Structure-activity relationship investigations of the modulating effect of core substituents on the affinity of pyrazoloquinolinone congeners for the benzodiazepine receptor

Article abstract of DOI:10.1016/S0014-827X(98)00072-X

A series of 6- and 7-substituted-2-arylpyrazolo [4,3-c] quinolin-3-ones was synthesized and tested in vitro for binding with the benzodiazepine receptor in competition with [³H]flunitrazepam. Electronic parameters (molecular electrostatic potential (MEP), charge distribution on the nitrogen atoms, dipole moment μ, and ionization potential (IP) were calculated for the compounds by semi-empirical quantum chemistry methods. Lipophilicity of the compounds, expressed as logarithm of the octanol-water partition coefficient (log P), was calculated by the program Pallas. A quantitative correlation of the biological data with molecular parameters revealed a significant dependence (r = 0.95) of the activity on hydrophobic constants of the substituents, log P, and magnitude of the MEP minimum associated with the carbonyl oxygen atom.

Full text of DOI:10.1016/S0014-827X(98)00072-X

Il Farmaco 53 (1998) 579±585
Structure±activity relationship investigations of the modulating effect of
core substituents on the af®nity of pyrazoloquinolinone congeners for the
benzodiazepine receptor
b,
a
b
Janina Karolak-Wojciechowskaa, Jerzy Lange , Waldemar KsiaÎzek , Maøgorzata Gniewosz ,
*
Ç
Søawomir Rumpc
a
 Â
Institute of General and Ecological Chemistry, Technical University, PL-90924 èodz, Poland
bChemistry Department, University of Technology, PL-00664 Warsaw, Poland
cMilitary Institute for Hygiene and Epidemiology, PL-01163 Warsaw, Poland
Received 26 June 1998; accepted 20 October 1998
Abstract
A series of 6- and 7-substituted-2-arylpyrazolo[4,3-c]quinolin-3-ones was synthesized and tested in vitro for binding with the benzodia-
zepine receptor in competition with [3H]¯unitrazepam. Electronic parameters (molecular electrostatic potential (MEP), charge distribution
on the nitrogen atoms, dipole moment m, and ionization potential (IP)) were calculated for the compounds by semi-empirical quantum
chemistry methods. Lipophilicity of the compounds, expressed as logarithm of the octanol±water partition coef®cient (log P), was calculated
by the program Pallas. A quantitative correlation of the biological data with molecular parameters revealed a signi®cant dependence
(r ˆ 0:95) of the activity on hydrophobic constants of the substituents, log P, and magnitude of the MEP minimum associated with the
carbonyl oxygen atom. q 1998 Elsevier Science S.A. All rights reserved.
Keywords: Pyrazoloquinolinones; Benzodiazepine receptor; QSAR; Substituent effects
1. Introduction
to attempt to delineate different models of the pharmaco-
phore for the BzR binding site [1±10]. In almost all of them,
the pyrazolo[4,3-c]quinolin-3-one family ligands were
considered as a subject for study and discussion.
Many members of this family have been synthesized and
pharmacologically tested to date [11±15]. However, only
three top activity compounds derived from the original
Yokoyama's synthesis [11,12] and commonly known
under the CGS code notation arouse particular interest,
although none of them have gained new drug approval
(Fig. 1).
CGS-9895 and CGS-9896 exhibit strong anxiolytic and
anticonvulsant activity with an almost total lack of neuro-
logical de®cits and are therefore classi®ed as partial agonists
[16]. On the contrary, CGS-8216 is an inverse agonist which
reveals anxiogenic, somnolytic, and proconvulsant actions
[17]. Attempts were made to rationalize this rather unusual
effect of the substituent R in terms of either tautomeric
equilibria [1] or reversed relative lipophilicity of both the
pyridine-fused benzene ring and the 2-phenyl ring [13].
Either concept calls for assuming the possibility of an
inverted orientation of the molecules within the receptor
pocket [13]. The problem does not seem to be de®nitively
Benzodiazepines and other chemical families capable of
binding with high af®nity to the benzodiazepine receptor
(BzR) have held the attention of researchers in the ®elds
of medicine, pharmacology, and chemistry for several
years. Many are used as indispensable therapeutic agents
to treat speci®c diseases in the central nervous system
(CNS). Many more are still the subject of pharmacological
and/or clinical testing and their fate is in the balance. Since
BzR is located on the GABAA receptor channel, the major
inhibitory neurotransmitter system in the CNS, it is a matter
of utmost importance to gain a better knowledge of its struc-
ture at the molecular level and thus to reach a more precisely
outlined target for the design of new active compounds.
Numerous structure±activity relationship studies that
recently addressed this problem have revealed that many
BzR ligands, in spite of marked differences in the structure
of their heterocyclic cores, have some geometric, electronic,
and/or hydrophilic/lipophilic structural features in common.
