The discovery of avanafil for the treatment of erectile dysfunction: A novel pyrimidine-5-carboxamide derivative as a potent and highly selective phosphodiesterase 5 inhibitor

Article abstract of DOI:10.1016/j.bmcl.2014.10.008

Novel pyrimidine-5-carboxamide derivatives bearing a 3-chloro-4-methoxybenzylamino group at the 4-position were identified as potent and highly selective phosphodiesterase 5 inhibitors. Among them, we successfully found 10j (avanafil) which exhibited a potent relaxant effect on isolated rabbit cavernosum (EC₃₀ = 2.1 nM) and a high isozyme selectivity.

Full text of DOI:10.1016/j.bmcl.2014.10.008

Bioorganic & Medicinal Chemistry Letters 24 (2014) 5460–5465
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The discovery of avanafil for the treatment of erectile dysfunction: A
novel pyrimidine-5-carboxamide derivative as a potent and highly
selective phosphodiesterase 5 inhibitor
Toshiaki Sakamoto a, Yuichi Koga a, Masataka Hikota a, Kenji Matsuki a, Michino Murakami b,
Kohei Kikkawa b, Kotomi Fujishige c, Jun Kotera c, Kenji Omori c,d, Hiroshi Morimoto a, , Koichiro Yamada a
⇑
a Medicinal Chemistry Research Laboratories II, Mitsubishi Tanabe Pharma Corporation, 2-2-50 Kawagishi, Toda, Saitama 335-8505, Japan
b Pharmacology Research Laboratories II, Mitsubishi Tanabe Pharma Corporation, 2-2-50 Kawagishi, Toda, Saitama 335-8505, Japan
c Advanced Medical Research Laboratories, Mitsubishi Tanabe Pharma Corporation, 2-2-50 Kawagishi, Toda, Saitama 335-8505, Japan
d Industry and Academia Cooperation Research Project, Laboratory of Target and Drug Discovery, Graduate School of Pharmaceutical Sciences, Nagoya University, Furo-cho,
Chikusa-ku, Nagoya 464-8601, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Novel pyrimidine-5-carboxamide derivatives bearing a 3-chloro-4-methoxybenzylamino group at the
4-position were identified as potent and highly selective phosphodiesterase 5 inhibitors. Among them,
we successfully found 10j (avanafil) which exhibited a potent relaxant effect on isolated rabbit cavernosum
(EC30 = 2.1 nM) and a high isozyme selectivity.
Received 4 September 2014
Revised 30 September 2014
Accepted 2 October 2014
Available online 28 October 2014
Ó 2014 Elsevier Ltd. All rights reserved.
Keywords:
PDE5
PDE6
Erectile dysfunction
Avanafil
Several orally available phosphodiesterase 5 (PDE5) inhibitors
have been developed and are prescribed as first-line therapy for
the treatment of erectile dysfunction (ED).1 Since PDEs consist of
a superfamily with 11 subfamilies (PDE1–11) and each of them
plays important roles in various tissues,2 it is necessary to develop
a PDE5 selective inhibitor in order to avoid adverse effects result-
ing from inhibition of the other PDEs. For example, the first
launched orally active PDE5 inhibitor, sildenafil showed high
efficacy, but visual side effects such as cyanopsia and visual distur-
bance were reported in some cases.3,4 These adverse effects are
indicated to be attributed to inhibition of PDE6, which is expressed
in photoreceptor cells of the retina and plays important roles for
the photo signal transmission.5
In a previous paper, we reported findings of 5-(3,4,5-trim-
ethoxybenzoyl)-4-amimopyrimidine derivatives as a novel chemi-
cal class of highly selective PDE5 inhibitors by scaffold hopping
from isoquinoline derivatives (Fig. 1).6 Although T-6932 (1) showed
a potent PDE5 inhibitory activity and an improved selectivity over
PDE6, its relaxant effect on isolated rabbit corpus cavernosum
was moderate (EC30 = 53 nM) owing to its high lipophilicity
(cLogP = 4.58). Therefore, the substituents at the 2- and 5-positions
of the pyrimidine ring were explored to potentiate the relaxant
effect while maintaining a high PDE5 selectivity against PDE6. In
this paper, we report the discovery of the novel, potent, and highly
selective PDE5 inhibitor avanafil 10j. The details of synthesis and
structure–activity relationships are described.
