The discovery and development of selective 3-fluoro-4-aryloxyallylamine inhibitors of the amine oxidase activity of semicarbazide-sensitive amine oxidase/vascular adhesion protein-1 (SSAO/VAP-1)

Article abstract of DOI:10.1016/j.bmcl.2012.04.111

A new class of 3-fluoroallyl amine-based SSAO/VAP-1 inhibitors is reported. These compounds have excellent selectivity over diamine oxidase, MAO-A and MAO-B. Synthesis and SAR studies leading to compound 28 (PXS-4159A) are reported. The pharmacokinetic profile of 28 in the rat, together with activity in a murine model of lung inflammation are also disclosed.

Full text of DOI:10.1016/j.bmcl.2012.04.111

Bioorganic & Medicinal Chemistry Letters 22 (2012) 3935–3940
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The discovery and development of selective 3-fluoro-4-aryloxyallylamine
inhibitors of the amine oxidase activity of semicarbazide-sensitive amine
oxidase/vascular adhesion protein-1 (SSAO/VAP-1)
Jonathan S. Foot a, , Mandar Deodhar a, Craig I. Turner a, Ping Yin a, Ellen M. van Dam a, Diego G. Silva a,
⇑
Aldo Olivieri b, Andrew Holt b, Ian A. McDonald a
a Pharmaxis Ltd., 20 Rodborough Road, Frenchs Forest, NSW 2086, Australia
b Department of Pharmacology, University of Alberta, Edmonton, AB, Canada T6G 2H7
a r t i c l e i n f o
a b s t r a c t
Article history:
A new class of 3-fluoroallyl amine-based SSAO/VAP-1 inhibitors is reported. These compounds have
excellent selectivity over diamine oxidase, MAO-A and MAO-B. Synthesis and SAR studies leading to com-
pound 28 (PXS-4159A) are reported. The pharmacokinetic profile of 28 in the rat, together with activity
in a murine model of lung inflammation are also disclosed.
Received 7 March 2012
Revised 18 April 2012
Accepted 23 April 2012
Available online 30 April 2012
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
SSAO
VAP-1
Allylamines
Amine oxidase
Semicarbazide-sensitive amine oxidase (AOC3, also known as
vascular adhesion protein-1 [VAP-1], primary amine oxidase or
plasma amine oxidase, EC 1.4.3.21) is a member of the copper-
dependent family of amine oxidases (AOC).1 SSAO/VAP-1 is a type
1 membrane-bound protein which has a distal adhesion domain
and a catalytic amine oxidase site proximal to the membrane. Both
sites have been shown to be critical for SSAO/VAP-1-mediated
inhibition of leukocyte rolling, adhesion and transmigration in
response to inflammatory stimuli.2 SSAO/VAP-1 catalyzes the oxi-
dation of primary amines to aldehydes with release of ammonia
and hydrogen peroxide (Scheme 1) and is sensitive to carbonyl-
reactive reagents (such as semicarbazide), but insensitive to selec-
tive flavin-dependent MAO-A and MAO-B enzyme inhibitors like
selegiline and clorgyline.
In humans four AOC genes are known; three (AOC-1–3) trans-
late into enzymatically active protein and one (AOC-4) into a trun-
cated inactive form.3 AOC-1 codes for a diamine oxidase (DAO)
found in the kidney, gut and lung and is involved in the metabo-
lism of endogenous histamine. AOC-2 encodes for retinal amine
oxidase which is found almost exclusively in the eye.4 AOC-3
encodes for the primary monoamine oxidase SSAO/VAP-1 and is
found in adipocytes, smooth muscle cells and endothelial cells,
and is highly expressed in the lung, aorta, liver and ileum.5
Membrane bound SSAO/VAP-1 is proteolytically cleaved by a
metalloprotease to release an active, soluble form of the protein.6
Endogenous substrates for SSAO/VAP-1 include methylamine and
aminoacetone.7 The enzyme also oxidizes other substrates, such
as benzylamine, which is often used as the substrate to assay
enzymatic activity. There are considerable substrate and inhibitor
differences amongst species so it can be challenging to design po-
tent inhibitors of human SSAO/VAP-1 with similar inhibitory
potency against the rodent and canine enzymes.8
Different classes of SSAO/VAP-1 inhibitors have been reported
in the literature and have been reviewed thoroughly elsewhere.9
Inhibition of SSAO/VAP-1 has been shown to be efficacious in ro-
dent models of eye inflammation, rheumatoid arthritis, carrageen
air-pouch inflammation, liver fibrosis, and stroke, amongst other
inflammatory models.10 In addition, some inhibitors have proven
efficacious in lung inflammatory responses.10b For example,
mofegiline (1) attenuates the LPS-induced lung inflammatory
response in mice over expressing human SSAO/VAP-1.11 Therefore
selective SSAO/VAP-1 inhibitors have the potential to treat lung
inflammation in humans, including neutrophil-driven processes,
such as severe asthma. The focus of this research project is to
develop inhibitors for this indication.
Mofegiline (Fig. 1. 1), which progressed to Phase 2 clinical trials
⇑
as an adjunct to
L-dopa for the treatment of Parkinson’s disease, is a
Corresponding author.
E-mail address: jonathan.foot@pharmaxis.com.au (J.S. Foot).
well known inhibitor of the amine oxidase activity of SSAO/VAP-1.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.04.111

