You are on page 1of 191

XIDATIVE TIDE

FOR
SOME IMPLICATIONS
~~~ CATECHOLAMINE TOXICITY

HAEN

VRIJE UNNERSITEIT TE AMSTERDAM

Thiols in oxidative stress


some implications for catecholamine toxicity

ACADEMISCH PROEFSCHRIFT
ter verkrijging van de graad van doctor aan
de Vrije Universiteit te Amsterdam,
op gezag van de rector magnificus
dr. C. Datema,
hoogleraar aan de faculteit der letteren,
in het openbaar te verdedigen
ten overstaan van de promoriecommissie
van de faculteit der scheikunde
op donderdag 21 december 1989 te 15.30 uur
in het hoofdgebouw van de universiteit, De Bcelelaan 1105

Guido Rembertus Michiel Marie Haenen


geboren te Axel

Drukkerij Elinkwijk B.V. -Utrecht


1989

Promotoren: prof. dr. A. Bast


prof. dr. N.P.E. Vermeulen
prof. dr. H. Sies
Referent:

of
The investigations described in this thesis were carried out in the Department
Pharmacochemistry, Divisions of Molecular Pharmacology and Molecular Toxicology, Free
University, Amsterdam,The Netherlands.

OO G.R.M.M. Haenen, Amsterdam, 1990. Second edition


or
All rights reserved. No part of this book may be reproduced, stored in a retrieval system,
,
or
otherwise
,
recording
transmitted in any form or by any means, mechanical, photocopying,
without the prior written permission of the holder of the copyright.

Aan Myriam

Contents
Chapter 1

Introduction.

PartI

Protection against lipid peroxidation.

Chapter 2

Lipid peroxidation, modulation by endogenous and exogenous antioxidants.

in: M.A. Boogaerts (Ed.)Proceedings third Benelux workshop on free radicals in biology
and medicine,in press.

Chapter 3

Initiarion and propagation in lipid ~roxidation.

29

Chapter 4

Protection against lipid peroxidarion by a microsomal glutathionedependent labile factor.

35

FEBS Lett. 159 (1983) 24-28

Chapter 5

4-Hydroxy-2,3-trans-nonenal srimulates microsomal lipid pero~darion by


reducing the glutathione-dependent protection.

41

Arch. Biochem. Biophys. 259 (1987)449-456.

Chapter 6

Is phospholipase A2 involved in the glutathione-dependent protection


against in vitro microsomal lipid peroxidation?

49

in: O. Hayaishi,E. Nikt, M.Kondo and T. Yoshikawa (Eds.) Medical, biochemical


and chemical aspects of free radicals(1989)Elsevier, Amsterdam, pp. 1291-1294.

Chapter 7

Effect of thiols on lipid peroxidarion in rat liver microsomes.

53

Chem. Biol. Interact., in press.

Chapter 8

Interplay between dihydrolipoic acid and glutathione in the protecrion


against microsomal lipid peroxidation.

63

Bochim. Biophys. Acta 963(1988) 558-561.

Chapter 9

Mechanism of the reaction of ebselen with endogenous thiols.

69

Molec. Pharmacol., in press.

Part II

Catecholamines in oxidative stress.

Chapter 10 Modularion of oxidative stress by catecholamines.

91
93

Submitted.

Chapter 11 Acrivation of the microsomal glutathione S-transferase by metabolites


of a-methyldopa.

105

Submitted.

Chapter 12 Catecholamines do not reduce hydrogen peroxide producrion by


scavenging superoxide radicals.
Arch. Int. Pharmacodyn. Therapie, in press.

117

Part III

Oxidative stress and receptorfunction.

123

Chapter 13 Quantification of agonist-receptor interacrion.

125

Chapter 14 The effect of hydrogen peroxide on (3-adrenoceptor funcrion in the heart.


Free Rad. Res. Comms.4(1988)243-249.

145

Chapter 15 Reducrion of(3-adrenoceptor funcrion by oxidative stress in the heart.


Submitted.

151

Chapter 16 Contribution of 4-hydroxy-2,3-trans-nonenal to the reduction of


~3-adrenoceptor function in the heart by oxidarive stress.
Life Sci. 45(1989) 71-76.

167

Summary.

175

Samenvatting.

176

Curriculum vitae.

179

List of publicarions.

180

Nawoord.

182

- Chapter 1 Introduction.
Radicals are molecules with one or more unpaired electron. Due to the unpaired electrons)
these molecules can be very reactive. Radicals may damage a wide variety of organic
compounds occurring in living organisms (e.g. DNA, proteins, carbohydrates and lipids).
Radicals can be part of large macro-molecular structures and be rather immobile, but small,
freely diffusible radicals also exist. The latter species are referred to as free radicals.
Oxygen free radicals are the main type of radicals formed in aerobic cells under normal
conditions. Fortunately, the cells are equipped with an array of defense mechanisms that protect
against free radicals. However, several patho-physiological conditions may give rise to an
unbalance between the producrion of and the protection against oxygen free radicals, called
oxidative stress (Bies, 1986). It has been suggested that oxidarive stress is involved in the
etiology of various biological processes such as inflammation, ageing, carcinogenesis,
ischemia-reperfusion damage and photobiological effects (Bies, 1986).
All living organisms require sulfur. Most of the sulfur is present as cellular thiols (SH)or
disulfides(SS)(Jocelyn, 1972). Thiol and disulfide groups are found at critical positions in the
"active sites" of many enzymes (Jocelyn, 1972). In recent years it has been shown that there is
an interplay between thiols and oxidative stress. On the one hand, thiols are involved in the
defense against oxidative stress; glutathione (y-glu-cys-gly) plays a pivotal role in the protection
against oxidative stress. On the other hand, damage to thiol groups in several enzymes appears
to be one of the important mechanisms of oxidarive stress-induced damage (Bies, 1986).
In this thesis, the interplay between thiols and oxidarive stress is further explored. In part I
the effect of endogenous and exogenous administered thiols on oxidative stress is examined.
Part II deals with the modularion of oxidative stress by catecholamines: In part III, the effect of
oxidarive stress on ~i-adrenoceptor funcrion is described.
In the first studies described in this thesis (part I), the effect of thiols on oxidative stress was
determined. We focussed on the effect thiols have on lipid peroxidation. During the process of
lipid peroxidarion,poly-unsatured membrane lipids are converted into lipid hydroperoxides in a
free iadical reaction. As a result of this process, various membrane functions can be inacrivated.
The mechanism of the protection by glutathione against in vitro lipid peroxidation was studied.
Moreover, we deternuned the effect of other low molecular weight thiols on lipid peroxidation.
Also the mechanism of the peroxidase activity of ebselen - a syntheric compound that mimics the
activity of the endogenous glutathione-peroxidases - is examined.
During certain types of oxidative stress catecholamines are released. Generally, it is
supposed that these catecholamines contribute to the damage inflicted by oxidative stress. A
distinction can be made between a direct contribution of catecholamines and a contriburion that
is mediated by receptor hyperstimulation, e.g. the (3-adrenoceptor of the heart.
In part II of this thesis, the direct modularion of oxidative stress by catecholamines was
deternuned. The effect of catecholamines on the process of lipid peroxidation and on damage to
thiol groups by oxidarive stress was examined.
In part III, the effect of oxidative stress on (3-adrenoceptor function was determined. It is

lrnown that free thiol groups are essenrial for (3-adrenoceptor function (Vauquelin and Maguire,
1980), and during oxidative stress free thiol groups are attacked. Therefore, we quantified the
effect of oxidarive stress on the ~i-adrenoceptor system in the heart. The implications of the
effect of oxidative stress on ~3-adrenoceptor function for the toxicity of catecholamines are
discussed.
References.
P.C. Jocelyn (1972)Biochemistry of the SH group, Academic press, New York,London.
H. Sies (1986) Biochemistry of oxidative stress, Angew. Chem. Int. Ed. Engl. 25, 1058-1071.
G. Vauquelin and M.E. Maguire (1980) Inactivation of ~i-adrenergic receptors by N-ethylmaleimide in S 49
lymphoma cells, Mol. Pharmacol. 18, 362-369.

- PartI P~otection against lipid pegoxidation

- Chapter 2 Lipid peroxidation:


Modulation by endogenous and exogenous antioxidants.
Introduction.
Free radicals are molecules with one or more unpaired electrons. Due to these unpaired
electrons, these molecules might be very reactive. `When they are produced in biological
systems, they are capable of reacting with various cellular constituents. In the living cell, free
radicals are continuously produced. Fortunately, the cell is equipped with an elaborate defense
system that provides protection against the devastating effects of free radicals [1, 2]. However,
numerous (patho-)physiological processes, e.g. cancer, ageing, inflammation, emphysema or
reperfusion injury, are associated with an enhanced radical production [1]. It has been
suggested that these pathologies emerge from free radical-mediated damage, because radical
producrion may overwhelm the defense system in these cases.
As a result of the sometimes extremely high reactivity of free radicals, most types of
biomoleculen are prone to free radical attack. The intracellular location, where free radicals
outnumber the defense mechanisms, largely determines the damage inflicted and hence the
nature of the pathology provoked. For example, in radiation damage the enhanced radical
production in the vicinity of the genes is responsible for the DNA mutations. Another important
target of free radicals appears to be the biomembrane. This is partially due to the fact that
numerous enzymes and enzyme systems that are responsible for radical formation are located in
membranes. In the membrane, the radicals may induce peroxidation of membrane lipids. In this
process of lipid peroxidation, the free radical (R') abstracts a hydrogen atom from a polyunsaturated fatty acid:

X- - Y +R'-~ X

y +RH

H
The formed lipid radical is stabilized, because the free electron is delocalized over various
resonance structures:

~ _

_/
Y ~~ X

Y ~~ X

~
Y

The lipid radical reacts rapidly with oxygen:

X \ - Y -~ O2-~

O
O~

~X

Subsequently, the lipid peroxyl radical abstracts a hydrogen atom from a second polyunsaturatedfatty acid and a lipid hydroperoxide and a new lipid radical are formed:

O
O~
X

O~I~
O~
~%
~
\ -~ e m
Y X'
H
Y'
X
Y X'
e
Y'

This new lipid radical, on its turn,reacts with oxygen and generates another lipid hydroperoxide
and a third lipid radical. This propagarion or chain reacrion is interrupted when two unpaired
electrons meet and form a pair (terminarion) or when a relatively unreacrive radical is formed.
The lipid hydroperoxides formed during lipid peroxidation are thought to be highly reactive.
Homolyric or heterolytic cleavage of the peroxide, may produce radicals that start a new radical
chain reaction. This process is called lipid hydroperoxide-dependent lipid peroxidation. In lipid
hydroperoxide-independent lipid peroxidation, the chain reacrion of lipid peroxidation is
induced by radicals that do not originate from lipid hydroperoxides (Fig. 1).
As a result of the deteriortion by lipid peroxidation, all membrane functions can be
inacrivated. The lipid hydroperoxides formed may cluster and thus produce pores in the
membrane for compounds and ions(Ca2+) that are normally kept out [3]. The free radicals,
lipid hydroperoxides and other products formed during lipid peroxidarion may also react with
enzymes, and by this reaction annihilate the catalytic function of these proteins (see chapters 4
and 5[4, 5]). Also signal transmission over the membrane may become hampered, e.g. because
LH
R~
L
OZ

LOOH-independent
lipid peroxidation
LOS

LH

1
LO`
HOS
LOOP
LH~
L'~

LOOH-dependent
lipid peroxidation

ai

LO(i
LH

Figure 1. Sequence ofreactions in lipid peroxidation. LH is a pply-unsaturated fatty acid.

membrane fluidity is reduced by lipid peroxidation (see chapters 14-16 [6]). Moreover
diffusible, cytotoxic aldehydes are produced during lipid peroxidation, and these aldehydes
amplify cell damage induced by lipid peroxidation (see chapter 5 [5]).
Although lipid peroxidation is a normal phenomenon in aerobic cells, it is balanced by
various protective systems. In the lipid membrane, vitamin E is the most prominent endogenous
antioxidant. Vitamin C and reduced glutathione(GSH)are endogenous antioxidants located in
the cytosol. Moreover, several enzyme systems -like superoxide dismutase and catalase render activated oxygen species into less reactive ones. It is also possible to provide protection
against free radicals by exogenous antioxidants like flavonoids, several antibiotics and other
compounds.
In this review, the endogenous modulation of lipid peroxidation is limited to the mechanism
of the activities of vitamin E, vitamin C and GSH. Special attention is given to the
interrelationship between these endogenous antioxidants. Apart from this, it will be indicated
that numerous drugs are potent antioxidants, although this is not often realized. Actually, the
effecrivity of some drugs may partially be related to their anrioxidant capacity.
Vitamin E.
Vitamin E of natural origin consists of a group of compounds, namely a-,(i-, y-, ~-, ~1-,
~2- and ~-tocopherol. Of these compounds RRR a-tocopherol processes the highest vitamin
potency in vivo [7]. The a-tocopherol molecule can be divided into two parts, a chroman head
and a phytyl chain (Fig 2). It is generally believed that the phytyl chain intercalates with the fatty
acid residues of phospholipids, while the chroman head -responsible for the antioxidant effect faces the cytosol, although the chroman ring is still located in the hydrophobic zone of the lipid
bilayer [8]. In the antioxidant activity of vitamin E, a radical (R') abstracts a hydrogen atom
from the aromatic hydroxyl group of the chroman head (ArOH) rather then from a poly
unsaturated fatty acid, and a chromanoxyl radical is formed (Ar0').
R'+ ArOH~RH + Ar0'
The chromanoxyl radical is fairly stable, due to delocalization of the unpaired electron (Fig 3).
The oxygen in the heterocyclic ring of the chroman ring of tocopherols is fixed in such a
position that there is a considerable overlap between the 2p-type orbital of a lone electron pair of
the oxygen and the aromatic ~-system [9, 10]. This permits stabilization of the chromanoxyl
radical by interacrion of the unpaired electron with a lone pair of oxygen. In this way the degree
of delocalization of the free radical is enhanced., because of a substantial contribution by the
energetically favorable resonance structure V (fig 3). Compounds with less orbital overlap
between a lone pair of the oxygen and the aromatic ~-system, were found to be less potent
anrioxidants compared to tocopherols [10].
Despite the fact that the unpaired electron is stabilized on the chroman head, also the phytyl

chroman
head

phytyl chain

HO
~ I

Figure 2. Structure of a-tocopherol.

o .

o'
~ ~

~ I

C16H33

II

I
O

~
/

III

C16H33

C16H33

~
~

O
C16H33

~
~

IV

