Synthesis and pharmacological characterization of functionalized 2-pyridones structurally related to the cardiotonic agent milrinone

Article abstract of DOI:10.1016/S0968-0896(03)00528-5

A new class of cardiotonic agents characterized by a 2-pyridone structure was synthesized. Appropriate sym-2-dimethylaminomethylene-1,3-diones reacted with methylcyanoacetate to afford the desired compounds. These derivatives were evaluated for their ability in inducing cardiotonic response on guinea pig isolated myocardial preparations. Compound 8b increased atrial contractility to an extent which is significantly higher than that of milrinone, the parent drug used as a reference compound. The pharmacological characterization and the docking studies performed on 8b highlighted its selective mechanism of action via type 3 PDE (PDE3) inhibition.

Full text of DOI:10.1016/S0968-0896(03)00528-5

Bioorganic & Medicinal Chemistry 11 (2003) 4749–4759
Synthesis and Pharmacological Characterization of
Functionalized 2-Pyridones Structurally Related to the
Cardiotonic Agent Milrinone
Paola Fossa,a Giulia Menozzi,a Paola Dorigo,b Maura Floreanib and Luisa Mostia,*
aDepartment of Pharmaceutical Sciences, University of Genova, Viale Benedetto XV 3, 16132 Genova, Italy
bDepartment of Pharmacology and Anaesthesiology, Pharmacology Section, University of Padova,
Largo Meneghetti 2, 25131 Padova, Italy
Received 4 June 2003; accepted 29 July 2003
Abstract—A new class of cardiotonic agents characterized by a 2-pyridone structure was synthesized. Appropriate sym-2-
dimethylaminomethylene-1,3-diones reacted with methylcyanoacetate to afford the desired compounds. These derivatives were
evaluated for their ability in inducing cardiotonic response on guinea pig isolated myocardial preparations. Compound 8b increased
atrial contractility to an extent which is significantly higher than that of milrinone, the parent drug used as a reference compound.
The pharmacological characterization and the docking studies performed on 8b highlighted its selective mechanism of action via
type 3 PDE (PDE3) inhibition.
# 2003 Elsevier Ltd. All rights reserved.
Introduction
its ability in mimicking the structural and electronic
features of cAMP at the active site of the enzyme, thus
inhibiting conversion of cAMP to 50-AMP.5 However,
some evidence suggests that these drugs can cause ino-
tropic action also through antagonism towards endo-
genous adenosine at the cardiac A1 receptor.6 Until
now, a positive inotropic agent with exclusive high
affinity for PDE3 or for A1 receptor has not yet been
synthesized.
Congestive heart failure (CHF) is a major cause of
death in patients with heart disease. For many years,
digitalis glycosides have been used for the treatment of
CHF. This use, however, is limited because of their
narrow therapeutic index and their propensity to cause
life-threatening arrhythmias (arrhythmogenic lability).
The search for orally active ‘non-glycoside’ cardiotonic
drugs displaying a greater safety profile and improved
efficacy on patient survival resulted in a new class of
cardiotonics endowed with a different mechanism of
action. Amrinone and milrinone1 are the prototypes of
a series of analogues, proposed as a replacement for
digitalis, referred to as inhibitors of cGMP-inhibited
cAMP phosphodiesterase (PDE), also called Type 3
PDE or PDE3.2 In particular, milrinone, which has
been approved by the United States Food and Drug
Administration and is used in the clinical treatment of
patients with severe heart failure,3 improves the hemo-
dynamic status of heart failure via inotropic/vasodila-
tory effects attributable to the increase in intracellular
cAMP level.4 This mechanism of action seems related to
In previous studies, we used milrinone as the template
for the synthesis of a number of 6-substituted 5-acyl-
1,2-dihydro-2-oxo-3-pyridinecarbonitriles 17 and related
derivatives in which the cyano groupwas substituted
with an acetyl function 2.8
*Corresponding author. Tel.: +39-010-353-8350; fax: +39-010-353-
8358; e-mail: mosti@unige.it
0968-0896/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00528-5

4750
P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
Among these derivatives, ketone 1a9 is the most inter-
esting inducing a significant increase in contractile force
of both guinea pig spontaneously beating atria and
electrically driven left atria, comparable to that of mil-
rinone, particularly at higher concentrations, while the
3-acetyl analogues 2 showed poor cardiac activity.8
The cardiac effect of all new derivatives has been inves-
tigated on guinea pig isolated myocardial preparations,
using milrinone as the reference compound. In the case
of 8b, which revealed itself as the most active inotropic
agent, the biochemical mechanism responsible for its
cardiac action has been also examined.
The mechanism of action of 1a refers to a competitive
inhibition of the PDE3, even if an antagonism towards
endogenous adenosine also seems to be involved.10
In addition, in order to give further proof to the
mechanism of action for 8b and 4b, molecular models of
the complex enzyme-inhibitor were generated for 8b, 4b,
cilostamide,13 a highly potent PDE3 inhibitor and mil-
rinone, using the theoretical model of PDE3A.11 A pre-
liminary study was also performed docking 8b and
milrinone in the virtual model of A1 adenosine recep-
tor14 in order to assess the binding data obtained on A1
receptor.
Moreover, as further investigation on the molecular
mechanism that leads to the effective PDE3 inhibition,
without antagonism at A1 receptor, we have recently
built a theoretical model of the catalytic site of PDE3A
by comparative modelling11 from the crystallographic
structure of the catalytic site of PDE4B2B and the site-
directed mutagenesis data available for PDE3A. On this
model, docking studies performed with milrinone and
different PDE3 inhibitors chosen from literature, let us
derive some useful information regarding the structural
requirements for selective enzymatic inhibition. The
lactamic function was confirmed by our study as the
primary binding site and as common anchor point for
all inhibitors: it displayed two hydrogen bonds with the
amino groupof Lys947 side chain and with the back-
bone carbonyl groupof Thr908. In particular, as con-
cern 2-pyridone derivatives, the substituent on position
3 seemed to be conditioned by the size of the hydro-
phobic pocket where the 2-pyridone ring is inserted
(residues Thr908, Leu910, His913, Lys947 and Ile951),
while the substituent on position 5 resulted free to
occupy a wide portion of the catalytic site.
Results and Discussion
Chemistry
In a previous paper,7 we reported the facile reaction of
open-chain and cyclic sym-2-dimethylaminomethylene-
1,3-diones with various dinucleophiles to give function-
alised pyrazoles, isoxazoles, pyrimidines and pyridines.
For the synthesis of the esters 4a–c,e–k (Scheme 1), we
adopted this fruitful reaction of both open-chain and
cyclic synthons 3a–k,15 obtained by us in very high
yields by refluxing a solution of the suitable 1,3-diones
in N,N-dimethylformamide dimethylacetal, and a
1,3-dinucleophile with C–C–N structure, namely the
methyl cyanoacetate.
Thus, to better characterize the binding site and to
visualize the nature of the molecular contacts between
2-pyridone inhibitors and PDE3A, a more detailed
knowledge of the interactions between this class of
ligands and the enzyme is required, especially as concern
substitution on positions 3 and 5. In fact, concerning
positions 4 and 6, a number of SAR and QSAR studies
have fully defined their structural characteristics.12a,b
This reaction usually occurs, first, between the nucleo-
philic methylenic group of the 1,3-dinucleophile and the
electrophilic extra-chain carbon atom of the synthons
3a–k. Subsequently, the dimethylamine, when released
in situ, mostly cause the reaction to proceed towards the
direct closure of the 2-pyridone ring without isolation of
intermediates. There were no problems concerning the
regiospecificity of the cyclization step, due to the sym-
metrical nature of the 1,3-dione. The cyclization step
afforded the desired esters 4, generally as a sole product,
in good yield. Only in the case of 3d, did the above
reaction not occur because of the strong steric hin-
drance caused by the tert-butyl group.
