Synthesis and evaluation of two series of 4′-aza-carbocyclic nucleosides as adenosine A2A receptor agonists

Article abstract of DOI:10.1016/j.bmcl.2009.11.131

The synthesis of two series of 4′-aza-carbocyclic nucleosides are described in which the 4′-substituent is either a reversed amide, relative to the carboxamide of NECA, or an N-bonded heterocycle. Using established purine substitution patterns, potent and

Full text of DOI:10.1016/j.bmcl.2009.11.131

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1219–1224
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and evaluation of two series of 40-aza-carbocyclic nucleosides as
adenosine A2A receptor agonists
*
David Beattie, Andrew Brearley, Zarin Brown, Steven J. Charlton, Brian Cox, Robin A. Fairhurst ,
John R. Fozard, Peter Gedeck, Paul Kirkham, Koremu Meja, Lana Nanson, James Neef, Helen Oakman,
Gillian Spooner, Roger J. Taylor, Robert J. Turner, Ryan West, Hannah Woodward
Novartis Institutes for BioMedical Research, Respiratory Diseases Area, Horsham, United Kingdom
Novartis Institutes for BioMedical Research, Respiratory Diseases Area, Basel, Switzerland
a r t i c l e i n f o
a b s t r a c t
The synthesis of two series of 40-aza-carbocyclic nucleosides are described in which the 40-substituent is
either a reversed amide, relative to the carboxamide of NECA, or an N-bonded heterocycle. Using estab-
lished purine substitution patterns, potent and selective examples of agonists of the human adenosine
A2A receptor have been identified from both series. The propionamides 14–18 and the 4-hydrox-
ymethylpyrazole 32 were determined to be the most potent and selective examples from the 40-reversed
amide and 40-N-bonded heterocyclic series, respectively.
Article history:
Received 25 October 2009
Revised 24 November 2009
Accepted 24 November 2009
Available online 1 December 2009
Keywords:
Ó 2009 Elsevier Ltd. All rights reserved.
Adenosine A2A receptor agonists
Carbocyclic nucleosides
Anti-inflammatory
Adenosine plays a key role in regulating many aspects of phys-
iology, in part, through the interaction with the four purinergic P1
receptors termed: A1, A2A, A2B and A3.1 This family of four G pro-
tein-coupled receptors has been, and continues to be the subject
of extensive investigations to identify selective agonists and antag-
onists as potential therapies for treating a wide range of diseases
involving: smooth muscle contraction, neurotransmission, platelet
aggregation, pain, wound healing, cardiac function, cardioprotec-
tion, sleep disorders and immune response.2 One example being
the activation of the adenosine A2A receptor which has been dem-
onstrated to result in broad spectrum anti-inflammatory activity,
directly effecting multiple inflammatory cell types in man.3 Our
interest related to the potential for adenosine A2A receptor agon-
ism as a treatment for the chronic inflammation associated with
the respiratory diseases asthma and chronic obstructive pulmon-
ary disease (COPD).4 In vivo studies supporting this possibility in-
clude the A2A receptor agonist CGS21680 having been shown to
inhibit ovalbumin and lipopolysaccharide (LPS) induced pulmon-
ary inflammation in rat and murine models.5 However, the hypo-
tensive effects associated with systemic adenosine A2A receptor
activation are predicted to lead to the requirement for direct deliv-
ery of compounds to the lung if effective treatments for asthma
and COPD are to be realised. This route of administration is antic-
ipated to provide the potential for agents with acceptable cardio-
vascular side-effect profiles in the case of ligands specifically
designed to maximise their local efficacy at the site of administra-
tion whilst also being associated with a controlled level of systemic
exposure.6 The challenge in identifying inhaled A2A receptor
agonists with such lung-targeted profiles has been highlighted
recently by the clinical data reported for GW328267X and the pre-
clinical studies leading to the identification of the clinical candi-
date UK432,097.7
To date essentially all reported agonists of the adenosine A2A
receptor have been based upon modifications of the endogenous
ligand adenosine.8 Changes to the 50-hydroxy residue and the
introduction of substituents into the purine base have led to the
identification of potent and selective A2A receptor agonists such
as CGS21680 and 2-hexynyl-50-N-ethylcarboxamidoadenosine
(HENECA), Figure 1.9 Both of these ligands are based upon the 50-
carboxamide modification of the potent pan adenosine receptor
agonist 50-N-ethylcarboxamidoadenosine (NECA).10
Further transformation of the 50-carboxamide into a heterocy-
clic residue has also been shown to be the basis for potent
adenosine A2A receptor agonists, as exemplified by the 2-ethyl-5-
tetrazolyl moiety of GW328267X and the analogues described in
related claims.11
As part of our interest in identifying lung-targeted adenosine A2A
receptor agonists for inhaled administration we were interested in
novel nucleoside templates that could offer advantages in terms of
biological activity, stability and synthetic accessibility. In particular,
carbocyclic nucleosides in which an amino group is introduced at
* Corresponding author.
