Novel orally bioavailable γ-secretase inhibitors with excellent in vivo activity

Article abstract of DOI:10.1021/jm900056p

The development of potent γ-secretase inhibitors having substituted heterocycles attached to a benzobicyclo[4.2.1]nonane core is described. This work led to the identification of [6S,9R,11R]-2′,3′,4′,5, 5′,6,7,8,9,10-decahydro-2-(5-(4-fluorophenyl)-1-methylpyrazol-3-yl) -5′-(2,2,2-trifluoroethyl)spiro[6,9-methanobenzocyclooctene-11, 3′-[1,2,5]thiadiazole] 1′,1′-dioxide (16), which has excellent in vitro potency (0.06 nM) and which reduced amyloid-β in APP-YAC mice with an ED50 of 1 mg/kg (po). 16 had a good pharmacokinetic profile in three preclinical species.

Full text of DOI:10.1021/jm900056p

J. Med. Chem. 2009, 52, 3441–3444
3441
Letters
This communication describes a novel series of highly potent
γ-secretase inhibitors that show excellent in vivo Aꢀ-lowering
efficacy.
Novel Orally Bioavailable γ-Secretase
Inhibitors with Excellent in Vivo Activity
The development, in these laboratories, of a series of
γ-secretase inhibitors having a bicyclo[4.2.1]nonane core has
been described.5,6 The most potent compound to be disclosed
was the trifluoroethyl-substituted cyclic sulfamide 1 which
inhibited γ-secretase in a SHSY5Y cell assay7 with IC50 ) 0.24
nM. In APP-YAC mice,8 a transgenic line that overexpresses
human ꢀAPP, 1 had an ED50 of 17 mg/kg at 4 h for reduction
of brain Aꢀ(40).
Linda E. Keown,*,† Ian Collins,† Laura C. Cooper,†
Timothy Harrison,† Andrew Madin,† Jayesh Mistry,†
Michael Reilly,† Mohamed Shaimi,‡ Christopher J. Welch,‡
Earl E. Clarke,§ Huw D. Lewis,§ Jonathan D. J. Wrigley,§
Jonathan D. Best,| Fraser Murray,| and Mark S. Shearman§
Departments of Medicinal Chemistry, Molecular and Cellular
Neuroscience, and in ViVo Neuroscience, The Neuroscience
Research Centre, Merck Sharp & Dohme Research Laboratories,
Terlings Park, Eastwick Road, Harlow, Essex, CM20 2QR, U.K.,
and Department of Process Research, Merck Research Laboratories,
126 East Lincoln AVenue, Rahway, New Jersey 07065
ReceiVed January 15, 2009
As part of a program to optimize the in vivo activity of 1,
compounds were targeted in which the allylic linker between
the bicyclic core and the terminal cyclic amine was replaced
by a heterocycle. In this communication, the synthesis and
biological activity of γ-secretase inhibitors containing thiazole,
triazole, imidazole and pyrazole linking groups will be described.
Abstract: The development of potent γ-secretase inhibitors having
substituted heterocycles attached to a benzobicyclo[4.2.1]nonane core
is described. This work led to the identification of [6S,9R,11R]-
2′,3′,4′,5,5′,6,7,8,9,10-decahydro-2-(5-(4-fluorophenyl)-1-methylpyra-
zol-3-yl)-5′-(2,2,2-trifluoroethyl)spiro[6,9-methanobenzocyclooctene-
11,3′-[1,2,5]thiadiazole] 1′,1′-dioxide (16), which has excellent in vitro
potency (0.06 nM) and which reduced amyloid-ꢀ in APP-YAC mice
with an ED50 of 1 mg/kg (po). 16 had a good pharmacokinetic profile
in three preclinical species.
