The discovery of fused pyrrole carboxylic acids as novel, potent d-amino acid oxidase (DAO) inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.04.020

The 'NMDA hypofunction hypothesis of schizophrenia' can be tested in a number of ways. DAO is the enzyme primarily responsible for the metabolism of d-serine, a co-agonist for the NMDA receptor. We identified novel DAO inhibitors, in particular, acid 1, which demonstrated moderate potency for DAO in vitro and ex vivo, and raised plasma d-serine levels after dosing ip to rats. In parallel, analogues were prepared to survey the SARs of 1.

Full text of DOI:10.1016/j.bmcl.2008.04.020

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 3386–3391
The discovery of fused pyrrole carboxylic acids as novel, potent
D-amino acid oxidase (DAO) inhibitors
Tim Sparey,a,* Pravien Abeywickrema,d Sarah Almond,b Nick Brandon,b Noel Byrne,d
Alister Campbell,a Pete H. Hutson,b,c Marlene Jacobson,c Brian Jones,a Sanjeev Munshi,d
Danette Pascarella,c Andrew Pike,a G. Sridhar Prasad,d Nancy Sachs,c Melanie Sakatis,a
Vinod Sardana,d Shankar Venkatramane and Mary Beth Younge
aDepartment of Medicinal Chemistry, Merck Sharp & Dohme, The Neuroscience Research Centre, Terlings Park, Harlow,
Essex CM20 2QR, UK
bDepartment of Molecular & Cellular Neuroscience, Merck Sharp & Dohme, The Neuroscience Research Centre,
Terlings Park, Harlow, Essex CM20 2QR, UK
cDepartment of Schizophrenia Research, Merck & Co., Inc., Sumneytown Pike, West Point, PA, USA
dDepartment of Structural Biology, Merck & Co., Inc., Sumneytown Pike, West Point, PA, USA
eDepartment of Medicinal Chemistry, Merck & Co., Inc., Sumneytown Pike, West Point, PA, USA
Received 29 January 2008; revised 8 April 2008; accepted 10 April 2008
Available online 13 April 2008
Abstract—The ‘NMDA hypofunction hypothesis of schizophrenia’ can be tested in a number of ways. DAO is the enzyme primarily
responsible for the metabolism of D-serine, a co-agonist for the NMDA receptor. We identified novel DAO inhibitors, in particular,
acid 1, which demonstrated moderate potency for DAO in vitro and ex vivo, and raised plasma D-serine levels after dosing ip to rats.
In parallel, analogues were prepared to survey the SARs of 1.
Ó 2008 Elsevier Ltd. All rights reserved.
Schizophrenia is a condition that affects up to 1% of the
population. Current therapies are inadequate and one
new approach is to address the ‘NMDA hypofunction’
hypothesis1 by developing drugs to enhance NMDA
receptor function, for example, elevation of brain gly-
cine through inhibition of the glycine transporter
(GLYT1).2
bit D-amino acid oxidase (DAO), the enzyme primarily
responsible for metabolising D-serine.4
A number of reports profile compounds including indole-
s,5a benzisoaxoles5b,c and pyrrole/pyrazole derivatives5c,d
with potent in vitro activity for DAO inhibition and ben-
efits in animal models of cognitive and positive symptoms
of schizophrenia. There are striking similarities between
these classes of compounds; they are all low molecular
weight, carboxylic acids.
An alternative approach is to increase the brain concen-
tration of D-serine, a co-agonist of the NMDA receptor,
and would therefore be expected to increase the activa-
tion of the NMDA receptor.
Described herein is our approach to the identification of
potent DAO inhibitors and follow-up studies utilising
X-ray crystallography to design analogues suitable for
in vivo profiling.
Clinical evidence to support this hypothesis is mixed,
but some benefit was seen in patients when D-serine
was used as an add-on to atypical anti-psychotics.3
One approach to increase brain D-serine levels is to inhi-
Keywords: Schizophrenia; D-Amino acid oxidase (DAO) inhibitor.
