Design and synthesis of tricyclic sulfones as γ-secretase inhibitors with greatly reduced Notch toxicity

Article abstract of DOI:10.1016/j.bmcl.2010.02.080

A novel series of tricyclic γ-secretase inhibitors was designed and synthesized via a conformational analysis of literature compounds. The preliminary results have shown that compounds in this new series have much improved in vitro potency and in vivo pro

Full text of DOI:10.1016/j.bmcl.2010.02.080

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2591–2596
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design and synthesis of tricyclic sulfones as
with greatly reduced Notch toxicity
c-secretase inhibitors
a
a
a
a
a
a
Ruo Xu a, , David Cole , Ted Asberom , Tom Bara , Chad Bennett , Duane A. Burnett , John Clader ,
Martin Domalski a, William Greenlee a, Lynn Hyde a, Hubert Josien a, Hongmei Li a, Mark McBriar a,
Brian McKittrick a, Andrew T. McPhail b, Dmitri Pissarnitski a, Li Qiang a, Murali Rajagopalan a,
Thavalakulamgar Sasikumar a, Jing Su a, Haiqun Tang a, Wen-Lian Wu a, Lili Zhang a, Zhiqiang Zhao a
*
a Merck Research Laboratories, 2015 Galloping Hill Road, Kenilworth, NJ 07033, United States
b Department of Chemistry, Duke University, Durham, NC 27708, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of tricyclic c-secretase inhibitors was designed and synthesized via a conformational anal-
ysis of literature compounds. The preliminary results have shown that compounds in this new series have
much improved in vitro potency and in vivo profiles. More importantly, they have greatly reduced Notch
Received 11 January 2010
Revised 17 February 2010
Accepted 19 February 2010
Available online 23 February 2010
related toxicity that was associated with previous
c
-secretase inhibitors.
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Alzheimer’s disease
c
-Secretase inhibitor
CYP
Notch toxicity
Since the 40–42 amino acid amyloid peptides (Ab) are critical
to quickly validate this new series with minimum synthetic effort,
the original cyclohexane ring was broken open to form a bicyclic
ring typified by 2. After the bicyclic series was quickly synthesized
and validated, the third ring was reintroduced to form tricyclic
compound 3.
factors in the onset and progression of Alzheimer’s disease (AD),1
reducing or eliminating the production of Ab through inhibition
of
c
-secretase has been suggested to be a useful treatment for
AD.2 Previous sulfonamide
c-secretase inhibitors from our labora-
tories have persistent Notch related toxicity in addition to P-450
enzyme inhibition and liver toxicity liabilities although they have
excellent in vitro potencies and in vivo efficacies.3 To address this
The straight forward synthesis of key compounds in bicyclic
series is shown in Scheme 2. Phenylsulfone 4 was synthesized in
quantitative yield from 2,3,6-trifluorobenzyl bromide and was
then alkylated with protected hydroxyalkyl bromide to provide 5.
The protecting group was removed and the product underwent
SNAr reaction when treated with a base to provide bicyclic sulfone
6. Alkylation at the 4-position of 6 with various alkyl halides gave
compounds 7. Alternatively, commercially available ketones 8 that
have various substitutions on the A ring were treated with thio-
phenol and reducing reagent to generate sulfide which was then
oxidized in situ to sulfones 9. These compounds were similarly
alkylated to give 10. The same synthetic schemes could also be
used to modify the A ring substitutions and B ring sizes.
In order to introduce substitutions at the 2- and 3-positions of the
B ring, different synthetic routes were developed as shown in
Scheme 3. Sulfone 4 was treated with n-butyl lithium, followed
by an optionally substituted epoxide to form compound 11.
Compound 11 was converted to 2-substituted compound 12 and
2,4-disubstituted compound 13 using similar reaction sequences
shown before. On the other hand, vinyl diester 14 was first
synthesized by condensation of 2,3,6-trifluorobenzaldehyde with
issue, we designed and synthesized a novel series of
c-secretase
inhibitors via conformational analysis of an earlier sulfone lead.
Compounds in this new series have comparable or better in vitro
and in vivo profiles compared to the previous compounds. Addi-
tionally, they completely eliminated P-450 enzyme inhibition
and liver toxicity problem of the sulfonamide series. More impor-
tantly, the compounds in this new series have greatly reduced
Notch related toxicity.
Previously literature has disclosed an aryl sulfone series as
potent
formational analysis of this compound led to the design of a novel
c
-secretase inhibitor as exemplified by 1.4 A detailed con-
bicyclic and tricyclic sulfone
Scheme 1.
c-secretase inhibitors as shown in
The basic design was to form a covalent bond between two car-
bon atoms that are geometrically very close to each other. In order
* Corresponding author. Tel.: +1 908 740 3483; fax: +1 908 740 7152.
E-mail addresses: ruo.xu@merck.com, ruo.xu@spcorp.com (R. Xu).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.02.080

