Aminothiazoles as γ-secretase modulators

Article abstract of DOI:10.1016/j.bmcl.2011.08.060

We herein report the discovery of a new γ-secretase modulator class with an aminothiazole core starting from a HTS hit (3). Synthesis and SAR of this series are discussed. These novel compounds demonstrate moderate to good in vitro potency in inhibiting amyloid beta (Aβ) peptide production. Overall c-secretase is not inhibited but the formation of the aggregating, toxic Aβ42 peptide is shifted to smaller non-aggregating Ab peptides. Compound 15 reduced brain Abβ42 in vivo in APPSwe transgenic mice at 30 mg/kg p.o.

Full text of DOI:10.1016/j.bmcl.2011.08.060

Bioorganic & Medicinal Chemistry Letters 21 (2011) 6554–6558
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Aminothiazoles as c-secretase modulators
Thomas Lübbers a, , Alexander Flohr a, Synese Jolidon a, Pascale David-Pierson c, Helmut Jacobsen b,
⇑
Laurence Ozmen b, Karlheinz Baumann b
a Discovery Chemistry, F. Hoffmann-La Roche Ltd, Grenzacher Strasse 124, 4070 Basel, Switzerland
b CNS Discovery, F. Hoffmann-La Roche Ltd, Grenzacher Strasse 124, 4070 Basel, Switzerland
c Non-Clinical Safety, F. Hoffmann-La Roche Ltd, Grenzacher Strasse 124, 4070 Basel, Switzerland
a r t i c l e i n f o
a b s t r a c t
Article history:
We herein report the discovery of a new c-secretase modulator class with an aminothiazole core starting
from a HTS hit (3). Synthesis and SAR of this series are discussed. These novel compounds demonstrate
Received 22 June 2011
Revised 10 August 2011
Accepted 11 August 2011
Available online 19 August 2011
moderate to good in vitro potency in inhibiting amyloid beta (Ab) peptide production. Overall -secretase
c
is not inhibited but the formation of the aggregating, toxic Ab42 peptide is shifted to smaller non-aggre-
gating Ab peptides. Compound 15 reduced brain Ab42 in vivo in APPSwe transgenic mice at 30 mg/kg p.o.
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Alzheimer‘s disease
c
-Secretase modulator
Amyloid-b peptide
Ab42 lowering
APPSwe mice model
Alzheimer’s disease (AD) is the most common dementia with
more than 5 million patients alone in the US and 24 million world-
wide.1 It is characterized by progressive memory loss together
with declining activities of daily living and neuropsychiatric symp-
toms or behavioral changes. There is currently no cure or preven-
tion for this disease. Brains of AD patients are characterized by
the deposition of extracellular neuritic plaques composed out of
amyloid beta (Ab) peptides, by the formation of intracellular neu-
rofibrillary tangles formed by hyperphosphorylated tau proteins
and by neuronal loss.2
In most cases the disease is sporadic and there is some evidence
that the accumulation of Ab proteins is linked with a reduced clear-
ance of Ab3,4 while in familiar AD, gene mutations directly or indi-
rectly increase the production of Ab42.5–7 Ab peptides of different
length are formed from the amyloid precursor protein (APP) by suc-
ulators do not inhibit the total activity of the enzyme, and the pro-
cessing of the other substrates is not affected as demonstrated for
notch which is beside APP the other most characterized substrate
of c
-secretase.10–12 It was shown first with a subset of NSAIDs that
modulation of c
-secretase is possible.13 It may lead to a viable ap-
proach for drug discovery.14,15
In a high through put screen we identified the inosine mono-
phosphate dehydrogenase inhibitor 3 (IMDPH2 IC50 = 400 nM) as a
moderately potent
c-secretase modulator which is structurally re-
lated to the former Phase 1 compound from Eisai, E2012, (1)16–18
and to the Torrey Pines c
-secretase modulator 219 (Fig. 1).
