Mechanism of inhibition of tumor necrosis factor production by chlorpromazine and its derivatives in mice

Article abstract of DOI:10.1016/S0014-2999(96)00728-5

In previous work, we reported that chlorpromazine inhibits tumor necrosis factor (TNF) production in endotoxin lipopolysaccharide-treated mice, and protects against lipopolysaccharide toxicity. Chlorpromazine is used as an antipsychotic and has several ef

Full text of DOI:10.1016/S0014-2999(96)00728-5

Ž
.
European Journal of Pharmacology 317 1996 369–376
Mechanism of inhibition of tumor necrosis factor production by
chlorpromazine and its derivatives in mice
a
a
a
Pietro Ghezzi a,), Silvio Garattini , Tiziana Mennini , Riccardo Bertini ,
a
a
a
a
Rene Delgado Hernandez , Fabio Benigni , Silvano Sacco , Malgorzata Skorupska ,
Manuela Mengozzi , Roberto Latini , Mami Kurosaki , Alain Lombet , Arnel Fradin ,
´
a
a
a
b
b
b
b
b
Jacqueline Bonnet , Yves Rolland , Jean-Daniel Brion
Laboratory of Neuroimmunology, Istituto di Ricerche Farmacologiche ‘Mario Negri’, 20157 Milan, Italy
a
b
Institut de Recherches SerÕier, Suresnes, France
Received 5 August 1996; revised 5 September 1996; accepted 6 September 1996
Abstract
Ž
.
In previous work, we reported that chlorpromazine inhibits tumor necrosis factor TNF production in endotoxin lipopolysaccharide-
treated mice, and protects against lipopolysaccharide toxicity. Chlorpromazine is used as an antipsychotic and has several effects on the
central nervous system. It acts on different neurotransmitter receptors and has other biochemical activities some of which, like inhibition
of phospholipase A2, might be responsible for the inhibitory effect on TNF production. To investigate the role of these actions in the
inhibition of TNF production by chlorpromazine, we have synthesized some chlorpromazine derivatives that do not have central activities.
Some of these analogs have lost their affinity for various receptors and their phospholipase A2 inhibitory activity, but still inhibit TNF
Ž
.
production. No correlation was found between TNF inhibition and the ability to inhibit nitric oxide NO synthase, whereas a good
correlation was evident between TNF inhibition and antioxidant activity.
Ž
.
Keywords: Chlorpromazine; TNF tumor necrosis factor ; Endotoxin
1. Introduction
Our interest was to identify the mechanism of this
inhibitory action. Chlorpromazine has a large number of
pharmacological activities: it is a histamine receptor antag-
onist, 5-HT receptor antagonist, adrenoceptor antagonist,
dopamine receptor antagonist, phospholipase A2 inhibitor
As early as 1954, chlorpromazine was reported to pro-
tect against endotoxic shock Chedid, 1954 . In a series of
studies, we have characterized this effect and observed that
Ž
.
´
Ž
.
protection against lipopolysaccharide toxicity was associ-
Baldessarini, 1985; Vadas et al., 1986 , and antioxidant
Ž
.
Ž
.
ated with inhibition of tumor necrosis factor TNF pro-
Slater, 1968; Jeding et al., 1995 . Some of these actions
might explain its inhibition of TNF production, and in fact,
other 5-HT receptor antagonists, a- not-b adrenoceptor
antagonists, histamine receptor antagonists and phospholi-
pase A2 inhibitors inhibit lipopolysaccharide-induced TNF
Ž
.
duction Gadina et al., 1991; Mengozzi et al., 1994 . Since
TNF is a key mediator in the lethal action of lipopoly-
saccharide, as demonstrated by the protective effect of
anti-TNF antibodies or inhibitors of TNF production like
Ž
.
Ž
Ž
.
glucocorticoids Bertini et al., 1989; Beutler and Cerami,
production in vivo Basedovsky and Del Rey, 1987 . How-
ever, the relative importance of the sedative and psy-
chotropic actions of chlorpromazine and, biochemically, of
its neurotransmitter receptor binding, has not been clari-
fied.
.
1986; Zimmerman et al., 1989 , this inhibitory action of
chlorpromazine might well be at the basis of its protective
effect.
In the present study we used a different approach,
synthesizing different analogs of chlorpromazine and ana-
lyzing their ability to inhibit TNF production. Some qua-
)
Ž
.
Corresponding author. Fax: 39-2 354-6277.
0014-2999r96r$15.00 Copyright q 1996 Elsevier Science B.V. All rights reserved.

(
)
370
P. Ghezzi et al.rEuropean Journal of Pharmacology 317 1996 369–376
ternary derivatives of chlorpromazine were synthesized in
order to obtain molecules that could not pass the blood–
brain barrier and therefore had no activity on the central
2. Materials and methods
2.1. Chlorpromazine deriÕatiÕes
Ž
.
nervous system CNS . These chlorpromazine derivatives
were synthesized on the basis of earlier reports or patents
indicating that quaternary derivatives or sulphones have no
The quaternary ammonium compounds 1 and 2 were
synthesized according to classical methods involving treat-
ment of chlorpromazine with bromoalkanes. Nitric acid
oxidation led to the formation of sulfoxide nitrates which,
after neutralization, were quaternized to bromoalkanes to
afford compounds 3 and 4. Compound 5 was obtained
from 4-chloro-2X-nitrodiphenylsulfide, which was reduced
Ž
.
pharmacological activity Costentin et al., 1987 . The com-
pounds were then characterized with respect to some in
vivo and in vitro activities to gain an idea which phar-
macological activities of chlorpromazine are responsible
for its inhibition of TNF production. The lack of central
activity of some of these analogs was checked using a
standard locomotor activity test to evaluate the sedative
activity.
