The discovery of novel, potent and highly selective inhibitors of inducible nitric oxide synthase (iNOS)

Article abstract of DOI:10.1016/j.bmcl.2011.02.061

By careful analysis of experimental X-ray ligand crystallographic protein data across several inhibitor series we have discovered a novel, potent and selective series of iNOS inhibitors exemplified by compound 8.

Full text of DOI:10.1016/j.bmcl.2011.02.061

Bioorganic & Medicinal Chemistry Letters 21 (2011) 2468–2471
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The discovery of novel, potent and highly selective inhibitors of inducible
nitric oxide synthase (iNOS)
David R. Cheshire a, , Anders Åberg d, Gunilla M. K. Andersson d, Glen Andrews b, Haydn G. Beaton a,
⇑
Timothy N. Birkinshaw a, Nigel Boughton-Smith b, Stephen Connolly a, Tony R. Cook a, Anne Cooper c,
Sally L. Cooper a, David Cox a, John Dixon a, Nigel Gensmantel c, Peter J. Hamley a, Richard Harrison a,
Paul Hartopp a, Helena Käck d, Paul D. Leeson a, Timothy Luker a, Antonio Mete a, Ian Millichip a,
David J. Nicholls b, Austen D. Pimm a, Steve A. St-Gallay c, Alan V. Wallace b
a Department of Chemistry, AstraZeneca Charnwood, Bakewell Road, Loughborough LE11 5RH, United Kingdom
b Department of Bioscience, AstraZeneca Charnwood, Bakewell Road, Loughborough LE11 5RH, United Kingdom
c Department of Physical and Metabolic Sciences, AstraZeneca Charnwood, Bakewell Road, Loughborough LE11 5RH, United Kingdom
d Discovery Chemistry, Pepparedsleden 1, 431 83 Mölndal, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
By careful analysis of experimental X-ray ligand crystallographic protein data across several inhibitor ser-
ies we have discovered a novel, potent and selective series of iNOS inhibitors exemplified by compound 8.
Ó 2011 Elsevier Ltd. All rights reserved.
Received 6 January 2011
Revised 14 February 2011
Accepted 14 February 2011
Available online 18 February 2011
Keywords:
Nitric oxide
Nitric oxide synthases
Competitive iNOS inhibitors
Whilst the in vivo pharmacokinetic properties of 1 were
Nitric oxide (NO), produced from
L
-arginine by the nitric oxide
suitable for establishing efficacy in a variety of disease models5–7
synthases, plays a crucial role in cellular signalling. Over-production
of NO is associated with a number of debilitating diseases; numer-
ous reports implicate excess NO––from the inducible nitric oxide
synthase isoform (iNOS)––in inflammatory conditions such as rheu-
matoid arthritis and asthma.1–4 Potent, selective and safe small
molecule inhibitors of iNOS would be valuable tools in further
understanding the potential of inhibiting this enzyme to deliver
therapeutic agents. We have previously reported on a variety of
competitive iNOS inhibitors containing a cis-amidine motif, which
1 is reactive toward glutathione (c-L-glutamyl-L-cysteinylglycine);
this raises the possibility of reactive metabolite formation and idi-
osyncratic adverse drug reactions in man. We concluded that a
new direction was required as we had explored the full potential
of competitive inhibitors containing the amidine motif. We set
out to investigate non-amidine inhibitors of iNOS in the hope that
this might provide a better starting point for a program aimed at
producing inhibitors with drug-like properties.
The search for non-amidine iNOS inhibitors presented two
major challenges. The first was to maintain potency despite
dispensing with the charge reinforced bi-dentate hydrogen bond
between the amidine and the conserved glutamic acid (Glu3778).