2592
R. Xu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2591–2596
Break the bond
H
O2
S
O2
S
R'
O2
S
R
F
F
X
X
F
Cl
Cl
Cl
Form a bond
F
F
F
2
1
F
F
F
A
X
X
B
R
C
R'
F
O2S
O2S
X = O, CH2 or N
Reintroduce
C ring
Cl
Cl
3
2
Scheme 1.
F
F
F
O
F
F
F
F
d
c
OTBS
a, b
F
O2S
F
O2S
F
O2S
F
Br
Cl
Cl
Cl
4
5
6
e
F
O
R3
F
O2S
Cl
7
X
X
R
R
X
f
e
R3
R
SO2
Cl
SO2
O
Cl
8
9
10
Scheme 2. Reagents and conditions: (a) 4-chlorobenzenethiol, NEt3; (b) MCPBA 98% for two steps; (c) BrCH2(CH2)nOTBS, THF, KOt-Bu, 28%; (d) TBAF; then NaH, 67%; (e) alkyl
bromide, KOt-Bu, THF; (f) 4-chlorothiophenol, BH3, then MCPBA.
dimethylmalonate and then reacted with 4-chlorothiophenol in the
presence of potassium carbonate to give 15. Reduction of the diester
15 with DIBALH, followed by the SNAr reaction gave compound 16.
During the B ring closure reaction, only the desired trans configura-
tion between the thioether and hydroxyl methyl groups was formed.
The thioether was oxidized to the sulfone, and standard functional
group manipulation generated compound 17 with various substitu-
tions on the 3- and 4-positions of B ring.
Considerable effort went into the development of a synthetic
route to synthesize the rather complex tricyclic compounds. The
most practical synthesis from this effort is shown in Scheme 4.
Chromene 22 was synthesized according to a patent procedure5

R. Xu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2591–2596
2593
F
F
F
F
R1
OH
R1
O
R1
R3
F
F
O
a
c
b
F O2S
F O2S
F O2S
F O2S
Cl
Cl
Cl
Cl
4
11
12
13
F
F
F
F
F
F
F
F
F
S
O
S
COOCH3
F
F
e
f, b
d
COOMe
COOMe
OH
COOCH3
O
Cl
Cl
14
15
16
g, h
F
O
R2
R3
F
O2S
Cl
17
Scheme 3. Reagents and conditions: (a) n-BuLi, substituted epoxide, 52% when R1 = Me; (b) NaH; (c) alkyl bromide NaH, THF; (d) dimethylmalonate, NEt3; MsCl;
(e) 4-chlorothiophenol, K2CO3, 72% for two steps; (g) MCPBA, 80%; (h) functional group manipulation and alkylation.
and was then oxidized with MCPBA gave the epoxide 23. 4-Chloro-
thiophenol was then added to 23 in a regio selective fashion in the
presence of indium chloride,6 resulted in the desired regio isomer
24 as the sole product. Oxidation of the thioether 24 to the sulfone,
followed by dehydration gave the vinyl sulfone 25. The key step in
this synthesis was the Diels–Alder reaction of compound 25 with
2-trimethylsilyloxy 1,3-butadiene.7 After an acidic work up of this
reaction, a single regio isomer ketone 26 was obtained. The ketone
was reduced to the hydroxyl group with sodium borohydride at
low temperature to yield only the desired trans isomer 27. It was
determined early on that the biological activities of these com-
pounds resided mainly in one of the enantiomers, so pure enanti-
omers were needed to fully evaluate these compounds. In the
absence of a viable asymmetric synthesis, chiral separation was
used to obtain pure enantiomers. After evaluating all intermediates
leading to the final product for feasibility of chiral separation, it
was found that the hydroxyl compound 27 gave the best separa-
tion with all existing chiral columns. The two enantiomers were
separated at this stage using chiral HPLC, and the hydroxyl group
of the desired enantiomer was converted to amino group in three
steps to yield amine 28. It should be noted that the C-8 chiral cen-
ter was completely inverted at the step when the mesyl group was
replaced by the azide group. As a result, only the compound with
the desired cis configuration between amino and sulfone groups
was obtained. Further derivatives of amine 29 were synthesized
as usual.
The absolute configuration of the active enantiomer was deter-
mined by X-ray crystallography of triflic amide (29c, R = CpSO2)
and the result is shown in Figure 1.8 Compounds with this absolute
stereo configuration were more active in our biological assays.
Thus, the detailed in vivo studies were all performed using only
pure enantiomers with this absolute configuration.
The bicyclic compounds were synthesized to validate this new
series and to optimize the type and size of newly formed B ring.
Since A ring substitutions were well defined by literature prece-
dents and our in-house data,3,4 they were kept constant at this
time and B ring substitutions were varied. The Ab40 IC50s of se-
lected compounds in this series are listed in Table 1.
The data for compounds 13a–13c show that substitution at the
2-position of B-ring is not tolerated. Even the small alkyl or
hydroxymethyl groups dramatically reduced the potency of result-
ing compounds. On the other hand, small alkyl groups on the 3-
and 4-positions improved the potency dramatically. A simple addi-
tion of an ethyl group at the 4-position improved the membrane
Ab40 IC50 from 1600 nM (6) to 72 nM (7b). A trans methyl group
at 3-position improved the membrane Ab40 IC50 from 1600 nM

