Potent and selective xanthine-based inhibitors of phosphodiesterase 5

Article abstract of DOI:10.1016/j.bmcl.2006.11.019

Inhibitors of PDE5 are useful therapeutic agents for treatment of erectile dysfunction. A series of novel xanthine derivatives has been identified as potent inhibitors of PDE5, with good levels of selectivity against other PDE isoforms, including PDE6. St

Full text of DOI:10.1016/j.bmcl.2006.11.019

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2376–2379
Potent and selective xanthine-based inhibitors of
phosphodiesterase 5
Nichola J. Arnold,a Ruth Arnold,a David Beer,a Gurdip Bhalay,a
Stephen P. Collingwood,a Sarah Craig,a Nicholas Devereux,a Mark Dodds,a
Andrew R. Dunstan,a Robin A. Fairhurst,a David Farr,a Joseph D. Fullerton,a
Angela Glen,a Sylvie Gomez,a Sandra Haberthuer,a Julia D. I. Hatto,a Colin Howes,a
Darryl Jones,a Thomas H. Keller,a Beate Leuenberger,b Heinz E. Moser,a Irene Muller,b
Reto Naef,b Paul A. Nicklin,a David A. Sandham,a,* Katharine L. Turner,a
Morris F. Tweed,a Simon J. Watsona and Mauro Zurinib
aNovartis Institutes of Biomedical Research, Horsham Research Centre, Wimblehurst Road, Horsham, West Sussex RH12 5AB, UK
bNovartis Institutes of Biomedical Research, Lichtstrasse 35, CH-4056 Basel, Switzerland
Received 18 October 2006; revised 4 November 2006; accepted 6 November 2006
Available online 10 November 2006
Abstract—Inhibitors of PDE5 are useful therapeutic agents for treatment of erectile dysfunction. A series of novel xanthine deriv-
atives has been identified as potent inhibitors of PDE5, with good levels of selectivity against other PDE isoforms, including PDE6.
Studies in the dog indicate excellent oral bioavailability for compound 21.
Ó 2006 Elsevier Ltd. All rights reserved.
The cGMP metabolising enzyme phosphodiesterase 5
(PDE5) plays a pivotal role in regulating male erectile
function.1 Sexual stimulation results in release of NO
from the cavernosal nerve, activating guanylate cyclase,
which leads to an increase in intracellular cGMP levels.
Relaxation of smooth muscle is thus facilitated, leading
to increased blood influx into the corpus cavernosum
and hence penile erection. Inhibition of PDE5 is a clin-
ically proven concept for treatment of male erectile dys-
function (MED), with sildenafil 1 the first to market of
three currently approved PDE5 inhibitors.1,2
disclose the discovery of novel, xanthine-based PDE5
inhibitors which exhibit good levels of in vitro potency
and selectivity over other PDE isoforms, including
PDE6, as well as an acceptable in vivo pharmacokinetic
profile in the dog. 3-Isobutyl-1-methylxanthine (IBMX)
2a has been long known as a general, non-selective PDE
O
N
O
HN
N
N
Notwithstanding the utility of 1, adverse effects such as
headaches, flushing and visual disturbance have been
noted in a small number of MED patients taking silde-
nafil,3 some of which could be associated with cross
reactivity on other PDE isoforms. Competitive inhibi-
tion of PDE6, which controls function of rod and cone
cells within the eye, may be responsible for some of the
ocular side effects observed.4 In this communication, we
1
O2S
N
N
O
N
H
N
2a, R = H (IBMX)
2b, R =
N
R
N
O
Cl
2c, R =
Keywords: Phosphodiesterase 5; Xanthine; Erectile dysfunction.
*
Corresponding
david.sandham@novartis.com
author.
Fax:
+44
1403
323837; e-mail:
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.11.019

