Optimization of purine based PDE1/PDE5 inhibitors to a potent and selective PDE5 inhibitor for the treatment of male ED

Article abstract of DOI:10.1016/j.bmcl.2005.02.083

In search of a PDE5 inhibitor for erectile dysfunction, an SAR was developed from a PDE1/PDE5 purine series of leads, which had modest PDE5 potency and poor isozyme selectivity. A compound (41) with PDE5 inhibition and in vivo activity similar to sildenafil was discovered from this effort. In addition, purine 41 demonstrated superior overall PDE isozyme selectivity when compared to the approved PDE5 inhibitors sildenafil, vardenafil, and tadalafil, which may result in a more favorable side-effect profile.

Full text of DOI:10.1016/j.bmcl.2005.02.083

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2365–2369
Optimization of purine based PDE1/PDE5 inhibitors to a
potent and selective PDE5 inhibitor for the treatment of male ED
Craig D. Boyle,* Ruo Xu,* Theodros Asberom, Samuel Chackalamannil, John W. Clader,
William J. Greenlee, Henry Guzik, Yuequing Hu, Zhiyong Hu, Claire M. Lankin,
Dmitri A. Pissarnitski, Andrew W. Stamford, Yuguang Wang, Jeffrey Skell,
Stanley Kurowski, Subbarao Vemulapalli, Jairam Palamanda, Madhu Chintala,
Ping Wu, Joyce Myers and Peng Wang
Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
Received 19 November 2004; revised 24 February 2005; accepted 28 February 2005
Abstract—In search of a PDE5 inhibitor for erectile dysfunction, an SAR was developed from a PDE1/PDE5 purine series of leads,
which had modest PDE5 potency and poor isozyme selectivity. A compound (41) with PDE5 inhibition and in vivo activity similar
to sildenafil was discovered from this effort. In addition, purine 41 demonstrated superior overall PDE isozyme selectivity when
compared to the approved PDE5 inhibitors sildenafil, vardenafil, and tadalafil, which may result in a more favorable side-effect
profile.
Ó 2005 Elsevier Ltd. All rights reserved.
O
N
O
N
Phosphodiesterase type 5 (PDE5) inhibitors increase lev-
els of cyclic guanylate cyclase (cGMP) in the nitric oxide
(NO) pathway of penile erection, and have been devel-
oped for the treatment of male erectile dysfunction
(ED).1 Our objective is to develop potent and selective
PDE5 inhibitors that improve upon the isozyme selecti-
vity profile of the three marketed PDE5 inhibitors silde-
nafil (Viagra),2 vardenafil (Levitra),3 and tadalafil
(Cialis)4 (see Fig. 1). For example, the leading drug on
the market, sildenafil, suffers from poor selectivity over
PDE6.5 Undesired visual side effects have been attrib-
uted to this low isozyme selectivity.6 Our ideal candidate
should improve upon the PDE6 selectivity of sildenafil
and vardenafil, and improve the PDE11 selectivity over
tadalafil. Increased isozyme selectivity may result in a
more favorable side-effect profile.
N
O
S
HN
O
S
HN
N
N
N
O
O
N
N
N
O
N
O
sildenafil
vardenafil
O
N
O
H
5
3 N
R
N
6
2
N
Bn
N
1
N
N
N
7
H
H
O
O
tadalafil
1: R = Me, PDE5 = 80 nM, PDE1/5 = 1.3
2: R = Et, PDE5 = 19 nM, PDE1/5 = 342
O
We had synthesized initial lead compounds in a previous
PDE1/PDE5 program for the treatment of hyperten-
sion.7 The tricyclic guanine 1 was discovered to have
modest PDE5 potency, but poor selectivity versus
Figure 1. Lead PDE5 inhibitors.8
PDE1. Based on a comparison of guanine analogs 1
and 2, we believed that the N-5 ethyl group could im-
prove PDE5 selectivity over compounds containing the
N-5 methyl group (Fig. 1).8 Most compounds in our
previous PDE1/PDE5 program had a methyl group at
N-5, and due to synthetic difficulty, very little SAR
Keywords: PDE5 inhibitor; Erectile dysfunction; Male ED; Purine.