The second challenge was to understand and maintain the selectiv-
ity against the endothelial NOS (eNOS) in order to avoid unwanted
blood pressure effects.9 The origins of the selectivity we had
achieved in 1 were, at the time, a subject of heated debate.10
AR-C102222 1 is greater than 2500-fold selective for iNOS versus
eNOS even though the three isoforms share almost complete ami-
no acid conservation, and structural similarity, around the arginine
binding site.
mimics the guanidinium of
L
-arginine.5 These investigations
resulted in the discovery of the potent, and iNOS selective spirocyclic
1,2-dihydro-4-quinazolinamines exemplified by 1 (AR-C102222).
O
F
H
N
N
N
CN
N
F
NH2
1 (AR-C102222)
Our hypothesis was that potency could be retained by keeping
an aromatic ring over the haem group, and selectivity could be
engineered by occupying the same binding site region accessed
⇑
Corresponding author. Tel.: +44 1509 644823.
E-mail address: dave.cheshire@astrazeneca.com (D.R. Cheshire).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.02.061

D.R. Cheshire et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2468–2471
2469
by 1. To expedite this we screened other similar non-amidines
whereupon compound 211 was revealed as a weak inhibitor (IC50
1.2 lM), but with little selectivity. A crystal structure showed that
a halogen (method B), both with a chiral alcohol. Depending on
the chiral alcohol further steps (i–iv) were required to complete
the synthesis.
the amino-carboxamide of 2 resides in the main channel of the en-
zyme cavity. The nitro-group displaces the conserved Glu377
which swings to point away from the haem to expose its side-chain
methylenes, thus forming a ligand induced binding pocket. The ni-
tro-group appears to hydrogen bond to the backbone–NH of
Met374 amide, which had previously bonded to the Glu377 car-
boxylate. A number of non-amidine small molecules have since
been reported to bind to the NOS enzymes using this new pocket,
for example substituted nitro-indazoles.12,13 It is proposed that
strong binding is achieved by stacking of the ligand over the haem
and formation of a number of hydrogen bonds to the protein back-
bone. This binding is confirmed by X-ray crystal structures.
Extension of the amino-methyl chain of compound 4 by one
carbon atom, to provide propylamine 5, results in a remarkable
90-fold improvement in potency; further chain homologation pro-
vided no additional benefit. Gratifyingly 5 is greater than 4500-fold
selective over eNOS and 210-fold selective against nNOS (Table 2).
However, this compound shows a very high drop in cellular po-
tency of around 400-fold, which is much higher than the 50-fold
expected due to competition with the intracellular concentration
of the NOS enzyme’s substrate L-arginine. As the molecule’s high
lipophilicity (calculated log P = 4.1) was thought to be responsible
one of the two chlorine atoms in 5 was successfully replaced with a
cyanide group providing improved cellular potency (6a, IC50
0.7 lM). The selectivity of 6a against eNOS was greater than
10,000-fold! The cyanide group also appears to make a hydrogen
bond interaction with the backbone –NH of Met374 amide.
A crystal structure of 6b, the des-N-methyl analogue of 6a,
shown in Figure 115 demonstrates that the compound binds with
the aryl group over the haem, with the cyanide group in the shifted
Glu377 pocket, and the benzyl group in the Gln263 shifted ‘gluta-
mine specificity pocket’. However, the position of the glutamine
residue indicates that this pocket is similar to, but not exactly iden-
tical to that previously reported.10
Me
Cl
Me
N
CONH2
NH2
NH2
O
O
N
NO2
H
Cl
2
3
4
The clue to establishing the required selectivity originated from
the crystal structure of compound 3. As expected the aminopyri-
Examining the chiral centre we found that the (R)-enantiomers
were consistently 10 to 100-fold more potent than the (S) enanti-
omers––this was confirmed periodically with key compounds and
crystal structures. Changing the linking atom from oxygen 6b to
sulphur 6c improved iNOS potency, but was accompanied by an
unacceptable reduction in nNOS selectivity; this observation is
general to the series. More sterically demanding non-polar groups,
for example trifluoromethyl 6d or methoxy 6e, are tolerated, but
with some loss of potency. Such large groups probably describe
the limit in size for substituents at the C-3 position (R1). Moving
the C-3 substituent to the C-4 (R2) position, as in 6f, removed much
of the iNOS potency. Replacing the phenyl ring––which projects to-
ward the Gln263 pocket––with a small alkyl chain 6g lowers po-
tency and removes most of the eNOS selectivity. This supports
the original hypothesis that selectivity is only achieved by moving
the glutamine residue in iNOS.
