1-(2-(2,2,2-Trifluoroethoxy)ethyl-1H-pyrazolo[4,3-d]pyrimidines as potent phosphodiesterase 5 (PDE5) inhibitors

Article abstract of DOI:10.1016/j.bmcl.2010.03.106

1H-Pyrazolo[4,3-d]pyrimidines were previously disclosed as a potent second generation class of phosphodiesterase 5 (PDE5) inhibitors. This work explores the advancement of more selective and potent PDE5 inhibitors resulting from the substitution of 2-(2,2,2-trifluoroethoxy)ethyl at the 1 position in the so-called alkoxy pocket.

Full text of DOI:10.1016/j.bmcl.2010.03.106

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3125–3128
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
1-(2-(2,2,2-Trifluoroethoxy)ethyl-1H-pyrazolo[4,3-d]pyrimidines as potent
phosphodiesterase 5 (PDE5) inhibitors
a
a
a
a
b
Michael B. Tollefson a, , Brad A. Acker , E. J. Jacobsen , Robert O. Hughes , John K. Walker , David N. A. Fox ,
*
Michael J. Palmer b, Sandra K. Freeman a, Ying Yu a, Brian R. Bond a
a Pfizer Global Research and Development, 700 Chesterfield Parkway, Chesterfield, MO 63017, United States
b Pfizer Global Research and Development, Sandwich, Kent, CT13 9NJ, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
1H-Pyrazolo[4,3-d]pyrimidines were previously disclosed as a potent second generation class of phos-
phodiesterase 5 (PDE5) inhibitors. This work explores the advancement of more selective and potent
PDE5 inhibitors resulting from the substitution of 2-(2,2,2-trifluoroethoxy)ethyl at the 1 position in
the so-called alkoxy pocket.
Received 10 February 2010
Revised 25 March 2010
Accepted 26 March 2010
Available online 3 April 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keyword:
Phosphodiesterase 5 inhibitors
Sildenafil (Fig. 1, 1), sold as ViagraÒ and RevatioÒ, is successfully
used in the treatment of erectile dysfunction (ED) and pulmonary
arterial hypertension (PAH). Sildenafil acts by competitive inhibi-
tion of phosphodiesterase 5 (PDE5).1 PDE5 regulates guanosine
cyclic 30,50-monophosphate (cGMP) levels by converting cGMP to
GMP. cGMP interacts with protein kinase G which leads to the
reduction of intracellular Ca+ levels and vasodilation.2 By inhibiting
PDE5, cGMP levels rise resulting in vasodilation of the vascular
endothelium allowing for the treatment of ED and PAH.
Sildenafil is a potent PDE5 inhibitor (IC50 = 3.5 nM) and selective
against other PDE isoforms. The most active isoform is PDE6
(IC50 = 30 nM) which is expressed in the retina and may be related
to visual effects which are sometimes reported. Sildenafil is quickly
absorbed and has a 4–6 h duration of effect.3 Efforts to improve upon
the selectivity of PDE5 inhibitors and extend their plasma half-life
have been previously reported.4 Compound 2 is an example of a
new class of PDE5 inhibitors with good potency (1.84 nM) and selec-
tivity (PDE6 IC50 = 946 nM) that served as a representative starting
point in this novel template.5 Our goal was to enhance the potency
of these pyrazolo[4,3-d]pyrimidines while enhancing the selectivity
to prevent possible side effects while extending the half-life to allow
for chronic dosing of a PDE5 inhibitor.
attachment of the alkoxy piece is done early in the synthesis.
A Mitsunobu reaction of the alkoxyethanol (3) piece is performed
with 4-nitro-1H-pyrazole-5-carboxamide (4) to afford the N-alkyl-
ated pyrazole (5). The nitro group in compound 5 is reduced to the
amine 6 via hydrogenation. Cyclization of compound 6 to the dione
7 was accomplished by reaction with carbonyldiimidazole. Finally
the common intermediate for all analogs was prepared by chlori-
nation of 7 using phosphorus oxychloride in the presence of a cat-
alytic amount of dimethylformamide (DMF) to produce the
required pyrazolopyrimidine dichloride 8.
The final analogs (10–29) can easily be prepared from dichlo-
ride 8 in a facile manner which allowed for the rapid development
of SAR for this series (Fig. 3). Typically the first amine added to
dichloride 8 was a deactivated amine, such as an aminopyridine.
