The discovery and synthesis of highly potent, A(2a) receptor agonists

Article abstract of DOI:10.1016/S0960-894X(00)00017-2

A series of N6,2-disubstituted adenosine analogues have been synthesized and their functional activity measured against A(2a) and A₁ receptors. Examples of compounds with both a lipophilic N6-substituent and amino-functionalized 2-position were highly active at the A(2a) receptor on the human neutrophil. (C) 2000 Published by Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00017-2

Bioorganic & Medicinal Chemistry Letters 10 (2000) 403±406
The Discovery and Synthesis of Highly Potent, A2a Receptor
Agonists
Suzanne E. Keeling, a,* F. David Albinson, b Barry E. Ayres, a Peter R. Butchers, a
C. Lynn Chambers, a Peter C. Cherry, a Frank Ellis, a George B. Ewan, a
Michael Gregson, a John Knight, a Keith Mills, b Paul Ravenscroft, b
Linda H. Reynolds, a Shahin Sanjar c and Michael J. Sheehan a
aMedicinal Sciences, Glaxo Wellcome Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts, SG1 2NY, UK
bChemical Development, Glaxo Wellcome Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts SG1 2NY, UK
cDisease Sciences, Glaxo Wellcome Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts SG1 2NY, UK
Received 16 November 1999; accepted 5 January 2000
AbstractÐA series of N6,2-disubstituted adenosine analogues have been synthesized and their functional activity measured against
A2a and A1 receptors. Examples of compounds with both a lipophilic N6-substituent and amino-functionalized 2-position were
highly active at the A2a receptor on the human neutrophil. # 2000 Published by Elsevier Science Ltd. All rights reserved.
Introduction
We noted that substitution at the 6-position of adenine
ribosides generally conferred A1 agonist activity
whereas 2-substitution improved activity at the A2a
receptor, although there were many subtleties. Two
pairs of compounds illustrate the marked in¯uence of
the position of substitution on activity: moving the
cyclopentyl substituent from N6 in N-cyclopentyl-2-
chloroadenosine 1 to the 2 position in 2-cyclopentyla-
minoadenosine 2 reduced the A1 activity by three orders
of magnitude and improved the A2a activity 3-fold. The
related hydroxylated compounds 3 and 4 exhibit
a similar relationship: in this case, the A2a activity
increased by two orders of magnitude and the A1 activity
is decreased ꢀ40-fold.
The actions of adenosine, and the family of four G-
protein coupled 7-trans-membrane receptors (A1, A2a,
A2b, A3) which facilitate its physiological e€ects, have
been of great interest for a number of years.1±3 Adeno-
sine is used clinically for the treatment of supraven-
tricular tachycardia,4 but its lack of selectivity coupled
with its rapid metabolism precludes its therapeutic use
to any great degree. Consequently much e€ort has been
focused on the search for metabolically stable and
selective purinergic receptor agonists.5 In this commu-
nication we will describe the discovery and synthesis of
a series of highly potent A2a purinergic agonists.
The ability to obtain potent agonists at the adenosine
A2a receptor which inhibited human neutrophil func-
tion7,8 o€ered us the opportunity to investigate further
the potential therapeutic application of A2a agonists in
the treatment of in¯ammatory diseases. Selective and
potent A2a agonists were required and our medicinal
chemistry strategy explored three main areas: 2-alkyla-
minoadenosines, 2-alkylamino-NECAs and 2-alkyla-
mino-N6-arylalkyl-NECAs.
Results and Discussion
The structure±activity relationships of adenosine analo-
gues using binding assays have been well-docu-
mented.5,6 As part of our own general investigation into
the structure±activity relationships of substituted ade-
nosines using functional assays, we have investigated
the e€ects on A2a and A1 activity of substitution at the 2
and N6 positions of the adenine ring, as shown in Table 1.
We found that compounds with cycloalkyl substituents
at the 2 position e.g. 2 and 6, were more potent at the
A2a receptor than the corresponding acyclic-alkylamino
compound 7, and the incorporation of a hydroxyl group
*Corresponding author. Fax: +44 1438-763438; e-mail: ser0079@
glaxowellcome.co.uk
0960-894X/00/$ - see front matter # 2000 Published by Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00017-2

