The rational design of inhibitors of nitric oxide formation by inducible nitric oxide synthase

Article abstract of DOI:10.1016/j.bmcl.2007.02.018

A series of compounds was rationally designed as inhibitors of dimer formation of the inducible isoform of nitric oxide synthase, and subsequent nitric oxide production. The conformation of two fragments obtained from a crystal structure was utilized to d

Full text of DOI:10.1016/j.bmcl.2007.02.018

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2505–2508
The rational design of inhibitors of nitric oxide
formation by inducible nitric oxide synthase
Marc Whitlow, Marc Adler, David Davey, Qinglan Huang, Sunil Koovakkat,
John F. Parkinson, Eric Pham, Mark Polokoff, Wei Xu,
Shendong Yuan and Gary Phillips*
Berlex Biosciences, 2600 Hilltop Drive, Richmond, CA 94804, USA
Received 18 January 2007; revised 5 February 2007; accepted 7 February 2007
Available online 9 February 2007
Abstract—A series of compounds was rationally designed as inhibitors of dimer formation of the inducible isoform of nitric oxide
synthase, and subsequent nitric oxide production. The conformation of two fragments obtained from a crystal structure was utilized
to design a tether connecting those same two fragments. The resulting compounds were potent dimerization inhibitors that bound to
the enzyme in a similar conformation as the fragments.
Ó 2007 Elsevier Ltd. All rights reserved.
Inducible nitric oxide synthase (iNOS) has been impli-
cated in a variety of immunological diseases such as
multiple sclerosis and rheumatoid arthritis. Two consti-
tutive nitric oxide synthase (NOS) isoforms have been
identified, endothelial and neuronal NOS (eNOS and
bNOS, respectively) which are important for normal
function. The search for a selective inhibitor of iNOS
has been a priority within the industry and the work
addressing a specific enzyme inhibitor has been diffi-
cult.1 We and others have described work, leading to
inhibitors of nitric oxide (NO) production, that does
not lead directly through enzyme inhibition, but results
by not allowing the enzyme to adopt the active dimeric
form.2,3
In our previous report, we described the optimization of
a series of imidazolepyrimidines and the discovery of 1,
with an IC50 of 0.2 nM in a whole cell assay.4 The crystal
structure of this inhibitor and the murine iNOS mono-
meric oxygenase domain (iNOS D114) in a complex
was discussed. The imidazole binds directly with the iron
and the compound loops back to position the benzodi-
oxolane moiety over the porphyrin. In fact, the soaking
H
O
N
N
O
O
N
H3C
N
N
N
1
Figure 1. Crystal structure of 4-(imidazol-1-yl)phenol and piperonyl-
amine overlapping (gold carbon atoms, PDB entry 2ORQ) with 1 and
bound to murine iNOSD114 (green carbon atoms, PDB entry 2ORO).
Keywords: Dimerization inhibitors; iNOS.
*
Corresponding author. Tel.: +1 925 372 7882; fax: +1 510 262
7844; e-mail: garybphillips@comcast.net
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.02.018