This knowledge, although still limited, has made it possible
* Corresponding author; e-mail: lange@ch.pw.edu.pl
0014-827X/98/$ - see front matter q 1998 Elsevier Science S.A. All rights reserved.
PII S0014-827X(98)00072-X

580
J. Karolak-Wojciechowska et al. / Il Farmaco 53 (1998) 579±585
Fig. 1. The known pyrazolo[4,3-c]quinolin-3-ones with top af®nity for the
benzodiazepine receptor.
settled, however. Even so, the cyclic substituent at N(2) is
generally recognized as the crucial determinant of both the
af®nity for the receptor and the activity pro®le and all
research programs are consequently on that basis. Much
less is known about the effects of substituents in the benzene
fragment of the heterocyclic core.
Scheme 1.
We have previously described [18] a quantitative rela-
tionship between the structure and in vitro af®nity for the
BzR in another group of ligands, namely 6-aryltriazolo[4,3-
b]pyridazines substituted in the phenyl and triazole
moieties. The excellent correlation obtained involved the
Hansch constants for both substituents and the ionization
potential of the molecule as the most signi®cant components
of the correlation equation. This means that in compounds
with roughly the same overall geometry and almost exactly
the same localization of the anchoring sites, the binding
capacity greatly depends on modulation of the electronic
properties of the molecules by the substituents.
In the present research we have synthesized a series of
pyrazolo[4,3-c]quinolin-3-ones carrying various 7- or 8-
substituents (Rl and R2, respectively); some also have p-
substituents (R3) in the 2-phenyl moiety (Table 1). All the
compounds have been subsequently tested for their af®nity
for the benzodiazepine receptor by determining in vitro the
inhibition of [3H]¯unitrazepam binding in rat brain homo-
genates. The results have been quantitatively correlated
with molecular hydrophobic and electronic parameters.
Table 1
Pyrazoloquinolinonesa
2. Chemistry
The synthetic route, which is shown in Scheme 1,
followed in general that used by Yokoyama [11,12]. Thus,
condensation of diethyl ethoxymethylenemalonate with an
appropriately substituted aniline yielded the corresponding
diethyl [(arylamino)methylene]malonates (1), most of
which have been synthesized previously. Compound 1
cyclized at , 2508C to the corresponding ethyl 7- (or 8-)
substituted-4-hydroxyquinoline-3-carboxylates (2). Subse-
quent replacement of the hydroxy function with chlorine
by treatment with phosphorus oxychloride yielded the
corresponding ethyl 7- (or 8-) substituted-4-chloroquino-
line-3-carboxylates (3) which, in the reaction with an appro-
priately substituted phenylhydrazine, gave the ®nal products
(4±7) (Scheme 1 and Table 1). In order to secure a maxi-
mum comparability of the biological results, three known
pyrazoloquinolines (CGS-8216, 4d, and 4e) were also
resynthesized as reference compounds and included in the
SAR and QSAR analysis and discussion.
No.
R1
R2
R3
M.p. (8C)b
Yield (%)
4a
4b
4c
4d
4e
4f
OC2H5
OCH3
Cl
H
H
289±292
300±301
389±392
396±400c
317±320d
368±369
364±366
360±362
334±336
389±393
356±358
352±356
305±308
344±347
57
63
72
65
78
84
68
61
74
87
70
66
54
73
H
H
H
H
H
Cl
OCH3
H
H
H
H
CF3
H
H
4g
4h
5b
5c
5f
CH3
H
H
CH3
OCH3
Cl
H
H
Cl
Cl
Cl
Cl
OCH3
F
H
CF3
CH3
CF3
CF3
H
5h
6f
H
H
7f
H
3. Results and discussion
a Analyses C, H, N for all compounds.
b All melting points with decomposition.
c Literature [13] m.p. . 4008 (dec.).
d Literature [13] m.p. 322±3258C.
The recent knowledge of the ligand BzR interaction
assumes for certain that at least two sites on the ligand
molecule are required to act as acceptors of protons from

J. Karolak-Wojciechowska et al. / Il Farmaco 53 (1998) 579±585
581
Fig. 2. Possible tautomeric forms of the three well known 2-arylpyrazolo[4,3-c]quinolin-3 ones and their calculated heat of formation. Arrows point to atoms
presumed to act as the proton acceptor sites.
the receptor protein [9]. However, the compounds investi-
gated may theoretically exist in three tautomeric forms, the
prospective binding sites in each being associated with
different atom pairs (Fig. 2). In order to estimate which
tautomer is thermodynamically preferred and therefore
most likely to exist in an aqueous phase, the structures of
the particular tautomeric forms of the three CGS ligands
were optimized by molecule modeling and the values of
the heats of formation (Hf) were calculated for each of
them. The calculations, carried out by using the semi-
empirical method AM1 from the MOPAC-93 program
[19], referred to geometry optimization in an aqueous
medium so the dielectric constant of water (78.4) was
included. As seen in Fig. 2, the tautomer A of the investi-
gated molecules is thermodynamically preferred; it was
assumed that the same preference was true for other conge-
neric compounds. Moreover, the choice of tautomer A is
consistent with opinions of other authors working in the
domain of BzR ligands [9,13±15].