Taking into account the foregoing statements and in order
to obtain new milrinone analogues with a better cardio-
tonic activity due to a unique mechanism of action, we
planned the synthesis of methyl esters of 6-substituted
5-acyl-1,2-dihydro-2-oxopyridine-3-carboxylic acids 4a–c,
e, 1,2,5,6,7,8-hexahydro-2,5-dioxoquinoline-3-carboxilic
acids 4f–i, 2,5-dihydro-2,5-dioxo-1H-[1]pyrindine-3-carb-
oxylic acid 4j, 2,5-dihydro-2,5-dioxo-1H-indeno[1,2-
b]pyridine-3-carboxilic acid 4k and also the synthesis of
6 - ethyl - 5 - propionyl - 1H - pyridine - 2 - one 8b and
1H-indeno[1,2-b]pyridin-2,5-dione 8k.
In a few cases, we have been able to isolate intermediate
that confirm the direct intervention of the dimethyl-
amine which, in particular cases, may act as a base to
form water soluble salts. In fact, the formation of eno-
lates 5j,k is probably due to a stronger acidity of the
enol form involved, which protonates the released
dimethylamine thus forming a stable salt. Enolates 5
always gave the desired 2-pyridones 4 by refluxing in
anhydrous toluene with a small quantity of acetic acid.
In these new milrinone analogues, the 2-pyridone
nucleus is characterized by the presence of the ester
instead of the cyano or acetyl function in position 3,
whereas substituents in positions 5 and 6 are varied by
homologation, ramification and also aromatic substitu-
tion, or are fused generating a rigidified bicyclic or tri-
cyclic structures. Position 4 has been maintained
unsubstituted in accordance with literature.12a,b
To our knowledge, between the title compounds, the
only one described in literature was 4a, obtained in 68%
yield by reaction of methyl 3-methoxymethylene-2,4-
pentanedione with methyl cyanoacetate in the presence
of sodium methoxide.16

P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
4751
Scheme 1.
Only 5k was isolated, purified and characterized as a
solid product, which resulted very soluble in water. The
careful acidification of 5k gave the corresponding
methyl-2-aminocarbonyl-3-(3-hydroxy-1-oxo-1H-inden-
2-yl)acrylate 6. All the attempts to isolate and identify 5j
failed due to the difficulty in purifying the crude vitreous
reaction residue. Esters 4 were converted into carboxylic
acids 7 by saponification with potassium hydroxide in
boiling ethanol followed by acidification. 7j was not
obtained due to the lability of 4j in the experimental
conditions.
Pharmacological characterization
The newly synthesized analogues 4a–c,e–k, 7b,k and
8b,k were tested for their effects on the force of con-
traction and frequency rate of guinea pig spontaneously
beating atria in comparison with the parent drug milri-
none (Tables 1 and 2). Milrinone and the new analogues
were always tested on myocardial preparations obtained
from reserpine-treated animals in order to avoid a cate-
cholamine release from adrenergic stores which could
mask tissue responses to experimental drugs. Among
the tested compounds, 4c,f,g,h, 7k and 8b increased the
force of contraction of atria in a concentration-depen-
dent way (from 1 mM to 1 mM). Even at the highest
concentrations tested (1 mM), the analogues did not
cause arrhythmias or any increase in resting tension.
Moreover, the effects of the new compounds were com-
pletely reversible, since washout of myocardial prepara-
tions restored pre-drug cardiac contractility. The results
in Table 1 show that the inotropic response was partic-
ularly evident at the highest concentrations tested for
4c,f,g,h and 7k. 8b was the most active inotropic agent,
increasing atrial contractility to an extent significantly
higher than that of milrinone. The new compounds did
not significantly affect the cardiac frequency, only 8b
caused a chronotropic effect comparable to that exerted
by milrinone (Table 2). For the sake of clarity, it must be
noted that, in the same myocardial preparations, the full
b-adrenoceptor agonist isoprenaline (0.2 mM) caused an
increase of 262ꢀ4.8% in the force of contraction and an
Finally, the carefully decarboxylation of acids 7 by
reflux in quinoline containing a catalytic amount of
copper powder afforded to 6 substituted-5-acyl-1H-pyr-
idin-2-ones 8.
Acids 7a,c,e–i were already prepared, with a different
procedure, by refluxing a solution of the corresponding
amides in concentrated hydrochloric acid for a certain
time.7 None of these acids showed inotropic or chrono-
tropic activity in the isolated atria preparations.9 Pyr-
idones 8a,c,e–i were also obtained by decarboxylation
of the corresponding pyridine and quinoline carboxylic
acids.7 Among these pyridinones only 8h, characterized
by a quinolinedione structure, induced a positive ino-
tropic effect, particularly at higher concentrations. The
pharmacological characterization of 8h suggested that
an antagonism towards endogenous adenosine seemed
to be involved in its contractile effect.9

4752
P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
Table 1. Effect of compounds 4a–c,e–k, 7b,k and 8b,k on the force of contraction of spontaneously beating atria isolated from reserpine-treated
guinea pigs: comparison with milrinone
Compd
Developed tension (% variation from basal value)
ꢁ6
.
3 10
ꢁ5
.
3 10 M
10ꢁ6
M
M
10ꢁ5
M
10ꢁ4
M
3 10ꢁ4 M
.
10ꢁ3
M
Milrinone
16.69ꢀ0.95
6.94ꢀ0.76
ꢁ3.33ꢀ0.21
5.51ꢀ0.30
ꢁ0.13ꢀ0.05
2.12ꢀ0.13
2.04ꢀ0.08
2.00ꢀ0.18
ꢁ0.20ꢀ0.01
0.00ꢀ0.00
0.00ꢀ0.00
ꢁ0.94ꢀ0.02
5.94ꢀ0.10
18.96ꢀ0.40
10.35ꢀ0.29
30.24ꢀ0.75
10.49ꢀ0.26
ꢁ5.43ꢀ0.61
11.96ꢀ0.41
2.51ꢀ0.25
5.98ꢀ0.10
4.48ꢀ0.36
5.10ꢀ0.30
1.65ꢀ0.03
0.00ꢀ0.00
0.00ꢀ0.00
ꢁ1.88ꢀ0.07
5.94ꢀ0.03
27.09ꢀ3.06
13.78ꢀ0.34
38.68ꢀ1.25
4.47ꢀ0.35
ꢁ4.68ꢀ0.71
19.13ꢀ0.33
ꢁ1.27ꢀ0.18
9.53ꢀ0.38
7.25ꢀ0.32
10.12ꢀ0.56
ꢁ4.02ꢀ0.05
0.00ꢀ0.00
0.00ꢀ0.00
ꢁ1.88ꢀ0.07
5.94ꢀ0.03
40.13ꢀ3.03
13.78ꢀ0.34
46.16ꢀ0.80
ꢁ12.38ꢀ0.97
ꢁ5.40ꢀ1.01
28.47ꢀ0.45
ꢁ11.33ꢀ0.56
18.79ꢀ0.53
12.45ꢀ0.27
16.42ꢀ0.32
ꢁ21.30ꢀ0.83
0.00ꢀ0.00
52.70ꢀ0.83
ꢁ30.43ꢀ0.64
ꢁ5.40ꢀ1.81
51.08ꢀ1.09
ꢁ33.12ꢀ0.47
23.57ꢀ0.61
17.67ꢀ1.07
21.52ꢀ0.77
ꢁ59.82ꢀ0.45
3.70ꢀ0.04
52.25ꢀ1.13
ꢁ26.07ꢀ0.76
ꢁ4.56ꢀ0.98
72.39ꢀ0.63
ꢁ41.17ꢀ0.63
28.69ꢀ0.96
25.11ꢀ0.85
37.65 ꢀ0.88
ꢁ92.96ꢀ0.92
3.90ꢀ0.03
41.63ꢀ1.33
ꢁ30.60ꢀ1.45
nd
54.12ꢀ0.57
ꢁ45.80ꢀ0.70
50.33ꢀ0.95
45.05ꢀ0.60
40.89ꢀ1.01
nd
4a
4b
4c
4e
4f
4g
4h
4i
4j
nd
nd
4k
7b
7k
8b
8k
0.00ꢀ0.00
0.00ꢀ0.00
0.00ꢀ0.00
ꢁ0.94ꢀ0.05
14.80ꢀ0.28
57.19ꢀ5.80
21.22ꢀ0.49
ꢁ0.94ꢀ0.05
20.96ꢀ0.58
80.40ꢀ5.94
14.30ꢀ0.40
ꢁ0.94ꢀ0.05
27.69ꢀ0.81
106.79ꢀ6.15
6.86ꢀ0.28
ꢁ0.17ꢀ0.01
58.21ꢀ1.50
145.64ꢀ8.41
2.91ꢀ0.18
Milrinone and the new analogues were added cumulatively to the bathing fluid, and inotropic effect was recorded for 5 min after they reached
maxima, before adding a higher concentration. The effect of each concentration of a compound was defined by the difference between the force of
contraction, before and after its addition to the bathing fluid, and was expressed as a percent variation in developed tension in respect to the con-
trols. The value of basal force of contraction was 8.01ꢀ0.56 mN. Results are means ꢀSE from six experiments for each compound carried out on
six different myocardial preparations. Negative value indicates a negative inotropic effect.