2506
M. Whitlow et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2505–2508
of the two key fragments in the form of 4-(imidazole-1-yl)-
phenol and piperonylamine gave a very nearly identical
orientation (Fig. 1). A superposition of the two struc-
tures shows that the two fragments overlap with 1, bind-
ing in the murine iNOS monomer active site. The
arylimidazole moieties overlap nearly completely, while
the benzodioxolanes occupy the same space, but are
slightly out of plane from one another. The adopted
conformation of these two fragments within the enzyme
suggested some alternative structures that would inter-
act in a similar manner and potentially disrupt the pro-
tein–protein interaction critical for the dimerization
process and subsequent activation of the enzyme
(Fig. 2). As we wanted the two fragments to be present-
ed to the enzyme in a U-shape, our initial attempts fo-
cused on a central template we had used in the past to
present functionality in a very similar fashion, 2,6-disub-
stituted heterocycles.5
N
N
O
O
X
R
Cl
Cl
Y
Z
O
X
R
O
Y
Z
Scheme 1. Reagents and conditions: (i) Sesamol, Cs2CO3, DMSO,
50 °C, 18 h; (ii) 4-(imidazole-1-yl)phenol, Cs2CO3, DMSO, 90 °C, 18 h.
O
O
O
O
a
b
N
OH
The compounds with both aromatic groups attached
through a heteroatom were prepared by standard meth-
ods (Scheme 1).6 Sequential reaction of the appropriate-
ly substituted phenols with the appropriately substituted
difluoropyridine or dichloropyrimidine gave the diphen-
oxyheterocycle. Addition to the pyrimidines typically
gave mixtures that were sometimes difficult to separate.
N
CN
OH
O
N
O
N
O
O
c
N
Cl
The pyrimidines with one of the aromatic groups at-
tached through a methylene were prepared through the
corresponding pyrimidine (Scheme 2). The nitrile of
3,4-(methylenedioxy)phenylacetonitrile was converted
to the amidine. Reaction of the amidine with diethyl
malonate gave the dihydroxypyrimidine, which was
reacted with phosphorus oxychloride to give the dichlo-
ropyrimidine. Sequential displacement of the chlorides
gave the desired compounds.
N
R
O
N
N
Cl
Scheme 2. Reagents and conditions: (a) (i)HCl, EtOH, 0 °C; (ii)NH3,
EtOH; (iii)diethyl malonate, sodium methoxide, EtOH, reflux, 5 h; (b)
POCl3, N,N-diethylaniline, 45 °C, 3 h; (c) (i)4-(imidazole-1-yl)phenol,
K2CO3, DMSO; (ii)amine, DMSO, 40 °C, 24 h.
Compounds were assayed for inhibition of iNOS in a
whole cell assay using human A172 cells induced with
cytokines, which has been described previously.4,7 NO
formation was monitored spectrophotometrically with
the Griess reagent. The A172 data in the tables refer
to the IC50 determination as an average of at least two
separate experiments.
Table 1. Inhibition of NO production in A172 cells7
O
O
O
X
Y
Z
R
N
O
N
The initially prepared compounds showed good activity
(Table 1). The unsubstituted pyridine, 2a, had an IC50 of
78 nM. Additional substitution of the pyridine, as in 2b,
was detrimental. Pyrimidines were also investigated as
the central core and as long as one of the key substitu-
ents was straddled by a nitrogen (2c–e), similar potency
2
Compound
X
Y
Z
R
A172 (nM) IC50
2a
2b
2c
2d
2e
2f
N
N
N
N
N
CH
CF
CH
CH
N
CH
CF
N
H
78
CH3
H
450
55
N
CH3
H
90
74
CH
N
CH
N
N
CCH3
H
>10,000
130
N
O
2g
N
H
O
N
O
N
O
N
O
N
O
was realized. The single 4,6-disubstituted compound, 2f,
was inactive. Placement of a hydrogen between the
two aromatic moieties may make the U-shaped confor-
mation less favorable. Unlike the pyridine, additional
Figure 2. Initial compound design hypothesis.

