Design and synthesis of imidazoline derivatives active on glucose homeostasis in a rat model of type II diabetes. 1. Synthesis and biological activities of N-benzyl-N'-(arylalkyl)-2-(4',5'-dihydro-1'H-imidazol-2'- yl)piperazines

Article abstract of DOI:10.1021/jm9608624

The physiopathology of non-insulin-dependent diabetes mellitus is associated with a dysfunction in the regulation of insulin secretion. The α₂-adrenoceptors have been reported to be involved in this alteration, although α₂-antagonist

Full text of DOI:10.1021/jm9608624

J . Med. Chem. 1997, 40, 3793-3803
3793
Design a n d Syn th esis of Im id a zolin e Der iva tives Active on Glu cose Hom eosta sis
in a Ra t Mod el of Typ e II Dia betes. 1. Syn th esis a n d Biologica l Activities of
N-Ben zyl-N ′-(a r yla lk yl)-2-(4′,5′-d ih yd r o-1′H-im id a zol-2′-yl)p ip er a zin es
Fre´de´ric Rondu,§,∇ Gae¨lle Le Bihan,§,∇ Xuan Wang,§,‡,∇ Aazdine Lamouri,§ Estera Touboul,§ Georges Dive,|
Toune`s Bellahsene,§ Bruno Pfeiffer,† Pierre Renard,† Be´atrice Guardiola-Lemaitre,# Dominique Manechez,#
Luc Penicaud, Alain Ktorza,‡ and J ean-J acques Godfroid*,§
Laboratoire de Pharmacochimie Mole´culaire and Laboratoire de Physiopathologie de la Nutrition, CNRS URA 307, Universite´
Paris 7-Denis Diderot, 2, place J ussieu, 75251 Paris Cedex, France, ADIR, 1, rue Carle Hebert, 92415 Courbevoie Cedex,
France, IRI Servier, 6, place des Pleiades, 92415 Courbevoie Cedex, France, Centre d’Inge´nierie des Prote´ines, Institut de
Chimie, Universite´ de Lie`ge, Belgium, and UPRESA, CNRS 5018/ -UPS, Toulouse, France
Received December 26, 1996X
The physiopathology of non-insulin-dependent diabetes mellitus is associated with a dysfunction
in the regulation of insulin secretion. The R2-adrenoceptors have been reported to be involved
in this alteration, although R2-antagonists containing an imidazoline ring may stimulate insulin
secretion independently of R2-adrenoceptor blockage. Recently, a new “imidazoline-binding
site” involved in the control of K+-ATP channels in the B cell has been proposed. In the course
of searching for new antidiabetic agents, 1-alkyl-2-(4′,5′-dihydro-1′H-imidazol-2′-yl)-4-ben-
zylpiperazines, 1-benzyl-2-(4′,5′-dihydro-1′H-imidazol-2′-yl)-4-alkylpiperazines, and 1-benzyl-
2-(4′,5′-dihydro-1′H-imidazol-2′-yl)-4-benzylpiperazines have been designed and evaluated as
potential adrenoceptor antagonists. Pharmacological evaluation was performed in vivo using
glucose tolerance tests performed on a rat model of type II diabetes obtained by injection of a
low dose (35 mg/kg) of streptozotocin (STZ). For some compounds, binding experiments were
performed on R2 adrenoceptors and I1 and I2 imidazoline-binding sites. The biological and
physicochemical data have been combined with molecular modeling studies to establish
structure-activity relationships. The most active compound was 1-(2′,4′-dichlorobenzyl)-2-
(4′,5′-dihydro-1′H-imidazol-2′-yl)-4-methylpiperazine (7f); intraperitoneal administration (100
µmol/kg) of 7f strongly improved glucose tolerance in STZ diabetic rats. This effect seemed at
least partly mediated by a significant increase of insulin secretion. Other compounds of the
same family (7b, 16f, 23b) have also shown potent activity. We found no correlation between
in vivo antihyperglycemic properties and in vitro affinities for R2-adrenoceptors or I1, and I2
binding sites. These compounds can be considered as antihyperglycemic agents potentially
useful for treatment of type II diabetes and are currently under complementary investigation.
In tr od u ction
Non-insulin-dependent diabetes mellitus (NIDDM) is
wide spread and is one of the most common chronic
diseases, being present in 4% of the Western World’s
population, half of whom are unaware of it. Although
the frequency of diabetes is increasing, only two classes
of oral hypoglycemic agents are available for the treat-
ment of NIDDM. For both, residual â-cell insulin
secretion is necessary for activity. The sulfonylureas,
such as tolbutamide and gliclazide, are hypoglycemic
compounds, which act mainly at the pancreas level by
stimulation of insulin secretion. Their mechanism of
action is well known.1 On the other hand, the bigu-
anides (particularly metformin) act mainly in the pres-
For many years, development of antidiabetic drugs
ence of endogenous insulin by decreasing gluconeogen-
has been largely focused on improvement of sulfo-
esis and increasing peripheral utilization of glucose2-5
.
nylurea derivatives, in order to increase the efficiency
and decrease side effects6 (in particular hypoglycemic
events). Among the most promising new approaches,
some imidazoline derivatives such as midaglizole,7,8
deriglidole,9-11 and efaroxan12 were reported to be R2-
antagonists acting via stimulation of insulin secretion
and to exhibit antihyperglycemic activity in vivo. In the
past few years, further studies have shown that the
insulin-secreting potency of some imidazoline deriva-
tives was not correlated with their R2-antagonistic
* Author to whom correspondence should be addressed. E-mail:
godfroid@paris7.jussieu.fr.
§ Laboratoire de Pharmacochimie Mole´culaire, Universite´ Paris
7-Denis Diderot.
‡ CNRS URA 307, Universite´ Paris 7-Denis Diderot.
† ADIR.
#
IRI Servier.
| Universite´ de Lie`ge.
UPRESA, CNRS 5018/-UPS.
∇ These authors contributed equally to this work.
X Abstract published in Advance ACS Abstracts, October 1, 1997.
S0022-2623(96)00862-X CCC: $14.00 © 1997 American Chemical Society

3794 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23
Rondu et al.
properties,13,14 and so the involvement of new putative
biological targets was suggested.
Ta ble 1. pKa Values of Idazoxan and Some Compounds of the
Series 1-Aryl-4-alkyl-2-(imidazolin-2)-yl)piperazines
a
b
compd
pKa
pKa
idazoxan
7b
7f
8.8
9.5
9.4
5.5
5.4
a
b
pKavalue of the imidazoline ring. pKa value of the less
hindered tertiary amino function of the piperazine. The basicity
of compounds was assessed by potentiometric titration in aqueous
solution.
the molecular modeling studies. Our initial purpose
was to synthesize new structures having the same
binding profile as idazoxan, i.e., affinity on both R2-
adrenoceptors and I2 binding sites (RX 821002 being a
specific R2-adrenoceptor antagonist). Consequently,
calculated 3-D electrostatic potential maps of RX 821002
and idazoxan were compared with that of PMS 81219
(compound 7f) which was one of our compounds showing
potent antihyperglycemic properties. Our hypothesis
was that the electronic distribution should be as close
as possible to idazoxan’s but not to RX 821002’s.
It has been proposed that some of these derivatives
could interact with imidazoline-preferring binding sites
(IPBS).15,16 There are two major subtypes of IPBS: I1,
which preferentially bind [3H]clonidine or [3H]-p-ami-
noclonidine, and I2, which preferentially bind [3H]-
idazoxan.17,18 Idazoxan was selected for molecular
modeling studies for its affinity for both R2-adrenocep-
tors and I2 IPBS, whereas its close analogue, RX 821002,
was chosen as a selective R2-antagonist.
Concerning the potential protonation, experimental
pKa values of compounds 7b,f and idazoxan were
determined. The results (Table 1) clearly demonstrate
that the imidazoline ring is the first protonation center.
The piperazine moiety is only very slighty ionized at
the physiological pH.
Moreover, calculations were performed. Whatever the
conformation could be, the protonation energy difference
between the imidazoline and piperazine nitrogens was
higher than 25 kcal/mol.20 Here, we made an abstrac-
tion of the possible protonation at physiological pH of
the imidazoline ring in this comparison.
In this paper, we report the synthesis and the
pharmacological evaluation of new antihyperglycemic
2-(4′,5′-dihydro-1′H-imidazol-2′-yl)piperazines having the
following general formula:
The 3-D electrostatic potential map is the “finger-
print” of a molecule, taking into account electronic
distribution, conformation, and volume. It is the best
correlation explaining the similarity of activities be-
tween molecules which, apparently, are structurally
different.
