Synthesis and biological activity of flurbiprofen analogues as selective inhibitors of β-amyloid1-42 secretion

Article abstract of DOI:10.1021/jm0502541

Flurbiprofen, a nonsteroidal antiinflammatory drug (NSAID), has been recently described to selectively inhibit β-amyloid_1-42 (Aβ42) secretion, the most toxic component of the senile plaques present in the brain of Alzheimer patients. The use of this NSAID in Alzheimer's disease (AD) is hampered by a significant gastrointestinal toxicity associated with cyclooxygenase (COX) inhibition. New flurbiprofen analogues were synthesized, with the aim of increasing Aβ42 inhibitory potency while removing anti-COX activity. In vitro ADME developability parameters were taken into account in order to identify optimized compounds at an early stage of the project. Appropriate substitution patterns at the alpha position of flurbiprofen allowed for the complete removal of anti-COX activity, while modifications at the terminal phenyl ring resulted in increased inhibitory potency on Aβ42 secretion. In rats, some of the compounds appeared to be well absorbed after oral administration and to penetrate into the central nervous system. Studies in a transgenic mice model of AD showed that selected compounds significantly decreased plasma Aβ42 concentrations. These new flurbiprofen analogues represent potential drug candidates to be developed for the treatment of AD.

Full text of DOI:10.1021/jm0502541

J. Med. Chem. 2005, 48, 5705-5720
5705
Synthesis and Biological Activity of Flurbiprofen Analogues as Selective
Inhibitors of â-Amyloid1-42 Secretion
Ilaria Peretto,§ Stefano Radaelli,§ Carlo Parini,§ Michele Zandi,§ Luca F. Raveglia,§ Giulio Dondio,§
Laura Fontanella,§ Paola Misiano,§ Chiara Bigogno,§ Andrea Rizzi,† Benedetta Riccardi,† Marcello Biscaioli,†
Silvia Marchetti,† Paola Puccini,† Silvia Catinella,† Ivano Rondelli,† Valentina Cenacchi,† Pier Tonino Bolzoni,†
Paola Caruso,† Gino Villetti,† Fabrizio Facchinetti,‡ Elda Del Giudice,‡ Nadia Moretto,| and Bruno P. Imbimbo*,†
Research &Development, Chiesi Farmaceutici S.p.A., Via Palermo 26/A, 43100 Parma, Italy, NiKem Research, Via Zambeletti
25, 20021 Baranzate, Milan, Italy, Neurobiology Unit, Research & Innovation, Via Svizzera 16, 35127 Padua, Italy, and
Department of Biochemistry and Molecular Biology, University of Parma, Parco Area delle Scienze 23/A, 43100 Parma, Italy
Received March 21, 2005
Flurbiprofen, a nonsteroidal antiinflammatory drug (NSAID), has been recently described to
selectively inhibit â-amyloid1-42 (Aâ42) secretion, the most toxic component of the senile plaques
present in the brain of Alzheimer patients. The use of this NSAID in Alzheimer’s disease (AD)
is hampered by a significant gastrointestinal toxicity associated with cyclooxygenase (COX)
inhibition. New flurbiprofen analogues were synthesized, with the aim of increasing Aâ42
inhibitory potency while removing anti-COX activity. In vitro ADME developability parameters
were taken into account in order to identify optimized compounds at an early stage of the
project. Appropriate substitution patterns at the alpha position of flurbiprofen allowed for the
complete removal of anti-COX activity, while modifications at the terminal phenyl ring resulted
in increased inhibitory potency on Aâ42 secretion. In rats, some of the compounds appeared to
be well absorbed after oral administration and to penetrate into the central nervous system.
Studies in a transgenic mice model of AD showed that selected compounds significantly
decreased plasma Aâ42 concentrations. These new flurbiprofen analogues represent potential
drug candidates to be developed for the treatment of AD.
Introduction
treatment and to depend on the specific chemical
structure of the NSAIDs being used, with nonaspirin
agents providing the lowest relative risk.3 Despite these
encouraging findings, all large, long-term, placebo-
controlled clinical trials aimed at reducing inflammation
in the brain of AD patients have produced negative
results.4 More recently, it has been shown that some
NSAIDs decrease the production of Aâ42 in vitro and
in vivo5-7 and counteract the progression of Aâ42
pathology in transgenic mouse models of AD.8-14 The
proposed mechanism for this activity is an allosteric
modulation of presenilin-1 (PS-1), the major component
of the γ-secretase complex, the enzyme responsible for
the formation of Aâ.15-18 The inhibition of Aâ42 produc-
tion is independent of the anti-cyclooxygenase activ-
ity5,19 and is related to the chemical structure of the
compounds, with some NSAIDs being active (ibuprofen,
sulindac, flurbiprofen, indomethacin, diclofenac) and
others not (naproxen, aspirin, celecoxib). This could
explain the negative results of the large AD trials
carried out so far, since they were performed with
compounds (naproxen, hydroxychloroquine, dapsone,
prednisone, rofecoxib and celecoxib) not able to decrease
Aâ42 production.20 Unfortunately, the generic use of
NSAIDs in AD is hampered by a significant gastrointes-
tinal toxicity associated with COX inhibition.
Alzheimer’s disease (AD) is the most common form
of dementia, affecting approximately 5% of the popula-
tion over the age of 65. The basic pathological abnor-
malities in AD are amyloid plaques, neurofibrillary
tangles and neuronal death. Their exact relationship is
still unclear, but the major therapeutic efforts are
presently directed in influencing amyloid metabolism.1
Amyloid plaques consist of a proteinaceous core com-
posed of 5-10 nm amyloid fibrils surrounded by dys-
trophic neurites, astrocytic processes and microglial
cells. The â-amyloid peptide (Aâ) consists of 39-42
amino acids generated by the cleavage of the amyloid
precursor protein (APP). The main form of Aâ contains
40 amino acids (Aâ40), but the carboxy-terminal-
extended species made of 42 residues (Aâ42) is also
produced. This longer form is more prone to aggregate
into fibrils and Aâ42 makes up the major component of
amyloid plaques. Plaques are in equilibrium with
soluble monomers and oligomers and may not be the
sole contributor to neuronal death, as there is evidence
that soluble amyloid is also potentially neurotoxic.2
Epidemiological studies have documented a reduced
prevalence of AD among patients using nonsteroidal
antiinflammatory drugs (NSAIDs). The protective ef-
fects of NSAIDs appear proportional to the length of
We tried to identify new NSAIDs analogues endowed
with potent and selective Aâ42 lowering activity but
devoid of COX inhibitory activity, thus suitable for
chronic use in AD patients. Specifically, we concentrated
our efforts on flurbiprofen and flurbiprofen derivatives,
since this compound has been reported as the most
* Corresponding author: Bruno P. Imbimbo, phone: +39 0521
279278, fax +39 0521 279579, e-mail b.imbimbo@chiesigroup.com.
† Chiesi Farmaceutici S.p.A.
§ NiKem Research.
‡ Research & Innovation.
| University of Parma.
10.1021/jm0502541 CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/06/2005

5706 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Peretto et al.
Table 1. Variations of the Aromatic Rings of Flurbiprofen
Scheme 1. Planned Chemical Variations on
Flurbiprofen
Aâ42
inhibition
IC50
compound
R1
R2
3-F
(@ 100 µM) (µM)a
flurbiprofen phenyl
30%
87%
75%
13%
58%
55%
79%
44%
60%
25%
96%
40%
11%
67%
42%
27%
76%
75%
57%
40%
2%
305
75
5a
5b
5c
5d
5e
5f
5g
5h
5i
3,4-dichlorophenyl
3,5-dichlorophenyl
2-chlorophenyl
3-chlorophenyl
4-chlorophenyl
4-cyclohexylphenyl
4-methylphenyl
4-CF3 phenyl
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
3-F
77
nda
nd
nd
21
nd
129
nd
83
active NSAID studied so far.6 It is well-known that the
R-enantiomer of flurbiprofen is much less active than
the S-enantiomer on both COX-1 (IC50 of 44 versus 0.03
µM) and COX-2 (IC50 of 123 versus 0.91 µM) activity.21
Indeed, the R-enantiomer (MPC-7869) is reported to be
in Phase 3 clinical development for Alzheimer’s disease
in the USA,20 with the rationale that the reduction of
COX inhibitory activity should minimize the gastrointes-
tinal toxicity potential. However, the IC50 of R-flurbi-
profen on Aâ42 (280 µM) is still higher than that on
both COX-1 (44 µM) and COX-2 (123 µM). Thus, if active
concentrations of R-flurbiprofen on Aâ42 are reached
in Alzheimer patients (i.e., around 280 µM) this will
necessarily imply that a significant anti-COX activity
is maintained. On the basis of these considerations, we
developed a chemistry program with the following
objectives: (a) to increase Aâ42 lowering properties, (b)
to remove anti-COX activity from flurbiprofen.
2-CF3 phenyl
5j
5k
5l
3-benzothienyl
3-thienyl
4-pyridyl
4-methoxyphenyl
3,4 methylendioxyphenyl
4-hydroxyphenyl
4-cyclohexyloxyphenyl
nd
nd
67
5m
5n
5o
13a
16d
6a
6b
6c
210
nd
41
4-(4-hydroxyphenyl)phenyl 3-F
35
4-CF3 phenyl
4-CF3 phenyl
4-CF3 phenyl
2-CF3
3-Cl
H
nd
nd
nd
a nd ) not determined.
pounds 11a-h (Table 5). The coupling was performed
in ligand-free conditions22 by reacting 10a,b with the
suitable boronic acids in aqueous sodium carbonate in
the presence of TBAB and Pd(OAc)2. Reactions were
performed at 120 °C under microwave or conventional
thermal heating in a sealed tube. These conditions gave
a faster reaction time, improved yields and simplified
the purification of crude materials.
Derivatives 13a-d were synthesized by alkylation or
by Mitsunobu reaction on the corresponding ester of 5o
or 12 that were obtained through a traditional Suzuki
coupling (Scheme 4).
Terphenyl derivatives 16a-d were prepared through
a combination of the two Suzuki protocols previously
employed (Scheme 5). Intermediate 10b was converted
into the corresponding iodo derivative 14 by a Ni-
mediated halogen exchange23 in order to allow a cleaner
Suzuki coupling reaction with 4-bromobenzeneboronic
and 4-bromo-3-fluorobenzeneboronic acids. Hence, 14
and 4 were submitted to a traditional Suzuki coupling
leading to 15a-c in about 60% yield, whereas the final
coupling was performed in ligand free conditions as
previously described, affording 16a-d.
Compounds featuring bioisosteres of the carboxylic
acid group (Table 3) were prepared according to Schemes
6 and 7. Hydroxamic acid 18 and acylsulfonamide 19
were prepared by reaction of the corresponding acyl
chloride 17 with ammonia, hydroxylamine or methane-
sulfonamide, respectively (Scheme 6).
Chemistry. Several possible chemical variations
were considered and implemented on the basis of
flurbiprofen structure (Scheme 1).
A first array of compounds was synthesized through
a Suzuki coupling reaction between 4 and the suitable
commercially available aryl or heteroaryl boronic acids.
Intermediate 4 was prepared in four steps as outlined
in Scheme 2; the anion of diethyl 2-methylmalonate was
reacted with 2,4-difluoronitrobenzene to obtain 1. After
catalytic hydrogenation, the resulting amine 2 was
converted into the corresponding diazonium salt that
was reacted with potassium iodide to obtain 3. Finally,
hydrolysis and decarboxylation of 3 afforded 4 in 57%
overall yield.
Suzuki couplings were performed on 1 mmol scale in
a parallel way by employing a Myriad Personal Syn-
thesizer. Intermediate 4 was reacted with the appropri-
ate boronic acids in 1,2-dimethoxyethane at 80 °C in the
presence of tetrakis(triphenylphosphine)palladium and
aqueous sodium carbonate to obtain 5a-o (Table 1).
The same reaction sequence was adopted to synthe-
size compounds 6a-c that feature different substituents
on the central aromatic ring.
After evaluation of the first array of biphenyl deriva-
tives, a new synthetic pathway was developed in order
to allow the synthesis of gem-disubstituted derivatives
(Scheme 3). 4-Bromo-3-fluorotoluene was brominated
with NBS to afford 7 that was reacted with potassium
cyanide to obtain 8. From this intermediate gem-
disubstituted derivatives 9a,b could be prepared by
reaction with the appropriate methyl iodide or 1,2-
dibromoethane. The cyano group was then hydrolyzed
to afford 10a,b that was submitted to a modified Suzuki
coupling reaction to obtain the expected biaryl com-
Tetrazole 21 and 2H-[1,2,4]-oxadiazol-5-one derivative
23 were obtained from nitrile 20 that in turn was
synthesized from the corresponding acid 5h in three
steps. Compound 21 was prepared by reacting 20 with
Bu3SnN3, whereas 23 was obtained by cyclization with
CDI of N-hydroxyamidine 22, obtained by reaction of
20 with hydroxylamine (Scheme 7).
Similarly, compounds 24, 25 and 26, featuring a
methyl ester, tert-butyl ester and a primary amide group

Flurbiprofen Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18 5707
Scheme 2a
a Reagents and conditions: (a) NaH, DMSO; (b) H2, Pd/C; (c) 1. NaNO2, HCl, 0 °C; 2. KI; (d) NaOH, EtOH-H2O, 100 °C, 57% (four
steps); (e) R1B(OH)2, Pd(PPh3)4, Na2CO3, DME-H2O, 80 °C.
Scheme 3a
a Reagents and conditions: (a) NBS, CCl4; (b) KCN, EtOH; (c) RX or Br(CH2)2Br, NaOH, toluene-H2O; (d) NaOH, EtOH-H2O, 100 °C;
(e) R1B(OH)2, Pd(OAc)2, Na2CO3, TBAB, H2O, 120 °C (MW or oil bath, sealed tube).
Scheme 4a
a Reagents and conditions: (a) 1. MeOH, H2SO4, reflux; 2. R5OH, DEAD, PPh3, THF, reflux; 3. KOH, THF/MeOH, reflux.
respectively, were obtained from the corresponding
carboxylic acid 5b by known procedures (Scheme 8).
Biology. The effects of the compounds on Aâ42 and
Aâ40 secretion were evaluated in human neuroglioma
cells (H4-APP695NL) using a validated immunoassay
(see Experimental Section). The effects of flurbiprofen
and of a representative analogue (11c) on other Aâ
species were studied using MALDI-TOF technique.