dine, an amidine mimic, binds to Glu377 in a similar way to L-argi-
nine, but it was noticed that the N-benzyl substituent, by shifting
the isoform conserved residue Gln263, forms a new pocket and
gives this simple compound (3, IC50 0.79 lM) a 12-fold selectivity
against eNOS. Examination of a large number iNOS ligands co-crys-
tallised with the oxygenase domain of mouse iNOS, and human or
bovine eNOS, revealed that an isoform specific pocket was created
by bulky inhibitors of human iNOS. Although Gln263 is invariant in
the three NOS isoforms the key to specificity is the ability of the
Gln263 residue in iNOS to shift and expand the binding pocket.
Due to differences in the second shell of amino acids around the
binding site this can only occur in the iNOS isoform.10
The combination of the fragments, retention of the basic amine
(to bind the haem carboxylic acids) and replacement of the
nitro-group resulted in 4, which is gratifyingly both potent (IC50
1.9 lM) and selective (50-fold against eNOS and nNOS, see Table
Compound 6a is stable in human microsomal and hepatocyte
preparations, but is not stable in rat hepatocytes; accordingly it
is not bioavailable when dosed orally to a rat. The principal routes
of metabolism were identified as N-de-methylation and oxidation
at C-4. These metabolic routes were blocked by substituting the C-
4 hydrogen atom with fluorine, and removing the N–Me group to
2). Compound 4 became the starting point for an optimisation
program which culminated in the discovery of exciting novel and
selective iNOS inhibitors.
The compounds were synthesised in essentially two steps
(Table 1). The first was a coupling reaction (Scheme 1); either a
Mitsunobu coupling (method A), or an aromatic displacement of
Table 1
Structures and synthesis methods
Compd
Structure
R3
Synthesis
U
R1
R2
R4
R5,a
X
Y
W
Coupling method
Final step
Saltb
5
C
C
C
C
C
C
C
C
C
N
N
Cl
Cl
Cl
Cl
–CF3
MeO–
H
Cl
Cl
H
H
H
H
H
H
Cl
H
F
Cl
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Et
Me
Me
H
OH
OH
F
F
F
OH
OH
F
F
Cl
Cl
O
O
O
S
O
O
O
O
O
S
–NMeBoc
–Cl
–N3
A
A
B
B
B
A
A
B
B
B
B
i
C4H4O4
HCl
C2H2O4
C4H4O4
HCl
C2H2O4
C4H4O4
HCl
C2H2O4
C4H4O4
C2H2O4
6a
6b
6c
6d
6e
6f
6g
6h
7
–CN
–CN
–CN
–CN
–CN
–CN
–CN
–CN
–CN
–CN
ii
iii
iii
i
i
iv
i
iii
iii
i
H
–N3
Me
Me
H
Me
H
–NMeBoc
–NMeBoc
Cl
–NMeBoc
–N3
Ph
Ph
Me
–CF3
H
H
H
H
–N3
–NHBoc
8
5-Isoxazoyl
O
a
Stereochemistry is R as drawn, for all compounds except 6g which is racemic.
Salts were prepared by (neutralising if required) adding the relevant acid (1 equiv)––fumaric (C4H4O4), oxalic (C2H2O4) acid or HCl in a suitable solvent at room
b
temperature.