The conformational space of RX 821002, at the AM1
level,21 has been analyzed using complete scans of the
two dihedral angles defining the relative position of the
methoxy and imidazoline groups. At each step, the
geometry was reoptimized following the procedure
implemented in GAUSSIAN 94.21
Taking into account the optimized geometry of RX
821002, Figure 1 shows that a major electrostatic
potential takes place around one of the oxygens of the
benzodioxane ring (isocontour at -20 kcal/mol) rein-
forced by the generating effect of the oxygen of the
methoxy substituent (visible with an isocontour at -40
kcal/mol, the red lines).
The comparison between 3-D electrostatic potential
maps of idazoxan and of PMS 812 (compound 7f) is clear
(Figure 2): two negative wells appear around the
nitrogens of the imidazoline ring (as in the case of RX
821002). Moreover, two separated electrostatic regions
are positioned around the nitrogens of piperazine and
These series were designed after our studies based on
comparative analysis of electronic distribution in both
Idazoxan and RX 821002. Despite their marked anti-
hyperglycemic properties, our new compounds were
devoid of significant affinities for both the R2-adreno-
ceptors and the I1 and I2 IPBS; therefore, we decided to
evaluate all of them in vivo for their antidiabetic
activity. This was achieved using glucose tolerance
tests performed on a rat model of type II diabetes
obtained by injection of a low dose (35 mg/kg) of
streptozotocine. Structural modification was systemati-
cally carried out in order to determine the necessary
structural requirements for potent in vivo activity. For
some compounds, secretion of insulin was also evalu-
ated. Some in vitro binding experiments were per-
formed on selected analogues to investigate a potential
correlation between in vivo and in vitro studies.
Molecu la r Mod elin g a n d Design of Th ese Ser ies
As mentioned in the Introduction, idazoxan and
methoxyidazoxan (RX 821002) were initially selected for

Imidazolines Active on Glucose Homeostasis
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23 3795
(for alkyl * CH3) in acetone with K2CO3 and catalytic
KI (for alkyl bromide) according to the method of J erzy
et al.23 Yields of compounds 8-11 (Scheme 1) and 20
and 21 (Scheme 2) increased using an excess of the alkyl
halide. Under the same conditions methyl iodide led
to a mixture of tertiary amine and quaternary am-
monium salt. Products 6 (Scheme 1) and 18 (Scheme
2) were then obtained using formaldehyde and formic
acid in methanol according to Icke.24
Alkylation of 2 (Scheme 2) with 1 equiv of substituted
benzyl chloride in refluxing toluene gave two esters (17
and 24) in 35% and 25% yields, respectively. These
compounds were easily separated on column chroma-
tography. Ester 24 can be prepared in a yield of 80%
using 2 equiv of benzyl chloride. All imidazolines were
prepared under the same conditions according to Neef
et al.25 with a minor modification. An excess of alumi-
num organic reagent was necessarily used in order to
avoid a mixture of open-chain amide and ring-closed
heterocycle.
F igu r e 1. Electrostatic potential maps of RX 821002 isocon-
toured at -20 (green lines), -30 (yellow lines), and -40 (red
lines) kcal/mol.
P h a r m a cologica l P a r a m eter s in Vivo:
Sign ifica n ce a n d Va lid ity
Antidiabetic properties of all the synthesized com-
pounds were quantified by their ability to improve
glucose tolerance during an intravenous glucose toler-
ance test (IVGTT) in diabetic rats. From the few
existing models, a mild diabetic experimental rat model
developed by Thibault et al.26 was chosen for our
studies. This diabetic rat presents moderate basal
hyperglycemia, glucose intolerance, and impairment of
glucose-induced insulin secretion which are the main
features in patients with NIDDM. Moderate diabetes
is obtained by iv injection of a low dose (35 mg/kg) of
streptozotocin (STZ). Two weeks later, measurement
of the basal glycemia and IVGTT was performed,
allowing the selection of animals with moderate diabe-
tes, i.e., basal glycemia comprised between 7 and 10 mM
and a ∆G ranging from 50 to 80 mM. IVGTT was then
performed following a single administration of 100 µmol/
kg of the synthesized compounds. The following pa-
rameters were taken into account: (i) G30, which is the
glycemia value at 30 min after glucose administration,
(ii) ∆G, which represents the increase in glycemia over
baseline integrated over a period of 30 min following
the glucose load, and (iii) K, which is the rate of glucose
disappearance between 5 and 30 min after glucose
administration. For some compounds, secretion of
insulin was also evaluated and expressed as ∆I which
represents the incremental plasma insulin values over
baseline integrated over a period of 30 min after the
glucose load. To correct for the slight variations in
response that appeared when glucose tolerance tests
were performed with different control diabetic rats, we
decided to express the results as a percentage (∆G*) of
variation of ∆G between treated and untreated diabetic
rats. In the same way, we defined G30*, K*, and ∆I*.
F igu r e 2. Comparison of the electrostatic potential maps of
idazoxan (left) and compound 7f (right), the lead compound of
the series 2-(4′,5′-dihydro-1′H-imidazol-2′-yl)piperazine, iso-
contoured at -20 (green lines), -30 (yellow lines), and -40
(red lines) kcal/mol.
the oxygens of dioxane. Electrostatic potential maps of
idazoxan and PMS 812 (compound 7f) may be super-
imposable if one ignores the phenyl versus benzyl ring
locations; the distances between the aromatic rings and
the heteroatoms must be different in the phenyl and
benzyl analogues. This is the reason why we chose to
develop studies in these series.
Ch em istr y
Scheme 1 shows a typical total synthesis of 1-aryl-4-
alkyl-2-(2-imidazolinyl)piperazines. The condensation
of 1,4-dibenzylethylenediamine and 2,3-dibromopropi-
onic acid ethyl ester, according to J ucker et al.22 with
minor modifications, gave after acidification ethyl 1,4-
dibenzyl-2-piperazinecarboxylate hydrochloride (1). The
N-debenzylation using catalytic hydrogenolysis is faster
with the dihydrochloride salt than with the free base.
The two nitrogen atoms of 2 do not present the same
reactivity. At low temperature (-10 °C) the alkylation
with 1 equiv of triphenylmethyl chloride led to the
monosubstitution on position 4. In these conditions the
ethyl 4-(triphenylmethyl)-2-piperazinecarboxylate (3)
was obtained with a very satisfactory yield. Reaction
of 3 with 1 equiv of substituted benzyl chloride afforded
the corresponding ethyl 1-(substituted benzyl)-4-(triph-
enylmethyl)-2-piperazinecarboxylate 4. Cleavage of the
protective group in dilute HCl yielded the ethyl 1-(sub-
stituted benzyl)-2-piperazinecarboxylate 5. Alkylation
of 5 was carried out by action of the suitable alkyl halide
∆Gtreated STZ rats - ∆Guntreated STZ rats
∆G* )
× 100
|
|
∆Guntreated STZ rats - ∆Gcontrol rats
To be considered as an effective antidiabetic, com-
pounds must have high a percentage of variation for
∆G*, G30*, and K*. Results around 100% or more mean
that the parameters for the treated diabetic animals are
close to those obtained with nondiabetic control animals.

3796 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23
Rondu et al.
Sch em e 1
P h
P h
P h
This is the case, among others, for 7f (PMS 812) for
which results have already been published19 and 7b for
which glycemia values of STZ-treated rats are very close
to those obtained with normal rats (Figures 3 and 4).
Compounds like 13b, which has virtually no influence
on the parameters, can be considered as inactive (Figure
5). Among the three parameters, G30, which is mea-
sured directly, is considered as the most relevant.
2). Substitutions by halogens or other substituents
(compounds 7b-p ) may result, depending on the sub-
stitution site, in a significant increase of activity com-
pared to the nonsubstituted compound 7a . The most
potent 2-monosubstituted compound was 7b, a 2-chloro
derivative, which proved to be much more active than
the nonsubstituted 7a (∆G* and G30* values respectively
of 81% and 103% for 7b compared to 0% and 23% for
7a ). Analogue 7m , which is the fluoro analogue of 7b,
exhibited almost the same potency. Monosubstitution
by other groups such as 2-methyl (7p ) or 2-methoxy (7l)
resulted in lower activity, while substitution by a
trifluoromethyl (7o) gave better results, almost equiva-
lent to those of 7b (Table 2).
The substituents of this 2-monosubstituted series can
be arranged in the following order of potency as anti-
hyperglycemic agents: Cl > F > CF3 > OMe > CH3.
Compound 7b was selected for pharmacomodulation.