2594
R. Xu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2591–2596
F
F
O
F
F
F
F
O
S
O
O
c, d
a
b
OH
F O2S
F
O
Cl
Cl
22
23
24
25
e
F
F
F
F
O
O
O
O
H
k
h, i, j
f, g
H
H
H
F
F
F
F
OH
NH2
O2S
NHR
O2S
O2S
O
O2S
Cl
Cl
Cl
Cl
29
28
27
26
Scheme 4. Reagents and conditions: (a) MCPBA 60%; (b) p-chlorothiophenol, lnCl3, 57%; (c) MCPBA; (d) mesyl chloride, NEt3, 64% for two steps; (e) 2-trimethylsilyoxy 1,3-
butadiene, trifluorotoluene, 160 °C, 48%; (f) NaBH4, À78 °C to rt, 67%; (g) Chirl OD Column; (h) mesyl chloride, NEt3; (i) NaN3; (j) PPh3, H2O, 44%for three steps; (k) various
conditions.
Table 1
F
Membrane Ab40 IC50 data for selected bicyclic compounds
O
F
H
O
R1
R2
O
O
F
O2S
S
N
R3
H
F
O2S
Cl
Figure 1. The crystal structure of 29c.
Cl
Compda
R1
R2
R3
Mem. IC50 (nM)
b
6
7a
7b
H
H
H
H
H
H
H
H
H
H
CH3
1600
808
72
1200
830
620
510
64
(6) to 64 nM (17b). Unfortunately, all attempts to further improve
the potency of the bicyclic compound was not successful. Introduc-
ing other substitutions on other positions of A and B ring also re-
sulted in less active compounds.
CH2CH3
CH3
CH2CH3
CH2CH3
H
H
CH3
CH3
CH2CH3
13a
13b
13cc
17a
17b
17c
17d
17e
trans CH3
trans CH2CH3
CH2OH
H
H
H
H
H
trans CH2OH
trans CH3
trans CH2OCH3
trans CH3
Compounds such as 7b and 17b were potent enough to warrant
further studies, especially when they were resolved to the corre-
sponding pure enantiomers. Unfortunately, further studies re-
vealed that all these compounds have poor pharmacokinetic
profiles most likely due to their high Log P values. Attempts to re-
duce their Log P by introducing polar groups at 3- and 4-positions
yielded only less active compounds. Since both literature and our
in-house work had shown that the p-chlorophenyl sulfone is one
of the best substitutions at bottom portion of the molecule, we
kept this part of the molecule constant for the scope of this study.
Additionally, a pyran ring proved to be optimal in a limited explo-
ration of various B rings. When the oxygen atom of the B ring was
replaced by carbon, nitrogen and sulfur, the resulting compounds
were much less active in general.9 This reason, coupled with the
fact that pyran B ring was much easier to access synthetically,
prompted us to pursue only the pyran as B ring in all future studies.
Once the design of the new series was validated in the bicyclic
series, we turned our attention to tricyclic compounds. The syn-
270
110
97
trans CH3
a
b
c
All compounds are racemic unless otherwise indicated.
Values are means of two experiments.
Mixture of diastereomers.
thetic effort of this complex ring system paid off in the end since
the resulting compounds were among the best we have obtained
in this series. A large library of compounds with substitutions at
8-position, including amides, sulfonamides, urea, carbonamides,
sulfones and heterocycles was synthesized and evaluated. The bio-
logical data for selected compounds are listed in Table 2.
As the data in Table 2 shows, many different substitutions at the
8-positon of the C ring were tolerated, resulting in compounds
with good to excellent in vitro activities. This provided a very good

