Norbornyllactone-substituted xanthines as adenosine A1 receptor antagonists

Article abstract of DOI:10.1016/j.bmc.2006.01.021

During the search for second-generation adenosine A₁ receptor antagonist alternatives to the clinical candidate 8-(3-oxa-tricyclo[3.2.1.0^2,4]oct-6-yl)-1,3-dipropyl-3,7-dihydro-purine-2,6-dione (BG9719), we developed a series of novel xanthines substituted with norbornyl-lactones that possessed high binding affinities for adenosine A₁ receptors and in vivo activity.

Full text of DOI:10.1016/j.bmc.2006.01.021

Bioorganic & Medicinal Chemistry 14 (2006) 3654–3661
Norbornyllactone-substituted xanthines as adenosine A1
receptor antagonists
William F. Kiesman,a,* Jin Zhao,a Patrick R. Conlon,a Russell C. Petter,a Xiaowei Jin,b
Glenn Smits,b Frank Lutterodt,b Gail W. Sullivanc and Joel Lindenc
aDepartment of Medicinal Chemistry, Biogen Idec, Inc., 14 Cambridge Center, Cambridge, MA 02142, USA
bDepartment of Pharmacology, Biogen Idec, Inc., 14 Cambridge Center, Cambridge, MA 02142, USA
cDepartment of Medicine (Cardiology) and Pharmacology, University of Virginia, Charlottesville, VA 22908, USA
Received 12 December 2005; revised 9 January 2006; accepted 10 January 2006
Available online 2 February 2006
Abstract—During the search for second-generation adenosine A1 receptor antagonist alternatives to the clinical candidate 8-(3-oxa-
tricyclo[3.2.1.02,4]oct-6-yl)-1,3-dipropyl-3,7-dihydro-purine-2,6-dione (BG9719), we developed a series of novel xanthines substitut-
ed with norbornyl-lactones that possessed high binding affinities for adenosine A1 receptors and in vivo activity.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
norbornyl-lactone with the xanthine ring in the exo-posi-
tion (Fig. 1).
Adenosine, the ubiquitous breakdown product of ATP,
modulates a variety of physiologic processes by its
action on four receptor subtypes: A1, A2A, A2B, and
A3.1 The A1 receptor has been identified as a mediator
in CNS, cardiovascular, metabolic, renal, and gastro-in-
testinal systems.2 Our attention has been focused upon
the development of a selective adenosine A1 receptor
antagonist for the treatment of congestive heart failure
and concomitant renal impairment. 8-(3-Oxa-tri-
cyclo[3.2.1.02,4]oct-6-yl)-1,3-dipropyl-3,7-dihydro-purine-
2,6-dione (aka BG9719)3 was found to facilitate diuresis
and stabilize the renal function in CHF patients on top
of standard therapy.4 These promising clinical data
prompted the initiation of a back-up program. Our inves-
tigations focused on the discovery of molecules with
equivalent biological activity and enhanced pharmaceuti-
cal properties including increased solubility and thermal
stability to those of the lead molecule, BG9719. One ap-
proach was to synthesize a series of norbornyl-based
compounds that place an oxygen atom in a position sim-
ilar to that of the epoxide oxygen in BG9719, such as the
2. Results
The synthesis of the xanthine derivatives started from
the commercially available racemic bicyclo[2.2.1]hept-
5-ene-2,3-dicarboxylic acid (Fig. 2).5 Iodolactonization
followed by selective reduction of the iodide via a free-
radical process gave lactone 2 in 42% yield over two
steps.6 HATU-mediated (O-(7-azabenzotriazol-1-yl)-
N,N,N0,N0-tetramethyluronium hexafluorophosphate)
coupling to diaminodialkyl uracils3 produced amides
3a–c, which after stepwise base-catalyzed cyclization to
the xanthine and then acid-catalyzed lactonization, gave
racemates 4a–c. The single enantiomer, compound 4d,
O
O
H
O
H
N
NH
NH
N
N
N
O
O
N
N
O
O
Keywords: Adenosine A1 receptor antagonist; Xanthine; BG9719; 1,3-
Dipropyl-3,7-dihydro-purine-2,6-dione.
BG9719
Figure 1. BG9719 and norbornyl lactone analog.
*
Corresponding author. Tel.: +1 617 679 2790; fax: +1 617 679
3635; e-mail: william.kiesman@biogenidec.com
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.01.021

W. F. Kiesman et al. / Bioorg. Med. Chem. 14 (2006) 3654–3661
3655
O
O
1. I2 , NaHCO3
OH
OH
H
2. AIBN, Bu3SnH
O
HO
O
O
HATU, NEt3, CH3CN,
diaminodialkyl uracil
1
2
O
R1
O
R1
O
N
N
N
N
1. KOH, i-PrOH
O
HN
O
N
H
2. HCl
N
H2N
R2
O
R2
O
O
O
4a: R1 = R2 = n-propyl
4b: R1 = n-propyl; R2 = Me
4c: R1 = 2-(2-pyridyl)ethyl;
R2 = n-propyl
3a: R1 = R2 = n-propyl
3b: R1 = n-propyl; R2 = Me
3c: R1 = 2-(2-pyridyl)ethyl;
R2 = n-propyl
4d: R1 = R2 = n-propyl
(from (S,S)-lactone)
KOH
DIBAL,
CH2Cl2, -78 ˚C
O
R1
O
R1
N
N
N
N
O
HN
O
HN
N
R2
N
R2
HO
O
H
KO
O
OH
5: R1 = R2 = n-propyl
6: R1 = R2 = n-propyl
TFA, Et3SiH
O
R1
N
N
O
HN
N
R2
O
7: R1 = R2 = n-propyl
Figure 2. Synthesis of lactone-based xanthines.