An excess of amine could be added and heated under conventional
or microwave conditions to afford selective addition to the C-7
position leading to monoamine product 9. It was possible to add
O
N
N
N
O
HN
O
N
N
HN
N
N
N
Initially we decided to investigate the ethoxyethyl moiety at the
N-1 position of the pyrazolopyrimidine due to its potential suscep-
tibility to metabolic degradation due to its alkoxy substituent and
its potential to influence selectivity as indicated by previous SAR.5
The synthesis of these analogs is illustrated in Figures 2 and 3. The
N
O S
O
N
N
NH
1
2
* Corresponding author. Tel.: +1 636 459 5493.
E-mail address: michael.tollefson@yahoo.com (M.B. Tollefson).
Figure 1. Structure of first generation (sildenafil, 1) and second generation
(pyrazolopyrimidine, 2) PDE5 inhibitor.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.03.106

3126
M. B. Tollefson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3125–3128
R1
O
R1
O
(300 pM) with improved PDE6 and PDE11 selectivity indicating
the alkoxy pocket is shallower in PDE6 and PDE11. Not surprisingly
compound 11 had poorer metabolic stability (58%) due to its high-
er lipophilicity. We prepared a bulkier i-propoxy analog (12) which
lost a little potency (860 pM) compared to 10 but did not improve
the selectivity as much as 11. We prepared the trifluoroethoxy ana-
log (13) in an effort to block metabolism at the ethoxy chain. This
was successful as our HLM stability assay improved to 91% even
though the log P increased. Another interesting observation with
13 was that the trifluoroethyoxy group greatly enhanced the selec-
tivity against PDE6 and PDE11.
We pursued the trifluoroethoxyethyl analogs further due to their
excellent selectivity and improved metabolic profile. We started by
exploring the SAR of the C3 and C7 positions (Table 2). The high level
of selectivity observed with the trifluoroethyoxyethyl analog 13 was
observed in these analogs as well. We explored the pyridine substi-
tuent’s effect on selectivity by adding a small methyl substituent in
an effort to keep log P’s manageable. The 6-methylpyridyl analogs
(14 and 15) exhibited greater PDE6 selectivity while maintaining
PDE5 potency compared to 13. When the methyl group is moved
to the 4 position of the pyridine (16 and 17) some PDE5 activity is
lost while PDE11 selectivity is reduced indicating subtle changes
in the binding pockets of the PDE isoforms. The methyl group at
the C3 position (14 and 16) exhibits slightly better potency and
selectivity compared to the ethyl group (15 and 17).
O
R1
H
N
HO
O
H2
N
3
NH2
N
O
N
O
DBAD/Ph3P
Pd/C
N
N
NO2
NH2
NH2
4
NO2
5
NH2
6
CDI
R1
O
R1
O
Cl
N
POCl3
O
N
N
cat. DMF
N
NH
O
N
N
H
N
Cl
8
7
Figure 2. Preparation of intermediate dichloride 8.
R1
R1
R1
O
O
O
R2
N
R2
N
H2NR2
HNR3R4
HN
N
HN
Cl
N
A p-fluoroaniline analog (18) exhibited greatly improved PDE5
potency (89 pM) while increasing PDE6 and PDE11 selectivity
and being metabolically stable. The downside of this type of analog
is the increase in log P, the potential for hERG interactions and con-
cerns over aniline derived reactive metabolites. Pyrimidine 19
modulates the basicity of the pyridine nitrogen and an enhance-
ment in PDE5 potency (150 pM) while improving PDE11 selectivity
while having excellent metabolic stability.
N
N
N
N
N
N
N
R3
N
Cl
N
Cl
R5
R5
R5
R4
8
9
10-29
Figure 3. Preparation of final analogs.
the second and more nucleophilic amine (HNR3R4) in excess to the
above reaction mixture to produce the final product (10–29) with
selective addition at C-5 position and no adducts observed from
the first amine. Extraction and reverse phase HPLC afforded pure
products for testing.