3936
J. S. Foot et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3935–3940
Scheme 1. SSAO/VAP-1 deaminates primary amines with re-oxidation of the reduced co-factor.
haloallylamine scaffold have been designed for most amine oxi-
dases; most recently LJP 1586, which was selected from a series
of inhibitors which included mofegiline (1),13 was reported to be
an effective inhibitor of SSAO/VAP-1, with less activity as an
MAO-B inhibitor.10b
This current Letter describes the design, synthesis and evalua-
tion of new, potent fluoroallylamine inhibitors of SSAO/VAP-1
which show greater than 100-fold selectivity against the other
amine oxidizing enzymes MAO-A, MAO-B and diamine oxidase.
Several crystal structures of SSAO/VAP-1 have been published
revealing an elongated binding pocket with the co-factor and
copper sitting deep within the pocket.14 Furthermore, modeling
indicated that if 1 were re-designed with the phenylethyl moiety
tethered to the 3-, rather than the 2-position of 3-fluoroallylamine,
to give the more linear compound (2), this should be readily
accepted by SSAO/VAP-1. These modeling studies did not
Figure 1. Structure of mofegiline (1).
In a mechanism reminiscent of its mode of action to inhibit
MAO-B,12 mofegiline is believed to be processed as a substrate by
SSAO/VAP-1 to generate a covalently attached, very reactive inter-
mediate (highlighted in Scheme 2) which is alkylated by a neigh-
boring nucleophilic residue. Loss of fluoride leads to irreversible
inhibition and explains, in part, the very high potency that can
be observed with this class of inhibitor. Inhibitors based on the
Scheme 2. Proposed mechanism-based inhibition of SSAO/VAP-1 by 1.
Scheme 3. Design of a lead with improved selectivity, rearrangement of the 2- to the 3-substituted fluoroallyamine scaffold.