was made by the same series of steps except the enantio-
merically pure (S,S)-norbornene-5,6-dicarboxylic acid
served as the starting material.5,6
Epoxidation of 1 with m-CPBA gave the exo-epoxide,
which after treatment with HCl in dioxane produced
exo-hydroxy endo-lactone 8 (Fig. 3). HATU coupling
of the acid with 5,6-diamino-1,3-dipropyl-1H-pyrimi-
dine-2,4-dione and cyclization gave compound 10. Iod-
olactonization of 1 enabled selective esterification of
the exo-carboxylic acid followed by reductive ring open-
ing of the iodolactone intermediate gave norbornene
11.5,6 Coupling to the uracil followed by base-promoted
cyclization gave norbornene carboxylate 12. Catalytic
hydrogenation in i-PrOH produced the saturated exo-
carboxylic acid 13, while esterification followed by cata-
lytic reduction gave exo-methyl ester 14.
Compound 4a served as a starting point for further
synthetic manipulation. Hydrolysis of the lactone
with KOH gave the potassium salt of the hydroxy
acid: 5, which was stable at basic pH and also stable
to the assay conditions used to determine binding
affinities.
Selective reduction of the lactone was accomplished with
DIBAL at low temperature and gave a mixture of lactols
in a 3:1 ratio (R:S).7 Further reduction of lactol 6 to
ether 7 proceeded cleanly under the influence of trieth-
ylsilane and TFA.8
The biological activities of the antagonists were evaluat-
ed by the following procedures. The primary screen

3656
W. F. Kiesman et al. / Bioorg. Med. Chem. 14 (2006) 3654–3661
O
high human adenosine A1 receptor binding affinity.
O
1. mCPBA
HO
Compounds were incubated with radioligand and ali-
quots of crude membrane suspensions were prepared
from either rat brain cortex (for rA1) or rat brain stria-
tum (for rA2A). Values of Ki were determined from con-
centration–response relationships for each compound to
displace binding of radioligand.9b
OH
OH
H
2. HCl/dioxane
O
HO
O
O
1
8
1. I2, NaHCO3
2. H2SO4, MeOH
3. Zn, AcOH
HATU, NEt3, CH3CN,
diaminodipropyl uracil
O
N
O
N
Compounds with sufficient hA1 and rA1 binding affinity
were then tested in a rapid oral activity rat diuresis
experiment. The antagonist was dosed in a 0.5% car-
boxymethylcellulose suspension orally. Urine volumes
were collected over time and urine sodium and potassi-
um concentrations were measured. Figures 4 and 5 dis-
play the results of the rat oral activity screen at 2 and
0.3 mg/kg, respectively.11
O
HO
O
N
H
OMe
H
H2N
O
O
HO
O
9
11
1. HATU, NEt3, CH3CN,
diaminodipropyl uracil
2. KOH, i-PrOH
1. KOH, i-PrOH
2. HCl/dioxane
The EC50 values of selected antagonists were also deter-
mined in isolated rat atria in order to rank the in vivo
activities.12 All experiments were conducted in accor-
dance with the NIH Guide for the Care and Use of Lab-
oratory Animals and the protocols were approved by
the Institutional Animal Care and Use Committee.
O
N
O
N
O
HN
OH
H
HO
N
N
NH
O
O
O
N
N
O
10
3. Discussion
12
1. H2SO4, MeOH
2. 5% Pd/C, H2, i-PrOH
The binding activity of the first member of the lactone
series, compound 4a, was 18 nM with 37-fold selectivity
for A1 versus A2A receptors. This result was somewhat
surprising to us in view of the relatively low A1 affinity
of the closely related trans-substituted norbornyl xan-
thine 13, where the xanthine occupied the endo-position
and the methylcarboxylate the exo-position. Evidently
locking the carboxylate into the compact lactone ring
system allowed a better fit into both the A1 and A2A
receptors as affinities for both receptors increased over
4-fold. Serendipity aside, this result spurred us on to
examine more closely a few readily made examples of
compounds containing a norbornyl framework with
exo-xanthine substitution. As in the case of BG9719,
the chirality of the [2.2.1] system had some effect on both
the binding affinity and selectivity. The single enantio-
mer, lactone 4d, exhibited a 3-fold increase in binding
affinity for the A1 receptor with a marginal increase in
affinity for the A2A receptor. Substitution of the n-pro-
pyl chain in the N1 position with a 2-(2-pyridyl)ethyl
group (4c) or truncation of the N3 propyl chain to a
methyl group (4b) decreased A1 and A2A affinity equally.
Placement of a polar carboxylic acid in either the endo-
(5) or exo-positions (12,13) led to poor A1 binding affin-
ity. Even introduction of the less polar hydroxyl group
in the 2-position (10) decreased A1 affinity about 3-fold
and A2A affinity by a factor of 6. This result suggested
that the pocket the norbornyl group occupies in both
the A1 and A2A receptors is particularly sensitive to po-
lar substitution.13 Reduction of the lactone to a mixture
of lactols (6) compromised A1 activity, while subsequent
reduction to the ether (7) restored A1 binding and selec-
tivity to levels almost identical to those for lactone 4a.