R. Xu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2591–2596
2595
Table 2
Biological data for tricyclic compounds
F
O
H
F
O2S
R
Cl
Compda
R
IC50 (nM) Mem.
IC50 (nM) Cell Ab40, Ab42
CYP 3A4c
(lM)
RR AUC (h ng/ml)
CRND8 Ab40 at 3 hd Inh. Plasma, Cortex
b
b
26
27
@O
OH
NH2
NHTf
NHAc
13
28
1700
27.1
107
17, 12
25, 18
—
108, 46
145, 107
1.2
9.1
—
>30
—
60
164
—
5052
1725
19% (Plasma)
44%, 55%
—
90%, 42%
—
28
29ae
29b
O
O
29ce
8.4
7.6
9.3, 4.3
68, 26
>30
—
1909
—
80%, 58% at 6 h
19% (Plasma)
S
HN
O
O
N
29de
S
HN
a
b
c
All compounds are racemic unless otherwise indicated.
Values are means of two experiments.
Values determined after 30 min pre-incubation with compound.
Dosed orally at 30 mg/kg.
d
e
Pure enantiomer, absolute structure is indicated in Figure 1.
3. (a) Josien, H.; Bara, T.; Rajagopalan, M.; Asberom, T.; Clader, J. W.;
Favreau, L.; Greenlee, W. J.; Hyde, L. A.; Nomeir, A. A.; Parker, E. M.;
Pissarnitski, D. A.; Song, L.; Wong, G. T.; Zhang, L.; Zhang, Q.; Zhao, Z.
Bioorg. Med. Chem. Lett. 2007, 17, 5330; (b) McBriar, M. D.; Clader, J.
W.; Chu, I.; Del Vecchio, R. A.; Favreau, L.; Greenlee, W. J.; Hyde, L. A.;
Nomeir, A. A.; Parker, E. M.; Pissarnitski, D. A.; Song, L.; Zhang, L.; Zhao,
Z. Bioorg. Med. Chem. Lett. 2008, 18, 215; (c) Li, H.; Asberom, T.; Bara, T.
A.; Clader, J. W.; Greenlee, W. J.; Josien, H. B.; McBriar, M. D.; Nomeir,
A.; Pissarnitski, D. A.; Rajagopalan, M.; Xu, R.; Zhao, Z.; Song, L.; Zhang,
L. Bioorg. Med. Chem. Lett. 2007, 17, 6290; (d) Asberom, T.; Zhao, Z.;
Bara, T. A.; Clader, J. W.; Greenlee, W. J.; Hyde, L. A.; Josien, H. B.; Li,
W.; McPhail, A. T.; Nomeir, A. A.; Parker, E. M.; Rajagopalan, M.; Song,
L.; Wong, G. T.; Zhang, L.; Zhang, Q.; Pissarnitski, D. A. Bioorg. Med.
Chem. Lett. 2007, 17, 511.
4. 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.;
Mortishire-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.
5. Sund, C.; Roue, N.; Lindstroem, S.; Antonov, D.; Sahlberg, C.; Jansson, K. WO
2005066131 A1; Chem. Abstr. 2005, 143, 153297.
handle to fine tune the overall profile of the final compounds. Fur-
ther testing has shown that these compounds had no CYP or liver
toxicity issues as seen in the earlier series. Selected compounds
were also tested in an acute efficacy assay in our CRND8 mouse
model.10 Overall, compound 29a had the best efficacy in this assay,
and it inhibited plasma Ab40 by 90% and cortex Ab40 by 42% three
hours after a single oral dose of 30 mg/kg. This compound also sig-
nificantly lowered plasma and cortical Ab40 after chronic dosing in
the CRND8 mouse and did not produce Notch related GI toxicity
under conditions where our earlier sulfonamide series produced
significant GI toxicity.11 The full results of this study will be pub-
lished in due course.
In summary, a novel series of tricyclic sulfone c-secretase inhib-
