3-Substituted gem-cyclohexane sulfone based γ-secretase inhibitors for Alzheimer's disease: Conformational analysis and biological activity

Article abstract of DOI:10.1016/j.bmcl.2006.04.019

Previously, chemistry effort on the gem-cyclohexane series of γ-secretase inhibitors has focused on the 4-position of the cyclohexane ring. Recently chemistry has been directed towards the 3-position and substitution here has also provided compounds with

Full text of DOI:10.1016/j.bmcl.2006.04.019

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3839–3842
3-Substituted gem-cyclohexane sulfone based c-secretase
inhibitors for Alzheimer’s disease: Conformational
analysis and biological activity
Richard A. Jelley,a,* Jason Elliott,a Karl R. Gibson,a Timothy Harrison,a Dirk Beher,b
Earl E. Clarke,b Huw D. Lewis,b Mark Shearmanb and Jonathan D. J. Wrigleyb
aDepartment of Medicinal Chemistry, Merck Sharp and Dohme Research Laboratories, The Neuroscience Research Centre,
Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
bDepartment of Molecular and Cellular Neuroscience, Merck Sharp and Dohme Research Laboratories,
The Neuroscience Research Centre, Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
Received 30 January 2006; revised 6 April 2006; accepted 8 April 2006
Available online 6 May 2006
Abstract—Previously, chemistry effort on the gem-cyclohexane series of c-secretase inhibitors has focused on the 4-position of the
cyclohexane ring. Recently chemistry has been directed towards the 3-position and substitution here has also provided compounds
with high c-secretase activity.
Ó 2006 Elsevier Ltd. All rights reserved.
Alzheimer’s disease (AD) is the major cause of dementia
in the developed world. It is characterized clinically by a
progressive decline in memory and cognitive function
which ultimately leads to death. For more than a dec-
ade, it has been proposed that the neurodegeneration
observed in AD is due to the accumulation and aggrega-
tion of amyloid b-peptide (Ab) into extracellular pro-
teinaceous plaques.1 The potential consequence of
plaque formation is the loss of synaptic and neuronal
function. Ab-peptides are formed from the sequential
cleavage of amyloid precursor protein (APP) by two
proteases (b- and c-secretase). It is the subsequent aggre-
gation and precipitation of Ab-peptide that leads to the
formation of amyloid plaques. One particularly compel-
ling approach to the treatment of AD would be the inhi-
bition of either of the two proteases involved in the
formation of Ab-peptide. Based on this hypothesis, a
chemistry programme was initiated to develop a selec-
tive inhibitor of c-secretase activity in order to reduce
Ab and delay the onset and progression of Alzheimer’s
disease.
In a previous publication from our laboratories we
reported the identification of a novel series of c-secre-
tase inhibitors.2 In the substituted gem-cyclohexane
series of c-secretase inhibitors, work had primarily
focused on the 4-position of the cyclohexane ring. Re-
cent chemistry has been directed towards substituted
analogues at the 3-position, for example, 1. This com-
munication describes the synthesis of compounds
which illustrate how the conformation of the substitu-
ents on the cyclohexane ring relates to in vitro c-secre-
tase activity.
R
F
Cl
S
O2
( )
1
F
The gem-substituted cyclohexanes were initially prepared
as outlined in Scheme 1. Reaction of d-valerolactone 2
with lithium bis(trimethylsilyl)amide and allylbromide
gave 3 in moderate yield. Facile reduction of the lactone
ring with lithium borohydride and subsequent activation
of the diol as the bis-mesylate 4 allowed double alkylation
of 132 to generate the gem-cyclohexane. The resulting
mixture of diastereoisomers cis-5 and trans-5 was
Keywords: gem-Cyclohexane; Sulfone; c-Secretase inhibitors; Alzhei-
mer’s disease; Conformational analysis; Biological activity.
*
Corresponding author. Tel.: +44 1279 600767; fax: +44 7970
075142; e-mail: richard_jelley@yahoo.co.uk
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.04.019