In H4 cells overexpressing human APP20 the HTS hit 3 lowered
Ab42 and Ab40 (Ab42 IC50 = 815 nM; Ab40 IC50 = 3700 nM)
whereas Ab38 was increased (Ab38 EC50 = 1500 nM) and total Ab
remained unchanged up to 20
say this compound did not show any activity up to the highest con-
centration of 25 M tested. In vitro potency may vary substantially
lM. In a cellular notch reporter as-
cessive proteolytic cleavage by BACE and
ses are prime targets for a disease modifying approach.
c
-secretase. These secreta-
-Secretase
c
l
may cleave many different substrates and the inhibition of the en-
when measured in different cellular assays (e.g., Ab42 IC50 of 1 in
zyme can potentially lead to mechanistically-based side effects. In
rat primary neurons = 42 nM).
the case of notch protein processing, a
came apparent in a recent phase III clinical trial.8,9 An alternative ap-
proach consists in the modulation of -secretase where the cleavage
of the APP C-terminal fragment is shifted resulting in the formation
of more soluble, smaller Ab peptides such Ab38 and Ab37 while the
c
-secretase substrate, this be-
The synthesis of the modulators exemplified in Scheme 1 pro-
ceeded smoothly.
The hydroxy derivate 12 was prepared from the corresponding
boc-protected aldehyde 11 via Grignard addition with 4-chloro-
phenyl magnesium chloride (Scheme 2).
c
formation of toxic, aggregating Ab42 is inhibited.
c
-Secretase mod-
The annelated aminothiazoles were prepared from 2-arylke-
tones 13 via regioselective bromination yielding 14 and subse-
quent condensation with the thiourea 5a (Scheme 3, exemplified
for compound 15).
⇑
Corresponding author. Tel.: +41 61 68 83095; fax: +41 61 68 86459.
E-mail address: thomas.luebbers@roche.com (T. Lübbers).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.08.060

T. Lübbers et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6554–6558
6555
Torrey Pines
Eisai
Figure 1. Non-NSAID
c
-secretase modulators and the HTS hit (3).
H
H
N
H
O
N
NH2
O
N
O
N
O
N
N
S
R
a
NO2
O
a, b
S
N
Cl
N
N
N
5a
11
4
5a R = H
b, c
c
NH2
5b R = Bn
+
Cl
H
N
OH
O
O
N
S
N
d, e
7
8
f, g
H
N
H
N
12
Cl
O
N
N
R
Cl
+
Cl
S
Scheme 2. Reagents and conditions: (a) BrCH(CHO)2, NaOAc, EtOH, refl., 12 h, 56%;
(b) Boc2O, DMAPcat, CH2Cl2, 0 °C, rt, 12 h, 49%; (c) (4-Cl-Ph)MgBr, THF, À78 °C, 0 °C,
30 min, 14%.
O
N
9
6a,b
h
and thus led to an increase of Ab production as compared to wild
type mice. 10a reduced brain Ab42 by 47% at a dose of 150 mg/
kg p.o. after 4 h and by 22% after 7 h. The unfavorable physico-
chemical properties and the low free fractions of 10a in mice (be-
low 0.1%) are probably the reason for the only weak in vivo
activity.
R
N
O
S
Cl
N
N
The SAR of the imidazole phenyl head group was investigated
with different aryl substituted aminothiazoles (Table 2). In vitro
potency was substantially lost when the R1- and R2-substituents
were missing (32). One of the substituents was sufficient to rein-
stall the in vitro activity (33 or 34). The 4-position on the imidazole
ring seemed to be critical for CYP inhibition since methylation
could strongly reduce this liability (34 and 35).
N
10a,b
Scheme 1. Reagents and conditions: (a) 4-methylimidazole, KOH, DMSO, 80 °C, 5 h,
45%; (b) 1 atm H2, 10% Pd/C, EtOH, rt, 4 h, 78%; (c) PhCHO, NaBH4, MeOH, rt, 12 h,
85%; (d) PhCONCS, THF, rt, 2 h; (e) K2CO3, H2O, MeOH, 12 h, rt, 82% over two steps;
(f) CuCl2, t-BuONO, MeCN, rt, 30 min; (g) DBU, Et2O, rt, 10 min; (h) EtOH, reflux,
12 h, 43–61%.
Conformational restriction through annelation of the benzylic
position to the aminothiazole ring led to excellent in vitro activity
(tetrahydrobenzothiazole derivatives, Table 3). The MeO- and F-
derivatives 15 and 37 were in vitro equally potent. The H- and
CN-derivatives (36 and 39) lost two-fold in vitro activity, all other
tested substituents even more.
However, all compounds from this annelated thiazole series
were very lipophilic, exhibited low solubilities and had in general
a CYP2D6 issue. CYP2D6 IC50 values for the MeO-, CN- and Cl-
Different head groups were prepared as shown in the following
Schemes 4 and 5 and converted to the final products as demon-
strated in the previous synthetic schemes.