This would indicate that although there is space for
the carbonyl in lactone 4a and the CH2 group in ether
7 the oxygen of the lactone 4a does not contribute to
5% Pd/C,
H2, i-PrOH
O
O
OMe
OH
H
H
NH
N
NH
N
O
N
O
N
N
N
O
O
13
14
Figure 3. Hydroxylactone- and exo-substituted-norbornyl xanthine
preparations.
consisted of a single-point assay performed in duplicate
on membranes derived from stably transfected HEK
(hA2A, hA2B, and hA3) or CHO-K1 (hA1 receptors) cells
expressing each of the four human adenosine receptor
subtypes (hA1, hA2A, hA2B, and hA3).9a Tubes were
incubated at room temperature for 2 h, filtered, and
counted in a c-counter. The readout gave the percent
(%) of specific radioligand binding; a high number de-
notes poor antagonist activity, while a low number indi-
cates displacement of the radiolabel from the receptor.
Compounds that displayed good A1 binding activity in
the single point assays were further evaluated to deter-
mine IC50 values. Kis were calculated from IC50 values.10
Full binding curves used antagonist concentrations from
10ꢀ11 to 10ꢀ5 M assayed in duplicate. Table 1 lists the
binding data for all of the compounds tested against
all four human adenosine receptor subtypes.
The binding affinities for rat A1 (rA1) and A2A (rA2A
receptors were determined for compounds that exhibited
)

W. F. Kiesman et al. / Bioorg. Med. Chem. 14 (2006) 3654–3661
3657
Table 1. Binding affinities for lactone-based xanthine derivatives on recombinant human adenosine receptors
Compoundc
Kia (nM) or % of specific radioligand bindingb
hA2A/hA1
hA1
hA2A
hA2B
hA3
BG9719
12
0.7 (rat)d
18
3.0 (rat)e
71
66
1660
1250 (rat)d
657
264 (rat)e
2330
3510
611
—
4810
—
138
1800 (rat)d
37
4a
802
—
(88%)
—
105 (rat)
33
53
4b
4c
4d
(69%)
(82%)
1540
—
(93%)
(93%)
>10,000
—
6.0
0.3 (rat)e
107
923
269 (rat)e
4270
154
860 (rat)
40
5
6
7
(70%)
(81%)
(40%)
—
(89%)
(95%)
(89%)
—
(38%)
20
1.0 (rat)e
62
13 (rat)e
508
144
(100%)
1510
>10,000 (rat)e
6280
>10,000 (rat)e
—
(94%)
—
76
>10,000 (rat)
101
>800 (rat)
10
2030
—
>10,000
—
—
12
13
14
—
(76%)
1080
—
—
42
(94%)
(76%)
74
3100
a Ki values are means of closely agreeing duplicate determinations (means 50%), each derived from seven antagonist concentrations assayed in
duplicate.
b The readout gave the percent (%) of specific radioligand binding; a high number denotes poor antagonist activity, while a low number indicates
displacement of the radiolabel from the receptor.
c All compounds shown are racemic mixtures, except 4d, which was prepared from the enantiomerically pure carboxylic acid 2.
d Rat receptor binding values from Ref. 3.
e Ki values were determined from concentration–response relationships for each compound to displace binding of radioligand to rat brain cortex
(for rA1) or rat brain striatum (for rA2A).
200.0
180.0
160.0
140.0
binding affinity. In contrast, no trend was observed
between the affinities for the rA2A and hA2A receptors,
except that the difference between the binding affinities
for the species was smaller.
120.0
100.0
80.0
60.0
40.0
20.0
0.0
The in vivo activities of the most potent binders were
then investigated. A fixed oral dosage of 2 mg/kg of
the antagonist was administered to rats in metabolic
cages. Urine output was monitored for 4 h and analyses
were performed to determine the amounts of excreted
sodium and potassium in the urine.
vehicle BG9719
4a
4d
Figure 4. Rat oral efficacy screen; measurement of urinary sodium
excretion per hour, over a 4 h period, caused by a 2 mg/kg oral dose of
compound in a 0.5% CMC suspension (see Ref. 11 for details).
After oral dosing, there were significant increases in
sodium excretion compared to vehicle in the rats receiv-
ing compounds 4a and 4d (Fig. 4). At this dose, the bio-
logical activity of the racemate was equivalent to that of
the single enantiomer. The magnitude of sodium mobi-
lization was also similar in both cases to that seen with
an equivalent dose of BG9719. Potassium excretion was
also similar between vehicle-treated animals and those
treated with all three compounds. This result was not
unexpected as A1-mediated natriuresis is potassium neu-
tral. Plasma concentrations of the compounds were also
taken at the 1-h time point. The lactones 4a and 4d had
plasma levels of 1600 and 2300 ng/mL compared to only
110 ng/mL for BG9719. This result was an early indica-
tion that the half-life of the lactone compounds in the
rat may be longer than that of BG9719 and that the par-
ent lactones could also have greater in vivo stability.
70
60
50
40
30
20
10
0
Vehicle BG9719
7
10
Figure 5. Rat oral efficacy screen; measurement of urinary excretion of
sodium per hour, in a 4-h period, of a 0.3 mg/kg oral dose of
compound in a 0.5% CMC suspension (see Ref. 11 for details).
the A1 binding affinity or the overall selectivity versus
the A2A receptor.