C16H33

Figure 3. Resonance structures of the a-tocopherol radical.

side chain of vitamin E affects its antioxidant activity. Decreasing the length of the phytyl chain
of a-tocopherol from C16 to C11, C6 or C1, decreased the concentration needed to inhibit lipid
peroxidation in vitro [8]. This effect has been ascribed to a decrease in mobility of the
chromans in the lipid bilayer with increasing length of the phytyl chain [8]. Also the better
orbital overlap between a lone pair of the oxygen and the aromatic ~t-system of the compound in
which the phytyl side chain has been replaced by a methyl group [10], is responsible for the
observation that in vitro the chroman with the shorter chain is the better antioxidant. However,
a-tocopherol -the compound with the longest chain (C16)-has the highest vitamin potency in
vivo [7, 9]. It has been suggested that the major, and probably only, role of the phytyl moiety
of tocopherols is to ensure that the chroman moiety is present at the places where biological
systems require protecrion against lipid peroxidation, i.e. in biomembranes. In fact, the phytyl
side chain has been shown to have the oprimal length -for penetration of chromans in a lipid layer
[l l]. Additionally, it has been reported that an increase in the length of the side chain decreases
the affinity of chromans for cytochrome P-450 [12]. This indicates that the compounds with the
smaller side chain are more rapidly metabolized by cytochrome P-450, which accounts for their
shorter half lives in vivo. Furthermore, it has been shown that y-tocopherol turns over very
rapidly, compared to a-tocopherol [13]. a-Tocopherol has the greatest vitamin E activity in
vivo of the various tocopherols, quite simply, because of superior absorption and retention [7].
In order to give an efficient protection, vitamin E has to be regenerated from the vitamin E
radical. It has been suggested that both vitamin C and GSH are able to mediate the regeneration
of vitamin E (vide infra).
Beside the antioxidant activity of vitamin E, it has been demonstrated that insertion of
vitamin E into lipid bilayers decreases membrane fluidity [14]. The decrease is similar to that
occurring after cholesterol enrichment. The effect of vitamin E is probably a result of an
interaction between vitamin E and phospholipids. It has been suggested that the free hydroxyl
group of the chroman head forms a hydrogen bound with one of the oxygen atoms of the polar
head of a phospholipid molecule, while the phytyl chain would be aligned parallel to the
phospholipid acyl chains [14]. This produces an increased molecular packing of the lipid
bilayer, resulting in the lower membrane fluidity observed after partition of vitamin E into a
membrane. Incorporation of either vitamin E or cholesterol into membranes decreased

membrane fluidity, but only vitamin E was able to protect against lipid peroxidation[14]. This
indicates that the effect on the physical state of the membrane does not contribute to the
antioxidant acrion of vitamin E.
Peroxidarion of membrane lipids is not the only factor which is responsible for cell death by
lipid peroxidation. There are several secondary mechanisms that amplify the initial focal injury
by lipid peroxidation and generate the end stage, widespread pathological alteration. The best
known amplification systems are the producrion of diffusible alkenals like 4-hydroxy-2,3trans-nonenal, the rise in intracellular calcium concentration and phospholipase A2 stimulation
[15]. By the action of phospholipase A2, phospholipids are split at the sn-2 position and
lysophospholipids and free fatty acids are produced. Overstimulation of phospholipase A2
induces an excessive deacylation of phospholipids, leading to an accumulation of
lysophospholipids followed by loss of membrane integrity with the final consequence of cellular
swelling and lysis [15].
There are several ways in which phospholipase A2 is activated by lipid peroxidation. For the
phospholipase AZ-dependent amplificarion of membrane damage, the elevation of the calcium
concentration in the cytosol produced by lipid peroxidation, resulting in a direct srimularion of
phospholipase A2,is of major unportance.
Although one might expect that vitamin E stimulates phospholipase A2 by reducing
membrane fluidity, it has been observed that vitamin E decreases phospholipase A2 acrivity
[16]. It has been reported that the hydroxyl group of the chroman head is crucial for the effect of
vitamin E on phospholipase A2 activity [16]. Inhibition of this secondary amplificarion
mechanism by vitamin E contributes to the prevention of irreversible damage, ven in the
presence of lipid peroxidarion.
Additionally,it has been shown that vitamin E restored membrane funcrion after damage has
been inflicted by phospholipase A2 [8]. Decreasing the length of the phytyl chain reduces the
ability of chromans to neutralize the damage caused by phospholipase A2 [8]. It is tempting to
suggest that this effect is the result of a "replacement" of the released fatty acid by vitamin E.
The chroman head forms a hydrogen bond with the free hydroxyl group of the glycerol moiety
of the lysophospholipid (at sn-2), while the phytyl chain aligns with the fatty acid at the sn-1
position. In this way vitamin E mimics a fatty acid, and the lysophospholipid-vitamin E
complex resembles a "normal" phospholipid.
As pointed out in this section, vitamin E is a potent radical scavenger, but it also
accomplishes other functions in the protecrion against lipid peroxidation.
Vitamin C.
The role of vitamin C (3-oxo-L-gulofuranolacton) in lipid peroxidation is known to be
ambivalent. Besides the various beneficial roles of vitamin C, also in preventing lipid
peroxidation by acting as antioxidant[17-20], vitamin C has a distinct pro-oxidant capacity [4,
21-25]. This is illustrated in figure 4. Lipid peroxidarion was measured with the thiobarbituric
acid assay, as described in chapter 4(reference 4), and expressed as the absorbance at 535 nm
versus 600 nm (D535-600) Vitamin C (0.2 mM)in the absence of iron did not induce lipid
peroxidation in rat liver microsomes, nor did 10 M Fe3+ in the absence of vitamin C. The
combinarion of 0.2 mM vitamin C and 10 M Fe3+, however,resulted in a rapid producrion of
thiobarbituric acid reactive material(Fig. 4). Also 10 M Fel+ alone induced lipid peroxidarion
(Fig. 4).
Addition.of vitamin C up to a concentration of 0.2 mM potentiated the 10 M Fel+-induced
lipid peroxidation, since the maximal amount of thiobarbituric acid reacrive material increased

1.5
5~
O

~ 1.0

4~

c~i

~ 0.5

2-~~ ~

10

20

30

40

50

time (rnin)
Figure 4. Time course of lipid peroxidation in eoritrol rat liver microsomes. Lipid peroxidation was
induced by 10 M Fe3+ (1), 10 M Fee+ (2), 0.2 mM vitamin C (3), 10 M Fe3+ and 0.2 mM vitamin
C (4) or 10 M Fel+ and 0.2 mM vitamin C (5).

(Fig. 5). Increasing the vitamin C concentrarion above 0.2 mM revealed the antioxidant capacity
of vitamin C, since a lag time in the rime course of lipid peroxidarion appeared. With high
concentrations of vitamin C, no lipid peroxidation was observed within the time span studied
(Fig. 5).
Vitamin C in concentrations up to 0.2 mM is able to induce lipid peroxidation in rat liver
microsomes. The pro-oxidant activity of vitamin C depends on the presence of metal ions like
copper or iron ions (Fig 4)[4, 23-26]. Lipid peroxidarion might be induced by the vitamin C
radical,(dehydroascorbate radical anion, vit C') produced during the iron catalyzed autoxidarion
of vitamin C, however, it s more likely that lipid peroxidation is the result of the reduction of
Fe3+ by vitamin C [24].
Fe3+ + Vit C ~ Fe2+ +Vit C'+ 2H+
The vitamin C radical is a relatively non-reactive species [27]. It mainly decays by
disproportionation, thereby terminating the propagation of free radical reacrions, resulting in the
production of vitamin C and dehydroascorbate(DHA)[27].
2 Vit C'+ 2 H+~Vit C + DHA
Alternatively, the vitamin C radical may also reduce another Fe3+ion [27].
Vit C'+ Fe3+
~ DHA + Fe2+
During the oxidation of vitamin C, hydrogen peroxide is also fornied [28, 29].
Vit C + 02~DHA + H2O2
These reactions provide all the ingredients for the Fenton reaction.
Fel+ + H2O2
~ Fe3+ + OH- + OH'
T'he role of the thus formed OH'-radical in iron-induced lipid peroxidation is still somewhat
obscure [30-33]. Due to its high reactivity, the place where the OH'radical is formed (i.e. near
the membrane or proteins or in the aqueous solvent) may determine the biological damage that is
provoked by this radical [25, 26], a phenomenon called site specific damage. The fact that Fel+
alone was also found to induce lipid peroxidation and that Tris, a well known OH' radical
scavenger, was used to buffer the incubation medium (iig 4), suggest that other radicals may be
10

2.0
H
ti:
c~ 1.0

a
~~
0

10

20

30

40

50

time (min)
Figure 5. Modulation of lipid peroxidation in control microsomes by vitamin C. The concentrations
vitamin C were for curves 1 - 10 respectively: 0, 0.025, 0.1, 0.2, 0.35, 0.6, 1, 1.5, 2, 4. All reactions
were started by the addition of 10 M Fel+.

involved. We use OH'to denote the free radical that can abstract a hydrogen atom from a poly
unsaturated fatty acid(LH)in the membrane, which will result in the peroxidarion of membrane
lipids.
LH+OH'~L'+H2O
L'+02~L00'
LOO'+ LH ~LOOH + L'
Concentrations of vitamin C above 0.2 mM were shown to provide protection against lipid
peroxidation. This can be brought about in various ways. Vitamin C might react with free
radicals that can initiate lipid peroxidation. It has well been documented that vitamin C reacts
with OZ'- [27, 29, 34] and OH'[14].
VitC+OH's Vit C'+H2O+H+
Vitamin C might also break the chain reaction of lipid peroxidation by reacting - with lipid
peroxyl radicals [17].
~ Vit C'+ LOOH + H+
Vit C +LOO'
However, due to the hydrophilic character of vitamin C,the reacrion with lipid peroxyl radicals
is not very efficient [18-20]. Vitamin C has been found to be a good antioxidant for
peroxidarions initiated in the aqueous phase, while it was less effective in trapping reactive
radicals in lipid membranes [20]. For the latter reaction a co-operation between the lipid soluble
vitamin E and vitamin C has been suggested [17-20]. Vitamin E (ArOH) has been shown to
protect against lipid peroxidation by scavenging in the lipid membrane radicals that might initiate
lipid peroxidation [35] or by chain-breaking the propagarion process of lipid peroxidation [1720].
ArOH + OH'~Ar0'+ H2O
ArOH + LOO'~Ar0'+ LOOH
The thus formed, relatively stable vitamin E radical might react with vitamin C, which results in
the regeneration of vitamin E [17-20].
Ar0'+ Vit C~ArOH +Vit C'+ H+
11

1.5
0
0
~ 1.0
m
0.5

11
0

10

20

30

40

50

time (min)
Figure 6. Modulation of lipid peroxidation in vitamin E-deficient microsomes by vitamin C. The
concentrations vitamin C were as in figure 5. All reactions were started by the addition of 10M Fel+.

In this way vitamin C would potentiate the protecrion by vitamin E against lipid peroxidation.
Several investigators have demonstrated unequivocally that there is a relation in the protection
against lipid peroxidarion between the antioxidants vitamin E and vitamin C [e.g. 17-20],
although it has been questioned whether this relationship is as direct as was shown in the last
equation [36]. Alternarively, it is possible that the anrioxidant effect of vitamin Cis, comparable
to its pro-oxidant activity, related to the reduction of iron. There is evidence that the
Fe3+ / Fe2+ ratio is important in lipid peroxidation [21-23, 32, 33]. A maximal rate of lipid
peroxidation is observed when the Fe3+/ Fe2+ratio is one. Both Fel+ and Fe3+ are required
for the catalysis of lipid peroxidation. Complete reduction to Fel+, or complete oxidation to
Fe3+ of all the iron prevents lipid peroxidation. Already in 1969 E.D. Wills [22] staters that if it
is assumed that vitamin C functions by reducing iron and that the occurrence of lipid
peroxidarion is deternlined by the ratio of Fel+ / Fe3+, then high concentrations of vitamin C
could produce aFel+/Fe3+ratio incapable to induce lipid peroxidation. This will result in the
inhibition of lipid peroxidation by high concentrations vitamin C. Recently this hypothesis has
gained further support [32].
To determine the contriburion of vitamin E to the antioxidant effect of vitamin C,the pro- and
anrioxidant acrivity of vitamin C in liver microsomes from vitamin E deficient rats was assessed.
Omission of vitamin E from the diet during 10 weeks,reduced the a-tocopherol levels in liver
microsomes from 1.51 nmol a-tocopherol per mg protein to 0.22 nmol a-tocopherol per mg
protein. Vitamin E depletion also reduced the level of poly unsaturated fatty acids (18:2, 18:3,
20:4 and 20:6). It has been shown that the dietary requirement of vitamin E has to be related to
the level of poly unsaturated fatty acid intake [7]. The overall effect of the deficient diet was that
the vitamin E content of the rat liver microsomes related to the amount of poly unsaturated fatty
acid dropped from 3.71 g a-tocopherol per g poly unsaturated fatty acid to 1.17 g a-tocopherol
per g poly unsaturated fatty acid. This indicates that these microsomes present a suitable model
to assess the effect of vitamin E.
As expected, vitamin E deplerion did not affect the pro-olcidant activity of vitamin C(Fig 6).
12

C1~1.

~
m 1.0

11

10

20

30

40

50

time (min)
Figure 7. Modulation of lipid peroxidation in heated microsomes by vitamin C. The concentrations
vitamin C were as in figure 5. All reactions were started by the addition of 10 .M Fel+.

The antioxidant activity of vitamin C was altered to some extent by vitamin E depletion.
Although high concentrarions of vitamin C were still able to protect against lipid peroxidation,
no lag time was induced by vitamin C in the concentrarion range of 0.2 to 1 mM,only the final
extent of lipid peroxidation declined (Fig 6).
To investigate whether an enzymatic component was involved in pro- or antioxidant actions
of vitamin C, heated microsomes were used. As shown in fig. 7, heating did not inhibit the prooxidant action of vitamin C. The effect of hearing of the microsomes on the antioxidant acrivity
of vitamin C was comparable to the effect of vitamin E depletion; the lag time disappeared but
vitamin C in high concentrations was still capable to decrease the extent of lipid pero~dation.
Combining the results obtained with different concentrations of vitamin C in control
microsomes, in vitamin E deficient microsomes and in heated microsomes, it is tempting to
suggest that the induction of a lag phase in lipid peroxidation by vitamin C in control
microsomes is due to an enzymatic interaction of vitamin E with vitamin C, since the lag time
