The geminal dimethyl analogue of Flurbiprofen as a novel Aβ42 inhibitor and potential Alzheimer's disease modifying agent

Article abstract of DOI:10.1016/j.bmcl.2006.01.033

The subtle modification of a selection of Aβ₄₂ inhibiting non-steroidal anti-inflammatory drugs (NSAIDs), through synthesis of the geminal dimethyl analogues, was anticipated to ablate their cyclooxygenase activity whilst maintaining Aβ₄₂ inhibition. Methylflurbiprofen 6 exhibited similar in vitro Aβ₄₂ inhibition to its parent NSAID Flurbiprofen and was further evaluated in the Tg2576 mouse model of Alzheimer's disease and an animal model of gastro-intestinal (GI) impairment, but proved unviable for further clinical development.

Full text of DOI:10.1016/j.bmcl.2006.01.033

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2219–2223
The geminal dimethyl analogue of Flurbiprofen as a novel
Ab42 inhibitor and potential Alzheimer’s disease modifying agent
Nicholas Stock,a,* Benito Munoz,a Jonathan D. J. Wrigley,b Mark S. Shearman,b
Dirk Beher,b James Peachey,c Toni L. Williamson,c Gretchen Bain,d Weichao Chen,a
a
e
a
´
Xiaohui Jiang, Rene St-Jacques and Peppi Prasit
aDepartment of Chemistry, Merck Research Laboratories, 3535 General Atomics Ct, San Diego, CA 92121, USA
bDepartment of Molecular and Cellular Neuroscience, Merck Neuroscience Research Centre,
Terlings Park, Harlow, Essex CM20 2QR, UK
cDepartment of In-Vivo Neuroscience, Merck Neuroscience Research Centre, Terlings Park, Harlow, Essex CM20 2QR, UK
dDepartment of Neurobiology, Merck Research Laboratories, 3535 General Atomics Ct, San Diego, CA 92121, USA
eDepartment of Pharmacology, Merck Frosst, 16711, Route Transcanadienne, Kirkland, Que., Canada
Received 6 January 2006; accepted 9 January 2006
Available online 7 February 2006
Abstract—The subtle modification of a selection of Ab42 inhibiting non-steroidal anti-inflammatory drugs (NSAIDs), through
synthesis of the geminal dimethyl analogues, was anticipated to ablate their cyclooxygenase activity whilst maintaining Ab42 inhi-
bition. Methylflurbiprofen 6 exhibited similar in vitro Ab42 inhibition to its parent NSAID Flurbiprofen and was further evaluated
in the Tg2576 mouse model of Alzheimer’s disease and an animal model of gastro-intestinal (GI) impairment, but proved unviable
for further clinical development.
Ó 2006 Elsevier Ltd. All rights reserved.
Alzheimer’s disease (AD) is a progressive, adult onset,
neurodegenerative disorder characterized by impair-
ments in memory and cognition. The disease displays
characteristic histological changes, including extracellu-
lar amyloid deposits and intracellular neurofibrillary
tangles (NFTs).1 The characteristic plaques present
within the AD brain are primarily composed of b-amy-
loid (Ab) peptides 40 and 42 amino acids in length. They
are derived from the proteolytic processing of the b-am-
yloid precursor protein (APP) through initial cleavage
by the b-secretase (BACE) and then c-secretase en-
zymes. The subsequent formation of Ab oligomers and
amyloid plaques is believed to contribute to neuronal
loss and ultimately failure of cognitive function. Bio-
chemical and genetic evidence suggests that an increased
production or decreased clearance of the longer Ab42
peptide appears to be a key contributor to AD.
In’t Veld et al., in a landmark epidemiological study
(Rotterdam), identified a link between prophylactic
NSAID use and a significant reduction in AD risk.2
Weggen and co-workers have identified a subset of
NSAIDs that lower the level of Ab42 peptide at thera-
peutically relevant doses in vitro.3 They further demon-
strated that Ibuprofen reduces the levels of Ab42 in a
transgenic mouse model.4 A number of reports also indi-
cate the potential beneficial effects of certain NSAIDs on
APP processing including potential mechanisms of ac-
tion through direct modulation of the c-secretase com-
plex.1e,4,5 However, whilst an attractive mechanism for
potential treatment of AD, prolonged NSAID usage is
associated with severe gastro-intestinal (GI) complica-
tions through chronic inhibition of the COX-1 enzyme.6
Thus, we sought to modify these agents to minimize
COX activity whilst maintaining their Ab42 inhibitory
action (Fig. 1).