Ž
.
Ž
.
by ZnrAcOH and treated with 3-chloropropyl dimeth-
ylamine under alkaline conditions. The structure of the
compounds investigated is shown in Fig. 1. They include
Ž
w Ž
Their effect on blood pressure was also tested, to see
whether they still had the hypotensive activity of chlor-
promazine. For technical reasons, blood pressure was stud-
ied in rats. These studies were done to clarify whether the
effect of chlorpromazine on TNF production was separate
from its central and cardiovascular effects.
two quaternary derivatives, compound 1 N- 3- 2-chlo-
.
x
rphenothiazin-10-yl propyl trymethyl-ammonium bro-
w Ž
.
Ž
.
mide and compound 2 N- 3- 2-chlorphenothiazin-10-yl
x
.
propyl butyl-dimethylammonium bromide ; two sulfox-
w Ž
Ž
ide-quaternary ammoniums, compound 3 N- 3- 2-chloro-
.
x
5-oxide phenothiazin-10-yl propyl butyl-dimethylam-
w Ž
.
Ž
Since phospholipase A2 inhibitors also inhibit TNF
monium bromide and compound 4 N- 3- 2-chloro-5-
Ž
.
.
x
production Spriggs et al., 1990 , and chlorpromazine is a
well-known inhibitor of this enzyme, chlorpromazine and
its analogs were tested for their ability to inhibit purified
phospholipase A2 in vitro. Inhibition of nitric oxide syn-
thase was also investigated, since some reports suggest
oxide phenothiazin-10-yl propyl trimethylammonium
bromide ; and an open-ring derivative, compound 5 N-2-
4-chlorophenyl thio phenyl N,N -dimethyl propane 1,3-
.
Ž
X
Ž
.
x
.
diammonium chloride .
Ž
.
nitric oxide NO is involved in lipopolysaccharide-induced
2.2. Animals and treatments
Ž
.
TNF production Moncada et al., 1991; Rojas et al., 1993 .
Finally, since other antioxidants have been shown to in-
hibit TNF production Peristeris et al., 1992; Zhang et al.,
1994 , and chlorpromazine is described to be a scavenger
of hydroxyl radicals Jeding et al., 1995; Slater, 1968 , we
also investigated the antioxidant activity of chlorpromazine
and its derivatives using an in vitro test for antioxidant
Ž
.
Male CD-1 mice 25 g body weight from Charles
Ž
Ž
.
River Calco, Como, Italy; Saint-Aubin, France were
used. Mice were housed five per cage and could feed ad
libitum. Dexamethasone phosphate was from Istituto Far-
macologico Milanese, Milan, Italy; lipopolysaccharide
.
Ž
.
Ž
phenol-extracted preparation from Escherichia coli
Ž
.
.
activity Halliwell et al., 1987; Aebi, 1974 .
O55:B5 was from Difco.
Ž
.
Fig. 1. Structure of chlorpromazine CPZ and of the synthetic CPZ derivatives.

(
)
P. Ghezzi et al.rEuropean Journal of Pharmacology 317 1996 369–376
371
Procedures involving animals and their care were con-
ducted in conformity with institutional guidelines that are
in compliance with national and international laws and
define specific binding are summarized in Table 1. Bind-
ing in the presence of various concentrations of test com-
pounds was expressed as a percentage of the specific
binding without drug. The IC50 values were established by
simultaneous nonlinear curve fitting according to the logis-
Ž
policies EEC Council Directive 86r609, OJ L 358, 1,
December 12, 1987; Italian Legislative Decree 116r92,
Gazzetta Ufficiale della Repubblica Italiana No. 40, Febru-
ary 18, 1992; NIH Guide for the Care and Use of Labora-
Ž
.
tic equation reported by De Lean et al. 1978 , using the
3
3
w
x
w
x
‘Allfit’ program.
H Prazosin,
H 8 OH-DPAT,
H SCH 23390,
w x
H 5-HT, and
3
3
3
3
.
w
x
w
x
tory Animals, NIH Publication No. 85-23, 1985 .
H spiperone,
x
w
For lipopolysaccharide toxicity studies, male, adult
H paroxetine were purchased from New England Nu-
Ž
.
22–24 g BALBrc mice were given lipopolysaccharide
clear.
as a single dose of 1 mgrmouse, i.p., with and without a
30-min pretreatment with the test compound, as previously
2.5. Locomotor actiÕity test
Ž
.
described Gadina et al., 1991 . Survival was evaluated 72
h later.
The sedative action of chlorpromazine and its deriva-
tives was measured by recording the spontaneous locomo-
tor activity of mice placed in a box representing a new
environment, equipped with infra-red sensors to measure
2.3. Inhibition of TNF production
Ž
In some experiments mice were treated i.p. with chlor-
the number of movements Activity Monitor, Digiscan-
Ž
.