was absent in heated liver microsomes and in liver microsomes for vitamin E deficient rats. The
effect of vitamin C on the final extent of lipid peroxidation was found to be non-enzymatic and
vitamin E-independent, since this effect is comparable in control microsomes, in heated
microsomes and in vitamin E deficient microsomes. Probably the latter effect is related to the
extent of the reduction of iron by vitamin C. High concentrations f vitamin C may reduce iron
completely, resulting in a Fel+/ Fe3+ratio incapable to induce lipid peroxidaton. Apparently,
both the extent of the reduction of iron as well as an interplay with vitamin E are involved in the
antioxidant acrivity of vitamin C in vitro.
In order to determine whether the effects observed were specific for vitamin C we used a
diastereoisomer of vitamin C,isoascorbate (3-oxo-D-gulofuranolacton). It was found that the
pro- and antioxidant activity of isoascorbate on lipid peroxidation were idenrical to those of
vitamin C. Moreover, the effects of vitamin C and isoascorbate were additive (data not shown).
A similar additive effect is found between vitamin C and cysteine (chapter 7[37]). Apparently,
in our in vitro system the effects of vitamin C on lipid peroxidation were not specific for
vitamin C. This indicates that there is no enzyme involved in the effects of vitamin C on lipid
13

peroxidation, in contrast to our previous suggesrion (vide supra). Moreover, on the one hand,
the effects of isoascorbate in vitro were equal to those of vitamin C. On the other hand,
isoascorbate has little to no vitamin potency in vivo, isoascorbate cannot provide the dietary
requirement for vitamin C [38, 39]. These data, taken together, question the physiological
importance of the protection against lipid peroxidarion afforded by vitamin C. In addirion,
cytosalic concentrations would favor glutathione over vitamin C as cytosolic antioxidant in most
tissues in vivo [36].
Glutathione.
Reduced glutathione (GSH)is a tripeptide (y-glu-cys-gly) with a free thiol group, that is
present in high concentrations in the liver of mammals. Interestingly,free radical production in
the liver can be high due to the abundant presence of enzymes giving rise to the production of
these reacrive molecules [2].Therefore, we suggest that due to the potent radical producing
capacity of the liver, the presence of a highly effecrive GSH-dependent defense system in this
organ can be explained teleologically.
The hydrophilic antioxidant GSH is not an efficient scavenger of free radicals, especially
when the radicals are formed in the lipid membrane [40]. In respect to the GSH-dependent
defense against lipid peroxidation, much attention has been given to the GSH-peroxidases. The
peroxidases catalyze the reducrion of hydroperoxides into their corresponding, less reacrive
alcohols at the expense of GSH,that is oxidized to GSSG.
GSH-P7 ROH + H2O + GSSG
ROOH + 2 GSH Under normal condirions, GSH is regenerated from GSSG by the GSH-reductase, that uses
NADPH as cofactor.
There are two classes of GSH-peroxidases, the selenium-dependent GSH-peroxidases and
the GSH-transferases, with different substrate selectivity (table 1). By the selenium-dependent
GSH-peroxidases, the peroxides are reduced in a cyclic ter-uni ping pong reaction, in which the
selenium moiety of the enzyme shuttles between the selenol form, the selenenic acid form and
the GSH-selenenyl sulfide form [41]. The selenium-dependent GSH-peroxidases can be
divided into the classical GSH-eeroxidase as fust described by Mills [42], and the phospholipid
hydroperoxide GSH-eeroxidase(PLHG-px) as first described by Ursini et al.[43](at that rime
denoted as PIP). Both selenium-dependent GSH-peroxidases use hydrogen pero~cide as well as
organic hydroperoxides as substrate. However, only the PLHG-px accepts phospholipid
hydroperoxides directly [43]. This is of special importance, because the poly unsaturated fatty
acids that are peroxidized during lipid peroxidation remain esteri~ed to glycerol. Most poly
unsaturated fatty acids are situated at the sn-2 position of phospholipids, and hence the lipid
hydropero~des that are generated by lipid peroxidation will also be located at the sn-2 position.
In order to detoxify phospholipid hydroperoxides by the classical selenium-dependent GSHperoxidase, the phospholipid hydroperoxide has to be deacylated by phospholpase A2, yielding
a lysophospholipid and a free lipid hydroperoxide [44]. The freed lipid hydroperoxide can be
converted into a lipid alcohol by the classical selenium-dependent GSH-eeroxidase. In this
sequence of reactions, phospholipase A2 does not amplify the damage produced by lipid
pero~dation - as described in the section on vitamin E - but it participates in the detoxication of
products formed by lipid peroxidation. The different actions of phospholipase A2 are due to the
different locations of the membrane where the enzyme is active. In amplifying cell injury,
phospholipase A2 deacylates phospholipids in parts of the membrane that are not peroxidized.
This can result in the accumulation of cytoto~cic concentrations of lysophospholipids at locations

14

selenium-dependent

hydrogen
peroxide
lipid hydroperoxide
phospholipid
hydroperoxide

selenium-independent

classical GSHpero~dase

PLHGpero~dase

cytosolic GSHtransferaces

microsomal GSHtransferace

+2

1 No duect substrate for the enzyme, however in combination with phospholipase A2 (PLA2), the phospholipid
hydroperoxide is accepted. By the action of PLA2, the phospholipid hydroperoxide is converted into a lipid
hydroperoxide.
2 Based on preluninary results of Mousialou and Morgenstem [53]

Table 1. Difference in substrate selectivity between the classical GSH-eeroxidase, the phospholipid hydroperoxide
GSH-eeroxidase (PLHG-px), the cytosolic GSH-transferaces and the microsomal GSH-transferase. When a
pero~de is accepted as substrate, this is depicted as +.

where no lipid peroxidation has occurred. In the detoxication of phospholipid hydroperoxides,


phospholipase A2 acts in the restricted areas of the membrane that have been peroxidized,
although also at these places an excessive producrion of lysophospholipids amplifies membrane
damage. Moreover, no vitamin E is anymore available there for the neutralization of the
cytotoxic lysophospholipids, since vitamin E has been consumed during lipid peroxidation.
There are several mechanisms by which phospholipase A2 is acrivated by lipid peroxidation
[15]. Firstly, lipid peroxidarion makes the membrane more rigid and more hydrophobic, which
facilitates binding and penetrarion of phospholipase A2. Before exerting its catalyric acrivity,
cytosolic phospholipase A2 has to bind to the membrane. Secondly, calcium pores are generated
in the membrane and calcium-ATP-ases aze inactivated, leading to an increase of the calcium
concentration in the cytosol, which directly activates phospholipase A2. Thirdly, vitamin E is
consumed during lipid peroxidation, and vitamin E is an inhibitor of phospholipase A2.
Addirionally, it has been stated that phospholipase A2 preferenrially removes oxidized fatty
acids from membrane phospholipids [45, 46], although no convincing evidence for this
statement has been put forth.
By lipid peroxidation the phospholipids especially in parts of the membrane that have been
peroxidized become more accessible to phospholipase A2. Only in the area that has been
peroxidized (i) the membrane fluidity drops, which increases the affinity of cytosolic
phospholipase A2 for this area and improves substrate availability for phospholipase A2, and
(ii) the endogenous inhibitor of phospholipase A2, vitamin E, is consumed. Elevarion of the
intracellular concentration of calcium makes the phospholipids in both the peroxidized and nonperoxidizedparts of the membrane more prone to deacylation by phospholipase AZ. This, to
some extent, selective activation of phospholipase A2 at the peroxidized spots of the membrane
favours the phospholipase A2-dependent repair mechanism over the phospholipase A2-mediated
amplificarion of toxicity. However, despite the activarion of phospholipase A2 by lipid
peroxidation, its activity remains much to low to provide an efficient protection against lipid
peroxidation in vitro via the detoxication of phospholipid hydroperoxides [47], especially when
it is taken into consideration that the authors who proposed this mechanism [42, 44,45] suggest
that phospholipid hydroperoxides are extremely reactive.
15

,OH
~N+
o
0 0~0
P,
O ~O

OH

PLHGpx _ ~"~
m-GSH-tr

o, ,o~o
P
" 'O

II

PLA2 I
0
-o

,OH
O

III + V

~ N+

o ,o~o
~P~
" O

PLA2
OH
GSH-px
PLHG-px

_o

- GSH-tr

m-GSH-tr

o,P,o~o
" s0

IV + V

`~
O

Figure 8. Reduction of phospholipid hydroperoxides by the GSH-peroxidases and phospholipase A2


(PLA2). During lipid peroxida[on, the poly-unsaturated fatty acids remain esterified to glycerol and
phospholipd hydroperoacides(i) are formed. The phospholipid hydroperoxide GSH-eeroxidase(PLHGpx)
and the microsomal GSH-transferace (m-GSH-tr) are able to reduce directly I and lipid peroxides (III).
For the reduction of I by the classical GSH-eeroxidase (GSH-px) or cytosolic GSH-transferaces(GSHtr), firstly I has to be hydrolyzed to III and a lysophospholid (V)in a PLA2 catalyzed reaction.
Subsequently, III is reduced, either by GSH-px,PLHG-px, GSH-tr or m-GSH-tr, to a lipid hydroxide
(IV). Finally, V can be reacylated by an acyltransferase to regenerate a phospholipid, and IV can be
consumed in the fatty acid oxidation. The phospholipid hydroxide(In,formed by the direct reduction ofI
by PLHG-pz or GSH-tr, has to be converted by PLA2 in IV and V, in order to restore the original
composition of the membrane.

The GSH-transferaces are a class of enzymes that catalyze the reaction of GSH with
electrophiles. Conjugation to GSH is a well known pathway in the biotransformation of both
endogenous compounds and xenobiotics. The cytotoxic aldehydes that are produced during
lipid peroxidation, and amplify cell damage induced by lipid peroxidation (i.e. alkenals like
4-hydroxy-2,3-trans-nonenal) are detoxified by a GSH-transferace catalyzed reaction with
glutathione [48].
The GSH-transferaces also catalyse the reaction between hydroperoxides and GSH. The
eeroxidase activity consists of two reactions, the first one is catalyzed by the GSH-transferaces,
the second one is anon-enzymatic reaction [49].
LOOH + GSH

GSH-tri

LOH + GS OH

GSOH + GSH~ GSSG + H2O


Most GSH-transferaces are, comparable to the selenium-dependent GSH-peroxidases, located
in the cytosol, but amembrane-bound GSH-transferase has also been described [50]. There is a
considerable variarion in eeroxidase activity of the different GSH-transferace isozymes [51].
Hydrogen peroxide is not reduced by the GSH-transferaces, but organic hydroperoxides are
16

1.5

3
i~

1.0

c~

0.5
~~ ~,
0

10

20

30

40

50

time(min)
Figure 9. Effect of glutathione or mercaptoethanol on lipid~eroxidaon in control microsomes. Lipid
peroxidation was induced with the combination of 10 M Fe +and 0.2 mM vitamin C.Further additions
were: none (1), 1 mM glutathione(2)or 1 mM mercaptoethanol (3).

accepted as substrate. Like the classical selenium-dependent GSH-peroxidase, phospholipase


A2 is needed to detoxify phospholipid hydroperoxides by the cytosolic GSH-transferases [52].
It has been reported that the membrane-bound GSH-transferase reduces phospholipid
hydroperoxides without the participation of phospholipase A2[53].
Despite the fact that the direct peroxidase activity of PLHG-px and the membrane-bound
GSH-transferase is independent of phospholipase A2, the latter enzyme is needed in the final
repair of the membrane (Fig. 8). In order to restore the original composirion of the membrane,
the frmed lipid hydroxide has to be removed from the phospholipid hydroxide in a
phospholipase A2-catalyzed reaction. The free lipid hydroxide can be metabolized by the fatty
acid oxidation process (lipid hydroxides are intermediates in fatty acid oxidarion), while the
lysophospholipid can be reacylated by an acyltransferase to regenerate the phospholipid.
Much emphasis has been placed on the idea that the detoxication of phospholipid
hydroperoxides by the GSH-peroxidases, either alone or in combination with phospholipase
A2,is the major - if not only - GSH-dependent protecrion mechanism against lipid peroxidation
[42, 45, 46, 52]. However, by this mechanism only protection is provided against the
decomposition of lipid hydroperoxides into free radicals(LOOH-dependent lipid peroxidation,
fig. 1). No protection is afforded against LOOH-independent lipid peroxidation. Probably, in
vitro induced lipid peroxidation consists for the greater part of LOOH-independent peroxidation
[5]. Nevertheless, it has well been characterized that liver microsomes of the rat possess an
efficient GSH-dependent protection mechanism against lipid peroxidation in vitro [4, 53-57].
In the incubarion system containing 10 M Fel+ and 0.2 mM vitamin C, addition of 1 mM
GSH delayed but did not prevent the occurrence of lipid peroxidation (Fig. 9). After a lag rime
lipid peroxidation starts, the final amount of thiobarbituric reactive material produced, equals the
amount in the incubarion system without GSH. The GSH-dependent protection proceeds via a
heat-labile factor, that is probably an enzyme [4, 53]. This enzyme is not identical to the
micrsomal GSH-transferace, nor is phospholipase A2 involved in the GSH-dependent
protection (chapter 5[5,47]). The fact that mercaptoethanol(Fig. 9)and dithiothreitol (chapter
17

7 [37]), are not able to replace GSH in the protecrion against microsomal lipid peroxidation,
while mercaptoethanol and dithiothreitol are equally effective or more effective than GSH in the
peroxidase acrivity of the PLHG-px [43] is one of the indications -that PLHG-px is not
responsible for the GSH-dependent protection. It has also been demonstrated that the GSHdependent protecrion is not due to a reducrion of iron (chapter 4 [4]).
Reddy et al. [55] observed that the GSH-dependent protection was absent in vitamin E
depleted mcrosomes. This made it tempring to suggest that the GSH-dependent protection was
due to a GSH-mediated regeneration of vitamin E(Vit E)from the vitamin E radicals (Vit E'), a
reaction catalyzed by the heat-labile factor that functions as a free radical reductase (chapter 4
[4])

~~

~lt ~

2GSSG
free radical
reductase

RH

Vit E'
MEMBRANE

GSH
~ CYTOSOL

We have shown that this free radical reductase contains an essenrial and vulnerable thiol moiety
itself (chapter 5 [5, 57]), and that the free radical reductase is selective for GSH (chapter 7
[37]). Moreover, the GSH-dependent protection in liver microsomes is susceptible to oxidative
stress. After a relatively moderate level of lipid peroxidation, addition of GSH does not provide
