Cyclic sulfamide γ-secretase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.06.084

A novel series of N-alkyl-substituted cyclic sulfamides were developed from a screening hit. Chemistries were developed which allowed surveys of N-alkyl groups and amines resulting in the identification of N-trifluoroethyl- substituted cyclic sulfamides with good in vitro and in vivo γ-secretase activity. One compound with subnanomolar activity elicited a reduction in brain Aβ40 after oral dosing in APP-YAC mice.

Full text of DOI:10.1016/j.bmcl.2005.06.084

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4212–4216
Cyclic sulfamide c-secretase inhibitors
b
c
d
a
b
Tim Sparey,a, Dirk Beher, Jonathan Best, Mirlinda Biba, Jose L. Castro, Earl Clarke,
*
´
Joanne Hannam,a Timothy Harrison,a Huw Lewis,b Andrew Madin,a Mark Shearman,b
Bindi Sohal,a Nancy Tsou,e Christopher Welchd and Jonathan Wrigleyb
aDepartment of Medicinal Chemistry, Merck Sharp and Dohme Research Laboratories, The Neuroscience Research Centre,
Terlings Park, Harlow, Essex CM20 2QR, UK
bDepartment of Cellular and Molecular Neuroscience, Merck Sharp and Dohme Research Laboratories,
The Neuroscience Research Centre, Terlings Park, Harlow, Essex CM20 2QR, UK
cDepartment of In Vivo Neuroscience; Merck Sharp and Dohme Research Laboratories, The Neuroscience Research Centre,
Terlings Park, Harlow, Essex CM20 2QR, UK
dDepartment of Process Research, Merck Research Laboratories, Rahway, NJ 07065, USA
eDepartment of Medicinal Chemistry, Merck Research Laboratories, Rahway, NJ 07065, USA
Received 14 February 2005; revised 23 June 2005; accepted 27 June 2005
Available online 27 July 2005
Abstract—A novel series of N-alkyl-substituted cyclic sulfamides were developed from a screening hit. Chemistries were developed
which allowed surveys of N-alkyl groups and amines resulting in the identification of N-trifluoroethyl-substituted cyclic sulfamides
with good in vitro and in vivo c-secretase activity. One compound with subnanomolar activity elicited a reduction in brain Ab40
after oral dosing in APP-YAC mice.
Ó 2005 Elsevier Ltd. All rights reserved.
Alzheimerꢀs disease (AD), the main form of dementia
found in the elderly population, is the fourth leading
cause of death and is predicted to affect 20 million peo-
ple worldwide within 20 years.1 Current therapies are
palliative, and include the acetylcholinesterase inhibitors
Aricept, Reminyl, and Exelon. One approach to identify
a therapy capable of preventing the progression of AD is
the inhibition of b-amyloid (Ab) production, specifically
Ab42.2 This peptide constitutes the major component of
extracellular plaques and is a metabolite of b-amyloid
precursor protein (b-APP), formed through sequential
cleavage by b-secretase and then c-secretase.
pounds. As previously reported, the first iteration from
screening hit 2a produced a number of potent com-
pounds, including 5-chlorothiophene sulfonamide 2b
and n-butyl sulfonamide 2c.3 Initially, we investigated
the incorporation of heteroatoms into the alkyl side chain
of 2c, in particular the replacement of a methylene with a
nitrogen atom. Thus, reacting amine 3 with the appropri-
ate sulfamoyl chloride 4 gave acyclic sulfamides 6 in one
step (Scheme 1). Alternatively, to explore a wider range
of amines more quickly, reaction with intermediate cate-
chol sulfamate 5 produced sulfamides 6 in good yield.
The endo stereochemistry of the NH2 group in 3 was al-
ready established4 but was confirmed through the obser-
vation of NOEs between the endo NH and pseudo axial
protons on the benzylic carbons of 6b.
Recently, we described sulfonamide derivatives of substi-
tuted bicyclo [4.2.1]nonanes such as 1, which are potent
c-secretase inhibitors.3 Compound 1 was identified by
optimization of screening hit 2a. In parallel with the opti-
mization of the sulfonamide substituent in 1, alternative
motifs to the sulfonamide were investigated. Here, we de-
scribe the evolution of a series of novel sulfamides with
potent c-secretase inhibitory activity and present data
illustrating the in vivo efficacy of this new class of com-
Keywords: Alzheimerꢀs disease; c-Secretase inhibitor; Cyclic sulfamide.
*
Corresponding author. Tel.: 44 1279 440298; fax: 44 1279 440187;
e-mail: tim_sparey@merck.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.06.084