404
S. E. Keeling et al. / Bioorg. Med. Chem. Lett. 10 (2000) 403±406
into the 2-cyclopentyl ring yielded A2a agonists more
potent than NECA 5, and with signi®cantly reduced A1
activity. This is illustrated by the pair of compounds 8
and 9, and notably the 40-N-ethylcarboxamide sub-
stituent present in compound 9 confers several fold
improvement in activity as compared to the 40-hydroxy-
methyl substituent in 8. Work investigating N6- sub-
stituted analogues in the 40-hydroxymethyl series had
resulted in the discovery of moderately active A2a ago-
nists (compounds 10±12), but with comparatively lower
A1 activity than the unsubstituted compound 2. Repla-
cement of the 40- hydroxymethyl group in 12 by the 40-
N-ethylcarboxamide substituent again increased the A2a
activity several fold (compound 13), as did hydroxyla-
tion of the 2-cyclopentylamino ring (14). Interestingly,
rat binding data was reported in the literature12 at this
time on N6- arylalkyl compounds, such as 15. This
compound has a Ki=10300 nM at A2 receptors on rat
striatal membranes and Ki=45000 nM at A1 on rat
whole-brain membranes, compared to its 2-H analogue
which had ®gures of 9.4 and 24 nM respectively. The
observed weakened binding anities when a 2-cyclo-
hexylamine substituent is present is contrasted by our
functional results on compound 10 where agonist activity
is retained at the A2a receptor on the human neutrophil.
Table 1. Relationship between substituent position and activityc
At this point, we had identi®ed several potent com-
pounds, but their solubilities were generally poor and
potentially a hindrance to their utility as biological
tools. We focused upon analogues of 13 and 14 and
sought to improve solubility within the series by incor-
porating a basic amino group in the 2 substituent.
Additionally, it had been reported13 that NECA analo-
gues with large 2 substituents possessed signi®cant A2
selectivity. The results are shown in Table 2.
a
b
X
Y
Z
A2a (HN) A1 (GPI)
1
2
3
Cl
HOCH2-
17
5.1
171
1
549
2
H2N-
HOCH2-
HOCH2-
Compounds 16±18 are examples of analogues which
showed signi®cantly improved aqueous solubility com-
pared to that of 13, in addition to activity at the A2a
receptor on the human neutrophil at least 10 times more
potent than NECA 5 itself.
Cl
H
4
H2N-
HOCH2-
1.8
84
5
6
H2N-
H2N-
EtNH.CO-
EtNH.CO-
1
2.2
1
80
Synthesis of A2a Agonists with Improved Solubility
The analogues 16±18, were prepared from the common
intermediate 23, synthesised using one of the two pro-
cedures shown in Scheme 1. Compound 22 was reached
either from guanosine hydrate 19 in six steps, or in four
steps from 2,6-dichloropurine 24 and a fully-protected
ribofuranose sugar 25. Tri-acetylation of 19 gave 20,
7
8
H2N-
H2N-
EtNH.CO-
HOCH2-
85
No data
1470
0.8
9
H2N-
EtNH.CO-
0.1
260
Table 2. A2a and A1 SAR for solubilising 2-substituents
10
11
12
13
14
1-Naphthyl-CH2NH- HOCH2-
25
39
No data
>1000
>612
3-ClPhEtNH-
Ph2CHCH2NH-
Ph2CHCH2NH-
Ph2CHCH2NH-
HOCH2-
HOCH2-
36
EtNH.CO-
EtNH.CO-
6.2
0.16
>5000
390
X
A2a (HN) A1 (GPI) Solubility pH 2.6 (mg/ml)
15
1-Naphthyl-CH2NH- HOCH2-
No data No data
13
16
6.2
>5000
263
0.024
0.08
>=1.87
aAgonist activities were measured in vitro using inhibition of human
neutrophil activation9,10 (HN).
bInhibition of guinea pig Ileum9,10 twitch (GPI).
cNECA (50-N-ethylcarboxamidoadenosine) 5 was used as an internal
standard9±11 and all agonist potencies are quoted as an equie€ective
concentration ratio (ECR) relative to NECA. ECR values were deter-
mined by dividing the EC50 of the test compound by the EC50 of
NECA.
17
18
0.05
0.03
369
672
1.57
>=1.82