5708 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Peretto et al.
Scheme 5a
a Reagents and conditions: (a) KI, Ni, DMF, 150 °C; (b) 4-bromophenylboronic acid or 4-bromo-3-fluorophenylboronic acid, Pd(PPh3)4,
Na2CO3, DME-H2O, 80 °C; (c) 4-fluorophenylboronic acid or 4-hydroxyphenylboronic acid, Pd(OAc)2, Na2CO3, TBAB, H2O, 120 °C, MW or
oil bath, sealed tube.
Scheme 6a
P450, total body clearance and penetration in cere-
brospinal fluid after continuous subcutaneous admin-
istration in rats.
Structure-Activity Relationships (SAR). Inves-
tigation of SAR for flurbiprofen analogues focused
initially on substitution at the terminal phenyl ring of
the molecule (Table 1). Derivatives bearing the 3-chloro
(compound 5d) or the 4-chloro substitution on the
terminal phenyl ring (compound 5e) showed an in-
creased activity for the inhibition of Aâ42 production
(about 50% at the concentration of 100 µM). On the
contrary, derivative 5c featuring the o-chloro substitu-
tion at the terminal phenyl ring was significantly less
active (13% inhibition at the concentration of 100 µM).
Combined 3,4-dichloro substitution (compound 5a)
showed a further improved activity, with an IC50 of 75
µM; compound 5b, featuring the 3,5-dichloro substitu-
tion, showed a similar potency (IC50 of 77 µM). Com-
parison of compounds 5h and 5i, bearing a trifluoro-
methyl substituent at position para and ortho, respect-
ively, seemed to confirm that para substitution is better
than ortho substitution in inhibiting Aâ42 production.
Substitution with alkoxy derivatives (compounds 5m,
5n, 13a) also led to an increase of Aâ42 inhibition, while
the corresponding hydroxy derivative 5o was less active;
this effect was more prominent with bulky alkoxy
groups (13a). Compound 13a also showed to selectively
inhibit Aâ42 formation, since Aâ40 inhibition (Table 1)
was less significant (25% inhibition at the concentration
of 100 µM).
a Reagents and conditions: (a) (COCl)2, DCM; (b) NH2OH, Et3N,
DCM; (c) MeSO2NH2, Et3N, DCM.
During the experiments on human neuroglioma cells,
the effects of compounds on cell viability were simul-
taneously assessed using the MTT cytotoxicity assay:
all compounds showed less than 15% of cytotoxicity at
the concentrations used for the test. The pharmacologi-
cal selectivity of selected compounds was evaluated on
COX-1 and COX-2 enzyme activity assay. The effects
of a few compounds on another substrate of γ-secretase,
Notch-1, were assessed using a luciferase assay. In vitro
ADME developability parameters of selected compounds
were explored by evaluating membrane permeability in
human colon adenocarcinoma (Caco-2) cell monolayers,
interactions with major isoforms of human cytochrome
Compounds bearing alkyl instead of alkoxy residues
followed the same trend as the alkoxy derivatives, with
bulky alkyl substituents (such as the cyclohexyl group,

Flurbiprofen Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18 5709
Scheme 7a
Scheme 8a
a Reagents and conditions: (a) (COCl)2, (b) DCM, MeOH or
tBuOH; (c) NH3, DCM.
a high clearance; the concentration in the cerebrospinal
fluid was about 1-2% of the concentration in plasma
for 5f and 16d (similarly to what was observed with
flurbiprofen) or below the detection limit for compound
13a.
Substitution at the central aromatic ring was studied
with derivatives 6a, 6b, 6c (Table 1) bearing the
4-trifluoromethyl substitution on the terminal phenyl
ring. Replacement of the fluorine atom with chlorine in
position 3 (derivative 6b) or with a trifluoromethyl
group in position 2 (derivative 6a) resulted in com-
pounds with an inhibition of Aâ42 production compa-
rable to 5h. On the contrary, unsubstituted derivative
6c was almost inactive at the concentration of 100 µM
(Table 2).
a Reagents and conditions: (a) (COCl)2, DCM; (b) NH3, DCM;
(c) (CF3CO)2O, Py, DCM; (d) Bu3SnN3, 100 °C; (e) NH2OH, MeOH,
70 °C; (f) CDI, dioxane, 100 °C.
compound 5f) displaying the best activity (IC50 ) 21
µM). Compound 5f showed also a moderate selectivity
between Aâ42 and Aâ40 inhibition (IC50 ) 21 µM for
Aâ42 versus IC50 ) 46.5 µM for Aâ40 inhibition). The
terphenyl derivative 16d similarly showed good activity
for Aâ42 inhibition (IC50 ) 35.3 µM).
Carboxylic acid group bioisosteres were also consid-
ered, to verify if a better ADME developability profile
could be obtained, especially for the penetration of
compounds in the cerebrospinal fluid. Derivatives bear-
ing the 4-trifluoromethyl substitution or the 3,5-dichlo-
rophenyl substitution on the terminal phenyl ring were
synthesized (Table 3). All carboxylic group bioisosteres
(compounds 21, 18, 19, 23) displayed a lower activity
in Aâ42 inhibition (Table 3) and significantly increased
interactions with P450 enzymes (Table 4). The concen-
tration in the cerebrospinal fluid was evaluated for
compounds 19 and 23 and was not greater than 2%.
Interestingly, two ester derivatives (methyl and tert-
butyl, compounds 24 and 25) were found to selectively
raise Aâ42 (242% and 259% increase in Aâ42 compared
to 45% and 20% inhibition of Aâ40 at 100 µM, respec-
tively), supporting the allosteric nature of the γ-secre-
tase modulation by NSAIDs.18 Indeed, other authors
have shown that simple chemical modifications of the
carboxylic moiety of NSAIDs can generate selective
Aâ42 raising agents.24 A further modification of the
carboxylic moiety with primary amide (compound 26)
led to the loss of selectivity in the Aâ42 inhibition (IC50s
of 64 and 12 µM on Aâ42 and Aâ40, respectively); the
metabolic clearance was very high (193 mL/h), but
penetration in the cerebrospinal fluid was improved
(19% of the plasma levels). However, this last observa-
Replacement of the terminal phenyl ring with other
heterocycles was also considered; while thienyl and
pyridyl derivatives 5k and 5l were only moderately
active, compound 5j bearing the 3-benzothienyl group
displayed good activity, with an IC50 of 83 µM.
The most interesting compounds were evaluated for
their activity toward COX-1 enzyme, P450 interactions
and permeability in the Caco2 cell layer model (Table
2). Benzothienyl derivative 5j and compound 5a bearing
the 3,4-dichloro substitution on the terminal phenyl ring
still showed a significant interaction with COX-1 (68%
and 78% inhibition at the concentration of 100 µM,
respectively), while introduction of bulkier substituents
such as in compounds 5f, 5h, 13a, 16d caused a
significant loss of COX-1 inhibitory activity. Significant
interactions with one or more isoforms of P450 enzyme
were observed for methoxy derivative 5m, cyclohexyl
derivative 5f, dichloro derivative 5a, cyclohexyloxy
derivative 13a, and terphenyl derivative 16d; trifluoro-
methyl derivative 5h showed no interactions at the
concentration of 100 µM. Permeability in the Caco-2
assay proved to be medium to high for all the tested
compounds.
The three most potent compounds (cyclohexyl deriva-
tive 5f, cyclohexyloxy derivative 13a and terphenyl
derivative 16d) were also tested in vivo in a steady-state
model of metabolic clearance. All the derivatives showed

5710 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Peretto et al.
Table 2. Selectivity Assays and ADME Developability Parameters for Selected compounds
Aâ40
Caco-2 clearance CSF/plasma
compound
inhibition
COX-1 inhibition
P450 interactions
absorption (mL/h)
ratio
flurbiprofen nd
IC50 ) 170 nM no interactions @ 100 µM
78% @ 100 µM inhibition >50% @ 100 µM for 2C19, 2C9*
65% @ 100 µM 2C9b
high
high
high
13.9
nda
nd
174
nd
nd
nd
148
85
1.3%
5a
nd
nd
5b
5f
5h
5j
5m
13a
16d
IC50 ) 94 µM
nd
IC50 ) 46 µM
18% @ 100 µM inhibition >50% @ 100 µM for 1A2, 2C19, 2D6; 2C9b medium
1.5%
nd
nd
nd
40% @ 100 µM no interactions @ 100 µM
high
high
high
high
high
nd
68% @ 100 µM 2C9b
nd
nd
IC50 ) 87 µM (2C19)
nd
25% @ 100 µM 20% @ 100 µM inhibition >50% @ 100 µM for 1A2, 3A4; 2C9b
below LLOQc
1.1%
nd
17% @ 100 µM IC50 ) 58 µM (2C19), 74 µM (3A4), 4.7 µM (2C9)
a nd ) not determined. b Activation instead of inhibition was observed due to interference with the assay. c LLOQ ) lower limit of
quantitation.
Table 3. Carboxylic Group Replacement of Flurbiprofen Derivativesa
a nd ) not determined.
Table 4. Selectivity Assays and ADME Developability Parameters for Selected Compounds
Aâ40
inhibition
Caco-2
absorption
clearance
(mL/h)
CSF/plasma
ratio
compound
COX-1 inhibition
P450 interactions
flurbiprofen
5h
21
nd
nd
nd
IC50 ) 170 nM
40% @ 100 µM
9% @ 100 µM
no interactions @ 100 µM
no interactions @ 100 µM
inhibition >50% @ 100 µM
for 1A2, 2C9, 2C19, 2D6, 3A4
inhibition >50% @ 100 µM
for 1A2, 2C9, 2C19, 2D6, 3A4
inhibition >50% @ 100 µM
for 2C9, 2C19, 3A4
nd
high
high
nd
13.9
nda
nd
1.3%
nd
nd
19
23
nd
nd
nd
nd
nd
nd
3.1
5.3
1.7%
2.1%
24
25
26
45% @ 100 µM
20% @ 100 µM
IC50) 12 µM
nd
nd
nd
nd
nd
nd
nd
nd
nd
19%
nd
nd
nd
193
a nd ) not determined.
tion needs confirmation with further series of related
compounds.
inhibitory activity (Table 6), as for compounds 11a, the
alpha-cyclopropyl derivative of flurbiprofen, or for 11c
(the corresponding alpha-methyl derivative 5a displayed
78% COX-1 inhibition at the concentration of 100µM).
The cyclopropyl substitution also seemed to moderately
increase the Aâ42 inhibitory activity (Table 5), as can
be observed comparing the activity results for 11a and
flurbiprofen, 11c and 5a, 11b and 5h. On the contrary,
cyclopropyl substitution did not significantly affect
Substitution pattern at the alpha position of flurbi-
profen was also considered (Table 5), in view of the
reported different activity of the R- and S- enantiomers
of flurbiprofen both on COX-1 activity and on Aâ42
inhibitory activity.19 Introduction of the cyclopropyl
substitution or the dimethyl substitution at the alpha
position caused an almost complete loss of COX-1