Aiming to improve its high lipophilicity of T-6932, pyrimidine-
5-carboxamide derivatives were designed. In the course of the
synthesis described in Scheme 1, the synthetic intermediate, ethyl
pyrimidine-5-carboxylate derivative 4, was found to show single
digit nanomolar activity for PDE5 (IC50 = 5.5 nM) and a slightly
improved relaxant effect (EC30 = 37 nM) compared to that of
T-6932.7 These results indicate that the 3,4,5-trimethoxybenzene
moiety, which caused the high lipophilicity of T-6932, was not cru-
cial for PDE5 inhibitory activity. Therefore, pyrimidine-5-carboxyl-
ate or carboxamide derivatives bearing more hydrophilic
substituents were designed and synthesized as in Scheme 1.
4-Aminopyrimidine derivative 3 was obtained by condensation
of commercially available 2 with 3-chloro-4-methoxybenzylamine.
After oxidation of the methylsulfanyl group with mCPBA, the
2-position of the pyrimidine ring was substituted with 2-hydroxy-
methylpyridine in the presence of sodium hydride to give 4. In this
reaction, 2-hydroxypyrimidine derivative 5 was obtained in 55%
⇑
Corresponding author. Tel.: +81 045 963 7164; fax: +81 045 963 7165.
E-mail address: morimoto.hiroshi@md.mt-pharma.co.jp (H. Morimoto).
http://dx.doi.org/10.1016/j.bmcl.2014.10.008
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

T. Sakamoto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5460–5465
OMe
Cl
5461
ring
opening
R
2
O
O
N
N
A
N
N
H
O
Intramolecular
hydrogen bond
(pseudo-ring)
N
CO2Me
OMe
5
MeO
MeO
OMe
OMe
OMe
isoquinoline
1; T-6932
EC30 = 53 nM
Selectivity
PDE6 / PDE5 = 2400
Figure 1. The structure and pharmacological profiles of lead compound T-6932 (1).
OMe
Cl
MeS
N
Cl
a
b
MeS
OMe
N
NH
N
N
CO2Et
CO2Et
OMe
Cl
2
3
OMe
Cl
O
N
NH
N
Cl
+
N
O
O
N
NH
CO2Et
HO
N
NH
CO2Et
+
N
O
N
N
N
4
5
6a
OMe
Cl
OMe
Cl
c
d
O
N
NH
4
N
O
N
NH
CO2H
N
N
O
N
X
R1
7
6b-j
Scheme 1. Reagents and conditions: (a) 3-chloro-4-methoxybenzylamine, Et3N, THF, rt (quant.); (b) (i) mCPBA, CHCl3, 0 °C, (ii) 2-hydroxymethylpyridine, NaH, THF, rt (4 33%,
5 55%, 6a 4%); (c) NaOH aq., EtOH, rt (81%); (d) EDC, HOBt, R1XH, DMF, rt (6b 49%, 6d 86%, 6e 78%, 6f 72%, 6g 65%, 6h 93%, 6i 91%, 6j 93%) or (i) SOCl2, CH2Cl2, rt, (ii) R1XH,
NaHCO3 aq., CH2Cl2, rt (6c 19%).
yield. In addition to this, 2-pyridylmethyl ester 6a was obtained by
concomitant ester exchange in 4% yield. Hydrolysis of the ester
group of 4, and subsequent condensation with various alcohols
or amines gave 6b–j.
The 2- and 5-positions of the pyrimidine ring were modified
according to the synthetic methods summarized in Scheme 2.
Hydrolysis of the ester group of 3, and the following condensation
with the corresponding amines afforded 9a–d, which were led to
10a–g, 10j, 10l, and 10m after oxidation of the methylsulfanyl
group. Alternatively, 10h, 10i, 10k, 10n, and 10o were synthesized
via the corresponding 2-amino-subustituted pyrimidine-5-carbox-
lates 12a–d, which were prepared from 3 or 2-hydroxypyrimidine
derivative 5.
by introduction of metabolically unstable functional groups into
the molecule. Faster onset of action should be accomplished by
improvement of solubility. Potentiation of the relaxant effect
should be accomplished by the reduction of the lipophilicity as
mentioned above. In order to satisfy these requirements, we
preferentially selected the substituents bearing heteroarylmethyl
moieties, amines, and primary or secondary hydroxyl groups.