Table 2. Effect of compounds 4a–c,e–k, 7b,k and 8b,k on the frequency rate of spontaneously beating atria isolated from reserpine-treated guinea
pigs: comparison with milrinone
Compd
Frequency (% variation from the basal value)
.
ꢁ6
.
3 10
10ꢁ6
M
M
10ꢁ5
M
3 10ꢁ5 M
10ꢁ4
M
3 10ꢁ4 M
.
10ꢁ3
M
Milrinone
3.01ꢀ0.24
ꢁ1.35ꢀ0.13
ꢁ16.66ꢀ2.21
ꢁ1.84ꢀ0.09
ꢁ1.18ꢀ0.19
ꢁ0.69ꢀ0.14
0.05ꢀ0.02
9.12ꢀ0.05
2.89ꢀ0.19
11.52ꢀ0.33
ꢁ6.71ꢀ1.26
ꢁ24.44ꢀ3.92
ꢁ0.33ꢀ0.05
ꢁ2.79ꢀ0.47
ꢁ0.17ꢀ0.02
3.24ꢀ0.17
18.22ꢀ0.84
ꢁ8.12ꢀ0.62
ꢁ24.44ꢀ3.21
0.60ꢀ0.33
24.46ꢀ0.75
ꢁ5.63ꢀ0.53
20.11ꢀ2.31
7.68ꢀ0.64
32.81ꢀ1.04
ꢁ7.52ꢀ0.55
19.56ꢀ2.44
7.49ꢀ0.44
39.20ꢀ0.90
0.00ꢀ000
nd
1.36ꢀ0.09
ꢁ41.00ꢀ0.87
8.34ꢀ0.60
0.99ꢀ0.05
10.60ꢀ0.75
nd
4a
4b
4c
4e
4f
4g
4h
4i
ꢁ19.44ꢀ1.45
ꢁ35.3ꢀ0.29
ꢁ1.36ꢀ0.15
ꢁ0.44ꢀ0.06
1.07ꢀ0.35
ꢁ11.49ꢀ0.60
0.28ꢀ0.08
ꢁ22.96ꢀ0.84
1.32ꢀ0.17
ꢁ38.83ꢀ0.87
6.70ꢀ0.36
ꢁ0.93ꢀ0.06
0.06ꢀ0.03
ꢁ0.59ꢀ0.07
2.60ꢀ0.23
0.77ꢀ0.13
ꢁ0.06ꢀ0.06
ꢁ0.17ꢀ0.05
0.00ꢀ0.00
ꢁ1.25ꢀ0.10
ꢁ0.93ꢀ0.11
0.00ꢀ0.00
ꢁ1.98ꢀ0.12
ꢁ10.46ꢀ0.94
ꢁ3.03ꢀ0.00
ꢁ6.46ꢀ0.51
ꢁ0.28ꢀ0.06
ꢁ10.43ꢀ0.30
8.34ꢀ0.12
3.83ꢀ0.23
ꢁ20.66ꢀ0.77
ꢁ9.09ꢀ0.01
ꢁ6.48ꢀ0.38
ꢁ0.28ꢀ0.04
ꢁ9.21ꢀ0.25
17.26ꢀ1.18
6.52ꢀ0.06
ꢁ36.10ꢀ0.71
ꢁ9.09ꢀ0.32
ꢁ0.00ꢀ0.00
ꢁ0.28ꢀ0.03
ꢁ0.56ꢀ0.06
21.94ꢀ2.23
6.52ꢀ0.06
ꢁ80.41ꢀ0.91
ꢁ10.01ꢀ0.41
0.00ꢀ0.00
4j
nd
nd
4k
7b
7k
8b
8k
0.00ꢀ0.00
0.00ꢀ0.00
ꢁ0.28ꢀ0.03
ꢁ7.82ꢀ0.24
1.14ꢀ0.03
ꢁ0.28ꢀ0.04
ꢁ7.82ꢀ0.64
3.85ꢀ0.10
ꢁ0.28ꢀ0.04
4.62ꢀ0.09
ꢁ0.28ꢀ0.05
34.98ꢀ0.20
41.28ꢀ3.42
16.41ꢀ0.17
32.88ꢀ2.29
11.07ꢀ0.13
5.64ꢀ0.10
4.45ꢀ0.09
5.54ꢀ0.07
Milrinone and the new analogues were added cumulatively to the bathing fluid, and chronotropic effect was recorded for 5 min after they reached
maxima, before adding a higher concentration. The effect of each concentration of a compound was defined by the difference between the frequency,
before and after its addition to the bathing fluid, and was expressed as a percent variation in frequency in respect to the controls. The value of basal
heart rates was 136ꢀ7 bpm. Results are means ꢀSE from six experiments for each compound carried out on six different myocardial preparations.
Negative value indicates a negative chronotropic effect.
increase of 57ꢀ3.0% in the heart rate over the basal
values, respectively. Since, among the tested com-
pounds, 8b seemed to be the most interesting one, fur-
ther experiments were carried out in order to
characterize its pharmacological action.
milrinone. EC50 values for 8b and milrinone, calculated
from these concentration–effect curves, were 126ꢀ4.10
and 27.35ꢀ0.10 mM, respectively.
In order to better elucidate the mechanisms responsible
for the positive inotropic effect of this new milrinone
analogue, its effect was also determined in the presence
of the b-adrenoceptor antagonist propranolol (10ꢁ7 M),
a1-adrenoceptor antagonist prazosin (5ꢂ10ꢁ9 M), H2-
histaminoceptor antagonist ranitidine (10ꢁ5 M), H1-
Firstly, as even the small effect of a compound on heart
frequency could influence the action of the substance on
the force of contraction of myocardial preparations, 8b
was studied in left atria electrically driven at 1 Hz.
Results (Table 3) confirmed that, also in these experi-
mental conditions, the milrinone analogue had a posi-
tive inotropic effect higher than that of milrinone.
However, for 8b, a higher inotropic activity did not
correspond to a higher potency compared to that of
histaminoceptor antagonist pyrilamine (10ꢁ7 M), Ca2+
-
channel blocker verapamil (10ꢁ7 M), and carbachol
(5ꢂ10ꢁ8 M), an agent known to selectively abolish the
increase in contractility of the atria induced by a rise of
cAMP levels.17,18 Moreover, the effect of 8b was also

P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
4753
Table 3. Effect of compund 8b on the force of contraction of electrically driven left atria isolated from reserpine-treated guinea-pigs
Compd
Developed tension (% variation from the basal value)
ꢁ6
.
3 10
ꢁ5
.
3 10
10ꢁ6
M
M
10ꢁ5
M
10ꢁ4
3 10ꢁ4 M
.