M. Whitlow et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2505–2508
2507
substitution of the pyrimidine gave a compound with
similar activity (2d).
These two conformational differences may contribute
to the decrease in activity. Additional differences, substi-
tution on the pyrimidine and position of the pyrrolidine,
may also contribute to the increase in activity of com-
pound 1.
In our earlier work on analogously substituted pyridines
for Factor Xa,5 we had found that the analogs that pre-
sented the binding substituent in the para position, as is
the imidazole, tended to be suboptimal from a potency
perspective, presumably due to an inability to best place
a substituent. In contrast, the 3-substituted compounds
on rotation could adopt a wider range of conformations
and were more potent. Preparation of a compound in
this case, 3, showed this not to carry over in this exam-
ple, as the compound did not attain an IC50 at 10 lM.
Taking into account the structural information and the
SAR of 1, we tried to effect the dihedral angle of the phe-
nyl and imidazole ring by introducing substitution on
the phenyl (Table 2). Addition of a small substituent,
such as fluoride (4a), gave a compound with similar
activity, and a larger substituent, trifluoromethyl (4c),
gave slightly less activity than the hydrogen substituted
compound, 2d. The intermediate chloride (4b) gave the
most active compound in the series.
O
N
O
N
To circumvent the chemical selectivity problem associat-
ed with the pyrimidine synthesis, an alternate set of
compounds was prepared, in which an aromatic moiety
was attached to the central template by a carbon
(Table 3). In addition, water-solubilizing groups could
be added in the 6-position of the pyrimidine to modulate
lipophilicity as the second chloride could be displaced
and the crystal structure did not indicate any detrimen-
tal steric interactions. Unfortunately, significant activity
was lost with chloro substituted analog, 5a. This loss of
activity was surprising in light of the activity of 2e and
2g. The addition of nitrogen-containing substituents
brought back some of the activity.
O
N
O
3
When the structures of the fragments are overlapped
with 2c, the overlap of the benzodioxolane moiety is
not as good as that of the previously reported com-
pound 1 (Fig. 3). This is consistent with the experimen-
tal results, as the activity difference between 2c and 1 is
approximately two orders of magnitude. The overall
structure is similar with the imidazole binding to the
iron and the benzodioxolane positioned over the por-
phyrin. Comparing the structure of 2c to the two inde-
pendent fragments, the imidazole overlaps well, but
the planes defined by the phenyl and the imidazole rings
are further from planarity (6° vs 13° for the dihedral an-
gle between the imidazole and phenyl rings). The benzo-
dioxolane is in the same area over the porphyrin, but the
conformation is different. Whereas the planes defined by
the benzodioxolane and phenyl of the two fragments are
closer to perpendicular (the angle between the plane nor-
mals is 65.1°), in 2c they are closer to parallel (35.2°).
Table 2. Inhibition of NO production in A172 cells7
O
O
N
O
N
O
H3C
R
N
N
4
Compound
R
A172 (nM) IC50
2d
4a
4b
4c
H
F
90
92
Cl
CF3
19
220
Table 3. Inhibition of NO production in A172 cells7
O
O
N
R
N
O
N
N
5
Compound
R
A172 (nM) IC50
5a
5b
5c
Cl
Morpholine
N(CH3)(CH2)3N(CH3)2
>10,000
830
220
Figure 3. Crystal structure of 4-(imidazol-1-yl)phenol and piperonyl-
amine (PDB entry 2ORQ) overlapping with 2c bound to murine
iNOSD114 (PDB entry 2ORR).