J. S. Foot et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3935–3940
3937
Scheme 4. Preparation of E and Z-fluoroallylamines.
accomplished by a Horner–Wadsworth–Emmons reaction of 4
with the N-Boc protected amino aldehyde (3), followed by DI-
Table 1
Initial lead identification16
BAL-H reduction to afford
a separable mixture of 5 and 6
a,b
Structure
SSAO IC50
(lM)
(Scheme 4).15 Reaction of 5 and 6 with phenols under Mitsunobu
conditions, followed by TFA deprotection and HCl salt formation,
led to geometric isomers 7 and 8, respectively. With unsubstituted
phenol (R = H), compounds 9 and 10 were prepared (Table 1); the
des-fluoro analogos 11 and 12 were also prepared by standard pro-
cedures. From this small series it can be deduced that the E-geom-
etry is more favorable than the Z, and that the 3-fluoro substitution
is essential. With this information in hand, a larger structure–activ-
ity relationship study was initiated to optimize potency and
selectivity.
F
F
O
9
0.29
NH2
O
O
10
20
NH2
The most potent compound, 9 (IC50 = 290 nM) was found to
have similar potency as an inhibitor of MAO-B (IC50 = 180 nM,
Table 2), with virtually no activity against MAO-A (IC50 >30 lM).
11
12
>100
82
NH2
NH2
The challenge, therefore, was to improve inhibitory potency
against SSAO/VAP-1 while reducing inhibition of MAO-B. The bind-
ing site of MAO-B is relatively narrow, such that large substituents
on the aromatic ring would be expected to restrict binding consid-
erably and thus lead to much weaker inhibition. In contrast the
binding site of SSAO/VAP-1 is relatively large and can accommo-
date more bulky groups. With this understanding, substituents
which break the planarity (with respect to the aromatic ring) were
expected to be much less efficient inhibitors of MAO-B. Using the
convergent synthetic route (Scheme 4), a range of phenols was
coupled with intermediate 5 to afford the derivatives listed in Ta-
ble 2.
O
a
Values are mean of three experiments.
Resorufin/Amplex Red assay readout of H2O2 production.
b
conclusively determine whether 2 would be an effective mecha-
nism-based inhibitor; however, the positioning of Asp386
suggested that turnover would occur. This deconvolution is shown
in Scheme 3. To enable rapid SAR studies, an oxygen atom was
incorporated into the side chain.
Synthetically, the first task was to prepare large amounts of
both isomers of the key intermediate (5 and 6). This was
These data show that simple substitution at either the 3 or 4-
positions with small electron-withdrawing and electron-donating
Table 2
Substituted aryl examples
a
a
a
Aryl R-substitution
SSAO IC50
(
lM)
MAO-A IC50
>30
(
lM)
MAO-B IC50 (lM)
9
None
0.29
0.16
0.24
0.25
0.02
0.21
0.18
0.07
0.19
0.04
0.15
1.07
0.18
0.31
0.45
0.03
0.16
6.40
0.07
0.01
0.09
0.13
0.03
>30
13
14
15
16
17
18
19
20
21
22
23
4-OMe
4-OPh
3-Cl
4-SMe
4-tBu
>30
4-OCF3
4-NO2
3,4-Difluoro
4-CF3
3-CF3
4-CONH2
>10
>30
a
Resorufin/Amplex Red assay readout of H2O2 production.

3938
J. S. Foot et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3935–3940
Table 3
Amide examples
Structure
Human amine oxidases (IC50
;
l
M)a
Mouse adipocytes (IC50
SSAO/VAP-1
0.01
; l
M)a
SSAO/VAP-1
MAO-A
MAO-B
DAO
>10
1
9
0.02
0.29
1.78
>30
0.01
0.18
F
O
O
23
1.07
0.37
0.14
0.26
0.02
0.01
0.08
0.04
>30
>30
H2N
O
NH2
NH2
F
24
18.99
10.96
2.21
1.64
1.12
0.62
2.98
N
O
F
O
25
N
NH2
O
F
O
26
N
NH2
O
F
O
H
N
27
>30
>30
>30
0.53
0.15
NH2
O
O
F
O
H
28
26.13
N
NH2
F
F
O
O
H
N
29
NH2
O
O
H
N
30 (racemic)
NH2
30a (R)
30b (S)
0.03
0.04
>30
>30
1.61
6.43
2.46
74.07
0.13
a
Resorufin/Amplex Red assay readout of H2O2 production.
groups did not influence inhibitory potency against SSAO/VAP-1 to
any great extent. However, compound 17, substituted with a bulky
tert-butyl group was much less potent against MAO-B while main-
taining potency against SSAO/VAP-1. It was also observed that the
introduction of the hydrophilic, polar carboxamide (23) led to a
considerable reduction in potency against SSAO/VAP-1 and espe-
cially MAO-B. Taking advantage of the observation that both bulky
and polar groups impart selectivity over MAO-B, a series of amides
were synthesized and evaluated. The tertiary amides 24–26
(Table 3) had the same profile as the unsubstituted amide 23. How-
ever, the secondary amides (27–29) were more promising, espe-
cially those with larger lipophilic groups. The most potent and
selective compounds in this series were 27 (20 nM [SSAO/VAP-
1]; 32-fold selective vs MAO-B), 28 (10 nM; 112-fold) and racemic
30 (40 nM; 75-fold).17
When compound 28 is docked into the active site of human
SSAO/VAP-1 prior to Schiff base formation with the TPQ co-factor,
it becomes apparent that the lipophilic cyclohexyl fits snugly into