PDE5 inhibitory activity of the synthesized compounds were
evaluated using the enzyme isolated from canine lung, and the
light-activated bovine retina PDE6 was used for evaluation of
PDE6 inhibitory activity in the process of compound screening.8
First, the influence of the substituents on the ester or amide
group for PDE5 inhibitory activity and selectivity was investigated.
As shown in Table 1, incorporation of hetero aromatic rings such as
pyridine and pyrimidine into the ester moiety of 4 maintained PDE5
inhibitory activity with reduced lipophilicity (6a; IC50 = 2.1 nM,
cLogP = 4.18, 6b; IC50 = 4.2 nM, cLogP = 3.22). The secondary
amide derivative 6c was equipotent to the corresponding ester
Our investigation toward the development of a novel and opti-
mal PDE5 inhibitor for the treatment of ED included incorporation
of the following characteristics: high selectivity for PDE5, appropri-
ately short duration of action to improve tolerability, and faster
onset of action. Short duration of action should be accomplished

5462
T. Sakamoto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5460–5465
OMe
OMe
Cl
OMe
Cl
Cl
c
R3
N
MeS
N
NH
N
NH
R2
a
b
MeS
N
NH
3
N
O
N
O
N
CO2H
HN
HN
R1
R1
10a-g, j, l, m
8
9a-d
9a, 10a-f (R1 = 2-pyridylmethyl)
9b, 10g, 10j (R1 = 2-pyrimidylmethyl)
9c, 10l (R1 = 3-hydroxypropyl)
9d, 10m (R1 = trans-4-hydroxycyclohexyl)
OMe
Cl
OMe
Cl
R3
N
Cl
N
NH
N
NH
R2
d
e
5
3
N
N
CO2Et
CO2Et
11
12a, b, d
OMe
Cl
OH
N
f
N
NH
N
CO2Et
12c
OMe
Cl
OMe
R3
N
R3
N
Cl
h
N
NH
R2
N
NH
R2
g
12a-d
N
O
N
CO2H
HN
R1
10h, i, k, n, o
13a-d
2
3
12a, 13a 10h
,
(R R N = 3-hydroxymethyl-piperidin-1-yl)
2
3
12b, 13b 10i
,
(R R N = 3-hydroxymethyl-pyrrolidin-1-yl)
2
3
12c, 13c, 10k
(R R N = (R)-2-hydroxymethyl-pyrrolidin-1-yl)
2
3
12d, 13d 10n, 10o
,
(R R N = (S)-2-hydroxymethyl-pyrrolidin-1-yl)
Scheme 2. Reagents and conditions: (a) NaOH aq., EtOH, rt (quant); (b) EDC, HOBt, R1NH2, DMF, rt (9a 94%, 9b 89%, 9c quant., 9d 88%); (c) (i) mCPBA, CH2Cl2 or CHCl3, 0 °C,
(ii) R2R3NH, CH2Cl2 or CHCl3 or THF or DMA, rt to 120 °C (10a 93%, 10b 72%, 10c 90%, 10d 66%, 10e 68%, 10f 91%, 10g 25%, 10j 52%, 10l quant., 10m 88%); (d) POCl3, PhNEt2,
100 °C (71%); (e) R2R3NH, DMA or THF, rt (12a 97%, 12b 91%, 12d quant.); (f) (i) mCPBA, CHCl3, 5 °C, (ii) (R)-prolinol, Et3N, THF, rt (12c 81%); (g) NaOH aq., DMSO or DMA, rt
(13a 96%, 13b 93%, 13c 89%, 13d 80%); (h) EDC, HOBt, R1NH2, DMF, rt (10h 69%, 10i 92%, 10k 74%, 10n 55%, 10o 58%).
derivative 6a; however, the N-methylated analogue 6e showed a
significant reduction of potency (6c, IC50 = 2.7 nM, 6e, IC50 = 85
nM). Besides the aromatic rings, incorporation of various functional
groups, such as ether, amine, and alcohol maintained PDE5 inhibi-
tory activity, but a series of 2-(2-pyridylmethyloxy)pyrimidine
derivatives 6a–j showed a lowered selectivity over PDE6 compared
to that of T-6932 (1, IC50 ratio: PDE6/PDE5 = 2400).