N. J. Arnold et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2376–2379
2377
O
N
tionalisation of isoquinolines lacking the 7-alkoxy
substituent, via formation of the 1,2-dihydroisoquino-
line, which was either isolated or reacted in situ.11 Meth-
od C employed a variant of the Bischler–Napieralski
procedure to assemble the 1-substituted isoquinoline
ring with the 4-acetate ester motif already installed.12
O
N
H
N
NH2
NH2
i, ii
O
N
N
O
N
O
O
O
R
N
OH
R=aryl,
heteroaryl
3
4
We established the initial structural requirements for
PDE5 activity13 (Table 1), with a small alkyl group at
N-1 of the xanthine being optimal (4 vs 7). Substitution
at N-7 in 5 resulted in a dramatic loss of activity, while
substitution of the methylene linker in 6 was also
detrimental.
Scheme 1. Parallel synthesis approach to xanthines. Reagents and
conditions: (i) acid, HATU or EDC; DIPEA, DMF, RT; (ii) NaOH,
MeOH, H2O, 60 °C.
inhibitor.5 Corbin6 described a series of 8-substituted
derivatives of IBMX such as 2b and 2c, which exhibited
low nM potency against PDE5 and modest selectivity
against PDE1. As no selectivity data against PDE6
had been reported, we were interested in re-evaluating
this class of compounds as a potential source of selective
PDE5 inhibitors.
We now probed the effect of variation of the xanthine
N-3 substituent on PDE5 activity, with selectivity profiling
on selected compounds against bovine retinal PDE6, hu-
man platelet PDE3 and bovine heart PDE1, as indicated
in Table 2. For comparison purposes, we also screened
the xanthine 2c reported by Corbin6 and were encour-
aged by the increased level of PDE6 selectivity exhibited
by 4. Within the series of N-3 alkyl substituents 8–10,
We have communicated a parallel synthetic method to
construct a library of C-8-substituted xanthines, react-
ing the diaminouracil 3 with a range of aryl- and hetero-
arylacetic acids (Scheme 1),7 which were either
commercially available or privileged structures obtained
from the Novartis Compound Archive. We were grati-
fied to identify the 4-isoquinolylmethyl derivative 48
with an IC50 of 9 nM for inhibition of human platelet
PDE5. We then set about a systematic investigation of
the structure–activity and selectivity relationships
around the xanthine–isoquinoline moiety in 4.
Table 1. PDE5 activity of 4 and derivatives 5-7
R2
O
R1
O
R3
O
N
N
N
O
N
N
Compound
R1
R2
R3
hPDE5
Diaminouracils analogous to 3 were either known com-
pounds or prepared using similar methodology.9 Iso-
quinoline-4-acetic acids were prepared by the three
methods outlined in Scheme 2. Method A utilised the
procedure of Bobbitt,10 for compounds with an activat-
ing 7-alkoxy group, method B was used for direct func-
4
5
6
7
Me
Me
Me
i-Bu
H
H
0.009
0.553
0.094
0.462
Me
H
H
Me
H
H
IC50 values are reported in lM.
Scheme 2. Synthesis of isoquinoline-4-acetic acids. Method A. Reagents and conditions: (i) toluene, reflux; (ii) NaBH4, EtOH, RT; (iii) glyoxylic
(R3 = H) or pyruvic (R3 = Me) acid, 6 N HCl, 100 °C. Method B reagents and conditions: (i) NaBH4, Ac2O–AcOH, 60 °C; (ii) glyoxylic acid, 6 N
HCl, 100 °C; (iii) NaEt3BH, THF–toluene, RT; (iv) ethyl glyoxalate, RT; (v) NaOH–H2O2, RT. Method C reagents and conditions: (i) CH3NO2,
tetramethylguanidine, 70 °C; (ii) SnCl2ÆH2O, EtOH, reflux; (iii) AcCl, Et3N, CH2Cl2, 0 °C; (iv) POCl3, MeCN, reflux; (v) Pd/C, decalin, reflux;
(v) LiOH, THF–H2O, RT.