any protection (chapter 4 [4]). This might be due to the consumption o vitamin E during the
initial phase of lipid peroxidation, or it might be caused by an inactivarion of the free radical
reductase by cytotoxic aldehydes, like 4-hydroxy-2,3-trans-nonenal, generated in the initial
phase (chapter 5 [5]).
Although it has been amply demonstrated that there is anefficient GSH-dependent protection
mechanism in rat liver microsomes, the exact mechanism of the protection is still unknown. We
know that there is an interplay between GSH and vitamin E in this protection and that there is
probably an enzyme involved in this interplay, but the exact nature of this enzyme is still
obscure. No unambiguous evidence has been produced to prove the participation of a free
radical reductase thus far. Nevertheless, all other explanarions for the microsomal glutathionedependent protecrion against lipid peroxidation can be rejected.
The interplay between GSH and vitamin C is complex. On the one hand, GSH is capable to
react with dehydroascorbate(DHA), yielding vitamin C and GSSG [58].
2 GSH + DHA ~ GSSG +Vit C
On the other hand the thiyl radical of GSH reacts with vitamin C yielding the semidehydroscorbate radical (Vit C')and GSH [59].
GS'+ Vit C~ GSH +Vit C'
Apparently the unpaired electron is more stabilized in the vitamin C radical than the thiyl radical
of GSH. This is probably due to the delocalization of the unpaired electron in the vitamin C
radical [27], a phenomenon not possible in the thiyl radical of GSH.
The physiological relevance of the interplay between GSH and vitamin C is not known yet,
although it has been reported that it is of minor importance in vivo[60]. In vitro, the reduction
of dehydroascorbate by GSH is responsible for the indirect pro-oxidant activity of GSH [4,
58]. GSH has no direct pro-oxidant acrivity (chapter 4 [4]), since it is not able to reduce free
18

iron at a measurable rate. When lipid peroxidation is generated in vitro by the combination of
iron and vitamin C, vitamin C becomes oxidized and dehydroascorbate is generated. GSH
reacts with dehydroascorbate and regenerates the consumed vitamin C. Via this reacrion, GSH
indirectly provides the reducing equivalents needed for the reduction of iron. This reacrion
cannot explain the antioxidant acriviry of GSH,since (i) only a low concentration of vitamin C
was used in these experiments (Fig. 8),(ii) other thiols that are equally effective as GSH in the
reducrion of dehydroascorbate, do not possess any anrioxidant acrivity [58] and (iii) GSH also
protects against lipid peroxidarion in incubation systems where no vitamin C is present[54,61].
Exogenous modulators oflipid peroxidation.
The noxious consequences of lipid peroxidation can be restrained. Interference with the lipid
peroxidation process can occur in several ways. Radical scavengers will interrupt the
propagation phase of lipid peroxidation (Fig. 1). Compounds might also inhibit radical
producing enzymes or mimic protective enzymes. Moreover protection can be afforded by
inhibiting the mechanisms that amplify cell injury by lipid peroxidation. Furthermore an
interaction with endogenous antioxidants like glutathione, vitamin E or vitamin C might be
involved in the effect of drugs on lipid peroxidation. There is already a wide array of antioxidants at the physician's disposal. It is however probably not common knowledge yet that a
large number of registered drugs have been characterized as potent anti-oxidants (Table 2). The
inhibition of lipid peroxidation may in some cases contribute to the established effectiveness of
these compounds.
Su~ydryl containing compounds are found among various pharmacological classes (Table
2)[57]. Lipoic acid and N-acetylcysteine have received most attention with regard to their

sulphydryl-containing compounds
liver-cirrhosis, polyneuropathy
lipoic acid
mucolytic toxicity,
mesna
oxazaphosphorine cytostatics
toxicity
angiotensie converting
inhibitor
captopril
enzyme
mucolytic
N-acetylcysteine
heavy metal-poisoning,
penicillamine
Wilson's disease

ref.
62
58

flavonoids
anti-inflammatory drugs
HZ-antagonists
(.~adreneceptor antagonists
calcium antagonists
anti-arrhythmic drugs
miscellaneous:
allopurinol
disodium cromoglycate
desferroxamine

74, 75
79
80
81
82
83

63
57
57

93
95
96

Table 2.Registered drugs with anti-oxidant activities


19

anti-oxidant potenrial. Interestingly enough, the mode of acrion of both compounds can be
explained via the replenishment of reduced glutathione. We established (chapter 8)[62] that
dihydrolipoic acid keeps glutathione in vitro in the reduced form. This explains the inhibition of
hepatic microsomal lipid pero~cidation by dihydrolipoic acid. Lipoic acid has also been reported
to be a singlet oxygen scavenger [64]. N-acetyl cysteine protects against free radical mediated
cellular damage due to its ability to act as a precursor for the cysteine portion of the tripeptide
glutathione [65] and appears to be a powerful scavenger of hypochlorous acid as well [66]. Not
all sulphur containing drugs have been tested for their anrioxidant activity, e.g. disulfuram and
6-mercaptopurine. The aldehyde dehydrogenase inhibitor disulfuram is metabolized to the very
potent anri-oxidant diethyldithiocarbamic acid [67], which, however, is also an inhibitor of
superoxide dismutase [68]. Moreover, aldehyde dehydrogenase is involved in the detoxication
of 4-hydroxy-2,3-trans-nonenal [69]. The drug 6-mercaptopurine is used as cytostaric antimetabolite [70]. A new syntheric sulfhydryl compound is 2-mercaptopropionylglycine, which
has been shown to be beneficial in alleviaring ischemia/reperfusion-induced cell-injury [71].
Prodrugs of 1-cysteine have been syntherized [72,73] with the aim to increase intracellular
glutathione levels.
Flavonoids are reported to decrease capillary permeability and fragility. Their precise
molecular pharmacological mode of acrion is not clear yet. Their prominent anti-oxidant
activities have been described [74]. Also their metal chelating properties might be ofimportance
in preventing free radical damage [75]. The anti-oxidant activity of other polyphenolic natural
pigments like polyhydroxynaphtoquinones [76], gossypol [77] or ellagic acid [78] is known,
but received remarkably little study.
A large series of anti-inflammatory drugs have been recently tested with regard to their ironbinding and hydroxyl radical scavenging actions, and were found to protect against site-specific
damage by the hydroxyl radical [79a]. It is thought that oxidarive free radical stress is involved
in inflammarions. The withdrawal of metal ions in combination with hydroxyl radical
scavenging activity undoubtedly contribute to the pharmacotherapeutic effectiveness of this class
of drugs. Recently, Van de Straat et al.[79b] tested paracetamol, and several3-monoalkyl and
3,5-dialkyl derivatives of paracetamol for their antioxidant capacity. Especially the dialkyl
derivatives were very potent inhibitors of lipid peroxidarion. It was found that the anrioxidant
acrivity of paracetamol and its derivatives highly correlated to their lipophilicity [79b].
We recently elaborated the anri-oxidant activity of histamine HZ-antagonists [80] and
suggested that this combination of action may be useful in the treatment of peptic ulcers. It is
interesting that of the compounds in use to ameliorate the damage after heart ischemia like
l3-adrenoceptor blocking drugs, calcium antagonists and anti-arrhythmic agents many have an
anti-oxidant acrion [81,82,83]. Since oxygen free radicals are involved during the course of
ischemic damage [84], the anti-oxidant acrivity of the drugs may contribute to their overall
phazmacological profile. In addition it has been reported that anti-arrhythmics inhibit
phospholipase A2[85], thus preventing an amplification mechanism of lipid peroxidation. It
might even be speculated that increasing the anri-oxidant behaviour with regards to specificity
and efficacy may lead to the development of better drugs for the pathologies involved.
Reperfusion injury may involve hydrogen peroxide or superoxide anion radical formation
[86,87]. In view of the use of(3-adrenoceptor antagonists in the treatment of angina and cardiac
arrhythmia the data published on the anri-oxidant [81] and the xanthine oxidase inhibitory
activity [88] by J3-Mockers might therefore be quite interesting.
As mentioned above, Cat+-ions play a central role in cell-injury in several pathophysiological
conditions. Cat+-entry-blocking activity may delay or prevent cytotoxicity induced by free
radicals [89]. It is suggested that it is not the inhibition of the physiological slow calcium inward
20

l0a
A

lOb
~ ~
~JI
v

Figure 10. Scavenging activity of the metabolite of c~clandelate, mandelic acid, compared to that of
mannitol. Hydroxyl radicals are generated by a ADP-Fe +-H2O2 mixture,and detected with 5,5-dimethyl1-pynoline-N-oxide (DMPO) using ESR. The scavengers were added to the incubation medium which
contains aADP-Fel+-H2O2 mixture with 34.5 mM DMPO,according to [91]. The height of the ESRsignal indicates the extent of DMPO-hydroxyl radical formation. The lower the height of the signal, the
higher the scavenging activity of the compound. Figure 10a: A: control system; B-F: the same as in A
but in the presence of 5,10, 15,20 and 25 mM mandelic acid. Figure lOb: As in figure l0a but instead of
mandelic acid now B: in the presence of25mM mannitol.

fluxes, but rather the inhibition of excessive calcium overflow occurring in pathological
situarions that should be the prevailing characteristic in the Cat+-antagonists in these condirions.
These so-called Cat+-overload blockers might include cyclandelate [89], flunarizine and
sabeluzole [90].
The anri-oxidant properties of these compounds are probably also important since their
intended therapeutical use is in situations in which free radicals are implicated. Cyclandelate
exerts its clinical activity in cerebral ischemia or hypoxia, at least partly through a calcium
modulatory effect [89]. Moreover, a major metabolite of cyclandelate, mandelic acid is a very
effective hydroxyl radical scavenger as indicated in fig. 10, in which it is compared with
mannitol. Rate constants for the hydroxyl radical scavenging reaction were (8.2-9.7)x109 and
(1.4-3.7)x109 M-ls-1 for mandelic acid and mannitol, respectively.
Several antibiotics [92] have been tested for their radical scavenging effect. Some could
inhibit the inactivation of al-anti-proteinase by hypochlorous acid. Thus the endogenous
inhibitor of elastase is preserved. In this way the ultimate occurrence of emphysema in lung
inflammation might be reduced.
Xanthine oxidase can be a major source for oxygen radical formation [87]. Allopurinol, an
inhibitor of this enzyme, also displays prominent hydroxyl radical scavenging activity [93].
21

Moreover, the compound may interfere with electron transfer [94]. A similar interference has
been observed for cromoglycate [95]. This compound inhibits mast cell degranulation and has
been found to compete with molecular oxygen for hydrated electrons, thus preventing the
formation of superoxide anion radicals n a radiolytic experimental set-up.
Metal complexation may preclude iniriarion of lipid peroxidarion. Efficient iron complexation
may be reached with e.g. desferroxamine [96]. Reacrivity of the Fe2+-chelate with 02 can be
prevented by using a hexadentate chelate. In this way no remaining co-ordinarion loci for
hydroxyl radical formarion are left [95]. However,it should be noted that in the interacrion of
desferroxamine with radicals anenzyme-damaging nitroxide can be formed [97].
As demonstrated, registered drugs sometimes unexpectedly, possess anti-oxidant
characterisrics. In fig. 11, a survey is provided, which cannot be complete however. Moreover
new compounds are being developed which are designed especially for this purpose. N-acetyldehydroalanines and particularly AD 20 (Fig. 11) are capable of stabilizing free radicals and
have been suggested to reduce the free radical mediated damage as adriamycin toxicity [98].
A novel series of compounds (with U 74006 F as a representative,fig. 11) was launched as
potent inhibitor of iron-dependent lipid peroxidation [99]. Most reports with these compounds
are directed towards use of these potential drugs in the treatment of head- and spinal-cordinjuries and strokes.
Ebselen is a newly designed anti-inflammatory agent with an unconventional mode of
action. This selenium containing compound has peroxidase activity with the endogenous thiol
glutathione as co-factor [100]. Interestingly, ebselen accepts lipid hydroperoxides as well as
phospholipid hydroperoxides as substrates. We found that other thiols then GSH can function
as co-factor as well and proposed a modified scheme of action of the GSH-peroxidase acrivity
of ebselen (chapter 9). Ebselen also inhibits the oxidative burst of alveolar macrophages [101],
which might be due to inhibition of protein kinase C [101,102] and contributes to its
pharmacodynamic activity. Not only glutathione peroxidase, but also superoxide dismutase
enzymatic activity has been mimicked, as for example with tetrakis--3,5diisopropylsalicylatodiaquodicopper (II)(abbreviated as Cu(II)2(3,5-DIPS)4)(Fig. 11), which
is lrnown to disproportionate superoxide anion radicals at the same rate as superoxide dismutase
[104]. The compound protects against the effects of irradiation. However, its overall
pharmacological action is probably not limited to its superoxide dismutase acrivity [104].
It is also possible to potenriate the protection against free radicals, by administration of the
endogenous enzymes superoxide dismutase and catalase. Chemically modified enzymes
superoxide dismutase or catalase are being developed as well, in order to extend their half lives
and to reduce their anrigenic acrivity compared to the natural form [105,106]. Recombinant
DNA techniques facilitate the production of these enzymes and may stimulate this approach in
the future.
Butylated hydroxytoluene (2,6-di-tert-butyl-l-hydroxytoluene) derivatives are being
developed as lipoxygenase inhibitors and radical scavengers [107, 108](Fig. 11). Recently, in
"an attempt to outdo nature", an apparent successful vitamin E-analogue (Fig. 11) has been
synthetized [109]. The characteristics that explain the chain-breaking anti-oxidant activity of
vitamui E (as mentioned above) have been optimized in this analogue. Also vitamin C-analogues
have been synthetized. A series of 2-O-alkylascorbic acid combine hydrophilic and lipophilic
properties in one molecule. 2-O-octadecylascorbic acid (Fig. 11) seems a promising scavenger
[110].

22

(~shc
0

(CH3}~C

NJ
N

~~
/N~/ \,N
~
I
CHZ

ethyl 3,5-di-tert-butyl-4-hydroxycinnamate

=C~
CH2
NH

~ 2~

"'
~;

~3 I

CH3SOZOH

~ ~

i ~2
I

vN

~h c=o

~3 ~

~OOOH

U-74006F

cic,
~

HO

~`~
~~~Yi~~3~3~3
~3