2470
D.R. Cheshire et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2468–2471
Table 2
Biological activities and selectivity profiles
log D
iNOS (
l
M)a
Sel. eNOSa
Sel. nNOSa
Cell (l
M)a
SET (lM)
NA (lM)
CY2D6 (lM)
hERG (lM)
1
4
5
0.79
NA
NA
1.7
0.8
1.3
1.3
0.9
0.9
0.4
0.8
0.7
ꢀ0.5
0.041
1.9
0.022
0.0090
0.0040
0.0040
0.071
0.110
0.960
0.051
>3000
>53
22
52
210
140
270
33
5
65
104
18
35
1
NA
NA
0.0
4.5
2.1
6.7
10.0
5.7
NA
44.0
4.4
2.6
70
NA
NA
0.0
0.5
15.0
17.0
28.0
12.0
NA
7.1
16.0
7.4
NA
NA
NA
1.0
0.4
4.1
1.6
0.3
NA
0.4
0.34
5
0.1
NA
NA
10
NA
NA
NA
NA
NA
NA
>10
NA
>10
NA
8.7
0.7
0.9
0.2
2.7
9.5
NA
3.6
0.7
0.6
0.72
>4500
>10,000
>10,000
>10,000
1400
6a
6b
6c
6d
6e
6f
6g
6h
7
NA
100
170
>10,000
>10,000
>10,000
0.0060
0.0040
0.0054
28
1.3
8
>100
>14
aFor nitric oxide enzyme and cellular screening methods.5,14
R1
R1
Table 3
Pharmacokinetic profile of compounds 1, 6a, 6h and 8
R2
R2
R4
U
R4
Compd Rat
Human
Rat dosed
Rat dosed orally
+
R5
Y
W
intravenously
X
X
N
H
R3
R3
Hepsa Micsa Hepsa Clb Vdss
T1/2 (h) Bioavailability (%)
c
1
5
NA
4
6
<3
1.4
<1
1
57
94
49
56
4.5
6.6
13
1.1
1.3
4.3
1.0
75
NA
57
49
Scheme 1. Synthesis of compounds 5–8. Coupling method A: Ph3P, DEAD, THF, 0–
25 °C 12 h. Coupling method B: Either NaH (1 equiv), in DMF (or THF), 1 h, 40 °C; or,
K2CO3 in DMF (or MeOH), 40 °C. Final steps: (i) 4 N HCl, in 1,4-dioxane (or THF),
25 °C; (ii) NaI, acetone, reflux, 18 h then 40% aq MeNH2, 25 °C, 5 h; (iii) Ph3P, 10% aq
THF (v/v), reflux, 1 h; (iv) NaN3, DMSO, 60 °C 16 h then Ph3P, 10% aq THF (v/v),
reflux, 1 h.
6a
6h
8
54
8.9
15
<3
3.4
a
Mics: microsomes, units of
cells.
lL/min/mg. Heps: hepatocytes, units of l
L/min/106
b
Clearance of drug from plasma, units of mL/min/kg.
Volume of distribution at steady state, units of L/kg.
c
potent CYP2D6 inhibitor (Table 2, IC50 0.34 lM); this proved a
general property of the series. These new iNOS inhibitors bear a
significant structural resemblance to the selective serotonin reup-
take inhibitor Fluoxetine™, and whilst Fluoxetine™ is not an inhib-
itor of any of the three NOS isoforms these compounds do bind to
the noradrenaline (NA) and serotonin transporters (SET). Com-
pounds were also tested for hERG activity (Table 2), but were
shown to be very weak inhibitors.
Removal of these unwanted properties was achieved simply by
lowering the compounds’ lipophilicity. This was initially attempted
by replacing the phenyl ring over the haem with a pyridine ana-
logue 7 which retained potency and selectivity for iNOS, but did
not abrogate the SET or NA binding. A meaningful reduction in
lipophilicity was achieved by replacing the phenyl ring projecting
toward the Gln263 pocket with a heterocyclic ring. The combina-
tion of a 3-trifluoro-2-pyridyl group and an isoxazole gave 8 which
is a potent and selective iNOS inhibitor, and gratifyingly removed
the NA, SET and CYP2D6 activities. The only shortcoming of 8
was the lack of nNOS selectivity; this is likely due to the smaller
size of the isoxazole (compared with the phenyl substituent) bind-
ing in the ‘glutamine selectivity pocket’.