Flurbiprofen Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18 5711
Table 5. Combined Substitutions: Terminal Phenyl Ring and Alpha Position
compound
R1
R4, R4′
H, Me
Aâ42 inhibition (@ 100 µM)
IC50 (µM)
flurbiprofen
11a
11b
11c
11d
11e
11f
11g
11h
11i
16a
16b
16c
13b
13c
13d
phenyl
30%
40%
55%
79%
49%
75%
52%
58%
77%
60%
99%
80%
75%
97%
28%
70%
305
nda
111
41
96.3
46
nd
70.6
nd
nd
31.6
33.6
57.9
11
phenyl
cyclopropyl
cyclopropyl
cyclopropyl
cyclopropyl
cyclopropyl
cyclopropyl
Me, Me
4-CF3 phenyl
3,4-dichlorophenyl
3-benzothienyl
4-(4,4-dimethylcyclohexyl)phenyl
4-tetrahydropyran-4-ylphenyl
4-CF3 phenyl
4-cyclohexylphenyl
4-cyclohexylphenyl
Me, Me
cyclopropyl
cyclopropyl
cyclopropyl
cyclopropyl
cyclopropyl
cyclopropyl
cyclopropyl
4-(4-hydroxyphenyl)phenyl
4-(4-hydroxyphenyl)-3-fluorophenyl
4-(4-fluorophenyl)-3-fluorophenyl
4-(3,3,5,5-tetramethylcyclohexyloxy)phenyl
4-(tetrahydropyran-4-yloxy)phenyl
4-(4-trifluoromethylcyclohexyloxy)phenyl
nd
20
a nd ) not determined.
Table 6. Selectivity Assays and ADME Developability Parameters for Selected Compounds
CSF/
plasma
ratio
Aâ40
inhibition
COX-1
inhibition
Caco-2
absorption
clearance
(mL/h)
compound
P450 interactions
flurbiprofen nd
IC50 ) 170 nM no interactions @ 100 µM
high
high
high
high
high
low
13.9
9.8
0.52
1.5
113.6
145.8
1.3%
11a
11b
11c
11d
11e
-1% @ 100 µM 26% @ 100 µM IC50 ) 83 µM (2C9)
22% @ 100 µM <1% @ 100 µM IC50 ) 71 µM (2C19)
14% @ 100 µM <1% @ 100 µM IC50 ) 38 µM (2C9), 25 µM (2C19), 48 µM (3A4)
-5% @ 100 µM <1% @ 100 µM IC50 ) 19 µM (2C9)
IC50 ) 101 µM 17% @ 100 µM IC50 ) 17 µM (2C9), 13 µM (2C19), 5.3 µM (2D6),
84 µM (3A4)
2.3%
1.8%
1.3%
2.6%
below LLOQc
11f
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
39
below LLOQ
11g
11h
11i
nd
nd
47% @ 100 µM 12% @ 100 µM IC50 ) 59 µM (3A4), 28 µM (2C9)
nd
nd
nd
12% @ 100 µM IC50 ) 59 µM (3A4), 2C9b
nd
nd
16a
16b
16c
13b
IC50 ) 409 µM <1% @ 100 µM IC50 ) 89 µM (2C9), 35 µM (2C19), 21.5 µM (2D6) high
nd
nd
IC50 ) 111 µM <1% @ 100 µM IC50 ) 74 µM (2C19), 58 µM (3A4)
32% @ 100 µM 18% @ 100 µM IC50) 7 µM (2C9), 46 µM (2C19),) 38 µM (3A4)
IC50 ) 25.5 µM <1% @ 100 µM IC50) 26 µM (1A2) 15 µM (2C9), 33 µM (2C19),
30 µM (2D6) 27 µM (3A4)
high
low
medium
109
8.1
134
below LLOQ
1.8
below LLOQ
13c
13d
nd
nd
nd
nd
low
70.5
6.6
1.1
0.75
IC50 ) 81 µM
<1% @ 100 µM IC50 ) 17.7 µM (2C9)
a nd ) not determined. b Activation instead of inhibition was observed due to interference with the assay. c LLOQ ) lower limit of
quantitation.
either P450 profile and Caco2 permeability, or metabolic
clearance and brain penetration.
This compound was selective toward COX-1 and showed
a low metabolic clearance, but interactions with P450
cytochrome were significant for three isoforms, with an
IC50 comparable to the Aâ42 inhibition activity. Com-
pound 11b, bearing the 4-trifluoromethyl substitution,
showed good selectivity toward COX-1 and also a better
developability profile (low clearance, only one significant
P450 interaction), but it was less active in Aâ42 inhibi-
tion (IC50 ) 111 µM, while inhibition of Aâ40 was only
22% at the concentration of 100 µM) if compared to 11c.
Benzothienyl derivative 11d displayed no interactions
with P450 cytochrome (except for 2C9 isoform, with IC50
) 19 µM), but only a moderate potency for Aâ42
inhibition and a very high clearance.
In view of these results, we reconsidered cycloalkyl
derivatives trying to improve the activity and develop-
ability profile of compound 5f (Table 1), which was one
of the most active initially synthesized. Modifications
of the terminal cyclohexyl moiety were introduced as
in compounds 11e and 11f, but they did not result in
significant improvement of the clearance profile or of
On the basis of SAR results of the first set of
compounds, a new set of derivatives was planned: the
carboxylic group was maintained, as well as the 3-fluoro
substitution at the central phenyl ring; the alpha-
cyclopropyl moiety was introduced in order to gain
COX-1 selectivity and a moderate activity enhancement,
avoiding the presence of a troublesome stereogenic
center. Functionalization of the terminal phenyl ring
was the key feature to be introduced in order to improve
the Aâ42 inhibitory activity; the substituents identified
and outlined in Table 1 improved the activity but were
not satisfactory for the interactions with P450 cyto-
chrome, metabolic clearance and penetration in the
cerebrospinal fluid.
A second array of compounds bearing combined
substitution was then synthesized and tested (Table 5).
Compound 11c, bearing the 3,4-dichloro substitution,
displayed a good Aâ42 activity; interestingly, no inhibi-
tion of the production of Aâ40 was observed (Table 6).

5712 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Peretto et al.
Table 7. Pharmacokinetic Parameters of Flurbiprofen and Novel Selected Compounds after Intravenous (iv) and Oral (po)
Administration to Ratsa
compound
dose (mg/kg)
AUC0-∞ (µg‚h/mL)
Cmax (µg/mL)
Tmax (h)
t1/2 (h)
CL (mL/h/kg)
F (%)
flurbiprofen
po
iv
po
iv
po
iv
po
iv
7.6
1.5
2.5
0.5
2.5
0.5
0.5
0.2
311
78
49.1
9.1
54.9
14.4
16.2
6.1
0.5
4.9
5.1
78.3
0.3
19.1
0.9
11b
11c
13d
2613
575
435
173
6
4.0
0.08
4.0
0.08
8.0
0.25
32.8
36.2
20.7
29.4
11.8
13.2
90.9
50.2
24.2
2.9
0.3
0.6
11
18.8
a Cmax ) maximum concentration obtained; t1/2 ) elimination phase half-life; CL ) total body clearance; F ) absolute oral bioavailability.
the P450 interactions. Similarly, all the terphenyl
derivatives 16a, 16b, 16c maintained good Aâ42 inhibi-
tion activity but displayed strong P450 interactions.
Optimization of alkoxy analogues of compound 13a
was then considered. While alkoxy moieties introduced
in compounds 13b and 13c resulted in worse develop-
ability profiles or in loss of potency, derivative 13d,
featuring the trifluoromethyl cyclohexyloxy derivative,
maintained a good Aâ42 inhibition activity, good selec-
tivity toward COX-1, only one interaction with P450
cytochrome (2C9 isoform) and low metabolic clearance;
permeability in the Caco-2 assay was low, as was the
cerebrospinal fluid/plasma concentration ratio.
The three most interesting compounds, 13d, 11b and
11c, were further profiled for their selectivity toward
COX-2. At the concentration of 300 µM, both 11b and
11c were found to be completely inactive. At the same
concentration 13d retained a 52% inhibitory activity on
COX-2 but at the concentration of 100 µM the compound
did not inhibit the enzyme.
Compounds 11b and 13d were also tested in the
CBF1-luciferase assay for their activity on Notch-1
cleavage. At the concentration of 100 µM, 11b and 13d
did not significantly affect Notch-1 signaling (2.1 (
15.1% and 9.3 ( 10.8% inhibition, respectively).
Since hepatotoxicity is an uncommon, but potentially
lethal adverse reaction caused by NSAIDs,25 we em-
ployed the MTT assay, that measures mitochondrial
integrity and activity, to estimate hepatoxicity potential
of compounds 11b, 11c and 13d. All compounds showed
low hepatotoxicity (10% at the concentration of 100 µM
Figure 1. Plasma levels of flurbiprofen and selected ana-
for 11b, IC50 ) 284 µM for 11c, and IC50 greater than
logues (11b, 11c and 13d) in the rat after intravenous (panel
1000 µM for 13d).
A) and oral (panel B) administration to rats. Plasma levels
are standardized for a dose of 1 mg/kg.
Pharmacokinetic profiles of compounds 11b, 11c and
13d were then evaluated after single intravenous and
oral administration in rats and compared to that of
flurbiprofen. Plasma concentration profiles standardized
for a 1 mg/kg dose versus time are shown in Figure 1
and main pharmacokinetic parameters are listed in
Table 7. After oral dosing, flurbiprofen appeared to be
quickly absorbed (Tmax ) 30 min) and cleared from
plasma with a half-life of around 5 h. Absolute oral
bioavailability was found to be 78.3% and total body
clearance 19.1 mL/h/kg. After oral administration,
compounds 11b, 11c and 13d appeared to be absorbed
slower than flurbiprofen (Tmax ) 4-8 h). Compounds
were eliminated slowly with half-lives ranging from 12
to 33 h. Total body clearances were also lower than that
of flurbiprofen (0.9, 2.9 and 18.8 mL/h/kg for 11b, 11c
and 13d, respectively). Absolute oral availability of
compounds 11b, 11c and 13d were 91%, 50% and 24%
and were in line with the apparent permeability values
derived by the Caco-2 model.
Selectivity in the inhibition of Aâ42 species production
versus other amyloid peptides was confirmed with
compound 11c, by MALDI-TOF analysis of cell super-
natant during the Aâ42 inhibition assay. More than 20
different Aâ species in cell supernatant were detected
(Figure 2). Compound 11c determined a dose dependent
decrease of Aâ42 and a simultaneous increase of the
Aâ38 species.
The effects of a 7-day treatment (12.5 mg/kg twice-
a-day) of R-flurbiprofen or compounds 13d, 11c and 26
on Aâ40 and Aâ42 plasma concentrations of transgenic
mice are summarized in Figure 3. Compared to baseline,
R-flurbiprofen did not decrease significantly either Aâ40
or Aâ42. Conversely, all the new compounds signifi-
cantly (p < 0.05) reduced Aâ42 concentrations with
mean differences compared to controls being -40% after
13d, -29% after 11c and -40% after 26. The corre-
sponding values for Aâ40 were -35%, -4% and -26%,

Flurbiprofen Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18 5713
Figure 2. MALDI-TOF spectra of supernatants of immunoprecipitated H4-APP695NL cell cultures incubated for 42 h in absence
(upper panel) and in the presence (lower panel) of 11c (50 µM).
Figure 3. Effects of 7-day dosing (12.5 mg/kg twice-a-day) of R-flurbiprofen or compounds 13d, 11c or 26 on Aâ40 (black columns)
and Aâ42 (white columns) concentrations in plasma of Tg2576 transgenic mice 3 h after the final dose. Columns represent mean
(( SEM) Aâ concentrations expressed as percentage of the vehicle control. *p < 0.05 compared to baseline.
respectively. Brain concentrations of Aâ were not sig-
nificantly affected by the short-term treatment of the
compounds (data not shown). In Study 1, employing
mice with 7-8 months of age (8 animals per treatment
group), there were no deaths after vehicle, 2 deaths after
R-flurbiprofen and 1 death after 13d. At Day 8, animals
treated with 13d had a significant weight loss (-2.8 (
0.4 g) compared to vehicle (-0.9 ( 0.7 g). In Study 2,
employing mice of 3-4 months of age (12 animals per
treatment group), there were four deaths after vehicle,
three deaths after 11c and three deaths after 26. There
were no significant differences between treatment groups
on body weight gain.
toxicity. The improvement in potency was reached
mainly with appropriate substitutions on the distal
phenyl ring of flurbiprofen, while selectivity toward
COX-1 and COX-2 was achieved by alpha-cyclopropyl
substitution.
Flurbiprofen is exclusively metabolized by CYP2C9
isoenzyme.26 We found that both racemic and R-flurbi-
profen interact bimodally with CYP2C9 with activation
at lower concentrations and inhibition at higher con-
centrations. Several flurbiprofen derivatives synthesized
during this project showed the same behavior. While
flurbiprofen does not interact with the other cytochrome
isoenzymes, many analogues showed marked inhibitory
activity, especially on CYP2C19. Structure-activity
relationships of these interactions were quite complex,
with aromatic substitution on the distal phenyl ring and
replacement of the carboxylic moiety with the tetrazole
group producing the major interferences with CYP
isoenzymes.
Discussion
The flurbiprofen analogues we synthesized appear
more potent than flurbiprofen and almost completely
inactive on COX-activity, especially regarding COX-1
activity, the most dangerous one for the gastrointestinal

