; Uemura, K.; Shimada, N.; Osawa, S.; Berezovska, O.;
Yokoshima, S.; Fukuyama, T.; Tomita, T.; Iwatsubo, T. EMBO J 2011, 30, 4815.
32. Pozdnyakov, N.; Murrey, H. E.; Crump, C. J.; Pettersson, M.; Ballard, T. E.; am
Ende, C. W.; Ahn, K.; Li, Y.-M.; Bales, K. R.; Johnson, D. S. J. Biol. Chem. 2013, 288,
9710.