E-mail address: robin.fairhurst@novartis.com (R.A. Fairhurst).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.11.131

1220
D. Beattie et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1219–1224
teric potential of these two 40-aza-carbocyclic nucleoside series in
NH2
O
which the 40-ethyl carboxamide residue of NECA is compared with
40-propionamide and 40-methylpyrazole moieties as representative
examples of the reversed amide and the N-bonded heterocyclic ser-
ies.12 Comparing the electron density surfaces, a high degree of spa-
tial alignment is evident between the high electron density regions,
highlighted in red, associated with the oxygen atoms of the carbonyl
residue of NECA and the carbonyl residue of the reversed amide
example. The carbonyl residue of NECA having been proposed to
be involved in an important hydrogen bonding interaction with
serine 277 on the seventh transmembrane spanning domain of the
activated conformation of the adenosine A2A receptor.13
N
N
N
OH
NH
O
N
N
O
H
HO
OH
CGS21680
NH2
N
N
NH
Similarly for the N-bonded heterocycle series, the imino-nitro-
gen of the pyrrazole ring exhibits a comparable high electron den-
sity surface which also closely matches the position of the carbonyl
residue of NECA. Thus, highlighting the need for an imino-nitrogen
adjacent to the site of attachment to the carbocyclic ring in the 40-
heterocyclic series, if this key interaction is to be maintained. Addi-
tionally, the substitution of the pyrazole residue in the 4-position
presents a methyl residue in a similar spatial orientation to the al-
kyl residues of both NECA and the reversed-amide series providing
the opportunity for homologation along a similar vector. Interest-
ingly, in contrast to the 40-aza-carbocyclic nucleoside series, de-
scribed above, the corresponding 40-aza-ribose analogues would
be anticipated to be of relatively low stability. The reason for the
instability of such 40-aza-ribose variants being as a consequence
of their doubly-anomeric nature, and to the best of our knowledge
no reports of either have been made.
O
N
N
O
HO
OH
HENECA
NH2
OH
N
N
N
N
O
N
N
N
N
N
H
HO
OH
GW328267X
Figure 1. Structures of the adenosine A2A receptor agonists CGS21680, HENECA and
GW328267X.
The preparation of 40-aza analogues of the carbocyclic nucleo-
side aristeromycin as antiviral agents, including reversed amide
analogues, has been reported previously.14 These derivatives have
been prepared starting from either: (1R,3S)-4-cyclopentene-1,3-
diol monoacetate in which the purine base is introduced via a pal-
ladium(0) catalysed allylic coupling reaction and the 40-amino sub-
stituent via a Mitsunobu reaction, with azide as the nucleophile, or
from the cycloaddition product 2-oxa-3-azabicyclo[2.2.1]hept-5-
ene in which cleavage of the N-O bond provides the 40-amino sub-
stituent directly as well as an allylic alcohol for the introduction of
the purine base via a palladium(0) allylic coupling reaction. In both
cases the 20,30-diol is introduced via a cis-dihydroxylation reaction
with addition to the double bond required to occur from the steri-
cally less hindered face. The nature of the substituents at the 10-
and 40-positions having been shown to be important in controlling
the facial selectivity of the cis-dihydroxylation reaction.15 We had
envisioned a similar approach starting from (1R,3S)-4-cyclopen-
tene-1,3-diol monoacetate, but using two palladium(0) catalysed
allylic coupling reactions to introduce both the purine base and
the 40-aza substituent, as outlined in Figure 3. This approach was
selected because it was anticipated to provide: the opportunity
for the shortest synthetic sequences, common intermediates to
the 40-position provide the opportunity to generate reverse-amide
variants of NECA and N-bonded heterocyclic variants of the
C-bonded 40-(2-tetrazolyl) residue present in GW328267X. The
electron density comparisons shown in Figure 2 highlight the isos-
4'-Substituent electron
Structure
density surface
NH2
N
N
O
O
N
N
N
H
HO
OH
NECA
NH2
N
O
N
N
HN
N
HO
OH
1. Palladium(0) catalysed allylic
coupling to introduce a purine
substituted at the 9-position
2. Activate for a palladium(0) catalysed
allylic coupling to introduce an N-
heterocycle, or ammonia equivalent
Propionamide example of
the reversed amide series
NH2
N
N
N
HO
O
O
N
N
N
HO
OH
3. cis-Dihydroxylation
under steric control
4-Methylpyrazole example of
the N-bonded heterocycle series
Figure 2. Electron density profiles of prototype reversed amide and N-bonded
heterocyclic 40-aza-carbocyclic nucleoside analogues in comparison with NECA.
Figure 3. Key aspects of the synthetic approach to the 40-aza-carbocyclic nucleo-
side series from (1R,3S)-4-cyclopentene-1,3-diol monoacetate.