Scheme 1. Synthesis of Aminothiazole 6a
Alzheimer’s disease (ADa) is a devastating neurodegenerative
disorder and is the most common cause of senile dementia in
the elderly. The disease is characterized by progressive cognitive
and memory deterioration that leads to difficulty in carrying
out everyday activities. As life expectancy in the developed
world continues to increase, the number of cases of AD is
predicted to rise significantly. Currently, the only treatments
approved for AD are palliative, and chief among these are
cholinesterase inhibitors.1 A large research effort is focused on
the discovery and development of therapies that could potentially
halt, or even reverse, the progression of AD.
The main pathological features of AD are neurofibrillary
tangles and neuritic plaques. The plaques are composed of 40-
to 42-amino acid peptides known as amyloid-ꢀ (Aꢀ) which are
generated from ꢀ-amyloid precursor protein (ꢀAPP) by sequen-
tial cleavage by the enzymes ꢀ- and γ-secretase. It has been
proposed that accumulation of Aꢀ is the trigger for progression
of AD and therefore, inhibiting production of this protein or
increasing its clearance could have therapeutic value.2 This
hypothesis has prompted the search for compounds that inhibit
ꢀ-secretase3 or γ-secretase.4
a Reagents: (a) Tf2O, pyridine, CH2Cl2, 0 °C to room temp; (b) CO,
Pd(OAc)2, dppp, Et3N, MeOH, DMSO, 80 °C; (c) 4 M NaOH, THF,
60 °C; (d) MeO2CCH2NH2.HCl, HBTU, Pr2NEt, MeCN, room temp to
* To whom correspondence should be addressed. E-mail: keown.linda@
yahoo.co.uk.
i
†
40 °C; (e) 4 M NaOH, THF, 50 °C; (f) 4-trifluoromethylpiperidine, HBTU,
Department of Medicinal Chemistry, Merck Sharp & Dohme Research
iPr2NEt, MeCN, 40 °C; (g) Lawesson’s reagent, toluene, reflux.
Laboratories.
‡
Merck Research Laboratories.
Department of Molecular and Cellular Neuroscience, Merck Sharp &
§
The synthesis of the 2,5-substituted thiazole 6 is outlined in
Scheme 1. The previously described phenol 26 was obtained in
homochiral form by supercritical fluid chromatography of the
benzyl ether derivative.9 The phenol was converted to the triflate
which was then carbonylated and hydrolyzed to give the
Dohme Research Laboratories.
|
Department of In Vivo Neuroscience, Merck Sharp & Dohme Research
Laboratories.
a Abbreviations: AD, Alzheimer’s disease; Aꢀ, amyloid-ꢀ; ꢀAPP,
ꢀ-amyloid precursor protein; APP-YAC, amyloid precursor protein-yeast
artificial chromosome.
10.1021/jm900056p CCC: $40.75
2009 American Chemical Society
Published on Web 05/11/2009

3442 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 11
Scheme 2. Synthesis of Glycolatoboronate 8a
Letters
Scheme 4. Synthesis of Imidazole 12a
a Reagents: (a) [Cu(OH)·TMEDA]2Cl2, 4-fluorophenylboronic acid,
CH2Cl2, O2, 46%; (b) boronate 8, Pd(PPh3)4, Cs2CO3, DME, H2O, reflux,
88%.
a Reagents: (a) bis(neopentyl glycolato)diboron, [1,1′-bis(diphenyl-
phosphino)ferrocene]dichloropalladium(II) complex with CH2Cl2, KOAc,
1,4-dioxane, DMSO, reflux, 88%.
Scheme 5. Synthesis of Pyrazole Regioisomers 15 and 16a
Scheme 3. Synthesis of 1,2,4-Triazole 10a
a Reagents: (a) NaNO2, HBr, H2O, 0 °C to room temp, 44%; (b) (4-F-
Ph)3Bi(OAc)2, Me2NC(dNH)NMe2, Cu(OAc)2, THF, 50 °C, air, 45%; (c)
boronate 8, Pd(PPh3)4, Cs2CO3, 1,2-dimethoxyethane (DME), H2O, reflux,
73%.
carboxylic acid 4. The absolute stereochemistry of the required
enantiomer was determined by X-ray crystallographic analysis
of the salt formed with (1R,2S)-ephedrine and is as shown in
Scheme 1, namely (6S,9R,11R). Coupling of homochiral car-
boxylic acid with glycine methyl ester was followed by
hydrolysis and a further amide coupling to give the acyclic
precursor 5; thiazole formation could then be achieved by
heating with Lawesson’s reagent.