This consisted of varying the position of the Cl on the
benzyl moiety. Compounds 7c (3-Cl) and 7d (4-Cl) were
found to be less active than 7b. Surprisingly, substitu-
tion with a 3-methoxy (7k ) gave better results than with
a 3-Cl (7c). Compound 7f, which is a 2,4-dichloro
Resu lts a n d Discu ssion
As previously mentioned, all the synthesized com-
pounds were evaluated via an ip administration (100
µmol/kg) in glucose tolerance tests performed on STZ
rats which exhibited moderate basal hyperglycemia and
glucose intolerance. The first tested imidazoline deriva-
tive, 7b, had a potent effect on the glycemia, which
resulted in values similar to those of nondiabetic rats
(Figure 3). This correlated very well with its effects on
the three parameters, with ∆G*, G30*, and K* values
around 100%. Analogues with substituents on the
aromatic ring of the 1-benzyl-2-(4′,5′-dihydro-1′H-imi-
dazol-2′-yl)-4-methylpiperazine were studied in order to
determine their influence on antidiabetic activity (Table

Imidazolines Active on Glucose Homeostasis
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23 3797
Sch em e 2
F igu r e 4. Effect of a single intraperitoneal administration
of 7f on glucose tolerance: (0) STZ rats treated with 7f, (])
untreated STZ rats, and (O) control rats. *p < 0.05, signifi-
cantly different from untreated STZ rats. 7b was given (100
µmol/kg ip) 15 min before an iv glucose load (0.5 g/kg). Plasma
glucose level (mmol/L) was decreased compared to untreated
diabetic rats and was not significantly different from the value
in normal control rats.
F igu r e 3. Effect of a single intraperitoneal administration
of 7b on glucose tolerance: (0) STZ rats treated with 7b, (])
untreated STZ rats, and (O) control rats. *p < 0.05, signifi-
cantly different from untreated STZ rats. 7b was given (100
µmol/kg ip) 15 min before an iv glucose load (0.5 g/kg). Plasma
glucose level (mmol/L) was decreased compared to untreated
diabetic rats and was not significantly different from the value
in normal control rats.
derivative, was found to be the most potent compound
among the disubstituted derivatives with an activity
even higher than that of 7b. The positions of the
substituent seem to be one of the most important factors
as shown by the results of the 3,4-dichloro derivative
7h which is clearly less active than 7f. These results
indicate that electronic and steric factors of substituents
seem to be strongly involved in the activity.

3798 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23
Rondu et al.
Ta ble 2. Variation of the Glycemia Parameters after ip
Administration of 100 µmol/kg 1-Benzyl-2-(4′,5′-dihydro-
1′H-imidazol-2′-yl)piperazines to Streptozotocin Ratsa
c
∆G*b G30
*
(%)
K*d ∆I*e
(%) (%)
compd
7a (8)
7b (5)
7c (6)
7d (4)
7e (5)
7f (8)
7g (9)
7h (7)
7i (5)
7j (6)
7k (5)
X
R
(%)
H
CH3
0
81
0
0
0
105
103
36
0
47
23
0
23
103
23
68 110
36
112 143 221
59 107
47
67
98 109
39
16
2-Cl
3-Cl
4-Cl
2,3-Cl
2,4-Cl
2,6-Cl
3,4-Cl
3,5-Cl
2-OCH3
3-OCH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
53
42
70 117
56
68
58
39
57
55
88
60
7l (7)
2,3-OCH3 CH3
F igu r e 5. Effect of a single intraperitoneal administration
of 13b on glucose tolerance: (0) STZ rats treated with 13b,
(]) untreated STZ rats, and (O) control rats. *p < 0.05,
significantly different from untreated STZ rats. 13b was given
(100 µmol/kg ip) 15 min before an iv glucose load (0.5 g/kg).
Plasma glucose level (mmol/L) was not decreased compared
to untreated diabetic rats and was significantly different from
the value in normal control rats.
7m (5)
7n (5)
7o (8)
2-F
CH3
CH3
CH3
CH3
H
ethyl
n-propyl
isopropyl
n-propyl
isobutyl
72
102
94
24
0
108
0
84
24
110
8
52 108
2,4-F
2-CF3
2-CH3
2-Cl
2-Cl
2-Cl
2-Cl
2,4-Cl
2,4-Cl
106
97
63
81
11
34
37
44
64
48
0
38
17
83
7p (8)
12b (4)
13b (4)
14b (8)
15b (6)
14f (7)
16f (4)
idazoxan (6)
midaglizole (6)
25
80 104
95 119 623
Regarding the substituents on piperazine nitrogens,
replacement of the methyl by a hydrogen (12b) resulted
in a total loss of activity. Substitution with an ethyl
(13b) seemed to have only little effect, while a propyl
(14b) altered the activity completely compared to the
N-methyl derivative 7b. Surprisingly, compounds with
a branched alkyl isopropyl (15b) or isobutyl (16f)
exhibited almost the same potencies as their N-methyl
counterparts 7b,f. Introduction of a second benzyl
substituent instead of the alkyl group on the piperazine
nitrogen atom (25a ,b,f) resulted in a total loss of activity
compared to the monobenzyl N-alkyl analogues (7a ,b,f).
Inverse analogues of the most active compounds 7b,f
and 15b were prepared by variation of the benzyl and
alkyl substituents on the piperazine nitrogens. The
activities of these 1-alkyl-2-(4′,5′-dihydro-1′H-imidazol-
2′-yl)-4-benzylpiperazine derivatives (Table 3, com-
pounds 19b,f, 23b) were in the same range for the three
parameters as that of their counterparts (Table 2).
Some compounds (7a ,b,f,k ,l, 14b, 16f) were also evalu-
ated for their capacity to stimulate the secretion of
insulin during IVGTT. It clearly appeared that the
most active compounds on ∆G*, G30*, and K* are also
those that strongly stimulate insulin release (7b, ∆I )
109%; 7f, ∆I ) 221%; 16f, ∆I ) 623%). Compounds less
active on glycemia parameters moderately stimulate the
secretion of insulin (7k , ∆I ) 88%; 7l, ∆I ) 60%; 14b,
∆I ) 25%; 7a , ∆I ) 16%).
48
121
79
98
32
a
Three-month-old male Wistar rats (250 g) treated with 35 mg/
b
kg iv streptozotocin. ∆G: incremental glycemia values over
baseline integrated over 30 min after glucose (0.5 g/kg iv)
administration. G30
c
: glycemia value 30 min after glucose ad-
d
ministration. K: rate of glucose disappearance between 5 and
30 min after glucose administration. e ∆I: incremental plasma
insulin values over baseline integrated over 30 min after glucose
administration. All results (asterisk) are expressed as a percent
of variation of the parameters between treated rats (with an ip
administration of 100 µmol/kg of the tested compound) and
untreated rats; n: number of experiments conducted with each
drug, indicated in parentheses.
tives, and more particularly those active on insulin
secretion, may interact with a novel type of imidazoline-
preferring site located on pancreatic Β cells. Compound
7f, which proved to be our most interesting compound,
is currently under complementary pharmacological
evaluation. Further chemical modulation is in progress
to obtain more active compounds and develop structure-
activity relationships. For this purpose, we have in-
vestigated some electronic and structural characteristics
of 7f, the most interesting compound, using a molecular
modeling approach.
The imidazoline substituent placed at the 2-position
of the piperazine ring of 7f plays a major role in the
twisted conformation. In practice, four conformations
(C11, C12, C21, C22) can be located on the energy
surface depending on the hybridization of the two
nitrogens with their respective lone pair “up” or “down”.
The in vitro binding experiments performed with our
compounds have shown that none of them had signifi-
cant affinity (>10-6 M) for the R2-adrenoceptors or I1
and I2 IPBS. This strongly suggests that the potent
insulin-secreting properties of the synthesized com-
pounds do not involve these receptors or binding sites.
This is in agreement with an increasing number of
publications showing that R2-adrenergic antagonists are
able to stimulate insulin secretion independently of R2
or/and I1 and I2 blockage.27-35 The imidazoline deriva-
Full geometry optimization of the four conformers has
been carried out at three levels of calculation: an
approximate AM136 method and an ab initio one using
minimal and double basis sets, MINI-1′ and 6-31G.38
37
∆G has been computed at 298.15 K and 1 atm (Table
4). From the geometrical point of view, these conform-
ers occupy a large part of the 3-D space mainly as a

Imidazolines Active on Glucose Homeostasis
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23 3799
Ta ble 3. Variation of the Glycemia Parameters after ip
Administration of 100 µmol/kg 4-Benzyl(or 1,4-dibenzyl)-2-(4′,5′-
dihydro-1′H-imidazol-2′-yl)piperazines to Streptozotocin Ratsa
to a new class of imidazoline derivatives having poten-
tial therapeutic interest in type II diabetes. In the
future, the exploration in the design of these new potent
imidazoline derivatives may allow a greater understand-
ing of the structural parameters required for activity
and may shed light on the role of imidazoline derivatives
in the regulation of insulin secretion.
