A solid-phase approach towards the synthesis of PDE5 inhibitors.

Article abstract of DOI:10.1016/S0960-894X(02)00296-2

PDE5 inhibitors based upon the xanthine scaffold 8 have been expediently synthesized using a solid-phase route. Attachment of the xanthine scaffold 8 using the Rink chloride linker 4 and N-1 functionalization using Mitsunobu chemistry is described. A library of compounds was produced in multi-milligram quantities with excellent purities and acceptable yields. The compounds were tested for their PDE5 inhibitory activity.

Full text of DOI:10.1016/S0960-894X(02)00296-2

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1973–1976
A Solid-Phase Approach Towards the Synthesis of PDE5
Inhibitors
David Beer, Gurdip Bhalay,* Andrew Dunstan, Angela Glen, Sandra Haberthuer
and Heinz Moser
Novartis Horsham Research Centre, Wimblehurst Road, Horsham, West Sussex RH12 5AB, UK
Received 28February 2002; revised 11 April 2002; accepted 26 April 2002
Abstract—PDE5 inhibitors based upon the xanthine scaffold 8 have been expediently synthesized using a solid-phase route.
Attachment of the xanthine scaffold 8 using the Rink chloride linker 4 and N-1 functionalization using Mitsunobu chemistry is
described. A library of compounds was produced in multi-milligram quantities with excellent purities and acceptable yields. The
compounds were tested for their PDE5 inhibitory activity. # 2002 Elsevier Science Ltd. All rights reserved.
Erectile dysfunction1 (ED) is defined as the inability to
achieve and maintain an erection to permit satisfactory
sexual intercourse. The major contributors to this con-
dition include: age, diabetes, heart disease, hormone
levels, hypertension, psychological influences and
physical injury such as damage to the spinal cord. ED
is a common and important problem; it has been esti-
mated that in the United States alone up to 30 million
men suffer from ED.2 Investigations to elucidate the
mecanism of ED have revealed the importance of
the enzyme phosphodiesterase 5 (PDE5). The cyclic
nucleotide phosphodiesterases (PDEs) are a super
family of enzymes that catalyze the hydrolysis of cyclic
nucleotides, cAMP and cGMP to their 50-nucleoside
monophosphates via cleavage of the phosphodiester
bond. PDE5 is the principal cGMP hydrolyzing enzyme
found in smooth muscle (corpus cavernosum) in the
penis. Upon sexual stimulation and under normal
physiological conditions the neurotransmitter nitric
oxide is released from non-adrenergic, non-cholinergic
neurons from nerve endings in the penis. This results in
the activation of soluble guanylyl cyclase, which con-
verts GTP to cGMP. Therefore the inhibition of PDE5
leads to the elevation of cGMP levels, which produces
greater relaxation of smooth muscle, producing an
increased blood flow in to the corpus cavernosum tissue
resulting in an erection.
success, for some patients there are some side-effects
associated with its use. Some of the problems have been
associated with: headaches, visual disturbances, nausea,
facial flushing and dyspepsia. These are thought to arise
from the inhibition of other PDEs. The search for more
selective4 PDE5 inhibitors is of interest to us and others.
We decided to start our program by investigating the
xanthine structural motif. From previous work, it was
known that 3-isobutyl-1-methylxanthine5 (IBMX) 2 is a
potent and non-selective inhibitor of PDE (Scheme 1).
We were interested in this class of compounds as a
potential source of selective PDE5 inhibitors (Scheme 2).
Scheme 1. Sildenafil 1 and isobutyl methylxanthine (IBMX) 2.
Sildenafil3 1 (ViagraTM) is an efficacious PDE5 inhibitor
used for the treatment of male ED and made an enor-
mously successful launch in 1998. Despite its great
Scheme 2. Preparation of the Rink chloride linker. Reagents: (i)
C2Cl6, PPh3, THF, rt, 6 h.
*Corresponding author. E-mail: gurdip.bhalay@pharma.novartis.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00296-2

