1,7- and 2,7-naphthyridine derivatives as potent and highly specific PDE5 inhibitors

Article abstract of DOI:10.1016/S0960-894X(03)00440-2

Novel 1,7- and 2,7-naphthyridine derivatives, designed by the introduction of nitrogen atom into the phenyl ring of previously reported 4-aryl-1-isoquinolinone derivatives, were disclosed as a new structural class of potent and specific PDE5 inhibitors. Among them, 2,7-naphthyridine 4c showed potent PDE5 inhibition (IC₅₀=0.23 nM) and one of the best PDE5 specificities against PDEs1-4,6 (>100,000-fold selective versus PDE1-4, 240-fold selective vs PDE6). This compound showed more potent relaxant effects on isolated rabbit corpus cavernosum (EC₃₀=5.0 nM) than Sildenafil (EC₃₀=8.7 nM). The compound 4c (T-0156) was selected for further biological and pharmacological evaluation of erectile dysfunction.

Full text of DOI:10.1016/S0960-894X(03)00440-2

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2341–2345
1,7- and 2,7-Naphthyridine Derivatives as Potent and Highly
Specific PDE5 Inhibitors
Tatsuzo Ukita,a,* Yoshinori Nakamura,a Akira Kubo,a Yasuo Yamamoto,a
Yasunori Moritani,a Kunio Saruta,a Takanori Higashijima,a Jun Kotera,b
b
Kotomi Fujishigeb Michino Takagi,b Kohei Kikkawab andKenji Omori
aDiscovery Research Laboratory, Tanabe Seiyaku Co., Ltd., 3-16-89, Kashima, Yodogawa, Osaka 532-8505, Japan
bDiscovery Research Laboratory, Tanabe Seiyaku Co., Ltd., 2-2-50, Kawagishi, Toda, Saitama 335-8505, Japan
Received28 January 2003; accepted19 April 2003
Abstract—Novel 1,7- and 2,7-naphthyridine derivatives, designed by the introduction of nitrogen atom into the phenyl ring of
previously reported 4-aryl-1-isoquinolinone derivatives, were disclosed as a new structural class of potent and specific PDE5
inhibitors. Among them, 2,7-naphthyridine 4c showedpotent PDE5 inhibition (IC 50=0.23 nM) andone of the best PDE5 specifi-
cities against PDEs1–4,6 (>100,000-foldselective versus PDE1–4, 240-foldselective vs PDE6). This compoundshowedmore
potent relaxant effects on isolatedrabbit corpus cavernosum (EC 30=5.0 nM) than Sildenafil (EC30=8.7 nM). The compound 4c
(T-0156) was selectedfor further biological andpharmacological evaluation of erectile dysfunction.
# 2003 Elsevier Science Ltd. All rights reserved.
Introduction
isoforms, notably PDE1 andPDE6. Thus the discovery of
new PDE5 inhibitors with improvedselectivities against
these PDE isoforms andgreater potency for PDE5 inhi-
Cyclic nucleotide phosphodiesterases (PDEs) regulate
physiological processes by the hydrolysis of intracellular
secondmessengers, cyclic guanosine monophosphate
(cGMP) andcyclic adenosine monophosphate (cAMP)
to their corresponding 50-nucleoside monophosphates.
PDEs have been classifiedinto at least 11 identified
families to date,1 basedon their amino acidsequences,
substrate specificities, endogeneous and exogeneous
regulators, andpharmacological properties. PDE5,
cGMP-binding cGMP-specific PDE, is the primary
enzyme that responsible tor the degradation of cGMP
in human corpus cavernosum, andinhibition of this
enzyme cause penile erection via relaxation of the vascular
6
bition have recently attractedconsiderable interest.
We have reporteda series of 1-isoquinolinone deriva-
tives as a new structural class of PDE5 inhibitors, and
selectedcompound 2 (T-1032) for further biological and
pharmacological evaluation.7
2,3
smooth muscle andcorpus cavernosum tissue.