10ꢁ3
M
Milrinone
8b
1.23ꢀ0.01
9.84ꢀ0.27
4.92ꢀ0.03
13.25ꢀ0.31
16.77ꢀ1.38
22.83ꢀ0.31
25.97ꢀ2.67
34.72ꢀ0.38
36.83ꢀ2.94
44.72ꢀ0.38
52.64ꢀ3.67
41.96ꢀ0.32
70.93ꢀ5.64
11.90ꢀ0.37
Milrinone and the new analogues were added cumulatively to the bathing fluid, and inotropic effect was recorded for 5 min after they reached
maxima, before adding a higher concentration. The effect of each concentration of a compound was defined by the difference between the force of
contraction, before and after its addition to the bathing fluid, and was expressed as a percent increase in developed tension over controls. The value
of basal force of contraction was 4.91ꢀ0.42 mN. Results are means ꢀSE from six experiments for each compound carried out on six different
myocardial preparations.
tested in the presence of adenosine deaminase (ADA, 2
U/mL), the enzyme that inactivates adenosine by meta-
bolising it to inosine. As shown in Table 4, the cardio-
tonic action of 8b does not involve activation of
adrenoceptors (both a- and b-adrenoceptors) or hista-
minoceptors (both H1- and H2-histaminoceptors), since
its effect was not influenced by the presence of prazosin,
propranolol, pyrilamine or ranitidine, respectively. An
increased uptake of extracellular Ca2+ through voltage-
sensitive Ca2+ channels, with consequent increase in
intracellular Ca2+ availability for the contractile
machinery, could be involved in the inotropic response to
8b. In fact, verapamil, at a concentration sufficiently high
to block slow-Ca2+ channels, inhibited the effect of the
new analogue. In Table 4, it is also shown that the
increase in cardiac developed tension, evoked by the new
milrinone analogue, was not modified by the presence of
adenosine deaminase. As a means of comparison, it can
be observed that, under the same experimental condi-
tions, the presence of adenosine-deaminase decreases the
positive inotropic effect of milrinone by about 90%.19
This interesting finding clearly suggests that the effect of
8b does not originate from its ability to antagonize the
influence exerted by endogenous adenosine on the heart.
Such a hypothesis is supported by the results of some
experiments carried out to evaluate the affinity of 8b
towards A1 and A2A receptors. Studies of displacement
of [3H]CHA from rat cortical membranes and of
[3H]CGS21680 from rat striatal membranes indicate very
low or no affinity of 8b for A1 (percent inhibition of spe-
cific binding at a concentration of 10 mM, %I=10.2) and
A2A adenosine receptors (percent inhibition of specific
binding at a concentration of 10 mM, %I=0.0).
However, the main finding of this set of experiments is
represented by the data obtained in the presence of car-
bachol (see Table 4), which clearly suggest an important
involvement of cAMP in the cardiac action of 8b, since,
at all tested concentrations, the effect of the new analo-
gue was significantly reduced when left atria were pre-
incubated with carbachol. This finding is confirmed by
the results reported in Figure 1, in which the inhibitory
effects of both milrinone and 8b on the activity of a
PDE3 isolated and partially purified from guinea pig
cardiac tissue are shown. 8b inhibited in a concen-
tration-dependent way the enzyme activity. When com-
pared to the parent drug, the new analogue appeared
rather similar to milrinone in its activity, its maximal
inhibition against PDE3 activity being 81.93ꢀ0.29%
versus 100% of milrinone. The potency of the new
compound was also quite similar to that of the parent
drug, as indicated by the comparison of Ki values
(4.67ꢀ0.04 and 8.64ꢀ0.22 mM for milrinone and 8b,
respectively), calculated by means of Cheng and Pru-
soff’s equation20 from the reported concentration–effect
curves. These results, together with those obtained in
the presence of carbachol, indicate that the positive
inotropic effect exerted by this new milrinone analogue
is related to an increase in cAMP levels obtained
through inhibition of PDE3 activity, even if the ability
of the compound to increase Ca2+availability through
voltage-dependent Ca2+channels cannot be excluded.
In this regard, an increase in Ca2+ entry through Ca2+
channels has been already proposed for milrinone.21
However, this effect may be a consequence of cAMP
increase due to PDE inhibition, as recently clearly
reviewed.22
Table 4. Effect of various antagonists on the positive inotropic effect exerted by increasing concentrations of compound 8b on the force of
contraction of electrically driven left atria isolated reserpine-treated from guinea pigs
Inhibitor (M)
Developed tension (ꢀ% over the basal value)
10ꢁ5 M 8b
ꢁ6
ꢁ3
.
3 10 M 8b
.
3 10 M 8b
—
11.90ꢀ0.30
36.83ꢀ2.94
75.45ꢀ6.12
Propranolol
Prazosine
Rantidine
Pyrilamine
Verpamile
Carbachol
(10ꢁ7 M)
11.61ꢀ0.25*
35.41ꢀ2.32*
73.45ꢀ6.12*
ꢁ9
.
(5 10 M)
(10ꢁ5 M)
(10ꢁ7 M)
(10ꢁ7 M)
10.31ꢀ0.99*
37.53ꢀ3.21*
76.37ꢀ3.22*
12.61ꢀ1.34*
34.98ꢀ3.01*
74.89ꢀ3.45*
12.45ꢀ1.55*
34.79ꢀ3.65*
75.21ꢀ4.01*
4.76ꢀ0.23 (ꢁ60%)**
7.14ꢀ1.03 (ꢁ40%)**
11.63ꢀ0.98*
16.58ꢀ1.57 (ꢁ55%)**
23.93ꢀ2.05 (ꢁ35%)**
36.83ꢀ2.45*
41.49ꢀ2.31 (ꢁ45%)**
52.81ꢀ3.54 (ꢁ30%)**
73.78ꢀ5.74*
ꢁ8
.
(5 10 M)
(2 U/mL)
Adenosine-deaminase
Results are means ꢀSE from six experiments for each compound carried out on six different myocardial preparations. *Not statistically significant
(p>0.05); **statistically significant (p<0.001) versus values obtained with the same concentration of 8b in the absence of antagonists.

4754
P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
Figure 2. Binding of milrinone, 4b, 8b and cilostamide into the cata-
lytic site of PDE3A. The protein is shown as ribbon. Milrinone, 4b, 8b
(sticks) and PDE3A residues involved in hydrogen bonding are
coloured as follows: green, carbon atom; red, oxygen atom; blue,
nitrogen atom; grey, hydrogen atom. For clarity, cilostamide (sticks) is
coloured in orange. Other PDE3A residues mentioned in the text and
not involved in hydrogen bonding, are coloured in blue.
Phe1004 and Leu1012. As a consequence, the cyano
group, which does not interact with the biological
counterpart due to its polar nature, could be replaced
by a hydrogen.
Figure 1. Concentration–effect curves for the inhibition of the activity
of PDE3 partially purified from guinea pig cardiac tissue by milrinone
and its analogue 8b. PDE3 activity was measured by the two-step
method,27 as previously described.10 Results are meansꢀSE from four
experiments in duplicate. Data are expressed as percent inhibition of
the basal enzyme activity (1.14ꢀ0.02 nmol cAMP hydrolyzed/mg of
protein/min).
Regarding the substituent in position 5, the acyl moiety
occupies a portion common to the pyridine ring of mil-
rinone. The carbonyl group, according to our calcula-
tions, is not able to display any hydrogen bond or polar
interactions with the enzymatic counterpart, such as in
the case of the pyridine ring of the reference compound.
However, it should be noted that the pyridine ring on
milrinone is able to partially shield residues Tyr751 and
His752, described as extremely important for catalysis
by site directed mutagenesis experiments.23 This finding
contributes to justify on molecular basis, the slightly
higher potency of milrinone as inhibitor in comparison
with 8b, as pointed out by the pharmacological results.
In addition, docking studies highlight that the sub-
stituent on position 5 in 2-pyridone derivatives has the
possibility to occupy a large pocket making extra
hydrogen bonds or additional hydrophobic interactions
into the PDE3A catalytic site. This finding could helpin
explaining the reduced potency displayed by milrinone
and 8b in comparison with inhibitors with a longer
molecular scaffold. For example cilostamide, a well-
known potent PDE3 inhibitor (Ki value vs PDE3=5
nM),13 properly occupies the catalytic site showing a
similar hydrogen bonding network than milrinone and
8b plus extensive van der Waals interactions with
Leu910 and Ile951 and a number of hydrophobic inter-
actions with the side chains of residues Phe989, Leu1000
and Phe1004.
For complete in vitro study of this new analogue, 8b
was also tested on other enzymes involved in the control
of cardiac contractility, that is sarcolemmal Na+/
K+ATPase and Ca2+ATPase, and sarcoplasmic reticu-
lum Ca2+ATPase. The analogue was devoid of any
effect on both Ca2+ATPases, whereas, at the highest
concentration tested (10ꢁ3 M), it inhibited the activity
of Na+/K+ATPase of 18ꢀ2% in respect of control
values.