5714 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Peretto et al.
It has been shown that different chemical classes of
prototypical γ-secretase inhibitors do not discriminate
between APP and Notch-1 cleavage sites.27 In vitro
studies showed that application of a peptidomimetic
γ-secretase inhibitor to fetal thymus organ cultures
results in inhibition of T cell development in a manner
consistent with loss or reduction of Notch-1 function.28
More recently, an in vivo study in TgCRND8 APP
transgenic mice showed that LY-411,575, a nonpeptidic
γ-secretase inhibitor, produces marked effects on lym-
phocyte development and on the intestine tissue mor-
phology.29 In addition, it has been shown that a 5-day
treatment with γ-secretase inhibitors with dibenza-
zepine and benzodiazepine structure causes dose-de-
pendent intestinal goblet cell metaplasia in rats.30 In
cell-based assays, sulindac sulfide has been shown not
altering either ꢀ or S3 cleavage sites.5 Additional in vitro
studies demonstrated that sulindac sulfide and ibupro-
fen preserve release of APP, Notch-1 and ErbB-4 intra-
cellular domains.6 Although these observations were
confirmed by other groups, at high concentrations of
NSAIDs an inhibition of the S3 cleavage site of Notch-1
is observed, reminiscent of classical γ-secretase inhibi-
tion.16 Thus, it appears that the separation of the
inhibitory effects on the various cleavage sites is con-
centration dependent leaving a “window of modulation”.
Our preliminary results indicated that compounds 11b
and 11c evaluated at concentrations (100 µM) near or
above the IC50 on Aâ42 secretion do not significantly
inhibit Notch-1 cleavage. In vivo studies are required
to confirm the lack of biological effects of these com-
pounds on Notch-1 processing.
The two main metabolites of flurbiprofen are the 4′-
hydroxy and 3′,4′-dihydroxy derivatives.31 Indeed, flur-
biprofen analogues with para and meta positions of the
distal phenyl ring protected from metabolic oxidation
showed a much lower total body clearance than that of
flurbiprofen. Successful protection of 3′ and 4′ positions
was reached with halogens or trifluoromethyl moiety
(11b and 11c). Interestingly, these compounds showed
higher inhibitory potency on Aâ42 secretion than flur-
biprofen. Total body clearances calculated with a single
blood sample taken at steady-state after continuous
subcutaneous infusion with osmotic mini-pumps were
found in good agreement with values calculated for
selected compounds (flurbiprofen, 11b, 11c, 13d) in
traditional experiments with single intravenous doses
and drug concentrations measured at multiple time
points. After intravenous or oral administration to rats,
11b and 11c were eliminated slowly with plasma half-
life of 20-36 h. The low total body clearance of 11b and
11c is probably due to their limited oxidative metabo-
lism, the 3′ and 4′ positions being protected from
hydroxylation.
new compounds was in the range of 1-2% of the
corresponding plasma levels. Interestingly, the primary
amide 26 appeared to penetrate more efficiently the
blood brain barrier with cerebrospinal fluid levels
reaching concentrations equal to 19% of those in plasma.
This suggests that this modification of the carboxylic
moiety could be a good strategy to improve CNS
penetration of these compounds. Further studies are
needed to address this important issue.
Studies in transgenic mice expressing the Swedish
mutated form of APP showed that a 7-day oral treat-
ment with compounds 13d, 11c, and 26 decreased
significantly Aâ42 plasma concentrations and to a lesser
extent Aâ40. Compound 11c appeared to be the most
selective inhibitor on the two Aâ peptides (-30% on
Aâ42 and -4% on Aâ40, compared to controls) and this
agreed with its in vitro activity profile (IC50 ) 41 µM
on Aâ42 and 14% inhibition of Aâ40 at 100 µM). Short-
term treatment with the new compounds, as well as
R-flurbiprofen, was not able to significantly modify Aâ
levels in the brain. A lack of central effects was also
recently observed with a 3-day treatment of racemic
flurbiprofen in Tg2576 transgenic mice.34 This could be
due to insufficient brain concentrations reached by these
compounds not compatible with their micromolar po-
tency. On the other hand, long-term treatments (3-6
months) with ibuprofen,8,9,11 indomethacin12 or flurbi-
profen derivatives10,14 have demonstrated a significant
reduction in brain Aâ pathology in transgenic mice.
Different reasons may explain the different efficacy of
selected NSAIDs on central Aâ after short-term and
long-term treatment. First, it could be that NSAIDs
accumulate within the neuronal membrane where
γ-secretase is localized. Alternatively, it is possible that
NSAIDs do not act centrally but instead reduce periph-
eral Aâ42 levels, which results in enhanced efflux of
Aâ42 from the brain.20 Compared to vehicle, the new
compounds appeared to be relatively well tolerated,
although no formal biochemistry or histopathological
examinations were done. Mortality rate was similar in
vehicle-treated animals and in animals treated with the
new compounds. However, at Day 8 animals treated
with 13d had a significant body weight loss of about 2
g compared to controls. Conversely, mean body weight
of animals treated with 11c and 26 were similar to that
of vehicle-treated animals along the entire study period.
In a 2-week toxicity study in mice, no behavioral,
laboratory or histopathological toxic effects were ob-
served with 11b at doses of 30 or 100 mg/kg/day.
Conclusions
Novel flurbiprofen analogues have been identified
with potent and selective inhibitory activity on Aâ42.35
The new Aâ42 lowering agents appear to be almost
devoid of anti-COX-1 and COX-2 activity. Compared to
flurbiprofen, the new compounds show improved po-
tency on Aâ42 (up to 27 times) and improved total body
clearance (up to 29 times). In rats, compounds 13d, 11b,
11c showed good oral bioavailability and long elimina-
tion half-life. Short-term studies in a transgenic mice
showed that compounds 13d, 11c and 26 decreased
significantly plasma Aâ42 concentrations. These new
flurbiprofen analogues are worthy of being tested after
long-term administration in animal models of AD to
Penetration in cerebrospinal fluid of flurbiprofen was
low with concentrations representing only 1.5% of the
corresponding plasma levels. This is probably linked to
the formation of a covalent adduct with proteins via
reactive acyl glucuronide metabolites that prevent from
crossing of the blood-brain barrier.32 Indeed the in vivo
plasma unbound percent fraction of flurbiprofen in rats
is 1.4%,33 a value in very good agreement with the
percent penetration in cerebrospinal fluid we found in
our studies. Penetration in cerebrospinal fluid for the

Flurbiprofen Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18 5715
mmol) in water (75 mL), keeping the temperature at 0 °C. The
reaction mixture was allowed to warm to room temperature
and stirred for 3 h, then extracted with ethyl acetate. The
combined layers were washed in sequence with 10% Na2S2O3
and brine, then dried over Na2SO4 and concentrated under
vacuum to afford 3 as a brown oil (37 g, 88%), which was used
without further purification in the next step. 1H NMR
(CDCl3): δ 7.70(dd. 1H); 7.14(dd, 1H); 6.94(dd, 1H); 4.23(q, 2H);
4.22(q, 2H); 1.83(s, 3H); 1.26(t, 6H). MS (EI; TSQ 700; source
180 °C; 70 V; 200 µA): 394 (M+); 321; 293; 276; 247; 166. Anal.
(C14H16FIO4) C, H, F, I.
2-(3-Fluoro-4-iodophenyl)propionic Acid (4). To a solu-
tion of 3 (37 g, 94 mmol) in ethanol (130 mL), 2 N NaOH (130
mL) was added. The mixture was heated to 100 °C for 8 h,
then concentrated under vacuum. The residue was dissolved
in water (300 mL) and washed with ethyl ether (2 × 100 mL),
then the aqueous layer was acidified (pH 2) with HCl and
extracted with ethyl acetate (3 × 150 mL). The organic extracts
were washed with brine, dried over Na2SO4 and concentrated
to afford a brown oil (27 g). Chromatography on silica gel
(methylene chloride) afforded a yellow solid that crystallized
from methanol/water 1:3 yielding 4 as a white solid (16 g, 58%).
1H NMR (CDCl3): δ 7.69(dd, 1H); 7.05(dd, 1H); 6.87(dd, 1H);
3.71(q, 1H); 1.50(d, 3H). MS (EI; TSQ 700; source 180 °C; 70
V; 200 µA): 294 (M+); 249. Anal. (C9H8FIO2) C, H, F, I.
1-Bromo-4-bromomethyl-2-fluorobenzene (7). To a so-
lution of 1-bromo-2-fluoro-4-methylbenzene (10.0 g, 53 mmol)
in carbon tetrachloride (100 mL) NBS (14.0 g, 80 mmol) was
added. The mixture was heated to 80 °C, then dibenzoyl
peroxide (0.1 g, 0.4 mmol) was added. The mixture was stirred
at this temperature for 1 h then cooled to room temperature
and washed in sequence with water and brine, dried over Na2-
SO4 and concentrated to afford an oil (16 g) that was used
without further purification in the next step.
(4-Bromo-3-fluorophenyl)acetonitrile (8). A solution of
crude 7 (16 g) and NaCN (2 g, 0.04 mol) in ethanol (100 mL)
was heated to 80 °C for 2 h, then cooled to room temperature
and concentrated under vacuum. The residue was suspended
in water and extracted with ethyl acetate. The organic solution
was washed with brine, dried over Na2SO4 and concentrated
to afford a brown oil. Chromatography over silica gel (hexane/
ethyl ether 7:3) yielded 8 as a light brown solid (5 g, 44% over
two steps). 1H NMR (CDCl3): δ 7.57(dd, 1H); 7.13(dd, 1H);
7.02(dd, 1H); 3.72(s, 2H). MS (EI; TSQ 700; source 180 °C; 70
V; 200 µA): 213 (M+); 134; 114; 107; 87. Anal. (C8H5BrFN) C,
H, N, Br, F.
verify their ability to ameliorate brain Aâ pathology and
could become viable clinical candidates once the ongoing
preclinical development activities are successfully com-
pleted.
Experimental Section
Chemistry. 1H NMR spectra were recorded on a Bruker
ARX 300 (300 MHz); chemical shifts are reported downfield
in parts per million (ppm) relative to TMS utilizing the solvent
peaks as the reference. EI Mass spectra analysis was recorded
on a Thermo Finnigan TSQ700 spectrometer; ESI-LCMS
analysis was performed on a Phenomex Luna C-18, 3 µm, 4.6
× 50 mm column, with an AQA Thermo Finnigan single
quadrupole instrument or with a Waters Micromass ZQ2000
instrument; high performance liquid chromatography (HPLC)
analysis was performed on a Shimadzu SCL-10A equipped
with an SIL-10AD injector and an SPD-M10A detector nor-
mally operating in a 200-360 nm range, with a Waters
Symmetry C-18, 3.5 µm, 4.6 × 75 mm column, using a 10 min
gradient of 0-100% solvent B, where solvent A is 90:10:0.05
CH3CN-H2O-TFA and solvent B is 90:10:0.05 H2O-CH3CN-
TFA; preparative HPLC purifications were performed on a
Waters SymmetryPrep C-18, using a 20 min gradient of
0-100% solvent B, where solvent A is 99.9:0.1 H2O-TFA and
solvent B is 99.9:0.1 CH3CN-TFA; reactions were monitored
by TLC using 0.25 mm Merck silica gel plates (60 F254); column
chromatography was performed on Merck silica gel 60 (particle
size 0.063-0.2 mm); flash chromatography was conducted
using a Biotage-Quad3 apparatus and prepacked silica gel
columns (KP-SIL, particle size 32-60 µm). Solid liquid
extraction (SLE) cartridges were purchased from Varian
(“Chem Elut”, 3 mL, unbuffered, part no. 12198003). Anhy-
drous solvents were purchased from Aldrich and used as
received. “Brine” refers to a saturated aqueous solution of
NaCl. Unless otherwise specified, solutions of common inor-
ganic salts used in workups are aqueous solutions. Common
reagents and solvents are abbreviated as follows: CDI, car-
bonyldiimidazole; DCM, dichloromethane; DEAD, diethyl azodi-
carboxylate; DME, 1,2-dimethoxyethane; DMF, N,N-dimeth-
ylformamide; DMSO, dimethyl sulfoxide; NBS, N-bromosuccin-
immide; TBAB, tetrabutylammonium bromide; TFA, trifluo-
roacetic acid; THF, tetrahydrofuran.
2-(3-Fluoro-4-nitrophenyl)-2-methylmalonic Acid Di-
ethyl Ester (1). Anhydrous DMSO (100 mL) was added over
30 min to a 60% dispersion of sodium hydride in mineral oil
(6.12 g, 153 mmol), then the mixture was cooled to 15 °C and
a solution of diethyl-2-methylmalonate (24.0 g, 138 mmol) in
anhydrous DMSO (100 mL) was added dropwise over 30 min.
A solution of 2,4-difluoronitrobenzene (20.3 g 128 mmol) in
anhydrous DMSO (200 mL) was added dropwise over 60 min
to the resulting mixture. After 30 min, the reaction mixture
was poured into crushed ice (5 kg) and extracted with ethyl
acetate (150 mL × 3). The organic solution was washed with
brine, dried over Na2SO4 and concentrated to afford a crude
oil (40.0 g). Chromatography on silica gel (hexane, then
hexane/ethyl acetate 9:1) afforded 1 as yellow oil (35.0 g, 87%).
1H NMR (CDCl3): δ 8.04(dd, 1H); 7.39(dd, 1H); 7.33(ddd, 1H);
4.26(q, 2H); 4.25(q, 2H); 1.88(s, 3H); 1.27(t, 6H). MS (EI; TSQ
700; source 180 °C; 70 V; 200 µA): 313 (M+); 268; 241; 213;
195; 184; 166. Anal. (C14H16FNO6) C, H, N, F.
1-(4-Bromo-3-fluorophenyl)cyclopropanecarbo-
nitrile (9b). To a solution of 8 (5 g, 23 mmol) and 1,2-
dibromoethane (3 mL, 35 mmol) in toluene (20 mL), 50%
NaOH (20 mL) and TBAB (1.6 g, 5 mmol) were added. The
mixture was vigorously stirred at room temperature for 4 h,
then diluted with water and extracted with ethyl acetate. The
organic solution was washed with 1 N HCl, then with brine,
dried over Na2SO4 and concentrated to afford a brown solid
(5.7 g). Chromatography on silica gel (hexane/ethyl ether 1:1)
1
afforded 9b as yellow solid (5.1 g, 91%). H NMR (CDCl3): δ
7.53(dd, 1H); 7.03(dd, 1H); 6.98(dd, 1H); 1.78(m, 2H); 1.40(m
2H). Anal. (C10H7BrFN) C, H, N, Br, F. Compound 9a was
obtained following the same procedure and employing MeI
instead of 1,2-dibromoethane.
1-(4-Bromo-3-fluorophenyl)cyclopropanecarboxylic
Acid (10b). To a solution of 9b (5.1 g, 21 mmol) in methanol
(10 mL) 35% NaOH (40 mL) was added and the mixture was
heated to 100 °C for 8 h. After cooling to room temperature,
the mixture was acidified (pH 2) with 2 N HCl, and the
precipitate formed was filtered, washed with water and
redissolved in 5% NaHCO3. Insoluble materials were filtered
off and the solution acidified with 2 N HCl. The precipitate
was filtered, washed with water and dried under reduced
2-(4-Amino-3-fluorophenyl)-2-methylmalonic Acid Di-
ethyl Ester (2). To a solution of 1 (35 g, 112 mmol) in ethanol
(500 mL), 10% Pd on charcoal (2.0 g) was added. The mixture
was hydrogenated for 1 h under atmospheric pressure, then
filtered through a Celite pad and concentrated to afford 2 as
a colorless oil (30.2 g, 95%). 1H NMR (CDCl3): δ 7.06(dd, 1H);
6.95(ddd, 1H); 6.72(dd, 1H); 4.22(q, 2H); 4.21(q, 2H); 1.80(s,
3H); 1.25(t, 6H). Anal. (C14H18FNO4) C, H, N, F.
2-(3-Fluoro-4-iodophenyl)-2-methylmalonic Acid Di-
ethyl Ester (3). To a solution of 2 (30 g, 106 mmol) in 6 N
HCl (180 mL) cooled to 0 °C, a solution of NaNO2 (7.3 g, 106
mmol) in water (40 mL) was added dropwise. The resulting
solution was added dropwise to a solution of KI (73 g, 440
1
pressure to afford 10b as a white solid (4.2 g, 76%). H NMR
(DMSO-d6): δ 7.60(dd, 1H); 7.34(dd, 1H); 7.13(dd, 1H); 1.45-
(m, 2H); 1.18(m, 2H). Anal. (C10H8BrFO2) C, H, Br, F.
Compound 10a was obtained starting from 9a and following
the same procedure.