Ž
.
promazine 4 mgrkg or its derivatives given at equimo-
Omnitech Electronics, Sufraco, France . Test compounds
.
lar doses to 4 mgrkg chlorpromazine . Thirty minutes
were injected i.p. 30 min before the test, at the doses
indicated, and locomotor activity was then evaluated for 15
min.
later, mice received an i.p. injection of 2.5 mgrmouse of
lipopolysaccharide. Blood was obtained from the retro-
orbital plexus under light ether anesthesia and serum was
prepared. Serum TNF was measured in a cytotoxicity
assay, using L929 cells in the presence of 1 mgrml of
2.6. Inhibition of phospholipase A2 in Õitro
Ž
Ž
.
actinomycin D, as previously described Aggarwal et al.,
Porcine pancreatic phospholipase A2 Sigma was incu-
.
1985 , with human recombinant TNF as standard.
bated for 1 h at 378C at a concentration of 5 ngrml with
In another set of experiments, mice were given the test
drugs orally, dissolved in water. Thirty minutes later, mice
were injected i.v. with 250 mg of lipopolysaccharide in
saline. One hour later, mice were anesthetized with isoflu-
60 000–80 000 cpmrsample of Escherichia coli mem-
3
w
x
branes radiolabelled with H oleic acid in a final volume
Ž
of 0.4 ml, as previously described Boraschi et al., 1991;
Davidson et al., 1987 . Test compounds were used at
various concentrations; water-insoluble compounds were
dissolved in dimethyl sulfoxide DMSO at a final concen-
tration of 5%. Inhibition of enzyme activity was calculated
in control samples containing equal amounts of DMSO.
.
Ž
.
rane Forene, Abbott Laboratories , and blood was ob-
Ž
Ž
.
tained by cardiac puncture and collected on heparin 50
Urml, Choay, France . Plasma TNF-a was assayed using
a commercially available enzyme immunosorbent assay
.
Ž
.
Ž
.
ELISA kit Genzyme .
2.7. Inhibition of NO synthase in Õitro
2.4. Neurotransmitter receptor binding affinity
Neuronal NO synthase activity was measured according
Ž
.
Binding assays were done on crude membrane prepara-
to Bredt and Snyder 1989 . Briefly, cerebella from Wistar
Ž
Ž
.
tions from rat brain areas, as described Mennini et al.,
rats 180–200 g were rapidly removed, cleaned and
minced in 20 volumes of ice-cold buffer 20 mM
.
Ž
1991; Testa et al., 1989 . The type of receptors analyzed,
the brain area used, radioligands and competitors used to
HEPESrTris pH 7.2 containing 0.32 M sucrose, 0.5 mM
Table 1
Binding assays: summary of methods
Ž
.
Ž
.
Receptor
Localization
Radioligand nM
Competitor nM
Ž .
Phentolamine 3
3
w
w
w
w
w
w
x
Ž
.
a-1 NA
DA-1
DA-2
5-HT-1A
5-HT-1B
5-HT uptake
Cortex
H Prazosin 0.2
H SCH23390 0.4
H Spiperone 0.2
H 8-OH-DPAT 1.0
H 5-HT 2.0
H Paroxetine 0.4
3
3
3
3
3
x
Ž
.
Ž
Ž
.
.
Ž .
Striatum
Striatum
Hippocampus
Striatum
Cortex
y cis-Flupentixol 10
y Sulpiride 100
x
Ž
.
Ž
.
x
Ž
.
Ž .
5-HT 1
x
Ž
.
Ž .
5-HT 10
x
Ž
.
Ž .
Fluoxetine 10
w
Ž .
.
5-HT, serotonin; DA-1, dopamine D1 receptor; DA-2, dopamine D2 receptor; NA, noradrenaline; 8-OH-DPAT, N,N-di 2,3 n -3H-propylamino -8-hy-
droxy-1,2,3,4-tetrahydronaphthalene.

(
)
372
P. Ghezzi et al.rEuropean Journal of Pharmacology 317 1996 369–376
Table 2
.
EDTA and 1 mM dithiothreitol . After homogenization
with a Dounce homogenizer, samples were centrifuged at
Effect of intraperitoneal administration of chlorpromazine and its deriva-
tives on TNF production in vivo
9000=g for 20 min and the supernatant was passed
through a column of 1 ml of Dowex AG50WX8 Na
form to remove endogenous arginine. The assay was
carried out in a total volume of 0.5 ml. The column eluate
Ž
.
q
Compound
TNF ngrml
Ž
Salineq
lipopolysaccharide
Compoundq
.
lipopolysaccharide
23" 4 a y77%
Ž
.
.
.
1
2
3
4
102"26
102"26
44" 8
44" 8
102"26
102"26
Ž
.
0.4 ml was supplemented with 2 mM NADPH, 0.45 mM
11" 4 a y89%
Ž
3
w
x
Ž
.
CaCl2 , 100 mM arginine and H arginine 1 mCirml as
tracer, and the test drug at various concentrations. After 1
h at 308C, two 0.2-ml samples of the incubation mixture
a
Ž
14" 7 y68%
20"11 a y54%
Ž
.
Ž
.
5
81"15 y20%
b
Ž
.
Ž
.
were applied to a Dowex column as above to remove
unreacted substrate, and eluted with 1 ml of water. NO
synthase activity was determined by counting the radioac-
tivity of the eluate containing the radiolabeled citrulline
formed , using 4 ml of Picofluor 40 Packard , in a
scintillation counter.
Chlorpromazine
-0.5
)95%
Ž
w Ž . x
N- 3- 2-chlorphenothiazin-10-yl propyl trymethyl-ammonium bro-
1
.
Ž
w
Ž
.