Exp er im en ta l Section
c
∆G*b
(%)
G30
(%)
*
K*d
(%)
compd (n)
19b (5)
X
R
Molecu la r Mod elin g. For all the studied products, the
geometry has been fully optimized with respect to all the 3N
- 6 degrees of freedom. The berny algorithm of the GAUSS-
IAN 94 package20 minimizes the first derivative of the energy
function calculated either at the AM1 semiempirical level or
at the ab initio level using minimal MINI-1′ and double 6-31G
basis sets. The nature of the equilibrium structures thus
obtained as true minima is given by all the positive eigenval-
ues of the second-derivative matrix.
By opposite to rigid rotor conformational analysis, the
relaxation procedure applied for RX 821002 requires a com-
plete reoptimization of the 3N - 8 other degrees of freedom
when a bidimensional map is concerned. The electrostatic
potential maps have been calculated at the ab initio level using
the MINI-1′ basis sets with the GAUSSIAN suite of pro-
grams.20
2-Cl
2,4-Cl
2-Cl
2-Cl
H
CH3
CH3
93
34
70
112
0
97
102
50
96
33
148
152
106
95
19f (8)
22b (5)
23b (5)
25a (10)
ethyl
isopropyl
43
25b (7)
2-Cl
0
0
36
22
25f (5)
2,4-Cl
0
18
idazoxan (6)
midaglizole (6)
8
32
48
121
79
98
a
Three-month-old male Wistar rats (250 g) treated with 35 mg/
b
kg iv streptozotocin. ∆G: incremental glycemia values over
baseline integrated over 30 min after glucose (0.5 g/kg iv)
administration. G30
ministration. K: rate of glucose disappearance between 5 and
30 min after glucose administration. All results (asterisk) are
expressed as a percent of variation of the parameters between
treated rats (with an ip administration of 100 µmol/kg of the tested
compound) and untreated rats; n: number of experiments con-
ducted with each drug, indicated in parentheses.
Ch em istr y. Gen er a l Meth od s. The purity of each com-
pound was checked by thin-layer chromatography on TLC
plastic sheets (silica gel 60F254, layer thickness 0.2 mm) from
Merck. Column chromatography purification was carried out
on silica gel 60 (particle size 0.063-0.200 mm) from Merck,
without any special treatment. All melting points were
determined in a digital melting point apparatus (Electrother-
mal) and are uncorrected. The structures of all compounds
c
: glycemia value 30 min after glucose ad-
d
1
were confirmed by IR and H NMR spectra. IR spectra were
Ta ble 4a
obtained with a Pye-Unicam SP3-200 infrared spectrometer,
and 1H NMR spectra were recorded in CDCl3 on a Brucker
AC 200 spectrometer using hexamethyldisiloxane (HMDS) as
an internal standard. All elemental analyses were within
(0.4% of theoretical values.
AM1
MINI-1′
6-31G
∆E
∆E
∆G
∆E
∆G
C11
C12
C21
C22
0.000
0.183
2.285
0.000
1.281
2.339
0.277
0.000
0.560
1.739
0.000
1.257
-1.099
1.399
0.000
-0.284
-0.303
0.962
Eth yl 1,4-Diben zyl-2-p ip er a zin eca r boxyla te Dih yd r o-
ch lor id e (1). To a hot (80 °C) stirred solution of N,N ′-
dibenzylethylenediamine (72 g, 0.3 mol) and triethylamine
(100 mL, 0.72 mol) in toluene (300 mL) was added dropwise,
but rapidly, ethyl 2,3-dibromopropionate (80 g, 0.31 mol in
toluene (300 mL)). After the addition, the reaction mixture
was stirred at 80 °C for 3 h, then cooled, and filtered. The
filtrate was washed with saturated aqueous sodium hydrogen
carbonate (200 mL). The organic layer was dried over MgSO4
and the solvent removed in vacuo. The crude product was
dissolved in anhydrous EtOH (400 mL) and saturated with
HCl gas. The addition of ether gave a precipitate of 1 (100-
110 g, 81-86%) as a white powder: mp 152-154 °C; IR film
(of free base) 1735 (νCdO), 1600 and 1585 (νCdC) cm-1; 1H NMR
(of free base) δ 7.4-7.1 (10H, m, aromatic H), 4.13 (2H, q, J )
7.1 Hz, COOCH2), 3.38-3.88 (4H, m, PhCH2), 3.32-3.21,
3.05-2.95, 2.69-2.27 (7H, 3m, piperazinic H), 1.19 (3H, t, J
) 7.1Hz, CH3).
-0.860
-0.258
a
Energy values are expressed in kcal/mol. The C11 conformer
is taken as reference. Negative values correspond to more stable
conformers.
result of the relative orientation of the dichlorophenyl
group. Remarkably, the relative energy differences lie
in a very small range as noted in Table 5. This feature
can be interpreted as a high capability of the molecule
7f to adopt the most effective conformation at the
receptor site including the ability of the imidazoline to
be protonated in its biological active conformation.20
Con clu sion
Eth yl 2-P ip er a zin eca r boxyla te Dih yd r och lor id e (2). A
shaken suspension of 1 (40 g, 0.1 mol) and 10% Pd-C (600
mg) in anhydrous ethanol (300 mL) was treated with H2 (under
pressure) at 40 °C overnight. After an addition of water until
dissolution of 2 the reaction mixture was filtered through
Celite. The addition of ether to the filtrate gave the precipitate
2 (20-23 g, 86-91%) as a white powder: mp 195-197 °C; IR
(in paraffin oil) 3380 (νN-H), 1730 (νCdO) cm-1; 1H NMR (of free
base) δ 4.13 (2H, q, J ) 7.1 Hz, COOCH2), 3.42-3.36, 3.19-
3.11, 2.98-2.68 (7H, 3m, piperazinic H), 1.84 (2H, br s, NH),
1.21 (3H, t, J ) 7.1 Hz, CH3).
Eth yl 4-(Tr iph en ylm eth yl)-2-piper azin ecar boxylate (3).
To a stirred solution of 2 (23.1 g, 0.1 mol) and triethylamine
(55 mL, 0.4 mol) in dry dichloromethane (400 mL) was added
dropwise triphenylmethyl chloride (28 g, 0.1 mol) in dry
dichloromethane (300 mL) at -10 °C. After addition, the
In conclusion, most of these 29 imidazoline derivatives
synthesized had a potent effect on po glucose tolerance,
and some of them increased significantly in vivo the
insulin secretion in mildly diabetic rats. Compounds
7b, 7f (PMS 812), and 16f had the highest potency, with
glycemia values similar to those obtained with normal
rats. These compounds do not have high affinity for R2-
adrenoceptors or the known imidazoline binding sites
I1 and I2. Possibly, the effect of 7f on the insulin
secretion results from its interaction with a novel type
of imidazoline-binding site. Molecular modeling efforts
with 7f suggest that this compound is able to adopt the
most effective conformation at the receptor site and is
characterized by high flexibility. Compound 7f belongs

3800 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23
Rondu et al.
Ta ble 5. Physical Data of 1-Aryl-4-alkyl-2-(imidazolin-2-yl)piperazines
recrystn
solventc
compd
X
R
mp (°C)a
% yieldb
formula
7a
7b
7c
7d
7e
7f
7g
7h
7i
7j
7k
7l
7m
7n
7o
H
2-Cl
3-Cl
4-Cl
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
166-168
160-162
180-182
198-200
158-160
168-170
182-184
170-172
167-169
168-170
178-180
160-162
160-162
166-168
163-165
156-158
160-162
162-164
160-163
179-181
155-157
118-120
55
70
60
49
58
60
43
43
47
36
38
41
37
34
45
51
22
30
74
40
54
38
D
C
D
D
C
D
D
D
D
D
D
D
D
D
D
D
C
C
D
B
A
D
C15H22N4‚2HCl‚2H2O
C15H21ClN4‚2HCl‚1H2O
C15H21ClN4‚2HCl‚1H2O
C15H21ClN4‚2HCl‚1H2O
C15H20Cl2N4‚2HCl‚1H2O
C15H20Cl2N4‚2HCl‚1H2O
C15H20Cl2N4‚2HCl‚1H2O
C15H20Cl2N4‚2HCl‚1H2O
C15H20Cl2N4‚2HCl‚1H2O
C16H24N4O‚2HCl‚1H2O
C16H24N4O‚2HCl‚1‚5H2O
C17H26N4O2‚3HCl‚2H2O
C15H21FN4‚3HCl‚2H2O
C15H20F2N4‚2HCl‚1H2O
C16H21F3N4‚2HCl‚1H2O
C16H24N4‚2HCl‚1.5H2O
2,3-Cl
2,4-Cl
2,6-Cl
3,4-Cl
3,5-Cl
2-OMe
3-OMe
2,3-OMe
2-F
2,4-F
2-CF3
2-CH3
2-Cl
7p
12b
13b
14b
14f
15b
16f
C14H19ClN4‚3HCl‚1H2O
C16H23ClN4‚2HCl
C17H25ClN4‚3HCl‚1H2O
C17H24Cl2N4‚2HCl‚1H2O
C17H25ClN4‚2HCl‚1H2O
C18H26Cl2N4‚2HCl‚1H2O
2-Cl
2-Cl
2,4-Cl
2-Cl
2,4-Cl
C2H5
nC3H7
nC3H7
iC3H7
iC4H9
a
b
All compounds had decomposition. Yields of the last step. c A, acetone; B, acetone/ethanol; C, acetone/methanol; D, dichloromethane/
hexane/methanol.