Next, the influence of the substituent at the 2-position of the
pyrimidine ring was investigated. In the previous paper, we
reported that the 5-(3,4,5-trimethoxybenzoyl)-pyrimidine deriva-
tives with a cyclic amine at the 2-position exhibited a high PDE5
selectivity over PDE6.6 Based on this, secondary amines bearing a
hydroxyl group were preferentially introduced at the 2-position
of the pyrimidine-5-carboxamide derivatives. As summarized in
Table 2, most of the synthesized compounds showed moderate
PDE5 inhibitory activities with IC50 values from 20 nM to 70 nM.
Among them, 10f and 10j bearing an (S)-2-hydroxymethyl-pyrroli-
din-1-yl group were found to show remarkably potent PDE5 inhib-
itory activity (10f, IC50 = 1.2 nM, 10j, IC50 = 5.2 nM). The
stereochemistry of the 2-hydroxymethyl group on the pyrrolidine
ring was important for PDE5 inhibitory activity, since 10j bearing
an (S)-2-hydroxymethyl-pyrrolidin-1-yl group was 7-fold more
potent than its (R)-enantiomer 10k (IC50 = 36 nM).
In addition, the (S)-2-hydroxymethyl-pyrrolidin-1-yl group
played a significant role for PDE5 selectivity. As summarized in
Table 3, all of the synthesized 5-position variants of 10f and 10j
showed excellent selectivity over PDE6, although these analogues
had various functional groups in the amide moiety. The selectivity
of 10f, 10j, 10m, and 10n was over 4000-fold, which was higher
than that of T-6932.
The relaxant effects on isolated rabbit corpus cavernosum of
these compounds were evaluated.7 Most of the tested compounds
showed improved relaxant effects compared to that of T-6932 as
summarized in Table 3. Among them, 10j (avanafil) showed the

T. Sakamoto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5460–5465
5463
Table 1
PDE5 inhibitory activity and selectivity over PDE6 of pyrimidine-5-carboxylate and carboxamide derivatives
OMe
Cl
O
N
NH
N
N
O
X
R1
6a-j
Compound
-XR1
IC50 (nM)a
2.1
Selectivity (PDE6/5)
1000
Compound
-XR1
IC50 (nM)a
15
Selectivity (PDE6/5)
N
N
O
N
H
OMe
O
6a
6f
240
630
O
6b
4.2
640
6g
N
5.6
N
H
N
N
N
N
N
H
N
H
OH
6c
2.7
12
480
720
6h
6i
13
12
420
480
OH
N
H
N
H
6d
OH
N
N
Me
6e
85
NT
6j
3.5
510
N
H
a
IC50 values of these compounds were calculated by an equation of the first degree using two data; just above and below 50% inhibition. Some of IC50 values were
calculated using PrismÒ, version 3.0.
Table 2
The influence of the substituent at the 2-position of the pyrimidine ring for PDE5 inhibitory activity
OMe
OMe
Cl
R3
N
Cl
R3
N
N
NH
N
NH
R2
R2
N
O
N
O
HN
HN
N
N
N
10a-f
10g-k
Compound
-NR2R3
IC50 (nM)a
32
Compound
-NR2R3
IC50 (nM)a
>100
Me
N
O
OH
10a
10g
N
HO
O
10b
10c
32
59
10h
10i
63
25
HO
HO
N
HO
HO
N
N
N
OH
OH
10d
10e
10f
54
10j
5.2
36
N
N
N
N
OH
OH
6.5
1.2
10k
N
OH
a
IC50 values of these compounds were calculated by an equation of the first degree using two data; just above and below 50% inhibition. Some of IC50 values were
calculated using PrismÒ, version 3.0.