2508
M. Whitlow et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2505–2508
Sigal, N. A.; Vergona, R.; Whitlow, M.; Young, T.; Devlin,
J. J. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 1506.
3. (a) Sennequier, N.; Wolan, D.; Stuehr, D. J. J. Biol. Chem.
1999, 274, 930; (b) Ohtsuka, M.; Konno, F.; Honda, H.;
Oikawa, T.; Ishikawa, M.; Iwase, N.; Isomae, K.; Ishii, F.;
Hemmi, H.; Sato, S. J. Pharmacol. Exp. Ther. 2002, 303, 52;
(c) Chida, N.; Hirasawa, Y.; Ohkawa, T.; Ishii, Y.; Sudo,
U.; Tamura, K.; Mutoh, S. Eur. J. Pharmacol. 2005, 509,
71; (d) Gahman, T. C.; Lang, H.; GHerbert, M. R.; Thayer,
A. M.; Hassig, C. A.; Noble, S. A.; Cousins, R. D.; Zhuang,
H.; Santos, C. R.; Chen, X. U.S. Patent Appl. 2006/
0116515.
4. Davey, D. D.; Adler, M.; Arnaiz, D.; Eagen, K.; Erickson,
S.; Guilford, W.; Kenrick, M.; Morrissey, M. M.; Ohlmey-
er, M.; Pan, G.; Paradkar, V. M.; Parkinson, J.; Polokoff,
M.; Saionz, K.; Santos, C.; Subramanyam, B.; Vergona, R.;
Wei, R. G.; Whitlow, M.; Ye, B.; Zhao, Z.; Devlin, J.;
Phillips, G. J. Med. Chem. 2007, in press.
Although we assumed these compounds were acting as
dimerization inhibitors, we had no data to differentiate
them from direct enzyme inhibitors. Compound 2c
was assayed in standard human enzyme assays8 and
found to have very low activity (>70 lM) against the
inducible isoform, certainly not sufficient to explain
the activity in the whole cell assay. No activity against
either bNOS or eNOS was observed at 100 lM in the
corresponding assays.
In summary, we have identified compounds that potent-
ly inhibit nitric oxide formation by cytokine induced
A172 cells. The compounds were designed by first
obtaining a crystal structure of the key fragments bound
in the enzyme. A tether was designed that could induce a
similar conformation of the two substituents. Although
the compounds could be further modified, an alternate
tether was identified that had significant advantages
and is reported in the subsequent communication.9
5. Phillips, G. B.; Davey, D. D.; Eagen, K. A.; Koovakkat, S.;
Liang, A.; Ng, H. P.; Pinkerton, M.; Trinh, L.; Whitlow,
M.; Beatty, A. M.; Morrissey, M. M. J. Med. Chem. 1999,
42, 1749.
6. Phillips, G. B.; Guilford, W. J.; Buckman, B. O.; Davey, D.
D.; Eagen, K. A.; Koovakkat, S.; Liang, A.; McCarrick,
M.; Mohan, R.; Ng, H. P.; Pinkerton, M.; Subramanyam,
B.; Ho, E.; Trinh, L.; Whitlow, M.; Wu, S.; Xu, W.;
Morrissey, M. M. J. Med. Chem. 2002, 45, 2484.
7. Assay:4 A172 Cells were harvested and plated. Eighteen to
24 h later, 222 U/ml of human interferon-gamma, 22 ng/ml
of human tumor necrosis factor-alpha, 2.2 ng/mL of human
interleukin 1-b, and the appropriate concentration of the
compound were added. After incubation for 18–24 h, an
aliquot of the culture medium was removed and tested for
nitrite concentration with Griess reagent. IC50 values were
calculated from a log-logit analysis of the data.
8. Bredt, D. S.; Snyder, S. H. Proc. Natl. Acad. Sci. U.S.A.
1990, 87, 682.
9. Wei, R. G.; Adler, A.; Davey, D.; Ho, E.; Mohan, R.;
Polokoff, M.; Tseng, J.-L.; Whitlow, M.; Xu, W.; Yuan, S.;
Phillips, G. Bioorg. Med. Chem. Lett. 2007, accepted for
publication.
Acknowledgment
The authors thank Jerry Dallas for his NMR work and
structure elucidation.
References and notes
1. (a) Tafi, A.; Angeli, L.; Venturini, G.; Travagli, M.; Corelli,
F.; Botta, M. Curr. Med. Chem. 2006, 13, 1929; (b) Salerno,
I.; Sorrenti, V.; Di Giacomo, C.; Romeo, G.; Siracusa, M.
A. Curr. Pharm. Des. 2002, 8, 177.
2. McMillan, K.; Adler, M.; Auld, D. S.; Baldwin, J. J.;
Blasko, E.; Browne, L. J.; Chelsky, D.; Davey, D.; Dolle,
R. E.; Eagan, K. A.; Erickson, S.; Feldman, R. I.; Glaser,
C.; Mallari, C.; Morrissey, M. M.; Ohlmeyer, M. H. J.;
Pan, G.; Parkinson, J. F.; Phillips, G. B.; Polokoff, M. A.;