T. Sparey et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4212–4216
4213
12 in good overall yield.4 Utilizing the same Strecker/
LAH sequence as described previously, diamine 13
was prepared. A byproduct was observed during the
reduction with LAH, identified as amine 14 (presumably
derived from loss of cyanide and reduction of the imine).
Amine 14 was removed by chromatography on silica.
Diamine 13 was then cyclized to cyclic sulfamide 15
using sulfamide and then functionalized by alkylation
with n-propyl bromide with high regioselectivity
(>95% by crude NMR). Benzyl ether 16 was deprotected
and converted to triflate 17, which was carbonylated, re-
duced and oxidized to give aldehyde 18. A Horner–Em-
mons reaction followed by reduction gave allylic alcohol
19, which was activated with PBr3 and reacted with an
appropriate amine. Subsequently, a shorter synthetic
sequence was developed, whereby triflate 17 was reacted
with vinyl boronate 215 and the resultant silyl ether
deprotected with TBAF.
Scheme 1. Reagents and conditions: (a) RNHSO2Cl 4, Et3N, MeCN,
reflux, 35–81%; (b) catechol sulfate, THF, 0 °C to rt, 81%; (c) amine,
dioxane, 80 °C, 66–80%.
To restrict the freedom of rotation of the sulfamide
group, conformationally constrained cyclic sulfamides
10 were synthesized. This was accomplished through a
Strecker reaction and LAH reduction sequence on ke-
tone 7, followed by cyclization of diamine 8 with sulfa-
mide in pyridine (Scheme 2). Subsequent alkylation of
unsubstituted cyclic sulfamide 9, using sodium hydride
in DMF as base and the appropriate alkyl halide,
furnished substituted cyclic sulfamides 10a–d. Good
regioselectivity was observed during the reaction, with
no monoalkylation observed on the endo nitrogen.
However, alkylation on both nitrogens gave rise to a
small quantity of 11 (with n-propyl bromide as
electrophile). The regiochemistry and stereochemistry
of 10b were confirmed by the observation of NOEs
between the N–H and the pseudo axial protons on the
7-membered ring.
In the synthesis of cyclic sulfamides 10, the direct alkyl-
ation route to the N-trifluoroethyl analogue 10d was
low yielding (10%). An improved route was required to
facilitate the evaluation of more functionalized com-
pounds. This was achieved by installation of the trifluor-
ethyl group prior to cyclic sulfamide formation. Thus,
trifluoroacetylation of diamine 13 using trifluoroacetic
anhydride followed by reduction with borane gave triflu-
oroethyl amine 22 (Scheme 4). The observed regioselec-
tivity was attributed to preferential reactivity of the
less-hindered primary amine. Cyclization gave sulfamide
23 in good yield. Side chains containing allylic amines
could be introduced from triflate 24 as described above,
allowing a range of amines to be incorporated efficiently.
In a further evolution of this route, boronate 25a was
synthesized6 and reacted with 4-(trifluoromethyl)piperi-
dine to provide boronate 25b, which was then coupled
with triflate 24 to furnish 26f in improved yield.
In an attempt to improve the potency of these com-
pounds, substituents appended to the benzene ring were
introduced following precedent from the sulfonamide
series.3 This was achieved regioselectively using the
known lead tetraacetate-mediated oxidation of ketone
7 (Scheme 3), and the phenol converted to benzyl ether
Some interesting structure–activity relationships were
observed for the compounds described above. We found
that the replacement of a methylene in compound 2c
(IC50 610 nM)7 with NH gave the more potent acyclic
sulfamide 6b (IC50 106 nM; Table 1). However, shorten-
ing (6a) or increasing (6c) the length of the sulfamide al-
kyl group was detrimental to potency, whereas replacing
the propyl group with a trifluoroethyl group resulted in
an equipotent compound (6d). A different trend was ob-
served for cyclic sulfamides, notably for the n-propyl
(10b) and n-butyl (10c) analogues which were equipo-
tent, but more interestingly 10d that displayed an order
of magnitude increase in potency. This may be attribut-
able to the conformational restriction imposed by the
cyclic sulfamide ring better orientating the lipophilic tri-
fluoroethyl group for binding (6d vs 10d). Compounds
9 and 11 were both found to be inactive (IC50s >
10,000 nM).
Compound 2b was found to have poor systemic expo-
sure in rat after oral dosing (Table 2). In comparison,
6b and 10b showed improved exposure. Given the phar-
macokinetic advantage of 10b over 6b, presumably due
to the removal of one sulfamide N–H, further efforts
focused on optimizing the cyclic sulfamide series.
Scheme 2. Reagents and conditions: (a) NaCN, NH4OH, ammonia,
NH4Cl MeOH, 0 °C, 70%; (b) LAH, THF, rt, 77%; (c) sulfamide,
pyridine, reflux, 24 h, 85%; (d) NaH, alkyl halide, DMF, rt, 24 h,
10–71%.