3840
R. A. Jelley et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3839–3842
OMs
gave the truncated alcohol, cis-6. Subsequent conversion
to the acid cis-7 was achieved using pyridinum dichro-
mate in DMF.
O
O
i
ii, iii OMs
O
O
( )
( )
2
3
4
By starting with the methyl analogue 10 (where a methyl
group was introduced at the 3-position of the cylohex-
ane ring after the initial allylation step) a second quater-
nary centre was generated. This ultimately gave
compound cis-12 after employing identical chemistry
to that described previously.
iv, v
F
F
Cl
Cl
( )
( )
S
S
+
O2
O2
An alternative route to 3-substituted gem-cyclohexanes
is illustrated in Scheme 2. Double addition of methyl
acrylate to 132 using sodium hydride and potassium
tert-butoxide as base gave the b-keto ester 14 in moder-
ate yield. The ketone was selectively reduced by conver-
sion to the enol mesylate and treatment with sodium
borohydride/nickel chloride to give a moderate yield of
the ester trans-15 as a 1:10 mixture of cis:trans diastere-
oisomers. This reaction was complicated by competing
reduction of the arylchloride moiety; however this by-
product was easily separated by chromatography. At-
tempts to epimerise the trans-ester to the cis-isomer
failed. However, by reducing the ester to the aldehyde
trans-16 and treating with potassium carbonate in meth-
anol, complete epimerisation to the cis-aldehyde cis-16
(where both the aryl sulfone and the aldehyde occupy
an equatorial position) was achieved.
F
F
trans-5
cis-5
vi
viii, ix
CO2Me
Cl
( )
OH
( )
F
F
Cl
S
S
O2
O2
F
F
cis-6
vii
x
cis-8
CO2H
( )
CO2Me
Cl
F
F
Cl
( )
S
S
O2
O2
F
F
cis-9
cis-7
Previous results with 4-substituted gem-cyclohexane
sulfones2 have demonstrated the importance of an axial-
ly disposed aryl and an equatorial sulfone for high
enzyme inhibition. The large sulfone group dictates the
O
O
xi
O
O
( )
3
10
ii-v
CO2Me
F
OH
O
F
F
Cl
Cl
( )
( )
O
F
S
S
vi-vii
O2
i
F
O2
O2S
Cl
SO2
( )
cis-12
11
F
F
Scheme 1. Reagents and conditions: (i) LHMDS, allyl bromide, THF,
À78 °C to rt, 35%; (ii) LiBH4, THF, 0 °C to rt, 71%; (iii) MsCl,
EtNiPr2, CH2Cl2, rt, 73%; (iv) 13, NaH, KOtBu, THF, rt; (v) separate
diastereoisomers, (1:1) 40%; (vi) methylacrylatetricyclohexylphos-
F
13
14
Cl
ii, iii
O
O
O
H
phine[1,3-bis(2,4,6-trimethyl
phenyl)-4,5-dihydroimidazol-2yli-
F
F
iv
dene][benzylidine]ruthenium(IV) dichloride, CH2Cl2, rt, 90%; (vii)
H2, 35 psi, Rh/C, EtOAc/EtOH, 70%; (viii) BH3ÆTHF, THF, 0 °C, 4 h;
(ix) NaOH, H2O2, 76%; (x) PDC, DMF, rt, 31%; (xi) NaH, MeI,
DMF, 75%.
Cl
SO2
Cl
SO2
( )-Trans
( )-Trans
F
F
trans-16
trans-15
v
O
obtained as a (1:1) mixture and separated using flash
chromatography. Each diastereoisomer was processed
separately using the same chemistry, however, only cis
structures are shown in Scheme 1. The terminal olefin
was functionalized using a cross-metathesis reaction with
methyl acrylate utilising Grubb’s second generation cat-
alyst.3 The unsaturated ester cis-8 was obtained. Hydro-
genation using rhodium on carbon gave the desired
saturated ester cis-9 without any cleavage of the aromatic
chlorine. Alternatively hydroboration of cis-5 using
borane–tetrahydrofuran complex at low temperature
H
F
Cl
SO2
( )-Cis
cis-16
F
Scheme 2. Reagents and conditions: (i) methylacrylate, NaH, KOtBu,
THF, rt, 45%; (ii) MsCl, EtNiPr2, CH2Cl2, rt; (iii) NaBH4,
NiCl2Æ6H2O, CH2Cl2/MeOH, 0 °C, 40% (1:10 cis:trans); (iv) DIBAl-
H, PhCH3, À78 °C, 77% (1:10 cis:trans); (v) K2CO3, MeOH, rt, 2 h,
100%.