5716 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Peretto et al.
General Procedure for Preparing Biaryl Compounds
5a-o. Reactions were performed in 2 mL vessels under argon
atmosphere employing a Myriad Personal Synthesizer. 2 M
Na2CO3 (0.75 mL) and 3 mL of a solution prepared dissolving
4 (4.7 g, 16 mmol) and Pd(PPh3)4 (0.92 g, 0.8 mmol) in DME
(70 mL) were added in sequence to 1 mmol of the suitable
arylboronic acid. Reaction vessels were maintained at 80 °C
for 15-24 h under stirring. Workup was performed by adding
2 N HCl (1.5 mL) and ethyl acetate (3 mL) and removing the
aqueous phase with SLE cartridges. The organic phase was
concentrated under vacuum and crude materials purified by
flash chromatography (DCM/MeOH 97:3) or preparative HPLC.
Products were obtained on average 50% isolated yield and
purity greater than 95% (HPLC-UV 215 nm).
was heated to 140 °C for 2 h, then cooled and diluted with 1
N HCl (100 mL) while decanting metallic Ni. The precipitate
was filtered, washed with water and redissolved in ethyl
acetate, then dried over Na2SO4, and concentrated to yield a
crude solid (0.84 g) that was employed without further
purification.
General Procedure for Preparing Compounds 15a-
c. To a solution of 14 or 4 (2 mmol), the suitable 4-bromoaryl-
boronic acid (2.3 mmol) and Pd(PPh3)4 (0.1 mmol) in DME (20
mL) in 2 M Na2CO3 (5 mL) were added and the mixture was
heated to 80 °C for 2 h. After extractions with ethyl acetate,
washings with brine and drying over Na2SO4, the residue was
chromatographed on silica gel (hexane/ethyl acetate 7:3) to
afford the expected products in 60% average yield.
General Procedure for Preparing Compounds 16a-
d. A suspension of the suitable acid 15a-c (1.0 mmol), the
suitable arylboronic acid (1.5 mmol), TBAB (1.5 mmol) and
Pd(OAc)2 (0.05 mmol) in 2 M K2CO3 was heated to 130 °C in
a sealed Pyrex tube for 1 h. 2 N HCl and ethyl acetate were
added and the mixture was filtered through a Celite pad. The
organic layer was separated, washed with brine, dried over
Na2SO4 and concentrated under vacuum. Crude materials
were purified by preparative HPLC. Final products were
obtained in 30% average yield and purity greater than 95%
(HPLC-UV 215 nm).
Compound 5a. MS (EI; TSQ 700; source 180 °C; 70 V; 200
µA): 312 (M+.); 267 (M+ - CO2).
1
Compound 5b. H NMR (DMSO): δ (DMSO): 12.45(s br,
1H, COOH); 7.65(dd, J ) 1.9, 1.9 Hz, 1H, 4′); 7.60(dd, J ) 1.9,
1.9 Hz, 2H, 2′ and 6′); 7,57(dd, J ) 8.5, 8.5 Hz, 1H, 6); 7,27-
(dd, J ) 12.3, 1.6 Hz, 1H, 3); 7,25(dd, J ) 8.5, 1.6 Hz, 1H,
5);3.80(q, J ) 7.2 Hz, 1H, CH3-CH); 1.41(d, J ) 7.2 Hz, 3H,
CH3-CH). MS (EI; TSQ 700; source 180 °C; 70 V; 200 µA):
312 (M+); 267 (M+ - CO2). Anal. (C15H11Cl2FO2) C, H, Cl,
F.
General Procedure for Preparing Biaryl Compounds
11a-h. A suspension of 10a or 10b (0.39 mmol), the appropri-
ate arylboronic acid (0.47 mmol), TBAB (125 mg, 0.39 mmol)
and Pd(OAc)2 (10 mg) in 2 M Na2CO3 (1 mL) was heated to
120 °C in a sealed tube for 1 h. 2 N HCl and ethyl acetate
were added and the mixture was filtered through a Celite pad.
The organic layer was separated, washed with brine, dried over
Na2SO4 and concentrated under vacuum. Crude materials
were purified by flash chromatography on silica gel (hexane/
ethyl acetate 8:2). Final products were obtained in average
70% isolated yield and purity greater than 95% (HPLC-UV
215 nm).
Compound 16a. 1H NMR (DMSO-d6): δ 12.48(s br, 1H,
COOH); 9.55(s, 1H, OH); 7.69(d, 2H, CH-Ar); 7.61-7.45(m, 5H,
CH-Ar); 7.28(m, 1H, CH-Ar); 7.25(s, 1H, CH-Ar); 6.87(d, 2H,
CH-Ar); 1.48(m, 2H, CH2-cyclopr.); 1.23(m, 2H, CH2-cyclopr.).
MS (EI; TSQ 700; source 180 °C; 70 V; 200 µA): 348 (M+);
302. Anal. (C22H17FO3) C, H, F.
2-(2-Fluoro-4′-trifluoromethylbiphenyl-4-yl)propio-
nyl Chloride (17). To a solution of 5h (1.5 g, 4.8 mmol) in
DCM (50 mL), oxalyl chloride (1.7 mL, 19.2 mmol) was added
dropwise and the solution was stirred overnight at room
temperature. Concentration yielded 17 as a yellow oil (1.6 g)
that was employed without further purification.
Compound 11a. 1H NMR (DMSO-d6): δ 12.41(s br, 1H,
COOH); 7.56-7.35(m, 6H, CH-Ar); 7.27(m, 1H, CH-Ar); 7.24-
(s, 1H, CH-Ar); 1.48(m, 2H, CH2-cyclopr.); 1.22(m, 2H, CH2-
cyclopr.). MS (EI; TSQ 700; source 180 °C; 70 V; 200 µA): 256
(M+); 210; 196. Anal. (C16H13FO2) C, H, F.
General Procedure for Preparing Alkoxy Derivatives
13a-d. 5o or 12 were submitted to esterification by refluxing
overnight a solution of the compound (3.8 mmol) and 98%
sulfuric acid (1 mL) in methanol (50 mL). After extractions
with ethyl acetate, washings with brine and drying over Na2-
SO4, the methyl ester was obtained as a white solid and
employed without further purification.
2-(2-Fluoro-4′-trifluoromethylbiphenyl-4-yl)-N-hydroxy-
propionamide (18). Triethylamine (0.60 mL, 4.9 mmol) was
added dropwise to a suspension of hydroxylamine hydrochlo-
ride (240 mg, 3.4 mmol) in DCM (5 mL). The mixture was
stirred for 30 min at room temperature, then the insoluble salt
was filtered off. To the clear solution, cooled to 0 °C, a solution
of 17 (200 mg, 0.60 mmol) in DCM (1 mL) was added dropwise.
After 1 h the mixture was extracted with ethyl acetate, washed
with brine and dried over Na2SO4.The crude material was
chromatographed on silica gel (DCM/MeOH 95:5) to afford 18
1
as a white solid (60 mg, 30%). H NMR (DMSO-d6): δ 10.68-
(s, 1H); 8.83(s, 1H); 7.83(d, 2H); 7.76(d, 2H); 7.54(dd, 1H);
7.32-7.23(m, 2H); 3.53(q, 1H); 1.39(d, 3H). MS (EI; TSQ 700;
source 180 °C; 70 V; 200 µA): 327(M+); 267. Anal. (C16H13F4-
NO2) C, H, N, F.
The methyl ester (3.6 mmol), triphenylphosphine (1.15 g,
4.3 mmol) and 3.6 mmol of the suitable alcohol were dissolved
in anhydrous THF (10 mL) under inert atmosphere. To this
mixture, a solution of DEAD (0.8 mL, 4.3 mmol) in THF (5
mL) was added dropwise. The reaction mixture was heated to
65 °C overnight, then diluted with ethyl acetate, washed in
sequence with 1 N HCl, 1N NaOH and brine, then dried over
Na2SO4 and concentrated to dryness. The residue was chro-
matographed on silica gel (hexane/ethyl acetate 95:5 to 90:
10) and then submitted to hydrolysis as follows: 1.6 mmol
were dissolved in a mixture of THF and MeOH 3:1 (15 mL),
then 2 N KOH (5 mL) was added and the solution was refluxed
overnight. Extractions with ethyl acetate, washings with brine
and drying over Na2SO4 yielded a crude material that was
purified by crystallization from ethanol to afford the expected
compounds in 25% average yield (three steps).
N-[2-(2-Fluoro-4′-trifluoromethylbiphenyl-4-yl)propio-
nyl]methanesulfonamide (19). To a solution of methane-
sulfonamide (120 mg, 1.26 mmol) and Et3N (0.2 mL, 1.64
mmol) in DCM (5 mL), a solution of 17 (200 mg, 0.60 mmol)
in DCM (1 mL) was added dropwise. After stirring for 1 h,
the mixture was extracted with ethyl acetate, washed with
brine and dried over Na2SO4. The crude material crystallized
from MeOH to afford 19 (85 mg, 36%) as a white solid. 1H NMR
(DMSO-d6): δ 11.93(s br, 1H); 7.84(d, 2H); 7.78(d, 2H); 7.59-
(dd, 1H); 7.32-7.24(m, 2H); 3.84(q, 1H); 3.24(s, 3H); 1.42(d,
3H). MS (EI; TSQ 700; source 180 °C; 70 V; 200 µA): 389-
(M+); 294; 267. Anal. (C17H15F4NO3S) C, H, N, F, S.
Compound 13a. 1H NMR (DMSO-d6): δ 11.80(s br, 1H,
COOH); 7.47-7.35(m, 3H, CH-Ar); 7.20-7.13(m, 2H, CH-Ar);
7.01(d, 2H, CH-Ar); 4.38(m, 1H, CH-cyclohex.); 3.63(q, 1H, CH-
benzylic); 1.95(m, 2H, CH2-cyclohex.); 1.73(m, 2H, CH2-cyclo-
hex.); 1.58-1.22(m, 6H, CH2-cyclohex.); 1.35(d, 3H, CH3). MS
(EI; TSQ 700; source 180 °C; 70 V; 200 µA): 342 (M+); 260;
215. Anal. (C21H23FO3) C, H, F.
1-(3-Fluoro-4-iodophenyl)cyclopropanecarboxylic Acid
(14). A suspension of 10b (1.0 g, 3.9 mmol), KI (6.5 g, 39 mmol),
Ni (2.5 g, 40 mmol) and I2 (50 mg, 0.2 mmol) in DMF (10 mL)
2-(2-Fluoro-4′-trifluoromethylbiphenyl-4-yl)propioni-
trile (20). A solution of 17 (1.3 g, 3.9 mmol) in DCM (10 mL)
was added dropwise to a saturated solution of gaseous am-
monia in DCM (100 mL). The mixture was stirred for 30 min
at room temperature, then concentrated, the residue washed
with water (3 × 20 mL) and desiccated to afford the corre-
sponding primary amide as white solid. The raw product was
dissolved in dioxane in 1,4-dioxane (15 mL) and pyridine (5
mL, 62 mmol) was added; the resulting solution was cooled to
0 °C and trifluoroacetic anhydride (5 mL, 35 mmol) was added