All the synthesized compounds (cf. Table 1) were subject
to molecular modeling, geometry optimization, and calcula-
tion of the molecular electrostatic potential (MEP) distribu-
tion. An illustrative example of MEP distribution
(compound 4c) is shown in Fig. 3. In every case, a deep
MEP minimum, marked as M1, was observed in the vicinity
of the carbonyl oxygen atom which is one of the potential
acceptors of hydrogen from the receptor protein. Another
minimum (M2) is associated with the N(1) atom; it is much
shallower and in some cases accompanied by an additional
M3 minimum. The Ml minimum is located almost exactly in
the plane of the tricyclic core, whereas M2 and M3 are to be
found above or below this plane but always in close vicinity
of the nitrogen atom. A slight de¯ection of the aryl substi-
tuent from the plane of the core seems to be the most plau-
sible explanation of the observed splitting of the M2
minimum into the M2±M3 pair. The distance between the
Ê
M1 and M2 minima, which amounts to approximately 7.5 A,
is an almost perfect match for that between their equivalents
(points marked H1 and H2) in the known models of the
benzodiazepine receptor [9,10].
The results of calculations considering the depths of the
M1 MEP minimum are given in Table 2. The M2 and M3
minima were considered of no importance to the qualitative
discussion and consequently not included in this table.
Determination of binding with BzR was carried out in
vitro by a standard procedure using [3H]¯unitrazepam as
the labeled ligand [20,21]. The results, expressed as %
competitive inhibition (I) of [3H]¯unitrazepam (1.5 nM)
binding by the investigated compounds (1 £ 1027 M con-
centration), are given in Table 2.
Correlation of the binding capacity (I) with the calculated
electronic parameter revealed a highly signi®cant depen-
dence on the depth of the M1 MEP minimum (kcal/mol)
as shown by the 2nd order equation (Eq. (1)).
log I ˆ 0:0055M12 2 0:2202M1 2 0:3089 ˆ log IPC
ꢀ1†
n ˆ 15; r ˆ 0:922; F ˆ 6:93; p ˆ 0:01; s ˆ 0:0483
The log I values calculated with the aid of this equation for
each individual compound of the series investigated are
Fig. 3. MEP distribution in 7-chloro-2-phenylpyrazolo[4,3-c]quinolin-3-
one (4c). M1 and M2 denote two MEP minima.

582
J. Karolak-Wojciechowska et al. / Il Farmaco 53 (1998) 579±585
Table 2
Calculated electronic parameters and partition coef®cients used in the correlation regressions and experimental data of af®nity for the benzodiazepine receptor
confronted with the calculated values
Compound
Hydrophobic constants p for substituents
MEP min. M1
(kcal/mol)
log P
Inhibition of [3H]¯unitrazepam
(1.5 nM) binding (%)
R1
0
R2
R3
Determined
Calculateda
CGS-8216
0
0
0
0
0
0
0
0
0
0
0
0
0
17.79
19.46
18.14
16.24
17.80
19.05
15.54
19.26
19.19
17.20
14.44
11.83
15.81
15.07
13.42
2.05
2.46
2.00
2.77
2.68
2.13
3.21
2.61
2.48
2.73
3.50
3.93
3.21
3.16
3.38
77.8
72.9
79.3
77.0
81.4
80.3
58.6
78.8
74.1
62.2
58.9
33.2
48.4
78.7
45.6
77.98
77.09
77.98
73.45
80.72
77.09
61.52
79.62
78.89
62.66
46.99
34.83
54.96
63.24
55.59
4a
4b
4c
4d
4e
4f
0.38
±0.02
0.71
0
0.71
0
±0.02
0.88
0
0
4g
4h
5b
5c
5f
0.56
0
0.56
±0.02
0.71
0.88
0.56
0.88
0.88
0
0.71
0
0.71
0.71
0
5h
6f
0
0.71
0
±0.02
0.14
7f
0
a Eq. (5) was used to calculate the inhibition data in this column.
used in the following correlation approaches as the principal
component data. They are denoted there as log IPC. Attempts
at correlating the binding ef®cacy with the calculated
HOMO and LUMO energies and with the dipole moment
values were unsuccessful. Analogously, positive correla-
tions with the Hammett and Swain±Lupton constants failed.
As shown in Table 2, the depth of the M1 MEP minimum
depends on the substituents and thus indirectly re¯ects their
effects on the binding capacity (I or log I) of the compounds.