.
x
mide ; 2 N- 3- 2-chlorphenothiazin-10-yl propyl butyl-dimethylam-
Ž
Ž
w
Ž
.
x
monium bromide ; 3 N- 3- 2-chloro-5-oxide phenothiazin-10-yl propyl
butyl-dimethylammonium bromide ; 4 N- 3- 2-chloro-5-oxide pheno-
thiazin-10-yl propyl trimethylammonium bromide ; 5 N-2- 4-chloro-
.
Ž
.
.
Ž
w Ž
.
x
.
Ž
Ž
X
.
x
.
phenyl thio phenyl N,N -dimethyl propane 1,3-diammonium chloride .
Drugs were administered i.p. in saline 30 min before lipopolysaccharide
2.8. HypotensiÕe actiÕity in rats
Ž
.
Ž
2.5 mgrmouse, i.p. . TNF was measured by bioassay. Data peak serum
.
TNF levels are means"S.E. from 4–6 mice per group. Percent inhibi-
tion is given in parentheses. P-0.05, b P-0.01 versus lipopoly-
a
Systolic blood pressure was measured non-invasively,
Ž
Ž
.
saccharide alone by Dunnett’s test.
using a tail cuff, in adult male CD rats Crl:CD SD BR,
.
215–240 g body weight, Charles River up to 50 min after
chlorpromazine or derivatives. The method is based on
recording the arterial pulse from the caudal artery with a
derivatives on serum TNF production. The test compounds
were administered i.p. 30 min before lipopolysaccharide
Ž
Ž
.
piezoelectric transducer applied to the tail BP recorder
2.5 mgrmouse, i.p. and serum TNF was measured 1 h
.
8005, Basile, Comerio, Italy . Rats were preconditioned
after lipopolysaccharide administration, using the cytotoxi-
for 10 min at 388C to obtain a measurable pulse.
2.9. Antioxidant actiÕity
city assay. No TNF was detectable in the absence of
Ž
.
lipopolysaccharide treatment data not shown . The dose of
Ž
.
chlorpromazine 4 mgrkg, equal to 12.6 mmolrkg and
the treatment schedule were chosen on the basis of previ-
Ž
.
The deoxyribose assay for hydroxyl radicals was per-
ously published data Gadina et al., 1991 . The TNF
Ž
.
Ž
.
formed as already described Halliwell et al., 1987 .
Briefly, the reaction was carried out in a total volume of 1
production of control groups lipopolysaccharide alone
obtained within the same experiment is reported for each
Ž
.
ml of potassium phosphate buffer 20 mM, pH 7.4 con-
compound. Chlorpromazine completely blocked TNF pro-
Ž
.
.
Ž
.
Ž
Ž
.
taining deoxyribose 5 mM , FeCl3 100 mM ; EDTA 104
mM , H2O2 1 mM , ascorbate 100 mM , and the test
duction )95% inhibition in three experiments , in agree-
.
Ž
Ž
.
ment with previous reports. All the chlorpromazine deriva-
drugs at various doses. Reaction mixtures were incubated
Ž
.
at 378C for 1 h. Thiobarbituric acid 1 ml of 1% wrv was
Table 3
Effect of oral administration of chlorpromazine and its derivatives on
TNF production in vivo
Ž
.
then added, plus 1 ml 2.8% wrv trichloroacetic acid.
The tubes were heated for 30 min at 1008C, cooled and the
absorbance was measured at 532 nm against the appropri-
ate blanks.
The possible effect of chlorpromazine and derivatives
on the stability of hydrogen peroxide was evaluated spec-
trophotometrically by monitoring the H2O2 concentration
Ž
.
Compound
TNF ngrml
% Inhibition
Ž
.
Vehicle water
60.9" 4.8
19.7" 6.7
–
68
72
73
60
41
95
a
1
2
3
4
5
a
16.9" 3.4
16.5" 8.5 a
a
24.3" 3.4
36.1"10.2
2.8" 0.4
Ž
.
Ž
.
at 240 nm Aebi, 1974 . Catalase Sigma; 100 mgrml
was used as a reference and caused complete degradation
a
Chlorpromazine
of H2O2 in this system.
Ž
w Ž . x
N- 3- 2-chlorphenothiazin-10-yl propyl trymethyl-ammonium bro-
1
.
Ž
w
Ž
.
x
mide ; 2 N- 3- 2-chlorphenothiazin-10-yl propyl butyl-dimethylam-
monium bromide ; 3 N- 3- 2-chloro-5-oxide phenothiazin-10-yl propyl
butyl-dimethylammonium bromide ; 4 N- 3- 2-chloro-5-oxide pheno-
thiazin-10-yl propyl trimethylammonium bromide ; 5 N-2- 4-chloro-
phenyl thio phenyl N,N -dimethyl propane 1,3-diammonium chloride .
Drugs were administered orally in water 30 min before lipopolysaccha-
ride 250 mgrmouse, i.p. . TNF was measured by ELISA. Data peak
serum TNF levels are means"S.E. from 4–6 mice per group. P-0.01
.
Ž
w
Ž
.
x
3. Results
.
Ž
w Ž
.
x
.
Ž
Ž
X
.
x
.
3.1. Inhibition of TNF production by chlorpromazine and
its deriÕatiÕes
Ž
.
Ž
a
.
Table 2 shows the effect of chlorpromazine and its
vs. control by Student’s t-test.

(
)
P. Ghezzi et al.rEuropean Journal of Pharmacology 317 1996 369–376
373
Ž
.
w Ž
tives, tested at doses equimolar to that of chlorpromazine,
significantly inhibited TNF production, to varying extents.