Docking studies
Results of docking studies on the theoretical model of
PDE3A11 for structure 8b, the most active compound
among those studied (Fig. 2), highlighted that this deri-
vative shares a common binding mode with milrinone,
inserting its lactamic ring in the hydrophobic pocket
defined by Thr908, Leu910, His913, Lys947 and Ile951.
The lactamic function, as in the case of milrinone, dis-
plays two hydrogen bonds with the biological counter-
part: the first with the amino group of Lys947 sidechain
and the second with the backbone carbonyl groupof
Thr908. On position 3, the absence of the cyano moiety
does not seem to positively or negatively influence the
binding mode of 8b. This result could be explained by
taking into account that 2-pyridone derivatives direct
the substituent on position 3 towards a pocket defined
mainly by non-polar residues such as Leu910, Ile951,
Furthermore, in order to give a preliminary investiga-
tion whether a negative interaction or a non-interaction
with the enzyme could helpin explaining the lacking of
pharmacological activity in derivative 4b, we also
applied our docking procedure in this case. The results
obtained with 4b (Fig. 2), suggest that substitution of

P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
4755
the cyano with an ester function is detrimental, since it
produces a different binding mode, probably due to the
steric hindrance of the ester function, longer than a
cyano moiety. The ester function, according to the
results of our calculations, makes the pyridone unable
to fully accommodate into the hydrophobic pocket
formed by Thr908, Leu910, His913, Lys947 and Ile951.
The molecule does not display hydrogen bonds with the
counterpart and occupies a part of the enzymatic pocket
closer to residues Tyr751 and Phe972. The consequent
lost of the interaction with residues Lys947 and Thr908
combined with the pharmacological profile displayed by
compound 4b could be seen as a further proof of the key
role played by these two amino acid residues, specific
for binding to the inhibitor.
directed mutagenesis experiments,24 which have fully
demonstrated the involvement of these two amino acid
residues in the antagonist binding.
Thus, the findings of our computational approach sup-
port the inhibition of PDE3 as the unique mechanism of
action for 8b and support the hypothesis of a double
mechanism of action for milrinone. Moreover, they
suggest an essential role in 2-pyridone derivatives for
the cyano moiety, which probably is one of the switch
able to turn their biological activity mainly towards the
adenosine receptor antagonism.19
Conclusion
In addition, since in vitro binding results on PDE3
highlight the enzyme inhibition as the unique mechan-
ism of action of 8b in comparison with milrinone, which
is reported to act also via antagonism at cardiac A1
adenosine receptor, we find it interesting to verify whe-
ther these biological data could be supported by the
results of an ‘in silico’ approach. Regarding the catalytic
site of PDE3A, the similar binding mode displayed by
milrinone and 8b, does not highlight any peculiar inter-
action able to explain the pure enzymatic mechanism of
action of this analogue. We thus performed docking
experiments on milrinone and 8b, using a theoretical
model of the human A1 adenosine receptor recently
developed by us.14 The results obtained (Fig. 3) show
that 8b does not have any particular affinity towards the
biomolecule, since it accommodates into the binding site
of the receptor without displaying any interaction with
the counterpart. Interestingly, milrinone is able to
establish a hydrogen bond between the nitrogen of lac-
tamic function and the backbone carbonyl groupof
residue Thr91 and an additional polar interaction
between the cyano groupin position 3 and the amino
groups (backbone or, alternatively, side chain) of
His278. These data are in agreement with results of site-
The present paper presents a combination of synthetic,
pharmacological and computational methods to gain
further insights into the molecular requirements for
selectively targeting 2-pyridone derivatives towards
PDE3 inhibition or A1 adenosine receptor antagonism.
Interestingly, through this approach, for the first time, a
new milrinone analogue has been synthesized, that is 8b,
endowed with a positive inotropic effect higher than
that of the parent drug but, differently from milrinone,
provided with high selectivity towards cardiac PDE3.
Pharmacological characterization and computational
results in fact, suggest PDE3 inhibition as a unique
mechanism of 8b cardiac actions. In addition, these
docking studies give further evidence to the key role
played in PDE3 by residues Thr908 and Lys947 in the
binding with the inhibitor. Differences in the interaction
of 2-pyridone derivatives with the enzyme and the A1
receptor, depending on the nature of substituent on
position 3, have been pointed out. Furthermore, some
considerations on the characteristics of substituent on
position 5 of the 2-pyridone structure, in order to
increase its inhibitory potency towards PDE3, have
been drawn. QSAR and 3-D-QSAR studies on this class
of milrinone analogues are in progress.
Experimental
Chemistry
Melting points were determined with a Fisher–Johns
apparatus and are uncorrected. IR spectra were regis-
tered on a Perkin-Elmer 398 spectrophotometer and are
.
expressed in cmꢁ1 1H NMR and 13C NMR spectra
were recorded on a Varian Gemini 200 (200 and
50.30 MHz, respectively) spectrometer. Chemical shifts
are reported as d values (ppm) downfield from internal
Me4Si in the solvent shown. Coupling constants (J) are
expressed in Hertz (Hz). The following NMR abbrevia-
tions are used: br (broad), s (singlet), d (doublet), t (tri-
plet), q (quartet), m (multiplet), ar (aromatic proton), ex
(exchangeable with D2O). Elemental analyses for C, H,
N, were performed on a Model EA 1110 Elemental
Analyzer (Carlo Erba Strumentazione) and results agree
withinꢀ0.3% with calculated values. All reagents were
of analytical grade.
Figure 3. Binding of milrinone and 8b into the active site of human A1
adenosine receptor. The protein is shown as ribbon. Ligand (sticks)
and A1 residues involved in hydrogen bonding are coloured as follows:
green, carbon atom; red, oxygen atom; blue, nitrogen atom; grey,
hydrogen atom.

4756
P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
Methyl esters of 6-substituted 5-acyl-1,2-dihydro-2-oxo-
pyridine-3-carboxylic acids 4a–c,e and 1,2,5,6,7,8-hexa-
hydro-2,5-dioxoquinoline-3-carboxylic acids 4f-i: general
direct procedure
Methyl 7-phenyl-1,2,5,6,7,8-hexahydro-2,5-dioxoquino-
line-3-carboxylate (4i). (93%); mp303–305; IR (KBr):
1
3300–2500, 1700, 1648, 1587 cmꢁ1. H NMR (DMSO-
d6) d 2.40–3.60 (m, 5H, CH-7+CH2-6+CH2-8), 3.79 (s,
3H, CH3O), 7.4 (s, 5H, ar), 8.47 (s, 1H, CH-4), 12.70 (br
s, 1H, NH ex).
Methyl cyanoacetate (9,9 g, 100 mmol) was added to a
solution of the suitable sym-2-dimethylaminomethylene-
1,3-diones 3 (10 mmol) in anhydrous MeOH (20 mL).
The mixture was stirred at reflux at different times
(4a,f,g,h,i, 7 h; 4b,c,e, 24 h), then the formed crude solid
was filtered, washed thoroughly with anhydrous Et2O
and recrystallized (4a,f,g,i from EtOH; 4b,c,e,h, from
EtOAc).
Dimethyl ammonium salts of enolate anions 5j,k: general
procedure. As mentioned above, during the synthesis of
pyridones 4j and 4k, the corresponding dimethylammo-
nium salts 5j and 5k were obtained as main products.
Their isolation from thick crude oily residues of the
reaction was carried out with a thorough washing and
handling with Et2O/EtOAc (3:1) at rt. The organic layer
was dried, filtered and evaporated under reduced pres-
sure: the solid residue was finally recrystallized from
EtOH.
Methyl 5-acetyl-6-methyl-1,2-dihydro-2-oxopyridine-3-
carboxylate (4a). (72%); mp202–203 ꢃC; IR (CHCl3):
3350–2500, 1739, 1705, 1655 cmꢁ1. 1H NMR (CDCl3) d
2.54 (s, 3H, CH3), 2.80 (s, 3H, CH3CO), 3.93 (s, 3H,
CH3O), 8.72 (s, 1H, CH-4), 12.94 (br s, 1H, NH ex).16
Dimethyl ammonium salt of enolate anion of methyl 2-
cyano-3-(2-hydroxy-5-oxocyclopenten-1-yl)acrylate (5j).