D. Beattie et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1219–1224
1221
X
X
X
O
N
N
N
tBuO
tBuO
O
N
N
N
N
EtO
iii), iv)
N
i), ii)
HO
OAc
N
+
N
Cl
O
N
H
N
Cl
N
Cl
O
1a-c
2a-c
HO
OH
3a,b
a, X = Cl
b, X = bis-(4-methoxyphenyl)-methylamino
c, X = 2,2-diphenylethylamino (DPEA)
NH2
NH2
O
O
N
N
N
N
N
v) vi)
vii)
3 b
HN
N
HN
N
Cl
N
HO
OH
HO
OH
4
5
Scheme 1. Synthesis of the 2-(1-hexynyl) 40-aza-carbocyclic reversed amide nucleoside 5. Reagents and conditions: (i) 1.0 equiv NaH, 0.05 equiv (Ph3P)4Pd, 0.15 equiv Ph3P,
50:1 THF/DMSO, 2 h, 50 °C (a 45–62%, b,c 64–84%); (ii) to prepare 2a: 4 equiv EtOC(O)Cl, 3 equiv pyridine, THF, 1 h 0 °C (67–75%), or to prepare 2b,c: 2 equiv
3-ethoxycarbonylbenzotriazole-1-oxide, cat. DMAP, 5:1 i-Pr2NEt/pyridine, 18 h, rt (65–86%); (iii) to prepare 3a: 1.1 eqiuv (Boc)2NH, 0.05 equiv Pd2(dba)3, 0.15 equiv Ph3P,
THF, 3 h, rt (23–50%), or to prepare 3b: 1.1 eqiuv (Boc)2NH, 0.05 equiv (Ph3P)4Pd, 0.15 equiv Ph3P, THF, 3 h, rt (32%); (iv) 0.15 equiv OsO4, 2 equiv NMMO, 10:1 THF/H2O, 18 h,
rt (58–89%); (v) 2:1 CH2Cl2/CF3CO2H, 18 h, rt (75%); (vi) 1 equiv propionyl chloride, 5 equiv i-Pr2NEt, THF, 3 h, rt (>90%); (vii) 10 eqiuv 1-hexyne, 0.13 equiv CuI, 0.13 equiv
(Ph3P)2PdCl2, 0.25 equiv Ph3P, 2:1 Et2NH/DMF, 3 h, 120 °C Biotage Initiator™ microwave (46%).
prepare both the reversed amide and N-bonded heterocyclic series
and favourable selectivities in the cis-dihydroxylation reaction.
To start exploring the two 40-aza-carbocyclic nucleoside series,
the reversed amide analogue of HENECA and N-bonded heterocy-
clic analogues of GW328267X were selected as the initial targets
for synthesis. Starting from (1R,3S)-4-cyclopentene-1,3-diol mono-
acetate, palladium(0) catalysed allylic coupling reactions with the
sodium anions of the purine derivatives 1a–c provided the prod-
ucts coupled at the 9-position in good yield, Scheme 1.16 Small
amounts of an isomeric product, presumed to be the 7-substituted
purine analogues, were identified in the crude reaction mixtures
from these allylic coupling reactions and were removed upon puri-
fication. Subsequent acylation to prepare the ethyl carbonate 2a
proceeded smoothly with ethyl chloroformate. Similarly, for the
6-substituted-amino analogues, acylation with 3-ethoxycarbonyl-
benzotriazole-1-oxide was found to give improved yields of the
carbonates 2b,c.17 Starting from 2b to prepare the reversed amide
series, a second palladium(0) catalysed allylic coupling reaction
X
NH2
OH
N
N
N
N
iii)
i), ii)
R2
N
R2
2b
N
N
Cl
N
N
H
HO
OH
HO
OH
6b
7-12
X
X
N
N
N
N
N
N
ii), iv)
i)
N
Y
R2
2c
R2
N
N
N
H
N
Cl
HO
OH
27c
28, 29
X
Cl
N
N
N
N
N
N
N
N
v)
vi), ii), iv)
Y
R2
2a
R2
N
N
N
H
Cl
HO
OH
30
31-33
Scheme 2. Synthesis of 40-aza-carbocyclic N-bonded heterocyclic nucleosides. Reagents and conditions: (i) 1.1 equiv N-heterocycle (R2–H), 0.05 equiv (Ph3P)4Pd, 0.15 equiv
Ph3P, THF, 3 h, rt (27–95%); (ii) 15 mol % OsO4, 2 equiv NMMO, 10:1 THF/H2O, 18 h, rt (20–71%); (iii) 5 equiv (S)-phenylalaninol, 1,2-dichlorobenzene, 3 h, 240 °C Biotage
Initiator™ microwave (10–33%); (iv) 5 equiv histamine derivative, DMSO, 8 h, 160 °C (31%), or 5 equiv histamine derivative, 1 equiv NaI, 1:1 CH3CN/NMP, 1–2 h, 200–240 °C
Biotage Initiator™ microwave (27–42%); v) 1.1 equiv N-heterocycle, 0.05 equiv Pd2(dba)3, 15 mol % Ph3P, THF, 1 h, 50 °C (68–72%); (vi) 1.1 equiv amine, 1.2 equiv i-Pr2NEt,
THF, 2–24 h, 35–50 °C (70–94%).

1222
D. Beattie et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1219–1224
with bis-(tert-butoxycarbonyl)-amine as the nucleophile intro-
duced the 40-aza functionality with the anticipated high level of re-
gio and stereoselectivity. Dihydroxylation of the 2,3-double bond
then proceeded under steric control with a high degree of facial
selectivity to give the intermediate 3b bearing all the key function-
ality of the targeted compounds. To prepare the reversed amide
analogue of HENECA, treatment of 3b with trifluoroacetic acid re-
moved the tert-butoxycarbonyl (Boc) and 4,40-dimethoxybenzhyd-
ryl (DMBH) protecting groups from the two amino residues.