In a second set of experiments, chlorpromazine and its
derivatives were tested on lipopolysaccharide-induced TNF
production, using a different procedure. Lipopolysaccha-
ride was administered i.p. at the dose of 250 mgrmouse,
the test compounds were administered orally, and TNF
production was measured with an ELISA specific for
TNF-a. Chlorpromazine and all its derivatives lowered
Chlorpomazine 4 mgrkg , N- 3- 2-chlorphenothiazin-10-
.
x
Ž
yl propyl trymethyl-ammonium bromide compound 1, 5
mgrkg , and N- 3- 2-chlorphenothiazin-10-yl propyl
butyl-dimethylammonium bromide compound 2,
mgrkg increased survival 77%, 80% and 40% respec-
tively, compared to 0% survival in mice treated with
lipopolysaccharide alone . N-2- 4-Chlorophenyl thio
phenyl N,N -dimethyl propane 1,3-diammonium chloride
.
w Ž
.
x
5
Ž
.
Ž
.
Ž
.
X
x
Ž
.
compound 5 at doses up to 100 mgrkg was not protec-
tive data not shown .
Ž
Ž
.
serum TNF levels in lipopolysaccharide-treated mice Ta-
.
Ž
x
w
ble 3 . Four selected compounds chlorpromazine; N- 3-
Ž
.
2-chlorphenothiazin-10-yl propyl trymethyl-ammonium
w Ž
3.2. Neurotransmitter receptor binding affinity
Ž
.
.
.
bromide compound 1 ; N- 3- 2-chlorphenothiazin-10-yl
x
Ž
propyl butyl-dimethylammonium bromide compound 2 ;
N-2- 4-chlorophenyl thio phenyl N,N -dimethyl propane
1,3-diammonium chloride compound 5 were also tested
Table 4 reports the affinities of chlorpromazine and
some of its derivatives on the most representative neuro-
transmitter receptors. Generally, the derivatives were less
active than chlorpromazine. For dopamine receptors, chlor-
promazine was about five times more active on dopamine
X
Ž
.
x
Ž
.
for their ability to protect against the lethal effect of
Ž
.
lipopolysaccharide 1 mgrmouse, i.p. in BALBrc mice.
Ž
.
Fig. 2. Hypotensive effect of chlorpromazine in rats and comparison with its derivatives. Rats were injected i.p. with 16 mgrkg chlorpromazine CPZ or
Ž
.
equimolar doses of its derivatives. Systolic blood pressure SBP was monitored up to 50 min. Each curve represents one rat.

(
)
374
P. Ghezzi et al.rEuropean Journal of Pharmacology 317 1996 369–376
Table 4
Neurotransmitter receptor binding affinity of chlorpromazine and its
derivatives
Receptor
Chlorpromazine Compound
1
2
3
4
5
a-1 NA
DA-1
DA-2
5-HT-1A
5-HT-1B
5-HT uptake 7.02
7.87
6.91
7.92
5.01
5.38
6.57 6.40 5.32 4.96 5.92
4.69 4.93 n.a.
5.22 5.76 n.a.
n.a.
n.a.
5.17
5.05
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
6.49 4.87 n.a.
n.a.
n.a.
n.a.
n.a.
5.20
n.a.
Ž
.
Fig. 3. Effect of chlorpromazine CPZ and its derivatives on phospholi-
Ž
w Ž . x
N- 3- 2-chlorphenothiazin-10-yl propyl trymethyl-ammonium bro-
1
Ž
.
pase A2 PLA2 in vitro. The activity of porcine pancreatic phospholi-
pase A2 was assayed by measuring the release of radioactivity from
membranes of E. coli radiolabeled with H oleic acid. Data are means
of triplicate samples variability was -10% . The percentage of control
activity no drug is shown. Typically, 5 ngrml phospholipase A2 in
control samples released 10000–15000 cpm of radioactivity from
60000–80000 cpm of membranes in 1-h incubation. Release of radioac-
tivity in blank samples no phospholipase A2 was 500–1000 cpm.
Quinacrine QNCR was used as a positive control.
.
Ž
w
Ž
.
x
mide ; 2 N- 3- 2-chlorphenothiazin-10-yl propyl butyl-dimethylam-
monium bromide ; 3 N- 3- 2-chloro-5-oxide phenothiazin-10-yl propyl
butyl-dimethylammonium bromide ; 4 N- 3- 2-chloro-5-oxide pheno-
thiazin-10-yl propyl trimethylammonium bromide ; 5 N-2- 4-chloro-
phenyl thio phenyl N,N -dimethyl propane 1,3-diammonium chloride .
5-HT, serotonin; DA-1, dopamine D1 receptor; DA-2, dopamine D2
receptor; NA, noradrenaline. Data are expressed as pIC50 M . Each drug
.
Ž
w
Ž
.
x
3
w
x
.
.
Ž
w Ž
Ž
.
x
.
Ž
Ž
X
Ž
.
.
x
.
Ž
.
Ž
.
Ž
was tested at 6–8 concentrations in triplicate variation coefficients were
Ž
.
-10% . n.a. not active denotes an IC50 higher than 10y4 M.
.
Ž
.
3.4. HypotensiÕe action of chlorpromazine and its deriÕa-
tiÕes in rats
D2 than on dopamine D1 receptor binding. Compounds 1
and 2 showed the same proportion of binding, while
compound 5 was about equipotent. However, they were all
As shown in Fig. 2, the dose of 16 mgrkg chlorproma-
zine reduced systemic blood pressure as early as 15 min
Ž
.
considerably less active by at least two log units than the
parent compound. Compounds 3 and 4 did not show any
Ž
.
after injection data from three separate rats are shown .