Q
~3

vitamin E analogue

~3

~~'.

\ I

c
_n,

~N

ebselen (PZ51)
~
~

HO

o(cx,~cx,
2-O-octadecylascorbic acid
tto

~
\~/
HOH

Cu(II)2(3,5-DIPS)4

Figure 11. Molecular structures of newly synthetised anti-oacidants.


In this review we outlined that the interaction between endogenous modulators of lipid
peroxidation like GSH, vitamin C and vitamin E is crucial in the prevention of free radical
induced pathologies. Until now, in the design of new protective drugs hardly any attention is
given to their interplay with endogenous anrioxidants. We expect that a growing understanding
of the lipid peroxidation process in pathophysiology, and in particular of the interplay between
endogenous protecrive components and drugs will add further impetus to the development of
new, specific and effective anti-oxidants.
References.
1
2
3
4

H. Sies,(Ed.), Oxidative Stress, Academic Press, New York -London, 1985.


A. Bast and G.R.M.M. Naenen, Cytochrome P-450 and glutathione, what is the significance of their
interrelationship in lipid peroxidation, Trends Biochem. Sci. 9(1984)510-513.
M. Crompton, A, Costi and L. Hayat., Evidence for the presence of a reversible, Cat+-dependent pore
activated by oxidative stress in heart mitochondria, Biochem. J. 245(1987)915-918.
G.R.M.M. Naenen and A Bast., Protection against lipid peroxidadon by a microsomal glutathione23

6
7
8

9
10

11
12

13

dependent heat labile factor, FEBS Lett 159(1983)24-28 (chapter 4).


G.R.M.M. Haenen, J.N.L. Tai Tin Tsoi, N.P.E. Vermeulen, H. Timmerman and A Bast., 4-Hydroxy-2,3trans-nonenal stimulates microsomal lipid peroxidation by reducing the glutathione dependent protection,
Arch. Biochem. Biophys. 259(1987)44956(chapter 5).
G.R.M.M. Haenen,P. van Dansfik, N.P.E. Vermeulen, H. Timmerman and A. Bast, The effect of hydrogen
permcide on (3-adrenoceptor function in the heart, Free Rad.Res. Comms.4(1988)243-249(chapter 14).
L.A. Witting, Vitamin E and lipid antioxidants in free-radical-initiated reactions, in W.A. Pryor (Ed.) Free
radicals in biology, Academic Press, New York, Vol. IV (1980) pp. 295-319.
A.N. Erin, N.G. Davitashvili, V.E. Kagan, E.I. Zakharova, I.K. Sarycheva and R.P. Evstigneeva, Role of
the isoprenoid chain of a-tocopherol in the protection of synaptosomes from lipid peroxidation and the
acon of phospholipase A2, Biokhimiya (English translation) 53(1988) 591-597.
G.W. Burton and K.U. Ingold, Autoxidation of biological molecules 1. The antioxidant activity of vitamin
E and related chain-breaking phenolic antioxidants in vitro, J. Am. Chem. Soc. 103(1981)fx}72-6477.
G.W. Burton, T. Doba, E.J. Gabe, L. Hughes, F.L. Lee, L. Prasad and K.U. Ingold, Autoxidation of
biological molecules 4. Maximizing the antioxidant activity of phenols, J. Am. Chem. Soc. 107 (1985)
7053-7065.
B. Maggio, A.T. Diplock and J.A. Lucy, Interactions of tocopherols and ubiquinones with monolayers of
phospholipids, Biochem. J. 161 (1977) 111-121.
E.A. Serbinova, V.E. Kagan, G.M. Koynova and T.S. Stoytchev, Interaction of natural inhibitors of free
radical lipid oxidation with cytochrome P-450, International conference on medical, biochemical and
chemical aspects of free radicals, Elsevier, Amsterdam (1988)IX-1-02.
I.R. Peake, H.G. Windmueller and J.G. Beiri, A comparison of the intestinal absorption, lymph and plasma
transport, and tissue uptake of a- and'y-tocopherols in the rat, Biochim. Biophys. Acta 260 (1972) 679-

688.
14 J.M. Patel and D.A. Edwards, Vitamin E, membrane order and antioxidant behavior in lung microsomes and
reconstituted lipid vesicles, Toxicol. Appl. Pharmacol. 96(1988) 101-114.
15 F.R. Ungemach, Pa[hobiochemical mechanism of hepatocellular damage following lipid peroxdation,
Chem. Phys. Lipids 45(1987) 171-205.
16 C.E. Douglas, A.C. Chan and P.C. Choy, Vitamin E inhibits platelet phospholipase A2, Biochim.
Biophys, Acta 876 (1986) 369-645.
17 E. Niki, T. Saito, A. Kawakami and Y. Kamiya, Inhibition of oxidation of methyl linolate in solution by
vitamin E and vitamin C, J. Biol. Chem. 259 (1984) 4177-4182.
18 M. Scarpa, A. Rigo, M. Maiorino, F. Ursini and C. Gregolin, Formation of a-tocopherol radical and
recycling of a-tocopherol by ascorbate during peroxidation of phosphatidylcholine liposomes, Biochem.
Biophys. Acta 801 (1984) 215-219.
19 D.C.Lieber, D.S. Kling and D.J. Reed, Antioxidant protection of phospholipid bilayers by a-tocopherol, J.
Biol. Chem. 261 (1986) 12114-12119.
20 T. Doba, G.W. Burton and K.U. Ingold, Antioxidant and co-oxidant activity of vitamin C, Biochem.
Biophys. Acta 835(1985) 298-303.
21 F.E. Hunter Jr., J.M. Gebicki, P.E. Hoffsten, J. Weinstein and A. Scott, Swelling and lysis of rat liver
mitochondria induced by ferrous ions, J. Biol. Chem. 238 (1963), 828-835.
22 E.D. Wills, Lipid peroxidation in microsomes, Biochem. J. 113(1969) 315-324.
23 L. Ernster, K. Nordenbrand and S. Orrenius, Microsomal lipid geroxidation, in K. Yagi (Ed.) Lipid
Peroxides in Biology and Medicine, Academic Press, Orlando, F1, 1982, pp. 777-784
24 D.J. Kornbrust and R.D. Mavis, Microsomal lipid peroxidation, Mol. Pharmacol. 17(1980)400-407.
25 A. Samuni, J. Aronovitch, D. Godinger, M. Chevion and G. Czapski, On the cytotoxicity of vitamin C and
metal ions, Eur. J. Biochem. 137(193) 119-124.
26 E. Shinar, T. Navok and M. Chevion, The analogous mechanisms of enzymatic inactivation induced by
ascorbate and superoxide in the presence of copper, J. Biol. Chem. 258(1983) 14778-14783.
27 B.H.J. Bielski and H.W. Richter, Some properties of the ascorbate free radical, Ann. N. Y. Acad. Sci. 258
(1975)231-237.
28 A.R. Morgan, R.L. Cone and T.M. Elgert, The mechanisms of DNA strand breakage by vitamin C and
superoxide and the protective roles of catalase and superoxide dismutase, Nucl. Acids Res. 3(1976) 11391149.
29 M. Scarpa, R. Stevanato, P. Viglino and A. Rigo, Superoxide ion as active intermediate in the
24

autooxidation of ascorbate by molecular oxygen, J. Biol. Chem. 2~8 (1983)6695-6697.


30 J.R. Bucher, M. Tien and S.D. Aust, The requirement for ferric in the initiation of lipid peroxidation by
chelated ferrous iron, Biochem. Biopys. Res. Comm. lIl (1983) 777-7~4.
31 A. Bast and M.H.M. Steegs, Hydroxyl radicals are not involved in NADPH dependent microsomal lipid
peroxidation, Experientia 42(1986)555-556.
32 J. M. Braugler, L.A. Duncan and R.L. Chase, The involvement of iron in lipid peroxidation, 7. Biol.
Chem. 261 (1986) 10282-1029.
33 G. Minotti and S.D. Aust, The requirement for iron (III) in the initiation of lipid peroxidadon by iron (II)
and hydrogen peroxide, J. Biol. Chem. 262(1987) 1098-1104.
34 N. Nishikimi, Oxidation of ascorbic acid with superoxide anion generated by the xanthine-xanthine oxidase
system, Biochem. Biophys. Res. Comm.63(1975) 463-768.
35 K.Fukuzawa, S. Takase and H. Tsukatani, The effect of concentration on the antioxidant effectiveness of atocopherol in lipid peroxidation induced by superoxide free radicals, Arch. Biochem. Biophys. 240 (1985)
117-121.
36 P.B. McCay, Vitamin E,interactions with free radicals and ascorbate, Ann. Rev. Nutr. 5(1985)323-340.
37 G.R.M.M. Haenen, N.P.E. Vermeulen, H. Timmerman and A. Bast, Effect of several thiols on lipid
peroxidation in rat liver microsomes, Chem.-Biol. Interact (1989) in press (chapter 7).
38 Martindale, The extra pharmacopoeia,28~ Edition, Th Phannaceutical Press, London, 1982, p 1287.
39 E. Degkwitz, S. Walsch, M,Dubberstein and J. Winter, Ascorbic acid and cytochromes, Ann. N. Y. Acad.
Sci. 258 (1975) 201-208.
40 L.R.C. Barclay, The cooperative antioxidant role of glutathione with alipid-soluble and water soluble
antioxidant during peroxidation of liposomes initiated in the aqueous phase and in the lipid phase, J. Biol.
Chem. 263(1988) 16138-16142.
41 L. Floh, G. Loschen, W.A. Guntzler and E. Eichele, Glutathione peroxidase V, the kinetic mechanism,
Hopper-Seyler's Z. Physiol. Chem. 353 (1971) 987-999.
42 G.C. Mills, Hemoglobin catabolism 1, Glutathione peroxidase, an erythrocyte enzyme which protects
hemoglobin from oxidative breakdown,J. Biol. Chem 232(1958) 189-197.
43 F. Ursini, M. Maiorino, M. Valente, L. Ferri and C. Gregolin, Purification from pig liver of a protein
which protects liposomes and biomembranen from peroxidative degradation and exhibits glutathione
peroxidase activity on phosphatidylcholine hydroperoxides, Biochim. Biophys. Acta 710(1982) 197-211.
44 P.B. McCay, D.D. Gibson, K.-L. Fong and K.R. Hornbrook, Effect of glutathione eeroxidase activity on
lipid peroxidation in biological membranes, Biochim. Biophys. Ac[a 431 (1976)459-468.
45 A. Sevenian, S.F. Muakkassah-Kelly and S. Montestruque, The influence of phospholipase A2 and
glutathione eeroxidase on the elimination of membrane lipid peroxides, Arch. Biochem. Biophys. 223
(1983)441-452.
46 F.J.G.M. van Kuyk, A. Sevenian, G.J. Handelman and E.A. Dratz, A new role for phospholipase A2:
Protection of membranes from lipid peroxidadon damage,Trends. Biochem. Sci.12(1987)31-34.
47 G.R.M.M. Haenen, N.P.E. Vermeulen, H. Timmerman and A. Bast, Is phospholipase A2 involved in the
glutathione-dependent protection against in vitro lipid peroxdadon? in: Medical biochemical and chemical
aspects of free radicals, Elsevier, Amsterdam (1989)in press.
48 P. lin, U.H. Danielson and B. Mannervik,4-Hydroxyalk-2-mals are substrates for glutathione transferase,
FEBS Lett. 179 (1985) 267-270.
49 J.R. Prohaska, The glutathione eeroxidase activity of the glutathione-S-transferases, Biochim. Biophys.
Acta 611 (1980) 87-98.
50 R. Morgenstem and J.W. Depierre, Membrane-bound glutathione transferaces, in: H. Sies and B. Ketterer
(Eds.), Glutathione conjugation, mechanisms and biological significance, Academic press, London-New
York (1988) pp. 157-174.
51 B. Ketterer, D.J. Meyer and A.G. Clark, Soluble glutathione transferase isozymes, in: H. Sies and B.
Ketterer (Eds.), Glutathione conjugation, mechanisms and biological significance, Academic press, LondonNew York(1988)pp. 74-135.
52 K.H. Tan, D.J. Meyer, J. Belin and B. Ketterer, Inhibition of microsomal lipid peroxidation by glutathione
and glutathione Iransferases B and AA,Biochem. J. 220(1984)2~3-252.
53 E. Mousialou and R. Morgenstern, Studies on the glutathione dependent inhibition of lipid peroxidation,
Eur. Workshop on Drug Metab., Konstanz FRG (1988) poster 3.15.
54 R.F. Burk, Glutathione-dependent protection by rat liver microsomal protein against lipid peroxidation,
25

Biochim. Biophys. Acta i5i (1983) 21-28.


55 C.C. Reddy, R.W. Scholz, C.E. Thomas and E.J. Massaro, Vitamin E dependent reduced glutathione
inhibition of rat microsomal lipid peroxidation, Life Sci. 31 (1982) 571-576.
56 M.E. Murphy and J.P. Kehrer, Simultaneous measurement of tocopherols and tocopherol q~inones in tissue
fractions using high-performance liquid chromatography with redox cycling electrochemical detection. J.
Chromatogr. 421 (1987) 71-82.
57 G.R.M.M. Haenen, N.P.E Vermeulen., J.N.L. Tai Tin Tsoi, H.M.N. Ragedi, H. Timmerman and A. Bast,
Activation of the microsomal glutathione-S-transferace and reduction of the glutathione dependent protection
by acrolein, Biochem. Pharmacol. 37(1988) 1933-1938.
58 A. Bast, G.R.M.M Haenen and E.M. Savenije-Chapel, Inhibition of rat hepatic microsomal lipid
peroxidation by mesna via glutathione, Arzneim. Forsch. 39(198?) 1043-1045.
59 L.C. Forni and R.L. Willson, Vitamin C and consecutive hydrogen atom and electron transfer in free radical
reactions: A novel catalytic role for glutathione, in: D.C. McBrein and T.F. Slater (Eds.) Protective agents
in cancer, Academic Press, New York-London (1983) pp. 161-173.
60 H.-W. Leung and P.E. Morrow, Interaction of glutathione and ascorbic acid in guinea pig lungs exposed to
nitrogen dioxide, Res. Comm. Chem. Pathol. Pharmacol. 31 (1981) 111-118.
61 A. Bast and G.R.M.M. Naenen, Cytochtome P-450 and vitamin E free radical reductase: Formation and
protection against free radicals, in: A. Crastes de Paulet, L. Douste Blazy and R.Paoletti (Eds)Free radicals,
lipoproteins and membrane lipids, Plenum Press,(1989)in press.
62 A. Bast and G.R.M.M. Naenen, Interplay between lipoic acid and glutathione in the protection against
microsomal lipid peroxidation, Biochim. Biophys. Acta 963(1988)558-561 (chapter 8).
63 T.M. Egan, J.O. Mints, K.G. Scrimgeour and J.D. Cooper, Captopril - A potential free radical scavenger:
inhibition of PMN NADPH oxidase, Clin. Invespg. Med. 11 (1988) 351-356.
64 H. Sies, St. Kaiser and P. Di Mescio, Thioctc acid and singlet oxygen, in: Neue biochemische,
pharmakologische and klinische Erkentnisse zur Ttiocts~ure. Abstract Intern. Workshop on thioctic acid in
Wiesbaden (FRG),February 1989.
65 J.R.Dawson, K. Norbeck, I. Anundi and P. Moldus, The effectiveness of N-acetylcysteine in isolated
hepatocytes, against the toxicity of paracetamol, acroleine and paraquat, Arch. Toacicol. 55(1984) 11-15.
O.I.
Aruoma,B. Halliwell, B.M. Hcey and J. Buller, The antioxidant action of N-acetylcysteine: Its reaction
66
with hydrogen peroxide, hydroxyl radical, superoxide, and hypochlorous acid,Free Rad. Biol. Med.6(1989)
593-597.
67 G.M. Bartoli, A. Muller, E. Cadenas and H. Sies, Antioxidant effect of diethyldithiocarbamate on
microsomal peroxidation assessed by low-level chemiluminescence and alkane production,FEBS Le[[. 164
(1983)371-374.
68 B.D, Goldstein, M.G. Rozen, J.C. Quintavilla and M.A. Amoroso, Decrease in mouse lung and liver
glutathione pero~dase activity and potentiation of the lethal effects of ozone and paraquat by the superoxide
dismutase inhibitor diethyldithiocarbamate, Biochem.Pharmacol. 28(1979)27-30.
69 D.Y. Mitchell and D.R. Petersen, The oxidaon of a-~i unsaturated aldehydic products of lipid peroxidation
by rat liver aldehyde dehydrogenases, Toxicol. Appl. Pharmacol.87(198 403-410.
70 W.C. Bowman and M.J. Rand, Textbook of Pharmacology, Blackwell Scientific Publications Second
Edition, 1980.
71 M. Kato and K.J. Kato, Effects of N-(2-mercaptopropionyl)glycine on ischemic-reperfused dog kidney in
vivo and membrane preparation in vitro, Mol. Cell. Biochem. 78 (1987) 151-159.
72 J.C. Roberts, H.T. Nagasawa, R.T. Zera, R.F. Fricke and D.J.W. Goon,Prodrugs of L-cysteine a protective
agents against acetaminophen-induced hepatotoxicity. 2-(Polyhydroxyalkyl)- and 2(polyacetoxyalkyl)thiazolidine-4(R)-carboxylic acids, J. Med. Chem.30(1987) 1891-1896.
73 J.M. Williamson and A. Meister, Stimulation of hepatic glutathione formation by administration of L-2oxothiazolidine-4-carboxylate, a 5-oxo-L-prolinase substrate, Proc. Nati. Acad. Sci. USA 78(1981) 936939.
74 A.K. Ratty and N.P. Das, Effect of flavanoids on nonenzymatic lipid peroxidation: Structure-activity
relationship, Biochem. Med Metabol. Biol. 39(1988)69-79.
75 I.B. Afanas'ev, A.I. Dorozhko, A.V. Brodski, V.A. Kostyuk and A.I, Potapovitch, Chelating and free radical
scavenging mechanisms in inhibitory action of rutin and quercetin in lipid peroxidation, Biochem.
Pharmacol. 38 (1989) 1763-1769.
76 A.V. Lebedev, L.V. Bogulslayskaya, D.O. Levitskii and O.B. Maksimov, Mechanisms of inhibition of
26