As a comparator to the previous work we prepared compound
10 with an ethoxyethyl substituent (Table 1). Compound 10 exhib-
ited good PDE5 potency6 (400 pM) and reasonable metabolic sta-
bility with 78% of the compound intact after a 30 min incubation
with human liver microsomes (HLM). We sought to see the effect
of extending the alkoxy substituent. When replacing the ethoxy
with an n-propoxy substituent (11) we observed similar potency
We examined the SAR at the C-5 position while maintaining the
pyrimidine at the C-7 position and the trifluoroethoxyethyl group
at the N-1 position (Table 3). Expanding the piperazine ring to
the homopiperazine shows a sharp drop off in potency (20 and
21) indicating the importance of ring size and placement of the
basic nitrogen. Methylation of the homopiperazine nitrogen (22)
does little to affect the PDE5 potency of these molecules. Introduc-
tion of the dimethylated aminopiperidine affords a nice improve-
ment in potency indicating that this pocket has room to place
larger substituents to maintain potency and selectivity. The intro-
duction of a methyl group at the 3-position of the piperazine (24
and 25) led to an improvement of PDE5 potency by presumably
taking advantage of enhanced lipophilic contact to the binding
pocket. When R5 is ethyl (25), PDE11 selectivity is greatly
improved compared to piperazine 19 while PDE6 selectivity is
maintained. The enantiomer of 24, 26 is sixfold less potent than
24 and shows some decrease in PDE6 and PDE11 selectivity. An
example of an acyclic analog (27) shows weaker PDE5 potency
compared to the cyclic analogs. Examination of a couple non-basic
analogs led to modest potencies indicating importance of potential
H-bonding interactions with the basic amine. The morpholine (28)
and piperazinone (29) analogs exhibited good potency and
selectivity.
Table 1
SAR at the N-1 position. (R5 = Et)
R1
O
HN
N
N
N
N
N
N
NH
With a basic center and an aromatic ring in the molecule we
were concerned about potential interactions with the human
ether-a-go-go-related gene (hERG) which have the potential to
lead to QT prolongation related toxicities which warranted further
study.7 Compounds 14, 19, 24 and 25 offered a good cross section
of compounds with a combination of good potency, selectivity and
metabolic stability. We initially used a competition assay using
radiolabeled dofetilide to determine the compound’s likelihood
to interact with the hERG channel (Table 4). At a compound con-
Compd R1
PDE5 inhibition PDE6/PDE5 PDE11/PDE5 HLM
ratiob
ratiob
stabilityc
a
IC50 (nM)
10
11
12
13
Et
0.40
0.30
0.86
24ꢀ
180ꢀ
50ꢀ
44ꢀ
160ꢀ
52ꢀ
78%
58%
nd
n-Pr
i-Pr
CH2CF3 0.65
150ꢀ
450ꢀ
91%
a
b
c
PDE5, PDE6 and PDE11 assay protocols can be found in Ref. 6.
Ratio of IC50’s.
Human liver microsome stability, % compound remaining after 30 min.
centration of 10 lM each molecule inhibited at a value less than

M. B. Tollefson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3125–3128
3127
Table 2
SAR at the C-3 and C-7 position
F3C
O
R2
HN
N
N
N
R5
N
N
NH
R2
R5
PDE5 inhibition IC50 (nM)
PDE6/PDE5 ratiob
PDE11/PDE5 ratiob
HLMc (%)
a
Compd
14
Me
0.78
420ꢀ
440ꢀ
83
63
75
80
N
N
15
16
17
Et
1.01
1.84
3.28
310ꢀ
510ꢀ
220ꢀ
440ꢀ
280ꢀ
170ꢀ
N
N
Me
Et
F
18
19
Et
Et
0.09
0.15
470ꢀ
120ꢀ
2200ꢀ
2800ꢀ
100
96
N
N
a
b
c
PDE5, PDE6 and PDE11 assay protocols can be found in Ref. 6.
Ratio of IC50’s.
Human liver microsome stability, % compound remaining after 30 min.