Flurbiprofen Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18 5717
1
dropwise. The reaction mixture was allowed to warm to room
temperature and stirred for 4 h. The mixture was then poured
into crushed ice (150 g) and extracted with ethyl acetate. The
organic solution was washed in sequence with 1 N HCl, 10%
NaHCO3 and brine, then dried over Na2SO4 and concentrated
to afford 20 as a yellow solid (0.9 g, 95%).
white solid (54 mg, 28%). H NMR (200 MHz, DMSO-d6 T )
303 K): 7.73-7.50 (m, 4H, Ar-H); 7.50-7.38 (br, 1H, NH);
7.33-7.18 (M, 2H, Ar-H); 6.98-6.81 (br, 1H, NH); 3.76-3.54
(q, J ) 7.05 Hz, 1H, CH); 1.42-1.20 (d, J ) 7.05 Hz, 3H, CH3).
MS (EI; TSQ 700; source 180 °C; 70 V; 200 µA): 334 (M+Na
+)
5-[1-(2-Fluoro-4′-trifluoromethylbiphenyl-4-yl)ethyl]-
2H-tetrazole (21). A suspension of 20 (0.5 g, 1.7 mmol) in
azido tributyltin (3.4 g, 10 mmol; caution: hazardous material)
was heated to 100 °C for 4 h, then cooled to room temperature
and quenched with a 1:1 mixture of MeOH and 37% HCl (10
mL). The mixture was stirred for 2 h at room temperature,
then diluted with water and extracted with ethyl acetate. The
organic solution was washed with brine, dried over Na2SO4
and concentrated to afford a residue that was washed with
ether and desiccated to afford 21 as a white solid (0.49 g, 86%).
1H NMR (CDCl3): δ 15.46(s br, 1H); 7.67(d, 2H); 7.60(d, 2H);
7.37(dd, 1H); 7.20-7.12(m, 2H); 4.52(q, 1H); 1.80(d, 3H). MS
(EI; TSQ 700; source 180 °C; 70 V; 200 µA): 336(M+); 267.
Anal. (C16H12F4N4) C, H, N, F.
2-(2-Fluoro-4′-trifluoromethylbiphenyl-4-yl)-N-hydroxy-
propionamidine (22). A suspension of 20 (250 mg, 0.85
mmol), hydroxylamine hydrochloride (65 mg, 0.94 mmol) and
NaHCO3 (80 mg, 0.95 mmol) in MeOH (2 mL) was refluxed
for 4 h, then concentrated and the residue was extracted with
ethyl acetate, washed with brine and dried over Na2SO4 to
afford a crude material (205 mg, 74%) that was employed
without further purification.
Biology. Aâ42 and Aâ40 Assay. Aâ42 and Aâ40 were
measured in the culture medium of H4-APP695NL cells, a
human neuroglioma cell line carrying the double Swedish
mutation of the APP695 isoform (K595N/M596L). Cells were
seeded onto 24-well plates (2 × 105 cell/well) and allowed to
grow to confluence for 24 h, in 5% CO2/95% air in a humidified
atmosphere. Increasing concentrations of the compounds were
added to the cells overnight in a final volume of 0.5 mL. DAPT
(N-[N-(3,5-difluorophenacetyl)-L-alanyl]-S-phenylglycine tert-
butyl ester) and flurbiprofen were used as positive controls at
1 µM and 100 µM final concentrations, respectively. DMSO
(1%) was used as negative control. At the end of the incubation,
100 µL of supernatants were removed and treated with a
biotinylated mouse monoclonal antibody (4G8, Signet Labo-
ratories Inc., Dedham, MA), specifically recognizing the 17-
24 amino acid region of Aâ and two rabbit polyclonal antibodies
(C-term 42 and C-term 40, Biosource International Inc.,
Camarillo, CA), specifically recognizing the C-terminus of Aâ42
and Aâ40, respectively. Antigen-antibody complexes were
recognized by TAG-donkey anti-rabbit IgG (Jackson Immu-
noResearch Laboratories, Soham, UK). Streptavidin-coated
magnetic beads captured the complexes and the signals were
read by an electrochemiluminescence instrument (Origen M8
Analyzer, BioVeris Corporation, Gaithersburg, MD).
Cytotoxicity in Human Neuroglioma Cells. The cyto-
toxicity potential of test compounds was assessed in the same
cells of the Aâ assay (H4-APP695NL) with the MTT assay.
3-[4,5-Dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide
(MTT) is a soluble pale yellow salt that is reduced by
mitochondrial succinate dehydrogenase to form an insoluble
dark blue formazan product to which the cell membrane is
impermeable. The ability of cells to reduce MTT provides an
indication of mitochondria integrity and activity and it may
be interpreted as a measure of viability and/or cell number.
After medium removal for Aâ42 and Aâ40 determination, cells
were incubated for 3 h with 500 µL culture medium containing
0.5 mg/mL MTT, at 37 °C, 5% CO2 and saturated humidity.
After removal of the medium, 500 µL of 100% DMSO were
added to each well. The amount of formed formazan was
determined reading the samples at 570 nm (background 630
nm) using a microplater reader (model 450, Biorad, Hercules,
CA).
MALDI-TOF Analysis of Aâ Species. H4-APP695NL cells
were treated with flurbiprofen (250 µM) or DAPT (1 µM) or
11c (50 and 100 µM) or DMSO (1%) and then incubated in 5%
CO2 at 37 °C for different periods (18 and 42 h). The
conditioned media were harvested and centrifuged (10 000 rpm
for 5 min). The supernatants (0.5 mL) of centrifuged condi-
tioned media were collected and treated for 18 h at 4 °C with
4 µg of monoclonal antibody selective for the 17-24 region of
Aâ (4G8, Signet Laboratories Inc., Dedham, MA). Precipitation
of the Aâ-4G8 complexes was obtained using 30 µL of protein
G/A-agarose beads (SantaCruz Biotechnology Inc., Santa Cruz,
CA) left for 3 h at 4 °C. Protein complexes were separated by
centrifugation at 10 000 rpm for 2 min. The agarose beads were
washed twice with ice-cold immunoprecipitation buffer (0.1%
n-octyl glucoside, 140 mM NaCl, 0.025% sodium azide and 10
nM Tris-HCl, pH ) 8) and once with 10 mM Tris-HCl, pH )
8, containing 0.025% sodium azide. Proteins were then ex-
tracted with 3 µL of saturated solution of R-cyano-4-hydroxy-
cinnamic acid in TFA/H2O/CH3CN (1:20:20). One µL of the
resulting solution was deposited on a stainless steel sample
holder and allowed to dry before analysis. MALDI-TOF
measurements were performed using a Voyager-DE PRO
instrument (Applied Biosystem, Foster City, CA). Linear,
positive operating mode was adopted. Ions, formed by a pulsed
UV laser beam (λ ) 337 nm), were accelerated to 25 keV.
Instrumental conditions were: grid voltage ) 93%, extraction
3-[1-(2-Fluoro-4′-trifluoromethylbiphenyl-4-yl)-ethyl]-
4H-[1,2,4]oxadiazol-5-one (23). A solution of 22 (200 mg,
0.61 mmol) and CDI (150 mg, 0.93 mmol) in 1,4-dioxane (5
mL) was heated to 100 °C for 30 min, then cooled to room
temperature, diluted with water and extracted with ethyl
acetate. The organic solution was washed with brine, dried
over Na2SO4 and concentrated to afford a crude material that
was washed with ether and desiccated to obtain 23 as a white
1
solid (125 mg, 58%). H NMR (DMSO-d6): δ 12.33(s br, 1H);
7.85(d, 2H); 7.78(d, 2H); 7.61(dd, 1H); 7.37(dd, 1H); 7.30(dd,
1H); 4.23(q, 1H); 1.57(d, 3H). MS (EI; TSQ 700; source 180
°C; 70 V; 200 µA): 352 (M+); 307; 293. Anal. (C17H12F4N2O2)
C, H, N, F.
2-(2-Fluoro-3′,5′-dichloromethylbiphenyl-4-yl)-propi-
onic Acid Methyl Ester (24). Thionyl chloride (130 mg, 1.09
mmol) was added dropwise to a solution, cooled to 0 °C, of 5b
(110 mg, 0.35 mmol) in MeOH (3 mL). The solution was stirred
overnight at room temperature, then was concentrated under
vacuum. The crude material was diluted in diethyl ether,
washed with NaHCO3 10% and dried over Na2SO4. The
concentration of the organic solution afforded a colorless oil
1
(72 mg; 63%). H NMR (200 MHz, CDCl3, T ) 303 K): 7.45-
7.39 (m, 2H, Ar-H); 7.38-7.30 (m, 2H, Ar-H); 7.20-7.15 (m,
1H, Ar-H); 7.15-7.08 (d,d, J ) 6.52 Hz, J ) 1.57 Hz, 1H, Ar-
H); 3.85-3.66 (q+s, J ) 7.30 Hz, 4H, O-CH3, CH-CH3); 1.60-
1.48 (d, J ) 7.30 Hz, 3H, CH3). MS (EI; TSQ 700; source 180
°C; 70 V; 200 µA): 349 (M+Na +).
2-(2-Fluoro-3′,5′-dichloromethylbiphenyl-4-yl)-propi-
onic Acid tert-butyl ester (25). In a solution, cooled to 0
°C, of 5b (110 mg, 0.35 mmol) in dichloromethane 5 drops of
H2SO4 96% were added and isobutylene was bubbled for 5 min.
The solution was stirred at room temperature in a sealed Pyrex
tube overnight. After concentration the crude material was
diluted in diethyl ether, washed with water and dried over
Na2SO4. The concentration of the organic solution afforded a
colorless oil (62 mg; 48%). 1H NMR (200 MHz, CDCl3, T ) 303
K): 7.47-7.40 (m, 2H, Ar-H); 7.39-7.31 (m, 2H, Ar-H); 7.20-
7.14 (m, 1H, Ar-H); 7.14-7.07 (d,d, J ) 13.30 Hz, J ) 4.85
Hz, 1H, Ar-H); 3.72-3.57 (q, J ) 7.26 Hz, 1H CH-CH3); 1.53-
1.37 (d+s, J ) 7.26 Hz 12H, 4CH3). MS (EI; TSQ 700; source
180 °C; 70 V; 200 µA): 391 (M+Na +)
2-(2-Fluoro-3′,5′-dichloromethylbiphenyl-4-yl)-propi-
onamide (26). The methyl ester 24 (200 mg, 0.61 mmol) was
stirred in a solution of 8 M NH3 in MeOH (5 mL) for 2 days.
After concentration the crude material was chromatographed
on silica gel (n-hexane/ethyl acetate 50:50) to afford 26 as a