This dose was chosen on the basis of preliminary experi-
ments showing that it had a reproducible hypotensive
effect. Equimolar doses of chlorpromazine derivatives had
no such effect.
dopamine D1 or dopamine D2 receptor binding up to the
y4
Ž
.
highest concentration tested 10 M .
For a-1 noradrenaline receptors, all the compounds
showed some affinity, ranging from 20–1000 times lower
than chlorpromazine, with the following rank order: chlor-
promazine)compound 1)compound 2)compound 5)
compound 3)compound 4.
3.5. Effect of chlorpromazine and its deriÕatiÕes on phos-
pholipase A2 actiÕity
Chlorpromazine itself had no affinity for 5-HT1A recep-
tors whereas compound 3 showed affinity for this receptor
subtype, higher than on the other receptors considered. For
5-HT1B receptors compound 5, like chlorpromazine, had
micromolar affinity. None of the other compounds inhib-
ited radioligand binding to 5-HT1A or 5-HT1B receptors
even at a concentration as high as 10y4 M. Finally, except
The ability of chlorpromazine or its derivatives to in-
hibit phospholipase A2 was tested in a completely cell-free
assay with porcine phospholipase A2. Quinacrine was
tested as a reference phospholipase A2 inhibitor. As shown
for chlorpromazine, none of the compounds tested inhib-
Table 5
3
w
x
ited H paroxetine binding.
Inhibition of rat cerebellar NO synthase
Ž .
M
Compound
Ki
3.3. SedatiÕe action of chlorpromazine and its deriÕatiÕes
1
2
3
4
)4
3.8=10y6
2.4=10y5
)4
To confirm the lack of central action of the quaternary
derivatives, compounds 1 and 2 were tested in a sponta-
neous locomotor activity test. Unlike chlorpromazine,
which at the dose of 2 mgrkg markedly depressed loco-
5
4.2=10y6
3.8=10y6
1.1=10y7
Chlorpromazine
N-Nitro-L-arginine
Ž
.
motor activity y58%, P-0.01 , compound 2 at the
Ž
w Ž . x
N- 3- 2-chlorphenothiazin-10-yl propyl trymethyl-ammonium bro-
1
same dose had no such effect, a depressing effect being
.
Ž
w
Ž
.
.
x
mide ; 2 N- 3- 2-chlorphenothiazin-10-yl propyl butyl-dimethylam-
monium bromide ; 3 N- 3- 2-chloro-5-oxide phenothiazin-10-yl propyl
butyl-dimethylammonium bromide ; 4 N- 3- 2-chloro-5-oxide pheno-
thiazin-10-yl propyl trimethylammonium bromide ; 5 N-2- 4-chloro-
Ž
.
observed only at 5 mgrkg y43%, P-0.01 , and com-
Ž
w
Ž
.
x
pound 1 was completely devoid of a sedative effect at
.
Ž
w Ž
Ž
doses up to 10 mgrkg y9%, not significant, at this
.
x
.
Ž
Ž
X
.
.
x
.
dose .
phenyl thio phenyl N,N -dimethyl propane 1,3-diammonium chloride .

(
)
P. Ghezzi et al.rEuropean Journal of Pharmacology 317 1996 369–376
375
in Fig. 3, chlorpromazine and quinacrine had comparable
phospholipase A2 inhibitory activity. Some of the test
compounds had inhibitory activity comparable to or better
As a compound which degrades H2O2 would have also
been active in this model, we investigated the direct effect
of chlorpromazine and its derivatives on H2O2 degrada-
tion. No degradation was observed with any of the com-
Ž
than that of chlorpromazine the quaternary derivatives
.
Ž
compound 1 and compound 2 , while others were not
pounds at a concentration as high as 10 mM data not
Ž
.
.
inhibitory at all the sulfoxide compounds 3 and 4 even at
shown .
Ž
.
the highest concentration tested 1 mM .
3.6. Effect of chlorpromazine and its deriÕatiÕes on NO
synthase actiÕity
4. Discussion
This paper shows that some derivatives of chlorproma-
zine, including quaternary salts and sulfoxides, maintain
the parent compound’s ability to inhibit TNF production
when administered to mice before lipopolysaccharide. This
activity correlates with their ability to protect mice from
the effect of given a lethal dose of lipopolysaccharide.
Interestingly, TNF production was inhibited when chlor-
promazine or its derivatives were administered orally or
i.p., which could be particularly important from the phar-
macological point of view. Lower serum TNF levels were
demonstrated by a TNF bioassay and ELISA, indicating
that the effect of chlorpromazine was not merely at-
tributable to interference in the TNF bioassay. The bioas-
say-confirmed activities of chlorpromazine were evident.
Inhibition of TNF production seemed maintained even
in compounds that had lost the sedative activity of chlor-
promazine and that, being quaternary salts, were unlikely
to cross the blood–brain barrier. These data suggest that
chlorpromazine does not require its central activity to exert
its TNF inhibitory activity.
The ability of chlorpromazine or its derivatives to in-
hibit rat cerebellar NO synthase is shown in Table 5.
Chlorpromazine and compounds 2 and 5 had a Ki in the
Ž
micromolar range about two times less than the reference
.
inhibitor, N-nitro-L-arginine . Compound 3 was one log
unit less active than chlorpromazine and compounds 1 and
y4
Ž
.