4214
T. Sparey et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4212–4216
Scheme 3. Reagents and conditions: (a) Pb(OAc)4, TFA, 75%; (b) BnBr, NaH, DMF, rt, 48 h, 83%; (c) NaCN, NH4OH, ammonia, NH4Cl, MeOH,
0 °C, 74%; (d) LAH, THF, 53%; (e) sulfamide, pyridine, reflux, 24 h, 89%; (f) NaH, n-propyl bromide, DMF, rt, 24 h, 59%; (g) 10% Pd on C, H2,
EtOAc, MeOH, 4 h, 99%; (h) Tf2O, pyridine, 0 °C to rt, 30 min, 89%; (i) CO, Pd(OAc)2, dppb, Et3N, MeOH, DMF, 56%; (j) DIBAL, THF, 95%; (k)
PDC, DCM, 89%; (l) methyl diethylphosphono acetate, LiOHÆH2O, THF, 79%; (m) 21, Pd(OAc)2, dppp, K2CO3, DMF, 80 °C, 14 h, 51%; (n) TBAF,
THF, rt, 79%; (o) PBr3, DCM, À5 °C, 1 h; (p) amine, DCM, rt, 63%.
To assess these improvements in more functionalized
sulfamides, 20a was synthesized and resolved8 for direct
comparison with 1. While the inhibitory activity was
reduced 7-fold, an order of magnitude increase in
exposure was observed after oral dosing (Table 3).
From this study, the optimum amine was identified as 4-
trifluoromethylpiperidine, which, when combined with
an N-trifluoroethyl cyclic sulfamide, produced
a
compound with subnanomolar potency (26f). Further
variations to the allylic side chain portion of these
compounds will be reported shortly.
N-Trifluoroethyl cyclic sulfamides containing allylic
amine side chains were synthesized to improve
To assess the efficacy of 26f, the racemate was resolved8
and the (À)-enantiomer10 (IC50 0.24 nM (n = 3), 99%
ee8; (+)-enantiomer IC50 111 nM, 99% ee8) dosed orally
to APP-YAC mice11 (Chart 1). In a dose–response
experiment, (À)-26f demonstrated a significantly im-
proved efficacy profile compared to 1 with an ID50 of
17 mg/kg p.o.
potency.
A 6-fold improvement in potency was
observed by replacing the n-propyl (20a) with a
trifluoroethyl group (26a; Table 4). To assess the
effects on potency of replacing the morpholine with
amines of different basicity and functionality, a range
of amines was incorporated. The data suggested that
a substituent in the 4-position of the piperidine
improved activity (26b vs 26c).
Summary
Good potency was observed for compounds with
heteroatoms and/or substituents at the 4-position of
the nitrogen heterocycle contributing to a pKa value
within the 6–7 range (26a, e, f, and 20b).
A novel series of cyclic sulfamides has been developed
from screening hit 2a. An improved pharmacokinetic
profile was demonstrated for simple sulfamides 6b and