Fee+-induced oxidation of phosphaddylcholine by polyhydroxynaphtoquinones, Eng, Transl. of Biokhimiya


53(1988) 598-603.
77 D.R. Janero and B. Burghardt,Protection of rat myocardial phospholipid against peroxidalive injury through
superoxide-(xanthine oxidase)-dependent, iron-promoted Fenton chemistry by the male contraceptive
gossypol, Biochem. Pharmacol. 37(1988) 3335-3342.
78 T. Osawa, A. Ide, J.-D. Su and M. Namiki, Inhibition of lipid peroxidation by ellagic acid, J. Agric. Food
Chem. 35(1987) 808-812.
79a O.I. Aruoma and B. Halliwell, The iron-binding and hydroxyl radical scavenging action of anti-inflammatory
drugs, Xenobiotica 18(1988)459-470.
79b R. van de Straat, G.J. Bijloo and N.P.E. Vermeulen,Paracetamol, 3-monoalkyl- and 3,5-dialkyl-substituted
derivatives, Biochem.Pharmacol. 37(1988) 3473-3476.
80 E. Rekka, J. Kolstee, H. Timmerman and A. Bast, The effect of some HZ-receptor antagonists on rat hepatic
microsomal cytcchrome P~~50 and lipid geroxidation in vitro, Eur. J. Med. Chem. ~4(1989)43-54.
81 I. Tong Mak and W.B. Weglicki, Protection by R-blocking agents against free radical-mediated sarcolemmal
lipid peroxidation, Circulation Res.~i (1988), 262-266.
82 D.R. Janero, B. Burghardt and R. Lopez, Protection of cardiac membrane phospholipid gainst oxidative
injury by calcium antagonists, Biochem. Phannacol. 37(1988)4197-4203.
83 E. Rekka,R. Mannhold, A. Bast and H. Timmerman, Molecular pharmacological aspects of andanhythmic
activity I. Group I and group III compounds and lipid peroxidadon,Biochem.Pharmacol.(submitted).
84 R.H.M. Julicher, L.B.M. Tijburg, L. Sterrenberg, A. Bast, J.M. Koorren and J. Noordhoek, Decreased
defence against free radicals in rat heart during normal reperfusion after hypoxic ischemic and calcium-free
perfusion, Life Sci. 35(1984) 1281-1288.
85 R. Nikolov and K. Koburova, Inhibition of phospholipase A2 in vitro by some anti-hypoxic drugs, Meth.
Find. Expfl Clin. Pharmacol. 6 (1984) 429-431.
86 J.M. Brown, M.A. Grosso, G.J. Whitman, A. Banerjee, L.S. Terada, J.E. Repine, A.H. Harken, The
coincidence of myocardial reperfusion injury and hydrogen peroxide production in the isolated heart, Surgery
105 (1989) 496-501.
87 A. van der Vliet, T.J.R. Tuinstra and A. Bast, Modulation of oxidative stress in the gastrointestinal tract and
effect on rat intestinal motility, Biochem.Pharmacol.(in press).
88 D.R. Janero, R. Lopez, J. Pittman and B. Burghardt, Propranolol as xanthine oxidase inhibitor:
Implications for antioxidant activity, Life Sci. 44 (1989) 1563-1573.
89 A. Bast, R. Leurs and H. Timmerman, Cyclandelate as a calcium modulating agent in rat cerebral cortex,
Drugs 33(1987) 67-74.
90 D.I. Wilhelm, G. Lcer, Th. Peters, Differential influence of various calcium entry blockers on ouabain
toxicity in rat left atria. 27th Spring Meeting Deutsche Pharmakologische Gesellschaft, Abstract 176,
Mainz 1976.
91 I. Zs. Nagy and R.A. Floyd, Electron spin resonance spectroscopic demonstration of the hydroxyl free
radical scavenging properties of dimethylaminoethanoi in spin trapping experiments confirming the
molecular basis for the biological effects of centrophenoxine, Geriair. 3(1984)297-310.
92 M. Wasil, B. Halliwell and C.P. Moorhouse, Scavenging of hypochlorous acid by tetracycline, rifampicin
and some other antibiotics: a possible antioxidant action of rifampicine and tetracycline? Biochem.
Pharmacol. 37(1988) 775-778.
93 P.C. Moorhouse, M. Grootveld, B. Halliwell, J.G. Quinlan and J.M.C. Gutteridge, Allopurinol and
oxypurinol are hydroxyl radical scavengers,FEBS Lett. 213(1987)23-28.
94 D.A.Peterson, B. Kelly and J.M. Gerrard, Allopurinol can act as an electron transfer agent. Is this relevant
during reperfusion injury? Biochem. Biophys. Res. Comm. 137(1986) 76-79.
95 A.J. Carmichael, C.M. Arroyo and L.G.Cockerharn, Reaction of disodium cromoglycate with hydrated
elech~ons, Free Rad. Biol. Med.4(1988)215-218.
of
96 F. Ursini, M. Maiorino, P. Hochstein and L. Ernster, Microsomal lipid peroxidation: mechanisms
31-36..
6
(1989)
Biol.
Med.
Rad.
Free
chelators.
anduon
iron
of
initiation. The role
97 M.J. Davies, R. Donkor, C.A. Dunster, C.A. Gee, S. Jonas and R.L. Wilson, Desferrioxamin (Desferal)
and superoxide radicals. Formation of anenzyme-damaging nitroxide, Biochem. J. 246(1987)725-729.
98 M. Praet, P. Buc Calderon, G. Pollakis, M. Roberfroid and J.M. Ruysschaert, A new class of free radical
scavengers reducing adriamycin mitochondrial toxicity, Biochem.Pharmacol.37(1988)4617-4b22.
27

99 J.M. Braughler, J.F. Pregenzer, R.L. Chase, L.A. Duncan, E.J. Jacobsen and J.M.
McCall, Novel 21amino steroids as potent inhibitors of iron-dependent lipid peroxidation, J. Biol Chem.
262 (1987) 1043810440.
100 A. Wendel, M. Fausel, H. Safayhi, G. Tiegs and R. Otter, Activity of PZ 51 in
relation to glutathione
peroxidase, Biochem. Pharmacol. 33(1984)3241-3245.
101 R. Leurs, H. Timmerman and A. Bast, Inhibition of superoxide anion radical production
by ebselen(PZ 51)
and its sulfur analogue(PZ 25) in guinea pig alveolar macrophages, Biochem. Int. 18
(1989)295-299.
102 I.A. Cotgreave, S.K. Duddy, G.E.N. Kass, D. Thompson and P. Moldus, Studies on
the anti-inflammatory
activity of ebselen. Ebselen interferes with granulocyte oxidative burst by dual inhibition
of NADPH
oxidase and protein kinase C? Biochem. Pharmacol. 38(1989)649-656.
103 H.Fisher and N. Dereu, Mechanism of the catalypc reduction of hydropero~cides by ebselen:
A selenium-77
study. Bull. Soc. Chim. Belg. 96 (1987) 7~7-768.
104 J.R.J. Sorensen, Copper complexes as 'radiation recovery' agents, Chem. Britain Febr.
(1989) 169-172.
105 T. Ogino, M. Inoue, Y. Ando, M. Awai, H. Maeda and Y. Morino, Chemical modificati
on of superoxide
dismutase. Extension of plasma half life of the enzyme through its reversible binding
to the circulating
albumin, Int. J. Peptide Protein Res, 32(1988) 153-159.
106 J.S. Beckman, R.L. Minor Jr., C.W. White, J.E. Repine, G.M. Rosen, B.A. Freeman,
Superoxide
dsmutase and catalase conjugated to polyethylene glycol increases endothelial enzyme
activity and oxidant
resistance, J. Biol. Chem. 264(1988) 6884-6892.
107 H. Shirota, K. Katayama, H. Ono, K. Chiba, S. Kobayashi, K. Terato, H.
Ikuta and I. Yamatsu,
Pharmacological properties of N-methoxy-3-(3,5-ditert-buty113-hydroxybenzylidene)-2-pyrrol
idone(E5110)a
novel nonsteroidal antrinflammatory agent, Agents Actions 21(1987)250-252.
108 K. Tajima, M. Sakamoro, K. Okada, K. Mukai, K. Ishizu, H. Sakurai and H. Mori,
Reaction of biological
phenolic antioxidants with superoxide generated by cytochrome P-450 model system, Biochem.
Biophys.
Res. Comm. 115 (1983) 1002-1008.
109 G.W. Burton and K.U. Ingold, Vitamin E. Application of the principles of physical organic
chemistry to
the exploration of its structure and funcpon, Accounts of Chem. Res. 19 (1986) 194-201.
110 K. Kato, S. Terao, N. Shimamoto and M. Hirata, Studies on scavengers of active
oxygen species 1.
Synthesis and biological activity of 2-O-alkylascorbic acids, J. Med. Chem. 31 (1988)793-798.

28

- Chapter 3 Initiation and propagation in lipid peroxidation.


Abstract. Lipid peroxidation is a radical reaction. Therefore, initiation and propagation mean
the generation and transfer of radicals respectively. Also other processes in lipid peroxidation
have been named initiarion or propagation. We propose to ban these alternarive definitions of
initiation and propagation in order to prevent confusion. Moreover, the importance of the
difference between lipid hydroperoxide-independent and lipid hydroperoxide-dependent lipid
peroxidation is outlined.
The process of lipid peroxidarion is involved in the etiology of various patho-physiological
conditions [1]. In this process poly unsaturated fatty acids (PUFA's)located in the membrane
are peroxidized by a free radical mechanism. Radical reactions can be divided in three types; (i)
iniriation, i.e. the generation of radicals, (ii) propagation, i.e. the transfer of radicals and (iii)
termination, i.e. the disappearance of radicals.
One of the primary events in lipid peroxidation is the abstraction of a hydrogen atom of a
PUFA (LH). This is considered to be the reaction of the PUFA with a radical. Although often
the hydroxyl radical(OH')is nominated as the radical responsible for this hydrogen abstraction,
the identity of the radical involved is still obscure [2]. Nevertheless, we also use OH'to denote
this radical:
LH + OH'~L'+ H2O
Usually, this reaction is referred to as being the initiation reaction in lipid peroxidation [2].
However, in this reaction no radicals are being formed, the radical is only being transfereed
from the hydroxyl radical to the PUFA (LH). Therefore, the reacrion should be characterized as
a propagation reaction. The actual iniriation reaction is the reaction that generates the hydroxyl
radical, e.g. the Fenton reaction:
Fel+ + H2O2~ Fe3+ + OH- + OH'
The Fenton reacrion has also been characterized as a propagation reaction, namely in the iron
catalyzed decomposition of hydrogen peroxide [3]. In principle, Fel+ is already a free radical
since it contains 4 unpaired electrons. In the iron catalyzed decomposition of hydrogen
peroxide, the reducrion of Fe3+ to Fel+ was denoted as the iniriation reaction [3]. There are
some indicarions that the reduction of iron is also involved in the onset of lipid peroxidation.
Nevertheless, in the reduction of iron no radicals are being produced. Actually, the number of
unpaired electrons in Fe3+(5)is higher, making it "more of a radical" than Fel+. Therefore, we
do not recommend to refer to the reduction of iron as the iniriation reaction, but rather to limit
the use of initiation to the Fenton reaction or Fenton-like reactions.
Beside the radical transfer between the hydroxyl radical and the PUFA which generates a
lipid radical, other propagation reactions are involved in lipid pero~darion, The lipid radical(L')
reacts with oxygenl.
L'+02--~ LOO'
when the
1 Strictly speaking, this reaction would be a termination reaction rather than a propagation reaction,
biradical nature of oxygen is being considered.
29

IiZOZ + Fe
'
Z

~
~ ~,

primary
initiation

propagation

;~
5~.,
a :~

.
~ ;~

~
Iy

Lom .

r
secondary
initiation

~2

Lv,

L~

L
LO

~LO

OZ yLO

IAOH

LOOH

LOOH
F~

r
Lo'
LOH
~~

,~ ~
i p~

~
O s~.,
ati

LH

~ p
~

,,

OH'+OH +Fe
fiZO

N Cj
pi
j.~

propagation

L~
02 Lod

O
2

L~ _
L.
~o
wox
i.00x

Fig. 1 Sequence of reaction in lipid peroxidation. LH is apoly-unsaturated fatty acid.