2378
N. J. Arnold et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2376–2379
Table 2. PDE5 activity and isoform selectivity: effect of xanthine N-3
variation
N-3 isobutyl group. 1-Substitution (16) gave excellent
PDE5 potency and selectivity over PDE6, however,
there was a significant loss of selectivity over PDEs
1 and 3. Conversely, there was a 10-fold loss in poten-
cy with the 3-methyl derivative 17. Conformationally
restricted dioxole 18 delivered a dramatic loss of
potency, suggesting the importance of spatial orienta-
tion of at least one of the substituents, although
unsubstituted isoquinoline 19 did retain reasonable
potency. We ultimately established the 6-substituent
(20 vs 21) as being critical to providing optimal
potency and selectivity.
O
H
O
N
N
4, 8-15
N
O
O
N
R
N
Compound
R
hPDE5 bPDE6 bPDE1 hPDE3
2c
4
0.025
0.009
0.065
0.389
—
—
0.417
0.299
Caco-2 permeabilities14, together with calculated PSA15,
were then assessed for a selection of compounds (Table
4). Good permeability was observed with the mono-
and dimethoxy isoquinoline derivatives 4, 11 and 21,
although further substitution on the isoquinoline in 16
appeared deleterious. The absence of permeability for
12 and 15 appeared to correlate with their higher PSA
values.16
8
0.030
0.020
0.028
0.008
0.006
0.033
0.042
0.005
0.333
0.823
1.13
—
—
9
1.32
—
1.64
—
10
The hydrochloride salt of compound 2117 was selected
for further profiling. Against a panel of additional
PDE isoforms, human recombinant PDE4B, PDE4D
and PDE7A all gave IC50 values >10 lM, while human
11
0.294
0.363
1.82
2.79
0.670
3.91
—
1.06
2.71
—
12 (racemic)
Table 3. PDE5 activity and isoform selectivity: effect of isoquinoline
substitution
OH
OMe
OMe
13
14
15
O
H
N
N
16-21
N
R
0.471
0.399
—
O
N
NHAc
3.44
3.13
SO2
Compound
R
hPDE5 bPDE6 bPDE1 hPDE3
N
H
O
O
IC50 values are reported in lM.
16
0.002 0.226
0.072
0.042
N
there was a slight loss of PDE5 potency, although the
level of selectivity against PDE6 was similar to 4. An
increase in steric bulk in 11 furnished superior selectivity
over PDEs 1 and 3, while carbinol derivative 12 (pre-
pared from 8 by treatment with 9-BBN–H2O2) gave
slightly superior PDE5 potency, as well as significantly
enhanced selectivity against PDE3. The effect of larger
polar functionality was explored in 13–15 with the latter
compound being prepared by NaOH hydrolysis of the
amide in 14 followed by treatment with isopropylsulfo-
nyl chloride. Although the sulfonamide functionality
in 15 provided the highest levels of potency and selectiv-
ity hitherto observed, this substituent did confer rather
low solubility on the molecule compared with the sim-
pler N-3 alkyl derivatives.
O
O
17
18
19
20
21
0.026 0.525
—
—
N
O
O
10
—
—
—
N
0.034 0.857
0.051 0.247
0.002 0.180
7.29
—
1.85
—
N
N
O
O
The isoquinoline moiety in 4 offered a plethora of substi-
tution possibilities, some of which we next examined in
Table 3.
2.85
2.22
N
At this stage, for reasons of synthetic convenience and
greater solubility (vide supra), we chose to retain the
IC50 values are reported in lM