Acylation of the free base form of the diamino intermediate gave
the 40-propionamide 4 which then underwent a Sonogashira cou-
pling with 1-hexyne to give the targeted product 5. Similarly to
prepare the N-bonded heterocyclic analogues of GW328267X the
ethyl carbonate 2b underwent palladium(0) catalysed allylic
coupling reactions with a series of N–H heterocycles to give the
corresponding N-allylated analogues, as shown in Scheme 2 with
the substituents detailed in Table 1. Isolated yields for these
N-allylation reactions ranged from good for the pyrazole analogues
(76–95%), to modest in the case of the pyridine-2-one (47%) and
indazole (27%) examples. In an analogous fashion to the prepara-
tion of 5, dihydroxylation of the N-allylated heterocycles
proceeded with a high level of facial selectivity to give the interme-
diates 6b. Finally, SNAr reactions to introduce (S)-phenylalaninol
into the purine 2-position of the intermediates 6b occurred under
elevated temperatures which also resulted in the removal of the
DMBH protecting groups to give directly the targeted products
7–12. The modest yields observed in these final coupling reactions
were not improved upon if a two-step approach was followed in
which the DMBH group was removed prior to displacement of
the 2-chloro substituent.
To further explore the 40-aza carbocyclic nucleosides, examples
bearing histamine 2-substituents were selected for synthesis in
both the reversed amide and N-bonded heterocyclic series. This
selection was based upon the NECA analogues having been re-
ported to be highly potent and selective adenosine A2A receptor
agonists with good solubility profiles.18 The reversed amide exam-
ples were prepared by an analogous sequence to that described
above as outlined in Scheme 3 with the substituents detailed in Ta-
ble 1. Starting from 3a, displacement of the 6-chloro substituent
with either 2,2-diphenylethylamine or 1-ethylpropylamine was
followed by Boc deprotection to give the 40-primary amines
13c,d. Subsequent acylation of 13c,d with a range of acid chlorides
Table 1
Purinergic P1 receptor binding and functional data
NH2
X
N
X
OH
O
N
N
N
N
N
R1
N
N
N
N
N
N
N
Y
Y
R2
HN
R2
N
N
N
N
N
N
H
H
H
HO
OH
HO
OH
HO
OH
General structure 7-12
General structure 14-26
General structure 28, 29, 31-33
Compound R1 (reversed amide)/R2 (N-bonded
X
Y
A2A Ki (nM) Neutrophil IC50 (nM) A1 EC50 (nM) A2B EC50 (nM) A3 EC50 (nM)
heterocycle)
NECA
25
16
30
38
63
CGS21680
GW328267
5
7
8
21
2.3
18
702
229
182
>10,000
5406
5491
5.2
4.6
5.2
5.4
8.0
39
27
382
16
7.9
19
0.4
12
nd
56
62
nd
nd
nd
5.9
4.7
5.0
2.7
8.3
160
67
542
60
40
>10,000
882
>10,000
nd
>10,000
>10,000
nd
>10,000
51
>10,000
nd
>10,000
>10,000
nd
nd
nd
>10,000
nd
nd
9866
nd
nd
nd
nd
nd
nd
nd
nd
613
4.2a
105
nd
Et
1-Pyrazolyl
4-Methyl-1-pyrazolyl
4-Ethyl-1-pyrazolyl
4-Phenyl-1-pyrazolyl
1-Indazolyl
1-Pyridin-2-one
Et
Et
Et
Et
Et
n-Pr
i-Pr
tert-Bu
CycloPr
CycloBu
CH2Ph
CH2CH2Ph
Et
4-Ethyl-1-pyrazolyl
5-Methyl-2-tetrazolyl
4-Ethyl-1-pyrazolyl
4-Hydroxymethyl-1-pyrazolyl
2-(4-Ethyl-1,2,3-triazolyl)
449a
776a
nd
9
10
11
12
14
15
16
17
18
19
20
21
22
23
24
25
26
28
29
31
32
33
nd
nd
nd
nd
139
42
44
1640
2520
nd
DPEA
DPEA Me
DPEA Et
EPA
EPA
DPEA i-Pr
DPEA i-Pr
DPEA i-Pr
DPEA i-Pr
DPEA i-Pr
DPEA i-Pr
DPEA i-Pr
H
>10,000
>10,000
nd
>10,000
>10,000
nd
nd
nd
>10,000
>10,000
nd
nd
nd
nd
nd
>10,000
4100
nd
Et
i-Pr
nd
nd
357a
480b
nd
1666
5472
nd
nd
nd
nd
nd
nd
>10,000
789a
nd
DPEA CH2CO2H 97
nd
nd
101
169
5.3
nd
nd
nd
nd
nd
>10,000
nd
DPEA Et
DPEA i-Pr
EPA
DPEA Et
DPEA Et
146
38
34
5.4
48
Et
All data are expressed as means from duplicates derived from 1 to 5 separate experiments.
