Discovery of γ-secretase inhibitors efficacious in a transgenic animal model of Alzheimer's disease

Article abstract of DOI:10.1016/j.bmcl.2006.10.011

Attachment of the cyclopropylcarbamate group to the piperidine core of γ-secretase inhibitors leads to a dramatic increase of their in vitro potency. Strategies for subsequent improvement of the in vivo pharmacokinetic profile of the series are discussed.

Full text of DOI:10.1016/j.bmcl.2006.10.011

Bioorganic & Medicinal Chemistry Letters 17 (2007) 511–516
Discovery of c-secretase inhibitors efficacious in a transgenic
animal model of Alzheimer’s disease
Theodros Asberom,a Zhiqiang Zhao,a Thomas A. Bara,a John W. Clader,a
William J. Greenlee,a Lynn A. Hyde,b Hubert B. Josien,a Wei Li,a
Andrew T. McPhail,c Amin A. Nomeir,d Eric M. Parker,b Murali Rajagopalan,a
Lixin Song,b Gwendolyn T. Wong,b Lili Zhang,b Qi Zhangb and
Dmitri A. Pissarnitskia,*
aDepartment of Chemical Research, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
bDepartment of Central Nervous System, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
cDepartment of Chemistry, Duke University, Durham, NC 27708-0346, USA
dDepartment of Drug Metabolism and Pharmacokinetics, Schering-Plough Research Institute, 2015 Galloping Hill Road,
Kenilworth, NJ 07033, USA
Received 31 August 2006; revised 5 October 2006; accepted 5 October 2006
Available online 10 October 2006
Abstract—Attachment of the cyclopropylcarbamate group to the piperidine core of c-secretase inhibitors leads to a dramatic
increase of their in vitro potency. Strategies for subsequent improvement of the in vivo pharmacokinetic profile of the series are
discussed. Resulting compounds significantly reduce Ab levels in TgCRND8 mice after a single PO dosing at 30 mpk.
Ó 2006 Elsevier Ltd. All rights reserved.
The most common degenerative brain disorder, Alzhei-
mer’s disease (AD), makes up to 70% of all cases of
dementia and is the third most common cause of death
in the United States.1 One of the major pathological
hallmarks of AD is abnormal extracellular deposition
of b-amyloid peptide (Ab) in the form of plaques in
the brain of AD patients. Although the exact cause of
AD is unknown, a large body of evidence suggests2 that
overproduction of amyloid peptides, especially Ab42, is
central to its pathogenesis. Amyloid peptides are pro-
duced by proteolysis of the amyloid precursor protein
(APP) by sequential action of b- and c-secretases.
Because of its central role in the generation of Ab pep-
tide, c-secretase was proposed as a target for the treat-
ment of AD.3
model of AD based on TgCRND8 mice,5 but its abil-
ity to reduce Ab40 in the plasma was significant only
when dosed subcutaneously at 100 mpk. We hypothe-
sized that the low efficacy of 1 was due to poor PK,
possibly resulting from the rapid in vivo hydrolysis
of the carbamate linker. In order to slow down this
metabolic process via the introduction of the steric
hindrance, molecule 2 featuring a gem-dimethyl group
was designed. We were pleased to see that not only
was the AUC6 of 2 increased by nearly a factor of
2, but also the potency of 2 was also enhanced by sev-
eral fold. Despite the improvement in the metabolic
stability of 2, its chemical stability in the acidic media
was not adequate. The compound decomposed on the
attempt to formulate it as a hydrochloric salt using
1 M HCl. This outcome is unsurprising in the view
of the similarity between the carbamate fragment of
2 and BOC protecting group, known to be cleavable
under acidic pH. We reasoned that by formation of
an additional bond between the two carbon atoms
of the gem-dimethyl group, thus converting it to the
cyclopropyl moiety, stability of the compound to the
acid would be increased.
Previously we described a series of c-secretase inhibi-
tors related to the sulfonylated piperidine 1.4 Com-
pound 1 was assessed in vivo in a transgenic animal
Keywords: Alzheimer’s disease; c-Secretase inhibitors; Kulinkovich
reaction.
*
Corresponding author. E-mail: Dmitri.Pissarnitski@spcorp.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.10.011