*
Corresponding authors. Tel.: +1 908 740 3503; fax: +1 908 740 7152
(C.D.B.); tel.: +1 908 740 3483; fax: +1 908 740 7152 (R.X.); e-mail
addresses: craig.boyle@spcorp.com; ruo.xu@spcorp.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.02.083

2366
C. D. Boyle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2365–2369
O
O
EtO2C
H2N
R
c
R
O
N
N
N
N
N
N
a, b
N
N
Bn
Bn
Bn
Cl
N
N
H
Bn
Bn
Bn
I
II
III
O
IV: R3= Bn
V: R3 = H
R
N
N
R3
d, e
N
Bn
f
N
N
R2
R1
Scheme 1. Synthesis of C-7/N-5 derivatives. Reagents and conditions: (a) RNCO, Et3N, toluene; (b) MeONa, MeOH; (c) POCl3, D; (d)
H2NCH(R1)CH(R2)OH, DIEA, NMP, 120 °C; (e) MsCl, Et3N; (f) Pd(OH)2/C, HCO2NH4, MeOH, D.
was developed at the N-5 position. Work in the current
project focused on developing SAR at C-7 and N-5, fol-
lowed by exploration of the N-1, N-3, and C-2 positions
to obtain selectivity versus PDE1 and PDE6 isozymes.
Table 2, minor modifications on N-5 substituents in
the N-1 benzyl series resulted in compounds with little
PDE5 potency (11–16) compared with the N-5 ethyl
derivative 6. Similarly, in the N-1 non-substituted series
(17–22), most non-ethyl compounds had poor PDE5 po-
tency. Only the closely related methyl (18) and isopropyl
(20) derivatives demonstrated some PDE5 activity, and
as shown by methyl analog 18, the PDE1/PDE5 selecti-
vity was poor compared to ethyl analog 10.
The synthesis of C-7 and N-5 derivatives is summarized
in Scheme 1. Aminoimidazole I9 was treated with an iso-
cyanate and cyclized to form the purine core II. The C-6
chloride III was obtained by reaction of II with POCl3.
Condensation of the chloride III by an aminoalcohol
followed by one-pot mesylation/cyclization afforded
the tricyclic guanine IV. Non-substituted N-1 deriva-
tives V were produced by debenzylation of IV.
Although the N-5 ethyl and C-7 benzyl compounds 6
and 10 exhibited promising PDE5 inhibition and selec-
tivity against PDE1, all of the derivatives displayed in
Tables 1 and 2 had poor selectivity versus PDE6. To ad-
dress this issue, modifications at N-1 and C-2 were
made, as outlined in Scheme 2. The C-2 position was
modified by halogenation of intermediate VI,11 and
alkynyl analogs were synthesized by a subsequent Sona-
gashira coupling reaction. A surprising result occurred
Several C-7 analogs were synthesized in the N-1 benzyl
(IV) and non-substituted N-1 series (V). In the N-1 ben-
zyl series (Table 1), some bulk was required for PDE5
potency as the n-propyl, phenyl, and benzyl derivatives
4–6 had similar binding data, but the ethyl compound
3 was inactive. However, in the non-substituted N-1 ser-
ies, the C-7 benzyl compound 10 had improved PDE5
activity and selectivity versus PDE1. Therefore, the C-
7 benzyl group was used in further SAR studies at other
positions in the molecule.10
Table 2. N-5 modificationsa
O
5
3N
R
N
With the C-7 benzyl group in place, a number of modi-
fications were made at the N-5 position. As shown in
1
N
N
N
7
R'
Table 1. C-7 modificationsa
O
5
3 N
Compound
R
R0
PDE5 IC50 PDE1/5 PDE6/5
(nM)a
N
1
N
N
N
7
11
6
Me
Et
Bn 6% I
Bn 32
—
—
5.3
—
—
—
—
—
—
R'
>313
—
—
R
12
13
14
15
16
17
18
10
19
20
21
22
allyl
i-Pr
c-Pr
Bn 17% I
Bn 13% I
Bn 12% I
Compound
R
R0 PDE5 IC50 PDE1/5 PDE6/5
(nM)a
—
—
CF3CH2 Bn 7% I
Bn 17% I
3
Et Bn 4% I
n-Pr (racemic) Bn 35
—
—
MeO
22H%H
—
—
4
>286
—
>313
4.9
1.3
5.3
I
H100
H4.5
5
Ph
Bn
Et
Bn 40
Bn 32
H33
Me
Et
11
2.3
1.8
6
>2222
—
—
7
>303
2.4
0.7
1.0
1.8
n-Pr
i-Pr
c-Pr
H25%
H22
H21%
I
—
8
n-Pr (racemic) H60
>167
59
>2222
4.0
9
10
i-Pr
Bn
H28
H4.5
I
I
—
—
—
—
CF3CH2 H34%
a Ref. 8. % I = % inhibition @ 100 nM.
a Ref. 8. % I = % inhibition @ 100 nM.