A straightforward estimation of the substituent effects was
obtained by using the classical Hansch method. In the ®rst
approximation, the subset of nine compounds with R3 ˆ H
(p ˆ 0) was investigated. With the standard values [22,23]
of the Hansch hydrophobic constants p (Table 2), the
following equation was generated.
nent is negative in this equation, R3 substituents with nega-
tive values of p (OCH3) should be expected to improve
af®nity for the receptor. Our investigations involved
compounds with four different R3 and, in fact, their af®nity
decreased in the following order of R3: OCH3 . H . F .
Cl. The af®nity of compounds with R1
ˆ
CF3, which
decreased in the order 6f . 4f . 7f . 5f, correlates
therefore very well with increasing p values of their R3
substituents: 2 0.02 , 0 , 0.14 , 0.71, respectively.
It is well known that the overall lipophilicity is one of the
most important molecular parameters to be used in QSAR
studies. The lipophilicity of the compounds investigated
was calculated as the logarithm of the octanol±water parti-
tion coef®cient (log P) by using the program Pallas [24].
The respective values are included in Table 2. Correlation
of log P with the experimental binding data (I) resulted in
the following equation
log I ˆ 20:0934pR1 1 0:0035pR2 1 1:9021
ꢀ2†
log I ˆ 20:1210 log P2 1 0:5324 log P 1 1:3114
ꢀ4†
n ˆ 9; r ˆ 0:781; F ˆ 10:2; p ˆ 0:01; s ˆ 0:0315
This shows that the af®nity of the investigated compounds
depends on both R1 and R2 substituents, although R2 with its
positive weight has a more profound effect. This can be seen
in Table 2 where 4e is shown as more active than 4b, 4d than
4c, and 4g than 4h. This conclusion is consistent with the
literature [13].
n ˆ 15; r ˆ 0:833; F ˆ 693; p ˆ 0:01; s ˆ 0:0583
In accord with this correlation, log P of the compounds
exhibiting highest af®nity for the receptor (log I
ù
2)
should fall into the range of 2±3. The highest value of
log P (3.93) was calculated for the compound with lowest
af®nity and with two highly lipophilic substituents (R ˆ
CF3 and R1 ˆ Cl). In the subset with R ˆ CF3, all
compounds are characterized by log P . 3. Although
acceptable from the point of view of statistics, the parabolic
relationship expressed by Eq. (4) may be misleading since,
among other reasons, compounds with a rather widely
different log P (e.g., in the range between 2 and 3.5) may
A similar regression with all 15 compounds gave
log I ˆ 20:1403pR1 2 0:0244pR2 2 0:2241pR3 1 1:9204
ꢀ3†
n ˆ 15; r ˆ 0:841; F ˆ 6:22; p ˆ 0:01; s ˆ 0:0703
This substantiates the conclusion derived from Eq. (2) that
the effect of R2 outweighs that of R1. Since the R3 compo-

J. Karolak-Wojciechowska et al. / Il Farmaco 53 (1998) 579±585
583
be closely similar in activity (log I). The linear approxima-
tion is statistically much inferior, however.
The regression expressed by Eq. (3) was ®nally supple-
mented by including log I values calculated for each indi-
vidual compound according to Eq. (1) and denoted in the
following equation as log IPC.
malonate (10.8 g, 0.05 mol) were heated for 2 h at 110±
1158C in an open ¯ask. The slightly yellowish mixture was
treated with charcoal and ®ltered while hot. The ®ltrate was
left standing for several days at room temperature and
slowly crystallized to yield the crude product which was
puri®ed by recrystallization from light petroleum (a 45±
608C fraction) or hexane.
Diethyl 2-[3-ethoxyphenylamino)methylene]malonate
(1a), m.p. 39±408C (light petroleum), yield 84%. Anal.:
Calc. For C16H21NO5: C, 62.53; H, 6.89; N, 4.56. Found:
C, 62.32; H, 6.78; N, 4.47%.
log I ˆ 20:022pR1 1 0:021pR2 2 0:097pR3
1 0:832 log IPC 1 0:451
ꢀ5†
n ˆ 15;
r ˆ 0:949; F ˆ 5:99;
s ˆ 0:0428
Diethyl 2-[(3-methoxyphenylamino)methylene]malonate
(1b), m.p. 43±458C (light petroleum), yield 79%. Anal.:
Calc. for C15H19NO5: C, 61.42; H, 6.53; N, 4.78. Found:
C, 61.32; H, 6.50; N, 4.73%.
This ®nal correlation combines with high statistical signi®-
cance the hydrophobic constants p of all three variable
substituents with the depth of the main molecular electro-
static potential minimum Ml.
Compounds 1c±1h were prepared analogously. Their
melting points were consistent with the literature data.