All functional assay EC50 data reflect agonist activity unless otherwise indicated.
nd; not determined.
a
Most potent activity of the compound determined to be >50% antagonism of the I-AB-MECA response.
Most potent activity determined to be a partial agonism, producing 48% the maximal response of I-AB-MECA.
b

D. Beattie et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1219–1224
1223
X
X
O
N
N
N
N
R1
N
N
N
N
Y
i), ii)
iii), iv)
H2N
HN
3 a
N
Cl
N
N
H
HO
OH
HO
OH
13c,d
14-26
Scheme 3. Synthesis of the 2-histamine 40-aza-carbocyclic reversed amide nucleosides 14–26. Reagents and conditions: (i) 1.1 equiv 2,2-diphenylethylamine or 1-
ethylpropylamine, 1.2 equiv i-Pr2NEt, THF, 24 h, 50 °C (91–95%); (ii) 2:1 CH2Cl2/CF3CO2H, 18 h, rt (74–75%); (iii) 1 equiv acid chloride (R1C(O)Cl), 5 equiv i-Pr2NEt, THF, 3 h, rt
(>90%); (iv) for the preparation of 14–25: 5 equiv histamine derivative, 1,2-dichlorobenzene, 1–2 h, 200–240 °C Biotage Initiator™ microwave (19–48%), or to prepare 26:
5 equiv histamine derivative, 1 equiv NaI, 1:1 CH3CN/NMP, 1 h, 200 °C Biotage Initiator™ microwave (41%).
and then displacement of the 2-chlorosubstituent with a series of
histamine derivatives (Y = H, Me, Et, i-Pr and CH2CO2C6H11) gave
the products 14–26.19 The carboxylic acid example 26 being iso-
lated directly from the SNAr reaction following an in situ hydrolysis
of the histamine acetic acid cyclohexyl ester derivative. Synthesis
of the N-bonded heterocyclic analogues is shown in Scheme 2 with
the substituents detailed in Table 1. Starting from the 6-DPEA
substituted carbonate 2c, 4-ethylpyrazole and 5-methyltetrazole
were introduced by allylic coupling reactions using the
(Ph3P)4Pd/Ph3P catalyst system to give the intermediates 27c. In
the case of 5-methyltetrazole a single regioisomer was obtained
exhibiting spectral properties consistent with the 2-substituted
derivative. Only approximately 10% of an isomeric product was
present in the crude allylic coupling reaction mixture for this
example, consistent with formation of a small amount of the
1-substituted tetrazole. Dihydroxylation of the intermediates 27c
was then followed by SNAr reactions to deliver the products 28
and 29. Similarly the 2,6-dichlorocarbonate 2a underwent allylic
coupling reactions with 4-ethylpyrazole, 4-hydroxymethylpyraz-
ole and 4-ethyl-1,2,3-triazole to give the intermediates 30 using
a Pd2(dba)3/Ph3P catalyst protocol. For the 1,2,3-triazole example
again predominantly a single product was obtained in the allylic
coupling reaction and determined to be the 2-substituted regio-
isomer by NMR spectroscopy. Dihydroxylation and SNAr reactions
with the intermediates 30 then yielded the targeted products
31–33.
the unsubstituted example 7. The larger phenyl residue in 10
was not tolerated and neither was the fusion of the phenyl ring
in the indazole example 11. Incorporation of pyridin-2-one in
example 12, as a cyclic reversed amide analogue, showed little
affinity for the adenosine A2A receptor. The weak activity of 12 is
also consistent with the requirement for an imino-nitrogen adja-
cent to the site of attachment to the carbocyclic ring for the N-
bonded heterocyclic series, as hypothesised from the electron den-
sity analysis.
Analysis of the data from the 2-histamine reversed-amide ana-
logues showed all the propionamide examples 14–18 to be potent
adenosine A2A receptor agonists with both the 2,2-diphenylethyla-
mino (DPEA) and 1-ethylpropylamino (EPA) purine 6-substituents.
High functional selectivity over the other P1 receptors was also ob-
served, in particular for the 6-EPA substituted examples 17 and 18
where greater than 100-fold selectivity was determined over the
closely related adenosine A3 receptor. The nature of the imidazole
substituent Y, as methyl, ethyl, isopropyl or in the unsubstituted
form in examples 14–18 was seen to have little impact on the over-
all P1 receptor profiles for these examples. Increasing the size of
the amide group reduced the affinity and functional activity at
the adenosine A2A receptor: the single homologation to the n-pro-
pyl, isopropyl and cyclopropyl analogues 19, 20 and 22 resulted in
a 10–30-fold drop in functional potency relative to the propiona-
mide. Interestingly, for the larger cyclopropyl and cyclobutyl ana-
logues 22 and 23 this also resulted in a drop in intrinsic efficacy
at the adenosine A3 receptor. These derivatives exhibited partial
agonist and antagonist adenosine A3 activities, respectively, rela-
tive to the fuller agonism seen for the propionamide examples
14 and 15. The histamine acetic acid derivative 26, with a carbox-
ylic acid residue in a similar position to that of CGS21680, surpris-
ingly showed a 20-fold drop in affinity towards the adenosine A2A
receptor relative to the alkyl substituted analogues 15 and 16.