63%); mp154–156 ꢃC; IR (KBr): 3300–2700, 2478,
Methyl 5-propionyl-6-ethyl-1,2-dihydro-2-oxopyridine-3-
carboxylate (4b). (88%); mp178–180 ꢃC; IR (CHCl3):
1
2200, 1708, 1675 cmꢁ1. H NMR (DMSO-d6) d 2.22 (s,
1
3360–2500, 1735, 1702, 1653 cmꢁ1. H NMR (CDCl3)
4H, 2CH2), 2.58 (s, 6H, N(CH3)2), 3.67 (s, 3H, CH3O),
7.65 (s, 1H, CH), 8.22 (near s, 2H, OH+NH ex).
d 1.18 (t, J=7, 3H, CH2CH3), 1.34 (t, J=7, 3H,
CH2CH3), 2.50–3.35 (m, 4H, 2 CH2CH3), 3.93 (s,
3H, CH3O), 8.75 (s, 1H, CH-4), 12.60 (br s, 1H, NH
ex).
Dimethyl ammonium salt of enolate anion of methyl
2-cyano-3-(3-hydroxy-1-oxo-1H-inden-2-yl)acrylate (5k).
(70%); mp166–167 ꢃC; IR (KBr): 3200–2700, 2480,
1
Methyl 5-isobutyryl-6-isopropyl-1,2-dihydro-2-oxopyri-
dine-3-carboxylate (4c). (83%); mp204–205 ꢃC; IR
2198, 1706, 1662 cmꢁ1. H NMR (DMSO-d6) d 2.57 (s,
6H, N(CH3)2), 3.69 (s, 3H, CH3O), 7.49–7.63 (m, 4H,
ar), 7.81 (s, 1H, CH), 8.18 (near s, 2H, OH+NH ex).
13C NMR (DMSO-d6) d 34.50, 52.44, 81.52, 107.91,
109.79, 120.53, 132.99, 138.65, 143.08, 169.50, 193.08.
1
(CHCl3): 3370, 3300–2500, 1738, 1705, 1658 cmꢁ1. H
NMR (CDCl3) d 1.17 (d, J=7, 6H, CH(CH3)2), 1.40 (d,
J=7, 6H, CH(CH3)2), 3.00–3.80 (m, 2H, 2 CH(CH3)2),
3.95 (s, 3H, CH3O), 8.61 (s, 1H, CH-4), 12.05 (br s, 1H,
NH ex).
Methyl 2-aminocarbonyl-3-(3-hydroxy-1-oxo-1H-inden-
2-yl)acrylate 6. An aqueous (20 mL) solution of the salt
5k (3 g, 10 mmol) was carefully acidified with 6 M HCl
at 0 ꢃC. The amorphous precipitate was filtered, washed
thoroughly with water, dried and purified by recrys-
tallization from MeOH. Yield 65%; mp165–167 ꢃC; IR
Methyl 5-benzoyl-6-phenyl-1,2-dihydro-2-oxopyridine-3-
carboxylate (4e). (54%); mp234–236 ꢃC; IR (CHCl3):
.
3350, 3200–2500, 1738, 1705, 1650 cmꢁ1 1H NMR
(CDCl3) d 3.92 (s, 3H, CH3CO), 7.20–7.70 (m, 10H, ar),
8.47 (s, 1H, CH-4), 12.30 (br s, 1H, NH ex).
1
(KBr): 3330, 3190, 1690, 1640, 1568 cmꢁ1. H NMR
(DMSO-d6) d 3.72 (s, 3H, CH3), 7.50–7.70 (m, 4H, ar),
7.84 (s, 1H, CH), 7.30–9.30 (very br s, 3H, OH+NH2,
ex). 13C NMR (DMSO-d6) d 51.85, 106.00, 106.93,
120.70, 132.93, 137.19, 138.86, 168.11, 171.50, 189.06.
Methyl 1,2,5,6,7,8-hexahydro-2,5-dioxoquinoline-3-car-
boxylate (4f). (80%); mp272–274; IR (CHCl ): 3200–
3
2500, 1738, 1680, 1640 cmꢁ1. 1H NMR (CDCl3) d 1.90–
2.40 (m, 2H, CH2-7), 2.45–2.75 (m, 2H, CH2-6), 2.95
(near t, J=6, 2H, CH2-8), 3.89 (s, 3H, CH3O), 8.72 (s,
1H, CH-4), 12.30 (br s, 1H, NH ex).
Methyl 2,5 - dihydro - 2,5 - dioxo - 1H - [1]pyrindine - 3 -
carboxylate (4j) and methyl 2,5-dihydro-2,5-dioxo-1H-
indeno[1,2-b]pyridine-3-carboxylate (4k): procedure via
enolate intermediates. A solution of suitable salts 5 (10
mmol) and glacial AcOH (0.6 g, 10 mmol) in anhydrous
toluene (200 mL) was refluxed in a Dean-Stark appara-
tus for 48 h. The mixture was washed with 1 M aqueous
NaOH solution and water, dried and evaporated under
reduced pressure. The residue was recrystallized from
EtOH.
Methyl 7-methyl-1,2,5,6,7,8-hexahydro-2,5-dioxoquino-
line-3-carboxylate (4g). (98%); mp263–265; IR
1
(CHCl3): 3470, 3300–2400, 1732, 1680, 1642 cmꢁ1. H
NMR (CDCl3) d 1.18 (d, J=6, 3H, CH3-7), 2.12–2.55
(m, 2H, CH2-8), 2.55–2.95 (m, 2H, CH2-6), 2.95–3.25
(m, 1H, CH-7), 3.87 (s, 3H, CH3O), 8.75 (s, 1H, CH-4),
13.03 (br s, 1H, NH ex).
Methyl 7,7-dimethyl-1,2,5,6,7,8-hexahydro-2,5-dioxoqui-
noline-3-carboxylate (4h). (70%); mp243–245; IR
Methyl 2,5 - dihydro - 2,5 - dioxo - 1H - [1]pyrindine - 3 -
carboxylate (4j). (57%); mp299–301 ꢃC; IR (KBr):
1
1
(CHCl3): 3100–2400, 1738, 1680, 1642 cmꢁ1. H NMR
3200–2400, 1735, 1662, 1648 cmꢁ1. H NMR (DMSO-
(CDCl3) d 1.16 (s, 6H, 2CH3), 2.48 (s, 2H, CH2-8), 2.95
(s, 2H, CH2-6), 3.92 (s, 3H, CH3O), 8.83 (s, 1H, CH-4),
12.80 (br s, 1H, NH ex).
d6) d 2.50–2.80 (m, 2H, CH2-7), 2.90–3.20 (m, 2H, CH2-
6), 3.88 (s, 3H, CH3O), 8.21 (s, 1H, CH-4), 12.90 (br s,
1H, NH ex).

P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
4757
Methyl 2,5-dihydro-2,5-dioxo-1H-indeno[1,2-b]pyridine-
3-carboxylate (4k). (98%); mp310–312 ꢃC; IR (KBr):
Pharmacological studies
1
3100–2500, 1715, 1690, 1638 cmꢁ1. H NMR (DMSO-
Animals. Dunkin–Hurtley male guinea pigs (350–400 g),
obtained from Harlan Italy (S. Pietro al Natisone,
Italy), were kept in controlled environmental conditions
(temperature: 23ꢀ2 ꢃC; light–dark cycle: 7 am to 7 pm).
Animals had free access to a standard laboratory diet
and to water. All animal-use procedures described in
this paper were in accordance with the National Insti-
tutes of Health guidelines for the Care and Use of
Laboratory Animals and comply with the ethical prin-
ciples and guidelines adopted by the European Com-
munity, law 86/609/CEE. The experimental protocol
was approved by the local veterinary committee.
d6) d 3.78 (s, 3H, CH3O), 7.55–7.75 (m, 3H, ar), 7.95–
8.05 (m, 1H, CH-6), 8.18 (s, 1H, CH-4), 13.75 (br s, 1H,
NH ex).