Compounds 4a, 4d, 7, and 10 were evaluated for rA1 and
rA2A receptor binding affinity. For A1 receptors, each of
the compounds tested in the rat receptor assays had
greater affinity for the rat versus the human receptors.
About a 6- to 20-fold improvement was seen in rA1
A similar but more challenging in vivo rat experiment
was performed that measured the urinary sodium excre-
tion per hour after oral administration of a lower
dose of 0.3 mg/kg of the antagonist. The results of this

3658
W. F. Kiesman et al. / Bioorg. Med. Chem. 14 (2006) 3654–3661
experiment comparing BG9719 with compounds 7 and
10 appear in Figure 5. The sodium excretion induced
by ether 7 was 37% greater than that of the BG9719
standard, while the hydroxylactone 10 result was similar
to vehicle.
5. Experimental
5.1. General methods
Unless otherwise stated, reactions were carried out un-
der nitrogen in oven-dried glassware. The HPLC meth-
od used to determine purity was performed on an
HP1100 system, YMC-ODS-AM C18 reversed-phase
column (4.6 · 100 mm)-guard column YMC-ODS-AM
S-5 120A (direct connect); 20–100% CH3CN/H2O gradi-
ent over 8 min, buffered with 0.1% TFA at 1.5 mL/min
flow rate, detector set at dual wavelength 214 and
254 nm. Reversed-phase HPLC was also used for pre-
Further work explored the functional activities of the
adenosine A1 receptor antagonists in two different iso-
lated rat atria assays. In the first assay, isolated spon-
taneously beating atria were treated with BG9719 and
compound 4a at 30 nM. Atria were then treated with
increasing concentration of N6-cyclopentyladenosine
(CPA) until the heart rate dropped to a minimum.
The concentration response curves to both compounds
showed a rightward shift with EC50 values of 127 nM
for BG9719 and 865 nM for compound 4a. In the sec-
ond assay, isolated rat atria were treated with a solu-
tion of CPA in high enough concentration to reduce
beating rate by 75%. The atria were then treated with
increasing concentrations of compounds BG9719 and
compound 4d, and their effects on heart rate were
determined. Both antagonists blocked the CPA-in-
duced reduction in heart rate with EC50 values of
7.9 nM for BG9719 and 71 nM for compound 4d. De-
spite the similarities in binding affinities for the three
molecules, BG9719 demonstrated superior in vivo
activity in the atria over both lactones.
1
parative purposes (LiChroprep C-18, 310 · 25 mm). H
and 13C NMR spectra were obtained using Bruker
300, 400, and 600 MHz NMR spectrometers. High-res-
olution mass spectroscopic data were obtained on a
THERMO ELECTRON LTQ FTMS. The data were
acquired at the positive, FULL scan and SIM scan FT
MS Mode, protonated molecular ion designated as
MH+. All the chemicals were supplied by Aldrich Chem-
ical Co., Inc., Milwaukee, WI, USA, unless otherwise
indicated.
5.2. 5-Oxo-4-oxa-tricyclo[4.2.1.03,7]nonane-9-carboxylic
acid (2)
1H NMR (300 MHz, CDCl3) d 4.76 (dd, 1H), 3.18 (s,
1H), 3.02 (s, 1H), 2.78 (s, 1H), 2.73 (s, 1H), 1.83–1.78
(m, 1H), 1.75 (d, 1H), 1.56 (m, 1H). 13C NMR
(600 MHz, CDCl3) d (ppm) 179.48, 177.08, 80.29,
49.89, 45.89, 42.43, 40.83, 37.89, 35.76. HRMS
m/z = 183.06531 (MH+), calcd 183.06519.
Since the product was targeted for delivery to acutely
decompensated congestive heart failure patients in the
hospital setting, the solubility of 4a in comparison to
BG9719 in prototype solutions suitable for IV adminis-
tration was determined. Both compounds exhibited
uniformly poor water solubility: BG9719, 0.18 mg/mL
in water; 0.16 mg/mL in 0.9% saline; 0.17 mg/mL
in D5W: 4a, 0.15 mg/mL in water; 0.09 mg/mL in 0.9%
saline; 0.09 mg/mL in D5W.
5.3. 8-(5-Oxo-4-oxa-tricyclo[4.2.1.03,7]non-9-yl)-1,3-
dipropyl-3,7-dihydro-purine-2,6-dione (4a)
To a stirred mixture of 0.500 g (2.74 mmol) of 5-oxo-4-
oxa-tricyclo[4.2.1.03,7]nonane-9-carboxylic acid, 0.748 g
(2.85 mmol) of 5,6-diamino-1,3-dipropyl-1H-pyrimi-
dine-2,4-dione hydrochloride,3 1.15 mL (8.22 mmol) of
NEt3, and 10 mL anhydrous acetonitrile was added
1.08 g (2.85 mmol) of HATU.15 The reaction solution
was stirred at rt for 1 h. The reaction mixture was concd
in vacuo and combined with 30 mL EtOAc and 30 mL
of 10% citric acid. The aqueous layer was separated
and washed twice with 40-mL portions of EtOAc. The
combined organic fractions were washed with 20-mL
portions of satd NaHCO3 and brine, and concd in vac-
uo. The resultant solid was combined, in a 50-mL
round-bottomed flask equipped with a condenser, with
a mixture of 5 mL of i-PrOH and 5.5 mL of 1 N KOH
(5.5 mmol) and heated to reflux. After heating for 1 h,
the reaction solution was concd in vacuo, taken up in
40 mL of water, and washed twice with 30-mL portions
of CH2Cl2. The aqueous layer was acidified with concd
HCl and the resultant precipitate was collected by suc-
tion filtration to give 0.155 g (15% yield) of an off-white
The lack of water solubility coupled with the discovery
of a more promising series of antagonists with increased
water solubility (>100-fold increase) and better in vivo
activity14 led to the discontinuation of the investigation
of this series of molecules and precluded the examina-
tion of the single enantiomers of ether 7, but it would
be reasonable to expect that the more active enantiomer
would be related to the more active lactone 4d and the
in vivo activity had the potential to be greater than that
of BG9719.