Assessing the data for the N-bonded heterocyclic 2-histamine ana-
logues: the methyl and ethyl substituted pyrazole, 1,2,3-triazole
and tetrazole examples 28, 29, 31 and 33 all showed only modest
affinity for the adenosine A2A receptor which translated into a sim-
ilar level of functional activity for the two examples tested. Of
greater interest was the 4-hydroxymethylpyrazole analogue 32
which exhibited sub 10 nM adenosine A2A receptor affinity and
functional activity. In terms of selectivity, greater than 100-fold
selectivity was observed over the other P1 purinoceptors for 32,
which again exhibited antagonist activity at the adenosine A3
receptor. Presumably the increase in activity and selectivity for
compound 32, relative to the other N-bonded heterocyclic deriva-
tives, is as a consequence of a favourable interaction of the
hydroxymethyl residue of the pyrazole moiety with the adenosine
A2A receptor.
To assess the activity of the compounds against the human
adenosine A2A receptor: affinity was measured in a radioligand
binding assay versus [3H]-NECA with membranes produced from
SF21 cells coexpressing recombinant human adenosine A2A recep-
tors and the human G protein subunits Gas2, b4 andg2. Functional
adenosine A2A activity was assessed by the ability to inhibit fMLP
induced release of reactive oxygen species from isolated human
neutrophils.20 Selectivity versus the other human P1 purinoceptors
was assessed in a series of functional assays: GTPcS for the aden-
osine A1 and A3 receptors expressed in Chinese hamster ovary
(CHO) cells, in both agonist and antagonist modes versus the li-
gands NECA and I-AB- MECA, respectively; and with a CRE-lucifer-
ase reporter gene assay measuring cAMP production in CHO cells
expressing the adenosine A2B receptor, with NECA as the reference
agonist. All the data are shown in Table 1.
Comparison of the data generated for the reversed amide HEN-
ECA analogue 5 with the human data reported for the parent com-
pound shows: a similar high level of affinity for, and functional
activity at, the adenosine A2A receptor; a comparable high level
of selectivity over the adenosine A1 and A2B receptors; and modest
10-fold selectivity over the adenosine A3 receptor.21 In contrast,
the most active 4-methyl and 4-ethylpyrazole N-bonded heterocy-
clic analogues 8 and 9 of GW328267X are 100-fold less active
adenosine A2A receptor agonists compared to the parent com-
pound. However, 8 and 9 do possess similar P1 receptor selectivity
profiles to GW328267X, including antagonist activity against the
I-AB-MECA activation of the adenosine A3 receptor. The pyrazole
4-methyl and 4-ethyl substituents proved more effective than
Interspecies cross-reactivity variation has been described for a
number of purinergic P1 receptor ligands as a consequence of vari-
ations in the amino acid receptor sequences.22 Additionally, our
primary interest related to the potential for adenosine A2A receptor
agonists as anti-inflammatory agents, where consistent responses
can be established across a wide range of human inflammatory cell

1224
D. Beattie et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1219–1224
3. (a) Haskó, G.; Pacher, P. J. Leukocyte Biol. 2008, 83, 447; (b) Palmer, T. M.;
Table 2
Trevethick, M. A. Br. J. Pharmacol. 2008, 153, S27.
Comparison of the inhibition of TNF-
a
release in rat and human LPS stimulated PBMC
4. vanden Berg, M.; Hylkema, M. N.; Versluis, M.; Postma, D. S. Drugs R D 2007, 8,
13.
assays
Compound
PBMC IC50 (nM)
5. (a) Bonneau, O.; Wyss, D.; Ferretti, S.; Blaydon, C.; Stevenson, C. S.; Trifilieff, A.
Am. J. Physiol. Lung Cell Mol. Physiol. 2006, 290, L1036; (b) Fozard, J. R.; Ellis, K.
M.; Villela Dantas, M. F.; Tigani, B.; Mazzoni, L. Eur. J. Pharmacol. 2002, 438, 183.
6. Trevethick, M. A.; Mantell, S. J.; Stuart, E. F.; Barnard, A.; Wright, K. N.; Yeadon,
M. Br. J. Pharmacol. 2008, 155, 463.
7. (a) Mantell, S. J.; Stephenson, P. T.; Monaghan, S. M.; Maw, G. N.; Trevethick, M.
A.; Yeadon, M.; Walker, D. K.; Selby, M. D.; Batchelor, D. V.; Rozze, S.;
Chavaroche, H.; Lemaitre, A.; Wright, K. N.; Whitlock, L.; Stuart, E. F.; Wright, P.
A.; Macintyre, F. Bioorg. Med. Chem. Lett. 2009, 19, 4471; (b) Mantell, S. J.;
Stephenson, P. T.; Monaghan, S. M.; Maw, G. N.; Trevethick, M. A.; Yeadon, M.;
Keir, R. F.; Walker, D. K.; Jones, R. M.; Selby, M. D.; Batchelor, D. V.; Rozze, S.;
Chavaroche, H.; Hobson, T. J.; Dodd, P. G.; Lemaitre, A.; Wright, K. N.; Stuart, E.
