3,4-Fused cyclohexyl sulfones as γ-secretase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2005.12.083

The 3,4-fused sulfamides, sulfonamides and sulfone have been identified as highly potent γ-secretase inhibitors. Evaluation of the SAR of substitution within these series has allowed the identification of a range of compounds which significantly reduce br

Full text of DOI:10.1016/j.bmcl.2005.12.083

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3073–3077
3,4-Fused cyclohexyl sulfones as c-secretase inhibitors
Duncan Shaw,a,* Jonathan Best,c Kevin Dinnell,a Alan Nadin,a Mark Shearman,b
Christine Pattison,a James Peachey,c Michael Reilly,a Brian Williams,a
Jonathan Wrigleyb and Timothy Harrisona
aDepartment of Medicinal Chemistry, Neuroscience Research Centre, Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
bDepartment of Molecular and Cellular Neuroscience, Neuroscience Research Centre, Terlings Park,
Eastwick Road, Harlow, Essex CM20 2QR, UK
cDepartment of In Vivo Neuroscience, Neuroscience Research Centre, Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
Received 12 December 2005; revised 23 December 2005; accepted 28 December 2005
Abstract—The 3,4-fused sulfamides, sulfonamides and sulfone have been identified as highly potent c-secretase inhibitors. Evalua-
tion of the SAR of substitution within these series has allowed the identification of a range of compounds which significantly reduce
brain Ab in transgenic mouse models and thus have potential as possible treatments for Alzheimer’s disease.
Ó 2006 Elsevier Ltd. All rights reserved.
Alzheimer’s disease (AD) is characterized by progressive
cognitive decline accompanied by the deposition of pla-
ques composed of the neurotoxic 40–42 amino acid pep-
tide amyloid-b (Ab) and by increased levels of soluble
Ab. One of the major hypotheses for the progression
of AD is that the extracellular accumulation of Ab is
the primary pathological event leading to neurodegener-
ation, dementia and ultimately death. In order to test
this hypothesis and potentially to provide pharmacolog-
ical agents to arrest the progress of the disease, the inhi-
bition of Ab production is of particular interest. The
release of Ab is the result of cleavage of b-amyloid pre-
cursor protein (b-APP) by two proteases. The first cleav-
age of b-APP by b-secretase produces a b-APP C-
terminal fragment which is cleaved within the cell mem-
brane by the aspartyl protease c-secretase to release Ab.
The identification of a selective orally active c-secretase
inhibitor has been targeted as an attractive way to test
the amyloid hypothesis.1
As part of our ongoing medicinal chemistry programme
the cyclohexane 12 (3 nM) and the sulfonamides 23
(4.2 nM) and ( )-33 (1.8 nM) were identified as potent
c-secretase inhibitors and previously disclosed.4–7
The tolerance for substitution of the cyclohexyl ring was
significant in that it allowed improved absorption prop-
erties to be achieved. The compounds did, however,
show a relatively short half-life in rodent (e.g., 2: F
31%, t1/2 1.1 h in rat) and high turnover in rat liver
microsomes (2 33%; 3 78%).8 We set out to synthesise
stable analogs of these sulfonamides which should also
take advantage of the known lipophilic binding pocket
identified from the X-ray crystal structure of the potent
inhibitor 4.9
O
H
SO2
N
HN
N
O2
S
O2
S
(+/-)
NHSO2Me
O2
S
NHSO2Me
O2
S
O2
S
F
F
F
F
F
(+/-)
Cl
Cl
Cl
F
F
F
Cl
5c
Cl
4
F
F
1
2
3
The relatively high potency of 2 and 3 compared to
many of the substituted cyclohexanes which we had
examined3 led us to believe that a single sulfonamide
binding site existed that could be accessed from either
Keywords: Sulfones; c-Secretase inhibitors; Alzheimer’s disease.