Thiazole-containing compounds were synthesized having a
cyclic amine or a 4-fluorophenyl ring as the terminal substituent.
The 4-fluorophenyl thiazole (7) was synthesized in a manner
similar to that outlined in Scheme 1, with the acyclic precursor
being the product of coupling between homochiral acid 4 and
2-amino-4′-fluoroacetophenone hydrochloride.
a Reagents: (a) (i) Bu3SnC(OEt)dCH2, Pd(OAc)2, LiCl, PPh3, DMF,
100 °C, (ii) 1 M HCl(aq.), 84%; (b) LHMDS, 4-fluorobenzaldehyde, THF,
-78 °C, 86%; (c) CeCl3 ·7H2O, NaI, MeCN, reflux, 69%; (d) MeNHNH2,
EtOH; (e) DDQ, THF, 86% over two steps.
The syntheses outlined in Schemes 3 and 4 were used to
generate a number of analogues and were found to be amenable
to scale-up.
The application of cross-coupling strategies proved vital for
rapid exploration of heterocycle structure-activity relationships,
and a key synthetic intermediate was the glycolatoboronate 8
derived from triflate 3 as shown in Scheme 2.
The fourth heterocycle to be highlighted is the methyl-3,5-
diarylpyrazole. This was initially synthesized as a 1:1 mixture
of regioisomers by condensation of methylhydrazine with the
appropriate racemic 1,3-diketone precursor (not shown). The
ratio of isomers could be improved to 5:1 in favor of regioisomer
16 by condensation of methylhydrazine with enone 14, available
in homochiral form, followed by oxidation of the crude product
mixture (Scheme 5).
The synthesis of the 1,3-diaryl 1,2,4-triazole 10 via a Suzuki
coupling is shown in Scheme 3. The preparation of the bromide
coupling partner utilizes methodology developed by Finet et
al.10 for the N-phenylation of azoles by triphenylbismuth
derivatives in the presence of catalytic copper diacetate.
3-Bromo-1,2,4-triazole was synthesized by diazotization and
subsequent bromination of 3-amino-1,2,4-triazole. Reaction with
bis(acetato)tris(4-fluorophenyl)bismuth11 under the published
conditions gave bromotriazole 9 which was coupled with the
homochiral boronate 8 to give the diaryltriazole 10.
Methylpyrazole regioisomer 16 could be synthesized selec-
tively (Scheme 6) via a Suzuki coupling of boronate 8. In this
case the required coupling partner was 1-methyl-3(5)-(4-
fluorophenyl)pyrazole nonaflate which was synthesized regi-
oselectively using methodology described in a publication from
these laboratories.13 Suzuki couplings were utilized to give
multigram quantities of a number of halogenated phenylpyrazoles.
The final step in the synthesis of imidazole 12 (Scheme 4)
was again a Suzuki coupling, and, in this case, the required
aryl heterocycle was prepared using the copper-catalyzed
N-arylation protocol developed by Collman.12

Letters
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 11 3443
Scheme 6. Optimized Synthesis of Methylpyrazole 16a
Table 2. Inhibitory Activities of Various Aryl-Substituted Imidazoles
and Pyrazolesa
a Reagents: (a) Pd(PPh3)4, Na2CO3, DMF, 100 °C, 100%.