4. Conclusions
A series of norbornyl-substituted lactones were inves-
tigated as replacements for the epoxynorbornane in
the clinical compound BG9719. These xanthine deriv-
atives in which the xanthine occupies the exo position
on the norbornyl ring system showed high A1 bind-
ing affinity and selectivity over the closely related
A2A receptor (notably compounds 4d: hA1 Ki = 6 nM;
154-fold selectivity and 7: hA1 Ki = 20 nM; 76-fold
selectivity). The lactones possessed similar if not bet-
ter in vivo activity to BG9719 in the rat diuresis
models.
1
solid. H NMR (300 MHz, CDCl3) d 4.90 (t, 1H), 4.10
(t, 2H), 4.00 (t, 2H), 3.48 (s, 1H), 3.38 (m, 2H), 2.80
(m, 1H), 2.10 (m, 1H), 1.95 (m, 1H), 1.85–1.65 (m,
5H), 1.55 (d, 1H), 0.95 (two t, 6H). 13C NMR
(600 MHz, CDCl3) d (ppm) 179.41, 156.11, 154.98,

W. F. Kiesman et al. / Bioorg. Med. Chem. 14 (2006) 3654–3661
3659
150.77, 148.70, 107.25, 80.27, 46.10, 45.59, 45.41, 43.78,
43.45, 43.44, 38.52, 35.43, 21.35, 21.33, 11.21, 11.06.
HRMS m/z = 373.18721 (MH+), calcd 373.18703.
HPLC retention time: 4.25 min, without YMC-ODS-
AM S-5 120A (direct connect guard column).
dried with Na2SO4 and concd in vacuo to give 0.201 g
(65% yield) of an off-white solid. H NMR (300 MHz,
1
CDCl3) d 5.40 (s, 1H), 5.15 (m, 1H), 4.55 (m, 1H),
4.10–3.90 (m, 4H), 3.15–2.70 (m, 3H), 2.50–2.40 (m,
1H), 1.90 (m, 1H), 1.80–1.50 (m, 5H), 1.35 (m, 2H),
0.90 (two t, 6H). HRMS m/z = 375.20272 (MH+), calcd
375.20268. HPLC retention time: 3.60 min.
The following compounds were made in an analogous
manner.
5.9. 8-(4-Oxa-tricyclo[4.2.1.03,7]non-9-yl)-1,3-dipropyl-
3,7-dihydro-purine-2,6-dione (7)
5.4. 3-Methyl-8-(5-oxo-4-oxa-tricyclo[4.2.1.03,7]non-9-
yl)-1-propyl-3,7-dihydro-purine-2,6-dione (4b)
Dissolved 6 (0.201 g, 0.537 mmol) in 2 mL TFA and
added Et3SiH (0.215 mL, 1.34 mmol) dropwise while
stirring. Stirred at rt for 1 h, concd in vacuo, and puri-
fied by prep HPLC isocratic method (40% CH3CN/
60% water at 15 mL/min, 40 min) retention time
22.44 min. Obtained 0.049 g (25% yield) of an off-white
solid. 1H NMR (300 MHz, CDCl3) d 4.40 (dt, 2H), 4.05
(t, 2H), 3.95 (t, 2H), 3.85–3.75 (m, 2H), 3.07 (q, 1H),
2.70 (m, 2H), 2.40 (s, 1H), 1.90 (d, 1H), 1.80–1.60 (m,
5H), 1.43 (d, 2H), 0.90 (two t, 6H). 13C NMR
(600 MHz, CDCl3) d (ppm) 157.88, 157.28, 155.66,
150.69, 106.68, 78.87, 73.18, 49.46, 46.42, 45.77, 44.70,
43.45, 41.06, 40.86, 35.83, 21.43, 21.31, 11.41, 11.14.
HRMS m/z = 359.20786 (MH+), calcd 359.20777.
HPLC retention time: 4.46 min.
0.058 g (15% yield) of an off-white solid. 1H NMR
(300 MHz, CDCl3) d 4.85 (m, 1H), 4.10 (t, 2H), 3.55
(s, 3H), 3.48 (s, 1H), 3.38 (m, 2H), 2.70 (m, 1H), 2.10
(m, 1H), 1.95 (m, 1H), 1.90–1.80 (m, 1H), 1.70–1.45
(m, 5H), 0.90 (t, 3H). m/z = 345.10 (MH+). HPLC reten-
tion time: 3.18 min.