reaction mixture was stirred overnight at -10 °C. The
reaction was stopped by addition of Na2CO3 (21 g) and water
(200 mL). The organic layer was separated, washed with
water, and dried over MgSO4. After addition of toluene (50
mL), solvents were evaporated and the crude product was
crystallized from hexane:ether (9:1, v/v) to give 26 g of a white
powder. The filtrate was purified by column chromatography
using petroleum ether:ether (7:3, v/v) as eluent giving about
6 g of 3. The total yield of this reaction was 80%: mp 112-
Meth od B: The crude product 4b obtained in the last step
was solubilized in ethanol (400 mL) and satured with HCl gas.
The mixture was stirred for 3 h at room temperature; then
the solvent was removed in vacuo. The residue obtained was
taken up in ether and washed with a 2 N HCl solution. The
aqueous layer was neutralized with a saturated Na2CO3
solution and extracted with ether. Drying over MgSO4,
filtration, and evaporation of CHCl3 gave a pure colorless oil
(5b) with the same yield as in method A.
1
113 °C; IR (in paraffin oil) 1740 (νCdO), 1590 (νCdC) cm -1; H
Gen er a l P r oced u r e for Eth yl 1-(Su bstitu ted ben zyl)-
4-a lk yl-2-p ip er a zin eca r boxyla te P r ep a r a tion . Meth od
A: Eth yl 1-(2′-Ch lor oben zyl)-4-m eth yl-2-p ip er a zin eca r -
boxyla te (6b). A mixture of 5b (28 g, 0.1 mol), 37% formal-
dehyde (12 mL, 0.16 mol), and formic acid (12 mL, 0.3 mol) in
CH3OH (120 mL) was refluxed for 20 h. After evaporation of
the solvent, the residue was taken up in ether and saturated
Na2CO3 solution was added until basic pH. The organic layer
was washed with water and dried over MgSO4, and the solvent
was removed in vacuo. The crude product was dissolved in
anhydrous EtOH (200 mL) and saturated by HCl gas. The
addition of ether gave a precipitate of 6b (32 g, 86%) as a white
powder: mp 166-168 °C; IR (film, of free base) 1745 (νCdO),
1600, 1580 (νCdC) cm-1; 1H NMR δ 7.47-7.06 (4H, m, aromatic
H), 4.15 (2H, q, J ) 7.2 Hz, COOCH2), νa ) 3.88, νb ) 3.71
(2H, q, AB spectrum, J ) 14.5 Hz, I2/I1 ) 2.33, CH2Ph), 3.38-
3.33, 3.07-2.95, 2.76-2.26 (7H, 3m, piperazinic H), 2.20 (3H,
s, NCH3), 1.22 (3H, t, J ) 7.2 Hz, CH3).
Meth od B: Eth yl 1-(2′-Ch lor oben zyl)-4-eth yl-2-p ip er a -
zin eca r boxyla te (8b). This compound was prepared from
5b following the general procedure described for 4b, but the
solution was stirred at 35 °C for 24 h. The crude product was
purified by column chromatography using petroleum ether:
ether (80:20, 60:40, and 50:50, v/v) as eluent giving 14.6 g of
8b as a pale yellow oil. Yield was about 70%: IR (film) 1755
(νCdO), 1600, 1580 (νCdC) cm-1; 1H NMR δ ppm 7.48-7.05 (4H,
2m, aromatic H), 4.13 (2H, q, J ) 7.2 Hz, COOCH2), νa ) 3.85,
νb ) 3.69 (2H, q, AB spectrum, J ) 14.5 Hz, I2/I1 ) 2.29, CH2-
Ph), 3.37-3.33, 3.04-2.98, 2.76-2.22 (9H, 3m, piperazinic H,
NCH2-CH3), 1.21 (3H, t, J ) 7.2 Hz, COOCH2-CH3), 0.98 (3H,
t, J ) 7.2 Hz, N-CH2-CH3).
NMR δ 7.7-7.05 (15H, m, aromatic H), other H gave unre-
solved signals between 4.5 and 1.1 ppm.
Gen er a l P r oced u r e for Eth yl 1-(Su bstitu ted ben zyl)-
4 -( t r i p h e n y l m e t h y l ) -2 -p i p e r a z i n e c a r b o x y l a t e
P r ep a r a tion : Eth yl 1-(2′-Ch lor oben zyl)-4-(tr ip h en ylm -
eth yl)-2-p ip er a zin eca r boxyla te (4b). A mixture of 3 (40
g, 0.1 mol), K2CO3 (40 g), KI (4 g), and 2-chlorobenzyl chloride
(19 g, 0.12 mol) in acetone (400 mL) was stirred for 15 h at
room temperature. After filtration and evaporation of the
solvent, the residue was taken up in ether, washed with water,
and dried over MgSO4. Ether was evaporated to give the crude
product 4b which was used in the next step without purifica-
tion.
Gen er a l P r oced u r e for Eth yl 1-(Su bstitu ted ben zyl)-
2-p ip er a zin eca r boxyla te P r ep a r a tion : Eth yl 1-(2′-Ch lo-
r oben zyl)-2-p ip er a zin eca r boxyla te (5b). Meth od A: All
of the crude product 4b obtained in the previous step was
treated with acetone (600 mL) and concentrated HCl (25 mL)
for 3 h at room temperature. The solvent was removed in
vacuo; the residue was solubilized in ether and washed with
water. The aqueous layer was neutralized with a saturated
Na2CO3 solution, extracted with ether, and dried over MgSO4.
Evaporation of the solvent gave a crude product which was
purified by column chromatography using first petroleum
ether:ether (30:70, v/v) and then ether to afford 19 g (80%) of
5b as a colorless oil: IR (film) 3330 (νN-H), 1740 (νCdO), 1605,
1
1580 (νCdC) cm-1; H NMR δ 7.47-7.10 (4H, m, aromatic H),
4.15 (2H, q, J ) 7.1 Hz, COOCH2), νa ) 3.74, νb ) 3.72 (2H, q,
AB spectrum, J ) 14.5 Hz, I2/I1 ) 4.45), 3.27-3.23, 3.15-2.76,
2.34-2.25 (7H, 3m, piperazinic H), 1.60 (1H, br s, NH), 1.23
(3H, t, J ) 7.1 Hz, CH3).

Imidazolines Active on Glucose Homeostasis
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23 3801
Ta ble 6. Physical Properties of 1-Alkyl-4-aryl-2-(imidazolin-2′-yl)piperazines and 1,4-Diaryl-2-(imidazolin-2′-yl)piperazines
crystn
compd
X
R
mp (°C)a
% yieldb
solventc
formula
19b
19f
22b
23b
25a
2-Cl
2,4-Cl
2-Cl
2-Cl
H
CH3
CH3
C2H5
140-142
160-162
138-140
158-160
150-152
62
60
64
58
57
F
C15H21ClN4‚3HCl‚1H2O
C15H20Cl2N4‚2HCl‚1H2O
C16H23ClN4‚3HCl‚1H2O
C17H25ClN4
D
B
C
E
iC3H7
C21H26N4
25b
2-Cl
86-88
67
A
C21H24Cl2N4‚3HCl‚2H2O
C21H22Cl4N4‚2HCl‚1H2O
25f
2,4-Cl
188-190
62
A
a
b
All compounds had decomposition. Yields of the last step. c A, acetone; B, acetone/ethanol; C, acetone/methanol; D, dichloromethane/
hexane/methanol; E, ether/hexane; F, ethanol/methanol.