Cyclooxygenase activity resides primarily in the (S)-
isomer (A) of NSAIDs, the (R)-enantiomer (B) being
largely COX inactive. However, in vivo racemization
(A ! B) is known for several NSAIDs7 and as such
might pose an issue on chronic administration of the
Keywords: Cyclooxygenase; Alzheimer’s disease.
*
Corresponding author. Tel.: +1 858 228 4666; fax: +1 858 228
4766; e-mail: nick.stock@amirapharm.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.033

2220
N. Stock et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2219–2223
Table 1. In vitro potencies of common NSAIDs at Ab42 inhibitiona
Compound
Ab42 % inhibition
SH-SY5Y SPA4CT
HEK APP695
Sulindac sulfide
Ibuprofen
73% at 30 lM
41% at 600 lM
53% at 100 lM
3% at 300 lM
9% at 30 lM
17% at 100 lM
23% at 100 lM
37% at 100 lM
54% at 40 lM
28% at 100 lM
73% at 100 lM
ND
Indomethacin
Naproxen
Sulindac sulfone
Diclofenac
0% at 40 lM
ND
R-Flurbiprofen
S-Flurbiprofen
51% at 100 lM
65% at 100 lM
Figure 1. Hypothesis to eliminate COX activity of NSAIDs whilst
maintaining Ab42 inhibitory activity.
ND, no significant inhibition of Ab40 was observed.
a Inhibitory action measured in SH-SY5Y and HEK cell lines over-
expressing either the direct substrate SPA4CT or the precursor
APP695, respectively (Ref. 9).
(R)-enantiomer. For example, (R)-Flurbiprofen is a
weak COX-1 inhibitor (IC50 ꢀ 44 lM in human whole
blood)8 and shows racemization in humans.7d We
envisaged that the dimethyl derivatives (C) would also
be poor COX inhibitors and retain Ab42 inhibitory ac-
tion. Furthermore, removal of the stereocenter obvious-
ly negates any complications from in vivo racemization.
R
R
OR1
OMe
ii)
O
O
R1 = H
R1 = Me (a)
(b)
OH
i)
R
We began our efforts by measuring the in vitro activity
of certain NSAIDs to inhibit Ab42 production and con-
firmed the published findings (see Table 1).3 Of note is
that both enantiomers of Flurbiprofen appear to signif-
icantly reduce the amount of Ab42 in vitro at high, but
clinically relevant, concentrations. Naproxen and
Diclofenac appeared to have negligible effect on Ab42
production at relevant concentrations which proved
surprising as these compounds constituted over 60% of
the prescribed drugs in the aforementioned ‘Rotterdam’
epidemiological study.2
iii)
O
(c)
Scheme 1. Reagents and conditions: (i) MeOH, H2SO4 (cat), reflux;
(ii) LDA, THF, DMPU, ꢁ78 °C then MeI; (iii) KOTMS, THF, 45 °C.
Table 2. Inhibitory potencies against Ab42 for geminal dimethyl
NSAID derivativesa
Parent
NSAID
Geminal dimethyl
derivative
Compound % inhibition
at 100 lM
(HEK cells)
Next, we synthesized the geminal dimethyl analogues of
a selection of these NSAIDs. We anticipated that these
compounds, being structurally similar to the parent
NSAID, should possess similar in vitro Ab42 inhibitory
action but much reduced COX activity. Synthesis of
these compounds is outlined (Scheme 1) along with
representative examples (Table 2).10 Fischer esterifica-
tion of the aryl propionic acid (a) followed by alkylation
under standard conditions yielded the geminal dimethyl
ester (b). Conversion of the ester to the acid (c) was
readily achieved by using potassium trimethylsilanolate
(KOTMS).11 However, synthesis of the geminal dimeth-
yl analogue (3) of Indomethacin required an alternative
synthetic procedure (Scheme 2).12
O
OH
Fenoprofen
Naproxen
1
2
0
0
O
OH
O
MeO
MeO
OH
O
N
Indomethacin
Ibuprofen
3
4
5
2
3
O
Cl
OH
These compounds were then assayed for their ability to
lower Ab42 levels (Table 2) in vitro. Only the dimethyl
analogues of Carprofen 5 and Flurbiprofen (Methylflur-
biprofen, 6) showed similar inhibitory activities to their
parent NSAID, the latter being the most potent (67%
inhibition at 100 lM).
O
Cl
OH
Carprofen
40
O
N
H
Despite the low level of in vivo racemization of (R)-
Flurbiprofen, chronic usage may still lead to gastro-
intestinal complications, especially if the therapeutic
OH
O
Flurbiprofen
6
67
F
'Methylflurbiprofen'
(R)-Flurbiprofen has recently entered clinical trials for prostate
cancer and an IND has been filed for its investigation as a treatment
for Alzheimer’s disease.
a Ab42 inhibitory activity measured at 100 lM in a HEK cell line
overexpressing APP695 (Ref. 9). No inhibition of Ab40 was observed.