*
Corresponding author at present address: Merck Sharp & Dohme,
Licensing & External Research Europe, Hertford Road, Hoddesdon
EN11 9BU, UK. Tel.: +44 1992 452838; fax: +44 1992 441907; e-
mail: tim_sparey@merck.com
Scheme 1. Reagents and condition: (a) EtO2CCH2N3, NaOEt, EtOH
(67%); (b) Toluene, reflux (77%); (c) NaOH (97%).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.04.020

T. Sparey et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3386–3391
3387
IC50 measurement and counter-screening assays. For
example, compound 1 was synthesised by a method de-
scribed previously,6 whereby condensation of furfural-
dehyde with ethyl azidoacetate followed by thermal
cyclisation of 31 furnished 1 in multi-gram quantities
(Scheme 1).
This methodology was adapted by using the appropriate
readily available heteroaromatic aldehydes to generate a
range of pyrrole carboxylic acid analogues to investigate
core changes represented by A (compounds 7–12) and
substituted derivatives B (compounds 13–18). To ex-
plore the SARs of other analogues represented by C
(compounds 19–25), aryl groups were introduced via Su-
zuki coupling of bromide 32 with the appropriate boro-
nic acid (Scheme 2).
Scheme 2. Reagents and conditions: (a) MeO2CCH2N3, NaOMe/
MeOH (70%); (b) Xylene, reflux (65%); (c) NaOH (95%); (d)
Pd(PPh3)4, Na2CO3, MeCN, H2O, reflux (43–78%).
For the preparation of fused heterocyclic pyrrole 12,
using SEM-protected pyrazole-2-carboxaldehyde in the
condensation and thermal cyclisation sequence afforded
33 which, after deprotection with HCl then saponifica-
tion, furnished 12. The utility of this method was en-
hanced through the availability of many N-substituted
pyrazole aldehydes which allowed the standard route to
be used to generate N-substituted acids 27–30 (Scheme 3).
Hits from the initial DAO inhibition assay7 were fol-
lowed up by performing further screening within the
Merck compound collection. An additional set of hits
was identified including a number of compounds previ-
ously reported as DAO inhibitors (e.g., compounds 3,5a
45b and 55c). A counter-screen assay confirmed selectiv-
ity over D-aspartate oxidase (DDO) the nearest homo-
logue of DAO. A selection of validated hits is shown
in Table 1.
Scheme 3. Reagents and conditions: (a) MeO2CCH2N3, NaOMe/
MeOH (60%); (b) Xylene (51%); (c) 2 N HCl in EtOH, reflux; (d)
NaOH (50–74%).
A HTS campaign using a subset of the Merck com-
pound collection identified a small number of DAO
inhibitors. Key hits were resynthesised for repeat testing,
Table 1. In vitro potencies of acids 1–57
Compound
h-DAO IC50 (nM)
h-DDO IC50 (nM)
1
2
3
141 73
>5000
343 84
>5000
>5000
745 377
4
5
507 236
>5000
>5000
238 99

3388
T. Sparey et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3386–3391
Table 2. In vitro potencies of compounds 1 and 6–187
From this subset of validated hits, compound 18 was
chosen to form the basis of a limited lead optimisation
campaign. Initially, heterocyclic variations and substi-
tuted analogues were synthesised to explore the SARs
of hit 1 with the aim of improving the potency for inhib-
iting DAO (Table 2).
Compound
h-DAO IC50 (nM)
1
141 73
>5000
The SAR of analogues 6–18 is intriguing. The removal
of the furan ring in 1 to pyrrole-2-carboxylic acid 6 re-
sulted in complete loss of potency. Replacing the furan
ring with other heterocycles such as thiophene analogues
8 and 10, and thiazole 9 resulted in compounds with
similar potencies. However, other heterocyclic replace-
ments such as oxazole 7, thiazole 11 and pyrazole 12
showed significant or complete losses in potency. Locat-
ing substituents on the furan or thiophene ring resulted
in two compounds (13 and 18) with reduced potency,
both with substituents on the 3-position of the fused het-
erocyclic core, but otherwise substituents on the 2-, 3- or
both positions lost potency (compounds 14–17). These
data suggest limited scope for the replacement of the
furan portion of the core with other heterocycles or
substituted heterocycles.