1
Representative NMR spectra are reported for 1a. H NMR
The results of the present QSAR analysis indicate that the
af®nity of the pyrazoloquinolinones investigated for the
benzodiazepine receptor depends not only on the total
hydrophobicity of the substituents but also on their position
in the bicyclic core. Thus, for the same substituent, the 8-
substituted compound is more active than the 7-substituted.
In compounds with the same nature and position of the core
substituents, the af®nity for the receptor is de®ned by the
para substituent in the 2-phenyl group and decreases with
increasing hydrophobicity of this substituent.
The overall effect of all substituents is most precisely
illustrated by the distribution of the MEP (Fig. 3). The M1
minimum of MEP close to the carbonyl oxygen atom acts as
a proton acceptor to produce a hydrogen bond which
anchors the ligand to the receptor protein. The depth of
this minimum is highly representative of the degree of af®-
nity.
(400 MHz, CDCl3): 1.33 (t, 3H, J ˆ 7:1 Hz, ether CH3),
1.37 (t, 3H, J ˆ 7:1 Hz, ester CH3), 1.41 (t, 3H, J ˆ 7:1
Hz, ester CH3), 4.01 (q, 2H, J ˆ 7:1 Hz, ether CH2), 4.25
(q, 2H, J ˆ 7:1 Hz, ester CH2), 4.30 (q, 2H, J ˆ 7:1 Hz,
ester CH2), 6.63±6.70 (m, 3H, arom. 4-, 5- and 6-CH), 7.21±
7.25 (m, 1H, arom. 2-CH), 8.49 (d, 1H, J ˆ 13:6 Hz, yCH),
10.96 (d, 1H, J ˆ 13:6, NH). 13C NMR (400 MHz, CDCl3):
14.32 and 14.75 (ester CH3), 14.46 (ether CH3), 60.03 and
60.32 (ester CH2), 63.63 (ether CH2), 93.58 (yCr), 103.74,
109.29, and 110.75 (arom. 4-, 5- and 6-CH), 130.90 (arom.
2-CH), 140.45 (arom. CNH), 151.74 (yCH), 160.26 (arom.
COEt), 165.65 and 168.98 (2 £ CyO).
4.1.3. 7-(6)-Substituted ethyl 4-hydroxyquinoline-3-
carboxylates (2)
The general procedure was as follows. Diethyl [(arylami-
no)methylene]malonate (0.03 mol) was rapidly added in one
portion with stirring to 75 ml of Dowtherm A preheated in
an open ¯ask under an argon blanket to 2508C. The mixture
was kept at this temperature for 20±30 min. Upon cooling,
100 ml of hexane were added to complete precipitation of
the product. Filtration yielded a tan product which was puri-
®ed by recrystallization from acetic acid or methoxyethanol
to give an off-white amorphous solid.
Ethyl 7-ethoxy-4-hydroxyquinoline-3-carboxylate (2a),
m.p. 280±2828C (AcOH), yield 78%. Anal.: Calc. for
C14H15NO4: C, 64.36; H, 5.79; N, 5.36. Found: C, 64.44;
H, 5.70; N, 5.28%.
4. Experimental
4.1. Chemistry
4.1.1. General methods
Melting points were determined in a BuÈchi apparatus and
are reported uncorrected. The 1H and 13C NMR spectra were
taken with Varian 200 and 300 MHz and Brucker 400 MHz
instruments with TMS as internal standard. Microanalyses
were carried out by Mrs E. Godzisz, Warsaw University of
Technology, on a Perkin Elmer C-H-N analyzer and were
within ^ 0.4% of the calculated values. Merck DC-Plas-
tikfolien with Kieselgel 60 were used in the purity checking;
the CHCl3/MeOH/AcOEt/satd.aq. NH3 (3:2:1:0.1) develop-
ing system was used. The reported melting points and
elemental analyses refer to recrystallized, chromatographi-
cally homogeneous compounds.
Compounds 2b±h were prepared analogously. Their
melting points were consistent with the literature data.
1
Representative NMR spectra are reported for 2a: H NMR
(300 MHz, CF3COOD): 1.45 (t, 3H, J ˆ 7:1 Hz, CH3), 1.49
(t, 3H, J ˆ 7:0 Hz, CH3), 4.25 (q, 2H, J ˆ 7:0 Hz, CH2),
4.58 (q, 2H, J ˆ 7:1 Hz, CH2), 7.33 (d, 1H, J ˆ 2:3 Hz, 8-
CH), 7.46 (dd, 1H, J6;8 ˆ 2:3 Hz, J6;5 ˆ 9:3 Hz, 6-CH), 8.44
(d, 1H, J ˆ 9:3 Hz, 5-CH), 9.10 (d, 1H, J ˆ 7:3 Hz, 2-CH).