Sildenafil (1, viagra1) is a first efficacious andorally active
PDE5 inhibitor, which is usedfor the treatment of male
erectile dysfunction.4 Despite its effectiveness, 1 shows
clinically significant adverse effects such as headache,
5
In order to obtain more potent and specific PDE5 inhi-
bitors with better pharmacokinetic properties as backup
compounds, we have done further synthetic studies on
compounds structurally relating to 2. In previous paper,7
we expectedthat isoquinolinone ring in 2 mimickedthe
purine nucleus of cGMP andthat the pendant aryl group
at the 4-position of 2 filleda space occupiedby the cyclic
nausea, cutaneous flushing, andvisual disturbances.
It
has been postulatedthat some of these side effects may be
ascribedto the modest selectivities towardthe other PDE
*Corresponding author. Tel.: +81-6-6300-2566; fax: +81-6-6300-
2564; e-mail: t-ukita@tanabe.co.jp
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00440-2

2342
T. Ukita et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2341–2345
phosphate group of cGMP. Beavo et al. reportedthat
N7-nitrogen of cGMP pointedout by an arrow in Figure
1 contributes to binding at the catalytic site of PDE5 as a
hydrogen acceptor.8 Thus, we envisionedthat the addi-
tional introduction of nitrogen atom to a suitable position
of isoquinolinone ring for the favorable interaction with
PDE5 might enhance PDE5 inhibition andimprove
selectivities over the other PDE isoforms. Futhermore, we
expectedthat this modification might leadto improve-
ment of PK profiles of compounds based on lowering the
log P value (clog P value: 2, 3.611; 3a, 2.652).9
derivatives (3 and 4) obtainedby the introduction of
nitrogen atom into the phenyl ring of isoquinolinone in 2.
Chemistry
The synthetic methods of 1,7-naphthyridine derivatives
3 are outlinedin Scheme 1. Lithiation of pyridine-2-
carboxamide 5 with n-BuLi followedby the trapping
with 3,4,5-trimethoxybenzaldehyde, and subsequent
cyclization under acidic condition gave 6 in 51% yield.
The oxidation of 6 with KMnO4 in aqueous KOH
This paper describes the synthesis and structure–activity
relationships (SARs) of 1,7- and2,7-naphthyridne
afforded keto acid 7 in a goodyield. Alkylation of
with di-tert-butyl bromomalonate followedby the
7
DBU-mediated cyclization, treatment with HCl/AcOEt
solution, andthe heating in dioxane gave a key inter-
mediate 8 in 55% yield. The reaction of 9, which is
obtainedby the hydrolysis of 8, with amines followed
by the esterification afforded 1,7-naphthyridine-3-car-
boxylate derivatives 10. The nucleophilic substitution
reactions of 10 with appropriate alkoxides or amine
gave 3a–i.
The syntheses of 2,7-naphthyridines 4 are summarized
in Scheme 2. 8-Methoxy-2,7-naphthyridine 16 was pre-
paredfrom N,N-diisopropyl-2-methoxynicotinamide 11
according to the analogous procedures for 10 as descri-
bedabove. The treatment of 16 with POCl3 afforded
8-chloro compound 17 in 79% yield. The addition-
elimination reaction of 17 with various types of nucleo-
philes such as alkoxides or amines gave 4a–e.
Figure 1.
Scheme 1. Reagents andconditions: (a) (1) n-BuLi, 3,4,5-trimethoxybenzaldehyde, THF, (2) concd HCl-dioxane; (b) KMnO4, pyridine, 2N NaO-
Haq; (c) (1) BrCH(COOt-Bu)2, KHCO3, DMF, (2) DBU, toluene, (3) 4N HCl/AcOEt, (4) dioxane, reflux; (d) 2N NaOHaq, DMF; (e) (1) R1NH2,
DMI or THF, (2) MeI, K2CO3, DMF or TMSCH2N2, MeOH, THF; (f) 4N HCl/AcOEt–CHCl3; (g) For 3a–f,h,i: R2H, NaH, THF or DMF. For
3g: R2H, DMF; (h) 4N HCl/AcOEt–CHCl3.