S. E. Keeling et al. / Bioorg. Med. Chem. Lett. 10 (2000) 403±406
405
Scheme 1. Synthesis of A2a agonists 16±18. X is de®ned as in Table 2. Reagents: (a) Ac2O, Et3N, MeCN, DMAP; (b) POCl3, PhNMe2, Et4N+Cl ,
MeCN, re¯ux; (c) t-BuONO, Pyridine.HCl, CuCl, DCM, 0±20 ꢁC; (d) Ph2CHCH2NH2, i-Pr2NEt, i-PrOH, re¯ux; (e) Na2CO3, MeOH, H2O; (f)
Me2C(OMe)2, TsOH, Me2CO; (g) KMnO4, KOH, Dioxan; (h) SOCl2, re¯ux; (i) EtNH2, DCM; (j) XH, DMSO, 120 ꢁC; (k) TFA, H2O; (l) 140 ꢁC,
neat; (m) K2CO3, MeOH.
which was converted to 21 using a two-step procedure
via the 6-chloro derivative followed by diazotization
and a modi®ed Sandmeyer reaction. Substitution of the
6-Cl by 2,2-diphenylethylamine in the presence of di-
isopropylethylamine in propan-2-ol was followed by
removal of the acetate protecting groups and reprotec-
tion of the 20,30-diol as the acetonide to yield 22 as its
p-toluenesulfonic acid salt. Alternatively, 26 was syn-
thesised directly by heating a mixture of 24 and 25, and
yielded 22 following steps similar to the conversion of
21. Oxidation of the free hydroxyl group yielded the
intermediate acid which was elaborated to the ethyl
amide 23. Substitution of the 2-Cl with a range of
amines at 120 ꢁC in DMSO and deprotection using
aqueous TFA, or vice versa, followed by puri®cation
using preparative HPLC yielded the series of com-
pounds 16±18 as their tri¯uoroacetic acid salts.
Conclusions
N6,2-disubstituted adenosine analogues have been syn-
thesized and shown to be potent A2a agonists at recep-
tors on the human neutrophil with low activity at the A1
receptors in the system studied. The improved physical
properties will allow their pharmacological e€ects to be
further evaluated for use as potential agents in the
treatment of in¯ammatory diseases.
Acknowledgements
The authors gratefully acknowledge the support of
Physical Sciences in this programme, and for helpful
discussions with Brian Cox and Richard P. C. Cousins
in the preparation of the manuscript.

406
S. E. Keeling et al. / Bioorg. Med. Chem. Lett. 10 (2000) 403±406
8. Cronstein, B. N.; Rosenstein, E. D.; Kramer, S. B.; Weiss-
References and Notes
mann, G.; Hirschhorn, R. J. Immunol. 1985, 135, 1366.
9. Gurden, M. F.; Coates, J.; Ellis, F.; Evans, B.; Foster, M.;
Hornby, E.; Kennedy, I.; Martin, D. P.; Strong, P.; Vardey, C.
J.; Wheeldon, A. Br. J. Pharmacol. 1993, 109, 693.
1. Daly, J. W. J. Med. Chem. 1982, 25, 197.
2. Jacobsen, K. A.; van Galen, P. J. M.; Williams, M. J. Med.
Chem. 1992, 35, 407.
3. Poulsen, S.-A.; Quinn, R. J. Bioorg. Med. Chem. 1998, 6,
619.
10. Kennedy, I.; Gurden, M.; Strong, P. Gen. Pharmacol.
1992, 23, 303.
4. Pantely, G. A.; Bristow, J. D. Circulation 1990, 82, 1854.
5. Trivedi, B. K.; Bridges, A. J.; Bruns, R. F. In Adenosine and
Adenosine Receptors; Williams, M., Ed.; Humana: Clifton, NJ,
1990; pp 57±103.
6. Siddiqi, S. M.; Jacobsen, K. A.; Esker, J. L.; Olah, M. E.; Ji,
X.; Melman, N.; Tiwari, K. N.; Secrist, III, J. A.; Schneller, S.
W.; Cristalli, G.; Stiles, G. L.; Johnson, C. R.; Ijzerman, A. P.
J. Med. Chem. 1995, 38, 1174 and pp 627±628 and references
therein.
11. The assays are fully described in ref 7. NECA 5 has mea-
sured EC50 of 5nM (4.1±6.1; n=46) on HN and 52 nM (4.4±
6.1; n=20) on GPI: EC50 ®gures for NECA are geometric
mean values, with 95% con®dence limits shown in parentheses
where n=number of observations. Where ECR ®gures are
based on n is >=2, arithmetic mean values are given.
12. Trivedi, B. K.; Bruns, R. F. J. Med. Chem. 1989, 32, 1667.
13. Francis, J. E.; Webb, R. L.; Ghai, G. R.; Hutchison, A. J.;
Moskal, M. A.; de Jesus, R.; Yokoyama, R.; Rovinski, S. L.;
Contrado, N.; Dotson, R.; Barclay, B.; Stone, G. A.; Jarvis,
M. F. J. Med. Chem. 1991, 34, 2570.
7. Marone, G.; Rosaria, P.; Sergio, V. Int. Arch. Allergy Appl.
Immunol. 1985, 77, 259.