F. Bioorg. Med. Chem. Lett. 2008, 18, 1284; (c) Luijk, B.; vanden Berge, M.;
Kerstjens, H. A. M.; Postma, D. S.; Cass, L.; Sabin, A.; Lammers, J.-W. J. Allergy
2008, 63, 75.
Human
Rat
NECA
CGS21680
GW328267
17
18
32
40
58
1.3
4.7
5.9
21
187
83
2.9
255
458
3591
All data are expressed as means from 2 to 5 separate experiments.
types.3 In contrast, reproducing these A2A responses in the equiva-
lent inflammatory cell types of lower species was found to be prob-
lematic. After reviewing several test systems an equivalent
adenosine A2A dependant response was established for the inhibi-
tion of the LPS induced release of tumour necrosis factor alpha
8. Diniz, C.; Borges, F.; Santana, L.; Uriarte, E.; Oliveira, J. M. A.; Gonçalves, J.; Fresco,
P. Curr. Pharm. Des. 2008, 14, 1968.
9. (a) Cristalli, G.; Eleuteri, A.; Vittori, S.; Volpini, R.; Lohse, M. J.; Klotz, K.-N. J.
Med. Chem. 1992, 35, 2363; (b) Hutchison, A. J.; Webb, R. L.; Oei, H. H.; Ghai, G.
R.; Zimmerman, M. B.; Williams, M. J. Pharmacol. Exp. Ther. 1989, 251, 47.
10. Stein, H. H.; Prasad, R. N. Adenosine-50-Carboxylic Acid Amides. U.S. Patent
3,914,415, Oct. 21, 1975.
11. (a) Bevan, N.; Butchers, P. R.; Cousins, R.; Coates, J.; Edgar, E. V.; Morrison, V.;
Sheehan, M. J.; Reeves, J.; Wilson, D. J. Eur. J. Pharmacol. 2007, 564, 219; b Chan,
C.; Cousins, R. P. C.; Cox, B. 2-(Purin-9-yl)-tertrahydrofuran-3,4-diol
derivatives. WO 99/38877, Aug. 5, 1999.
(TNF-a) from isolated human and rat peripheral blood mononu-
clear cells (PBMC).23 Using this test system it was possible to assess
the extent of any cross-reactivity differences between human and
rat as a consequence of receptor sequence variation. Cross-reactiv-
ity data for the most interesting 40-azacarbocyclic nucleoside
examples and the reference compounds are shown in Table 2.
The three reference compounds exhibited modest 2–5-fold shifts
to lower IC50 values in the rat relative to the human PBMC assay.
In contrast, the reversed amide analogues 17 and 18 demonstrated
54- and 78-fold shifts in activity, respectively. Even more dramatic
was the 171-fold shift observed for the N-bonded heterocyclic
example 32. These data highlight that an understanding of inter-
species cross-reactivity, as a consequence of amino acid sequence
differences, will be an important parameter for the preclinical opti-
misation of both the N-bonded heterocycle and reversed amide 40-
aza carbocyclic nucleoside series.
12. Electron density calculations were carried out using Jaguar 7.5 from the
Schrödinger 2008 modelling suite software to build and minimise the 40-
substituents. For the minimisation: density functional theory was used with
**
the B3LYP functional and 6-311G basis sets. The default settings were used
with an accuracy level of ‘accurate’. For the final geometry the electrostatic
potential was calculated and displayed. For the images the range of the colour
ramp was set the same for all of the three substituents shown in Figure 2.
13. (a) Tuccinardi, T.; Ortore, G.; Manera, C.; Saccomanni, G.; Martinelli, A. Eur. J.
Med. Chem. 2006, 41, 321; (b) Kim, S.-K.; Gao, Z.-G.; Van Rompaey, P.; Gross, A.
S.; Chen, A.; Van Calenbergh, S.; Jacobsen, K. A. J. Med. Chem. 2003, 46, 4847.
14. (a) Ando, T.; Kojima, K.; Chahota, P.; Kozaki, A.; Milind, N. D.; Kitade, Y. Bioorg.
Med. Chem. Lett. 2008, 18, 2615; (b) Yang, M.; Schneller, S. W. Bioorg. Med.
Chem. Lett. 2005, 13, 877; (c) Ghosh, A.; Miller, M. J.; De Clercq, E.; Balzarini, J.
Nucleosides Nucleotides 1999, 18, 217; (d) Hegde, V. R.; Seley, H. K.; Schneller, S.
W. J. Org. Chem. 1998, 63, 7092; (e) Mulvihill, M. J.; Miller, M. J. Tetrahedron
1998, 54, 6605.
15. (a) Donohoe, T. J.; Blades, K.; Moore, P. R.; Waring, M. J.; Winter, J. J. G.;
Helliwell, M.; Newcombe, N. J.; Stemp, G. J. Org. Chem. 2002, 67, 7946; (b)
Ghosh, A.; Ritter, A. R.; Miller, M. J. J. Org Chem. 1995, 60, 5808.