6
7
8
>5000
245 170
9
10
11
12
13
14
15
16
516 174
269 77
4204 436
>5000
The crystal structure of human DAO bound to com-
pound 1 was determined by replacing the bound benzo-
ate from the crystals of the ternary complex with FAD.9
Compound 1 occupies the pocket adjacent to bound
FAD, where benzoate is known to bind.10 The carboxyl-
ate group of 1 hydrogen bonds with guanidyl group of
Arg283 and the hydroxyl of Tyr228 (see Fig. 2 in Ref.
10). In addition, the N8 atom of 1 interacts with the
main-chain NH of Gly313. The volume of the binding
pocket occupied by 1 is not significantly different from
that occupied by benzoate, despite some conformational
differences in side chains observed around the binding
pocket.
The information from X-ray studies and the low average
molecular weight of hits identified from the HTS cam-
paign hinted at a small binding site consistent with an
enzyme capable of metabolising small amino acids such
as D-serine. Despite the losses in potency found when
small substituents were located off the core of acids 1
and 8, an attempt was made to explore space further
from the heterocyclic cores. The X-ray structure of 1:
DAO suggested space behind Tyr228 where hydropho-
bic groups could be accommodated if appended a suit-
able distance away from the core. Firstly, simple
derivatives of 1 were synthesised where a small range
of aromatic rings were appended from the two available
positions on the furan ring. All compounds synthesised
lost potency (Table 3).
1426 161
>5000
>5000
>5000
When consideration was given to the potency of 5,
where the position of the aryl group and linker con-
ferred increased potency compared with the parent (5
vs 6), it seemed plausible that potency could be gained
by accessing the same space in the simple acids presented
here. Thus, a simple selection of compounds was pre-
pared with groups appended from the nitrogen atom
adjacent to the pyrrole ring of compound 12 (com-
pounds 27–30). These compounds resulted in no gain
17
18
>5000
1153 236

T. Sparey et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3386–3391
3389
Table 3. In vitro potencies of acids 19–257
h-DAO IC50 (nM)
R
h-DAO IC50 (nM)
1
19
20
141 73
>5000
H
1
141 73
—
23
—
>5000
>5000
21
22
—
>5000
>5000
—
24
—
25
>5000
—
>5000
in potency relative to the NH parent (IC50s > 5 lM).
These data along with the results for compounds 19–
25 (Table 3) suggested that groups appended from the
core of acids such as 1 and 12 were ineffective at improv-
ing or even maintaining potency. This could be due to
the incorrect positioning of the appendages in 19–30 rel-
ative to the side chain of 5.
(Fig. 1c). After 24 h, enzyme activity had returned to base-
line levels in both cerebellum and kidney (data not
shown). Plasma levels of 1 after ip dosing at 25, 75, or
150 mg/kg revealed dose proportionality (Fig. 1d).
Compound 1 was identified from an HTS campaign and
found to significantly inhibit DAO activity in vitro. Proof
of principle in animals was demonstrated because 1 sig-
nificantly inhibited ex vivo DAO enzyme activity in kid-
ney and cerebellum. Consequently, D-serine levels in the
plasma were increased. The work done more than 30
years ago showed that systemic administration of D-
but not L-serine to rats induced nephrotoxicity, which
manifested as acute necrosis of the proximal tubules
and consequent aminoaciduria, proteinurea and glucosu-
ria.12 These effects were most noticeable 2 h after admin-
istration and were not apparent by 4–5 days when the
tubules were re-lined by new epithelium indicating that
the toxicity was reversible. Interestingly, others showed
that D-serine nephrotoxicity was attenuated by sodium
benzoate, an inhibitor of DAO.13 This study indicated
that the mechanism of D-serine nephrotoxicity may be re-
lated to the by-products of D-amino acid metabolism
(e.g., H2O2). These effects will require further monitoring
with the development of such inhibitors. This compound
therefore represents a useful tool to further explore the
role of D-serine in modulating central NMDA receptor
function. Limited lead optimisation efforts identified
equipotent analogues to 1 including compound 8. The re-
sults of the in vivo pharmacological characterisation of 8,
including the effects on central D-serine levels, will be re-
ported shortly (Fig. 2).14
Compounds 1 and 8 were further profiled to assess suit-
ability for in vivo studies to establish if a correlation ex-
isted between the compound exposure in plasma, DAO
enzyme inhibition and changes in D-serine levels. Firstly,
no off-target activities were found for 1 and 8 for the inhi-
bition of P450 enzymes (IC50s > 10 lM for CYP3A4,
CYP2D6 and CYP3C9) or binding to ion channels (IKR
or sodium channels; data not shown). In addition, com-
pound 1 was found not to be an antagonist at the glycine
site of NMDA receptors (data not shown).
