A new series of potent benzodiazepine gamma-secretase inhibitors.

Article abstract of DOI:10.1016/S0960-894X(02)00909-5

A new series of benzodiazepine-containing gamma-secretase inhibitors with potential use in the treatment of Alzheimer's disease is disclosed. Structure-activity relationships of the pendant hydrocinnamate side-chain which led to the preparation of highly

Full text of DOI:10.1016/S0960-894X(02)00909-5

Bioorganic & Medicinal Chemistry Letters 13 (2003) 179–183
A New Series of Potent Benzodiazepine ꢀ-Secretase Inhibitors
Ian Churcher,a,* Kate Ashton,a John W. Butcher,b Earl E. Clarke,c Timothy Harrison,a
Huw D. Lewis,c Andrew P. Owens,a Martin R. Teall,a Susie Williamsa and
Jonathan D. J. Wrigleyc
aDepartment of Medicinal Chemistry, The Neuroscience Research Centre, Merck, Sharp and Dohme Research Laboratories,
Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
bDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 4, Sumneytown Pike, West Point, PA 19486, USA
cDepartment of Biochemistry and Molecular Biology, The Neuroscience Research Centre,
Merck, Sharp and Dohme Research Laboratories, Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
Received 28 August 2002; revised 4 October 2002; accepted 22 October 2002
Abstract—A new series of benzodiazepine-containing g-secretase inhibitors with potential use in the treatment of Alzheimer’s dis-
ease is disclosed. Structure–activity relationships of the pendant hydrocinnamate side-chain which led to the preparation of highly
potent inhibitors are described.
# 2002 Elsevier Science Ltd. All rights reserved.
The 40–42 amino acid peptide amyloid-b (Ab) is the
major component of the extracellular proteinaceous
plaques seen in Alzheimer’s disease (AD) and much
evidence suggests a pivotal role for Ab in the disease
process.1 In particular, individuals possessing auto-
somal dominant mutations in the genes encoding for
amyloid-b precursor protein (bAPP) or the membrane-
bound protein homologues presenilin 1 and 2 have ele-
vated Ab levels and suffer from aggressive forms of
early onset AD.2 These observations have led to the
hypothesis that Ab, either in its soluble form or when
aggregated into fibrils and subsequently plaques, is
responsible for neuronal toxicity and cell death.3,4
Inhibition of Ab synthesis is thus an attractive target for
therapeutic intervention in AD.
A whole-cell g-secretase inhibition assay using SHSY5Y
cells in which human g-secretase catalyzes the break-
down of the overexpressed exogenous substrate
hspbA4CTF has been developed.7 Using this assay,
screening of the Merck sample collection identified 1
which inhibited the secretion of Ab with an IC50=33 nM.
Ab is derived by processing of the 695–770 residue, type
I transmembrane protein bAPP.5 The major metabolic
pathway of bAPP involves cleavage by the proteases
a-secretase and g-secretase leading to non-amyloido-
genic fragments.6 Alternative processing by stepwise
cleavage mediated by b-secretase and g-secretase leads
to the production of Ab and it is inhibitors of the latter
enzyme that were targeted in the current work.
Initial studies directed towards the optimization of 1
involved changes to the aromatic substitution pattern in
the hydrocinnamoyl C-3 side-chain. A synthesis of the
benzodiazepine which allows introduction of the side-
chain at a late stage is outlined in Scheme 1.8
Dimethyl anthranilate was transformed into benzodia-
zepinedione 29 which in turn allowed the introduction
of the C-5 benzamide substituent in protected form by
way of addition of the appropriate Grignard reagent.10
Subsequent deprotection and cyclization gave benzo-
diazepine 3. The C-3 amino substituent was introduced
*Corresponding author. Tel.:+44-1279-440000; fax:+44-1279-440390;
e-mail: ian_churcher@merck.com
0960-894X/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00909-5