5464
T. Sakamoto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5460–5465
Table 3
Selectivity and relaxant effect of pyrimidine-5-carboxamide derivatives bearing an (S)-2-hydroxymethyl-pyrrolidin-1-yl group at the 2-positon
OMe
OH
Cl
N
N
NH
N
O
HN
R1
10f, j, l, m, n, o
Compound
-R1
IC50 (nM)a
Selectivity (PDE6/5)
4300
EC30 (nM)b
22
10f
1.2
N
N
N
10j
5.2
6.6
1.8
4000
1500
2.1
OH
OH
10l
4.9
10m
>5600
>100
10n
10o
0.90
6.1
7700
2000
6.8
2.7
OMe
N
O
Me
a
IC50 values of these compounds were calculated by an equation of the first degree using two data; just above and below 50% inhibition. Some of IC50 values were
calculated using PrismÒ, version 3.0.
b
EC30 values were determined from the logarithmic concentration–inhibition curve. The value is given as the average of at least two experiments, where the variation from
the mean value is 30% or less.
most potent relaxant effect with an EC30 value of 2.1 nM and 4000-
fold selectivity over PDE6. These results were superior to those of
sildenafil (EC30 = 8.7 nM, PDE6/PDE5 = 500); therefore, 10j was
selected for further biological and pharmacological evaluation.
PDE5, PDE6, PDE2, PDE10, and PDE11 have two GAF domains in
their N-terminal half and constitute the GAF-PDE family.2 The
overall amino acid sequence similarity in the GAF family is high
compared with that of PDE family without GAF domain; therefore,
improvement of the selectivity against GAF-PDE family was
expected to be a critical issue in developing ideal PDE5-specific
inhibitors. Nevertheless, it is very interesting that avanafil showed
excellent and well-balanced selectivity against all other phospho-
diesterases (PDE1–11) in comparison to the other marketed PDE5
inhibitors.9 The selectivity ratio against trypsin-activated PDE6 of
avanafil was 121-fold, which was higher than those of sildenafil
(16-fold) and vardenafil (21-fold). Although tadalafil showed the
highest selectivity against PDE6 (550-fold) among the tested com-
pounds, the selectivity against PDE11 was moderate (25-fold). In
contrast, avanafil showed superior selectivity for PDE11 (>19231-fold)
to that of tadalafil, and also showed over 1000-fold selectivity ratio
against the other PDEs.
metabolism of its hydroxyl group, actual major metabolites of
avanafil in human plasma were found to be oxides of the pyrroli-
dine ring.12 Thus, the (S)-2-hydroxymethyl-pyrrolidin-1-yl group
at the 2-position of the pyrimidine ring plays an important role
for superior profiles of avanafil.
In conclusion, we successfully found avanafil (10j) as a novel
chemical class of potent and highly selective PDE5 inhibitor for
the treatment of ED. The modification to reduce the lipophilicity
of T-6932 led to the discovery of avanafil (cLogP = 2.36). Avanafil
showed improved relaxant effect on isolated rabbit corpus cav-
ernosum with an EC30 value of 2.1 nM, which was about 25 times
more potent than that of T-6932. The key structure of avanafil was
the (S)-2-hydroxymethyl-pyrrolidin-1-yl group on the pyrimidine
ring, which plays an important role for superior profiles, such as
high selectivity over PDE6 and short duration of action. Avanafil
showed a high selectivity against all other phosphodiesterases
(PDE1–11) and faster onset of action in comparison to the other
marketed PDE5 inhibitors. After clinical trials, avanafil was
approved by the Korea Food and Drug Administration, the Food
and Drug Administration (USA), and the European Medicine
Agency in 2011, 2012, and 2013, respectively.
In the evaluation of potentiating effect on the pelvic nerve stim-
ulation induced tumescence in anesthetized dogs, the times to
peak response for avanafil and sildenafil after intraduodenal
administration were 10 and 30 min, respectively.9 These results
indicate that avanafil has more rapid onset of action than sildenafil.
In addition, avanafil exhibited relatively short duration of action
correlated with the plasma drug concentration.