13C NMR (400 MHz, CF3COOH-external acetone-d6
lock): 12.3 and 12.6 (2 £ CH3), 64.5 and 65.7 (2 £ CH2),
4.1.2. Diethyl 2-[(arylamino)methylene]malonates (1)
The general procedure was as follows. The appropriate
aniline derivative (0.05 mol) and diethyl ethoxymethylene-

584
J. Karolak-Wojciechowska et al. / Il Farmaco 53 (1998) 579±585
100.0 (8-C), 104.0, 114.2, 122.5 (6-C), 126.5 (5-C), 142.8
144.7 (2-C), 167.6, 168.1, 172.3; chemical shift data without
assignments refer to quaternary aromatic carbon atoms.
was ®ltered, thoroughly washed with ether, and dissolved in
15 ml of 2 N NaOH. The solution was stirred for 1 h with
ethyl ether, the ether solution was separated and discarded,
whereas the aqueous solution was neutralized to pH 8±9 by
adding solid NH4Cl to precipitate the crude product. The
®nal puri®cation was done by recrystallization from ethanol.
Anal.: 4a. Calc. for C18H15N3O2: C, 70.81; H, 4.95; N,
13.76. Found: C, 70.62; H, 4.89; N, 13.80%. 4b. Calc. for
C17H13N3O2: C, 70.09; H, 4.50; N, 14.42. Found: C, 70.23;
H, 4.37; N, 14.38%. 4c. Calc. for C16H10ClN3O: C, 64.98; H,
3.41; N, 14.21. Found: C, 65.11; H, 3.28; N, 14.12%. 4f.
Calc. for C17H10F3N3O: C, 62.01; H, 3.06; N, 12.76. Found:
C, 61.87; H, 2.96; N, 12.70%. 4g. Calc. for C17H13N3O: C,
74.17; H, 4.76; N, 15.26. Found: C, 74.05; H, 4.84; N,
15.19%. 4h. C17H13N3O: C, 74.17; H, 4.76; N, 15.26.
Found: C, 74.27; H, 4.70; N, 15.24%.
4.1.4. 7-(6)-Substituted ethyl 4-chloroquinoline-3-
carboxylates (3)
The general procedure was as follows. The appropriately
6- or 7-substituted ethyl 4-hydroxyquinoline-3-carboxylate
(2) (0.025 mol) was cautiously added to ice-cooled phos-
phorus oxychloride (17.5 ml) and the mixture was re¯uxed
under argon for 5 h. Upon cooling, the mixture was poured
onto crushed ice, neutralized with ammonia, and extracted
with methylene chloride as rapidly as possible. The organic
layer was washed with saline, dried with MgSO4, and
evaporated to yield the crude product which was puri®ed
by recrystallization from petroleum ether or hexane.
Ethyl 4-chloro-7-ethoxyquinoline-3-carboxylate (3a),
m.p. 120±1228C (petroleum ether, 45±608C fraction),
yield 60%. Anal.: Calc. for C14H14ClNO3: C, 60.11; H,
5.04; N, 5.01. Found: C, 60.17; H, 4.92; N, 4.98%.
Ethyl 4-chloro-7-methoxyquinoline-3-carboxylate (3b),
m.p. 130±1328C (hexane), yield 67%. Anal.: Calc. for
(Cl3H12ClNO3): C, 58.77; H, 4.55; N, 5.27. Found: C,
58.85; H, 4.48; N, 5.20%.
Ethyl 4-chloro-6-methoxyquinoline-3-carboxylate (3e),
m.p. 96±988C (petroleum ether), yield 54%. Anal.: Calc.
for (C13H12ClNO3): C, 58.77; H, 4.55; N, 5.27. Found: C,
58.88; H, 4.39; N, 5.18%.
Ethyl 4-chloro-6-methylquinoline-3-carboxylate (3g),
m.p. 64±668C (petroleum ether), yield 71%. Anal.: Calc.
for (C13H12ClNO2): C, 62.53; H, 4.84; N, 5.61. Found: C,
62.44; H, 4.78; N, 5.58%.
4.1.6. 7-(6)-Substituted 2-arylpyrazolo[3,2-c]quinolin-3-
ones (5±7)
The general procedure was as follows. 4-Chlorophenyl-
hydrazine hydrochloride (0.005 mol) was suspended in 60
ml of xylene and the mixture was stirred for 2 h with 5 ml of
aqueous ammonia. The xylene layer was separated and
distilled until moisture stopped passing with the distillate.
The appropriately substituted ethyl 4-chloroquinoline-3-
carboxylate (0.003 mol) was then added; the reaction
continued as above to yield the corresponding 5.
An analogous procedure starting with 4-methoxyphenyl-
hydrazine and 4-¯uorophenylhydrazine was used to prepare
6 and 7, respectively.
Data for compounds 5±7 are given in Table 1.
1
The representative H NMR spectrum is reported for 5a
Compounds 3c, 3d, 3f, and 3h were prepared analo-
gously. Their melting points were consistent with the litera-
ture data.