Gen er a l P r oced u r e for Eth yl 4-(Su bstitu ted ben zyl)-
2-p ip er a zin eca r boxyla te P r ep a r a tion : Eth yl 4-(2′-Ch lo-
r oben zyl)-2-p ip er a zin eca r boxyla te (17b). To a refluxing
solution of 2 (23 g, 0.1 mol) and triethylamine (55 mL, 0.40
mol) in toluene (200 mL) was added dropwise 2-chlorobenzyl
bromide (16.1 g, 0.1 mol) in toluene (200 mL). The reaction
mixture was refluxed with stirring for 1 h. After cooling, the
solution was washed with water and dried over MgSO4, and
the solvent was removed in vacuo. The residue was purified
by column chromatography using first petroleum ether:ether
(70:30, v/v) and then ether yielding 10 g (25%) of 24b and 10
g (35%) of 17b as a pale yelow oil. 17b: IR (film) 3347(νN-H),
Gen er a l P r oced u r e for Im id a zolin e P r ep a r a tion : 1-(2′-
Ch lor oben zyl)-4-m eth yl-2-(4′,5′-d ih yd r o-1′H-im id a zol-2′-
yl)p ip er a zin e (7b). Ethylenediamine (4.5 mL, 0.067 mol)
was added dropwise to a stirred solution of trimethylaluminum
(58 mL, 0.17 mol) in toluene (70 mL) at 0 °C. After the
addition, the reaction mixture was heated at 50 °C and 6b
(9.8 g, 0.033 mol) in toluene (70 mL) was gradually added. The
mixture was refluxed for 5 h (under an argon atmosphere).
After stirring at room temperature overnight, the mixture was
cooled in an ice bath and treated dropwise with water (200
mL) diluted with methanol (80 mL). After filtration and
solvent evaporation the residue was taken up in dichlo-
romethane and washed with water. The organic layer was
washed with dilute HCl, and the aqueous layer was neutral-
ized by NaHCO3 and extracted twice with ether and then with
CH2Cl2 several times. The organic layers were dried over
MgSO4, and the solvents were removed in vacuo. The residue
crystallized in hexane/ether giving 7 g (70%) of 7b as a white
1
1730 (νCdO), 1594 (νCdC) cm-1; H NMR δ 7.41-7.08 (4H, m,
aromatic H), 4.12 (2H, q, J ) 7.1 Hz, COOCH2), 3.56 (2H, s,
CH2Ph), 3.53-2.24 (7H, m, piperazinic H), 1.85 (1H, s, NH),
1.18 (3H, t, J ) 7.1 Hz, CH3).
Gen er a l P r oced u r e for Eth yl 4-(Su bstitu ted ben zyl)-
2-p ip er a zin eca r boxyla te P r ep a r a tion . Meth od A: Eth yl
4-(2′-Ch lor ob en zyl)-1-m et h yl-2-p ip er a zin eca r b oxyla t e
(18b). 18b was prepared via the general procedure as for 6b.
The crude product was dissolved in anhydrous ethanol (200
mL) and saturated with HCl gas. The addition of ether and
chloroform gave a precipitate of 18b (31.2 g, 89%) as a white
powder: mp 155-156.5 °C; IR (film, of the free base) 1750
powder: mp 98-99 °C; IR (in KBr) 3150 (νN-H), 1620 (νCdN
)
cm-1; 1H NMR δ 7.41-7.14 (4H, m, aromatic H), 5.3 (1H, br s,
NH), νa ) 3.75, νb ) 3.39 (2H, q, AB spectrum, J ) 14 Hz,
CH2Ph), 3.79-3.38 (5H, m, imidazolinic and piperazinic H),
2.81-2.12 (6H, 3m, piperazinic H), 2.22 (3H, s, NCH3).
Compounds 7a -i,m -o gave 1H NMR spectra similar to that
of 7b. 1H NMR spectra of other compounds (particular
signals): δ 7j, 3.43 (3H, s, OCH3); 7k , 3.67 (3H, s, OCH3); 7l,
3.34 (6H, s, OCH3); 7p , 2.43 (3H, s, C6H4CH3); 12b, 3.4 (1H,
br s, piperazinic NH); 13b, 0.97 (3H, t, J ) 7.1 Hz, NCH2-
CH3); 22b, 1.01 (3H, t, J ) 7.1 Hz, NCH2CH3); 14b,f, 1.51 (2H,
m, NCH2CH2-CH3), 0.82 (3H, t, J ) 7.2 Hz, N(CH2)2CH3); 15b,
0.96 (3H, d, J ) 6.5 Hz, CHCH3); 0.95 (3H, d, J ) 6.5 Hz,
CH-CH3); 16f, 1.98 (2H, d, J ) 7.1 Hz, CH2N), 1.69-1.63 (1H,
m, CH), 0.79 (6H, d, J ) 6.5 Hz, CH-(CH3)2). The synthesized
imidazolines are listed in Tables 5 and 6.
1
(νCdO), 1600, 1580 (νCdC) cm-1; H NMR δ 7.39-7.08 (4H, m,
aromatic H), 4.3 (2H, q, J ) 7.1 Hz, COOCH2), 3.56 (2H, s,
CH2Ph), 3.0-2.35 (7H, m, piperazinic H), 2.29 (3H, s, NCH3),
1.19 (3H, t, J ) 7.1 Hz, CH3).
Meth od B: Eth yl 4-(2′-Ch lor oben zyl)-1-eth yl-2-p ip er a -
zin eca r boxyla te (20b). 20b was prepared via the general
procedure as for 4b. The crude product was purified by column
chromatography using petroleum ether:ether (90:10, v/v) as
eluent giving 20b as a yelow oil: IR (film) 1745 (νCdO), 1600
(νCdC) cm-1 1H NMR δ 7.4-7.08 (4H, m, aromatic H), 4.11
;
1H NMR spectra of *1-benzyl derivative esters or imidazo-
lines (series I) gave an AB system due to the CH2 group. In
(2H, q, J ) 7.1 Hz, COOCH2), 3.55 (2H, s, CH2Ph), 3.24-2.27
(9H, m, piperazinic H, NCH2), 1.17 (3H, t, J ) 7.1 Hz,
COOCH2CH3), 1.02 (3H, t, J ) 7 Hz, NCH2CH3).
1
series II (**4-benzyl derivatives) these H signals are chemi-
cally equivalent and spectra showed a singlet. In each series,
other signals are similar.
Gen er a l P r oced u r e for Eth yl 1,4-Bis(su bstitu ted ben -
zyl)-2-p ip er a zin eca r boxyla te P r ep a r a tion : Eth yl 1,4-Bis-
(2′-ch lor oben zyl)-2-p ip er a zin eca r boxyla te (24b). It was
prepared via the general procedure as for 17b using 2 equiv
of 2-chlorobenzyl bromide. The crude product was purified by
crystallization of the hydrochloride salt. After basification 26
g (80%) of 24b was obtained as a colorless oil: IR (film, of free
P h ar m acology. An im als an d Tr eatm en ts. Three-month-
old male Wistar rats (Iffa-Credo, France) weighing about 250
g were used in all the experiments. The animals were housed
in wire-bottomed cages and maintained at 21 ( 2 °C in a room
with a 12 h fixed light-dark schedule. Water and standard
laboratory chow (UAR, Villemoisson-sur-Orge, France) were
freely available.
base) 1745 (νCdO), 1605, 1585 (νCdC) cm-1 1H NMR δ 7.46-
;
7.05 (8H, m, aromatic H), 4.11 (2H, q, J ) 7.2 Hz, COOCH2),
νa ) 3.91, νb ) 3.75 (2H, q, AB spectrum, J ) 14.6 Hz, I2/I1 )
2.21, CH2Ph), 3.53 (2H, s, PhCH2), 3.39-3.35, 3.18-3.08,
2.92-2.84, 2.60-2.39 (7H, 4m, piperazinic H), 1.16 (3H, t, J
) 7.2 Hz, CH3).
Moderate diabetes was obtained by iv injection of a low dose
(35 mg/kg) of streptozotocin (STZ) dissolved in a citrate buffer
(Thibault et al.26) under ketamine hydrochloride anesthesia
(75 mg/kg ip; Imalgene, Me´rieux, France). These rats were

3802 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23
Rondu et al.
called STZ rats. Control rats received an injection of citrate
buffer under the same conditions. Glucose homeostasis and
insulin secretion were assessed by a glucose tolerance test
performed 2 weeks after STZ injection.
carried out for 45 min at 25 °C. Nonspecific binding was
defined in the presence of yohimbine (10 µM) in [3H]RX 821002
binding assays, either phentolamine (10 µM) or guanabenz (5
µM) in [3H]-p-aminoclonidine binding assays, and either
tolazoline (10 µM) or amiloride (10 µM) in [3H]idazoxan binding
assays. For each drug, six concentrations from 10-4 to 10-11
M were used in triplicate. Incubations were terminated by
vacuum filtration over Whatman GF/B glass fiber filters using
a cell harvester. The filters were washed three times with
the buffer, placed in scintillation vials, covered with 2 mL of
scintillation cocktail (Pico-Fluor, Packard Instrument), and
counted (Packard 2000 CA). Protein was assayed by the
Bradford method.
