Potent pyrrolidine- and piperidine-based BACE-1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.10.116

Based on lead compound 1 identified from the patent literature, we developed novel patentable BACE-1 inhibitors by introducing a cyclic amine scaffold. Extensive SAR studies on both pyrrolidines and piperidines ultimately led to inhibitor 2f, one of the m

Full text of DOI:10.1016/j.bmcl.2007.10.116

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 414–417
Potent pyrrolidine- and piperidine-based BACE-1 inhibitors
U. Iserloh,a,* Y. Wu,a J. N. Cumming,a J. Pan,a L. Y. Wang,a A. W. Stamford,a
M. E. Kennedy,b R. Kuvelkar,b X. Chen,b E. M. Parker,b C. Stricklanda and J. Voigta
aDepartment of Chemical Research, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
bDepartment of Neurobiology, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
Received 23 July 2007; revised 2 October 2007; accepted 5 October 2007
Available online 6 November 2007
Abstract—Based on lead compound 1 identified from the patent literature, we developed novel patentable BACE-1 inhibitors by
introducing a cyclic amine scaffold. Extensive SAR studies on both pyrrolidines and piperidines ultimately led to inhibitor 2f,
one of the most potent inhibitors synthesized to date.
Ó 2007 Elsevier Ltd. All rights reserved.
cyclize
Alzheimer’s disease (AD)1 is a neurodegenerative disor-
F
der associated with the accumulation of amyloid plaques
comprised of Ab40,42 peptides, along with neurofibrillary
tangles in the brain. These peptides arise from cleavage
of the amyloid precursor protein (APP) by b-amyloid
cleaving enzyme-1 (BACE-1),2 hence inhibitors of this
aspartyl protease could have therapeutic benefits for
the treatment of AD.
OMe
N
H
1
OH
OH
F
O
O
R'
Pr2N
N
H
R
2
n
N
H
Figure 1. Design of azetidine/pyrrolidine/piperidine-based BACE-1
inhibitors 2 (n = 0–2).
In the course of our program to develop patentable no-
vel inhibitors of BACE-1, we investigated conformation-
ally constrained versions of the hydroxyethylamine
motif found in many aspartyl protease inhibitors.3
Based on our analysis of the binding mode for the pro-
totypical inhibitor 1 complexed to BACE-1,4,5 we
decided to integrate the methylene groups flanking the
secondary amine into an azetidine, pyrrolidine, or piper-
idine ring (2, Fig. 1). Computational modeling of these
ring-constrained amines suggested a similar binding
mode as for 1, while simultaneously rigidifying the con-
formational flexibility of the R0-side chain.
A
R'
Ar
Xc
n
Boc
4
5
N
O
OH
OH
R'
Ar
R'
n
Boc
H2N
N
Ar
Bn2N
OH
N
B
3
Figure 2. Retrosynthetic options toward amino alcohol 3 (Ar = 3,5-
difluorophenyl; Xc = Evans oxazolidinone auxiliary).
From a retrosynthetic perspective, we considered the
stereoselective synthesis of key amino alcohol 3 via
two complementary routes (Fig. 2): formation of bond
A would require an Evans aldol coupling between an
enantiopure Boc-protected prolinal (or piperidine-2-
aldehyde) and an enolate derived from an appropriate
acyloxazolidinone (Xc = auxiliary) to give intermediate
4 with three contiguous stereocenters in their required
all-syn arrangement.6,7 Another approach to piperidines
through formation of bond B would establish the cor-
rect syn-stereochemistry of 5 via a Felkin-Anh-con-
trolled addition8 of substituted 2-lithiopyridines onto
(S)-N,N-(dibenzyl)-3,5-difluorophenylalaninal, followed
Keywords: Alzheimer’s disease; b-Secretase; Aspartyl protease inhib-
itors; Peptidomimetics.