Corresponding author. Tel.: +44 1279 440427; fax: +44 1279
440390; e-mail: Duncan_shaw@merck.com
*
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.12.083

3074
D. Shaw et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3073–3077
the 3 or 4 positions of the cyclohexane ring. A 3,4-fused
cyclic sulfamide was postulated to take full advantage of
this binding site, introduction of an N-substituent would
then potentially allow us to access the lipophilic binding
site. It was hypothesised (Fig. 1) that the ethyl substitut-
ed 3,4-fused cyclic sulfamide ( )-5c (orange carbons)
would overlay both the hydrogen bond acceptor (SO2)
and the lipophilic group with the same groups in the
X-ray crystal structure of 4 (grey carbons).10
A versatile cyclic sulfamide synthesis was developed
(Scheme 1) which allowed exploration of the lipophilic
binding pocket by variation of the nitrogen substituent
at the final stage (see Tables 1–3).
The cyclohexanone 62 was alkylated with SEM–Cl to
give, as a single isomer,11 the protected a-hydroxymethyl
ketone ( )-7.12 Stereoselective reduction with sodium
borohydride gave the anti-alcohol which underwent
smooth azide displacement (via the mesylate) and subse-
quent reduction to give the syn-amine ( )-8.
Deprotection of the alcohol was followed by selective
sulfonylation of the amine to afford the N,N-dim-
ethylsulfamide. Oxidation of the primary alcohol to
the aldehyde facilitated the introduction of an alkyl
substituted amine ( )-9. This aminosulfamide under-
went a facile cyclisation to the cyclic sulfamide ( )-5
Figure 1. Overlay of the X-ray structure of 4 (grey carbons) with the
minimised structure of the 5c (orange carbons).
Cl
Cl
F
O2S
F
O2S
i
F
F
(+/-)
(+/-)
O
TMS
O
O
6
7
ii-v
Cl
Cl
F
O2S
F
O2S
vi-x
F
F
(+/-)
NHR
(+/-)
O
TMS
HN
SO2
NH2
NMe2
xi
8
9a-h
Cl
F
O2S
F
(+/-)
HN
NR
S
O2
5a-h
Scheme 1. Reagents and conditions: (i) LHMDS, THF, SEM–Cl, À78 °C; 60%; (ii) NaBH4, À20 °C, IPA, 80%; (iii) MsCl, NEt3, DCM, 97%; (iv)
NaN3, DMF, 90 °C, 75%; (v) PPh3, THF, H2O, 95%; (vi) BF3ÆOEt2, DCM, 95%; (vii) Me2NSO2Cl, NEt3, DMA, DCM, 80%; (vii) Dess–Martin
periodinane, DCM, 80%; (ix) RNH2, EtOH; (x) NaBH4; (xi) THF, reflux, 70% (3 steps).

D. Shaw et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3073–3077
3075
Table 1. Variation of properties with the introduction of N-substit-
uents
An asymmetric synthesis was developed based on the
chiral deprotonation13 of the ketone 6 and trapping as
a trimethylsilyl enol ether followed by a Tin (IV) chlo-
ride catalysed alkylation with SEM–Cl to give 10
(Scheme 2). This chemistry facilitated identification of
the enantiomerically pure cyclopropyl cyclic sulfamide
11 (IC50 0.04 0.01 nM) for in vivo evaluation.
Cl
F
O S
2
F
(+/-)
The stability of the unsubstituted ring system 5a and the
strong relationship between turnover and the N-substi-
tuent of the cyclic sulfamide led us to believe this was
a major site of metabolism. We therefore targeted
the carbon-linked derivatives, the cyclic sulfonamides
12a–c (Scheme 3) and the cyclic sulfone 13 (Scheme 4).
HN
NR
S
O
2
R
Compound
IC50 (nM)14
Microsomal turnover (%)8
H
5a
5b
5c
5d
5e
5f
6.18 ( 2.0)
0.90 ( 0.3)
0.40 ( 0.1)
1.17 ( 0.7)
0.41 ( 0.1)
0.19 ( 0.2)
0.26 ( 0.04)
1.39 ( 0.5)
9
17
46
63
0
Me
Et
The enantiomerically pure amine 14 obtained as out-
lined in Schemes 1 and 2 was N-sulfonylated and then
allylated on nitrogen under forcing conditions to give
the allyl sulfonamide 15. Deprotection and activation
as the tosylate gave the cyclisation precursor 16. Treat-
ment with butyl lithium achieved a clean cyclisation to
the sultam. The sultam was deprotonated at low temper-
ature before treating with alkyl iodide and allowing to
warm to room temperature overnight. This gave very
good selectivity (>19:1) for the anti-alkyl product. Nick-
el catalysed deprotection gave the desired cyclic sulfa-
mide containing an equatorial alkyl group.