N. Stock et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2219–2223
2221
O
Br
MeO
ii)
i)
MeO
N
N
H
OH
O
H
OH
O
MeO
iii)
MeO
N
N
O
H
3
Cl
Scheme 2. Reagents: (i) ZnCl2, BuLi then 2-bromoisobutyryl bromide; (ii) 2.4 M NaOH, EtOH; (iii) KHMDS (2.5 equiv), THF, 4-chlorobenzoyl
chloride.
Table 3. Pharmacokinetic parameters for Methylflurbiprofen (6)a
22.5
-11%
-13%
20.0
Rat
Mouse
Dog
Monkey
17.5
15.0
12.5
10.0
7.5
IV
Cl (mL/min/kg)
Vdss (L/kg)
1.3
0.15
0.7
0.28
0.11
11.3
3.8
0.3
8.0
0.22
4.6
T1/2
3.4
PO
Cmax (lM)
Tmax (h)
F%
15.2
2.0
79
24.4
0.25
81
76.4
0.33
87
27.3
0.25
90
**
5.0
-85%
a Methylflurbiprofen dosed at 2 mg/kg po and iv in four species.
2.5
0.0
dose is high. The degree of racemization is also species
specific, possibly presenting problems with chronic
dosing paradigms in pre-clinical animal studies.7a,b
Based on this information, we envisaged Methylflurbi-
profen would be superior due to its similar in vitro
activity at Ab42 inhibition, lack of activity against
COX, and inherent achirality. Encouragingly, Meth-
ylflurbiprofen was inactive in a human whole blood
assay against COX-1 (IC50 > 100 lM). (For compari-
son (S)-Flurbiprofen has an IC50 ꢀ 0.2 lM against
COX-1 in human whole blood.) Furthermore, Meth-
ylflurbiprofen showed no inhibitory activity against
the Notch receptor cleavage at 200 lM, a potential
benefit compared to direct c-secretase inhibition13
and the pharmacokinetic parameters are excellent
(see Table 3), showing good bioavailability (79–90%),
half-life (3.4–11.3 h), and low clearance (0.3–3.8 mL/
min/kg) in four species. Ancillary pharmacology also
showed excellent dose proportionality, no adverse tox-
icology, no adverse effects in a canine cardiovascular
model, and no significant activity when screened
against a panel of other enzyme targets (data not
shown).
2.0
-20%
-18%
1.5
1.0
0.5
0.0
**
-80%
Figure 2. Brain levels of Ab40 (top) and Ab42 (bottom). Ibuprofen and
Methylflurbiprofen dosed at 21 mg/kg (qid) for 3 days in Tg2576 mice
(n = 8). A c-secretase inhibitor (positive control) was given as a single
10 mg/kg dose on the final day (n = 4). Soluble Ab40 and Ab42 peptide
was extracted from brain homogenate using DEA, see Refs. 4 (dosing)
and 15 (extraction) for assay details.
Utilizing Ibuprofen as a positive control, Methylflurbi-
profen was then tested in vivo using the Tg2576 trans-
genic mouse14 model to ascertain its ability to inhibit
the production of Ab42 peptide. Subchronic 3 day dos-
ing at 21 mg/kg/day (qid) revealed no significant efficacy
in lowering levels of Ab42. In vitro data indicated Meth-
ylflurbiprofen as a more potent compound than Ibupro-
fen but surprisingly no difference in activity is observed
between them in this assay (Fig. 2). No significant effects
on Ab40 levels were observed.
Presumably, to achieve in vivo efficacy these compounds
have to enter the brain, and as such we measured the brain
and plasma exposure levels of Methylflurbiprofen 6 and
various NSAIDs upon oral administration (Table 4).
Brain penetration for these compounds in mice proved
low, with an average brain to plasma ratio of 3%

2222
N. Stock et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2219–2223
Table 4. Brain and plasma levels of common NSAIDs and
Methylflurbiprofena
Continued use of racemic Flurbiprofen poses obvious
gastro-intestinal complications and use of (R)-Flurbi-
profen as a clinical agent for the treatment of AD might
also encounter similar complications due to in vivo race-
mization. Methylflurbiprofen has reduced activity to-
ward the COX enzyme and an inherent lack of
racemization liability. However, despite being as potent
in vitro at Ab42 inhibition as (R)-Flurbiprofen and more
potent than Ibuprofen, it appears only marginally effica-
cious in vivo at lowering Ab42 levels in the Tg2576
mouse model. Coupled with the observation that high
doses of 6 cause significant GI leakage, this compound
proved unviable for further development.