*
Corresponding author. Tel.: +1 908 740 7375; fax: +1 908 740
7152; e-mail: ulrich.iserloh@spcorp.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.10.116

U. Iserloh et al. / Bioorg. Med. Chem. Lett. 18 (2008) 414–417
Table 1. BACE-1 binding affinity and cellular activity
415
F
F
Pd/C, H2; (COCl)2;
O
O
Ar
nBuLi, 96%
F
O
N
H
N
O
7
8
CO2H
R'
O
F
6
OBn
n
HN
N
H
O
OBn
9
Ar
Xc
LiOOH;
DPPA
OH
N
OHC
Boc
O
Pr2N
N
Boc
72 %
OBn
NEt3, Bu2BOTf
80 %
O
OH
10
OBn
Compound n (ring system) R0
BACE-1 Cell
a
IC50
(nM)
Ar
Ar
O
a
IC50
(nM)
LiOH
75 %
N
H2N
N
HN
O
Boc
Boc
11
OH
2a
2b
2c
2d
2e
2f
0 (Azetidine)
1 (Pyrrolidine)
H
770
65
89
187
5
2800
330
12
H
OH
OBn
Ar
1450
780
13, HOBt,
EDCI; TFA
OMe
OBn
OPh
H
HN
N
O
150
165
H
OH
Pr2N
O
3
46%
2g
2h
2i
2 (Piperidine)
130
70
14
240
2f
nPr
1200
522
Bn
Scheme 1. Synthesis of BACE-1 inhibitor 2f as a representative
example for the Evans aldol route (Ar = 3,5-difluorophenyl; Xc = oxa-
zolidinone auxiliary).
2j
2-MeOPh CH2 1350
n/d
n/d
2k
2l
3-MeOPh CH2
4-MeOPh CH2
Ph(CH2)2
Chx
330
440
191
610
630
2000
8
2000
1850
4200
2000
3000
106
2m
2n
2o
2p
2q
2r
2s
2t
by reduction of the pyridine moiety to the desired piper-
idine scaffold.
ChxCH2
OAc
OEt
A specific example illustrating the implementation of
the Evans aldol route is shown for BACE-1 inhibitor
2f (Scheme 1), but served well for a host of other cyc-
lic amines described in Table 1. The acyloxazolidinone
8 required for the Aldol coupling was prepared in
three steps (96% overall yield) from commercially
available 3,5-difluorocinnamic acid 6 based on the
literature.9 Boc-protected 4-benzyloxyprolinal 9 was
generated in five straightforward steps10 from cis-4-
hydroxy-D-proline, and was subsequently coupled with
the boron-enolate derived from 8 in 80% isolated
yield.
OBu
7
160
858
OiPr
120
6
OCH2cPr
O(CH2)2OMe
O(CH2)2NMe2
OCH2Chx
121
110
2u
2v
2w
4
26
482
567
12
a See Refs. 15 and 16 for details of in vitro and cell assays.
ethanol)14 resulted in selective formation of adduct 16.
Treatment with sodium ethoxide afforded intermediate
17, which was subsequently deprotected to the amino
alcohol and converted to amide 18. Reduction of the
pyridine ring with platinum(IV) oxide/hydrogen in ace-
tic acid gave two isomers 2q and 19, which were sepa-
rated by silica gel chromatography.
Removal of the oxazolidinone auxiliary from aldol
adduct 1011 with lithium hydroperoxide provided a b-
hydroxy acid, which was exposed to diphenylphosphoryl
azide (DPPA). Upon Curtius-rearrangement, the incipi-
ent isocyanate was readily captured by the hydroxyl
group to furnish oxazolidinone 11 in 72% yield (two
steps).12 Hydrolysis with lithium hydroxide gave the de-
sired amino alcohol 12 in 75% yield, and was subse-
quently coupled to 5-methyl-N,N-dipropyl-isophthalic
acid 134a in the presence of HOBt/EDCI (or HATU/
HOAt). Final deprotection with trifluoroacetic acid
(TFA) furnished BACE-1 inhibitor 2f in 46% (two
steps).