Binding results were analyzed by linear regression, and the
curves were obtained with Graph PAD program (Institute for
Scientific Information, Philadelphia, PA). Ki was calculated
with the Cheng-Prusoff equation.
An a lytica l Meth od s. Plasma glucose was determined
using a glucose analyzer (Beckman Inc, Fullerton, CA).
Plasma immunoreactive insulin (IRI) concentration was de-
termined with a radioimmunoassay kit (CEA, Gif-sur-Yvette,
France). The lower limit of the assay was 19.5 pmol/L with a
coefficient of variation within and between assays of 6%.
In tr a ven ou s Glu cose Toler a n ce Test (IVGTT). Glucose
was dissolved in 0.9% saline and given by the saphenous vein
route (0.5 g/kg) in rats under pentobarbital anesthesia (60 mg/
kg ip; Clin-Midy, France). Blood samples were collected
sequentially from the tail veins before and 5, 10, 15, 20, and
30 min after the injection of glucose. They were then centri-
fuged, and the plasma was separated. Plasma glucose con-
centration was immediately determined in a 10 µL aliquot,
and the plasma left was kept at -20 °C until radioimmunoas-
say of insulin.
Dr u g Ad m in istr a tion a n d An tid ia betic Activity. The
2-(4′,5′-dihydro-1′H-imidazol-2′-yl)piperazine derivatives were
tested by a single intraperitoneal administration at 100 µmol/
kg in the STZ rat 20 min before the IVGTT. All the compounds
were used in the form of hydrochloride salts and were water-
soluble. Antidiabetic activity of the compounds was evaluated
using two parameters: ∆G, which represents the increase in
glycemia over baseline integrated over a period of 30 min
(IVGTT) following the glucose load, and K, which is the rate
of glucose disappearance between 5 and 30 min (in the case of
IVGTT), after glucose administration. The K coefficient was
calculated only during IVGTT. Insulin secretion during
IVGTT (∆I) was calculated as the incremental plasma insulin
values over baseline integrated over 30 min after the glucose
load.
Results in Figures 3-5 are expressed as a mean ( SEM.
The significance of differences between means was evaluated
by a one-way analysis of variance (ANOVA), and results were
considered significant at p < 0.05.
Tissu e a n d Mem br a n e P r ep a r a tion . Cerebral cortex
was obtained from whole bovine brains and homogenized in
20 volumes of ice-cold 50 mM Tris-HCl buffer (pH 7.4). The
homogenate was centrifuged twice at 48000g for 25 min at 4
°C. The pellet (used for R2 binding assays) was resuspended
in a phosphate buffer (pH 7.4), flash-frozen, and stored at
-80 °C untilneeded for R2 binding assays.
Reticular nucleus from calf bulbis was homogenized in ice-
cold 50 mM Tris-HCl buffer (pH 7.7) containing 5 mM EDTA.
The homogenate was centrifuged at 500g for 10 min at 4 °C.
The pellet (P1) was resuspended in the same buffer and
centrifuged again. The combined supernatants were centri-
fuged at 50000g for 25 min at 4 °C. The resulting pellet (P2)
was resuspended in 50 mM Tris-HCl buffer (pH 7.7) containing
0.1 mM p-methanesulfonyl fluoride, incubated for 30 min at
25 °C, then centrifuged again in the same conditions, resus-
pended in 50 mM Tris-HCl (pH 7.7), flash-frozen, and stored
at -80 °C until I1 imidazoline binding assays.
Renal cortex was obtained from male New Zealand white
rabbits and homogenized in ice-cold preparation buffer (20 mM
NaHCO3) (Coupry et al.39). The homogenate was centrifuged
at 40000g for 30 min at 4 °C. The pellet was resuspended in
50 mM Tris-HCl buffer containing 0.5 mM EDTA (pH 7.4),
centrifuged again, resuspended in the same buffer, flash-
frozen, and stored at -80 °C until I2 imidazoline binding
assays.
Ack n ow led gm en t. A portion of the calculations was
performed on a IBM RS 6000 cluster at the Computer
Center IDRIS-CNRS. The UNIX version of GAUSSIAN
is implemented on the FPS 522 installed at the Centre
d’Inge´nierie des prote´ines, University of Liege. Potential
electrostatic contours were drawn on an IBM RS 6000
computer using BIOSYM package software at the CCR
Paris 6-Pierre et Marie Curie/Paris 7-Denis Diderot
Universities. This work was supported in part by the
Belgian Programme on Interuniversity Poles of Attrac-
tion initiated by the Belgian State, Prime Minister’s
Office, Services Fe´de´raux des Affaires Scientifiques,
Techniques et Culturelles (PAI No. 9), and the Fonds
de la Recherche Me´dicale (Contract No. 3.4531.92).
Georges Dive is Chercheur Qualifie´ of the FNRS,
Brussels.
Refer en ces
(1) Schmid-Antomarchi, H. S.; De Weille, J .; Fosset, M.; Lazdunski,
M. The receptor for antidiabetic sulfonylureas controls the
activity of the ATP-modulated K+ channel in insulin-secreting
cells. J . Biol. Chem. 1987, 262, 15840-15844.
(2) Bailey, C. J .; Nattrass, M. Treatment - Metformin. Bailiere’s
Clin. Endocrinol. Metab. 1988, 2, 455-476.
(3) J ackson, R. A.; Mawa, M. I.; J aspan, J . B.; Sim, B. M.; Di Silvio,
L.; Featherbe, D.; Kurtz, A. B. Mechanisms of metformin action
in non-insulin-dependent diabetes. Diabetes 1987, 36, 632-640.
(4) Hermann, L. S.; Melander, A. Biguanides: Basic aspects and
clinical uses. In International Textbook of Diabetes Mellitus;
Alberti, K. G. M. M., DeFronzo, R. A., Keen, H., Zimmet, P., Eds.;
J ohn Wiley & Sons Ltd.: New York 1992, pp 773-795.
(5) Campbell, I. W. Sulphonylureas and Metformin: Efficacy and
Inadequacy in New Antidiabetic Drugs; Bailey, C. J ., Flatt, P.
R., Eds.; Smith-Gordon: London, 1990, pp 33-51.
(6) Berger, W. Incidence des effets secondaires se´ve`res sous trait-
ement par des sulfonylure´es et des biguanides. (Incidence of
severe side effects under sulfonylureas and biguanides treat-
ment.) J . Diabe´tol. Hoˆtel-Dieu, Flammation Ed., Paris 1986,
243-253.
Bin d in g Assa ys. [3H]RX 821002 (48 Ci/mmol), [3H]-p-
aminoclonidine (50 Ci/mmol), and [3H]idazoxan (42 Ci/mmol)
were obtained from Amersham (Buckinghamshire, U.K.) and
phentolamine, guanabenz, tolazoline, idazoxan, yohimbine,
clonidine, naphazoline, and amiloride from Sigma (St. Louis,
MO). Radioligand binding assays with [3H]RX 821002, [3H]-
p-aminoclonidine, or [3H]idazoxan for determination of specific
binding to R2-adrenoceptors and I1 and I2 imidazoline-binding
sites, respectively, were performed by a modification of meth-
ods previously described (Coupry et al.,39 Reis et al.,40 Uhlen
and Wikberg41). Membranes were slowly thawed and diluted
to a concentration of 0.4 mg of protein/mL for renal and
cerebral cortex and 0.7 mg of protein/mL for calf bulbis.
(7) Kameda, K.; Ono, S.; Aboko, Y. Hypoglycemic action of 2-[2-(4′,5′-
dihydro-1′H-imidazol-2′-yl)]-1-phenylethylpyridine dihydrochlo-
ride sesquihydrate (DG-5128), a new hypoglycemic agent. Arz-
neim. Forsch. Drug Res. 1982, 32, 39-44.
(8) Kashiwagi, A.; Harano, Y.; Susuki, M.; Kojima, H.; Harada, M.;
Nishio, Y.; Shiegeta, Y. New R2-adrenergic blocker (DG-5128)
improves insulin secretion and in vivo glucose disposal in
NIDDM patients. Diabetes 1986, 35, 1085-1089.
(9) Use of pyrrolo[3,2,1-hi]indole derivatives in the manufacture of
a
medicament for the treatment of diabetes. EP 0311467;
Synthelabo.
(10) Annual Drug Data Report 1992, 14, 995.