180
I. Churcher et al. / Bioorg. Med. Chem. Lett. 13 (2003) 179–183
Table 1. Structure–activity relationships of hydrocinnamate-based
C-3 substituents
Entry
R
IC50,
nM (n)
Entry
R
IC50,
nM (n)
6
340 (3)
17
18
19
20
21
22
23
24
25
26
27
29 (6)
7
305 (3)
93 (4)
5100 (3)
1600 (3)
8
9
15 (3)
1340 (3)
Scheme 1. (i) BrCH2COBr, DCM; (ii) aq NaOH, 0 ꢀC; (iii) NH3,
MeOH; (iv) Boc2O, NaH, THF; (v) 2-(4-bromophenyl)-4,4-dimethyl-
4,5-dihydro-oxazole, Mg, I2, THF; (vi) HCl, EtOAc then aq NaHCO3;
10
11
12
13
14
15
16
22 (3)
6800 (3)
i
(vii) KOtBu, AmONO, PhMe; (viii) EtNCO, Et3N, THF; (ix) H2/Pd/
C, MeOH; (x) benzylchloroformate, Na2CO3, aq dioxan; (xi) 1 N HCl/
dioxan; (xii) AcCl, Et3N, DCM; (xiii) NaOH, aq THF; (xiv) NH4Cl,
EDC, HOBt, Et3N, DMF; (xv) HBr/AcOH; (xvi) RCO2H, EDC,
HOBt, Et3N, DCM.
12 (6)
>10,000 (3)
>10,000 (3)
400 (3)
3200 (3)
40 (3)
by nitrosylation and hydrogenation of the derived car-
bamate11 which, after protection gave racemic amino-
benzodiazepine 4. Unmasking of the carboxamide
followed by benzyl carbamate cleavage gave the versa-
tile, racemic amine 5 which was coupled with a range of
carboxylic acids to give the amides 6–27. The required
hydrocinnamic acids were available from rhodium-cat-
alyzed hydrogenation of the corresponding cinnamic
acids or via alkylation of the anion of ethyl acetate fol-
lowed by hydrolysis.
60 (3)
180 (3)
110 (3)
110 (3)
>10,000 (3)
5000 (3)
The structure–activity relationships of a variety of C-3
amide substituents are summarized in Table 1. In gen-
eral, more polar functionality was poorly tolerated
(compounds 18–23) relative to unsubstituted hydro-
cinnamate 6. Of the more lipophilic substituents, halo-
gens in particular gave the most potent compounds.
Investigation of representative halogen substituent pat-
terns showed 2,4- and 3,4-dihalo (e.g., 11, 17) to be
amongst the best and these were selected for further
elaboration. Variation of chain length of the linking
group was briefly investigated but both truncation
(1!26) and homologation (6!27) were attended with
large decreases in potency.
the geminally substituted analogue 30 was poorly toler-
ated. The four stereoisomers of 29 were prepared indi-
vidually by reaction of a single enantiomer of the
requisite a-methyl hydrocinnamate13 with racemic ben-
zodiazepine 5 followed by separation of the resultant
diastereomers. This demonstrated that it was the (S)-
configured methyl hydrocinnamate that was preferred
and also that the activity resided mainly in one stereo-
isomer at the benzodiazepine C-3 centre. Subsequent
work (vide infra, compounds 34 and 35) suggested that
it was the (S)-configured benzodiazepine that was pre-
ferred but at this stage this was not proven.
Concurrently, investigations into reducing the flexibility
of the side-chain were carried out (Table 2).12 Whilst use
of the cinnamate side-chain in 28 resulted in a large
reduction in potency, introduction of a methyl group at
the a-position of the hydrocinnamate, as in 29, gave an
increase. Introduction of a further a-substituent as in
The use of an (S)-2-methyl-(2,4-dichloro) hydro-
cinnamate side-chain had thus shown that substituted