itors was designed and synthesized. Compounds in this new series
have comparable or better in vitro activities and in vivo efficacies
compared to our previously reported compounds. Additionally,
they have completely eliminated P-450 enzyme inhibition and li-
ver toxicity problem that was seen in our previous sulfonamide
series. More importantly, further studies revealed that they have
greatly reduced mechanism based Notch side effects compared to
6. Ranu, B. C.; Mandal, T. Can. J. Chem. 2006, 84, 762.
7. The procedure for synthesis tricyclic compound 33.
F
F
O
O
other c-secretase inhibitors.
H
°
1. p-CF3Ph, 160
C
+
F O2S
F
OTMS
Acknowledgements
2. HCl
O
O2S
We thank John Hunter for his unwavering support of this pro-
ject; Teresa Andreani, Mark Liang, Tao Meng Jianshe Kong and Jesse
Wong for scaling up of intermediates; T. M. Chan, Alexei Buevich,
Andy Evans and Rebecca Osterman for determining and confirming
stereochemistry of the compounds via NMR studies.
Cl
Cl
26
25
Compound 25 (22 g, 64 mmol) was dissolved in 300 ml trifluorotoluene and 2-
trimethylsilyloxy-1,3-butadiene (36.5 g, 260 mmol) was added. The reaction
was heated to 160 °C in a sealed tube for 7 h. The reaction was cooled to room
temperature and solvent was removed. The residue was dissolved in 200 ml
THF and 5 ml 1 N HCl solution was added. The reaction was stirred for 0.5 h
at room temperature. EtOAc (200 ml) was added and the organic layer was
washed with water (200 ml), brine (200 ml), dried over Na2SO4 and concen-
trated. The residue was purified by column (EtOAc/hexane from 0/100 to 50/
References and notes
1. (a) Hardy, J. A.; Selkoe, D. J. Science 2002, 297, 353; (b) Selkoe, D. J. Neuron 1991,
6, 487; (c) Hardy, J. A.; Higgins, G. A. Science 1992, 256, 184.
2. (a) Nguyen, J.; Yamani, A.; Kiso, Y. Curr. Pharm. Des. 2006, 12, 4295;
(b) Harrison, T.; Churcher, I.; Beher, D. Curr. Opin. Drug Disc. 2004, 7,
709.
50 in 45 min). Yield: 12.6 g, 48%. 1H NMR (CDCl3 400 MHz)
J = 8.1 Hz, 2H), 7.52 (d, J = 8.8 Hz, 2H), 7.14 (m, 1H), 6.64 (m, 1H), 5.27 (dd,
d 7.63 (d,

2596
R. Xu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2591–2596
J = 11.7 and 2.9 Hz, 1H), 4.13 (d, J = 11.7 Hz, 1H), 3.18 (d, J = 12.4 Hz, 1H), 2.79
10. Hyde, L. A.; McHugh, N. A.; Chen, J.; Zhang, Q.; Manfra, D.; Nomeir, A. A.; Josien,
H.; Bara, T.; Clader, J. W.; Zhang, L.; Parker, E. M.; Higgins, G. A. J. Pharmacol.
Exp. Ther. 2006, 319, 1133.
11. Wong, G. T.; Manfra, D.; Poulet, F. M.; Zhang, Q.; Josien, H.; Bara, T.; Engstrom,
L.; Pinzon-Ortiz, M.; Fine, J. S.; Lee, H. J.; Zhang, L.; Higgins, G. A.; Parker, E. M. J.
Biol. Chem. 2004, 279, 12876.
(dt, J = 12.4 and 3.7 Hz, 1H), 2.4–2.57 (m, 4H), 2.04–2.17 (m, 1H).
8. CCDC 766356 contains the supplementary crystallographic data for this
structure. These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
9. Unpublished in-house data.