J. S. Foot et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3935–3940
3939
Figure 2. Docking of unbound 28 into active site prior to Schiff base formation. Solvent exposed entrance to the left, molecular surface is shown as contours. Topaquinone co-
factor (off-copper) is shown in purple, lipophilic residues adjacent to cyclohexyl group in green.18
the binding pocket and makes positive interactions with the neigh-
boring lipophilic phenylalanine and tyrosine resides (Fig. 2). This
would lead to a lower binding constant (KD), and equal or perhaps
enhanced kinact, and hence to an overall increase in inhibitory
potency.
In summary, a novel series of mechanism-based inhibitors of
the amine oxidase activity of SSAO/VAP-1 in the allylamine class
has been developed. These compounds were optimized to afford
improved selectivity over MAO-B, achieved through modification
Compound 28 is a time-dependent, mechanism-based inhibitor
of the amine oxidase activity SSAO/VAP-1 (Fig. 3). A Kitz and Wil-
son plot19 determined the apparent KI and kinact values to be
1.64
for mofegiline (1.61
l
M and 0.46 minÀ1, respectively, which are similar to those
l
M and 1.03 minÀ1). In vitro, the inhibition
was shown to be irreversible after 24 h dialysis and the turnover
number determined to be approximately 10 (data not shown). Fur-
ther experiments to explore fluoride elimination were limited by
the large quantity of enzyme required.12
Compound 28 was shown to be inactive (IC50 >10 lM) against a
panel of traditional off-targets, inactive in the AMES test, did not
inhibit CYP isozymes (data not shown) and had an acceptable
pharmacokinetic profile (rat; bioavailability = 62%, half-life
(po) = 97 min.). Compound 28 (PXS-4159A) was chosen for further
evaluation.
Having established that 28 had an acceptable pharmacoki-
netic profile, the pharmacodynamic properties were evaluated
in the mouse and rat. SSAO/VAP-1 amine oxidase activity was
determined 6 h after an oral dose in gonadal adipose tissue,
and compared against the more potent 1. The EC50 values for
28 after a single oral dose were 10 mg/kg for both mouse and
rat which is reflective of the weaker potency against the rodent
enzyme.
Compound 28 was then tested in a mouse model of airway
inflammation. Mice were challenged with LPS to induce a predom-
inantly neutrophilic response.20 Pre-treatment with 28 (100 mg/
kg, ip) 3 h before LPS resulted in a marked decrease in cellular infil-
tration into the bronchoalveolar spaces (Fig. 4A). BAL fluid con-
tained up to 80% fewer neutrophils compared to vehicle alone,
indicating that in treated mice the recruitment of cells to the air-
ways was diminished (Fig. 4B). This reduction followed the same
trend as the positive treatment control, dexamethasone (Dexa).
Cytokine levels were also measured. Whereas vehicle treated mice
Figure 3. Kitz and Wilson analysis of 28. (A) Representative plot of time-dependent
inhibition of SSAO/VAP-1 by 28. First order inhibition rate constants at each
concentration of 28 are calculated as slope  2.303. (B) Replot of inhibition rate
constants from three experiments. Mean inhibition constants SEM shown in
figure.
showed high levels of TNFa (Fig. 4C), significant decreases were
observed with dexamethasone as well as with 28, consistent with
the diminished neutrophil influx into the airways.