5.5. 8-(5-Oxo-4-oxa-tricyclo[4.2.1.03,7]non-9-yl)-3-pro-
pyl-1-(2-pyridin-2-yl-ethyl)-3,7-dihydro-purine-2,6-dione
(4c)
0.010 g (10% yield) of a yellow oil. m/z = 435.96 (MH+).
HPLC retention time: 3.47 min.
5.6. 8-(5-Oxo-4-oxa-tricyclo[4.2.1.03,7 ]non-9-yl)-1,3-
dipropyl-3,7-dihydro-purine-2,6-dione (4d)
5.10. 2-Hydroxy-5-oxo-4-oxa-tricyclo[4.2.1.03,7]nonane-
9-carboxylic acid (8)16 8-(2-Hydroxy-5-oxo-4-oxa-tricy-
clo[4.2.1.03,7]non-9-yl)-1,3-dipropyl-3,7-dihydro-purine-
2,6-dione (10)
Starting from the (S,S)-norbornene-5,6-dicarboxylic
acid, 0.177 g (35% yield) of an off-white solid. 1H
NMR (300 MHz, CDCl3) d 4.90 (t, 1H), 4.10 (t, 2H),
4.00 (t, 2H), 3.48 (s, 1H), 3.38 (m, 2H), 2.80 (m, 1H),
2.10 (m, 1H), 1.95 (m, 1H), 1.85–1.65 (m, 5H), 1.55 (d,
1H), 0.95 (two t, 6H). HRMS m/z = 373.18714 (MH+),
calcd 373.18703. HPLC retention time: 4.37 min.
1
Off-white solid H NMR (400 MHz, MeOH-d4) d 4.40
(s, 1H), 3.90 (t, 2H), 3.85 (t, 2H), 3.65 (s, 1H), 3.15 (s,
1H), 3.05 (s, 1H), 2.70 (s, 1H), 1.90 (m, 2H), 1.70 (q,
2H), 1.60 (q, 2H), 0.85 (two t, 6H). HRMS
m/z = 389.18207 (MH+), calcd 389.18195. HPLC reten-
tion time: 3.72 min.
5.7. Potassium-3-(2,6-dioxo-1,3-dipropyl-2,3,6,7-tetrahy-
dro-1H-purin-8-yl)-6-hydroxy-bicyclo[2.2.1]heptane-2-
carboxylate (5)
5.11. 3-(2,6-Dioxo-1,3-dipropyl-2,3,6,7-tetrahydro-1H-
purin-8-yl)-bicyclo[2.2. 1]hept-5-ene-2-carboxylic acid
(12)
Combined 0.113 g of 4a (0.303 mmol) and 0.020 g of
KOH (0.303 mmol) in 2 mL of water and heated to
50 ꢁC for 24 h. Added 2 mL of CH2Cl2 and stirred vig-
orously for 2 h. Separated layers and concd water layer
To a stirred mixture of 2.50 g (12.7 mmol) of 11, 3.33 g
(12.7 mmol) of 5,6-diamino-1,3-dipropyl-1H-pyrimi-
dine-2,4-dione hydrochloride,3 8.8 mL (50.8 mmol) of
(i-Pr)2NEt, and 30 mL anhydrous DMF was added
4.83 g (12.7 mmol) of HATU.15 The reaction solution
was stirred at rt for overnight. The reaction mixture
was concd in vacuo to a volume of 10 mL and combined
with 150 mL EtOAc. The resultant solution was washed
three times with 20 mL of 1 N HCl and twice with satd
brine. The organic fraction was concd in vacuo and the
resultant solid was combined, in a 250-mL round-bot-
tomed flask equipped with a condenser, with a mixture
of 60 mL of i-PrOH, 4.20 g KOH, and 35 mL of water
and heated to reflux. After heating for 3 h, the reaction
was concd to a 30 mL volume and acidified with 1 N
HCl. The mixture was extracted with three 25-mL por-
tions of EtOAc, dried over Na2SO4, and concd in vacuo
to give a solid which after reprecipitation from a NaOH
solution with 1 N HCl gave 2.72 g (58% yield) of an off-
white solid. 1H NMR (300 MHz, CDCl3) d 9.35 (s, 1H),
1
in vacuo to give quant. yield of an off-white solid. H
NMR (300 MHz, DMSO-d6) d 4.10 (m, 1H), 3.95 (t,
2H), 3.85 (t, 2H), 2.90 (d, 1H), 2.40 (m, 2H), 2.10 (dt,
1H), 1.70 (q, 2H), 1.55 (m, 3H), 1.20–1.00 (m, 2H),
0.95 (two t, 6H). HRMS m/z = 391.19770 (MH+), calcd
391.19760. HPLC retention time: 3.54 min.
5.8. 8-(5-Hydroxy-4-oxa-tricyclo[4.2.1.03,7]non-9-yl)-1,3-
dipropyl-3,7-dihydro-purine-2,6-dione (6)
To
a
chilled solution (ꢀ78 ꢁC) of 4a (0.200 g,
0.537 mmol) in 3 mL of CH2Cl2 was added 1.07 mL
(1.07 mmol) of a 1.0 M solution of DIBAL in CH2Cl2.
The reaction mixture was stirred at ꢀ78 ꢁC for 2 h and
quenched by the addition of 3 mL of satd NaHCO3.