Replacement of the oxazole moiety in the initial HTS hit 3 by the
methyl-imidazole group led to a four-fold increase of potency (Ta-
ble 1, 10a). Alkylation of the nitrogen linker atom led to loss of po-
tency (10b). Introduction of a hydroxyl group at the benzylic
position was tolerated (12). The position of the benzylic moiety
at the thiazole ring seems to be not critical for in vitro potency
since 31 was as active as 10a with the benzyl moiety next to the
sulfur atom and not to the nitrogen atom as in 31.
derivatives were below 0.2
lM and for all other derivatives below
0.5 M. The modulator 15 was tested in a mouse PK experiment
l
(Fig. 2). It had moderate oral bioavailability, moderate volume of
distribution and clearance. The brain/plasma ratio was variable
but the compound was not a PgP substrate.
Compound 10a was tested in vivo in transgenic APPSwe mice
which overexpress human APP with the Swedish FAD mutations21

6556
T. Lübbers et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6554–6558
H
N
O
O
NH2
O
a
Br
+
S
N
N
13
14
5a
b
H
N
O
N
N
S
N
15
Scheme 3. Reagents and conditions: (a) Br2, CHCl3, rt, 3 h, crude; (b) EtOH, refl., 1 d, 53%.
O
NH2
NO2
O
NO2
N+
S
N
a, b
O
a
S
O
+
R
N
N
N
Br
R
O
25
27
29
26
28
30
16
17
18
NO2
NO2
b, c
H
O
B
c
N
O
N
O
+
O
N
N
S
O
N
Br
N
19
20
NO2
O
NO2
O
d-f
H
H
O
O
N
N
O
S
N
N
N
21
Scheme 5. Reagents and conditions: (a) POCl3, MeCN, 90 °C, 2 days, 21%; (b) t-
BuO(NMe2)2C, 110 °C, 3 h, quant; (c) MeC(NH)NH2ÁHCl, NaOMe, MeOH, refl., 4 h,
80%; (d) SOCl2, refl., 5 h, quant; (e) CH2(CO2Me2)2, MgCl2, NEt3, PhMe, rt, 2 h, then
aq. HClconc, rt, 2 h, 49%; (f) PhI(OAc)2, TrifOH, MeCN, refl., 3 h, 56%.
NH2
NH2
Br
d
O
B
O
+
N
N
N
Table 1
N
Selected SAR around compound 10a
22
23
24
Compound
Ab42 IC50 (nM)
H
Scheme 4. Reagents and conditions: R = H, Me (a) KOAc, Pd(PPh3)4cat, AcNMe2,
microwave, 160 °C, 2 h, 42%/62%; (b) SnCl2, EtOH, refl., 1 h, 91%/75%; (c) Pd(OAc)2,
KF, MeOH, 140 °C, microwave, 30 min, 16%; (d) Pd(PPh3)4cat, Na2CO3, MeCN, H2O,
3 h, 100 °C, 89%.
N
S
S
10a
214
N
N
Cl
bn
N
We performed a dose response experiment with 15 in the APP-
Swe mice model. A good correlation between brain exposure and
in vivo efficacy could be established (Fig. 3). The compound exhib-
ited an Ab42 ED50 of roughly 40 mg/kg. At 100 mg/kg p.o. the expo-
sure varied quite substantially probably because of variable
absorption of the compound caused by the poor physicochemical
properties of 15.
We found that the imidazole and oxazole head group could be re-
placed by several other heterocycles. Good in vitro potency was
obtained when the nitrogen atom in the head group was either in
meta-position in the case of five-membered hetaryl rings or in
para-position for six-membered hetaryl rings. The oxazoles 43, 49,
50, pyrazole 21 and thiazoles 46, 47 showed good in vitro activity
whereas the oxadiazoles 44, 45 lost in vitro activity. Even the bism-
ethyl pyrimidine 55 retained in vitro activity. The pyridine 53 and
pyridazine 54 derivatives exhibited good in vitro potency (Table 4).