I. Churcher et al. / Bioorg. Med. Chem. Lett. 13 (2003) 179–183
Table 2. Introduction of conformational restrictions into hydro-
cinnamate side-chain
181
IC50, nM (n)
4700 (3)
28
Scheme 2. (i) RCO2H, EDC, HOBt, DCM.
Table 3. Determination of preferred stereochemistry using C-5
phenyl benzodiazepine
IC50, nM (n)
29
8 (3)
34
>10,000 (2)
30
400 (8)
820 (3)
320 (6)
220 (9)
relationships of the aromatic substituents on the hydro-
cinnamate moiety mirrored that seen previously, with
halogen substituents again giving the best results (35–
37; cf. Table 1).
31
diastereomer A
diastereomer B
42 (3)
920 (4)
In order to recover potency back to the levels seen with
the C-5 benzamide substituent, a variety of a-sub-
stituted hydrocinnamates were screened at the C-3
position and the activity of representative compounds is
summarized in Table 4.
The necessary carboxylic acids were prepared in racemic
form via alkylation of the relevant substituted acetate
ethyl ester by 3,4-dichlorobenzyl bromide followed by
saponification. Amide bond formation with (S)-33 gave
a pair of diastereomers which usually could be sepa-
rated chromatographically. In cases where the sub-
stituted cinnamate was prepared as a defined, single
enantiomer, alkylation was carried out utilizing the
Evans benzyl oxazolidinone13 as auxilliary or else com-
mercially available substituted phenylalanines were
employed and subsequently modified.
32
diastereomer A
diastereomer B
500 (3)
1.9 (4)
benzodiazepines could function as highly potent
g-secretase inhibitors.
In addition to the above work, investigation into the
effect of carboxamide deletion was undertaken utilizing
the readily available C-5 phenyl benzodiazepine 3314
allowing the synthesis of amides 34–52 as shown in
Scheme 2. Immediately evident was that removal of the
carboxamide moiety resulted in a substantial reduction
in inhibitory activity (Table 3). Further, the availability
of both enantiomers15 of 33 showed the C-3 (S)
stereoisomer to be preferred. The structure–activity
A range of compounds were found to possess potency
similar to or better than methyl analogue 36. Basic cen-
tres were well tolerated (40, 41, 47) as were other, lipo-
philic substituents. In particular, aromatic and
heteroaromatic groups (e.g., 43–44) gave improved
potency. In general, both stereoisomers of the side-chain
showed similar levels of inhibition and the marked ste-
reochemical preference noted with the isomers of 31 and
32 (Table 2) was not usually observed.

182
I. Churcher et al. / Bioorg. Med. Chem. Lett. 13 (2003) 179–183
Table 4. Structure–activity relationships of a-substituted hydrocinnamates
Entry
R
IC50, nM (n)
Entry
46
R
IC50, nM (n)
38
440 (3)
(diast. A)
1360 (3)
(diast. A)
39
40
41
42
240 (3)
(diast. B)
47
48
49
50
190 (3)
(diast. B)
405 (4)
350 (4)
750 (3)
72 (3)
180 (3)
100 (3)
43
37 (3)
(1:1 diast mix)
51
52
15 (4)
44
45
100 (3)
(1:1 diast mix)
3.8 (4)
640 (3)
(1:1 diast mix)
The optimal a-substituent was found to be an (R)-con-
figured 4-fluorophenyl (51, 15 nM) and substitution of
the chlorine atoms for fluorine (52) gave a further
increase in potency to a level which now matched that
seen with the carboxamide-substituted benzodiazepines.
Acknowledgements
The authors wish to thank Drs. M. S. Shearman and D.
Beher for their valuable contributions and Beth Oxley
for assistance in screening.
In summary, we have developed a new series of benzo-
diazepine g-secretase inhibitors containing a substituted
hydrocinnamate C-3 substituent. High potency could be
attained by use of a benzodiazepine bearing a C-5
carboxamide group. A further series of compounds
utilizing a C-5 phenyl-substituted benzodiazepine was
developed in which the potency lost by omission of the
carboxamide moiety was successfully recovered by ela-
boration of the hydrocinnamate side-chain. In particular,
the (R)-3-(3,4-difluoro)-2-(4-fluorophenyl)-propionyl sub-
stituent allowed the preparation of a further series
of highly potent, benzodiazepine-derived g-secretase
inhibitors.
References and Notes
1. For recent reviews on the role of b-amyloid in Alzhei-
mer’s disease see: Coughlan, C. M.; Breen, K. C. Pharmacol.
Ther. 2000, 86, 111. Jhee, S.; Shiovitz, T.; Crawford, A. W.;
Cutler, N. R. Exp. Opin. Invest. Drugs 2001, 10, 593. Samba-
murti, K.; Grieg, N. H.; Lahiri, D. K. NeuroMolec. Med.
2002, 1, 1.
2. Reviews on the presenilins see: Sisodia, S. S.; St George-
Hyslop, P. H. Nat. Rev. Neurosci. 2002, 3, 281. Steiner, H.;
Haass, C. Nat. Rev. Molec. Cell Biol. 2000, 1, 217. Wolfe,
M. S. J. Neurochem. 2001, 76, 1615.
3. Selkoe, D. J. Physiol. Rev. 2001, 81, 741.