6-Substituted 5-acyl-1,2-dihydro-2-oxopyridine-3-car-
boxylic acids (7b) and 2,5-dihydro-2,5-dioxo-1H-
indeno[1,2-b]pyridine-3-carboxylic acid (7k): general pro-
cedure. Potassium hydroxide (1.68 g, 30 mmol) dis-
solved in 95% EtOH (20 mL) was added to a solution of
the corresponding ester 4 (10 mmol) in the same solvent
(30 mL). The resulting solution was boiled for 24 h (48 h
in the case of 4k), the solvent was evaporated under
reduced pressure and the residue was dissolved with
water (20 mL). The solution was washed with Et2O and
acidified with 6 N HCl at 0 ꢃC (p Hꢄ1) and the solid
residue was filtered, washed with water, dried in a
vacuum oven at 100 ꢃC and purified by recrystallization
from EtOH (7k) or from EtOAC (7b).
In order to obtain myocardial tissues depleted in endo-
genous catecholamines, the animals were treated intra-
peritoneally daily for 2 days with reserpine (2 mg/kg)
before sacrifice. At the time of sacrifice, the animals
were anaesthetised with inhalation of methoxyflurane
and then killed by cervical dislocation; the heart being
rapidly removed.
7a,c,e–i, already prepared by us following a different
synthetic way,7 were also synthesized according the
above procedure thus confirming not only the satisfac-
tory yield but also further proving their structure.
Assessment of inotropic and chronotropic activities on
isolated atria preparations. The atria were separated
from ventricles and suspended vertically in a 30 mL
organ bath containing a physiological salt solution
constantly gassed by 95% O2 and 5% CO2, at 30 ꢃC.
The bathing solution contained (in mM): NaCl 120,
KCl 2.7, CaCl2 1.36, MgCl2 0.09, NaH2PO4 0.4,
NaHCO3 12 and glucose 5.5. When left atria were used,
they were suspended vertically in the organ bath and
electrically driven at 1 Hz by square-wave pulses just
above threshold voltage, of 0.6–0.9-ms duration (S44
stimulator, Grass Instrument Corporation, Quincy,
MA, USA).
5-Propionyl-6-ethyl-1,2-dihydro-2-oxopyridine-3-car-
boxylic acid (7b). (81%); mp185–187 ꢃC; IR (CHCl3):
3500–2500, 1735, 1690, 1630 cmꢁ1. 1H NMR (CDCl3) d
1.21 (t, J=7, 3H, CH2CH3), 1.40 (t, J=7, 3H,
CH2CH3), 2.75–3.50 (m, 4H, 2CH2CH3), 9.04 (s, 1H,
CH-4), 12.60 (br s, 1H, NH ex), 13.22 (near s, 1H,
CO2H ex).
2,5-Dihydro-2,5-dioxo-1H-indeno[1,2-b]pyridine-3-car-
boxylic acid (7k). (99%); mp >465 ꢃC; IR (KBr): 3300–
1
2500, 1725, 1697, 1640 cmꢁ1. H NMR (DMSO-d6) d
7.60–7.80 (m, 3H, ar), 8.05–8.15 (m, 1H, CH-6), 8.32 (s,
1H, CH-4).
Resting tension was adjusted to about 10 and 5 mN in
spontaneously beating atria and electrically driven left
atrium, respectively. The developed tension was recor-
ded isometrically by means of a high-sensitivity force
transducer (type DYO for isolated auricles; Basile,
Comerio, Varese, Italy) connected to a writing oscillo-
graph (Unirecord System, model 7050, Basile, Comerio,
Varese, Italy). The initial equilibration period was
40–60 min for each preparation. Since the atria were
isolated from reserpine-treated animals, before the
beginning of the experiments, depletion of catechol-
amines was verified by lack of any positive inotropic
effect induced by the addition of tyramine (1.5 mM).
Experiments were performed only in preparations that
did not respond to tyramine. Milrinone and its analo-
gues were added cumulatively and the responses caused
by each drug concentration were recorded upto the
maximum response before a higher concentration was
added. The effects of each compound on the force of
contraction and the frequency were expressed as the
percent increase over controls (ꢀ%). Milrinone and
their analogues were dissolved in dimethyl sulfoxide
(DMSO). The final concentration of DMSO in the
medium did not influence the basal activity of the atrial
preparations. Where indicated, propranolol (0.1 mM) or
prazosin (5 nM) or ranitidine (10 mM) or pyrilamine (1
6-Substituted-5-acyl-1H-pyridin-2-ones (8b,k). The title
pyridinones were prepared according to the procedure
already described by us.7 A solution of 7b,k (10 mmol)
in quinoline (15 mL) containing copper powder (0.20 g)
was refluxed for 6 h. The mixture was filtered hot, the
liquid was cooled and CHCl3 (25 mL) was added. The
added CHCl3 caused the immediate precipitation of the
product, which was filtered and recrystallized from Et2O
(8b) or EtOH (8k).
6-Ethyl-5-propionylpyridine-2(1H)-one (8b). (40%); mp
140–142 ꢃC; IR (CHCl3): 3100–2600, 1680, 1650 cmꢁ1
.
1H NMR (CDCl3) d 1.17 (t, J=7, 3H, CH2CH3), 1.32
(t, J=7, 3H, CH2CH3), 2.82 (q, J=7, 2H, CH2CH3),
3.05 (q, J=7, 2H, CH2CH3), 6.45 (d, J=10, 1H, CH-
3), 7.89 (d, J=10, 1H, CH-4), 12.75 (br s, 1H, NH
ex).
1H-Indeno[1,2-b]pyridine-2,5-dione (8k). (20%); mp
352–355 ꢃC; IR (KBr): 3000–2400, 1710, 1658, 1640
1
cmꢁ1. H NMR (DMSO-d6) d 7.45–7.72 (m, 4H, ar),
8.07 (s, 1H, CH-4).

4758
P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
mM) or verapamil (1 mM) or carbachol (0.05 mM) or
adenosine-deaminase (2U/mL) were added to the bath-
ing medium 20 min before the inotropic agents.
computational procedure already described.11,14 Mole-
cular structures of ligands (4b, 8b, milrinone, cilosta-
mide) were built and energy minimized within
MacroModel.32 Atomic charges were determined with
semi-empirical AM1 calculations, using Spartan ’O2.33
Calculations were carried out on a SGI O2 5000 work-
station and on a standard personal computer running
under Linux.
Assay of soluble PDE3 activity from guinea pig heart.
PDE3 was isolated and partially purified from guinea
pig heart using the procedure by Weishaar et al.25 Pro-
tein content was determined by the method of Lowry et
al.26 using serum albumin as a standard. PDE activity
was measured by the two-step procedure of Thompson
et al.27 in a 0.4 mL medium containing 40 mM Tris–HCl
(pH 8.0), 5 mM MgCl2, 1 mM EGTA, 0.4 mM
[3H]cAMP and about 30–50 mg of proteins, as pre-
viously described.10 PDE3 from guinea pig heart had
apparent Km for cAMP of 1.33ꢀ0.15 mM and a Vmax of
4.54ꢀ0.29 nmol cAMP hydrolyzed/mg protein/min. Ki
values were calculated according to Cheng and Prusoff’s
equation20 [Ki=IC50/(1+C/KD)] where IC50 is the con-
centration of the compound that gives 50% of the
obtainable inhibition, C is the substrate concentration
used (0.4 mM cAMP) and KD (1.33 mM) its dissociation
constant. When assayed at 0.4 mM cAMP, the activity of
our preparations of PDE3 was 1.14ꢀ0.02 nmol cAMP
hydrolyzed/mg of protein/min. Milrinone and 8b analo-
gue were dissolved in DMSO, the final concentration of
which in the reaction mixture did not exceed 1%. A
similar amount of DMSO was added to the control
samples. This amount did not affect PDE 3 activity.