10b
12
>40,000
Cl
Cl
Cl
O
N
H
N
OH
S
N
N
S
339
136
N
H
N
31

T. Lübbers et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6554–6558
6557
Table 2
100
SAR around the imidazole phenyl head group
individual mice
averaged data
brain/plasma: 1.2
Br
82
H
80
R1
N
N
N
S
R2
O
60
40
20
0
65*
N
10 mg/kg
100 mg/kg
35**
Compound
R1
R2
Ab42 IC50 (nM)
CYP3A4 IC50 (lM)
32
33
34
35
H
MeO
H
H
H
Me
Me
>5000
789
<0.2
<0.2
1.2
30 mg/kg
15**
880
898
100 mg/kg
MeO
2.2
50 100
500 1000
5000 10000 30000
Log (brain exposure) [ng/g]
Table 3
In vitro activity of annelated thiazoles
Figure 3. correlation of brain Ab42 levels with brain exposures of 15 in APPSwe
mice in vivo (p.o. after 4 h).
H
N
R
N
N
Table 4
S
Influence of the head group on the in vitro potency
N
H
N
R
N
S
Compound
R
Ab42 IC50 (nM)
36
37
15
38
39
40
41
42
H
F
MeO
HO
CN
Me
Cl
CF3
133
52
44
209
123
733
675
>1250
R
Ab42 IC50 (nM) (compound)
O
N
N
O
N
N
44 (15)
O
N
N
MeO
N
86 (43)
3431 (44)
999 (45)
S
S
NC
N
N
N
N
3358 (48)
Dose (iv; po) [mg/kg]
Clp [mL/min/kg]
Vdss [L/Kg]
0.7, 7.2
18.3
4.6
244 (21)
O
252 (46)
212 (47)
O
N
N
N
t½ (po) [h]
2.5
S
N
N
N
N
F [%]
25
H
AUCN0-24 po [μM h kg/mg]
231
Brain/plasma
0.4-2.1
133 (49)
651 (51)
5548 (52)
127 (50)
Figure 2. Pharmacokinetic profile of 15 in mice.
N
N
N
N
N
N
N
BrH
483 (55)
BrH
The pyrazole derivative 21 had a low plasma exposure and was
in vivo inactive in the APPSwe mice model (100 mg/kg p.o. 4 h). In
contrast the methyl oxazole derivative 43 showed a 47% reduction
of brain Ab42 in the same model at 100 mg/kg p.o. The plasma
exposure was in the same range as for the methyl imidazole 15
but the brain/plasma ratio was lower (0.34). The plasma exposure
of the mono-methyl thiadiazole 46 and of the pyridine derivative
53 were low and therefore these compounds were inactive
in vivo at 100 mg/kg p.o. The pyridine 53 and the pyridazine deriv-
atives 54 inhibited strongly the CYP2D6 enzyme.
376 (54)
319 (53)
2868 (56)
Acknowledgments
We thank S. Herzig, N. Schinkel, N. Benedetti, M. Mayer, T. Me-
ier, M. Bender and R. Schubenel for their dedicated technical
assistance.
In summary, we explored the SAR around a
c-secretase HTS hit
(3). We could improve in vitro activity of the open-chain
tase modulators substantially when we fixed the conformation of
c
-secre-
References and notes
1. Jellinger, K. A. J. Neural Transm. 2006, 113, 1603.
2. Finder, V. H. J. Alzheimer Dis. 2010, 22, S5.
3. Alvira-Botero, X.; Carro, E. M. Curr. Aging Sci. 2010, 3, 219.
4. Mawuenyega, K. G.; Sigurdson, W.; Ovod, V.; Munsell, L.; Kasten, T.; Morris, J.
C.; Yarasheski, K. E.; Bateman, R. J. Science 2010, 330, 1774.
5. Scheuner, D.; Eckman, C.; Jensen, M., et al Nat. Med. 1996, 2, 864.
the benzyl ring through annelation to the aminothiazole ring.
The
c-secretase modulator 15 was in vivo active at 30 mg/kg.
Replacement of the imidazole head group by different heteroaryl
moieties did not provide superior compounds in respect to their
in vitro or in vivo potency.

6558
T. Lübbers et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6554–6558
6. Borchelt, D. R.; Thinakaran, G.; Eckman, C. B., et al Neurons 1996, 17, 1005.
7. Duff, K.; Eckman, C.; Zehr, C., et al Nature 1996, 383, 710.
8. Saxena, U. Expert Opin. Ther. Targets 2010, 14, 273.
20. Cellular
c
-secretase assay: Human neuroglioma H4 cells overexpressing human
APP were plated at 30,000 cells/well/200
containing 10% FCS, 0.2 mg/l Hygromycin B and incubated for 2 h at 37 °C, 5%
ll in 96-well plates in IMDM media
9. Imbimbo, B. P.; Panza, F.; Frisardi, V.; Solfrizzi, V.; D’Onofrio, G.; Logroscino, G.;
Seripa, D.; Pilotto, A. Expert Opin. Investig. Drugs 2011, 20, 325.