C. D. Boyle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2365–2369
2367
during the boron tribromide promoted demethylation of
intermediate VII. The desired product VIII was isolated
along with the debenzylated compound IX and the N-3
substituted regioisomer X. Presumably, debenzylation
of VIII generated a benzyl bromide, which subsequently
alkylated the N-3 position of IX. In a separate experi-
ment, benzylation of IX occurred regioselectively at
the N-3 position to provide product X.
of the analogous non-substituted N-1 benzyl com-
pounds. However, the p-hydroxybenzyl derivatives 31–
33 were quite interesting. While the N-1 p-hydroxybenz-
yl compound 31 did not have an affect on PDE inhibi-
tion when compared with analog 6, both the C-2
bromo (32) and phenylacetylenyl (33) compounds had
improved PDE5 potency and selectivity versus PDE1
and PDE6.
A comparison of C-2 and N-1 substituents is outlined in
Table 3. The N-1 benzyl compounds 23–27 generally did
not show any improvement in PDE5 inhibition or selec-
tivity compared to compound 6, but derivative 26
showed that a C-2 phenylacetylene moiety could provide
improved selectivity over PDE6 while maintaining
PDE5 potency. As shown by analogs 28–30, the p-meth-
oxybenzyl group did not improve upon the binding data
Although N-1 substituted compounds such as 32 and 33
had improved PDE6 selectivity over sildenafil, some of
the N-3 substituted compounds had improved in vitro
profiles, which more closely matched our project goals
(Table 4). As was the case in the N-1 series, substitution
with a p-hydroxybenzyl group generally produced an in-
crease in PDE5 inhibition. In addition, C-2 substituents
with electron withdrawing inductive effects provided
O
O
a
or
N
N
N
N
c
N
R
N
N
a, b
N
N
N
OMe
OMe
Bn
Bn
VI
VII
OH
O
N
O
N
O
N
N
N
N
N
N
N
R
R
+
+
R
N
N
H
N
N
N
N
OH
Bn
Bn
VIII
IX
X
Bn
Scheme 2. Synthesis of N-1/N-3/C-2 derivatives. Reagents and conditions: (a) NBS, NCS, or NIS; (b) phenylacetylene, CuI, PdCl2(PPh3)2; (c) BBr3.
Table 3. N-1/C-2 modificationsa
O
5
3N
N
R'
1
N
R
N
N
7
Compound
R
R0
PDE5 IC50 (nM)a
PDE1/5
>313
PDE6/5
5.3
6
Bn
Bn
Bn
Bn
Bn
32
23
24
25
H200
Cl
Br
—
5.0
60
145
—
—
7.3
4.5
26
27
Bn
Bn
33
74
>303
—
15.0
7.4
Ph
Ph
28
29
p-MeOBn
p-MeOBn
Bn
Br
38
48
—
—
7.5
7.8
30
p-MeOBn
20
—
4.5
Ph
Ph
31
32
p-HOBn
p-HOBn
Bn
Br
28
7.8
—
>1587
7.9
51.3
33
p-HOBn
6.5
>1538
23.0
a Ref. 8.

2368
C. D. Boyle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2365–2369
Table 4. N-3/C-2 modificationsa
600
500
O
N
R
5
3N
400
300
200
100
0
N
R'
1
N
N
7
0.25
0.5
1
2
3
4
Hr
Figure 2. Rat PK of compound 41. AUC = 351 nMÆh, Tmax = 0.25 h.13
Compound
R
R0
PDE5 PDE1/5 PDE6/5
IC50
(nM)a
34
35
Bn
Bn
Bn
I
23% I
32% I
—
—
—
—
600
500
400
300
200
100
0
36
Bn
40
—
3.6
Ph
37
38
p-MeOBn Bn
p-MeOBn Br
130
23
—
—
1.8
4.8
39
p-MeOBn
14
—
1.7
Ph
Ph
0.25
0.5
1
2
3
4
40
41
p-HOBn
p-HOBn
Bn
Br
29
2.5
—
>4000
7.4
72.0
Hr
Figure 3. Rat PK of sildenafil. AUC = 413 nMÆh, Tmax = 0.25 h.14
42
p-HOBn
5.3
—
19.8
43
44
45
p-HOBn
p-HOBn
p-HOBn
Cl
2.3
15
>4348
—
—
93.5
40.0
93.9
(Fig. 3),14 which features a rapid onset and fast clear-
ance. In addition, purine 41 had similar results in an oral
dog PK study, in which the AUC was 1156 nMÆh, and
the Tmax was 0.75 h.15 These results were sufficient to
warrant further testing of inhibitor 41 in a dog in vivo
model of efficacy.