1974
D. Beer et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1973–1976
Synthesis of the desired xanthine scaffold6 8 (Scheme 3)
started from the uracil 5. Reaction with sodium nitrite
in acetic acid and aqueous ethanol at 0 ꢀC for 1 h
produced the nitroso intermediate (75%), which was
subjected to a dithionite reduction using sodium
thiosulphite in water giving the diamino uracil 6 (97%).
The amide 7 was produced by direct condensation of
4-bromophenylacetic acid and the diamino uracil 6 by
heating the powdered mixture at 140 ꢀC to achieve
molten form and held there for 1 h (80%). Finally, the
construction of the xanthine heterocycle was completed
by heating 7 with a solution of sodium hydroxide and
methanol, giving the xanthine product 8 (75%) (Scheme
4). To initiate the program and to provide a rapid
understanding of the SAR of 8 we required a robust
method which would allow for the swift generation of
xanthine 8 analogues differing at R1 (Scheme 5). In
order to construct a series of compounds that differed at
N-1 (xanthine numbering, see 2) we first attempted to
attach xanthine 8 to commercially7 available 2-chloro-
trityl chloride resin 9 (Scheme 4) by treating with a
solution9 of 8 in DMF and Hunig’s base. Under these
conditions we failed to anchor 8 to the solid support.
Similar disappointment was borne using 10a and 10b. In
the case of 9 we reasoned that the steric challenge posed
by the 2-chlorotrityl chloride linker 9 was the primary
reason for this failure, so we turned our focus towards
the Rink chloride linker8 4 (Scheme 2). Transformation
of the Rink amide linker 3 to the chloro analogue 4
(Scheme 2) was achieved following a reported8 method
using a mixture of hexachloroethane and triphenylphos-
phine in THF (rt, 6 h). This was followed by the addi-
tion of a solution9 of xanthine 8. Evidence for the
attachment of the xanthine 8 was obtained using FT-IR
spectroscopy, which showed
a
C¼O stretch at
1687 cmꢁ1. To determine loading, the support bound
xanthine 11 was cleaved from the resin by exposing 11
to a solution of TFA in DCM (5% v/v) for 30 min at rt.
This gave the expected xanthine 13 (R1=H) as the tri-
fluoroacetate salt. The loading was calculated through
mass balance and typically ranged between 40 and 50%.
Scheme 3. Synthesis of xanthine 8. Reagents: (i) NaNO2, AcOH/
EtOH/H2O (1:3:2 v/v/v)), 0 ꢀC, 1 h; (ii) Na2S2O4, H2O, 85 ꢀC, 1 h; (iii)
4-bromophenylacetic acid, 140 ꢀC, 1 h; (iv) NaOH (2 M), MeOH/H2O
(7:1 v/v), 70 ꢀC, 3 h.
The loading is remarkable considering the highly dilute
conditions9 necessitated by the poor solubility of the
xanthine scaffold 8. The point of attachment of xan-
thine 8 with the solid support could be via N-7 or N-9
(xanthine numbering, see 2). Either way, it provides for
an anchored scaffold which upon cleavage from the
solid support releases the compound. The preparation
Scheme 4. Linkers that failed to tether the xanthine scaffold 8.
Scheme 5. Solid-phase route for the synthesis of 13. Reagents: (i) DMF, Hunig’s base, RT, 6 h; (ii) R1-OH, DBAD, PPh3, THF, rt, 1 h; (iii) TFA
5% in DCM, rt, 30 min.