512
T. Asberom et al. / Bioorg. Med. Chem. Lett. 17 (2007) 511–516
ii
i
iii
(racemic)
O
N
Ar1
N
COOMe
Br
N
COOMe
COOMe
Br
N
COOH
iv
COOMe
5
4
N
F
N
S
v
R
R
O
OH
Ar1
N
Ar1
N
Ar1
N
O
O
H
SO2
SO2
Ar2
Ar2
8
6
7
Cl
vi
O
vii
NR1R2
1, R=H
O
O
Ar1
N
Ar1
N
Membr. Ab40 IC50 = 11.2 nM
Rat AUC 0-6 h = 1376 hr.ng/mL (@10 mpk)
O
O
SO2
Ar2
SO2
Ar2
C6H4-NO2-p
2
9
2, R=Me
2, NR1R2 = 4-piperidinopiperidine,
Ar1=3-Fluorophenyl,
Membr. Ab40 IC50 = 4.4 nM
Rat AUC 0-6 h = 2429 hr.ng/mL (@10 mpk)
Ar2=4-chlorophenyl
O
O
N
C6H4-NO2-p
=
"BOC"
O
O
O
tert-butyloxycarbonyl
vi
viii
OH
O
7
Ar1
N
Ar1
N
SO2
SO2
Ar2
N
Ar2
N
11
10
cyclopropyloxycarbonyl
O
vii
NR1R2
O
O
Ar1
SO2
Synthesis of 2 and its cyclopropyl analogs 3 is depict-
ed on Scheme 1. Commercial 2-bromopicolinic acid
was esterified using methanol and thionyl chloride to
provide ester 4. Suzuki coupling of 4 with arylboronic
acids Ar1B(OH)2 was conducted using sodium carbon-
ate as the base in the 6:3 mixture of toluene and eth-
anol as the solvent at 90 °C. Resulting methyl ester 5
had up to 10% contamination of ethyl ester stemming
from the transesterification with the solvent. However,
without the ethanol or when it was replaced with
methanol, the reaction completely failed. Catalytic
hydrogenation of the pyridine ring took precedence
over the hydrogenation of Ar1 providing piperidine 6
selectively. Sulfonylation of the latter leading to 7
was achieved with arenesulfonylchlorides at elevated
temperatures in pyridine as the solvent. Reaction of
7 with excess of methyl Grignard reagent in THF fur-
nished tertiary alcohol 8, which was converted to 4-
nitrophenyl carbonate 9. Final compound 2 was ob-
tained from 9 by the displacement of 4-nitrophenol
by 4-piperidinopiperidine. Cyclopropylcarbonates of
type 3 were obtained in a similar manner from alcohol
10, available in a good yield (72%) from ester 7 by
Kulinkovich cyclopropanation.7 Slow addition of
methylmagnesium bromide to a chilled mixture of
the ester 7 and titanium tetraisopropoxide in THF
was essential for the success of this transformation.
Ar2
3
Scheme 1. Reagents and conditions: (i) MeOH, SOCl2, 0 °C to reflux,
5 h; (ii) Ar1B(OH)2, cat Pd(PPh3)4, Na2CO3, toluene/EtOH = 6:3,
90 °C, 10 h; (iii) H2, cat PtO2, MeOH/AcOH = 4:1; (iv) Ar2SO2Cl, Pyr,
70 °C, overnight; (v) MeMgBr, THF; (vi) 4-nitrophenylchlorocarbon-
ate (2 equiv), pyridine (3 equiv), CH3CN/THF = 1:3, 1–3 days; (vii)
HNR1R2, THF, 80 °C, 12 h; (viii) EtMgBr (3 equiv), Ti(OPr-i)4 (0.3
equiv), THF, 0 °C.
(NR1R2 = NMe2), where the dimethylamino group
likely originated from a molecule of DMF.
Initial profiling of the biological activity of cyclopropyl
carbonates 3 using membrane-based preparations c-se-
cretase8 for IC50 determinations demonstrated an
improvement of potency by a factor of 5–18 as
compared to the previously described4 analogs of 1
(Table 1).
As was described previously, both NMR data and ab ini-
tio calculations point to the chair-like conformation of
the piperidine ring in 1 with both side chains occupying
diaxial positions.4 In the course of the present work, we
were able to obtain X-ray quality crystals of compound
11 (Ar1 = 3-chlorophenyl, Ar2 = 4-chlorophenyl). As
indicated by X-ray analysis, the presence of the cyclo-
propyl moiety did not drastically alter the overall shape
of the inhibitors (Fig. 1) and the bis-diaxial chair-like
conformation was preserved.9
Due to the steric hindrance, transformation of 11 to
3 at room temperature was impractically slow, espe-
cially with crowded amines HNR1R2. We screened
a number of solvents to amend the process for paral-
lel synthesis and found that the majority of reactions
were completed in THF after an overnight stirring at
80 °C. Although the use of DMF as the solvent
resulted in shorter reaction times, the products were
While Table 1 lists the data for the inhibition of Ab40
production measured using membrane-based prepara-
tions of c-secretase,8 it was interesting to assess the cell
permeation properties of the series in the assays utilizing
whole cells. We were pleased to see that most of the
compounds preserved the potency well upon the transfer
often contaminated by dimethylaminocarbonate
3