Table 3
SAR of the C-5 position
F3C
O
N
HN
N
N
N
N
N
R5
N
R3
R4
NR3R4
R5
PDE5 inhibition IC50 (nM)
PDE6/PDE5 ratiob
PDE11/PDE5 ratiob
HLMc
a
Compd
N
N
N
N
20
Me
9.79
35ꢀ
204ꢀ
nd
nd
nd
NH
NH
N
21
22
Et
Et
3.15
6.48
69ꢀ
635ꢀ
287ꢀ
309ꢀ
23
Et
0.35
285ꢀ
5650ꢀ
31%
N
N
N
N
24
25
26
27
Me
Et
0.14
0.07
0.94
9.52
312ꢀ
125ꢀ
187ꢀ
>210ꢀ
14,700ꢀ
27,000ꢀ
2100ꢀ
76 mind
42 mind
68%
NH
NH
Me
Et
NH
N
H
>210ꢀ
nd
NH2
(continued on next page)

3128
M. B. Tollefson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3125–3128
Table 3 (continued)
NR3R4
R5
Et
PDE5 inhibition IC50 (nM)
3.92
PDE6/PDE5 ratiob
PDE11/PDE5 ratiob
HLMc
nd
a
Compd
N
28
107ꢀ
5100ꢀ
O
O
N
29
Et
2.96
>675ꢀ
>675ꢀ
34%
NH
a
PDE5, PDE6 and PDE11 assay protocols can be found in Ref. 6.
Ratio of IC50’s.
Human liver microsome stability, % compound remaining after 30 min, nd = not determined.
HLM half-life.
b
c
d
pressure after oral and IV administration of compound. For exam-
ple after IV dosing of 1.5 mg/kg of compound 25, we observed a
maximum decrease in mean arterial blood pressure (MAP) of
27 mmHg at 2 h post dose. The MAP remained at a reduction of
23 mmHg 6 h post dose. The measured free fraction of compound
25 at 6 h post dose indicates blood levels 72-fold over the PDE5
IC50 levels. When compound 25 is dosed at 5 mg/kg orally, the
MAP decreased by 27 mmHg at its maximum effect. The reduction
in MAP was sustained over 20 h post dose. At 20 h the free plasma
exposure of compound 25 is 25-fold over the measured IC50 levels
indicating its promise for once a day dosing.
We have presented a number of potent and selective PDE5
inhibitors that exhibit promising PK profiles consistent with pro-
jected once a day dosing in humans. More importantly these com-
pounds exhibit good oral efficacy as measured by the reduction of
MAP, which is sustained as predicted by the PK profile allowing for
further investigation of the chronic dosing of PDE5 inhibitors.
Table 4
Safety and in vivo pharmacology
Compd
Dofetilidea (%)
hERGb
IV Dog PKc
Cl
SHRd
t1/2
Vdss
14
40.0
10.7
24.7
18.8
560 nM
nd
nd
20.2
22
nd
nd
+
++
++
19e
24e
25f
5.1
900 nM
1.4
lM
6.7
4.9
5.1
8.0
9.4
7.8
lM
18.5
a
Percent inhibition of [3H]-dofetilide binding to the hERG protein stably
expressed on HEK-293 cells following a 10 lM dose of test compound.
b
hERG patch clamp electrophysiology assay, IC50
.
c
Compound dosed at 0.2f–0.5e mpk in 10 kg beagles. Halflife (t1/2) in h, clearance
(Cl) in mL/min/kg, volume of distribution (Vdss) in L/kg.
d
Compound dosed orally in spontaneously hypertensive rats (SHR) while mon-
itoring MAP, + = decrease of 10–15 mmHg, ++ decrease of >15 mmHg.
50%. As expected, the removal of the basic center leads to a loss of
dofetilide binding (% inh @ 10 lM: 28, 1%; 29, 3%). Generally the
References and notes
pyridyl analogs tended to have higher dofetilide binding, presum-
ably due to the pyridine’s higher log P and the pyridine being more
basic than the pyrimidine.