T. Ukita et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2341–2345
2343
Scheme 2. Reagents andconditions: (a) (1) n-BuLi, TMEDA, 3,4,5-trimethoxybenzaldehyde, THF, (2) AcOH-dioxane; (b) KMnO4, KOHaq; (c) (1)
BrCH(COOt-Bu)2, KHCO3, DMF, (2) DBU, toluene, (3) AcOH, reflux; (d) 2N NaOHaq, MeOH; (e) (1) 2-methyl-4-picolylamine, THF, (2) MeOH,
DEAD, PPh3, CH2Cl2; (f) POCl3, DMF, CHCl3; (g) For 4a–c: RH, NaH, THF or CH2Cl2. For 4d,e: RH, dioxane; (h) 4N HCl/AcOEt–CHCl3.
Table 1. Structures andPDE inhibitions of 1,7-naphthyridine deri-
vatives
Biological Results and Discussion
The compounds reported in this paper were firstly eval-
uatedfor the inhibitory activities against the three differ-
ent isoforms of PDEs isolatedfrom canine lung (PDE1
andPDE5), andbovine retina (PDE6) shown in Tables 1
and2 . Compounds selected on the basis of the PDE5
inhibitory activities were next evaluatedthe relaxant
effects on rabbit corpus cavernosum, andinvestigatedfor
the three additional different isoforms of PDEs isolated
from bovine adrenal gland (PDE2), canine heart (PDE3),
andcanine lung (PDE4) shown in Table 3.
1
CompdR
R2
n
PDE inhibition, IC50, nMa
PDE1
PDE5 PDE6
Table 1 summarizes the SARs of 1,7-naphthyridine
derivatives 3a–i. In comparison to the isoquinolinone,
T-1032 (2, Table 3), 1,7-naphthyridine 3a showed
increasedPDE5 inhibitory activity andimprovedselec-
tivity over PDE1. As a result of the modification of the
substituent at the 7-position, we foundthat 7-picolyloxy
derivatives 3b–d exhibitedpotent PDE5 inhibitory
activities. The comparison of compounds 3b–d showed
that the order of potencies toward PDE5 was 4-picolyl
(3d)>3-picolyl (3c)>2-picolyl (3b). These results indi-
cate that the spatial disposition of basic nitrogen atom
at the 7-position might be important to exert high PDE5
inhibition. Introduction of methyl group in the ortho
position to the pyridine nitrogen of picolyl group of 3d,
3a
3b
3c
3d
3e
3f
2
2
2
2
2
2
2
1
1
40,000
0.51
22
230
38
20,000 11
12,000
45,000
33,000
ndb
2.3
0.38
0.22
1.2
11
8.6
ndb
22
which was expectedto reduce the inhibitory potency
10
3g
3h
3i
41,000
22,000
>100,000
1.8
towardP450,
did not affect PDE5 inhibitory activity
(3e, IC50=0.22 nM). Next we examinedthe effect of
substituents at the 2-position of 3. One carbon elonga-
tion of the 2-picolyloxy group of 3e resultedin the loss of
activity (3f, IC50=1.2 nM). The replacement etheral oxy-
gen of 3e at the 2-position with N–H group resultedin the
decrease of activity (3g, IC50=1.8 nM). 2-Pyrazinyloxy
compound 3h exhibitedalmost the same potency as 3e,
although 2-pyrimidinyloxy compound 3i hadlower
activity than 3e (3h, IC50=0.29 nM; 3i, IC50=1.9 nM).
0.29
1.9
7.7
42
aIC50 values were determined from the logarithmic concentration–
inhibition curve (at least four points). The value is given as the mean
of at least two duplicate experiments.
bnd=not determined.