The bioactivity of these novel BACE-1 inhibitors was
assessed in standard in vitro15 and cellular16 assays,
with select results shown in Table 1. The effect of ring
size was probed in a series of unsubstituted analogs
2a, 2b, and 2g (R0 = H, n = 0–2). While the azetidine
2a shows only moderate activity (BACE-1
IC50 = 770 nM, cellular IC50 = 2800 nM), the increased
molecular volume of pyrrolidine 2b and piperidine 2g
greatly improved in vitro activity (BACE-1
IC50 = 65 nM and 130 nM, respectively). This level of
potency compares favorably to the lead inhibitor 1
(BACE-1 IC50 = 8 nM) given that inhibitors 2b and
2g are not garnering binding affinity from the 3-
methoxybenzyl substitution.
An alternate route to piperidine-based BACE-1 inhibi-
tors is shown in Scheme 2, exemplified by the synthesis
of compound 2q. Addition of (S)-N,N-dibenzyl-3,5-
difluorophenylalaninal 1413 to 2-lithio-4-chloropyridine
(regioselectively formed from 4-chloropyridine 15 and
n-butyllithium in the presence of 2-(dimethylamino)

416
U. Iserloh et al. / Bioorg. Med. Chem. Lett. 18 (2008) 414–417
fort to explore the SAR around 4-benzylpiperidine 2i,
15, nBuLi,
DMAE,
Cl
Ar
Ar
we evaluated methoxybenzyl analogs 2j–2l, phenethyl
derivative 2m, as well as the cyclohexyl (2n) and cyclo-
hexylmethyl (2o) variants. However, reduced in vitro
and cellular activity was observed in each case. Finally,
we decided to explore the SAR of 4-alkoxypiperidines
2p–2w. Gratifyingly, small alkoxy groups such as eth-
oxy, butoxy, cyclopropylmethoxy, and methoxyethoxy
showed single-digit nanomolar IC50 with good cellular
activity (cell IC50 = 106–160 nM).
H
Bn2N
Bn2N
N
-78 °C,43%
O
OH
14
16
OEt
Pd(OH)2/C, H2,
72%; EDCI,
Ar
Bn2N
17
NaOEt
N
90 °C
66 %
13, HOBt, 63%
OH
OEt
Ar
PtO2/H2,
AcOH
O
O
Inspection of the X-ray structure for 4-ethoxypiperidine
˚
derivative 2q reveals a 2.4 A hydrogen-bond from the
Pr2N
N
N
H
29%
OH
18
Thr72 backbone NH of the enzyme flap to the oxygen
atom of the 4-ethoxypiperidine, while the S20 pocket
remains unoccupied (Fig. 4). Energetically, the
favorable interaction of the ether oxygen with the flap
NH compensates for the absence of hydrophobic
interactions that are harnessed by inhibitors such as 2i
with larger substituents in the S20 pocket, thus enabling
improved cellular activity. By comparison, the 4-benzyl-
oxypyrrolidine 2e cannot establish a similar hydrogen-
bonding interaction due to its different trajectory into
the S20 enzyme subsite.
OEt
OEt
Ar
Ar
O
R
O
N
N
H
N
N
H
R
H
H
OH
OH
2q
19
Scheme 2. Synthesis of BACE-1 inhibitor 2q via addition to N,N-
dibenzyl-3,5-difluorophenylalaninal (Ar = 3,5-difluorophenyl).
BACE-1 inhibitors 2e and 2f were further profiled
against a panel of relevant human aspartyl proteases
(Table 2). Benzyloxypyrrolidine 2e exhibited low selec-
tivity against cathepsin D (Ki = 36 nM) and cathepsin
E (Ki = 5 nM), while the phenoxypyrrolidine 2f was
slightly more selective (291 nM and 24 nM,
respectively).
In conclusion, we have successfully demonstrated the
potential of novel pyrrolidine- and piperidine-based
BACE-1 inhibitors. As conformationally constrained
versions of hydroxyethylamine-type peptidomimetics,
their cyclic amine core allows for targeted substitution
into the S20 subsite. In addition, a hydrogen-bonding
interaction has been discovered that significantly
improved potency for 4-alkoxypiperidines that do not
Figure 3. Overlay of BACE-1 X-ray structures for 1 (green) and 2f
(yellow). Excellent overlap is observed in the S1, S2, and S3 enzyme
subsites, and a similar space is accessed in the S20 pocket.