(400 MHz, DMSO-d6): 1.50 (t, 3H, CH3), 4.17 (q, 2H, CH2),
7.41 (dd, 1H, J8;6 ˆ 2:1 Hz, J8;9 ˆ 9:4 Hz, 8-CH), 7.5±7.8
(m, 6H), 8.17 (d, 1H, J ˆ 9:4 Hz, 9-CH), 8.97 (s, 1H, 4-
CH).
Representative NMR spectra are reported for 3a: 1H
NMR (400 MHz, CDCl3): 1.46 (t, 3H, J ˆ 7:0, ether
CH3), 1.52 (t, 3H, J ˆ 7:1, ester CH3), 4.22 (q, 2H,
J ˆ 7:0, ether CH2), 4.48 (q, 2H, J ˆ 7:1, ester CH2), 7.32
(dd, 1H, J6;8 ˆ 2:3 Hz, J6;5 ˆ 9:3 Hz, 6-CH), 7.41 (d, 1H,
J ˆ 2:3 Hz, 8-CH), 8.29 (d, 1H, J ˆ 9:3 Hz, 5-CH), 9.16 (s,
1H, 2-CH).
Anal.: 5b Calc. for C17H12ClN3O2: C, 62.68; H, 3.71; N,
12.90. Found: C, 62.49; H, 3.67; N, 12.85%. 5c. Calc. for
C16H9Cl2N3O: C, 58.20; H, 2.75; N, 12.73. Found: C, 58.02;
H, 2.81; N, 21.37%. 5f. Calc. for C17H9ClF3N3O: C, 56.14;
H, 2.49; N, 11.55. Found: C, 55.97; H, 2.38; N, 11.60%. 5h.
Calc. for C17H12ClN3O: C, 65.92; H, 3.90; N, 13.57. Found:
C, 66.09; H, 3.78; N, 13.48%. 6f. Calc. for C18H12F3N3O2: C,
60.17; H, 3.37; N, 11.69. Found: C, 59.94; H, 3.28; N,
11.65%. 7f. Calc. for C17H9F4N3O: C, 58.80; H, 2.61; N,
12.10. Found: C, 58.64; H, 2.53; N, 12.19%.
13C NMR (400 MHz, CDCl3): 14.6 and 14.3 (2 £ CH3),
64.2 and 61.8 (2 £ CH2), 108.1 (8-C), 121.1, 120.5, 121.8
(6-C), 126.7 (5-C), 143.5, 150.9 (2-C), 151.7, 162.0, 164.6
(CyO); chemical shift data without assignments refer to
quaternary aromatic carbon atoms.
4.2. Receptor tests
4.1.5. 7-(6)-Substituted 2-phenylpyrazolo[3,2-c]quinolin-3-
ones (4)
Rat brains were homogenized at 08C in 20 vols. of 50 mM
Tris±HCl (pH 7.4) and the homogenates were incubated for
1 h at 378C and centrifuged at 20 000 £ g. Samples of the
homogenate (800 ml corresponding to 13.3 mg of the brain
tissue) were mixed with 1000 ml of [3H]¯unitrazepam
(speci®c activity 81 Ci/mM) and 100 ml of the tested
compound (1027 M). The mixture was incubated for 2 h
The general procedure was as follows. The appropriately
6- or 7-substituted 4-chloroquinoline-3-carboxylate (3)
(0.0036 mol) and 0.45 g (0.0045 mol) of phenylhydrazine
was re¯uxed in 20 ml of xylene for 4 h. A bright yellow
product quickly began to precipitate. Ethyl ether was added
to the cooled mixture to complete precipitation and the solid

J. Karolak-Wojciechowska et al. / Il Farmaco 53 (1998) 579±585
585
[9] W. Zhang, K.F. Koehler, P. Zhang, J.M. Cook, Development of a
comprehensive pharmacophore model for the benzodiazepine recep-
tor, Drug Des. Discov. 12 (1995) 193±248.
at 58C and ®ltered through a glass ®ber ®lter (Whatman GF/
C). The ®lter was washed twice with 5-ml portions of Tris,
dried, and immersed in 10 ml of the scintillation liquid
(POPOP 50 mg, PPO 4 g, methanol 20 ml, and toluene
1000 ml). Radioactivity was measured in a Betamatic II
Kontron b-scintillation counter and expressed as % inhibi-
tion of binding of the labeled ¯unitrazepam.
[10] H.O. Villar, M.F. Davies, G.H. Loew, P.A. Maguire, Molecular
models for recognition and activation at the benzodiazepine receptor:
a review, Life Sci. 48 (1991) 593±602.
[11] N. Yokoyama, B. Ritter, A.D. Neubert, 2-Arylpyrazolo[4,3c]quino-
lin-3-ones: novel agonist, partial agonist and antagonist of benzodia-
zepines, J. Med. Chem. 25 (1982) 337±339.