nPr
iPr
cPr
cBu
tBu
14
68
15
5g
5h
Table 2. Variation of biological properties with the cyclic sulfona-
mides and sulfones
F
The enantiomerically pure ketone 1015 was homologated
using a Horner–Emmons reaction to give the vinyl
sulfone which could be stereospecifically reduced using
F
O S
X
2
SO
2
R
Cl
Cl
Cl
Compound IC50 (nM)14 Microsomal turnover (%)8
0.08 ( 0.02) 20
F
O2S
F
O2S
X
R
i-ii
F
F
NH Me 12a
NH Et 12b
NH nPr 12c
CH2 Et 13
O
0.06 ( 0.05)
7
TMS
0.24 ( 0.05) 32
0.76 ( 0.06) 15
O
O
6
10
see scheme 1
with extrusion of dimethylamine by heating in
tetrahydrofuran.
Cl
F
O2S
The resultant racemic cyclic sulfamides exhibited potent
c-secretase inhibition, indicating that the lipophilic
binding pocket had been accessed. For unbranched alkyl
groups optimal activity was seen with ( )-5c (R = Et),
while small branched alkyl groups, for example, ( )-5f
(R = cPr) were of greatest potency. The larger
alkyl-substituted analogues showed an increased
turnover in rat liver microsomes (R = cBu ( )-5g) as
did the straight chain analogues (R = Et ( )-5c and
R = nPr ( )-5d).
F
N
NH
S
O2
11
Scheme 2. Reagents and conditions: (i) BuLi, LiCl, [S-(R*,R*)]-(À)-
bis-a-methylbenzylamine, À100 °C, TMS–Cl, THF; (ii) 1 mol% SnCl4,
À40 °C, SEM–Cl, DCM 75%.
Table 3. Variation of in vivo efficacy properties
Compound
In vitro IC50 (nM)14
[Plasma] EC50 (nM)18
[Brain] EC50 (nM)18
Brain/plasma ratio
ED50 at 4 h (mg/kg)18
11
0.04 ( 0.01)
0.06 ( 0.05)
0.76 ( 0.06)
104 ( 42)
380 ( 89)
761 ( 216)
20 ( 4)
88 ( 16)
327 ( 90)
0.16
0.18
0.41
11.4 ( 1.2)
2.8 ( 1.2)
3.9 ( 1.2)
12b
13

3076
D. Shaw et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3073–3077
Cl
out as in Scheme 3. A cyclisation to the phosphonyl-
3,4 fused cyclic sulfone was achieved in good yield.
Cl
F
O2S
F
O2S
i-ii
F
F
A second Horner–Emmons reaction was carried out to
provide the unsaturated sulfone 19. Reduction of the
double bond by hydrogenation gave the undesired syn-
alkyl ring substituent (>19:1 ratio), this compound
showing no inhibition of c-secretase at concentrations
up to 9 lM. The desired anti-alkyl compound 13 could
be isolated cleanly after a base catalysed epimerisation.
O
O
TMS
TMS
NH2
N
O2S
Me
14
15
iii-iv
The sulfonamides were highly potent inhibitors of
c-secretase and showed a similar SAR of straight chain
alkyl substitution to that seen with the sulfamides (i.e.,
ethyl was optimal). There was, however, no further ben-
efit seen from the introduction of branched alkyl groups
(data not shown). The ethyl substituted cyclic
sulfonamide 12b was also optimal in terms of rat micro-
somal stability and showed reduced turnover relative to
the corresponding sulfamides and the acyclic sulfona-
mides 2 and 3.
Cl
Cl
F
O2S
F
O2S
v-vii
F
F
TsO
N
NH
O2S
Me
R
S
O2
12a-c
16
Scheme 3. Reagents and conditions: (i) MsCl, TEA, DCM, 95%; (ii)
allyl bromide, NaH, DMF, 70 °C, 70%; (iii) BF3ÆOEt2, DCM, 95%; (iv)
The ethyl substituted sulfone 13 showed somewhat
reduced in vitro inhibition of c-secretase compared to
the sulfonamide 12b but was stable in rat liver
microsomes.
n
TsCl, pyridine, DMAP, 80%; (v) BuLi, THF, À30 °C to rt, 85%; (vi)
LiHMDS, RI, À78 °C to rt, 90%. (vii) DIBAL-H, (dppp)NiCl2,
toluene, 70%.