Subsequently, the lipid peroxyl radical(LOO') generated reacts with another PUFA,and a lipid
hydroperoxide(LOOH)and a new lipid radical (L') are being formed:
LOO'+ LH~LOOH + L'
The new lipid radical is also converted into a lipid hydroperoxide after subsequent propagation
reactions with oxygen and a third PUFA.In these reactions the radical is transferred to the third
PUFA, and eventually the radical is transferred again to a next PiTFA, while the third PiTFA is
converted into a lipid hydroperoxide (fig 1).
In principle, only one initiarion reacrion is needed to peroxidize all PLTFA's in a membrane.
However;several termination reacrions occur, e.g.:
LOO'+ L ~LOOL
LOO'+ LOO'~LOOL +02
Moreover,in a radical transfer a relarively stable radical can be formed. The classical example of
such a reaction is the reacrion of vitamin E with a lipid peroxyl radical:
LOO'+ Vit E ~LOOH + Vit E'
The vitamin E radical (Vit E') is fairly unreactive, because the free electron is effecrively
delocalized over the chromanoxyl head of vitamin E molecules [4]. In principle, the scavenging
of radicals by vitamin E does not stop the chain reaction of lipid peroxidation, and therefore this
is not a termination reaction. However, the radical is supposed to be transferred from the lipid
bilayer to the aqueous phase in a propagation reacrion of the vitamin E radical with GSH or
vitamin C. Subsequenfly, terminarion takes place in the aqueous phase.
The lipid peroxides formed during lipid peroxidation are considered to be very reactive, they
can iniriate another radical chain by e.g. homolyric of heterolyric cleavage:
LOOH~LO'+ OH'
LOOH-~ LOO'+ H'
with
and bring about the same sequence of reactions as
can
react
PUFA's
These radicals
described above (fig 1). Lipid peroxidation induced by radicals originating from lipid
30

hydroperoxides is called LOOH-dependent lipid peroxidation or chain branching. In LOOHindependent lipid peroxidation the radical that abstracts a hydrogen atom of the first PiTFA is
not derived from a lipid hydroperoxide [5]. In both processes, the propagation and termination
reactions of the PUFA's are identical. The only difference is the initiation reaction (fig 1). The
initiation reaction in LOOH-independent lipid peroxidation is called primary initiarion, while in
LOOH-dependent lipid peroxidarion it is called secondary iniriation. T'he importance of LOOHindependent and LOON-independent lipid peroxidation was expounded by Seingen et al. [6].
Unfortunately, they defined LOOH-independent lipid peroxidation as iniriation, and LOOHdependent lipid peroxidation as propagation, terms that already have another meaning in free
radical reactions. Despite a previous attempt of us [5] to correct the definitions of Seingen et al.
[6], initiation and propagation difined according to Seingen et al.[6] is still frequently used [7,
8].
There is no consensus about the relative contributions of LOOH-independent lipid
peroxidarion and LOOH-dependent lipid peroxidarion to the overall process of lipid peroxidarion
[6, 7, 9]. The contribution of each process is also dependent on the system used to study lipid
peroxidation. In the autoxidation of PiIFA's there is probably mainly LOON-dependent lipid
peroxidation, while in iron catalyzed peroxidation in vitro of e.g. microsomes we tentatively
believe their is a higher contribution of LOON-independent lipid peroxidation. This is partially
based on the time course of lipid peroxidation. In autoxidation the production of lipid
hydroperoxides in time is exponential, and it takes several days to reach completion. In contrast,
in vitro iron catalyzed lipid peroxidation in microsomes is not exponential in time and is
completed after several minutes. This fast lipid peroxidation is probably a result of the extensive
radical stress that is forced upon the lipid membrane.
It is impossible to discriminate between the two processes of lipid peroxidation in a simple
experiment, because during LOOH-independent lipid peroxidation lipid hydroperoxides are
produced, and there is no LOOH-independent lipid peroxidation without LOOH-dependent lipid
peroxidation. Despite claims that either LOOH-independent [7] or LOOH-dependent lipid
peroxidation [6] is the most important process in vitro, in our opinion the exact contribution of
both processes to in vitro lipid peroxidarion remains to be established.
The relative contriburions of LOOH-dependent and LOOH-independent lipid peroxidarion are
not purely of academic interest, there are some important practical implications. Several
protective mechanisms are based on the detoxication of lipid hydroperoxides. Namely, the
glutathione peroxidases (GSH-px) catalyze the conversion of lipid hydroperoxides into the
corresponding less reacrive alcohols(LOH):
GSH-~ LOH + H2O + GSSG
LOOH + 2 GSH
The formed lipid alcohols(LOH)are not active in any initiarion reaction. An essenrial drawback
of the protection by some of the GSH-px's is that during lipid peroxidation phospholipid
hydroperoxides(PLOOH) are being produced. During lipid peroxidation the PITFA's remain
esterified to glycerol [9, 10]. Several GSH-px's display no catalytic activity with these
phospholipid hydroperoxides. The phospholipid hydroperoxide (PLOOH) first has to be
hydrolyzed to a free lipid hydroperoxide in a phospholipase A2(PLA2)-mediated reaction (fig 2)
[9, 10]. The free lipid hydroperoxide is accepted by all GSH-px's as substrate. Tan et al. [10]
suggested that the combinarion of a membrane-bound phospholipase AZ and a cytosolic
31

PLOOH
PLA2

2 GSH

OOH
GSSG

GSH-px
secondary
initiation

LOH + H2O

Fig. 2 Scheme of reactions in the protection against lipid peroxidtion by the combination of
phospholipase A2(PLA2)and glutathione peroxdase (GSH-px). During lipid peroxidation phospholipid
hydroperoxides(PLOOH)are being produced. Before these hydroperoxides can be detoxified by some
GSH-px's, the phospholipid hydropermcides have to be hydrolyzed, to the free lipid hydroperoxide(LOOI
and a lysophospholipid, in a PLA2-mediated reaction. The free lipid hydroperoxide(LOOD reacts with
GSH in a reaction catalyzed by the GSH-px, and an unreactive lipid alcohol(LOIN is formed. In order to
protect against lipid peroxidation, the combination of PLA2 and GSH-px has to compete with the
secondary initiation.

GSH-px is able to protect against in vitro lipid peroxidation of liver microsomes. However, if
our assumption is correct that in vitro microsomal peroxidarion consists for the greater part of
LOOH-independent lipid peroxidarion, this automatically indicates that no effective protectin
by the combination of phospholipase A2 and GSH-px is possible. By the combination of the
enzymes, only protection is provided against LOOH-dependent lipid peroxidation.
Nevertheless, let us assume that in vitro microsomal lipid peroxidation is primarily the result of
LOOH-dependent lipid peroxidation. Since in vitro lipid peroxidation is a very fast process,
this would imply that secondary initiation proceeds very fast. In order to protect, the
combination of phospholipase A2and GSH-px has to compete with this fast secondary initiarion
(fig 2). In this respect it has to be noted that phospholipase A2-activity in liver microsomes is
extremely low. In the liver of the rat the total activity of phospholipase A2 assosiated with the
microsomal fraction is less than 5 mU [11]. Moreover, the phospholipase AZ activity in liver
microsomes is calcium-dependent[11] and no calcium is present in the experiments performed
by Tan et al. [10] on control microsomes. Under these circumstances, the activity of the
microsomal phospholipase A2 is practically absent. Therefore, the detoxication of lipid
hydroperoxides by p_hospholipase A2 and GSH-px cannot compete with the relatively fast
secondary initiation reactions. This indicates that the protection against microsomal peroxidation
by the combination of phospholipase A2 and the cytosolic GSH-tr as proposed by Tan et al.
[10] does not seem very likely. Moreover, this mechanism has been challenged on other points
also [12].
This commentary aims to restrict the future use of the terms initiarion and propagation in lipid
peroxidation to reactions where radicals are being formed or transferred respectively. Alternative
definitions of initiation and propagation should be banned, in order to prevent confusion. It
should be noted that radicals do not initiate, they only propagate or terminate. Knowledge of the
32

relarive contributions of LOOH-independent and LOOH-dependent lipid peroxidation to the


overall process of lipid peroxidation is important to fully apprehend the defense mechanisms
against lipid peroxidarion. In our opinion in vitro lipid pero~dation consists for the greater part
of LOOH-independent lipid peroxidation. Therefore, it has to be anticipated that the protection
based on the dtoxication of lipid hydroperoxides by the GSH-px's cannot be effective in vitro.
In vivo, the relative contribution of LOOH-dependent lipid peroxidation might be more
substantial, indicating that in vivo the role of the GSH-px's might be more important than
based on in vitro studies.
References.
H. Sies(Ed.) Oxidative Stress, Academic Press, New York-London, 1985.
J.M.C. Gutteridge,in: B. Halliwell (Ed.) Oxygen radicals azd tissue i,~ju~-y, 1987, p~. 9-1~
H.B. Danford, J. Free Rad. Biol. Med. 3(1987)405-421.
G.W. Burton, T. Doba, E.J. Gabe, L. Hughes, F.L. Lee, L. Prasad and K.U. Ingoid, J. Am. Chem. Soc.
107 (1985) 7053-7065.
5 A. Bast and G.R.M.M. Hamen,Trends Biochem. Sci. 9(1984)510-513.
6 B.A. Seingen, J.A. Buege, F.A. O'Neal and S.D. Aast, Photochem. Photobiol. 58 (1978) 803-809.
7 H.W. Davis, T. Suzuki and J.B. Schenkman, Arch. Biochem. Biophys. 252(1987)218-228.
8 K.H. Cheeseman, S. Emery, S.P. Maddix, T.F. Slater, G.W. Burton and K.U. Ingold, Biochem. J. 250
(1988)247-252.
9 P.B. McCay,D.D. Gibson, K.L. Fong and K.R. Hombrook, Biochim. Biophys. Acta 431 (1976) 459-468.
10 K.H. Tan, D.J. Meyer, J. Belin and B. Ketterer, Biochem. J. 220 (1984)243-252.
11 A.J. Aarsman, J.G.N. de Jong, E. Arnoldussen, F,W, Neys, P.D. Wassenaar and H, van den Bosch, J. Biol
Chem. 264 (1989) 10008-10014.
12 G.R.M.M. Hamen, N.P.E. Vermeulen, H. Timmerman and A. Bast, in: O. Hayaishi, E. Niki, M. Kondo
and T. Yoshikawa (Eds.) Medical, biomedical and chemical aspects of free radicals, Elsevier, Amsterdam,
1989, pp. 1291-1294 (chapter 6).
1
2
3
4

33

34

- Chapter 4 Protection against lipid peroxidation by a microsomal


glutathione-dependent labile factor
Abstract. Glutathione (GSH) protects rat liver microsomes against ascorbic acid (0.2
mM)/ferrous iron (10 M)-induced lipid peroxidation for some time. The inhibitory effect of
GSII is concentrarion-dependent (0.1-1.0 mM). Our data suggest that GSH acts by preventing
inirial radical formation rather than via radical scavenging or GSH-peroxidase activity. A labile
GSH-dependent factor is involved in he inhibition of nucrosomal lipid peroxidation by GSH,
inasmuch as hearing the microsomes abolishes the GSH effect. We found that besides heating,
lipid peroxidation also destroys the GSH-dependent factor. Consequently, continuous radical
stress will produce lipid peroxidation, despite the presence of GSH. Moreover, a detrimental
effect of in vivo-induced lipid peroxidation (CC14-treatment) on the GSH-dependent factor was
observed. The implications of the present data for the genesis of and the protection against
peroxidative damage are discussed.
Introduction.
Lipid peroxidation has been implicated as a major process in cellular damage [1]. Reduced
glutathione (GSH) is considered to be a potent inhibitor of lipid peroxidation, but the
mechanisms by which it carries out this function are not clear. T'he protecrive acrion of cytosolic
glutathione peroxidase and glutathione transferace has been questioned. Both enzymes do not
reduce microsomal lipid peroxides as a measurable rate [2,3]. It has also been stated that the
glutathione-dependent cytosolic factor which inhibits lipid peroxidation in biological membranes
does so by prevenring radical attack on the polyunsaturated fatty acids. Here, the activity of
GSH in microsomal lipid peroxidarion is evaluated.
Materials and methods.
Chemicals. Glutathione and [hiobarbituric acid were obtained from Merck (Dannstadt). All other chemicals
used were of analytical grade purity.
Preparation of microsomes.Male Wistar rats (T.N.O., Zeist), 200-250 g, were killed by decapitation. Livers
were removed and homogenized (1:2, w/v) in ice-cold phosphate buffer (50 mM,pH 7.4) containing 0.1 mM
EDTA. The homogenate was centrifuged at 10 000 x g (20 min at 4C). Subsequently the supernatant was
centrifuged at 10 000 x g (20 min) and again at 65 000 x g (60 min). The microsomal pellet was resuspended
in the phosphate buffer(2 g liver/ml) and stored at -80C. Before use the microsomes were thawed and diluted 5fold with ice-cold Tris-HCl buffer (50 mM, pH 7.4) containing 150 mM KCl and washed twice with
centrifugation at 115 000 x g(40 min). Finally the pellet was resuspended in the Tris buffer and used.
Pretreatment of rats. Carbon tetrachloride (CCl4)pretreatment consisted of the oral administration of a
CC14/corn oil (50%, v/v) solution (5 ml/kg). Water and food were provided ad libitum. After 18 h liver
microsomes were prepared as described above, but were less diluted in order to get the same protein concentration
as in the control incubations. Starvation consisted of a period of 18 h food deprivation before decapitation. Water
was given ad libitum. Phorone treatment was as in [4], 250 mg/kg i.p., and decapitation after 2 h.
Incubation conditions. Microsomes (final cone. 1/g liver/ml) were incubated at 37C, with shaking air being
freely admitted in Tris-HCl/KCl(50 mM/150 mM,pH 7.4). Ascorbic acid and GSH were neutralized with KOH
before addition. Reacpons were started by adding a freshly prepared FeSO4 solution.
Spectral measurements. Lipid peroxidation was assayed by measuring thiobarbituric acid (TBA)reactive

35

12 .