[12] N. Yokoyama, Pyrazoloquinolines, US Patent 4 312 870 (1982);
Chem. Abstr. 96 (1982) 181283v.
Acknowledgements
[13] R.I. Fryer, P. Zhang, R. Rios, Z.Q. Gu, A.S. Basile, P. Skolnick,
Structure±activity relationship studies at the benzodiazepine receptor
(BzR): comparison of the substituent effects of pyrazoloquinolinone
analogs, J. Med. Chem. 36 (1993) 1669±1673.
The research was carried out under grant nos. 4 4404 91
02 and 4 P05F 013 09 from the Committee of Scienti®c
Research, Poland. This ®nancial support is gratefully
acknowledged.
[14] S. Takada, H. Shindo, T. Sasatani, A. Matsushita, M. Eigyo, K. Kawa-
saki, S. Murata, A new thienylpyrazoloquinoline: a potent and orally
active inverse agonist to benzodiazepine receptors, J. Med. Chem. 30
(1987) 454±455.
[15] H. Shindo, S. Takada, S. Murata, M. Eigyo, A. Matsuhita, Thienyl-
pyrazoloquinolines with high af®nity to the benzodiazepine receptors:
continuous shift from inverse agonist to agonist properties depending
on the size of the alkyl substituent, J. Med. Chem. 32 (1989) 1213±
1217.
References
[1] H.O. Villar, E.T. Uyeno, L. Toll, W. Polgar, M.F. Davies, G.H. Loew,
Molecular determinants of benzodiazepine receptor af®nities, Mol.
Pharmacol. 36 (1989) 589±600.
[16] D.A. Bennett, Pharmacology of the pyrazolo-type compounds:
agonist, antagonist and inverse agonist actions, Physiol. Behav. 41
(1987) 241±245.
[2] G.M. Crippen, Distance geometry analysis of the benzodiazepine
binding site, Mol. Pharmacol. 22 (1982) 11±19.
[17] L.H. Jensen, E.N. Petersen, C. Braestrup, Audiogenic seizures in
DBA/2 mice discriminate sensitively between low ef®cacy benzodia-
zepine receptor agonists and inverse agonists, Life Sci. 33 (1983)
393±399.
[3] A.K. Ghose, G.M. Crippen, Modeling the benzodiazepine receptor
bonding site by the general three-dimensional structure-directed
quantitative structure±activity relationships method Remotedisc,
Mol. Pharmacol. 37 (1990) 725±734.
[18] J. Karolak-Wojciechowska, J. Lange, W. Kwiatkowski, M. Gniewosz,
J. Plenkiewicz, Bicyclic [b]-heteroannulated pyridazine derivatives.
II. Structure±activity relationships in the 6-aryltriazolo[4,3-b]pyrida-
zine ligands of the benzodiazepine receptor, Bioorg. Med. Chem. 2
(1994) 773±779.
[4] P.W. Codding, A.K.S. Muir, Molecular structure of Ro15-1788 and a
mode for the binding of benzodiazepine receptor ligands: structural
identi®cation of common features in antagonists, Mol. Pharmacol. 28
(1985) 178±184.
[5] R.I. Fryer, C. Cook, N.W. Gilman, A. Walser, Conformational shifts
at the benzodiazepine receptor related to the binding of agonists,
antagonists and inverse agonists, Life Sci. 39 (1986) 1947±1957.
[6] P.A. Borea, G. Gilli, V. Bertolasi, V. Ferretti, Stereochemical features
controlling binding and intrinsic activity properties of benzodiazepine
receptor ligands, Mol. Pharmacol. 31 (1987) 334±344.
[19] MOPAC (version 6.0), Quantum Chemistry Program Exchange No.
455, Bloomington, IN, 1990.
[20] R.F. Squires, C. Braestrup, Benzodiazepine receptors in rat brain,
Nature 266 (1977) 732±734.
È
[21] H. Mohler, T. Okada, Benzodiazepine receptor: demonstration in the
central nervous system, Science 198 (1977) 849±851.
[22] R.A. Coburn, QSAR-PC Programme, Biosoft, Cambridge, UK.
[23] C. Hansch, A. Leo, Substituent Constants for Correlation Analysis in
Chemistry and Biology, Wiley, New York, 1979.
[7] S. Tebib, J.J. Bourguignon, C.G. Wermuth, The active analog
approach applied to the pharmacophore identi®cation of benzodiaze-
pine receptor ligands, J. Comput.-Aided Mol. Des. 1 (1987) 153±170.
[8] C.R. Gardner, A review of recently-developed ligands for neuronal
benzodiazepine receptors and their pharmacological activities, Prog.
Neuro Psychopharmacol. Biol. Psychiatry 16 (1992) 755±781.
[24] PALLAS For Windows 1.2, CompuDrug Chemistry, Budapest,
Hungary, 1995.