Cl
Cl
F
O2S
With these data in hand we set about evaluating the
in vivo properties of the compounds using the
APP-YAC mouse model.16 The brain levels of DEA-
extractable Ab17 at 4 h post-oral dosing, relative to vehi-
cle and a positive control c-secretase inhibitor together
with the plasma and brain drug concentrations were
determined for the key compounds 11, 12b and 13.
The sulfone 13 was considered an interesting molecule
to investigate in vivo because the absence of the sulfon-
amide NH was postulated to offer an improvement in
terms of brain penetration.
F
O2S
i-ii
F
F
O
O
TMS
TMS
O
SO2Me
10
17
iii-v
Cl
Cl
F
O2S
F
O2S
F
In assessing the relative merits of the compounds we not-
ed that the in vitro IC50 and brain level required for effi-
cacy tracked well with each other. The highly potent
cyclic sulfamide 11 had efficacy at very low brain concen-
trations, but relatively poor plasma levels were obtained
after oral dosing leading to an ED50 of 11.4 mg/kg. The
cyclic sulfonamide 12b was shown to be efficacious at
lower doses due to the improved pharmacokinetics
observed. The cyclic sulfone 13 required higher brain
levels to show efficacy but the expected improvement in
brain penetration and the good pharmacokinetics meant
that this compound compared well with the cyclic sulfon-
amide 12b in terms of dose required for efficacy.
F
vi-vii
TsO
S
P(OEt)2
S
O2
O2
O
viii-ix
19
18
Cl
F
O2S
F
S
O2
13
Separate evaluation showed that the compounds de-
tailed in this communication inhibit notch cleavage at
concentrations comparable to those required to inhibit
Ab production, consistent with data from several exist-
ing series of c-secretase inhibitors.19
Scheme 4. Reagents and conditions: (i) nBuLi, MeSO2CH2P(O)(OEt)2,
THF À78 °C, 78%; (ii) L-SelectrideTM, THF À40 °C, 90%; (iii)
LiHMDS, diethyl chlorophosphonate, 68%; (iv) BF3ÆOEt2, DCM,
85%; (v) TsCl, pyridine, DMAP, 90%; (vi) BuLi, À78 °C, THF 76%;
(vii) LiHMDS, acetaldehyde 90%; (viii) 10% Pd/C EtOH, H2 (1 atm),
95%; (ix) KOtBu, THF, 89%.
In conclusion, we have shown that by a combination of
conformational restriction and exploitation of a well-
defined lipophilic binding pocket, we can reliably
achieve excellent in vitro c-secretase inhibition. In addi-
L-SelectrideTM to give the syn-sulfone 17. At this stage
the methyl sulfone was activated as the phosphonate
and then O-deprotection and activation were carried

D. Shaw et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3073–3077
3077
5. Castro, J. L.; Churcher, I.; Dinnell, K.; Harrison, T.;
Kerrad, S.; Nadin, A. J.; Oakley, P. J.; Owens, A. P.;
Shaw, D. E.; Teall, M. R.; Williams, B. J.; Williams, S.
WO2002081435.
6. Churcher, I.; Dinnell, K.; Harrison, T.; Kerrad, S.; Nadin,
A. J.; Oakley, P. J.; Shaw, D. E.; Teall, M. R.; Williams,
B. J.; Williams, S. WO2003018543.
7. Churcher, I.; Harrison, T.; Kerrad, S.; Oakley, P. J.;
Shaw, D. E.; Teall M. R.; Williams, S. WO2004031137.