N. J. Arnold et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2376–2379
2379
Table 4. Caco-2 permeability and calculated PSA
selectivity shows a favourable in vivo pharmacokinetic
profile in the dog.
Compound
Caco-2 Papp/10À5cm sÀ1
Net flux A–B
2
˚
Calculated PSA/A
4
11
12
15
16
21
3.18
2.34
0
88.37
85.61
References and notes
109.37
130.49
87.73
1. (a) Stamford, A. W. In Annual Reports in Medicinal
Chemistry; Doherty, A. M., Ed.; Academic Press: New
York, 2002; p 53; (b) Doggrell, S. A. Exp. Opin.
Pharmacother. 2005, 6, 75.
0
0.09
1.38
79.35
2. (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.
3. Osterloh, I. H.; Riley, A. Br. J. Clin. Pharmacol. 2002, 53,
219.
Table 5. Microsomal clearance of 21
Rat clint
(lL/min/mg)
Human clint
(lL/min/mg)
Dog clint
(lL/min/mg)
162
4
6
4. Laties, A. M.; Zrenner, E. Prog. Retin. Eye Res. 2002, 21,
485.
5. Beavo, J. A.; Rogers, N. L.; Crofford, O. B.; Hardman, J.
G.; Sutherland, E. W.; Newman, E. V. Mol. Pharmacol.
1970, 6, 597.
Table 6. In vivo pharmacokinetics of 21
iv T1/2 (min) iv Cl (ml/min/kg) Vss (L/kg) % F
6. Sekhar, K. R.; Grondin, P.; Francos, S. H.; Corbin, J. D.
In Phosphodiesterase Inhibitors; Academic Press: New
York, 1996; p 137.
Rata 21.0 0.9
Dogb 63.4 12.0
118.4 16.9
28.7 9.2
2.9 0.4
2.1 0.3
ꢀ1c
103 20d
Data are means SEM.
a n = 5.
7. (a) Beer, D.; Bhalay, G.; Dunstan, A.; Glen, A.; Haber-
thuer, S.; Moser, H. Bioorg. Med. Chem. Lett. 2002, 12,
1973; (b) Arnold, R.; Beer, D.; Bhalay, G.; Baettig, U.;
Collingwood, S. P.; Craig, S.; Devereux, N.; Dunstan, A.;
Glen, A.; Gomez, S.; Haberthuer, S.; Howe, T.; Jelfs, S.;
Moser, H.; Naef, R.; Nicklin, P.; Sandham, D.; Stringer,
R.; Turner, K.; Watson, S.; Zurini, M. Bioorg. Med.
Chem. Lett. 2002, 12, 2587.
8. Selected data for 4. NMR (400 MHz, DMSO-d6 + TFA-
d1) dH 0.93 (3H d J 6), 2.30 (m 1H), 3.32 (3H s), 3.91 (2H d
J 8), 4.10 (3H s), 4.16 (3H s), 4.81 (2H s), 7.79 (1H s), 8.01
(1H s), 8.60 (1H s), 9.59 (1H s). HRMS: C22H25N5O4
requires 423.1907; found 423.1889.
9. Merlos, M.; Gomez, L.; Vericat, M. L.; Bartroli, J.; Garcia-
Rafanell, J.; Forn, J. Eur. J. Med. Chem. 1990, 25, 653.
10. (a) Bobbitt, J. M.; Winter, D. P.; Kiely, J. M. J. Org.
Chem. 1965, 30, 2459; (b) Dyke, S. F.; Sainsbury, M.;
Moon, B. J. Tetrahedron 1968, 24, 1467.
11. (a) Katayama, H.; Ohkoshi, M.; Yasue, M. Chem. Pharm.
Bull. 1980, 28, 2226; (b) Minter, D. E.; Re, M. A. J. Org.
Chem. 1988, 53, 2653.
12. Marivet, M. C.; Bourguignon, J.-J.; Lugnier, C.; Mann, A.;
Stoclet, J.-C.; Wermuth, C. G. J. Med. Chem. 1989, 32,
1450.
13. Inhibition of PDE5 enzyme activity was determined with a
[3H]cGMP scintillation proximity assay using yttrium
silicate beads (Amersham Pharmacia Biotech). Assays
for other PDE isoforms were conducted similarly. Data
generally represent the means of at least two experiments
and individual IC50 values were generally within 2-fold of
each other.
14. Wang, Z.; Hop, C. E. C. A.; Leung, K. H.; Pang, J.
J. Mass Spectrom. 2000, 35, 71.
15. Ertl, P.; Rohde, B.; Selzer, P. J. Med. Chem. 2000, 43, 3714.
16. Artursson, P.; Bergstroem, C. A. S. Methods Princ. Med.
Chem. 2003, 18, 341.
b n = 3.
c Dosed intraduodenally.
d Dosed orally.
platelet PDE2 gave an IC50 of 4.6 lM. In addition, 21
was profiled in the MDS Pharma Services Spectrum
ScreenTM consisting of 111 receptors, where the only
activity observed was at the adenosine A1 receptor (Ki
2.3 lM) and the adenosine transporter (Ki 2.8 lM).
In vitro microsomal clearance experiments showed a
much higher lability in rat compared with dog and hu-
man microsomes (Table 5).
In vivo, (Table 6) a short half-life and high clearance
were observed in the rat after 0.8 mg/kg iv dosing. When
dosed intraduodenally at 3.3 mg/kg, very low plasma
levels of 21 were observed, suggesting a high first-pass
metabolism. Contrastingly, excellent oral bioavailability
was exhibited in the dog after dosing at 2.5 mg/kg, along
with an improved half-life after 0.2 mg/mg iv dosing.
These properties were considered appropriate for fur-
ther evaluation in the MED indication, where a relative-
ly short duration of action in man was desirable.
In conclusion, we have shown that xanthines containing
an 8-(4-isoquinolylmethyl) substituent can function as
potent inhibitors of PDE5. Optimal activity is achieved
with a methyl group at N-1 of the xanthine. At the N-3
position, a variety of simple alkyl and polar substituents
are tolerated, although Caco-2 permeability is favoured
only in alkyl derivatives. It is also essential for N-7 to be
unsubstituted. On the isoquinoline ring, the 6-position
has been identified as critical to delivering maximal
potency and selectivity against other PDE isoforms.
Furthermore, compound 21 with optimal potency and
17. Selected data for 21. NMR (400 MHz, DMSO-d6) d 0.95
(3H d J 7), 2.30 (1H m), 3.34 (3H s), 3.91 (2H d J 8), 4.07
(3H s), 4.59 (3H s), 7.45 (1H d J 9), 7.66 (1H s), 8.19 (1H d
J 9), 8.51 (1H s), 9.22 (1H s). HRMS: C21H23N5O3
requires 393.1801; found 393.1801.