Moreover, additional potency can be regained through
appropriate substitution of the pyrrolidine ring as is
evident from 4-substituted pyrrolidines 2c–2f. While 4-
hydroxy and 4-methoxy analogs 2c and 2d exhibit
diminished potency relative to its parent, bulkier 4-ben-
zyloxy or 4-phenoxy groups render 2e (BACE-1
IC50 = 5 nM) and 2f (BACE-1 IC50 = 3 nM) equipotent
with lead 1.
An overlay of X-ray structures for 1 and 2f (Fig. 3, flap
removed for clarity) reveals the expected overlap on the
nonprime side (S3 through S1 enzyme subsites) along
with the conformational constraint imposed by the pyr-
rolidine ring.5 As a result, the 4-benzyloxy group accesses
a space similar to that occupied by the methoxybenzyl
moiety present in inhibitor 1.
As for the pyrrolidines, substitution of the piperidine
ring can also lead to more potent compounds. While
2h (R = Pr, 70 nM) and 2i (R = Bn, 14 nM) show im-
proved in vitro activity relative to 2g, their increased
lipophilicity also reduced the cellular activity. In an ef-
Figure 4. Close-up view of BACE-1 X-ray structure for 2q (green) and
flap (yellow) in the S20 enzyme subsite. A hydrogen-bond is formed
from the flap Thr72 NH to the oxygen atom of the 4-ethoxypiperidine.

U. Iserloh et al. / Bioorg. Med. Chem. Lett. 18 (2008) 414–417
417
Table 2. Detailed profile for novel BACE-1 inhibitors 2e and 2f
3. Leung, D.; Abbenante, G.; Fairlie, D. P. J. Med. Chem.
2000, 43, 305.
F
4. (a) Maillaird, M.; Hom, R.; Gailunas, A.; Jagodzinska, B.;
Fang, L. Y.; John, V.; Freskos, J. N.; Pulley, S. R.; Beck,
J. P.; Tenbrink, R. E. PCT Int. Appl., WO 2002002512,
2002; (b) Freskos, J. N.; Fobian, Y. M.; Benson, T. E.;
Bienkowski, M. J.; Brown, D. L.; Emmons, T. L.; Heintz,
R.; Laborde, A.; McDonald, J. J.; Mischke, B. V.;
Molyneaux, J. M.; Moon, J. B.; Mullins, P. B.; Prince,
D. B.; Paddock, D. J.; Tomasselli, A. G.; Winterrowd, G.
Bioorg. Med. Chem. Lett. 2007, 17, 73.
5. X-ray coordinates for the complex of BACE-1 with
inhibitors 1, 2e, 2f, and 2q have been deposited in the
Protein Data Bank (http://www.rcsb.org), and can be
accessed under PDB ID 2QK5, 2QMD, 2QMF, and
2QP8.
R'
F
N
H
HN
O
OH
O
Pr2N
2e R' = OBn
R' = OPh
2f
Assay, IC50
2e (R = OBn)
2f (R = OPh)
BACE-1 IC50 (nM)
BACE-2 IC50 (nM)
5
45
3
54
6. (a) Shioiri, T.; Hayashi, K.; Hamada, Y. Tetrahedron
1993, 49, 1913; (b) Hayashi, K.; Hamada, Y.; Shioiri, T.
Tetrahedron Lett. 1991, 49, 7287.
7. Pettit, G. R.; Burkett, D. D.; Barkoczy, J.; Breneman, G.
L.; Pettit, W. E. Synthesis 1996, 719.
Cell IC50 (HEK 293) (nM)
Cathepsin D Ki (nM)
Cathepsin E Ki (nM)
Pepsin Ki (nM)
150
36
165
291
24
5
158
232
8. Reetz, M. T.; Reif, W.; Holdgrun, X. Heterocycles 1989,
28, 707.
exploit binding in the S20 pocket. Further SAR studies
will be reported in the subsequent paper in this
issue.17
9. (a) Kruse, L. I.; Kaiser, C.; DeWolf, W. E.; Frazee, J. S.;
Ross, S. T.; Wawro, J.; Wise, M.; Flaim, K. E.; Sawyer, J.
