Design, synthesis and biological activity of β-carboline-based type-5 phosphodiesterase inhibitors

Article abstract of DOI:10.1016/S0960-894X(03)00159-8

The SAR of a series of β-carboline derived type 5 phosphodiesterase inhibitors has been explored and we have discovered compounds with excellent levels of PDE5 potency and selectivity over PDE6. However, the series exhibits low levels of selectivity over PDE11, a phosphodiesterase with unknown function.

Full text of DOI:10.1016/S0960-894X(03)00159-8

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1425–1428
Design, Synthesis and Biological Activity of ꢀ-Carboline-Based
Type-5 Phosphodiesterase Inhibitors
Graham N. Maw,* Charlotte M. N. Allerton, Eugene Gbekor and William A. Million
Department of Discovery Chemistry, Pfizer Global Research and Development, Sandwich Laboratories (IPC351),
Ramsgate Road, Sandwich, Kent, CT139NJ, UK
Received 28 November 2002; revised 5 February 2003; accepted 10 February 2003
Abstract—The SAR of a series of b-carboline derived type 5 phosphodiesterase inhibitors has been explored and we have dis-
covered compounds with excellent levels of PDE5 potency and selectivity over PDE6. However, the series exhibits low levels of
selectivity over PDE11, a phosphodiesterase with unknown function.
# 2003 Elsevier Science Ltd. All rights reserved.
Phosphodiesterases (PDEs) play critical roles in the
modulation of the important secondary messenger cyclic
nucleotides, adenosine cyclic 30,50-phosphate (cAMP)
and cyclic guanasine 50-phosphate (cGMP).1,2 The cGMP
specific PDE5 enzyme controls the levels of intracellular
cGMP and plays a critical role in its modulation, duration
of action and physiological function. PDE5 is the major
PDE isozyme in penile corpus cavernosum tissue and
plays a key role in the control of penile erection. Sexually
stimulated penile erection is a result of trabecular smooth
muscle relaxation.3À5 During sexual stimulation nitric
oxide (NO) is released from the non-adrenergic, non
cholinergic nerves within the penis supplying the corpus
cavernosum smooth muscle and activates guanylate
cyclase to form cGMP via cyclisation of GTP. The NO
diffuses into penile smooth muscle cells and binds to the
heme of soluble guanylate cyclase, resulting in stimulation
of cGMP synthesis, which, in turn, reduces intracellular
calcium following binding to cGMP-dependant protein
kinases and cGMP-dependant ion channels. Inhibitors of
PDE5 amplify the neuronal NO/cGMP pathway involved
in human corpus cavernosal relaxation and therefore
enhance the normal process leading to penile erection.6À8
5 (PDE5),10 and therefore slows cGMP breakdown.
This enhances the action of nitric oxide and cGMP, and
facilitates penile erection in individuals suffering from
erectile dysfunction (ED).11
As part of our continued efforts in this area we have
investigated several templates other than the pyrrazolo-
pyrimidine present in sildenafil, 1. We have previously
reported that sildenafil, 1, tadalafil, 2, and vardenafil, 3,
(Fig. 2) exhibit different PDE selectivity profiles, in par-
ticular in their relative affinities for PDE5 and PDE11.12
Sildenafil, 1, and vardenafil, 3, show excellent PDE5
selectivity over PDE11 (760- and 1160-fold, respectively)
whereas tadalafil exhibits only a modest 5-fold selectivity.
It has been suggested that PDE11 protein localization in
man and mouse, coupled with the known role for
cGMP and cAMP in sperm physiology and testosterone
production, suggests a local role for PDE11 in testis
development and/or spermatogenesis.13 The presence in
acidophils also suggests a role in modulating secretion
of prolactin and/or growth hormone, hence an indirect/
endocrine effect on lactation, luteal function, steroido-
genesis, immunoregulation and growth, for example, as
well as on the testis and spermatogenesis. The physiological
significance of PDE11 inhibition is not clearly understood
but could have a role in sperm function.14 Thus, in the
future PDE11 inhibitors might be explored for novel
indications relating to hormonal or other functions, but
when developing PDE5 inhibitors to modulate vascular
function it was thought desirable to maintain high
selectivity of PDE5 effects over PDE11.
The launch of ViagraTM (sildenafil, 1) (Fig. 1) as the
first oral treatment for male erectile dysfunction revo-
lutionised the treatment of this disease.9 Sildenafil is a
potent and selective inhibitor of phosphodiesterase type
*Corresponding author. Tel.: +44-1304-648454; fax: +44-1304-
651817; e-mail: graham_maw@sandwich.pfizer.com
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00159-8