Brain
(lM)
Plasma
(lM)
Brain/plasma
ratio
Ibuprofen
18
41
3
592
1188
266
887
296
87
0.03
0.04
0.01
0.03
0.006
0.03
0.03
0.68
0.02
R-Flurbiprofen
Indomethacin
Naproxen
27
1.7
2.7
41
43
13
Sulindac sulfide
Mefanamic acid
Fenoprofen
1358
64
Acetaminophen
Methylflurbiprofen (6)
849
a 50 mg/kg po dosing in mice (n = 3). Drug levels measured at Tmax
.
Acknowledgments
(excluding acetaminophen). This indicated that the
brain levels achieved within our transgenic study did
not achieve the in vitro IC50 of Ab42 inhibition. This
highlighted an apparent disconnect between brain levels,
IC50, and in vivo efficacy, and appears contrary to pub-
lished data.4 Similar findings have also been observed by
Lanz and co-workers.16
The authors thank Merryl Cramer, Janice Brunner, and
Dan Cui for assistance with PK studies.
References and notes
1. (a) Hardy, J. A.; Higgins, G. A. Science 1992, 256, 184;
(b) Selcoe, D. J. Clin. Neurosci. Res. 2001, 1, 91; (c)
Golde, T. E. J. Clin. Invest. 2003, 111(1), 1; (d) Evin,
G.; Weidemann, A. Peptides 2002, 23, 1285; (e) Gaspa-
rini, L.; Ongini, E.; Wenk, G. J. Neurochem. 2004, 91,
521.
2. In’t Veld, B. A.; Ruitenberg, A.; Hofman, A.; Launer, L.
J.; van Duijn, C. M.; Stijnen, T.; Breteler, M. M. B.;
Stricker, B. H. C. N. Engl. J. Med. 2001, 345(21), 1515;
For a review of epidemiological data see: Etminan, M.;
Gill, S.; Samii, A. Br. Med. J. 2003, 327.
3. 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. Nature 2001, 414, 212; For further
examples of the in vitro Ab42 activity of NSAIDs see:
Morihara, T.; Chu, T.; Ubeda, O.; Beech, W.; Cole, G. M.
J. Neurochem. 2002, 83, 1009; Takahashi, Y.; Hayashi, I.;
Tominari, Y.; Rikimaru, K.; Morohashi, Y.; Kan, T.;
Natsugari, H.; Fukuyama, T.; Tomita, T.; Iwatsubo, T.
J. Biol. Chem. 2003, 278(20), 18664; Gasparini, L.;
Rusconi, L.; Xu, H.; del Soldato, P.; Ongini, E.
J. Neurochem. 2004, 88, 337.
Epidemiological data suggest compounds of this nature
only reveal efficacy at reducing plaque within the AD
brain after chronic usage. Therefore, GI tract liabilities
from prolonged administration of Methylflurbiprofen 6
were analyzed in a radioactive Cr leakage assay (Fig. 3).
Human whole blood COX-1 analysis of Methylflurbipro-
fen indicated an IC50 > 100 lM. However, a 5 day,
100 mg/kg (bid) dosing regimen in rat showed significant
leakage within the 51Cr assay, the effects being compara-
ble to a 1 day, 30 mg/kg (bid) dose of Ibuprofen. This dose
was chosen to analyze the consequence of elevated plasma
levels potentially achieved upon chronic dosing. Whilst
the COX-1activity of 6 is markedly reduced, plasmalevels
in this assay (average of 430 lM, 1 h post-final dose)
clearly exceed those needed for COX-1 inhibition, thus
producing gastro-intestinal lesions.
In summary, epidemiological data available show a
tantalizing link between chronic NSAID usage and a
reduction in the risk of contracting Alzheimer’s disease.
4. 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. J. Clin. Invest. 2003,
112(3), 440; In vivo analysis of Ibuprofen in a mouse model
of AD has also been reported by: Lim, G. P.; Yang, F.;
Chu, T.; Chen, P.; Beech, W.; Teter, B.; Tran, T.; Ubeda,
O.; Hsaio Ashe, K.; Frautschy, S. A.; Cole, G. M.
J. Neurosci. 2000, 20(15), 5709.
´
5. (a) Lleo, A.; Berezovska, O.; Herl, L.; Raju, S.; Deng, A.;
Bacskai, B. J.; Frosch, M. P.; Irizarry, M.; Hyman, B. T.