5718 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Peretto et al.
delay time ) 300 ns and grid wire ) 0.3%. External mass
calibration was performed by using the Calibration Mixture
2 of Sequazime Peptide Mass Standard Kit (Applied Biosys-
tem, Foster City, CA), based on the average values of [M +
H]+ of angiotensin I, ACTH (clip 1-17), ACTH (clip 18-39),
ACTH (clip 7-38) and bovine insulin at m/z 1297.51, 2094,-
46, 2466.72, 3366.19 and 5734.59, respectively.
COX-1 Assay. Freshly taken rat blood (0.5 mL) was
preincubated for 1 h at 37 °C with 0.5 µL of vehicle (DMSO,
0.3% final concentration) or test compounds (100 µM or at
different concentrations). Blood was then stimulated with the
Ca2+ ionophore A23187 (calcimycin; 1 µL, 50 µM) for 30 min.
Afterward, samples were centrifuged at 12 000 × g for 3 min
at 4 °C and the plasma obtained was immediately frozen at
-80 °C, until analysis. Concentrations of [3H]thromboxane B2
were determined by radioimmunoassay, using a commercial
kit (NEK007, Perkin-Elmer, Milan, Italy).
COX-2 Assay. Freshly taken human blood (0.5 mL) was
preincubated at 37 °C in the presence of aspirin (0.5 µL, 10
µg/mL, final concentration in the sample) to inactivate COX-
1. Thirty minutes later, blood was treated with lipopolysac-
charide from Escherichia Coli (1 µL, 100 µg/mL, final concen-
tration in the sample) plus 0.5 µL of vehicle (DMSO, 0.4% final
concentration) or test compounds (100 µM). Afterward, incuba-
tion was continued for 18 h at 37 °C and was terminated by
centrifugation at 12 000 × g for 3 min at 4 °C. Plasma was
separated and immediately frozen at -80 °C, until analysis.
Concentrations of [125I]prostaglandin E2 were determined by
radioimmunoassay, using a commercial kit (NEK020, Perkin-
Elmer, Milan, Italy).
Membrane Permeability. Intestinal drug absorption of
test compounds was estimated in vitro using human colon
adenocarcinoma (Caco-2) cell monolayers (Areta International,
Gerenzano, Italy) as previously described.36 Cells were sus-
pended in DMEMB and seeded (2 × 106 cells in 0.3 mL) in
24-well plates (Transwell, Costar, Cambridge, MA) using
polycarbonate microporous cell culture inserts. Cells were
grown for 21 days. The integrity of cell monolayers was
checked by measuring transepithelial electrical resistance
(1000 Ω) and by the transport of the paracellular leakage
marker sodium fluorescein. One hundred µL of test compound
solution (50 µM in 1% DMSO) was applied to the apical side.
After 2 h incubation at 37 °C the apical and basolateral side
solutions were removed and analyzed by liquid chromatogra-
phy and tandem mass spectrometry with selected reaction
monitoring. The apparent permeability (Papp) was calculated
as Papp ) CbVb/(C0‚At), in which Cb ) basolateral test compound
concentration at time t (µM), Vb ) basolateral volume (cm3),
C0 ) apical test compound concentration at t ) 0 (µM), A )
filter surface area (cm2), t ) time (s). The rank order of
apparent permeability of the test compounds was compared
with that of known reference compounds tested in the same
experiment as internal standards including flurbiprofen, pro-
pranolol and cimetidine. Papp values below the limit of 10 nm/s
were classified as “low”, “medium” when 10 ) Papp ) 50 nm/s
and “high” when Papp > 50 nm/s.
CYP2C19, 7-methoxy-4-trifluoromethylcoumarin (MFC) or
dibenzylfluorescein (DBF) for CYP2C9, 3-[2(N,N-diethyl-N-
methylamino)ethyl]-7-methoxy-4-methylcoumarin (AMMC) for
CYP2D6 and 7-benzyloxy-4-(trifluoromethyl)-coumarin (BFC)
for CYP3A4. Reaction was terminated at various times,
depending on the assays, by addition of a 4:1 acetonitrile: 0.5
M Tris base solution. Known inhibitors for each isoenzyme
(furafylline for CYP1A2, sulfaphenazole for CYP2C9, tranyl-
cypromine for CYP2C19, quinidine for CYP2D6 and ketocona-
zole for CYP3A4) were tested in all assays as positive controls.
Plates were read on a fluorometer (Fluoroskan Ascent, Thermo
Electron Corporation, Helsinki, Finland) at the appropriate
emission/excitation wavelengths. Median inhibitory concentra-
tion (IC50) was determined by nonlinear regression (GraFit
software, Erithacus Software, Horley, UK).
CBF1-Luciferase Notch Assay. Human embryonic kid-
ney 293 (HEK293) cells were seeded in 96-well plates and
grown at 37 °C in the presence of 5% CO2 to subconfluency
with Dulbecco modified Eagle’s minimum essential medium
(DMEM) medium supplemented with 10% fetal bovine serum,
100 units/mL penicillin and 100 µg/mL streptomycin. Cells
were transiently cotransfected with a truncated form of
Notch-1 (Notch-1 ∆ extracellular domain, N1∆EC) and a
C-promoter binding factor 1 (CBF1)-firefly luciferase reporter
construct using a cationic lipid transfection agent (Lipo-
fectamine 2000, Invitrogen, Carlsbad, CA). N1∆EC contains
the cytoplasmic portion and the membrane-spanning region
of Notch-1 and leads to rapid accumulation of the C terminus
of Notch1 in the nucleus and to transcriptional activation of
the CBF1 reporter construct. Eight hours after the cotrans-
fection, vehicle or test compounds were added to the cells for
an additional period of 36 h. The assay was performed in
triplicate using a commercial kit (LucLite, Perkin-Elmer,
Milan, Italy) and a luminometer (LKB 1251, LKB-Wallac,
Turku, Finland). Data were normalized to total protein content
and expressed as counts per second/µg protein.
Cytotoxicity in Human Hepatocytes. The cytotoxicity
in human hepatocytes was assessed by the MTT assay.
Cryopreserved human hepatocytes (25 000-30 000 cells/well)
were seeded into 96-well microplates and treated overnight
with the test compounds or 1% DMSO (vehicle). Cell medium
was then removed and the cells were incubated for 2 h in the
presence of 100 µL of MTT (0.5 mg/mL), at 37 °C, 5%CO2 and
saturated humidity. After removal of the medium, 100 µL of
100% DMSO were added to each well. The amount of formed
formazan was determined reading the samples at 570 nm
(background 630 nm) using a microplater reader (model 450,
Biorad, Hercules, CA).
Pharmacokinetic Studies. Metabolic stability and central
nervous system penetration of the new compounds were
determined in rats after continuous subcutaneous infusion by
osmotic mini-pumps (model 2001, Alzet Osmotic Pumps,
Cupertino, CA) using the cassette dosing technology. This
approach consists of the simultaneous administration of low
doses of two to six different compounds to a single animal and
estimating total body clearance by measuring the compound
concentration in a single plasma sample collected at steady-
state. Brain penetration was estimated by measuring com-
pounds in the cerebrospinal fluid (CSF) and calculating the
CSF to plasma ratio (as reported in Tables 1a, 2a, 3a). Doses
from 2 to 65 µg/h were infused for 4-7 days to rats. One mL
of plasma and 100-150 µL of CSF were collected at the end
of infusion. Three to four animals were used for each dose
session. Samples were prepared by adding 300 µL of acetoni-
trile and 40 µL of phosphoric acid 40% to 100 µL of plasma or
CSF and placing the mixture in a vortex for 20 s. Then,
samples were centrifuged at 14 000 rpm for 15 min and the
supernatants (30 µL) were injected into the HPLC system.
Equipment systems with fluorescence (Waters 470, Waters,
Guyancourt, France) or mass spectrometry (API 2000, Applied
Biosystems, Foster City, CA) detectors were used. The chro-
matographic conditions were adapted to each mixture of
compounds to obtain good peak separation and detection
sensitivity. A mixture of ammonium formate (20 mM) buffer-
Interaction with Cytochrome P450 Isoenzymes. The
assay was performed in a 96-well microtiter plates as pre-
viously described.37 Microsomes from baculovirus-insect cells
(Supersomes, Gentest Corporation, Woburn, MA) express-
ing human cytochrome P450 isoforms (CYP1A2, CYP2C9,
CYP2C19, CYP2D6 and CYP3A4) were utilized. Test com-
pounds or standard inhibitors (dissolved in DMSO or aceto-
nitrile) were serially diluted in a solution containing 16.3 µM
NADP+, 0.83 mM glucose-6-phosphate, 0.83 mM MgCl2, 0.4
U/mL of glucose-6-phosphate dehydrogenase and 0.1 mg/mL
microsomal protein prepared from wild-type baculovirus-
infected insect cells. Appropriate control wells without inhibi-
tor and without microsomes were also added. The plates were
then pre-warmed at 37 °C for 10 min, and the reaction was
started by adding pre-warmed enzyme/substrate mix (cyto-
chrome P450 isoforms with their specific substrates in a 0.35
M potassium phosphate buffer at pH 7.4). Specific substrates
were 3-cyano-7-ethoxycoumarin (CEC) for CYP1A2 and