Supporting Material
Compd
Formula
Analysis % calcd/found
C
H
N
4b
4c
4e
4f
C12H15NO4
C14H19NO4
C20H15NO4
C11H11NO4
C12H13NO4
C13H15NO4
C17H15NO4
C10H9NO4
C14H9NO4
C12H16N2O4
C16H16N2O4
C14H11NO5
C11H13NO4
C13H7NO4
C10H13NO2
60.75
60.82
63.38
63.37
72.06
71.80
59.72
59.53
61.27
61.11
62.64
62.20
68.68
68.59
57.97
58.15
65.88
65.99
57.13
57.01
63.99
63.80
61.54
61.49
59.18
59.32
64.74
64.74
67.02
67.27
6.37
6.37
7.22
7.21
4.54
4.50
5.01
5.11
5.57
5.55
6.06
6.07
5.08
5.04
4.38
4.46
3.55
3.41
6.39
6.45
5.37
5.50
4.06
4.07
5.87
5.92
2.93
3.01
7.31
7.24
3.91
3.88
5.90
5.91
5.28
5.22
4.20
4.14
6.33
6.14
5.95
5.80
5.62
5.45
4.71
4.69
6.76
6.73
5.49
5.42
11.10
11.02
9.33
9.28
5.13
5.11
6.27
6.26
5.81
5.78
7.82
7.95
6.79
6.92
4g
4h
4i
Assay of sarcolemmal Na+/K+ATPase and Ca2+AT-
Pase activity. Cardiac sarcolemmal vesicles were pre-
pared from guinea pig ventricular tissue by the method
of Slaughter et al.28 Na+/K+ATPase and Ca2+ATPase
activities were measured as previously described.29
Control values for Na+/K+ATPase and Ca2+ATPase
activities were about 220 and 58 nmol ATP hydrolyzed/
mg protein/min, respectively.
4j
4k
5j
5k
6
Assay of sarcoplasmic reticulum Ca2+ATPase activity.
A crude cardiac membrane vesicles preparation enri-
ched in sarcoplasmic reticulum was obtained by the
method of Jones et al.30
7b
7k
8b
8k
Ca2+-stimulated ATPase activity was assayed as pre-
viously described.31 Control values for sarcoplasmic
reticulum Ca2+ATPase was about 290 nmol Pi formed/
mg of protein/min.
.
C12H7NO2 1=H2O 69.90
71.04
Statistical analysis. Data are expressed as arithmetic
meansꢀSEM. Comparisons between means were cal-
culated by one-way analysis of variance (ANOVA). A p
value of less than 0.05 was considered statistically sig-
nificant. The EC50 values (where EC50 values were the
concentrations that gave half of the maximum effect
obtainable with the compound) and IC50 values (see
above) were calculated by means of GraphPad Prism
3.03 Software from the concentration–response data,
evaluated by sigmoidal curve fitting.
Acknowledgements
Financial support from MIUR (Italy) and University of
Genova is gratefully acknowledged.
References and Notes
Docking studies
1. Farah, A. E.; Alousi, A. A. Life Sci. 1978, 22, 1139.
2. Nicholson, C. D.; Challiss, R. A. J.; Shahid, M. Trends
Pharmacol. Sci. 1991, 12, 19.
Docking studies on PDE3A catalytic site and A1
adenosine receptor were performed according to the

P. Fossa et al. / Bioorg. Med. Chem. 11 (2003) 4749–4759
4759
3. Kikura, M.; Levy, J. H. Int. Anestesiol. Clin. 1995, 33, 21.
4. Kishi, T.; Nakahashi, K.; Ito, H.; Taniguchi, S.; Takaki,
M. Clin. Exp. Pharmacol. Physiol. 2001, 28, 737.
5. Moos, W. H.; Humblet, C. C.; Sircar, I.; Rithner, C.;
Weishaar, R. E.; Bristol, J. A.; McPhail, A. T. J. Med. Chem.
1987, 30, 1963.
6. Dorigo, P.; Maragno, I. Br. J. Pharm. 1986, 87, 623.
7. Mosti, L.; Schenone, P.; Menozzi, G. J. Heterocyclic Chem.
1985, 22, 1503.
8. Lo Presti, E.; Boggia, R.; Feltrin, A.; Menozzi, G.; Dorigo,
P.; Mosti, L. Il Farmaco 1999, 54, 465.
9. Dorigo, P.; Gaion, R. M.; Belluco, P.; Fraccarollo, D.;
Maragno, I.; Bombieri, G.; Benetollo, F.; Mosti, L.; Orsini, F.
J. Med. Chem. 1993, 36, 2475.
19. Floreani, M.; Fossa, P.; Gessi, S.; Mosti, L.; Borea, P. A.;
Dorigo, P. Naunyn-Schmiedeberg’s Arch. Pharmacol. 2003,
367, 109.
20. Cheng, Y. C.; Prusoff, W. H. Biochem. Pharmacol. 1973,
22, 3099.
21. Alousi, A. A.; Canter, J. M.; Montanaro, M. J.; Fort,
D. J.; Ferrari, R. A. J. Cardiovasc. Pharmacol. 1983, 5,
792.
22. Keef, K. D.; Hume, J. R.; Zhong, J. Am. J. Physiol. Cell
Physiol. 2001, 281, C1743.
23. Zhang, W.; Ke, H.; Tretiakova, A. P.; Jameson, B.; Col-
man, R. W. Protein Sci. 2001, 10, 1481, and references cited
therein..
24. Fredholm, B. B.; Ijzerman, A. P.; Jacobson, K. A.; Klotz,
K.-N.; Linden, J. Pharmacol. Rev. 2001, 53, 527.
25. Weishaar, R. E.; Burrows, S. D.; Kobylarz, D. C.; Quade,
M. M.; Evans, D. B. Biochem. Pharmacol. 1986, 35, 787.
26. Lowry, O. H.; Rosebrough, N. J.; Farra, A. L.; Randall,
R. L. J. Biol. Chem. 1951, 193, 265.
10. Dorigo, P.; Gaion, R. M.; Belluco, P.; Borea, P. A.;
Guerra, L.; Mosti, L.; Floreani, M.; Carpenedo, F. Gen.
Pharmacol. 1992, 23, 535.
11. Fossa, P.; Giordanetto, F.; Menozzi, G.; Mosti, L. Quant.
Struct. Act. Relat. 2002, 21, 267.
12. (a) Boggia, R.; Forina, M.; Fossa, P.; Mosti, L. Quant.
Struct. Act. Relat. 1997, 16, 201. (b) Altomare, C.; Cellamare,
S.; Summo, L.; Fossa, P.; Mosti, L.; Carotti, A. Bioorg. Med.
Chem. 2000, 8, 909.
27. Thompson, W. J.; Brooker, G.; Appleman, M. M. Meth.
Enzymol. 1974, 38, 205.
28. Slaughter, R. S.; Sutko, J. L.; Reeves, J. P. J. Biol. Chem.
1983, 258, 3183.
13. Hidaka, H.; Endo, T. Adv. Cyclic Nucleotide Protein
Phosphoryl. Res. 1984, 16, 245.
14. Giordanetto, F.; Fossa, P.; Menozzi, G.; Schenone, S.;
Bondavalli, F.; Ranise, A.; Mosti, L. J. Comput. Aided Mol.
Des. 2003, 17, 39.
29. Floreani, M.; Forlin, A.; Pandolfo, L.; Petrone, M.; Bel-
lin, S. J. Pharmacol. Exp. Ther. 1996, 278, 763.
30. Jones, L. R.; Besch, H. R., Jr.; Watanabe, A. M. J. Biol.
Chem. 1977, 252, 3315.
31. Floreani, M.; Santi Soncin, E.; Carpenedo, F. Naunyn-
Schmiedeberg’s Arch. Pharmacol. 1989, 339, 448.
32. MacroModel version 7.0, Schrodinger LLC. Mohamadi,
¨
15. Schenone, P.; Mosti, L.; Menozzi, G. J. Heterocyclic
Chem. 1982, 19, 1355.
16. Baker, S. R.; Cromble, L.; Dove, R. V.; Slack, D. A. J.
Chem. Soc. Perkin Trans. 1 1979, 677.
17. Endoh, M. Jpn. J. Pharmacol. 1979, 29, 855.
18. Endoh, M. J. Cardiovasc. Pharmacol. 2002, 40, 323.
F. Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J.
Comput. Chem. 1990, 11, 440.
33. Spartan ’O2; Wavefunction Inc.: Irvine, CA, USA