Nat. Med. 2004, 10, 1065; (b) Beher, D.; Clarke, E. E.;
Wrigley, J. D. J.; Martin, A. C. L.; Nadin, A.; Churcher,
I.; Shearman, M. S. J. Biol. Chem. 2004, 279, 43419.
6. Singh, G.; Ramey, D. R.; Morfeld, D.; Shi, H.; Hatoum,
H. T.; Fries, J. F. Arch. Int. Med. 1996, 156, 1530; Fries. J.
F. J. Rheumatol. 1991, 18, 7; Anderson, R. N. et al.
Centers for Disease Control and Prevention/National
Center for Health Statistics Report of Final Mortality
Statistics, 1995; Monthly Vital Statistics Report 1997,
145(11); Wolfe, M. M.; Lichtenstein, D. R.; Singh, G. N.
Engl. J. Med. 1999, 340, 1888.
Figure 3. GI permeability as measured using a 51Cr-EDTA excretion
assay. Methylflurbiprofen was dosed at 100 mg/kg (bid) for 5 days and
Ibuprofen at 30 mg/kg (bid) for 1 day (n = 6). Plasma levels for
Methylflurbiprofen at 1 h post-final dose averaged 430 lM, Assay
details are described in Ref. 17.

N. Stock et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2219–2223
2223
7. (a) Menzel-Soglowek, S.; Geisslinger, G.; Beck, W. S.;
Brune, K. J. Pharm. Sci. 1992, 81(9), 888; (b) Knihinicki,
R. D.; Day, R. O.; Graham, G. G.; Williams, K. M.
Chirality 1990, 2, 134; (c) Forland, S. C.; Wechter, W. J.;
Witchwoot, S.; Clifford, K. H.; Arnett, L.; Cutler, R. E.
J. Clin. Pharmacol. 1996, 36, 546; (d) Geisslinger, G.;
Lotsch, J.; Menzel, S.; Kobal, G.; Brune, K. Br. J. Clin.
Pharmacol. 1994, 34, 392.
8. Geisslinger, G.; Muth-Selbach, U.; Coste, O.; Vetter, G.;
Schro¨dter, A.; Schaible, H.-G.; Brune, K.; Tegeder, I.
J. Neurochem. 2000, 74, 2094.
9. Clarke, E. E.; Shearman, M. S. J. Neurosci. Methods 2000,
102, 61.
Chirgadze, N. Y.; Clawson, D. K.; Hartley, L. W.;
Johnson, L. M.; Jones, N. D.; McKinney, E. R.; Mihelich,
E. D.; Olkowski, J. L.; Schevitz, R. W.; Smith, A. C.;
Snyder, D. W.; Sommers, C. D.; Wery, J.-P. J. Med.
Chem. 1996, 39, 5119.
13. De Strooper, B.; Annaert, W.; Cupers, P.; Saftig, P.;
Craessaerts, K.; Mumm, J. S.; Schroeter, E. H.; Schrijvers,
V.; Wolfe, M. S.; Ray, W. J.; Goate, A.; Kopan, R. Nature
1999, 398, 518.
14. Hsiao, K.; Chapman, P.; Nilsen, S.; Eckman, C.; Harigaya,
Y.; Younkin, S.; Yang, F.; Cole, G. Science 1996, 274, 99.
15. (a) Kawarabayashi, T.; Younkin, L. H.; Saido, T. C.;
Shoji, M.; Hsaio Ashe, K.; Younkin, S. G. J. Neurosci.
2001, 21(2), 372; (b) Best, J. D.; Jay, M. T.; Out, F.; Ma,
J.; Nadin, A.; Ellis, S.; Lewis, H. D.; Pattison, C.; Reilly,
M.; Harrison, T.; Shearman, M. S.; Williamson, T. L.
J. Pharmacol. Exp. Ther. 2005, 313, 902.
10. A subset of these compounds has been investigated in
another study: Sakaguchi, K.; Green, M.; Stock, N.;
Reger, T. S.; Zunic, J.; King, C. Arch. Biochem. Biophys.
2004, 424, 219.
11. Laganis, E. D.; Chenard, B. L. Tetrahedron Lett. 1984, 25,
5831.
16. Lanz, T. A.; Fici, G. J.; Merchant, K. M. J. Pharmacol.
Exp. Ther. 2005, 312, 399.
12. For C-acylation of the indole scaffold, see: Dillard, R. D.;
Bach, N. J.; Draheim, S. E.; Berry, D. R.; Carlson, D. G.;
17. Khazaenina, T.; Jamali, F. Clin. Exp. Rheumatol. 2000,
18(2), 187.