R. A. Jelley et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3839–3842
3841
Cl
conformation; it adopts an equatorial position, even if
this results in an axial 4-substituent (Fig. 1). As a result,
both cis and trans diastereoisomers are potent c-secre-
tase inhibitors.
Cl
O2S
F
O2S
F
CO2Me
CO2Me
F
Analysis of the SAR (Table 1) shows a different pattern
for 3-substituted gem-cyclohexane sulfones.
F
cis-9
Cl
A
B
For most examples, the cis-isomer is more potent than
the corresponding trans-isomers. Analysis of H NMR
1
data for the cis isomers, such as cis-9 (Fig. 2),5 shows
unsurprisingly, that the molecules adopt a conformation
(A) in which both the sulfone and the 3-substituent are
equatorial. The alternate conformer (B) is extremely
unfavourable as it would place both large substituents
axial; unfavourable 1,3-diaxial interactions would fur-
ther destabilise this conformer. The preferred conforma-
tion therefore fits the pharmacophore model; the 3-
equatorial substituent is well tolerated and the molecules
are good inhibitors of c-secretase.
Cl
CO2Me
O2S
F
O2S
F
CO2Me
F
F
A
B
cis-12
Figure 2.
If the compound is di-substituted at the 3-position (cis-12)
1,3-diaxial interactions occur in both conformations A
and B (Fig. 2) but the energetic benefit of placing the large
sulfone group equatorial dominates and the active con-
formation A is again the only one present in solution.5
In the case of trans isomers, such as trans-9
(Fig. 3), 1H NMR analysis shows that the sulfone
no longer dominates the conformation;5 it is forced
into an axial position. This is in marked contrast
to the 4-substituted cyclohexanes and can be attrib-
uted to the added presence of unfavourable 1,3-diax-
ial interactions (A). With the sulfone axial (B), the
molecules no longer fit well into the pharmacophore
and therefore show relatively poor enzyme
inhibition.
Cl
Cl
R
O2S
O2S
F
F
There are some notable exceptions however. For smaller
3-substituents (e.g., allyl, trans-5), the trans-isomer
shows improved inhibition. In the case of the smallest
group (carboxaldehyde, trans-16) the difference in
potency between cis and trans actually reverses. Again,
this can be explained by analysis of 1H NMR data
R
F
F
A
B
cis
Figure 1.
Table 1. c-Secretase inhibition for 3-substituted sulfones
Cl
Cl
R1
R2
F
O2S
F
Cl
F
O2S
F
S
O2
( )
1
CO2Me
F
F
CO2Me
trans-9
A
B
4
IC50 (nM)
Entry
R1
R2
Figure 3.
cis-5
trans-5
cis-6
H
44
85
H
H
8
OH
Cl
trans-6
cis-7
H
112
24
Cl
CO2H
H
trans-7
cis-8
H
380
10
f
F
F
SO2
c
g
f
H
CO2Me
c
b
a
e
O2S
i
trans-8
cis-9
H
210
3
a
b
h
d
H
CO2Me
e
H
d
F
trans-9
cis-16
trans-16
cis-12
trans-12
H
376
77
O
H
O
CHO
H
F
H
CH3
55
10
A
B
CO2Me
1628
Figure 4.

3842
R. A. Jelley et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3839–3842
(Fig. 4);5 it is evident that both conformations A and B
are in rapid equilibrium as the observed NOE’s between
Hb–Hc and Hd–He are not both possible within a single
conformer. This reflects the decreasing importance of
unfavourable 1,3-diaxial interactions as the 3-substitu-
ent is reduced in size. With the equatorial sulfone con-
formation (A) present in solution, enzyme inhibition is
recovered.
References and notes
1. Hardy, J.; Selkoe, D. J. Science 2002, 356; Thorsett, E. D.;
Latimer, L. H. Curr. Opin. Chem. Biol. 2000, 377; Citron,
M. Mol. Med. Today 2000, 392; Vassar, R.; Citron, M.
Neuron 2000, 419.
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, 2685.
3. BouzBouz, S.; Cossy, J. Org. Lett. 2001, 1451.
In conclusion, we have further explored the effect of
conformation on enzyme inhibition for the cyclohexyl
sulfone c-secretase inhibitors and demonstrated that
substitution at the 3-position is well tolerated, provid-
ed that it does not destabilise the equatorial sulfone
conformation through 1,3-interactions. This was a
key discovery which allowed the further development
of this series to useful compounds for the treatment
of Alzheimer’s disease. This work will be reported
subsequently.
4. SH-SY5Y cells stably overexpressing the bAPP C-termi-
nal fragment SPA4CT (Dyrks, T.; Dyrks, E.; Monning,
U.; Urmoneit, B.; Turner, J.; Beyreuther, K. FEBS Lett.
1993 335, 89) are induced with sodium butyrate prior to
plating. Compounds are added at a range of concentra-
tions after 2 h and incubated overnight. Aliquots of
conditioned media are removed for analysis by a homo-
geneous time resolved fluoresence (HTRF) assay (Clarke,
E. E; Shearman, M. S. J. Neurosci. Methods 2000, 61).
All IC50 data shown are geometric means of a minimum
of three independent results. Cell viability is measured by
a colorimetric cell proliferation assay CellTitre 96TM AQ
assay, Promega.
5. Samples were dissolved in chloroform-d (100% D). All
experiments were acquired on a Bruker AV 500 MHz NMR
instrument equipped with a 3 mm triple resonance probe
(1H/19F/13C) with Z gradients. A temperature of 300 K was
maintained throughout the data acquisition.
Acknowledgment
The authors thank Steve Thomas for his support in
obtaining spectral data.