Flurbiprofen Analogues
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18 5719
(3) Etminan, M.; Gill, S.; Samii, A. Effect of nonsteroidal antiin-
flammatory drugs on risk of Alzheimer’s disease: systematic
review and meta-analysis of observational studies. Br. Med. J.
2003, 327, 128-132.
(4) Launer, LJ. Nonsteroidal antiinflammatory drugs and Alzheimer
disease: what’s next? JAMA 2003, 289, 2865-2867.
(5) Weggen, S.; Eriksen, J. L.; Das, P.; Sagi, S. A.; Wang, R.;
Pietrzik, 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. A subset of NSAIDs lower amyloidogenic Aâ42 indepen-
dently of cyclooxygenase activity. Nature 2001, 414, 212-216.
(6) Eriksen, J. L.; Sagi, S. A.; Smith, T. E.; Weggen, S.; Das, P.;
McLendon, D. C.; Ozols, V. V.; Jessing, K. W.; Zavitz, K. H.; Koo,
E. H.; Golde, T. E. NSAIDs and enantiomers of flurbiprofen
target γ-secretase and lower Aâ42 in vivo. J. Clin. Invest. 2003,
112, 440-449.
(7) Gasparini, L.; Rusconi, L.; Xu, H.; del Soldato, P.; Ongini, E.
Modulation of â-amyloid metabolism by non-steroidal anti-
inflammatory drugs in neuronal cell cultures J. Neurochem.
2004, 88, 337-348.
(8) Lim, G. P.; Yang, F.; Chen, P.; Beech, W.; Teter, B.; Tran, T.;
Ubeda, O.; Hsiao-Ashec, K.; Frautschy, S. A.; Cole, G. M.
Ibuprofen suppresses plaque pathology in a mouse model for
Alzheimer’s disease. J. Neurosci. 2000, 20, 5709-5714.
(9) Lim, G. P.; Yang, F.; Chu, T.; Gahtan, E.; Ubeda, O.; Beech, W.;
Overmier, J. B.; Hsiao-Ashec, K.; Frautschy, S. A.; Cole, G. M.
Ibuprofen effects on Alzheimer pathology and open field activity
in APPsw transgenic mice. Neurobiol. Aging 2001, 22, 983-991.
(10) Jantzen, P. T.; Condor, K. E.; Di Carlo, G.; Wenk, G. L.; Wallace,
J. L.; Rojiani, A. M.; Coppola, D.; Morgan, D.; Gordon, M. N.
Microglial activation and â-amyloid deposit reduction caused by
a nitric-oxide-releasing non-steroidal anti-inflammatory drug in
amyloid precursor protein plus presenilin-1 transgenic mice. J.
Neurosci. 2002, 22, 2246-2254.
(11) Yan, Q.; Zhang, J.; Liu, H.; Babu-Khan, S.; Vassar, R.; Biere,
A. L.; Citron, M.; Landreth, G. Anti-inflammatory drug therapy
alters â-amyloid processing and deposition in an animal model
of Alzheimer’s disease. J. Neurosci. 2003, 23, 7504-7509.
(12) Quinn, J.; Montine, T.; Morrow, J.; Woodward, W. R.; Kulhanek,
D.; Eckenstein, F. Inflammation and cerebral amyloidosis are
disconnected in an animal model of Alzheimer’s disease. J.
Neuroimmunol. 2003, 137, 32-41.
(13) Sung, S.; Yang, H.; Uryu, K.; Lee, E. B.; Zhao, L.; Shineman,
D.; Trojanowski, J. Q.; Virginia Lee, V. M. L.; Pratico`, D.
Modulation of NF-κB activity by indomethacin influences Aâ
levels but not APP metabolism in a model of Alzheimer disease.
Am. J. Pathol. 2004, 165, 2197-2206.
(14) van Groen, T.; Kadish, I. Transgenic AD model mice, effects of
potential anti-AD treatments on inflammation and pathology.
Brain Res Brain Res Rev. 2005, 48, 370-378.
(15) Weggen, S.; Eriksen, J. L.; Sagi, S. A.; Pietrzik, C. U.; Ozols, V.;
Fauq, A.; Golde, T. E.; Koo, E. H. Evidence that nonsteroidal
anti-inflammatory drugs decrease amyloid â42 production by
direct modulation of γ-secretase activity. J. Biol. Chem. 2003,
278, 31831-31837.
(16) Takahashi, Y.; Hayashi, I.; Tominari, Y.; Rikimaru, K.; Moro-
hashi, Y.; Kan, T.; Natsugari, H.; Fukuyama, T.; Tomita, T.;
Iwatsubo, T. Sulindac sulfide is a noncompetitive γ-secretase
inhibitor that preferentially reduces Aâ42 generation. J. Biol.
Chem. 2003, 278, 18664-18670.
(17) Beher, D.; Clarke, E. E.; Wrigley, J. D.; Martin, A. C.; Nadin,
A.; Churcher, I.; Shearman, M. S. Selected non-steroidal anti-
inflammatory drugs and their derivatives target γ-secretase at
a novel site. Evidence for an allosteric mechanism. J. Biol. Chem.
2004, 279, 43419-43426.
(18) Lleo, A.; Berezovska, O.; Herl, L.; Raju, S.; Deng, A.; Bacskai,
B. J.; Frosch, M. P.; Irizarry, M.; Hyman, B. T. Nonsteroidal
anti-inflammatory drugs lower Aâ42 and change presenilin 1
conformation. Nat Med. 2004, 10, 1065-1066.
acetonitrile-methanol was used as mobile phase for the
fluorescence detector, while a mixture of ammonium formate
(20 mM) buffer-acetonitrile was used for the mass spectrom-
etry detector API 2000. Pharmacokinetic profile after single
intravenous and oral administration in rats was determined
for flurbiprofen, 11b, 11c and 13d. Plasma samples (100 µL)
were drawn in groups of 4 animals at 0.083, 0.25, 0.5, 1, 2, 4,
8, 24 h after intravenous administration and at 0.5, 1, 2, 4, 8,
24, 48 and 72 h after oral administration of the compounds.
Intravenous doses were 0.2 mg/kg for 13d, 0.5 mg/kg for 11b
and 11c and 1.5 mg/kg for flurbiprofen. Oral doses were 0.5
mg/kg for 13d, 2.5 mg/kg for 11b and 11c and 7.6 mg/kg for
flurbiprofen. To permit a proper comparison between different
compounds, plasma levels were standardized for a 1 mg/kg
dose.
Studies in Transgenic Mice. The effects of selected
compounds of Aâ secretion in vivo were studied in transgenic
mice (Tg2576) overexpressing the human APP gene with the
Swedish double mutation (K670N/M671L) under the tran-
scriptional control of the hamster prion promoter.38 Animals
(B6SJLF17J strain) were bought from Taconic (Germantown,
NY). All experiments were performed in compliance with the
guidelines of the European Union for the use of laboratory
animals. In Study 1, groups of 8 animals of 7-8 months of
age were used. In Study 2, groups of 12 animals of 3-4 months
of age were employed. One day before starting treatments,
animals were anesthetized with ether and blood samples (150
µL) were collected via retro-bulbar puncture in EDTA-coated
tubes for measuring baseline Aâ40 and Aâ42 concentrations.
Vehicle (7.5 mL/kg of Kool-Aid) or test compounds (12.5 mg/
kg in Kool-Aid) were administered, by oral gavage, twice-a-
day for 7 days. The dose used (12.5 mg/kg twice-a-day) was
selected based on preliminary tolerability experiments with
R-flurbiprofen. On Day 8, animals were given a dose of 25 mg/
kg of the test drugs and sacrificed 3 h later. Blood samples
were again collected in EDTA-coated tubes and plasma
separated by centrifugation (800 × g for 20 min) and stored
at -80 °C until assay. The brains were quickly removed on
an ice-cold plate and fronto-parietal cortex and hippocampus
were dissected, immediately frozen on dry ice and stored at
-80 °C. Total Aâ was extracted from brain tissues as described
by Lanz and colleagues.34 In brief, cortices and hippocampi
were homogenized in 10 vol/weight of 5 M guanidium HCl in
50 mM Tris HCl, pH 8.0, agitated by rotation for 3 h at 4 °C,
diluted 1:10 with phosphate-buffered saline. These diluted
homogenates were spun at 35 000 × g (25 min at 4 °C). Aâ40
and Aâ42 concentrations in supernatants were assayed by
ELISA using commercial kits (Genetics Company, Zurich,
Switzerland). All ELISA’s detections were conducted in du-
plicate.
Acknowledgment. This study was supported by
Chiesi Farmaceutici, Parma, Italy. The authors thank
Dr. Alberto Cerri and Dr. Renzo Mena (NiKem Re-
search) for the analytical support and Dr. Roberto
Forlani (NiKem Research) for the library design. We
also thank Dr. Mariella Fusco (Research & Innovation)
for for performing animal experiments and Dr. Alberta
Leon (Research & Innovation) for her scientific advice.
(19) Morihara, T.; Chu, T.; Ubeda, O.; Beech, W.; Cole, G. M. Selective
inhibition of Aâ42 production by NSAID R-enantiomers. J.
Neurochem. 2002, 83, 1009-1012.
(20) Imbimbo, B. P. The potential role of non-steroidal anti-inflam-
matory drugs in treating Alzheimer’s disease. Expert Opin.
Investig. Drugs 2004, 13, 1469-1481.
(21) Geisslinger, G.; Schaible, H. G. New insights into the site and
mode of antinociceptive action of flurbiprofen enantiomers. J.
Clin. Pharmacol. 1996, 36, 513-520.
(22) Leadbeater, N. E.; Marco, M. Rapid and amenable Suzuki
coupling reaction in water using microwave and conventional
heating. J. Org. Chem. 2003, 68, 888-892.
Supporting Information Available: Spectroscopic data
and elemental analyses for compounds 5c-o, 6a-c, 11b-i,
13b-d, and 16b-d. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Hardy, J.; Selkoe, D. J. The amyloid hypothesis of Alzheimer’s
disease: progress and problems on the road to therapeutics.
Science 2002, 29, 353-356.
(2) Gong, Y.; Chang, L.; Viola, K. L.; Lacor, P. N.; Lambert, M. P.;
Finch, C. E.; Krafft, G. A.; Klein, W. L. Alzheimer’s disease-
affected brain: presence of oligomeric Aâ ligands (ADDLs)
suggests a molecular basis for reversible memory loss. Proc. Natl.
Acad. Sci. U.S.A. 2003, 100, 10417-10422.
(23) Yang, S. H.; Li, C. S.; Cheng, C. H.. Halide Exchange reactions
between aryl halides and alkali halides catalyzed by nickel
metal. J. Org. Chem. 1987, 52, 691-694.

5720 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 18
Peretto et al.
(24) Kukar, T.; Murphy, M. P.; Eriksen, J. L.; Sagi, S. A.; Weggen,
S.; Smith, T. E.; Ladd, T.; Khan, M. A.; Kache, R.; Beard, J.;
Dodson, M.; Merit, S.; Ozols, V. V.; Anastasiadis, P. Z.; Das, P.;
Fauq, A.; Koo, E. H.; Golde, T. E. Diverse compounds mimic
Alzheimer disease-causing mutations by augmenting Aâ42
production. Nat. Med. 2005, 11, 545-550.
metabolites in plasma and urine. J. Chromatogr. 1989, 494,
173-182.
(32) Georges, H.; Jarecki, I.; Netter, P.; Magdalou, J.; Lapicque, F.
Glycation of human serum albumin by acylglucuronides of
nonsteroidal anti-inflammatory drugs of the series of phenyl-
propionates. Life Sci. 1999, 65, PL151-PL156.
(33) Nagao, T.; Tanino, T.; Iwaki, M. Stereoselective pharmacoki-
netics of flurbiprofen and formation of covalent adducts with
plasma protein in adjuvant-induced arthritic rats. Chirality
2003, 15, 423-428.
(34) Lanz, T. A.; Fici, G. J.; Merchant, K. M. Lack of specific amyloid-
â1-42 suppression by nonsteroidal anti-inflammatory drugs in
young, plaque-free Tg2576 mice and in guinea pig neuronal
cultures. J. Pharmacol. Exp. Ther. 2005, 312, 399-406.
(35) PCT/EP2004/001595, “1-phenyl-2-monoalkyl carboxylic acid de-
rivatives for the treatment of neurodegenerative diseases”; PCT/
EP2004/001596, “1-phenylalkanecarboxylic acid derivatives for
the treatment of neurodegenerative diseases”.
(36) Biganzoli, E.; Cavenaghi, L. A.; Rossi, R.; Brunati, M. C.; Nolli,
M. L. Use of a Caco-2 cell culture model for the characterization
of intestinal absorption of antibiotics. Farmaco 1999, 54, 594-
599.
(37) Stresser, D. M.; Blanchard, A. P.; Turner, S. D.; Erve, J. C.;
Dandeneau, A. A.; Miller, V. P.; Crespi, C. L. Substrate-
dependent modulation of CYP3A4 catalytic activity: analysis
of 27 test compounds with four fluorometric substrates. Drug
Metab. Dispos. 2000, 28, 1440-1448.
(25) Rabinovitz M.; Van Thiel D. H. Hepatotoxicity of non-steroidal
anti-inflammatory drugs. Am. J. Gastroenterol. 1992, 87, 1696-
1704.
(26) Tracy, T. S.; Marra, C.; Wrighton, S. A.; Gonzalez, F. J.;
Korzekwa, K. R. Studies of flurbiprofen 4′-hydroxylation. Ad-
ditional evidence suggesting the sole involvement of cytochrome
P450 2C9. Biochem. Pharmacol. 1996, 52, 305-309.
(27) 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 â-APP cleavages.
Biochemistry 2003, 42, 7580-7586.
(28) Hadland, B. K.; Manley, N. R.; Su, D.; Longmore, G. D.; Moore,
C. L.; Wolfe, M. S.; Schroeter, E. H.; Kopan, R. γ-secretase
inhibitors repress thymocyte development. Proc. Natl. Acad. Sci.
U.S.A. 2001, 98, 7487-7491.
(29) 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. Chronic treatment with
the γ-secretase inhibitor LY-411, 575 inhibits â-amyloid peptide
production and alters lymphopoiesis and intestinal cell dif-
ferentiation. J. Biol. Chem. 2004, 279, 2876-2882.
(30) Milano, J.; McKay, J.; Dagenais, C.; Foster-Brown, L.; Pognan,
F.; Gadient, R.; Jacobs, R. T.; Zacco, A.; Greenberg, B.; Ciaccio,
P. J. Modulation of Notch processing by γ-secretase inhibitors
causes intestinal goblet cell metaplasia and induction of genes
known to specify gut secretory lineage differentiation. Toxicol.
Sci. 2004, 82, 341-358.
(38) Hsiao, K. K.; Borchelt, D. R.; Olson, K.; Johannsdottir, R.; Kitt,
C.; Yunis, W.; Xu, S.; Eckman, C.; Younkin, S.; Price, D. Age-
related CNS disorder and early death in transgenic FVB/N mice
overexpressing Alzheimer amyloid precursor proteins. Neuron
1995, 15, 1203-1218.
(31) Knadler, M. P.; Hall, S. D. High-performance liquid chromato-
graphic analysis of the enantiomers of flurbiprofen and its
JM0502541