D. Beer et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1973–1976
1975
Table 1. PDE 5% inhibition data for the R1 derivatives of 13 tested
at 1 mmol and also showing chemical yields and purities
Scheme 6. O-Alkylated ‘by-product’.
No
R1
PDE5 Inhibition (%)10 Yield (%)11 Purity (%)12
14
15
16
H
Me
42
93
93
47
45
47
97
98
97
1 h) and the appropriate alcohol14 giving 12. Resin 12
was then exposed to the cleavage solution [TFA (5% v/
v), 30 min, RT] releasing the required products in
useful yields and high purities (Table 1).
17
18
57
44
97
6846
98
Under the Mitsunobu conditions used for N-1 alkyl-
ation, we were prepared for the possibility of O-alkyl-
1
ation. Pleasingly we saw no evidence (using H NMR
and 13C NMR) for the production of the unwanted
O-alkylated product 32 (Scheme 6).
19
20
21
22
23
24
25
26
27
28
29
30
35
4
46
45
46
44
96
96
97
97
From the percentage PDE5 inhibition data (Table 1),
some interesting features are evident. The unsubtituted
xanthine 14 shows moderate activity and it is clear that
small alkyl groups are favoured, with methyl 15 and
ethyl 16 analogues displaying the same value (93%).
Short chain bulk, as in the cyclobutyl derivative 20 and
heteroaromatic derivatives 26, 27 have a deleterious
effect on inhibition. Derivatives 22–25 showed an
improved solubility profile in a range of solvents, but
only the alcohol derivative 22 retained good activity.
42
88
3846
41
59
22
20
33
15
13
96
In conclusion, we have broadened the utility of the Rink
chloride linker to allow for the loading of xanthine
heterocycles. And using Mitsunobu chemistry we have
been able to synthesize a range of analogues differing at
N-1. Moreover, using an appropriate scaffold this
methodology could be used to perform further changes
on the heterocycle. The preliminary data suggest that
xanthines containing a methyl, ethyl or hydroxymethyl
substituent (15, 16 and 22, respectively) at N-1 may
serve as potent inhibitors of PDE5. Further work will
be reported in due course.
45
45
46
47
46
46
45
98
98
97
97
97
98
96
References and Notes
1. (a) Rotella, D. P. Drugs Future 2001, 26, 153. (b) Maw, G.
N. In Annual Reports in Medicinal Chemistry; Bristol, J. A.,
Ed.; Academic: New York, 1999; p 71. (c) Bivalaqua, T. J.;
Champion, H. C.; Hellstrom, W. J. G.; Kadowitz, P. J. Trends
Pharmacol. Sci. 2000, 21, 484.
2. Feldman, H. A.; Goldstein, I.; Hatzichristou, D. G.; Krane,
R. J.; McKinlay, J. B. J. Urol. 1994, 151, 54.
3. (a) Terrett, N. K.; Bell, A. S.; Brown, D.; Ellis, P. Bioorg.
Med. Chem. Lett. 1996, 6, 1819. (b) Eardley, I. Br. J. Urol.
1998, 81, 122.
31
3844
97
4. (a) Osterloh, I. H.; Collins, M.; Wicker, P.; Wagner, G. Int.
J. Clin. Pract., Suppl. 1999, 102, 3. (b) Marmor, M. F.; Kess-
ler, R. Surv. Ophthalmol. 1999, 44, 153.
5. (a) Beavo, J. A.; Rogers, N. L.; Crofford, O. B.; Hardman,
J. G.; Sutherland, E. W.; Newman, E. V. Mol. Pharmacol.
1970, 6, 597. (b) Sekhar, K. R.; Grondin, P.; Francos, S. H.;
of the N-1 derivatives 14–31 (Table 1) was achieved by
reaction of the solid supported xanthine 11 (Scheme 5)
under the Mitsunobu13 conditions of triphenylphos-
phine, di-tert-butyl azodicarboxylate in THF (rt, for

1976
D. Beer et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1973–1976
Corbin, J. D. In Phosphodiesterase Inhibitors; Schudt, C.,
Dent, G., Rabe, K. F. Eds.; Academic: New York, 1996; p 137.
6. Merlos, M.; Gomez, L.; Vericat, M. L.; Bartroli, J.; Garcia-
Rafanell, J.; Forn, J. Eur. J. Med. Chem. 1990, 25, 653.
7. Novabiochem, e-mail uk@nova.ch
beads (Amersham Pharmacia Biotech). % inhibition values
quoted were obtained from testing at 1 mmol.
11. Yield calculated using mass balance.
12. Purity calculated using HPLC 254 nm.
13. Mitsunobu, O. Synthesis 1981, 1.
8. Garigipati, R. S. Tetrahedron Lett. 1997, 38, 6807.
9. The xanthine 8 was poorly soluble in DMF and required
the addition of Hunig’s base to aid dissolution.
10. Inhibition of PDE5 enzyme activity was determined with a
[3H]cGMP scintillation proximity assay using yttrium silicate
14. For 22 and 23, 2-tert-butoxyethanol and 2-(hydroxyethyl)
carbamic acid tert-butyl ester were used respectively in the
Mitsunobu reaction. With the initial product off resin, the
final compound was obtained by treatment with 95% TFA in
DCM for 1 h (rt).