L.; Erickson, R. W.; Ezekiel, M.; Ohlstein, E. H.;
Berkowitz, B. A. J. Med. Chem. 1987, 30, 486; (b) Gage,
J. R.; Evans, D. A. Org. Synth. 1989, 68, 83.
Acknowledgments
10. Synthesis of 9: BnOH, TsOH; Boc2O (70%, two steps);
BnBr, NaH (71%); LiBH4 (81%); Swern oxidation (98%);
see: Russell, M. G. N.; Beer, M. S.; Stanton, J. A.; Sohal,
B.; Mortishire-Smith, R. J.; Castron, J. L. Bioorg. Med.
Chem. Lett. 1999, 9, 2491.
11. Stereochemistry for 10 was confirmed by its cyclization to
the dihydropyrrolizine, followed by extensive NMR anal-
ysis (see Ref. 7).
12. Takacs, J. M.; Jaber, M. R.; Vellekoop, A. S. J. Org.
Chem. 1998, 63, 2742.
13. Prepared from commercially available Boc-3,5-difluor-
ophenylalanine: Reetz, M. T.; Drewes, M. W.; Schwic-
kardi, R. Org. Synth. 1999, 76, 110.
The authors are thankful to Drs. A. Buevich and T.-M.
Chan for NMR analyses, and Dr. William Greenlee for
continued support. Use of the IMCA-CAT beamline 17-
ID at the Advanced Photon Source was supported by
the companies of the Industrial Macromolecular Crys-
tallography Association through a contract with the
Center for Advanced Radiation Sources at the Univer-
sity of Chicago. Use of the Advanced Photon Source
was supported by the U.S. Department of Energy, Office
of Science, Office of Basic Energy Sciences, under Con-
tract No. W-31-109-Eng-38.
14. Choppin, S.; Gros, P.; Fort, Y. Eur. J. Org. Chem. 2001,
603.
15. BACE-1 enzymatic inhibition was determined using an
APP-derived peptide containing the Swedish mutant: Ken-
nedy, M. E.; Wang, W.; Song, L.; Lee, J.; Zhang, L.; Wong,
G.; Wang, L.; Parker, E. Anal. Biochem. 2003, 319, 49.
16. Compound IC50s for inhibition of Ab40 production were
determined by incubating HEK-293 cells, stably transfec-
ted with the human APP cDNA containing both Swedish
and London FAD mutations, with increasing concentra-
tions of BACE-1 inhibitors. Ab40 levels were measured in
the cell culture media using an Ab1–40 specific ELISA
assay: Zhang, L.; Song, L.; Terracina, G.; Liu, Y.;
Pramanik, B.; Parker, E. Biochemistry 2001, 40, 5049.
17. Iserloh, U.; Pan, J.; Stamford, A. W.; Kennedy, M. E.;
Zhang, Q.; Zhang, L.; Parker, E. M.; McHugh, N. A.;
Favreau, L.; Strickland, C.; Voigt, J. Bioorg. Med. Chem.
Lett. 2008, 18, 418.
References and notes
1. General reviews of AD: (a) Nguyen, J.; Yamani, A.; Kiso,
Y. Curr. Pharm. Des. 2006, 12, 4295; (b) Zimmermann,
M.; Gardoni, F.; Di Luca, M. Drugs Aging 2005, 22, 27;
(c) Citron, M. Nat. Rev. Neurosci. 2004, 5, 677.
2. BACE-1 is also known as b-secretase. Recent efforts toward
BACE-1 inhibitors have been amply reviewed: (a) Durham,
T. B.; Shepherd, T. A. Curr. Opin. Drug Discovery Dev.
2006, 9, 776; (b) Guo, T.; Hobbs, D. W. Curr. Med. Chem.
2006, 13, 1811; (c) Thompson, L. A.; Bronson, J. J.; Zusi, F.
C. Curr. Pharm. Des. 2005, 11, 3383; (d) Baxter, E. W.;
Reitz, A. B. Ann. Rep. Med. Chem. 2005, 40, 35; (e)
Cumming, J. N.; Iserloh, U.; Kennedy, M. E. Curr. Opin.
Drug Discovery Dev. 2004, 7, 536.