Table 1. Inhibitory Activities of Heterocycles
IC50, nM (n)
imidazoles
Ar
pyrazoles
2-fluorophenyl
3-fluorophenyl
4-fluorophenyl
2,4-difluorophenyl
3,4-difluorophenyl
4-chlorophenyl
3,4-dichlorophenyl
0.17 (2)
0.82 (2)
0.03 (2)
0.06 (2)
0.12 (3)
0.08 (4)
0.42 (4)
0.21b (2)
0.46b (2)
0.06 (7)
0.06 (2)
0.06 (3)
0.07 (2)
0.41 (2)
a All compounds are homochiral unless otherwise noted. b Given IC50
values are for racemic compounds.
Figure 1. Reduction of brain Aꢀ(40) 4 h after oral dosing of imidazole
12 and pyrazole 16 to APP-YAC mice.
generally well tolerated in vitro. The 1,3-disubstituted 1,2,4-
triazole 10 was found to be a potent γ-secretase inhibitor (IC50
) 0.15 nM) while the N-substituted imidazole 12 was found to
be more potent still (IC50 ) 0.03 nM). The N-methylpyrazole
regioisomers 15 and 16 inhibited the γ-secretase enzyme to
different extents with the more potent isomer 16 (IC50 ) 0.06
nM) having in vitro activity comparable to that of the imidazole.
On the basis of these results, aryl-substituted imidazoles and
pyrazoles were synthesized and representative examples are
shown in Table 2. It can be seen that 4-substitution is generally
preferred and that several disubstitution patterns are also well
tolerated. Further in vitro experiments showed that pyrazole 16,
a representative example of this class of compounds, inhibited
production of Aꢀ(40) and Aꢀ(42) with similar IC50 values and
also showed no selectivity over Notch cleavage.14
Two of the most potent compounds in vitro, the 4-fluorophe-
nyl-substituted imidazole 12 and pyrazole 16, were evaluated
in vivo. The compounds were dosed orally, as suspensions in
0.5% methocel, at 3 mg/kg to APP-YAC mice8 and, 4 h after
dosing, the level of soluble Aꢀ(40) in the forebrain was
measured. As shown in Figure 1, the imidazole 12 caused a
66% reduction and the pyrazole 16 an 84% reduction in brain
Aꢀ(40) relative to vehicle-treated animals.
a All compounds are homochiral unless otherwise noted.
All of the heterocycle-linked compounds described above
were found to be potent inhibitors of γ-secretase in vitro (Table
1) with subnanomolar IC50 values.
The thiazole analogue 6 of the allylic amine 1 retained in
vitro potency (IC50 ) 0.34 nM), and further changing the
structure to replace the terminal piperidine by a substituted
phenyl moiety (e.g. 4-fluorophenyl, 7) was also tolerated. It was
observed that, in general, the aryl-substituted thiazoles were less
active than the aminothiazoles in ion channel counterscreens
and therefore, in subsequent investigation of heterocyclic linkers,
compounds having a terminal aromatic ring were targeted. For
ease of comparison, only the 4-fluorophenyl-substituted het-
erocycles are shown in Table 1; this substitution pattern was
Compound 16 was also dosed orally at 0.3 and 1 mg/kg to
APP-YAC mice, and a dose-dependent reduction in Aꢀ(40) was
observed (Figure 1) with an ED50 of approximately 1 mg/kg.

3444 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 11
Letters
intervention in Alzheimer’s disease. Curr. Opin. Drug DiscoVery DeV.
2004, 7, 709–719.
(5) Lewis, S. J.; Smith, A. L.; Neduvelil, J. G.; Stevenson, G. I.; Lindon,
M. J.; Jones, A. B.; Shearman, M. S.; Beher, D.; Clarke, E.; Best,
J. D.; Peachey, J. E.; Harrison, T.; Castro, J. L. A novel series of
potent γ-secretase inhibitors based on a benzobicyclo[4.2.1]nonane
core. Bioorg. Med. Chem. Lett. 2005, 15, 373–378.