2344
T. Ukita et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2341–2345
Table 2. Structures and PDE inhibitions of 2,7-naphthyridine deri-
vatives
4c displayed enhanced PDE5 potency (IC50=0.23
nM) andimprovedselectivities over PDE1 andPDE6
(PDE5/PDE1 >100,000, PDE5/PDE6=240). In com-
parison with the corresponding 1,7-naphthyridine 3i
(IC50=1.9 nM, PDE5/PDE6=22), 2,7-naphthyridine
4c showedmore potent PDE5 inhibitory activity and
improvedPDE5/PDE6 selectivity. These results may
imply that the location of the hydrogen bond accep-
tor atoms (nitrogen andcarbonyl oxygen) of
2,7-naphthyridine 4 is more suitable than that of
1,7-naphthyridine 3 for the favorable interaction with
PDE5. Althouth the introduction of secondary amine
(4d) resultedin markedloss of PDE activities, incor-
poration of primary amine (4e) maintainedthe mod-
est PDE5 inhibition.
CompdR
n
PDE inhibition, IC50, nMa
PDE1
PDE5
PDE6
4a
4b
4c
4d
4e
0
2
1
0
2
25,000
0.31
4.4
Next we selectedsix compounds ( 3e,i and 4b,c,e) on the
basis of PDE5 potencies andisozyme selectivities over
PDE1 and PDE6 for further evaluation of three addi-
tional PDEs (PDE2, PDE3, andPDE4) inhibitory
activities andrelaxant effects on isolatedrabbit corpus
cavernosum, as shown in Table 3. These compounds
showedmodest to goodselectivities for PDE5 against
PDEs1-4. It shouldbe notedthat PDE5 selectivities of
2,7-naphthyridine 4b,c,e versus PDE6 have been greatly
20,000
>100,000
>100,000
19,000
0.52
0.23
94
72
56
1400
280
enhanced(IC
ratio>100), since the visual dis-
50
turbances associatedwith Sildenafil 1 are believedto be
the results of its low selectivity over PDE6 (IC50
ratioꢀ8). To our knowledge, compound 4c is one of the
most potent andspecific PDE5 inhibitors disclosedto
date. These compounds also possessed the relaxant
effects on rabbit corpus cavernosum, andcompound 4c
showedbetter efficacy than 1 (4c, EC30=5.0 nM ; 1,
EC30=8.7 nM).
2.6
aIC50 values were determined from the logarithmic concentration–
inhibition curve (at least four points). The value is given as the mean
of at least two duplicate experiments.
As shown in Table 2, 2,7-naphthyridine 4a exhibited
almost the same PDE5 inhibitory activity andiso-
zyme selectivities over PDE1 andPDE6 as 1,7-naph-
thyridine 3e. One carbon elongation of the
substituent at the 8-position improvedthe selectivity
over PDE6 without decreasing PDE5 potency (4b:
IC50=0.52 nM, PDE6/PDE5=140), in contrast to
1,7-naphthyridine 3f. These results suggest that
appropriate modification of 2,7-naphthyridines at the
8-position may be expectedto leadmore selective
compounds with high potency. Among the com-
pounds we synthesized, 8-pyrimidinyloxy compound
Conclusion
Novel 1,7- and 2,7-naphthyridine derivatives, designed
by the introduction of nitrogen atom into the phenyl
ring of previously reported4-aryl-1-isoquinolinone
derivatives, were disclosed as a new structural class of
potent andspecific PDE5 inhibitors. Among them, 2,7-
naphthyridine 4c was apparently more potent
(IC50=0.23 nM) andspecific ( >100,000-foldselective
vs PDE1-4, 240-foldselective vs PDE6) than Sildenafil
Table 3. PDE Inhibitions andrelaxant effects on isolatedrabbit corpus cavernosun
PDE inhibition, IC50, nMa
Relaxant effect
30, nMb
CompdPDE1
PDE2
PDE3
PDE4
PDE5
PDE6
EC
8.6
42
72
56
280
29
3e
3i
4b
4c
4e
33,000
>100,000
20,000
>100,000
19,000
270
40,000
25,000
>100,000
>100,000
78,000
>100,000
>100,000
>100,000
>100,000
5300
30,000
19,000
50,000
67,000
450
0.22
1.9
0.52
0.23
2.6
53
37
22
5.0
14
8.7
7.9
1 (Sildenafil)
2 (T-1032)
43,000
9700
>100,000
>100,000
11,000
3300
3.6
1.0
3000
28
aIC50 values were determined from the logarithmic concentration–inhibition curve (at least four points). The value is given as the mean of at least
two duplicate experiments.
bEC30 values were determined from the logarithmic concentration–inhibition curve. The value is given as the average of at least two
experiments.