I. Churcher et al. / Bioorg. Med. Chem. Lett. 13 (2003) 179–183
11. Bourrain, S.; Showell, G. A. Synthesis 1994, 505.
183
4. Selkoe, D. J. Nature 1999, 99, A23.
5. Selkoe, D. J.; Podlisny, M. B.; Joachim, C. L.; Vickers,
E. A.; Lee, G.; Fritz, L. C.; Oltersdorf, T. Proc. Natl. Acad.
Sci. U.S.A. 1988, 85, 7341.
6. For a review of the secretases see: Wolfe, M. S. J. Med.
Chem. 2001, 44, 2039. For recent speculation as to the identity
of g-secretase see: Tsai, J.-Y.; Wolfe, M. S.; Xia, W. Curr.
Med. Chem. 2002, 9, 1087.
7. Clarke, E. E.; Shearman, M. S. J. Neurosci. Meth. 2000,
102, 61.
8. Churcher, I.; Nadin, A. J.; Owens, A. P., PCT Int. Appl.
2002, WO 0230912.
12. Castro Pineiro, J. L.; Churcher, I.; Guiblin, A. R.; Harri-
son, T.; Kerrad, S.; Madin, A.; Nadin, A. J.; Owens, A. P.;
Sparey, T. J.; Teall, M. R.; Williams, S. PCT Int. Appl. 2001,
WO 0190084.
13. The required (R) and (S)-3-(2,4-dichlorophenyl)-2-
methylpropionic acids were prepared by the alkylation of (S)-
(+)-4-benzyl-3-propionyl-2-oxazolidinone and (R)-(ꢁ)-4-benzyl-
3-propionyl-2-oxazolidinone respectively followed by LiOH/
H2O2 hydrolysis cf. Evans, D. A.; Ennis, M. D.; Mathre, D. J.
J. Am. Chem. Soc. 1982, 104, 1737.
14. Bock, M. G.; DiPardo, R. M.; Evans, B. E.; Rittle, K. E.;
Veber, D. F.; Freidinger, R. M.; Hirshfield, J.; Springer, J. P.
J. Org. Chem. 1987, 52, 3232.
9. Siegl, P. K. S.; Goldberg, A. I.; Goldberg, M. R.; Chang, P.
I. PCT Int. Appl. 1997, WO 9749690.
10. Meyers, A. I.; Temple, D. L.; Haidukewych, D.; Mihelich,
E. D. J. Org. Chem. 1974, 39, 2787.
15. Reider, P. J.; Davis, P.; Hughes, D. L.; Grabowski, E. J. J.
J. Org. Chem. 1987, 55, 955.