16. Siddiqi, S. M.; Jacobson, K. A.; Esker, J. L.; Olah, M. E.; Ji, X.; Melman, N.; Tiwari,
K. N.; Secrist, J. A., III; Schneller, S. W.; Cristalli, G.; Stiles, G. L.; Johnson, C. R.;
Ijzerman, A. P. J. Med. Chem. 1995, 38, 1174.
In summary, reversed amide and N-bonded heterocycle 40-aza
carbocyclic nucleosides were identified as scaffolds of interest
through in silico electron density comparisons with the known
adenosine A2A receptor agonist NECA. An efficient synthetic route
to these two 40-aza carbocyclic nucleoside series is described
which utilises two successive palladium(0) allylic coupling reac-
tions as the key transformations. Following this approach, highly
potent and selective agonists of the human adenosine A2A receptor
have been identified in both the N-bonded heterocycle and re-
versed amide series by employing established purine substitution
patterns. Using these established purine substitution patterns the
propionamides 14–18 and the 4-hydroxymethylpyrazole 32 were
determined to be the most potent and selective examples from
the 40-reversed amide and 40-N-bonded heterocyclic series, respec-
tively. An assessment of the interspecies variability highlighted rel-
atively poor rat cross-reactivity for the most potent examples at
the human adenosine A2A receptor from both 40-aza carbocyclic
nucleoside series. The optimisation of these two 40-aza carbocyclic
nucleoside series as lung-targeted adenosine A2A receptor agonists
for use as inhaled anti-inflammatory agents will be the subject of
future publications.
17. Wuts, P. G. M.; Ashford, S. W.; Anderson, A. M.; Atkins, J. R. Org. Lett. 2003, 5,
1483.
18. Keeling, S. E.; Albinson, F. D.; Ayres, B. E.; Butchers, P. R.; Chambers, C. L.;
Cherry, P. C.; Ellis, F.; Ewan, G. B.; Gregson, M.; Knight, J.; Mills, K.; Ravenscroft,
P.; Reynolds, L. H.; Sanjar, S.; Sheehan, M. J. Bioorg. Med. Chem. Lett. 2000, 10,
403.
19. Jain, R.; Cohen, L. A. Tetrahedron 1996, 52, 5363.
20. Briefly, the neutrophil assay was carried out in the following manner:
neutrophils were isolated from human blood by centrifuge and resuspended in
Hank’s balanced salt solution containing 0.1% bovine serum albumin to give a
suspension of 4 ꢀ 106 cells/ml (neutrophil purity >97%) and 50
l
l dispensed per
l) and after a 30 min
l of a 400 M solution) and
l of a 4 M solution) were added.
assay well (SigmacoteÒ treated). Test article was added (50
l
preincubation at 37 °C solutions of lucigenin (50
l
l
then formylated tripeptide MLP (fMLP) (50
l
l
Release of reactive oxygen species was assessed from the chemiluminescence
generated over the 4 min incubation period following fMLP addition.
21. Volpini, R.; Costanzi, S.; Lambertucci, C.; Taffi, S.; Vittori, S.; Klotz, K.-N.;
Cristalli, G. J. Med. Chem. 2002, 45, 3271.
22. (a) Melman, A.; Gao, Z.-G.; Kumar, D.; Wan, T. C.; Gizewski, E.; Auchampach, J.
A.; Jacobsen, K. A. Bioorg. Med. Chem. Lett. 2008, 18, 2813; (b) Cappellacci, L.;
Franchetti, P.; Vita, P.; Petrelli, R.; Lavecchia, A.; Costa, B.; Spinetti, F.; Martini,
C.; Klotz, K.-N.; Grifantini, M. Bioorg. Med. Chem. 2008, 16, 336; (c) Gao, Z.-G.;
Blaustein, J. B.; Gross, A. S.; Melman, N.; Jacobson, K. A. Biochem. Pharmacol.
2003, 65, 1675; (d) Linden, J.; Jacobson, K. A. Dev. Cardiovasc. Med. 1998, 209, 1.
23. Briefly, the PBMC assays were carried out in the following manner:
mononuclear cells were isolated from blood, resuspended with 10% fetal calf
Acknowledgements
The authors would like to thank Dr. Brian Everatt, Clive Aldcroft,
Thomas Lochmann and Dr. John Tyler for analytical support.
serum to give a 1 ꢀ 106 cells/ml suspension and 100
ll dispensed per well. Test
article was added (50 ll) and after a 10 min preincubation LPS was added
References and notes
(50 ll of a 10 ng/ml solution). After incubating for a further 20 h at 37 °C TNF-a
levels were assessed by ELISA. The addition of the adenosine A2A selective
antagonist ZM-241,385 (10 nM) to the assays from both species produced the
anticipated displacement of the CGS21680 dose–response curves, consistent
with the response being adenosine A2A receptor mediated.
1. Fredholm, B. B.; Ijzerman, A. P.; Jacobson, K. A.; Klotz, K.-N.; Linden, J. Pharmacol.
Rev. 2001, 53, 527.
2. Jacobsen, K. A.; Gao, Z. G. Nat. Rev. Drug Disc. 2006, 5, 247.