8. Rat microsomal turnover was determined as follows: test
compound was incubated at 1 lM concentration with rat
microsomal protein 5 mg/mL and NADPH at 37 °C for
20 min. Samples were analyzed by LC–MS to identify
percentage of parent remaining.
tion, we have been able to reduce microsomal turnover
and achieve good plasma levels after oral dosing. We
have identified three different stable core structures with
complementary in vivo properties. In the case of the cyc-
lic sulfamides we achieved in vivo potency at low brain
levels. With the cyclic sulfonamides we were able to add
to this good pharmacokinetics after oral dosing. The
cyclic sulfones offer an alternative series where improved
brain–plasma ratio could lead to a more desirable profile
for our target central inhibition of c-secretase.
Acknowledgment
9. Dinnell, K.; Teall, M.; Oakley, P.; Harrison, T.; Shaw, D.;
Shearman, M.; Williams, B., in preparation.
The authors thank Beth Oxley and Robert Newman for
their assistance in screening, Peter Hunt for the molecu-
lar modelling.
10. Manual overlay using the SYBYL (Version 8) molecular
modelling program by TRIPOS.
11. As determined by H NMR using the nuclear Overhauser
1
effect.
12. Crich, D.; Lim, L. B. L. Synlett 1990, 117.
13. (a) Hodgson, D. M.; Stent, M. A. H. Top. Organomet.
Chem. 2003, 5, 1; (b) Honda, T.; Kimura, N.; Tsubuki, M.
Tetrahedron: Asymmetry 1993, 4, 21.
14. Clarke, E. E.; Shearman, M. S. J. Neurosci. Methods 2000,
102, 61, all IC50 data shown are the average of a minimum
of three independent results.
15. Enantiomeric excess determined by chiral HPLC (90%),
the absolute stereochemistry of the enantiomers with
in vivo activity was determined by X-ray crystallography,
performed on a derivative of 12b. The enantiomer of
compound 11 was prepared and shown not to be a c-
secretase inhibitor in vitro.
16. 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.
References and notes
1. (a) Harrison, T.; Churcher, I.; Beher, D. Curr. Opin. Drug
Discov. Devel. 2004, 7, 709; (b) Hardy, J. A.; Higgins, G.
A. Science 1992, 256, 184.
2. Teall, M.; Oakley, P.; Harrison, T.; Shaw, D.; Kay, E.;
Elliott, J.; Gerhard, U.; Castro, J. L.; Shearman, M.; Ball,
R. G.; Tsou, N. N. Bioorg. Med. Chem. Lett. 2005, 15, 2685.
3. Churcher, I.; Beher, D.; Best, J. D.; Castro, J. L.; Clarke,
E. E.; Gentry, A.; Harrison, T.; Hitzel, L.; Kay, E.;
Kerrad, S.; Lewis, H. D.; Morentin-Gutierrez, P.; Morti-
shire-Smith, R.; Oakley, P. J.; Reilly, M.; Shaw, D. E.;
Shearman, M. S.; Teall, M. R.; Williams, S.; Wrigley, J. D.
J. Bioorg. Med. Chem. Lett. 2006, 16, 280.
4. Sulfonamide 3 was obtained from the aldehyde 20
previously described by Jelley, R.; Beher, D.; Elliott, J.;
Gibson, K.; Clarke, E. E.; Harrison, T.; Lewis, H. D.;
Shearman, M.; Wrigley, J. D. J., in preparation as
outlined below. (i) NH4OAc, NaBH3CN; (ii) MsCl,
DCM, Et3N.
17. Best, J.; Jay, M.; Ohtu, F.; Ma, J.; Nadin, A.; Ellis, S.;
Lewis, H.; Pattison, C.; Reilly, M.; Harrison, T.; Shear-
man, M.; Williamson, T.; Atack, J. J. Pharmacol. Exp.
Ther. 2005, 313, 902.
18. EC50 obtained using
a dose response titration to
determine the plasma and brain concentrations neces-
sary to achieve a 50% reduction in DEA-extractable
brain Ab, 4 h after administration:ED50 defined as the
oral dose required to achieve these plasma and brain
levels.
Cl
Cl
F
O2S
F
O2S
i-ii
F
F
(+/-)
(+/-)
19. Lewis, H. D.; Perez Revuelta, B.; Nadin, A.; Neduvelil, J.;
Harrison, T.; Pollack, S. J.; Shearman, M. S. Biochemistry
2003, 42, 7580.
NHSO2Me
CHO
3
20