1426
G. N. Maw et al. / Bioorg. Med. Chem. Lett. 13(2003) 1425–1428
Our initial strategy was to introduce a basic nitrogen
into the molecule as a means of improving aqueous
solubility and to assess its impact on PDE5 potency and
selectivity. To this end we found that it is possible to
react compound 6 with a range of primary diamines
(RNH2) to afford compounds 7–12, Scheme 1, Table 1.
Table 1. Initial SAR results for compounds 7–12 investigating the
impact of introducing basic nitrogena
Figure 1. Structure of sildenafil.
Compd
R
IC50 (nM)
PDE6
PDE5
6.7
PDE11
2
7
CH3
126037
8224
Figure 2. Structure of tadalafil and vardenafil.
134
53
8
56
2966
NT
The aims of these SAR studies were to investigate whe-
ther PDE5 selectivity, especially over PDE11 could be
built into this b-carboline template. In this paper we
wish to report our efforts in this regard.
9
26
1619
NT
NT
950
85
The N-methyl group in compound 2 was targeted for
structural modification on the basis of synthetic acces-
sibility and thus the ability to prepare a number of
analogues for in vitro evaluation and rapid SAR gen-
eration. d-Tryptophan methyl ester hydrochloride 4
was treated with piperonal under Dean–Stark condi-
tions followed by treatment with trifluoroacetic acid to
give the carboline methyl carboxylate 5 as shown in
Scheme 1. The key intermediate for the remainder of the
synthesis 6 was prepared via the acylation of 5 with
chloroacetyl chloride.
10
11
164
17
NT
12
2.1
249
5.5
NT, Not tested.
aPhosphodiesterase type 5 (PDE5) was obtained from human corpus
cavernosum tissue, PDE6 from canine or bovine retina and PDE11
from recombinant SF9 expression as previously described.11,15,16
Scheme 1. Reagents and conditions: (i) piperonal, toluene, Dean–
Stark followed by trifluoroacetic acid, (ii) chloroacetyl chloride, tri-
ethylamine, (iii) RNH2, ethanol, 50 ꢀC.
Scheme 2. Reagents and conditions: (i) Pd/C, ammonium formate,
ethanol, (ii) RCHO, sodium triacetoxyborohydride, acetic acid, THF.

G. N. Maw et al. / Bioorg. Med. Chem. Lett. 13(2003) 1425–1428
1427
Table 2. Optimisation of the pyrrolidine N-substiuent, SAR for
compounds 12–30
Table 3. Investigation of benzdioxolane replacements and the effect
on PDE11 activity
Compd
Ar
—
IC50 (nM)
PDE11
3.5
PDE5
2.6
Selectivity
1.4
Compd
R
IC50 (nM)
27
31
PDE5 PDE6 6/5sel PDE11 11/5 sel
12
2.1
249
118
NT
2.7
6.7
6.5
2.5
0.5
15
16
57
>104 >175 NT
32
33
12.8
13.2
1770134
2660169
31
43
2.3
2.7
17
18
19
15.7
4.0NT
7.1
15.6
11.8
0.8
11.8
22
3
1340188
226
3
34
35
28.6
127
44
39
1.5
0.3
20
21
22
23
1.5
153
2.01
2.4
1.4
1.3
1.9
271
181
125
144
36
37
132
71
0.5
—
2.0358
0.7
4.2
2.1
>1000
ꢁ1000
173
640
1.4
2.2
2
ation pyrrolidine 12 showed an 8-fold increase in
enzyme inhibition potency.
24
0.8
494
2.6
Compound 12 was selected for further SAR investiga-
tion through optimisation of the N-substituent on the
pyrrolidine ring. The compound was smoothly deben-
zylated under transfer hydrogenation conditions over
palladium-on-carbon to give 13. This compound was
then converted into a range of N-alkyl derivatives
through straightforward reductive alkylation in the pre-
sence of a range of aldehydes and sodium triacetoxy-
borohydride, Scheme 2.
25
26
27
28
29
1.2
357
NT
807
1601
211
300
3.5
48
2.9
2.4
1.9
5.7
19.8
2.5
316
189
4.7
48
8.4
From Table 2 it can be seen that replacement of an sp2
hybridised carbon with an sp3 carbon in the position
beta to the pyrrolidine nitrogen results in the loss of 6–
20-fold in activity relative to compound 12 (e.g., com-
pounds 15 and 16). Activity is retained with the N-allyl
derivative 18 and the SAR remains quite flat with a range
of polar phenyl-group isosteres (i.e., compounds 20–25).
4.0846
4.7
1.2
1.2
30
3.6
620172
4.3
NT, not tested.
Excellent PDE5 selectivity over PDE6 is maintained
throughout the series (150–650-fold). Indeed, com-
pound 24 combines sub-nanomolar PDE5 inhibition
activity with >600-fold selectivity over PDE6.
Introduction of flexible linkers such as found in 7 and 8
afforded good levels of PDE5 inhibition. However,
providing some degree of conformational rigidity into
the tether group led to an increase of PDE5 inhibition in
compounds 9, 11 and 12. Interestingly, the R-configur-
Acylation of the pyrrolidine nitrogen as the N,N-dime-
thylcarbamoyl derivative 26 leads to a loss in PDE5