(11) J onas, J . C.; Garcia-Banado, M. J .; Angel, I.; Henquin, J . C. The
imidazoline SL 84.0148 shows stereoselectivity in blocking R2-
adrenoceptors but not ATP-sensitive K+ channels in pancreatic
Β-cells. Eur. J . Pharmacol. 1994, 264, 81-84.
Assays were conducted in
a total volume of 525 µL in
polypropylene tubes, and each tube contained 500 µL of
membrane suspension, 20 µL of radioligand, and 5 µL of drug.
Incubation was initiated by the addition of membrane and

Imidazolines Active on Glucose Homeostasis
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23 3803
(12) Berritge, T. L.; Doxey, J . C.; Roach, A. G. Comparison of the
effects of efaroxan and glibenclamide on plasma glucose and
insulin levels in rats. Eur. J . Pharmacol. 1992, 213, 213-218.
(13) O¨ stenson, C. G.; Cattaneo, A. G.; Doxey, J . C.; Efendic, S. Alpha-
adrenoceptors and insulin release from pancreatic islets of
normal and diabetic rats. Am. J . Physiol. 1989, 257, E439-E443.
(14) Chan, S. L. F. Role of alpha-2 adrenoceptors and imidazoline-
binding sites in the control of insulin secretion. Clin. Sci. 1993,
85, 671-677.
(15) Schultz, A.; Hasselblatt, A. An insulin-releasing property of
imidazoline derivatives is not limited to compounds that block
R-adrenoceptors. Naunyn-Schmiedeberg’s Arch Pharmacol. 1989,
340, 321-327.
(16) Michel, M. C.; Brodde, O. E.; Schnepel, B.; Behrendt, T.; Tschada,
R.; Motulsky, H. J .; Insel, P. A. (3H)Idazoxan and some other
alpha-2-adrenergic drugs also bind with high affinity to a non
adrenergic site. Mol. Pharmacol. 1989, 35, 324-330.
(17) Atlas, D.; Diamant, S.; Zonnenschein, R. Is imidazoline site a
unique receptor? Am. J . Hyper. 1992, 5, 83S-90S.
(18) Ruffolo, R. R. H.; Bondinelli, W. E.; Hieble, J . P. R- and â-
Adrenoceptors: From the gene to the clinic. Structure-activity
relationships and therapeutic applications. J . Med. Chem. 1995,
38, 3681-3716.
(19) Wang, X.; Rondu, F.; Lamouri, A.; Dokhan, R.; Marc, S.; Touboul,
E.; Pfeiffer, B.; Manechez, D.; Renard, P.; Guardiola-Lemaitre,
B.; Godfroid, J . J .; Ktorza, A.; Penicaud, L. Effect of S-21663
(PMS 812), an imidazoline derivative, on glucose tolerance and
insulin secretion in a rat model of type II diabetes. J . Pharmacol.
Exp. Ther. 1996, 278, 82-89.
(20) 6-31G results (see ref 38). All 24 combinations of the mono- and/
or diprotonations of the molecule were studied. These results
will be published elsewhere and are available if requested.
(21) Frisch, M. J .; Trucks, G. W.; Head-Gordon, M.; Gill, P. M. W.;
Wong, M. W.; Foresman, J . B.; J ohnson, B. G.; Schlegel, H. B.;
Robb, M. A.; Replogle, E. S.; Gomperts, J . L.; Andres, J . L.;
Raghavachari, K.; Binkley, J . S.; Gonzalez, C.; Martin, R. L.;
Fox, D. J .; Defrees, D. J .; Baker, J .; Stewart, J . J . P.; Pople, J .
A. GAUSSIAN 94; Gaussian Inc.: Pittsburgh, PA, 1992.
(22) J ucker, E.; Rissi, E. U¨ ber C-substituerte piperazinederivate. (On
C-substituted piperazine derivatives.) Helv. Chem. Acta 1962,
273, 2383-2401.
(23) J erzy, L.; Pietraseiwick, M.; Duszynska, B.; Cegla, M. T.
Structure-activity relationship studies of central nervous system
agents. Effect of the hydrocarbon chain on the affinity of
4-substituted 1-(3-chlorophenyl)piperazines for 5-HT1A receptor
site. J . Med. Chem. 1992, 35, 2369-2374.
(24) Icke, R.; Wisegarver, N. B. â-Phenylethyldimethylamine. Organic
Syntheses; Wiley: New York, 1955; Collect. Vol. 3, pp 723.
(25) Neef, G.; Eder, U.; Sauer, G. One step conversions of esters to
2-imidazolines, benzimidazoles, and benzothiazoles by aluminum
organic reagents. J . Org. Chem. 1981, 46, 2824.
(28) Chan, S. L. F.; Brown, C. A.; Scarpello, K. E.; Morgan, N. G.;
Pancreatic â-cells express an imidazoline binding site which is
distinct from I1- and I2-sites. Int. Symp. Imidazoline Receptors
1994, P3.
(29) Chan, S. L. F.; Dunne, M. J .; Stillings, M. R.; Morgan, N. G.
The alpha-2 adrenoceptor antagonist efaroxan modulates K+/
ATP channels in insulin-secreting cells. Eur. J . Pharmacol. 1991,
204, 41-48.
(30) Chan, S. L. F.; Hurst, R. D.; Morgan, N. G. Effects of alpha-2
adrenergic antagonists on insulin secretion from rat pancreatic
islets. Biochem. Soc. Trans. 1988, 16, 1005-1006.
(31) Chan, S. L. F.; Morgan, N. G. Stimulation of insulin secretion
by efaroxan may involve interaction with potassium channels.
Eur. J . Pharmacol. 1990, 176, 97-101.
(32) Chan, S. L. F.; Stillings, M. R.; Morgan, N. G. Mechanisms
involved in stimulation of insulin secretion by the hypoglycemic
alpha-adrenergic antagonist, DG-5128. Biochem. Biophys. Res.
Commun. 1991, 176, 1545-1551.
(33) Olmos, G.; Kulkarni, T. N.; Haque, M.; MacDermot, J . Imida-
zolines stimulate release of insulin from RIN-5AH cells inde-
pendently from I1 and I2 receptors. Eur. J . Pharmacol. 1994,
262, 41-48.
(34) Brown, C. A.; Lowett, A. C.; Smith, S. A.; Morgan, N. G.;
Stimulation of insulin secretion by imidazoline compounds is not
due to interaction with non-adrenoceptor idazoxan binding sites.
Br. J . Pharmacol. 1993, 108, 312-317.
(35) Renouard, A.; Widdowson, P. S.; Cordi, A. (3H)-Idazoxan binding
to rabbit cortex recognizes multiple imidazoline I2-type recep-
tors: pharmacological characterization and relationship to
monoamine oxidase. Br. J . Pharmacol. 1993, 109, 625-631.
(36) Dewar, M. J . S.; Zoebish, E. G.; Healy, E. F. A new general
purpose quantium mechanical molecular model. J . Am. Chem.
Soc. 1985, 107, 3902.
(37) Dive, G.; Dehareng, D.; Ghuysen, J . M. Energy analysis on small
to medium sized H-bonded complexes. Theoret. Chim. Acta 1993,
85, 409-421.
(38) Hehre, W. J .; Ditchfield, R.; Pople, J . A. Theoretical investiga-
tions on the solvation process. J . Chem. Phys. 1972, 56, 2557-
2562.
(39) Coupry, I.; Podevin, R. A.; Dausse, J . P.; Parini, A. Evidence for
imidazoline binding sites in basolateral membranes from rabbit
kidney. Biochem. Biophys. Res. Commun. 1987, 147, 1055-1060.
(40) Reis, D. J .; Reunathan, S.; Meeley, P. M. P. Imidazoline receptors
and clonidine-displacing substance: implication in regulation of
arterial pressure, neuroprotection and medullo-adrenal secre-
tion. Am. J . Hyper. 1992, 5, 45S-46S.
(41) Uhlen, S.; Wikberg, J . Delineation of rat kidney alpha-2A and
alpha-2B adrenoceptors with 3H-RX821002 radioligand bind-
ing: computer modelling reveals that guanfacine is an alpha-
2A selective compound. Eur. J . Pharmacol. 1991, 202, 235-243.
(26) Thibault, C.; Laury, M. C.; Bailbe, D.; Ktorza, A.; Portha, B.
Effect of prolonged in vivo glucose infusion on insulin secretion
in rats with previous glucose intolerance. Endocrinology 1992,
130, 2521-2527.
(27) Chan, S. L. F.; Brown, C. A.; Morgan, N. G. Stimulation of insulin
secretion by the imidazoline alpha-2 adrenoceptor antagonist
efaroxan is mediated by a novel, stereoselective, binding site.
Eur. J . Pharmacol. 1993, 230, 375-378.
J M9608624