The reaction mixture was allowed to warm to rt, diluted
with 4 mL of water and extracted with three 5-mL por-
tions of CH2Cl2. The combined organic fractions were

3660
W. F. Kiesman et al. / Bioorg. Med. Chem. 14 (2006) 3654–3661
J.; Suzuki, F.; Nonaka, H.; Ishii, A. J. Med. Chem. 1992,
35, 924–930.
6.30 (m, 1H), 6.20 (m, 1H), 4.15–3.95 (m, 3H), 3.39 (s,
1H), 3.33 (s, 1H), 2.70 (d, 1H), 1.85–1.40 (m, 7H),
0.95 (m, 6H). m/z = 373.09 (MH+). HPLC retention
time: 3.97 min.
3. Profiled as CVT-124 (a.k.a.1,3-dipropyl-8-[2-(5,6-ep-
oxy)norbornyl]xanthine) in Pfister, J. R.; Belardinelli, L.;
Lee, G.; Lum, R. T.; Milner, P.; Stanley, W. C.; Linden,
J.; Baker, S. P.; Schreiner, G. J. Med. Chem. 1997, 40,
1773–1778.
5.12. 3-(2,6-Dioxo-1,3-dipropyl-2,3,6,7-tetrahydro-1H-
purin-8-yl)- bicyclo[2.2.1]heptane-2-carboxylic acid (13)
4. Wolff, A. A.; Skettino, S. L.; Beckman, E.; Belardinelli, L.
Drug Dev. Res. 1998, 45, 166–171.
1H NMR (300 MHz, CDCl3) d 4.10 (dd, 2H), 4.00 (dd,
2H), 3.75 (t, 1H), 3.05 (d, 1H), 2.78 (m, 2H), 1.88–1.60
(m, 5H), 1.60–1.38 (m, 4H), 1.25 (d, 3H), 0.95 (two t,
6H). HRMS m/z = 375.20274 (MH+), calcd 375.20268.
HPLC retention time: 4.35 min.
5. Kiesman, W. F.; Petter, R. C. Tetrahedron: Asymmetry
2002, 13, 957, describes methods for the asymmetric
synthesis of the (R,R)-norbornene-5,6-dicarboxylic acid.
6. (a) Janssen, A. J. M.; Klunder, A. J. H.; Zwanenburg, B.
Tetrahedron 1991, 47, 5513; (b) Hamanaka, N.; Seko, T.;
Miyazaki, T.; Naka, M. Tetrahedron Lett. 1989, 9, 2399.
7. (a) Winterfeldt, E. Synthesis 1975, 7, 617; (b) Takacs, J.
M.; Helle, M. A.; Seely, F. L. Tetrahedron Lett. 1986, 27,
1257.
5.13. 3 -(2,6-Dioxo-1,3-dipropyl-2,3,6,7-tetrahydro-1H-
purin-8-yl)-bicyclo[2.2. 1]heptane-2-carboxylic acid
methyl ester (14)
8. Kraus, G. A.; Molina, M. T.; Walling, J. A. J. S. Chem.
Commun. 1986, 1568.
Dissolved 12 (0.050 g) in 5.0 mL of anhydrous MeOH
and added 3 drops concd H2SO4. The solution was stir-
red at rt for 24 h. The reaction mixture was concd in
vacuo and the resultant oil was dissolved in 10 mL
EtOAc. The organic layer was washed with 10 mL of
satd NaHCO3 and then 10 mL of brine, dried over
Na2SO4, and concd to dryness. The solid product was
then dissolved in 6.0 mL of degassed i-PrOH and com-
bined with 8 mg of 5% Pd/C. The reaction mixture
was flushed 3 times with hydrogen and stirred for 20 h
under a balloon of hydrogen at rT. The reaction mixture
was then filtered through a plug of silica and concd in
9. (a) For initial screening a solution of the antagonist
(1 lM) was incubated with membranes in 50 mM HEPES,
pH 7.4, 1 mM EDTA, 5 mM MgCl2, and 1 U/mL aden-
osine deaminase. DMSO was included in all assays
excepting A3 at a final concentration of 5%. Radioligands
consisted of: A1, 0.3 nM 125I-aminobenzyladenosine (125I-
ABA); A2A, 0.7 nM 125I-ZM241385; A2B, 0.5 nM 125I-3-
(4-aminobenzyl)-8-phenyloxyacetate-1-propyl-xanthine;
and A3, 0.6 nM 125I-ABA. Nonspecific binding was
measured in the presence of 50 lM xanthine amino
congener or 10 lM BW-1433 (A3).; (b) Compounds were
incubated at room temperature for 90 min with radioli-
gand (2 nM 3H-CPX for A1; 0.5–1.2 nM 3H-ZM241385
for A2A), 50 mM Tris–HCl buffer (pH 7.4), adenosine
deaminase (2 U/mL), and 100-lL aliquots of crude mem-
brane suspensions (10–20 lg protein) prepared from either
rat brain cortex (for A1) or rat brain striatum (for A2A).
Incubations were terminated by addition of ice-cold
50 mM Tris–HCl buffer and collection of membranes
onto Whatman GF/C glass fiber filters by vacuum
filtration. Membrane-bound radioactivity was quantified
by liquid scintillation counting. Values of Ki were deter-
mined from concentration–response relationships for each
compound to displace binding of radioligand, using
GraphPad Prism (GraphPad, San Diego, CA).