A feature of this series of novel iNOS inhibitors is their stability
in in vitro metabolising systems (human and rat), and their suit-
ability for in vivo testing in rat models on inflammatory disease
(Table 3). The series had rat half lives of greater than 1 h with mod-
erate to good bioavailability after oral dosing. A significant fraction
of drug is free in plasma, and all have acceptable solubility in
buffered aqueous media under equilibrium conditions at 37 °C.
Compound 6h demonstrated significant efficacy in a collagen in-
duced arthritis disease model, in female Dark Agouti rats, when
Figure 1. X-ray0 structure of 6b15 into the oxygenase domain of mouse iNOS,
resolution 2.65 ÅA. The Glu371 and Gln257 (mouse numbering) positions when L-
arginine is bound are shown in white. This emphasises the movements of these
residues on binding of our non-amidine inhibitors. Surface representations in white
and green illustrate the effect the position of Gln257 has on the size of the binding
pocket. The Gln257 position found in the work reported previously for amidine like
inhibitors is shown in purple.10
give the primary amine 6h. As expected a marked improvement in
the half-life and bioavailability (Table 3) was observed. Whilst 6h
maintained cellular potency, selectivity over nNOS was somewhat
reduced; this observation is seen in a number of analogues con-
taining fluorine at C-4. To address concerns about reactive metab-
olites rats were dosed with radio-labelled 6h; no metabolites
consistent with displacement of either the fluorine or chlorine by
glutathione were observed. Further profiling revealed that 6h is a
dosed twice daily at 75 lmol/kg. Pronounced analgesic activity
was observed, and histological analysis showed significant reduc-
tion in inflammation and erosion scores.16

D.R. Cheshire et al. / Bioorg. Med. Chem. Lett. 21 (2011) 2468–2471
2471
Chem. 2004, 47, 3320; Beaton, H.; Hamley, P.; Nicholls, D. J.; Tinker, A. C.;
Wallace, A. V. Bioorg. Med. Chem. Lett. 2001, 11, 1023; Beaton, H.; Boughton-
Smith, N.; Hamley, P.; Ghelani, A.; Nicholls, D. J.; Tinker, A. C.; Wallace, A. V.
Bioorg. Med. Chem. Lett. 2001, 11, 1027.
In conclusion we have demonstrated that careful analysis of
experimental crystallographic ligand–protein data across several
inhibitor series allowed a novel, potent and selective series of iNOS
inhibitors, exemplified by 6h and 8, to be designed. The series has
excellent properties for further investigation of the role of induc-
ible nitric oxide synthase in inflammatory disease models. Further
publications will discuss the journey to provide compounds with a
profile more suitable for human dosing.
8. Human iNOS numbering is used throughout this Letter. For practical reasons
most X-ray structures were determined from soaking the ligand into the
crystallised oxygenase domain of mouse iNOS.
9. Ruetten, H.; Thiemermann, C. In Nitric Oxide and Septic Shock in Nitric Oxide;
Ignarro, L. J., Ed.; Academic Press: San Diego, CA, 2000. 747.
10. Through
a collaboration, a thorough understanding of the origins of this
isoenzyme specificity have been published. Garcin, E. D.; Arvai, A. S.; Rosenfeld,
R. J.; Kroeger, M. D.; Crane, B. R.; Andersson, G.; Andrews, G.; Hamley, P. J.;
Mallinder, P. R.; Nicholls, D. J.; St-Gallay, S. A.; Tinker, A. C.; Gensmantel, N. P.;
Mete, A.; Cheshire, D. R.; Connolly, S.; Stuehr, D. J.; Aberg, A.; Wallace, A. V.;
Tainer, J. A.; Getzoff, E. D. Nat. Chem. Biol. 2008, 4, 700.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.02.061.