To correlate D-serine levels with ex vivo tissue DAO en-
zyme inhibition, compound 1 was dosed at 25, 75 and
150 mg/kg ip to rats, and D-serine levels were measured
out to 24 h in the plasma whilst enzyme activity was deter-
mined in kidney and cerebellum at 1, 2 and 4 h.11 A dose
dependent increase in the D-serine levels versus the vehicle
was not observed in plasma; however, at 150 mg/kg of 1, a
significant increase in D-serine levels was observed versus
the vehicle at 4, 8, 16 and 24 h (Fig. 1a). Ex vivo measure-
ment of DAO enzyme inhibition revealed significant inhi-
bition versus the vehicle at all doses and time points of 1 in
kidney (Fig. 1b). In the cerebellum, significant inhibition
was observed except for the 25 mg/kg ip dose at 1 h

3390
(a) Plasma D-serine
T. Sparey et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3386–3391
vehicle
450
400
350
300
250
200
150
100
50
***
25 mg/kg
***
75 mg/kg
***
***
**
150 mg/kg
*
*
0
1
2
4
8
16
24
Time (hours)
(b) Ex vivo DAO activity in kidney
150
125
100
75
50
***
***
25
0
***
***
***
***
***
1
***
**
2
4
-25
Time (hours)
Figure 2. Interactions of compound 1 with hDAO. Part of the bound
FAD is shown in cyan. The dotted lines represent hydrogen bonding
interactions. The coordinates of the structure of hDAO bound to
compound 1 have been deposited in Protein Data Bank (accession
number 3CUK).
(c) Ex vivo DAO activity in cerebellum
110
100
90
***
**
80
70
60
50
40
30
20
10
0
*
Acknowledgments
***
***
***
***
The authors thank Luli Duguma, Ellen Crawford and
chemists at WuXi PharmaTech Co., Ltd (Shanghai,
China).
***
1
2
4
Time (hours)
References and notes
(d) Plasma levels of 1
1000
1. Lindsley, C. W.; Shipe, W. D.; Wolkenberg, S. E.;
Theberge, C. R.; Williams, D. L., Jr.; Sur, C.; Kinney,
G. G. Curr. Top. Med. Chem. 2006, 6, 771.
2. Sur, C.; Kinney, G. G. Curr. Drug Targets 2007, 8, 643.
3. (a) Tsai, G.; Yang, P.; Chung, L. C.; Lange, N.; Coyle, J.
T. Biol. Psychiatry 1998, 44, 1081; (b) Heresco-Levy, U.
Expert Opin. Emerg. Drugs 2005, 10, 827.
4. (a) Nagata, Y. Experimentia 1992, 48, 753; (b) Hashimoto,
A.; Nishikawa, T.; Konno, R.; Niwa, A.; Yasumura, T.;
Oka, T.; Takahashi, K. J. Neurosci. Lett. 1993, 152, 33.
5. (a) WO2007070892, filed 11th December, 2006.; (b)
WO2005089753, filed 11th March, 2005.; (c)
WO2003039540, filed 12th November, 2002.; (d) Adage,
T.; Trillat, A. C.; Quattropani, A.; Perrin, D.; Cavarec, L.;
25 mg/kg
75 mg/kg
150 mg/kg
100
10
1
0
1
2
3
4
5
6
7
8
´
Shaw, J.; Guerassimenko, O.; Giachetti, C.; Greco, B.;
Time (hours)
Chumakov, I.; Halazy, S.; Roach, A.; Zaratin, P. Eur.