T. Ukita et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2341–2345
2345
1. 4c also showedmore potent relaxant effects on iso-
latedrabbit corpus cavernosum ( 4c, EC30=5.0 nM; 1,
EC30=8.7 nM) than 1. 4c (T-0156) was selectedfor
further biological andpharmacological evaluation of
erectile dysfunction.
13729. (b) Corbin, J. D.; Francis, S. H.; Webb, D. J. Urology
2002, 60 (suppl 2B), 4.
3. (a) Czarniecki, M.; Ahn, H.-S.; Sybertz, E. J. Annu. Rep.
Med. Chem. 1996, 31, 61. (b) Eardley, I. Exp. Opin. Invest.
Drugs 1997, 6, 1803. (c) Stamford, A. W. Annu. Rep. Med.
Chem. 2002, 37, 53.
4. (a) Langtry, H. D.; Markham, A. Drugs 1999, 57, 967. (b)
Moreland, R. B.; Goldstein, I.; Kim, N. N.; Traish, A. Trends
Endocrinology Metab. 1999, 10, 97.
References and Notes
5. (a) Moreira, S. G.; Brannigan, R. E.; Spitz, A.; Orejuela,
F. J.; Lipshultz, L. I.; Kim, E. D. Urology 2000, 56, 474. (b)
Kloner, R. A. Am. J. Cardiol. 2000, 86 (2A), 57 F. (c) Bressler,
S. Surv. Ophthalmol. 1999, 44, 153.
1. (a) Francis, S. H.; Turko, I. V.; Corbin, J. D. Pro-
gress in Nucleic Acid Research and Molecular Biology
2000, 65, 1. (b) Beavo, J. A.; Conti, M.; Heaslip, R. J. Mol.
Pharmacol. 1994, 46, 399. (c) Beavo, J. A. Physiol. Rev. 1995,
75, 725. (d) Juilfs, D. M.; Soderling, S.; Burns, F.; Beavo, J. A.
Rev. Physiol. Biochem. Pharmacol. 1999, 135, 67. (e) Conti,
M.; Jin, S.-L. C. Prog. Nucl. Acid Res. Mol. Biol. 2000, 63, 1.
(f) Fujishige, K.; Kotera, J.; Michibata, H.; Yuasa, K.; Take-
bayashi, S.; Okumura, K.; Ohmori, K. J. Biol. Chem. 1999,
274, 18438. (g) Fawcett, L.; Baxendale, R.; Stacey, P.;
McGrouther, C.; Harrow, I.; Soderling, S.; Hetman, J.; Beavo,
J. A.; Phillips, S. C. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
3702.
6. Kingman, S. Drug Disc. Today 2000, 5, 320.
7. Ukita, T.; Nakamura, Y.; Kubo, A.; Yamamoto, Y.; Mor-
itani, Y.; Saruta, K.; Higashijima, T.; Kotera, J.; Takagi, M.;
Kikkawa, K.; Omori, K. J. Med. Chem. 2001, 44, 2204.
8. Beltman, J.; Becker, D. E.; Butt, E.; Jensen, G. S.;
Rybalkin, S. D.; Jastorff, B.; Beavo, J. A. Mol. Pharmacol.
1995, 47, 330.
9. PCModels version 4.72; Daylight Chemical Information
Systems, Inc., Mission Viejo, CA.
10. Chiba, M.; Jin, L.; Neway, W.; Vacca, J. P.; Tata, J. R.;
Chapman, K.; Lin, J. H. Drug. Metab. Dispos. 2001, 29, 1.
2. (a) Corbin, J. D.; Francis, S. H. J. Biol. Chem. 1999, 274,