(6) Sparey, T.; Beher, D.; Best, J.; Biba, M.; Castro, J. L.; Clarke, E.;
Hannam, J.; Harrison, T.; Lewis, H.; Madin, A.; Shearman, M.; Sohal,
B.; Tsou, N.; Welch, C.; Wrigley, J. Cyclic sulfamide γ-secretase
inhibitors. Bioorg. Med. Chem. Lett. 2005, 15, 4212–4216.
(7) Clarke, E. E.; Shearman, M. S. Quantitation of amyloid-ꢀ peptides in
biological milieu using a novel homogeneous time-resolved fluores-
cence (HTRF) assay. J. Neurosci. Methods 2000, 102, 61–68.
(8) 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. Introduction
and expression of the 400 kilobase amyloid precursor protein gene in
transgenic mice. Nat. Genet. 1993, 5, 22–30.
(9) For chromatography conditions, see Supporting Information.
(10) Fedorov, A. Y.; Finet, J.-P. N-Phenylation of azole derivatives by
triphenylbismuth derivatives/cupric acetate. Tetrahedron Lett. 1999,
40, 2747–2748.
(11) Combes, S.; Finet, J.-P. A convenient synthesis of triarylbismuth
diacetates. Synth. Commun. 1996, 26, 4569–4575.
(12) Collman, J. P.; Zhong, M. An efficient diamine ·copper complex-
catalysed coupling of arylboronic acids with imidazoles. Org. Lett.
2000, 2, 1233–1236.
Pharmacokinetic evaluation of 16 in rat (F ) 35%, t1/2
2 h), dog (F ) 42%, t1/2 > 10 h), and rhesus (F ) 45%, t1/2
16 h) demonstrated that the compound had a good pharmaco-
kinetic profile across the range of preclinical species. Further
in vivo studies will be reported elsewhere.
In conclusion, a number of heterocycle-containing com-
pounds have been shown to be potent inhibitors of γ-secretase
in vitro. Furthermore, the imidazole 12 and pyrazole 16 were
highly efficacious in APP-YAC mice. The pyrazole 16, with
an ED50 of approximately 1 mg/kg for reduction of Aꢀ(40)
in this model, is one of the most potent γ-secretase inhibitors
reported to date.
)
)
Acknowledgment. The authors thank Dirk Beher, Dawn
Bloomfield, Suzanne Wood, and Michael Reader for support,
and Ben Munoz for critical reading of the manuscript.
Supporting Information Available: Experimental details and
spectral data for all new compounds. This material is available free
of charge via the Internet at http://pubs.acs.org.
References
(13) Bourrain, S.; Ridgill, M.; Collins, I. Regioselective rapid analogue
syntheses of 1-methyl-3,5-diarylpyrazoles via palladium-catalysed
coupling to 3(5)-pyrazolyl nonaflates. Synlett 2004, 5, 795–798.
(14) Lewis, H. D.; Perez Revuelta, B. I.; Nadin, A.; Neduvelil, J. G.;
Harrison, T.; Pollack, S. J.; Shearman, M. S. Catalytic site-directed
γ-secretase complex inhibitors do not discriminate pharmacologically
between Notch S3 and ꢀ-APP cleavages. Biochemistry 2003, 42,
7580–7586.
(1) Cummings, J. L. Alzheimer’s disease. N. Engl. J. Med. 2004, 351,
56–67.
(2) Hardy, J.; Selkoe, D. J. The amyloid hypothesis of Alzheimer’s disease:
progress and problems on the road to therapeutics. Science 2002, 297,
353–356.
(3) Ghosh, A. K.; Kumaragurubaran, N.; Hong, L.; Koelsh, G.; Tang, J.
Memapsin 2 (ꢀ-secretase) inhibitors: drug development. Curr. Alzhe-
imer Res. 2008 Apr, 5 (2), 121–131.
(4) Harrison, T.; Churcher, I.; Beher, D. γ-Secretase as a target for drug
JM900056P