3940
J. S. Foot et al. / Bioorg. Med. Chem. Lett. 22 (2012) 3935–3940
Figure 4. Evaluating compound 28 in a murine LPS model of lung inflammation. (A) Total cells in BALF; (B) neutrophil count; (C) TNFa levels. (n = 10 animals/group; P values;
⁄0.05, ⁄⁄0.01, ⁄⁄⁄0.001 as analysed by one-way ANOVA).
10. (a) Xu, H. L.; Salter-Cid, L.; Linnik, M. D.; Wang, E. Y.; Paisansathan, C.;
Pelligrino, D. A. J. Pharmacol. Exp. Ther. 2006, 317, 19; (b) O’Rourke, A. M.; Wang,
E. Y.; Miller, A.; Podar, E. M.; Scheyhing, K.; Huang, L.; Kessler, C.; Gao, H.; Ton-
Nu, H.; McDonald, M. T.; Jones, D. S.; Linnick, M. D. J. Pharmacol. Exp. Ther. 2008,
324, 867.
11. Yu, P. H.; Lu, L.; Fan, H.; Kazachkov, M.; Jiang, Z.; Jalkanen, S.; Stolen, C. Am. J.
Pathol. 2006, 168, 718.
12. Milczek, E. M.; Bonivento, D.; Binda, C.; Mattevi, A.; McDonald, I. A.;
Edmondson, D. E. J. Med. Chem. 2008, 51, 8019.
of a distal amide group. Compound 28 exhibited excellent potency
and selectivity in vitro, a promising pharmacokinetic profile and
activity in an animal model of lung inflammation, and has been ad-
vanced for preclinical toxicology evaluation.
References and notes
1. Salmi, M.; Jalkanen, S. Science 1992, 257, 1407.
13. Allyl amine MAO inhibitors, Bey, P. US4454158.
2. (a) O’Sullivan, J.; Unzeta, M.; Healy, J.; O’Sullivan, M. I.; Davey, G.; Tipton, K. F.
Neurotoxicology 2004, 25, 303; (b) Jalkanen, S.; Karikoski, M.; Mercier, N.;
Koskinen, K.; Henttinen, T.; Elima, K.; Salmivirta, K.; Salmi, M. Blood 2007, 1864,
110.
14. (a) Jakobsson, E.; Nilsson, J.; Ogg, D.; Kleywegt, G. J. Acta Cryst. 2005, D61, 1550;
(b) Ernberg, K.; McGrath, A.; Peat, T.; Adams, T.; Xiao, X.; Pham, T.; Newman, J.;
McDonald, I.; Collyer, C.; Guss, J. Acta Cryst. 2010, F66, 1572.
15. Separated by normal phase chromatography; the lack of selectivity in the
Horner–Wadworth–Emmons reaction is due to the inductive effect of the
fluorine ‘overriding’ the electron-withdrawing power of the ester group.
16. Enzyme sources: recombinant human MAO-A and MAO-B were purchased
from Sigma–Aldrich, recombinant human SSAO/VAP-1 (truncated; residues
34–763) was prepared by CSIRO, Parkville, Australia; full length human
recombinant DAO was provided by Professor Mitchell Guss, University of
Sydney.
3. Schwelberger, H. G. J. Neural Transm. 2007, 114, 757.
4. (a) Imamura, Y.; Kubota, R.; Wang, Y.; Asakawa, S.; Kudoh, J.; Mashima, Y.;
Oguchi, Y.; Shimizu, N. Genomics 1997, 40, 277; (b) Kaitaniemi, S.; Elovaara, H.;
Gron, K.; Kidron, H.; Liukkonen, J.; Salminen, T.; Salmi, M.; Jalkanen, S.; Elima,
K. Cell. Mol. Life Sci. 2009, 66, 2743.
5. Andres, N.; Lizcano, J. M.; Rodriguez, M. J.; Romera, M.; Unzeta, M.; Mahy, N. J.
Histochem. Cytochem. 2001, 49, 209.
6. (a) Kurkijarvi, R.; Adams, D. H.; Leino, R.; Mottonen, T.; Jalkanen, S.; Salmi, M. J.
Immunol. 1998, 161, 1549; (b) Stolen, C. M.; Yegutkin, G. G.; Kurkijarvi, R.; Bono,
P.; Alitalo, K.; Jalkanen, S. Circ. Res. 2004, 95, 50.
7. Precious, E.; Gunn, C. E.; Lyles, G. A. Biochem. Pharmacol. 1988, 37, 707; (b)
Lyles, G. A.; Chalmers, J. Biochem. Pharmacol. 1992, 43, 1409.
8. Yu, P. H.; Zou, D. M.; Davis, B. A. Biochem. Pharmacol. 1994, 47, 1055.
9. (a) Dunkel, P.; Balogh, B.; Meleddu, R.; Maccioni, E.; Gyires, K.; Mátyus, P. Expert
Opin. Ther. Patent 2011, 21, 1453; (b) McDonald, I. A.; Foot, J.; Yin, P.; Flening, E.;
van Dam, E. M. Annu. Rep. Med. Chem. 2007, 42, 229.
17. Compound 28 was also evaluated against the AOC-2 gene product, (human
transient) retinal amine oxidase and found to be inactive (IC50 >10 lM).
18. Figure generated using Molecular Operating Environment (MOE), 2011.10;
Chemical Computing Group Inc., 1010 Sherbooke St. West, Suite #910,
Montreal, QC, Canada H3A 2R7, 2011.
19. (a) Kitz, R.; Wilson, I. B. J. Biol. Chem. 1962, 237, 3245; (b) Holt, A.; Smith, D. J.;
Cendron, L.; Zanotti, G.; Rigo, A.; Di Paolo, M. L. Mol. Pharmacol. 2008, 73, 525.
20. Harrod, K. S.; Mounday, A. D.; Whitsett, J. A. Am J. Physiol. Lung Cell. Mol. Physiol.
2000, 278, L631.