T. Sparey et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4212–4216
4215
Scheme 4. Reagents and conditions: (a) Trifluoroacetic anhydride, THF, À78 °C to rt, 24 h, 90%; (b) borane, THF, reflux, 2 h, 60–82%;
(c) sulfamide, pyridine, reflux, 24 h, 91%; (d) H2, Pd on C, EtOAc/MeOH, 6 h, 93%; (e) Tf2O, pyridine, 0 °C to rt, 1 h, 88%; (f) 21, Pd(OAc)2,
dppp, K2CO3, DMF, 80 °C, 14 h, 50%; (g) TBAF, THF, rt, 3 h, 79%; (h) i—PBr3, DCM, À5 °C, 1 h, ii—amine, DCM, rt, 4 h, 45–83%; (i)
4-(triflouromethyl)piperidine, K2CO3, MeCN, rt, 95%; (j) 25b, Pd(OAc)2, dppp, K2CO3, DMF, 80 °C, 70%.
Table 1. Inhibitory activities of sulfamides 6 and 10
Table 2. Rat systemic exposures for 2b, 6b, and 10b
Compound
Dose AUC
(mpk) (systemic)
(lM h)
Compound
IC50,
nM (n)
R
Compound
IC50,
nM (n)
2b
6b
30
30
1
0.01
0.6
6a
6b
6c
6d
394(2)
132(3)
766(2)
134(3)
Ethyl
10a
10b
10c
10d
429(2)
138(2)
122(2)
17(5)
n-Propyl
n-Butyl
Trifluorethyl
10b
0.1
10b compared with sulfonamide 2b. Functionalization
of these sulfamides by appendage of allylic amines to
the aromatic ring produced more potent inhibitors (20
and 26) in which improved pharmacokinetic properties
were maintained. In a dose–response experiment, (À)-
26f showed a significantly improved efficacy profile in
APP-YAC mice compared with sulfonamide 1 and elic-
ited a dose-dependent reduction in brain Ab40 with an
ID50 of 17 mg/kg p.o.
Table 3. Profiles of 1 and 20a
Compound
IC50, nM (n)
Dose (mpk)
4
AUC (systemic) (lM h)
1a
1(4)
0.38
20aa
7(2)
1
1.19
a Data shown for single enantiomers (absolute stereochemistry is unknown).

4216
T. Sparey et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4212–4216
Table 4. Inhibitory activities of cyclic sulfamides 20 and 26
Acknowledgments
The authors thank Steve Thomas and John Atack for
their help and support.
9
Compounda
R
R1
IC50
nM (n)
,
pKa
References and notes
20a
20b
26a
26b
26c
26d
26e
26f
n-Propyl
19(5)
7.1
6.7
6.8
7.9
7.5
5.2
6.1
6.4
1. Jacobsen, J. S. Curr. Top. Med. Chem. 2002, 2, 343.
2. Esler, W. P.; Wolfe, M. S. Science 2001, 293, 1449.
3. Lewis, S. J.; Smith, A. L.; Neduvelil, J. G.; Stevenson, G.
I.; Lindon, M. L.; Jones, A. B.; Shearman, M. S.; Beher,
D.; Clarke, E.; Best, J. D.; Peachey, J. P.; Harrison, T.;
Castro, J. L. Biorg. Med. Chem. Lett. 2005, 15(2), 373.
4. Belanger, P. C.; Young, R. N.; Scheigetz, J.; Dufresne, C.;
Springer, J. P. J. Org. Chem. 1982, 47, 4329.
5. Hoffmann, R. W.; Dresely, S. Synthesis 1988, 2, 103.
6. Kamabuchi, A.; Moriya, T.; Miyaura, N.; Suzuki, A.
Synth. Commun. 1993, 23, 2851.
7. Compounds were screened using a whole cell c-secretase
inhibition assay in SHSY5Y cells: Clarke, E.; Shearman,
M. S. J. Neurosci. Methods 2000, 102, 61.
n-Propyl
2.0(2)
3.4(2)
250(2)
18(2)
Trifluoroethyl
Trifluoroethyl
Trifluoroethyl
Trifluoroethyl
Trifluoroethyl
Trifluoroethyl
21(2)
8. Compounds were resolved using supercritical fluid chro-
matography: OD column eluting with 30–50% MeOH
containing 25 mM i-BuNH2 at 50 ml/min.
4(2)
9. Determined by titration using
autotitrator.
a
Sirius GLpKa
0.5(4)
10. Crystallographic data (excluding structure factors) for
(À)-26f have been deposited with the Cambridge Crystal-
lographic Data Centre as supplementary publication
number CCDC 263202. The absolute stereochemistry is
as drawn in Table 4.
a For comparison purposes, data shown for racemates.
11. Lamb, B. T.; Sisodia, S. S.; Lawler, A. M.; Slunt, H. H.;
Kitt, C. A.; Kearns, W. G.; Pearson, P. L.; Price, D. L.;
Gearhart, J. D. Nat. Genet. 1993, 5, 22.
Chart 1. In vivo efficacy of 1 and (À)-26f in YAC mice.