10. De Strooper, B. et al Nature 1999, 398, 518.
CO2 prior to adding test compounds. Compounds for testing were dissolved in
100% Me2SO yielding in a 10 mM stock solution. Typically 12
solutions were further diluted in 1000 l of IMDM media (w/o FCS). Sub
sequential 1:1 dilutions gave ten point dose response curve. Hundred
ll of these
l
11. Douglas, W. E. Neuroscientist 2010, 16, 614.
a
12. Golde, T. E. et al Science 2009, 324, 603.
microliter of each dilution was added to the cells in 96-well plates. Appropriate
controls using vehicle only and reference compound were applied to this assay.
The final concentration of Me2SO was 0.4%. After incubation for 22 h at 37 °C,
13. Weggen, S.; Eriksen, J. L.; Das, P.; Sagi, S. A.; Wang, R.; Pietrizik, C. U.; Findlay, K.
A.; Smith, T. E.; Murphy, M. P.; Bulter, T.; Kang, D. E.; Marquez-sterling, N.;
Golde, T. E.; Koo, E. H. Nature 2001, 414, 212.
5% CO2, 50 ll supernatant was transferred into round-bottom 96-well
14. Review Oehlrich, D.; Berthelot, D. J.-C.; Gijsen, H. J. M. J. Med. Chem 2011, 54,
669.
polypropylene plates for detection of Ab42. Fifty microliter assay buffer
(50 mM Tris/HCl, pH 7.4, 60 mM NaCl, 0.5% BSA, 1% TWEEN 20) was added to
15. Review Pettersson, M.; Kauffman, G. W.; am Ende, C. W.; Patel, N. C.; Stiff, C.;
Tran, T. P.; Johnson, D. S. Expert Opin. Ther. Patents 2011, 21, 205.
16. Portelius, E.; Van Broeck, B.; Andreasson, U.; Gustavsson, M. K.; Mercken, M.;
Zetterberg, H.; Borghys, H.; Blennow, K. J. Alzheimer Dis. 2010, 21, 1005.
17. Nagy, C.; Schuck, E.; Ishibashi, A.; Nakatani, Y.; Rege, B.; Logovinsky V. Eisai
Inc., 10th International Conference on Alzheimer’s & Parkinson’s Diseases,
March 9–13, 2011, Barcelona, Spain Poster Session 2. Abstract# 839.
18. Hashimoto, T.; Ishibashi, A.; Hagiwara, H.; Murata, Y.; Takenaka, O.; Miyagawa
T. 13th International Conference on Alzheimer’s Disease (ICAD 2010), July 10–
15, 2010, Honolulu, Hawaii, Abstract No. P3–307.
the wells followed by the addition of 100
(ruthenylated Ab42-specific antibody BAP15 0.0625
50 l of a premix of capture antibody (biotinylated 6E10 antibody, 1
and Steptavidin-coated magnetic beads (Dynal M-280, 0.125 mg/mL) were
preincubated for 1 h at room temperature before adding the assay plates. Assay
plates were incubated on a shaker for 3 h at room temperature and finally read
in the Bioveris M8 Analyser according to the manufacturer’s instructions
(Bioveris).
l
l
of detection antibody
g/mL in assay buffer).
g/mL)
l
l
l
21. APP Swe mice model: Selected compounds were administered p.o. to ca.
3 month old transgenic mice overexpressing APP-Swe under the control of
mouse Thy-1 promoter. After 4 h brain homogenates (1% DEA) were measured
for Ab38, 40, 42 and total Ab by LP ECL. For PK and PD studies compound levels
were measured in brain and plasma using LC–MS/MS analytics.
19. Kounnas, M. Z.; Danks, A. M.; Cheng, S.; Tyree, C.; Ackerman, E.; Zhang, X.; Ahn,
K.; Nguyen, P.; Comer, D.; Mao, L.; Yu, C.; Pleynet, D.; Digregorio, P. J.;
Velicelebi, G.; Stauderman, K. A.; Comer, W. T.; Mobley, W. C.; Li, Y.-M.; Sisodia,
S. S.; Tanzi, R. E.; Wagner, S. L. Neuron 2010, 67, 769.