T. Asberom et al. / Bioorg. Med. Chem. Lett. 17 (2007) 511–516
Table 1. Influence of the cyclopropyl group on the potency of inhibitorsa
513
Ar1
NR1R2
Analogs of 1 Ab40 IC50 (nM)
Analogs of 3 Ab40 IC50 (nM)
Potency ratio
3-F–Ph
11.2
0.7
16.0
5.0
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
3-F–Ph
26.8
5.3
4.2
7.9
6.1
8.5
15.4
36.1
8.5
74.8
9.4
110.0
152.5
164.5
17.9
18.0
10.7
a All tested compounds were racemic mixtures.
from the membrane to the whole cell-based assay.10
However, molecules containing piperazines with a free
NH group on the right-hand side (19, 24, and 25)
showed a larger relative shift in potency as compared
with N-alkylated piperazines (20 and 27). For the major-
ity of the compounds, inhibition of Ab42 production fol-
lowed the trend of Ab40, with potency remaining in the
single- or low double-digit nanomolar range.
Preliminary information on the pharmacokinetics of the
series in the rat was obtained by analysis of the 6 h AUC
after oral dosing at 10 mpk (Table 2).6 When the NR1R2
group lacked a second basic nitrogen (e.g., 12 and 13),
the compounds did not have any measurable plasma lev-
els. For other compounds, the following strategies to im-
prove PK were successfully used: (a) blocking metabolic
‘hot spots’ with fluorine atoms (15); (b) introduction of a
methyl group or a bridge to create steric hindrance in
the vicinity of the carbamate linker (16 and 21); (c)
Figure 1. X-Ray crystal structure of the compound 11 (Ar1 = 3-
fluorophenyl, Ar2 = 4-chlorophenyl). Equatorial hydrogen atoms are
shown for clarity.

514
T. Asberom et al. / Bioorg. Med. Chem. Lett. 17 (2007) 511–516
Table 2. In vitro data and AUC0–6 h after PO dosing at 10 mpk in rata
Compound
Ar1
Ar2
NR1R2
Ab40 membrane (nM)
Ab40 cell (nM)
AUC (h ng/mL)
0
12
3-F–Ph
4-Cl–Ph
3.7
7.7
13
14
15
16
17
18
19
20
21
3-F–Ph
3-F–Ph
3-F–Ph
3-F–Ph
3-F–Ph
3-F–Ph
3-F–Ph
3-F–Ph
3-F–Ph
4-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4-Cl–Ph
4.5
0.8
2.3
2.6
5.2
3.6
5.3
1.9
2.1
10.7
5.0
0
802
5.5
1668
2878
3072
2090
1811
2087
7209
9.7
21.5
20.0
120.4
7.0
28.7
22
3-F–Ph
4-Cl–Ph
0.9
5.1
1768

T. Asberom et al. / Bioorg. Med. Chem. Lett. 17 (2007) 511–516
515
Table 2 (continued)
Compound
Ar1
Ar2
NR1R2
Ab40 membrane (nM)
Ab40 cell (nM)
AUC (h ng/mL)
23
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
3,5-Di-F–Ph
4-Cl–Ph
0.3
5.8
1076
6206
N/A
783
24
25
26
27
28
29
30
4-Cl–Ph
7.9
106.0
153.5
17.0
7.5
4-Cl–Ph
16.8
2.7
4-Cl–Ph
4-Cl–Ph
1.1
1401
5001
5053
23064
4-Cl–Ph
2.4
5.0
3,5-Di-F–Ph
3,5-Di-F–Ph
3.5
12.9
24.7
16.8
a All tested compounds were racemic mixtures.
pre-oxidation of the side chain by introduction of a
hydroxyl (20, 21, and 28); (d) replacement of 4-chloro-
phenylsulfonyl Ar2 group on the bottom part of the
molecule with alternative patterns of fluorination (29
and 30). 3,5-Difluorinated phenyl as the Ar1 substituent
often provided compounds with enhanced potency and
metabolic stability as compared to their monofluorinat-
ed counterparts (e.g., 28 vs 20).
Figure 2 illustrates the efficacy of five representative
members of the series in the transgenic animal model
of AD. Young (6 weeks of age) TgCRND8 mice5 which
have not yet developed the amyloid plaques were used in
the experiments. A group of four mice was used per each
compound with a vehicle-treated group as the negative
control. Following oral dosing at 30 mpk, a substantial
reduction of Ab40 was observed in the plasma of the
drug-treated mice after a 3 h period of time. The remain-
Figure 2. Reduction of plasma Ab4 in TgCRND8 mice after an oral
dosing at 10 mpk as compared to the vehicle-dosed animals. Each bar
represents an average of four animals.