Neuropsychopharmacol. 2007. August 1 (Epub ahead of
print).
Figure 1. Plasma D-serine, DAO enzyme activity and plasma concen-
tration over time following administration of 25, 75 or 150 mg/kg of 1
ip to rats. (a) D-serine levels in plasma; (b) Ex vivo DAO enzyme
activity in kidney; (c) Ex vivo DAO enzyme activity in cerebellum (the
effects of 1 on plasma D-serine and DAO enzyme activity were
analysed using a two-way ANOVA to analyse time and treatment
6. Krutosikova, A.; Kovac, J.; Dandarova, M.; Lesko, J.;
Ferik, S. Collect. Czech. Chem. Commun. 1981, 45, 2564.
7. Assayed in CHO cells transfected with DAO as described
in Brandish, P. E.; Chiu, C. S.; Schneeweis, J.; Brandon,
N. J.; Leech, C. L.; Kornienko, O.; Scolnick, E. M.;
Strulovici, B.; Zheng, W. J. Biomol. Screen. 2006, 115,
481, All compounds were screened a minimum of 4 times.
effect followed by a Bonferroni post test for individual group
comparisons. *p < 0.05, **p < 0.01 or ***p < 0.001 dose vs vehicle).
(d) Plasma levels of 1 out to 8 h.

T. Sparey et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3386–3391
3391
8. For compound 1: (400 MHz 1H NMR DMSO-d6) 6.58
(1H, s), 6.69 (1H, s), 7.76 (1H, s), 11.51 (1H, s), 12.36 (1H,
s); (13C NMR, DMSO-d6) 94.3, 98.1, 123.1, 127.4, 145.6,
147.5, 161.3; m/z (APÀ) 150 [MÀH]+.
10. Kawazoe, T.; Tsuge, H.; Pilone, M. S.; Fukui, K. Protein
Sci. 2006, 1512, 2708.
11. Determination of ex-vivo DAO activity was completed
with n = 4–5 rats per group according to Almond, S.
L.; Fradley, R. L.; Armstrong, E. J.; Heavens, R. B.;
Rutter, A. R.; Newman, R. J.; Chiu, C. S.; Konno, R.;
Hutson, P. H.; Brandon, N. J. Mol. Cell. Neurosci.
2006, 324, 324.
9. Prasad, S.; Munshi, S. Human DAO was crystallized as a
ternary complex with bound FAD and benzoate. DAO (aa 1–
347) at 2 mg/ml in buffer containing 50 mM sodium phos-
phate, pH 6.6, 200 mM FAD and 200 mM benzoate, was
mixed with equal volume of crystallisation buffer containing
100 mM Tris-HCl pH 7.7, 150 mM Tri-Potassium Citrate and
15% (w/v) PEG 3350. The crystals obtained at room
12. Carone, F. A.; Ganote, C. E. Arch. Pathol. 1975, 99,
658.
13. Williams, R. E.; Lock, E. A. Toxicology 2005, 207, 35.
14. Uslaner, J. M.; Smith, S. M.; Pascarella, D.; Yao, L.;
McNaughton, C. H.; Huszar, S. L.; Mullins, C. H.;
Kandebo, M.; Hinchliffe, R. M.; Jacobson, M. A.; Sparey,
T. J.; Jones, A. B.; Brandon, N.; Venkatraman, S.; Young,
M. B.; Sachs, N. A.; Hutson, P. H. J. Pharmacol. Exp.
Ther., submitted for publication.
˚
temperature diffracted X-rays to 2.5 A resolution and
belonged to space group C2 with cell dimensions a =
˚
˚
˚
187.80 A, b = 51.2 A, c = 153.3 A, a = 90.0°, b = 110.4°,
c = 90.0°. The asymmetric unit contains four molecules
associated as 2 dimers. Crystals were soaked in solution
containing compound 1 for 48 hours prior to data collection.