CH3
I
3.3
a Ref. 8. % I = % inhibition @ 100 nM.
compounds with low single digit nanomolar PDE5 po-
tency. The greatest selectivity over PDE6 was imparted
by C-2 halogen substituents, as demonstrated by ana-
logs 41, 43, and 45, all of which had PDE6 selectivity
10-fold greater than sildenafil and excellent PDE5 po-
tency. Halogen derivatives 41 and 43 were also tested
for PDE2, PDE3, and PDE4 inhibition, and both com-
pounds had >1 lM potency for all three isozymes.8 Fur-
thermore, the bromo analog 41, sildenafil, vardenafil,
and tadalafil were tested against PDE isozymes 1–
11.8,12 Not only did purine 41 have good selectivity ver-
sus the PDE1 and PDE6 isozymes, but it had >200-fold
selectivity versus all other PDE isozymes (Table 5). This
represents an overall selectivity profile for purine 41 that
is superior to the three marketed PDE5 inhibitors. Spe-
cifically, our purine 41 had superior selectivity over
PDE1 and PDE6 compared with sildenafil and vardena-
fil, and the PDE11 selectivity of compound 41 was sig-
nificantly greater than that of tadalafil.
We utilized a standard in vivo model of erectile function
in which the pelvic nerve of a dog is electrically stimu-
lated in order to mimic activation by NO, facilitating
the nitric oxide/cGMP pathway of erectile function.16
Addition of a PDE5 inhibitor such as sildenafil increases
the intracavernosal pressure (ICP), which causes an
erection. Administration of purine 41 in this assay
(Fig. 4) produced a dose-dependent increase in ICP (re-
ported as % of mean blood pressure). These results are
similar to those observed with sildenafil in the anesthe-
tized dog model.16
In conclusion, modification of several regions of the
modestly potent PDE1/PDE5 inhibitors 1 and 2 resulted
in the discovery of a series of compounds, which exhibi-
ted single digit nanomolar PDE5 potency and isozyme
selectivity greater than the approved drugs sildenafil,
vardenafil, and tadalafil. This in vitro profile may result
in fewer side effects related to PDE isozyme inhibition.
The PDE5 inhibitor 41 showed oral bioavailability in
both rat and dog with a rapid onset and clearance
The PDE5 inhibitor 41 was subjected to PK studies in
rat and dog. As demonstrated by the AUC curve in Fig-
ure 2,13 the oral rat PK has a similar profile to sildenafil
Table 5. PDE isozyme selectivity comparisona
Compd
PDE1
PDE2
PDE3
PDE4
PDE5
PDE6
PDE7
PDE8
PDE9
PDE10
PDE11
41
>10
1.1
>10
>10
>10
>10
>10
>10
2.1
1.4
9.6
4.5
>10
0.0025
0.0035
0.0002
0.004
0.18
>10
>10
>10
>10
>10
>10
>10
>10
>10
3.4
2.2
3.1
0.54
2.3
Sildenafil
Vardenafil
Tadalafil
0.025
0.001
0.98
0.23
>10
0.46
>10
0.79
>10
0.13
0.033
>10
a Refs. 8 and 12. All numbers are IC50 values reported as micromolar.

C. D. Boyle et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2365–2369
2369
Neustadt, B.; Puchalski, C.; Pula, K.; Silverman, L.;
Smith, E.; Stamford, A.; Tedesco, R. P.; Tsai, H.;
Tulshian, D.; Vaccaro, H.; Watkins, R. W.; Weng, X.;
Witkowski, J. T.; Xia, Y.; Zhang, H. J. Med. Chem. 1997,
40, 2196.
100
80
60
40
20
0
8. See the following reference for the conditions of the
PDE1–PDE6 binding assays described in this paper:
Wang, Y.; Chackalamannil, S.; Hu, Z.; Boyle, C. D.;
Lankin, C. M.; Xia, Y.; Xu, R.; Asberom, T.; Pissarnitski,
D.; Stamford, A. W.; Greenlee, W. J.; Skell, J.; Kurowski,
S.; Vemulapalli, S.; Palamanda, J.; Chintala, M.; Wu, P.;
Myers, J.; Wang, P. Bioorg. Med. Chem. Lett. 2002, 12,
3149. These are assays involving human PDE, whereas the
data reported in Ref. 7 are from bovine PDE assays, which
explains why the IC50 values reported for compounds 1
and 2 are different in this paper than the published values
in Ref. 7.