These profiles, that is, fast onset of action and shorter plasma
half-life than the other marketed PDE5 inhibitors were observed
in clinical trials.10,11 Although the (S)-2-hydroxymethyl-pyrroli-
din-1-yl group was introduced in the expectation of rapid
References and notes
1. Bruzziches, R.; Francomano, D.; Gareri, P.; Lenzi, A.; Aversa, A. Expert Opin.
Pharmacother. 2013, 14, 1333.
2. Omori, K.; Kotera, J. Circ. Res. 2007, 100, 309.
3. Marmor, M. F.; Kessler, R. Surv. Ophthalmol. 1999, 44, 153.
4. Giuliano, F.; Jackson, G.; Montorsi, F.; Martin-Morales, A.; Raillard, P. Int. J. Clin.
Pract. 2010, 64, 240.
5. Cote, R. H. Int. J. Impot. Res. 2004, 16, s28.
6. Sakamoto, T.; Koga, Y.; Hikota, M.; Matsuki, K.; Murakami, M.; Kikkawa, K.;
Fujishige, K.; Kotera, J.; Omori, K.; Morimoto, H.; Yamada, K. Bioorg. Med. Chem.
Lett. 2014, 24, 5175.

T. Sakamoto et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5460–5465
5465
7. The initial resting isomeric tension of each tissue strip was adjusted to 1.5 g by
gradual incremental stretching in 10 mL of organ bath chambers containing
physiological salt solution (PSS) at 37 0.5 °C, continuously aerated with 95%
[3H]cGMP plus unlabeled cGMP or in 50 mM Tris–HCl (pH8.0). Reaction
mixtures were incubated at 37 °C for 30 min and then boiled for 1.5 min.
Subsequently 100
added and incubated at 37 °C for 30 min. The reaction was stopped by adding
500
ll of 1 mg/ml solution of Crotalus atrox snake venom was
O
2 and 5% CO2. The contractile response to high-KCl (120 mM) PSS was checked
twice. Phenylephrine (PE) (5 Â 10À6 M) was added into each organ bath in
order to obtain a tonic contraction. After the PE contractile response was
stabilized, test compound (10À10–10À6 M) or vehicle was added to the
preparation at an interval of 30 min. Papaverine hydrochloride was added
into each organ bath chamber (final concentration, 10À4 M) to confirm the
maximal relaxation of the tissue strips at the end of experiment. The
composition of PSS was as follows (mM): NaCl 118, KCl 4.7, MgSO4 1.2, CaCl2
1.5, KH2PO4 1.2, NaHCO3 25.0, dextrose 11.0, EDTA-2Na 0.023 (pH 7.3 or 7.4).
8. PDE assay was done by the radiolabeled nucleotide method. PDE enzyme
ll methanol. Resultant solutions were applied to a Dowex (1 Â 8200–400)
column (volume 0.225 ml). Eluate radioactivity was measured with 5 ml
aqueous scintillation cocktails. DMSO solution without the compounds served
as a control. The final concentration of DMSO in the reaction mixture was 1%
volume per volume.
9. Kotera, J.; Mochida, H.; Inoue, H.; Noto, T.; Fujishige, K.; Sasaki, T.; Kobayashi,
T.; Kojima, K.; Yee, S.; Yamada, Y.; Kikkawa, K.; Omori, K. J. Urol. 2012, 188, 668.
10. Goldstein, I.; MoCullougt, A. R.; Jones, L. A.; Hellstrom, W. J.; Bowden, C. H.;
DiDonato, K.; Trask, B.; Day, W. W. J. Sex Med. 2012, 9, 1122.
11. Peterson, C.; Swearingen, D. J. Sex Med. 2006, 3(suppl 3), 253.
12. Kedia, G. T.; Uckert, S.; Assadi-Pour, F.; Kuczyk, M. A.; Albrecht, K. Ther. Adv.
Urol. 2013, 5, 35.
(100 ll) in 50 mM Tris–HCl (pH 8.0) was added to 200 ll assay buffer
composed of 50 mM Tris–HCl (pH 8.0), 12.5 mM MgCl2, 10 mM 2-
mercaptoethanol and 0.825 mg/ml bovine serum albumin. The enzyme
reaction was started by adding 200 ll substrate solution containing