11. Cowart, M.; Kowaluk, E. A.; Daanen, J. F.; Kohlhaas, K. L.; Alexander, K. M.;
Wagenaar, F. L.; Kerwin, J. F., Jr. J. Med. Chem. 1998, 41, 2636; For a similar
discovery see; Le Bourdonnec, B.; Leister, L. K.; Ajello, C. A.; Cassel, J. A.; Seida,
P. R.; O’Hare, H.; Gu, M.; Chu, Guo-H.; Tuthill, P. A.; DeHaven, R. N.; Dolle, R. E.
Bioorg. Med. Chem. Lett. 2008, 18, 336.
References and notes
12. Elguero, J.; Alkorta, I.; Claramunt, R. M.; Lopez, C.; Sanz, D.; Santa Maria, D.
Bioorg. Med. Chem. Lett. 2009, 17, 8027.
1. For an overview of NO biology see Perspectives on NO Physiology and Pathology;
Richardson, V. J., Wallace, A. V., Eds.; Transworld Research Network: Kerela,
India, 2009.
2. Vallance, P.; Leiper, J. Nat. Rev. Drug Disc. 2002, 1, 939.
3. Bian, K.; Murad, F. Front. Biosci. 2003, 8, D264.
4. Thippeswamy, T.; McKay, J. S.; Quinn, J. P.; Morris, R. Histol. Histopathol. 2006,
21, 445.
5. Tinker, A. C.; Beaton, H. G.; Boughton-Smith, N.; Cook, A. R.; Cooper, S. L.;
Fraser-Rae, L.; Hallam, K.; Hamley, P.; McInally, T.; Nicholls, D. J.; Pimm, A. D.;
Wallace, A. V. J. Med. Chem. 2003, 46, 913.
13. Cottyn, B.; Acher, F.; Ramassamy, B.; Alvey, L.; Lepoivre, M.; Frapart, Y.; Stuehr,
D.; Mansuy, D.; Boucher, Jean-L.; Vichard, D. Bioorg. Med. Chem. 2008, 16, 5962.
14. For the preparation of compounds, excepting 6f and 4, including structural data
and screening methods: Compounds 5, 6a–b, 6d and 6h: Cheshire, D.; Connolly,
S.; Cox, D.; Hamley, P.; Mete, A.; Pimm, A. WO2001062704, 2001; 6c, Cheshire,
D.; Connolly, S.; Cox, D.; Millichip, I. WO 2001062721, 2001: 6g, Cheshire, D.;
Connolly, S.; Cox, D.; Mete, A. WO2001062714, 2001; 6e, Cheshire, D.;
Connolly, S.; Cox, D.; Hamley, P.; Mete, A.; Pimm, A. WO2001062704, 2001;
7
and 8, Birkinshaw, T.; Cheshire, D.; Connolly, S.; Luker, T.; Mete, A.
6. LaBuda, C. J.; Koblish, M.; Tuthill, P.; Dolle, R. E.; Little, P. J. Eur. J. Pain 2006, 10,
505.
7. Connolly, S.; Aberg, A.; Arvai, A.; Beaton, H. G.; Cheshire, D. R.; Cook, A. R.;
Cooper, S.; Cox, D.; Hamley, P.; Mallinder, P.; Millichip, I.; Nicholls, D. J.;
Rosenfeld, R. J.; St-Gallay, S. A.; Tainer, J.; Tinker, A. C.; Wallace, A. V. J. Med.
WO2003011830, 2003. Compounds 6f and 4 were made using the methods
described in these references.
Supplementary Data.
A procedure for 6h can be found in the
15. RCSB Protein Data Bank; entry code 2Y37.
16. Unpublished data.