Baseline
10 µg/kg, i.v. bolus
30 µg/kg, i.v. bolus
100 µg/kg, i.v. bolus
1
2
4
8
16
Stimulation Frequency (Hz)
9. See Ref. 7 for the synthesis of imidazole I.
Figure 4. Effects of compound 41 on intracavernosal pressure (ICP, as
a percent of mean blood pressure, MBP) following pelvic nerve
stimulation in an anesthetized dog model of penile erection.
10. The (R)-configuration at C-7 was preferred for PDE5
inhibition. For example the (S)-enantiomers of com-
pounds 6 and 10 had PDE5 inhibition of 11% and 12%
@ 100 nM, respectively.
similar to sildenafil. In addition, purine 41 demonstrated
oral efficacy in the anesthetized dog erectile function
model. Further optimization of this series of purine
PDE5 inhibitors will be disclosed in other
publications.17
11. Imidazole VI was synthesized with the same chemistry
used to produce imidazole I (Ref. 7). Triethylorthoformate
was used for the formation of VI in place of the
triethylorthoacetate employed for the synthesis of I.
12. The following are conditions used for the PDE7-PDE11
binding assays: for PDE7, PDE8, and PDE11 assays,
Amershamꢀs cAMP-PDE SPA Assay kit was used, with
the following additions: 0.1 lM cAMP for PDE7, 0.05 lM
cAMP for PDE8, and 0.0125 lM cAMP for PDE11. For
PDE9 and PDE10 assays, Amershamꢀs cGMP-PDE SPA
Assay kit was used with the following additions: 0.15 lM
cGMP and 100 lM MnCl2 (PDE9) and 0.7 lM cGMP
(PDE10).
Acknowledgements
The authors wish to thank Professor Ronald Breslow,
Dr. John Hunter, Dr. John Piwinski, and Dr. Catherine
Strader for helpful discussions. We also thank Dr. Pra-
dip Das for obtaining analytical data.
13. Rats were orally dosed (0.4% w/v methyl cellulose) with
compound 41 at 10 mg/kg. Plasma levels were sampled for
24 h (n = 2).
References and notes
14. Sildenafil PK in rat was measured using the same
conditions as Ref. 13.
15. Dogs were orally dosed (0.4% w/v methyl cellulose) with
compound 41 at 5 mg/kg. Plasma levels were sampled for
4 h (n = 3).
1. Haning, H.; Niewo¨hner, U.; Bischoff, E. Prog. Med.
Chem. 2003, 41, 249.
2. Terett, N. K.; Bell, A. S.; Brown, D.; Ellis, P. Bioorg. Med.
Chem. Lett. 1996, 6, 1819.
3. Keating, G. M.; Scott, L. J. Drugs 2003, 63, 2673.
4. Curran, M. P.; Keating, G. M. Drugs 2003, 63, 2203.
5. Ballard, S. A.; Gingell, C. J.; Tang, K.; Turner, L. A.;
Price, M. E.; Naylor, A. M. J. Urol. 1998, 159, 2164.
6. Goldstein, I.; Lue, T. F.; Padma-Nathan, H.; Rosen, R.
C.; Steers, W. D.; Wicker, P. A. New Engl. J. Med. 1998,
338, 1397.
7. Ahn, H.-S.; Bercovici, A.; Boykow, G.; Bronnenkant, A.;
Chackalamannil, S.; Chow, J.; Cleven, R.; Cook, J.;
Czarniecki, M.; Domalski, C.; Fawzi, A.; Green, M.;
16. Carter, A. J.; Ballard, S. A.; Naylor, A. M. J. Urol. 1998,
160, 242. An ICP versus stimulation frequency graph for
sildenafil at doses of 1–100 lg/kg is published in this
paper. The results are similar to those of compound 41
shown in Figure 4.
17. In vitro optimization of a similar series of compounds in
conjunction with this work has recently been disclosed:
Pissarnitski, D. A.; Asberom, T.; Boyle, C. D.; Chackal-
amannil, S.; Chintala, M.; Clader, J. W.; Greenlee, W. J.;
Hu, Y.; Kurowski, S.; Myers, J.; Palamanda, J.; Stamford,
A. W.; Vemulapalli, S.; Wang, Y.; Wang, P.; Wu, P.; Xu,
R. Bioorg. Med. Chem. Lett. 2004, 14, 1291.
Gundes, A.; Ho, G.; Laudicina, M.; Lindo, N.; Ma, K.;
¨
Manna, M.; McKittrick, B.; Mirzai, B.; Nechuta, T.;