4 were completely inactive Ki )10 M .
3.7. Antioxidant ability of chlorpromazine and its deriÕa-
tiÕes
The antioxidant activity was tested as the scavenging of
hydroxyl radicals generated by a mixture of ascorbic acid,
H2O2 , and ferric-EDTA. The sugar deoxyribose is de-
graded on exposure to hydroxyl radicals, setting off a
series of reactions that eventually result in the formation of
malondialdehyde, which is detected as a pink chromogen.
Fig. 4 shows the antioxidant capability of chlorpromazine
and its derivatives, measured as inhibition of deoxyribose
degradation. It can be seen that all drugs tested in dose
When the receptor activities of the chlorpromazine
derivatives are taken into consideration, it can be seen that
the various chemical modifications reduce the affinity for
all the receptors considered by several orders of magni-
tude. It therefore seems reasonable to conclude that the
antinoradrenergic, antidopaminergic and 5-HT receptor an-
tagonist activities of chlorpromazine are not essential for
its inhibitory effect on TNF production.
Ž
.
response from 125 to 1000 mM showed a similar scav-
enging activity, compared to that of the reference antioxi-
Ž
.
dant, DMSO 10% vrv , with the only exception being
compound 5, which displayed very weak antioxidant activ-
ity.
Most of the derivatives had lost the hypotensive action
of chlorpromazine, which might be an undesirable effect in
the prevention of endotoxic shock.
Finally, as regards the phospholipase A2 inhibitory
activity of chlorpromazine, compounds 3 and 4, which
were good TNF inhibitors, had no phospholipase A2 in-
Ž
hibitory activity unlike compounds 1 and 2 whose activity
.
was comparable to that of chlorpromazine . A similar
dissociation was observed between TNF inhibitory activity
and the effect on NO synthase activity.
From the data reported here it seems that chlorproma-
zine’s ability to bind to neurotransmitter receptors is not
involved in its inhibition of TNF, nor is inhibition of
phospholipase A2 or NO synthase. The question therefore
remains what pharmacological activities of chlorpromazine
are involved. Since reactive oxygen intermediates have
also been implicated in TNF induction, which is inhibited
Ž
.
Fig. 4. Antioxidant activity of chlorpromazine CPZ and its derivatives
in vitro. The antioxidant activity was assayed by measuring the release of
thiobarbituric acid reactivity A532 from deoxyribose. Data are ex-
pressed as the percent inhibition of control activity no drugs, 100% of
Ž
.
Ž
Ž
.
by antioxidants such as N-acetylcysteine Peristeris et al.,
degradation .

(
)
376
P. Ghezzi et al.rEuropean Journal of Pharmacology 317 1996 369–376
Ž
Costentin, J., M. Petit and S. Dollfus, 1987, Les Neuroleptiques Ellipses
.
1992; Zhang et al., 1994 , we evaluated the reported
.
Publishers p. 178.
Ž
antioxidant activity of chlorpromazine Slater, 1968; Jed-
ing et al., 1995 compared to that of its derivatives. Our
Davidson, F.F., E.A. Dennis, M. Powell and J.R. Glenney Jr., 1987,
Inhibition of phospholipase A2 by lipocortins and calpactins, J. Biol.
Chem. 262, 1698.
De Lean, A., P.J. Munson and D. Rodbard, 1978, Simultaneous analysis
of families of sigmoidal curves: application to bioassay, radioligand
assay and physiological dose-response curves, Am. J. Physiol. 235,
E97.
Gadina, M., R. Bertini, M. Mengozzi, M. Zandalasini, A. Mantovani and
P. Ghezzi, 1991, Protective effect of chlorpromazine on endotoxin
toxicity and TNF production in glucocorticoid-sensitive and glucocor-
ticoid-resistant models of endotoxic shock, J. Exp. Med. 173, 1305.
Halliwell, B., J.M.C. Gutteridge and O.I. Aruoma, 1987, The deoxyribose
method: a simple ‘test-tube’ assay for determination of rate constants
for reaction of hydroxyl radicals, Anal. Biochem. 165, 215.
Jeding, I., P.J. Evans, D. Akanmu, D. Dexter, J.D. Spencer, O.I. Aruoma,
P. Jenner and B. Halliwell, 1995, Characterization of the potential
antioxidant and pro-oxidant actions of some neuroleptic drugs,
Biochem. Pharmacol. 49, 359.
Mengozzi, M., G. Fantuzzi, R. Faggioni, A. Marchant, M. Goldman, S.
Orencole, B.D. Clark, M. Sironi, F. Benigni and P. Ghezzi, 1994,
Chlorpromazine specifically inhibits peripheral and brain TNF pro-
duction and up-regulates interleukin 10 production in mice, Immunol-
ogy 82, 207.
Mennini, T., A. Bizzi, S. Caccia, A. Codegoni, C. Fracasso, E. Frittoli, G.
Guiso, I.M. Padura, C. Taddei, A. Uslenghi and S. Garattini, 1991,
Comparative studies on the anorectic activity of d-fenfluramine in
mice, rats and guinea pigs, Arch. Pharmacol. 343, 483.
Moncada, S., R.M.J. Palmer and E.A. Higgs, 1991, Nitric oxide: physiol-
ogy, patophysiology and pharmacology, Pharmacol. Rev. 43, 109.
Peristeris, P., B.D. Clark, S. Gatti, R. Faggioni, A. Mantovani, M.