1428
G. N. Maw et al. / Bioorg. Med. Chem. Lett. 13(2003) 1425–1428
activity. Isomerisation of this substituent leads to 27
and the restoration of excellent PDE5 inhibition activ-
ity. Simple exploration of the amide SAR (compounds
28–30) again shows a reasonably flat PDE5 SAR.
However, these compounds retain good selectivity over
PDE6 inhibition.
We also acknowledge technical assistance provided by
the Structural and Separation Sciences group.
References and Notes
1. Beavo, J. A. Physiol. Rev. 1995, 75, 725.
2. Beavo, J. A.; Conti, M.; Heaslip, R. J. Mol. Pharmacol.
1994, 46, 399.
3. Haher, A.; Mayer, M.; Steif, C. G.; Jonas, U.; Forsmann,
W. G. World J. Urol. 1996, 15, 32.
4. Lerner, S.; Melman, A.; Christ, G. J. Urol. 1993, 149, 1246.
5. Burnett, A. L. J. Urol 1997, 157, 320.
6. Maw, G. N. Ann. Rep. Med. Chem. 1999, 34, 71. Stanford,
A. W. Ann. Rep. Med. Chem. 2002, 37, 53.
However, Table 2 also clearly shows that the series is
not selective for PDE5 over PDE11 with the compounds
tested showing up to a very modest 3-fold selectivity. We
were interested to find out whether this was a generic
limitation of the template and prepared a range of ana-
logues containing benzdioxolane replacements, Table 3.
These compound were prepared using chemistry pre-
viously described, vide supra.
7. Cartledge, J.; Eardley, I. Curr. Opin. In CPNS Invest. Drugs
1999, 2, 240.
8. Eardley, I. Exp. Opin. Invest. Drugs 1997, 12, 1803.
9. Eardley, I. Br. J. Urol. 1998, 81, 122.
10. Terrett, N. K.; Bell, A. S.; Brown, D.; Ellis, P. Bioorg.
Med. Chem. Lett. 1996, 6, 1819.
11. Ballard, S. A.; Gingell, C. J.; Tang, K.; Turner, L. A.;
Price, M. E.; Naylor, A. M. J. Urol. 1998, 159, 2164.
12. Gbekor, E.; Bethell, S.; Fawcett, L.; Mount, N.; Phillips,
S. C. J. Urol. 2002, 167, 967.
These analogues retained respectable levels of PDE5
inhibition and it is interesting to note that the
PDE5 enzyme tolerates both electron withdrawing and
electron donating substituents in the para-position of
the aryl group, Table 3. However, this series of com-
pounds was also found to be non selective over PDE11.
13. Baxendale, R.; Burslem, F.; Phillips, S. J. Urol. 2001, 165,
340.
14. Baxendale, R.; Wayman, C.; Turner, L.; Phillips, S. C. J.
Urol. 2001, 165, 223.
Summary
15. Fawcett, L.; Baxendale, R.; Stacey, P.; McGrouther, C.;
Harrow, I.; Soderling, S.; Hetman, J.; Beavo, J. A.; Phillips,
S. C. PNAS 2000, 97, 3702.
Extensive SAR exploration of the b-carboline template
has identified a range of compounds exhibiting excellent
PDE5 potency and selectivity over PDE6. However,
these compounds are not selective for PDE5 over
PDE11. In addition, although the compounds showed
improved aqueous solubility compared to 2, they did
not exhibit other biopharmaceutic improvements. In
light of these results, and the as yet undefined role of the
PDE11 enzyme especially in relation to testis development
and/or spermatogenesis, we focused our efforts on
exploration of new templates. The results of these studies
will be the subject of further communications.
16. PDE activity was measured using a Scintillation Proximity
Assay (SPA)-based method as previously described.15 The
effect of PDE inhibitors was investigated by assaying a fixed
amount of enzyme in the presence of varying inhibitor con-
centrations and low substrate, (cGMP in a 3:1 ratio unla-
belled to [3H]-labeled at
a
concnꢁ1/3 Km) such that
IC50 ffi Ki. The final assay volume was made up to 100 mL
with assay buffer [20mM Tris–HCl pH 7.4, 5 mM MgCl , 1
2
mg/mL bovine serum albumin]. Reactions were initiated with
enzyme, incubated for 30–60 min at 30 ꢀC and terminated
with 50 mL yttrium silicate SPA beads (Amersham). Plates
were shaken, settled and then counted on a TopCount plate
reader (Packard, Meriden, CT) Radioactivity units were
converted to% activity of an uninhibited control (100%),
plotted against inhibitor concentration and inhibitor IC50
values obtained using the ‘Fit Curve’ Microsoft Excel
extension.
Acknowledgements
We would like to thank Dr. N. Mount for useful dis-
cussions and Ms. E. Evrard for technical contributions.