1. Salonia, A.; Rigatti, P.; Montorsi, F. Curr. Med. Res. Opin. 2003, 19, 241.
2. Kulkarni, S. K.; Patil, C. S. Methods Find. Exp. Clin. Pharmacol. 2004, 26, 789.
3. Guazzi, M.; Samaja, M. Curr. Med. Chem. 2007, 14, 2181.
To confirm the relevance of the dofetilide competition binding
assay we took select compounds into a hERG patch clamp assay
to determine the compound’s interaction with the hERG channel
(Table 4). We found that the rank order of activity observed with
the dofetilide assay was the same for the hERG assay. Once again
the pyridine 14 had the greatest hERG activity (560 nM). Note that
it appears that the hERG activity is underestimated by the dofeti-
lide assay for these molecules. Pyrimidine 19 showed 10.7% inhibi-
4. (a) Hughes, R. O.; Walker, J. K.; Cubbage, J. W.; Fobian, Y. M.; Rogier, D. J.;
Heasley, S. E.; Blevis-Bal, R. M.; Benson, A. G.; Owen, D. R.; Jacobsen, E. J.;
Freskos, J. N.; Molyneaux, J. M.; Brown, D. L.; Stallings, W. C.; Acker, B. A.;
Maddux, T. M.; Tollefson, M. B.; Williams, J. M.; Moon, J. B.; Mischke, B. V.;
Rumsey, J. M.; Zheng, Y.; MacInnes, A.; Bond, B. R.; Yu, Y. Bioorg. Med. Chem. Lett.
2009, 19, 4092; (b) Hughes, R. O.; Walker, J. K.; Rogier, D. J.; Heasley, S. E.; Blevis-
Bal, R. M.; Benson, A. G.; Jacobsen, E. J.; Cubbage, J. W.; Fobian, Y. M.; Owen, D.
R.; Freskos, J. N.; Molyneaux, J. M.; Brown, D. L.; Acker, B. A.; Maddux, T. M.;
Tollefson, M. B.; Moon, J. B.; Mischke, B. V.; Rumsey, J. M.; Zheng, Y.; MacInnes,
A.; Bond, B. R.; Yu, Y. Bioorg. Med. Chem. Lett. 2009, 19, 5209; (c) Owen, D. R.;
Walker, J. K.; Jacobsen, E. J.; Freskos, J. N.; Hughes, R. O.; Brown, D. L.; Bell, A. S.;
Brown, D. G.; Phillips, C.; Mischke, B. V.; Molyneaux, J. M.; Fobian, Y. M.; Heasley,
S. E.; Moon, J. B.; Stallings, W. C.; Rogier, D. J.; Fix, D. N. A.; Palmer, M. J.; Ringer,
T.; Rodriquez-Lens, M.; Cubbage, J. W.; Blevis-Bal, R. M.; Benson, A. G.; Acker, B.
A.; Maddux, T. M.; Tollefson, M. B.; Bond, B. R.; MacInnes, A.; Yu, Y. Bioorg. Med.
Chem. Lett. 2009, 19, 4088.
5. Palmer, M. J.; Bell, A. S.; Fox, D. N. A.; Brown, D. G. Curr. Top. Med. Chem. 2007, 7,
405.
6. (a) Tollefson, M.B. WO07054778.; (b) Brown, D. L.; Owen, D. R.; Phillips, C.;
Palmer, M. J.; Bell, A. S.; Freskos, J. N.; Fobian, Y. M.; Walker. J. K.; Hughes, R.;
Jacobsen, E. J.; Tollefson, M. B.; Brown, D. G.; Mischke, B. V.; Molyneaux, J. M.
WO07020521.
7. (a) Aronov, A. M. Curr. Opin. Drug Disc. Dev. 2008, 11, 128; (b) Shamovsky, I.;
Connolly, S.; David, L.; Ivanova, S.; Norden, B.; Springthorpe, B.; Urbahns, K. J.
Med. Chem. 2008, 51, 1162.
8. The Pfizer Institutional Animal Care and Use Committee reviewed and approved
the animal use in these studies. The animal care and use program is fully
accredited by the Association for Assessment and Accreditation of Laboratory
Animal Care, International.
tion at 10 lM in the dofetilide assay where it had an IC50 = 5.1 lM
in hERG. Therefore the hERG safety margin was less than originally
predicted by the dofetilide assay. The ratio of PDE5 IC50 to hERG
IC50 for compound 19 is over 34,000ꢀ.
We examined the pharmacokinetic (PK) properties of these
molecules using dogs (Table 4).8 Compounds in this series gener-
ally exhibit excellent solubility, presumably due to the basic nitro-
gen. Pyrimidines 19, 24 and 25 have similar profiles with moderate
clearance (ꢁ20 mL/min/kg) and volume (8–9 L/kg) leading to ter-
minal half-lifes of 5–7 h, consistent with once a day dosing in
human.
These compounds were taken into an in vivo model of efficacy
(Table 4). We used spontaneously hypertensive rats (SHR) which
were monitored for compound levels and blood pressure. Encour-
agingly, pyrimidines 19, 24 and 25 all exhibited a lowering of blood