Mengozzi, S.F. Orencole, M. Sironi and P. Ghezzi, 1992, N-Acetyl-
cysteine and glutathione as inhibitors of tumor necrosis factor produc-
tion, Cell. Immunol. 140, 390.
.
study shows a good correlation between antioxidant activ-
ity and inhibition of TNF production. In fact compound 5,
which had no antioxidant activity, only slightly inhibited
TNF induction. The antioxidant activity seems therefore
the characteristic of chlorpromazine that plays an impor-
tant role in the inhibition of serum lipopolysaccharide-
induced TNF. The antioxidant activity of chlorpromazine
Ž
is at the level of that of hydroxyl radicals this paper and
..
Ž
Jeding et al. 1995 , but it does not scavenge superoxide
Ž
.
Ž
radicals Jeding et al., 1995 or hydrogen peroxide this
.
paper .
In conclusion, the present report indicates the utility of
structural modification of chlorpromazine to investigate
the mechanism of its inhibitory effect on TNF production.
Acknowledgements
R.B. was on leave of absence from and supported by
Dompe S.p.a., L’Aquila, Italy. Supported by a grant from
Ministero della Sanita-Istituto Superiore di Sanita, Rome,
Italy, within the 1994 AIDS Project. We thank Luca
Pariotti for excellent technical assistance.
´
`
`
References
Aebi, H., 1974, Catalase, in: Methods of Enzymatic Analysis, Vol. 2, ed.
Ž
.
H.U. Bergmeyer Verlag Chemie, Weinheim p. 673.
Rojas, A., J. Padron, L. Caveda, M. Palacios and S. Moncada, 1993, Role
`
Aggarwal, B.B., W.J. Khor, P.E. Hass, B. Moffat, S.A. Spencer, J.
Henzel, S. Bringman, G.E. Nedwin, D.V. Goeddel and R.N. Harkins,
1985, Human tumor necrosis factor. Production, purification and
characterization, J. Biol. Chem. 260, 2345.
Baldessarini, R.J., 1985, Drugs and the treatment of psychiatric disorders,
in: The Pharmacological Basis of Therapeutics. 7th edn., Vol. 7, eds.
A. Goodman Gilman, L.S. Goodman, T.W. Rall and F. Murad
of nitric oxide pathway in the protection against lethal endotoxemia
afforded by low doses of lipopolysaccharide, Biochem. Biophys. Res.
Commun. 191, 441.
Slater, T.F., 1968, The inhibitory effects in vitro of phenothiazines and
other drugs on lipid peroxidation systems in rat liver microsomes and
their relationships to liver necrosis produced by carbon tetrachloride,
Biochem. J. 106, 155.
Ž
.
Macmillan, New York, NY p. 387.
Spriggs, D.R., M.L. Shermann, K. Imamura, M. Mohri, C. Rodriguez, G.
Robbins and D.W. Kufe, 1990, Phospholipase A2 activation and
autoinduction of tumor necrosis factor gene expression by tumor
necrosis factor, Cancer Res. 50, 7101.
Testa, R., G. Abbiati, R. Cesarani, G. Restelli, A. Vanasia, D. Barone, M.
Gobbi and T. Mennini, 1989, Profile of in vitro binding affinities of
neuroleptics at different rat brain receptors: cluster analysis compari-
son with pharmacological and clinical profiles, Pharm. Res. 7, 571.
Vadas, P., E. Stefanski and W. Pruzanski, 1986, Potential therapeutic
efficacy of inhibitors of human phospholipase A2 in septic shock,
Agents Action 19, 194.
Basedovsky, H. and A. Del Rey, 1987, Neuroendocrine and metabolic
responses induced by interleukin-1, J. Neurosci. Res. 18, 172.
Bertini, R., M. Bianchi and P. Ghezzi, 1989, Protective effect of chloro-
promazine against the lethality of interleukin 1 in adrenalectomized or
actinomicin D-sensitized mice, Biochem. Biophys. Res. Commun.
165, 942.
Beutler, B. and A. Cerami, 1986, Chechectin and tumor necrosis factor as
two sides of the same biological coin, Nature 320, 584.
Boraschi, D., L. Villa, P. Ghiara, A. Tagliabue, M. Mengozzi, E. Solito,
L. Parente, J. Van Damme and P. Ghezzi, 1991, Mechanism of acute
toxicity of IL-1beta in mice, Eur. Cytokine Netw. 2, 61.
Bredt, D.S. and S.H. Snyder, 1989, Nitric oxide mediates glutamate-lin-
ked enhancement of cGMP levels in cerebellum, Proc. Natl. Acad.
Sci. USA 86, 9030.
Zhang, H., H. Spapen, D.N. Nguyen, M. Benlabed, W.A. Buurman and
J.L. Vincent, 1994, Protective effect of N-acetyl-L-cysteine in endo-
toxemia, Am. J. Physiol. 266, H1746.
Zimmerman, R.J., B.J.J. Marafino, A. Chan, P. Landre and J.L. Winkel-
hake, 1989, The role of oxidant injury in tumor cell sensitivity to
recombinant human tumor necrosis factor in vivo. Implications for
mechanisms of actions, J. Immunol. 142, 1405.
Chedid, L., 1954, Actions comparees de la promethazine, de la chlor-
´
´
´
promazine et de la cortisone chez la souris recevant des doses
mortelles d’une endotoxine bacterienne, C.R. Soc. Biol. 148, 1039.