1
vacuo to give 0.048 g (>99% yield) of a white solid. H
NMR (300 MHz, CDCl3) d 8.65 (s, 1H), 4.14 (dd,
2H), 4.04 (dd, 2H), 3.77 (t, 1H), 3.70 (s, 3H), 3.30 (d,
1H), 2.82 (m, 1H), 2.67 (m, 1H), 1.90–1.58 (m, 5H),
1.55–1.28 (m, 4H), 1.22 (d, 1H), 0.95 (two t, 6H). 13C
NMR (600 MHz, CDCl3) d (ppm) 166.01, 154.54,
153.89, 150.28, 150.17, 105.71, 51.18, 45.51, 45.04,
44.14, 43.09, 42.08, 41.64, 40.82, 37.40, 27.76, 20.44,
20.27, 10.32, 10.18. HRMS m/z = 389.21845 (MH+),
calcd 389.21833. HPLC retention time: 5.49 min.
10. Linden, J. J. Cyclic Nucleotide Res. 1982, 8, 163–172.
11. Rat oral efficacy screen: the rats were placed into
metabolic cages and dosed by gavage with various doses
of antagonist. The doses and group sizes were: vehicle
(0.5% carboxymethylcellulose [CMC]) (n = 3); antagonist:
2.0 mg/kg for 4a and 4d and 0.3 mg/kg for 7 and 10
(n = 3). Urine was collected for 4 h after dosing. Urine
volume was measured gravimetrically, and sodium (Na)
and potassium (K) concentrations were determined by
flame photometry. Urine flow (UV), sodium excretion
(UNaV), and potassium excretion (UKV) were calculated
and are shown as units per hour, as an average for the 4-h
collection period.
Acknowledgment
We thank Dr. John Shryock (UFla) for determining rat
receptor affinities for selected compounds, Dr. Herman
van Vlijmen for molecular modeling advice, and
Dr Maria Pellegrini and Dr Xiaoping Hronowski for
aid in structural characterization.
References and notes
12. Isolation of atria from rat heart. Hearts were removed from
the rats, placed in petri dishes containing Krebs–Henseleit
(Krebs) buffer pre-warmed to 37 ꢁC, and bubbled with
95% O2/5% CO2. The composition of Krebs buffer was
118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO4, 25 mM
NaHCO3, 1.2 mM KH2PO4, 2.5 mM CaCl2, and 11 mM
glucose, pH 7.4. The right atrium was dissected and
cleaned of surrounding myocardial and vascular tissue.
Two lengths of thread were attached at opposite ends of
the atrium. One thread anchored the tissue to a glass rod
and the other was connected to an isometric force
1. Fredholm, B. B.; IJzerman, A. P.; Jacobson, K. A.; Klotz,
K. N.; Linden, J. Pharmacol. Rev. 2001, 53, 527–552.
2. Belardinelli, L.; Shryock, J. C.; Zhang, Y.; Scammells, P.
J.; Olsson, R.; Dennis, D.; Milner, P.; Pfister, J.; Baker, S.
P. J. Pharmacol. Exp. Ther. 1995, 275, 1167–1176; Dhalla,
A. K.; Shryock, J. C.; Shreeniwas, R.; Belardinelli, L.
Curr. Top. Med. Chem. 2003, 3, 369–385; For reviews of
A1 SAR see: Muller, C. E.; Stein, B. Curr. Pharm. Des.
¨
1996, 2, 501–530; Jacobson, K. A.; van Galen, P. J. M.;
Williams, M. J. Med. Chem. 1992, 35, 407–422; Shimada,

W. F. Kiesman et al. / Bioorg. Med. Chem. 14 (2006) 3654–3661
3661
transducer. The tissue was suspended in a water-jacketed
reservoir warmed to 37 ꢁC and bubbled with 95% O2/5%
CO2. A pre-load tension of 2 g was applied using a pre-
calibrated Gould recorder. Hung tissue was washed with
warm, oxygenated Krebs buffer, while maintaining 2 g of
tension. Baseline heart rate was measured on Ponemah
software from Gould Instruments (Valley View, Ohio).
Determination of EC50 of A1 antagonists using the CPA
dose-reversal paradigm. Isoproterenol (30 nM) was added
to all baths containing atria to increase the baseline heart
rate to between 350 and 400 beats per minute (bpm).
Following rate stabilization, 130 nM CPA was added to
baths to cause a 75% reduction in atrial beating rate
(control 0). Increasing concentrations of antagonist were
then added to the baths until the rate was restored to
maximum, and the effective concentration at which 50%
response was obtained (EC50) was determined. Five atria
were used in this experiment.
13. Unpublished results the subject of
publication.
a forthcoming
14. Petter, Russell C.; Kiesman, William F.; Conlon, Patrick
R.; Kumaravel, Gnanasambandam; Ensinger, Carol L.;
Dowling, James; Peng, Bo; Smits, Glenn; Jin, Xiaowei;
Lutterodt, Frank A.; Fu, Kai; LePage, Doreen; Jayaraj,
Andrew; Gill, Alan; Costa, Don; Wortham, Kathy;
Porter, Kaela; Linden, Joel; Sullivan, Gail. Abstracts of
Papers, 224th National Meeting of the American Chem-
ical Society, Boston, MA, American Chemical Society:
Washington, DC, 2002; MEDI-417.
15. HATU = O-(7-azabenzotriazol-1-yl)-N,N,N0,N0-tetramethyl-
uronium hexafluorophosphate.
16. Beilstein Registry # 73036-39-2.