516
T. Asberom et al. / Bioorg. Med. Chem. Lett. 17 (2007) 511–516
Sinning, L.; Smith, E. M.; Parker, E. M.; Vaccaro, H. A.;
Voigt, J. H.; Zhang, L.; Zhao, Z. Bioorg. Med. Chem. Lett.
2006, doi:10.1016/j.bmcl.2006.09.094.
ing levels of the Ab40 constituted 2–52% of those ob-
served in the vehicle-treated group. Reduction of the lev-
els of both Ab4 and Ab42 in the cortex of the brain
tracked well with the effect in the plasma (data not
shown and will be published separately).11
5. Chishti, M. A.; Yang, D. S.; Janus, C.; Phinney, A. L.;
Horne, P.; Pearson, J.; Strome, R.; Zuker, N.; Loukides,
J.; French, J.; Turner, S.; Lozza, G.; Grilli, M.; Kunicki,
S.; Morissette, C.; Paquette, J.; Gervais, F.; Bergeron, C.;
Fraser, P. E.; Carlson, G. A.; George-Hyslop, P. S.;
Westaway, D. J. Biol. Chem. 2001, 276, 21562.
6. Following the oral dosing at 10 mpk in HPBCD or MC,
AUC in rats was measured over a period of 6 h, using
cassette-accelerated rapid rat screen (CARRS): Korfm-
acher, W. A.; Cox, K. A.; Ng, K. J.; Veals, J.; Hsieh, Y.;
Wainhaus, S.; Broske, L.; Prelusky, D.; Nomeir, A.;
White, R. E. Rapid Commun. Mass Spectrom. 2001, 15,
335.
In conclusion, we discovered a novel series of c-secretase
inhibitors potentially useful as a therapy for the Alzhei-
mer’s disease. Several representative compounds of the
series were tested in TgCRND8 mouse serving as the
animal model of AD and demonstrated the ability to re-
duce Ab levels in vivo.
Acknowledgments
7. (a) Lysenko, I. L.; Kulinkovich, O. G. Russ. J. Org. Chem.
2001, 37, 1238; (b) Racouchot, S.; Sylvestre, I.; Ollivier, J.;
Kozyrkov, Y. Y.; Pukin, A.; Kulinkovich, O.; Salauen, J.
Eur. J. Org. Chem. 2002, 13, 2160.
8. Zhang, L.; Song, L.; Terracina, G.; Liu, Y.; Pramanik, B.;
Parker, E. Biochemistry 2001, 40, 5049.
9. Crystallographic data have been deposited with Cam-
bridge Crystallographic Data Centre as supplementary
publication number CCDC 619153.
The authors are thankful to Drs. A. Buevich, T.-M.
Chan, R. Osterman, and A. Evans for NMR analyses,
Dr. J. Voigt for molecular modeling and valuable dis-
cussions, and the group of Dr. J. Wong for the scale
up of synthetic intermediates.
10. Efficacy of c-secretase inhibitors in intact cells was
measured using HEK293 cells expressing human
APP with both Swedish and London mutations.
The cells were grown in 96-well plate with 100 ll
media per well, and were changed to fresh media
and incubated with c-secretase inhibitor for 4 h. Ten
micro litres of conditioned media was used to
measure Ab40 using ECL technology as described in
Ref. 8.
References and notes
1. Ewbank, D. C. Am. J. Public Health 1999, 89, 90.
2. Hardy, J.; Selkoe, D. J. Science 2002, 297, 353.
3. Josien, H. Curr. Opin. Drug Discov. Devel. 2002, 5,
513.
4. Pissarnitski, D. A.; Asberom, T.; Bara, T. A.; Buevich, A.
V.; Clader, J. W.; Greenlee, W. J.; Guzik, H. S.; Josien, H.
B. W.; McEwan; McKittrick, B. A.; Nechuta, T. L.;
11. Hyde, L. A.; Zhang, L., in preparation.