O
O
~
N

0.8

0.4

time (min)
Fig. 1. Influence of GSH on the time course of lipid peroxidation. Microsomes (1.3 mg protein/ml) were
incubated with 0.2 mM ascorbate. [GSH] was 0 mM (t), 0.1 mM (), 0.2 mM (~), 0.5 mM (o),
1 mM (~). Reactions were started with addition of Fel+ (10 M). Data represent one example out of
least 4 duplicate experiments.

material. An aliquot of the incubation (0.3 ml) was stopped by mixing with ice-cold TBA-trichloroacetic acidHCl-butylhydroxytoluene(BHT) solution (2 ml). After heating (15 min, 80C) and centrifugation (15 min) the
absorbance at 535 nm vs 600 nm was determined. The TBA-Irichloroacetic acid-HCl solution was prepared by
dissolving 41.6 mg TBA/10 ml trichloroacetic acid (16.8% w/v in 0.125 N HCI). To 10 ml TBA-trichloroacetic
acid-HCl 1 ml BHT(1.5 mg/ml ethanol) was added. The added chemicals did not interfere with the assay in the
concentrations used. Thiol deterrnination was performed after [5]. Corrections were made for the absorbance
produced by iron. Microsomal protein was assayed as in [6], using bovine serum albumin as standard.

Results.
GSH protects rat liver microsomes against lipid peroxidarion promoted by ascorbic acid and
ferrous iron in aconcentration-dependent manner (fig. 1). This protective effect of GSH is
abolished by heating the microsomes in boiling water for 90 s (fig. 2). This indicates the
involvement of(a) microsomal heat-labile factors)in the GSH effect [7].
To investigate the GSH consumption during lipid peroxidation the TBA-test has been
combined with a thiol deternunation (fig. 3). The results of fig. 1 and 3 show that GSH delays
but does not prevent lipid peroxidation. After a lag time a strong increase in TBA-reactive
material is observed, whereas GSH-levels remain relarively high. In order to explain the rise in
lipid peroxidation which eventually occurs, notwithstanding the presence of GSH, we thought
that the GSH-dependent lipid peroxidation inhibiting factors) may be destroyed during initial
but persistent stress. To investigate the notion, the GSH protecrion after moderate lipid
peroxidation has been examined (fig. 4). No protection by GSH is observed (fig. 4)indicating
that the factors)is(are) indeed destroyed during lipid peroxidation.
Furthermore, we investigated the stability of the factors) in vivo. Pretreatment of rats with
CC14 resulted in a diminished protective effect of GSH in microsomal incubates compared with
the effect observed in hepatic microsomes obtained from oil pretreated (control) rats. Also,
36

'.6,

O
O

~~~

time (min)
Fig. 2. Time course of lipid peroxidation. The reaction mixture consisted of heated microsomes (1.3 mg
protein/ml, before heating) 0.2 mM ascorbate (~), 0.2 mM ascorbate and 1 mM GSH (), no addition
(o), 1 mM GSH (). All reactions were initiatec with addition of Fel+. Data represent one example out
of at least 4 duplicate experiment.

GSH depletion induced by pretreatment with phorone [4] or by starvation [8], reduced the
protection by GSH in microsomes, compared with results established in hepatic mcrosomes
from untreated rats (not shown).

i.00

~---=a--------_~
~~'~-_

F
E
~.~5

=
l7

~/

0.50

0.25

25

10

15

20

30

45

time (min)
()
and thiol concentrations (- - -). Microsomes (1.2 mg
Fig. 3. Time course of lipid peroxidation
protein/ml) were incubated with 0.2 mM ascorbate (~), 0.2 mM ascorbate and 1 mM GSH(t) with (1)
as the corresponding thiol concentration, no addition (o), 1 mM GSH(p), with (0)as the corresponding
thiol concentration. All reactions were started with addition of Fel+. Data represent one example out of 2
duplicate experiments.
37

O
O
M
`I

time (min)
Fig. 4. Time course of lipid peroxidation. Microsomes (1.1 m~ protein/ml) were incubated with 0.05 mM
ascorbate. The reaction was started with the addition of Fe + (10 lvn. After 12.5 min the following
addirions were made: 0.05 mM ascorbate and 1 mM GSH (t), 0.05 mM ascorbate and the same volume of
Tris buffer which was needed for GSH (o), 1 mM GSH (), none (D). Appropriate corrections were
made for changes in volume.

Discussion.
The effect of GSH on ascorbate/iron induced microsomal lipid peroxidation is studied.
Ascorbic acid (0.2 mM)does not stimulate lipid peroxidation in the absence of iron, nor does
ferric iron (10 M)in the absence of ascorbic acid. Ascorbic acid added in combination with
ferric iron results in substantial lipid peroxidation (not shown). Taking this into account, as well
as the fact that ferrous iron is capable of promoting lipid peroxidation (fig. 2,3), the role of
ascorbic acid appears to be maintenance of iron in the reduced (ferrous) form [9]. The
srimulation of lipid peroxidation produced by GSH in a system containing heated microsomes,
ascorbic acid and ferrous iron (fig. 2)can be explained by the regeneration of ascorbic acid by
GSH,inasmuch as GSH is able to reduce dehydroascorbic acid to ascorbic acid. This reaction,
srimulate, indicaring that GSH itself does not reduce trivalent iron at a measurable rate [11](fig.
2). In the system containing ascorbic acid,ferrous iron and microsomes, ascorbic acid becomes
irreversibly oxidized since it is shown that addition of GSH, 12.5 min after ferrous iron, does
not enhance formation of TBA-reactive material (fig. 4)[12].
Microsomal lipid peroxidarion is inhibited by GSH (fig. 1). GSH acts by prevenring radical
formation. The protection by GSH is not due to radical scavenging or GSH-peroxidase acrivity
because in that case:
(i) A decline in GSH levels should be observed during the lag phase. The decrease in GSH
levels per se will lead to enhancement of lipid peroxidation. However, we did not observe a
substantial consumption of GSH (fig. 3) which would bring GSH levels under scavenging
concenirarion.
(ii) Maximal lipid peroxidation should not be the same in peroxidising microsomes with and

38

100.
L
Q1
-~
r0

E
a~
u

50.

.,
0

2.5

10
15
5
20
incubation time (min.)

Fig. 5. Percentage change in lipid peroxidation using a system containing microsomes, 0.2 mM
ascorbate, i mM GSH and 10M Fel+ compared with the lipid peroxidation without GSH, measured at
several incubation times. Each data point represents the mean of a duplicate experiment in which liver
microsomes were prepared &om one rat. Oil-prelrated control rats(o)(1.3-1.6 mg protein/ml) and CC14pretreated rats(~)(1.2-1.9 mg protein/ml).

without GSH. Our results, however, demonstrated the same extent of peroxidarion in both
systems (fig. 1,3).
Protecrion by GSH proceeds via aheat-labile microsomal factor (fig. 2). Apparently, also a
continuous radical stress (fig. 1,3) or lipid peroxidation (fig. 4) affects the GSH-dependent
factor.
Recently, the influence of a-tocopherol (vitamin E)on ascorbate/iron-induced peroxidation
in liposomes was thoroughly studied [13]. One molecule of a-tocopherol was found to be able
to protect even 220 molecules of polyunsaturated fatty acids. Moreover, it was stated that atocopherol may prevent radical formation without undergoing any measurable oxidation itself.
Our results indicate that GSH also acts via inhibition of radical formation. It is therefore
tempting to suggest an interaction between the water-soluble GSH and the lipid-soluble vitamin
E comparable to the presumed linkage between vitamins C and E [14]:

R~

vit. E

2 GSSG

labile
facto r

RH

vit.

GSH

39

Liver microsomes of vitamin E-deficient rats are devoid of a protective effect by GSH [15]. The
conclusion was therefore reached that inhibition by GSH is vitamin E-dependent. However,
vitamin E depletion has been found to produce an enhanced ethane exhalarion in rats [16[. Since
the latter is an index of lipid peroxidarion, the inability of GSH to protect might be explained by
the destruction of the labile factor. In our opinion so far no direct evidence for the vitamin E
dependency of the protective effect by GSH has been adduced. If vitamin E does not play a
role, it is certainly not the only component since in heated microsomes the inhibition by GSH is
eliminated, whereas hearing has no effect on the vitamin E level[15].
Attempts have been made to isolate GSH-dependent protective factor(s). In [17,18] the
investigators claimed to have isolated a GSH-dependent peroxidation inhibiting protein (PIP).
Peculiarly, ho~x~ever, they observed no protective efieci by GSH in intact microsomes.
Furthermore the action of PIP was ascribed to its capacity to reduce hydroperoxide derivatives
of phospolipids, at the expense of GSH. Our data suggest another protecrive mechanism by
GSH.
We also demonstrated the instability of the GSH-dependent protective factors) in vivo. A
decreased protection by GSH is obtained in vitro, using hepatic microsomes from CC14pretreatedrats (fig. 5).
The chain reacrion of microsomal lipid peroxidation might be iniriated after destruction of the
GSH-dependent peroxidarion inhibiting factor(s). Our results suggest that in that stage the
progressive membrane damage cannot be reversed via a direct or indirect increase of GSH
levels. This is of importance in order to be able to provide adequate measures against
hepatotoxic effects of many xenobiotics which give lipid peroxidation.
References.
1 Plaa, G.L. and Witschi, H.(1976) Annu. Rev. Pharmacol. Toxicol. 16, 125-141.
2 Gibson, D.D., Hombrook, K.R. and McCay,P.B.(1980) Biochim. Biophys. Acta 620, 572-582.
3 McCay,P.B., Gibson, D.D. and Hornbrook, K.R.(1981) Fed. Prot. FASEB 40, 199-205.
4 Younes, M. and Siegers, C.P.(1981) Chem.-Biol. Interact. 34, 257-266.
5 Ellman, G.L.(1959) Arch. Biochem. Biophys. 82, 70-77.
6 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J.(1951) J. Biol. Chem. 193-265-275.
7 Burk, R.F.(1982) biochem. Pharmacol. 31, 601-602.
8 Wendel, A., Feuerstein, S. and Konz, K.H.(1979) Biochem. Phannacol. 28, 2052-2055.
9 Kombrust, D.J. and Marvin, R.D.(1980) Mol. Pharmacol. 17, 400-407.
10 Leung, H.W. and Morrow,P.E.(1981) Res. Comm. Chem. Path. Pharmacol. 31, 111-118.
11 Rowley, D.A. and Halliwell, B.(1982)FEBS Lett. 138, 33-36.
12 Haase, G. and Dunkley, W.L.(1969) J. Lipid Res. 10, 555-576.
13 Fukuzawa, K., Tokumura, A., Oachi, S. and Tsukatani, H.(1982) Lipids, 511-513.
14 Packer, J.E., Slater, T.F. and Wilson, R.L.(1979) Nature 278, 737-738.
15 Reddy, C.C., Scholz, R.W., Thomas, C.E. and Massaroet E.J.(1982) Life Sci. 31, 571-576:
16 Tappel, A.L. and Dillard, J.(1981) Fed. Proc. FASEB 40, 174-178.
17 Ursini, F., Maiorino, M., Valente, M., Ferri, L. and Gregolin, C.(1982) Biochim. Biophys. Acta 710, 197211.
18 Maiorino, M., Ursini, F., Leonelli. M., Finato, N. and Gregolin, C.(1982) Biochem. Internat. 5, 575-583.

40

- Chapter 5 4-IIydroxy~2,3-trans-nonenal stimulates microsomal lipid


peroxidation b~ reducing the glutathione-dependent protection.
Abstract. Glutathione (GSH) protects liver microsomes against lipid peroxidation. This is
probably due to the reducrion of vitamin E radicals by GSH,a reacrion catalyzed by a membrane
bound protein. Pretreatment of liver microsomes with 0. or 1 mIVI 4-hydroxy-2,3-transnonenal(HNE), a major product of lipid peroxidation, reduces the GSH-dependent protection.
GSH and vitamin E concentrations are not affected by ths pretreatment. Pretreatment with 0.1
mM N-ethyl maleimide(NEM),a syntheric sulfhydryl reagent, resulted in a similar reducrion as
with HNE of the GSH-dependent protecrion against lipid peroxidarion. The reducrion of the
GSH-dependent protection by HNE and NEM is probably the result of inactivation of the
membrane bound protein by covalent binding to an essential SH group on the protein. If the
GSH-dependent protecrion would proceed via the microsomal GSH transferase, pretreatment
with NEM, which activates the microsomal GSH transferace, should enhance the GSHdependent protection. Actually, a decrease in the GSH-dependent protection is found.
Apparently the GSH-dependent protection does not proceed via the microsomal GSH
transferace. Also the microsomal phospholipase A2 is not involved, since addition of 0.1 mM
mepacrine, an inhibitor of phospholipase A2, did not preclude the GSH-dependent protection.
Once the process of lipid peroxidation, either in vivo or in vitro, has started, the protecrion of
liver mcrosomes by GSH is less effective. This might be the result of formed HNE. In this
way an endproduct of lipid pero~dauon srimulates the process that generates this product.
Introduction.
The process of lipid peroxidation has been related to phenomena such as inflammation,
aging, cancer and toxicity of xenobiotics [1]. During lipid peroxidation polyunsaturated fatty
acids are peroxidised by a free radical mechanism. First a hydrogen atom is abstracted from a
polyunsaturated fatty acid (LH). The fatty acid radical (L') formed reacts with oxygen, and
subsequently the lipid peroxyl radical (LOO) abstracts a hydrogen atom from another
polyunsaturated fatty acid. The latter, converted into a fatty acid radical, then reacts with oxygen
and abstracts a hydrogen atom from a third polyunsaturated fatty acid. The lipid hydroperoxides
(LOOH)thus formed are thought to be highly reacrive. They can initiate another radical chain
reaction (LOOH-dependent lipid peroxidarion), as discussed in chapter 3. Apart from the lipid
hydroperoxides, other highly cytotoxic products are formed. 4-Hydroxy-2,3-trans-nonenal
(HNE)seems to be of major importance in this respect [2].
Fortunately, the living cell has defense mechanisms which can provide protection against the
devastating effects of lipid peroxidation. In the liver the most important components in these
defense mechanisms are glutathione(GSH)and vitamin E. Liver microsomes have been shown
to be protected against lipid peroxidation by a membrane bound factor [3]. As we have
postulated [4], this protection might be due to reduction of the vitamin E radicals by GSH. We
also found that a continuous radical stress, both in vitro and in vivo, diminishes this protecrive
GSH effect [4].
41

Here we report on a study, which was designed to investigate whether HNE might be
responsible for the reduced GSH-dependent protection after oxidative stress.
Materials and methods.
Chemicals. Glutathione(GSI~ and thiobarbituric acid(TBA)were obtained from Sigma. N-ethyl maleimide
(NE1Vn was obtained from Aldrich. 4-hydroxy-2,3-trans-nonenal(HrIE) was synthesized according to Leurs et al.
[5]. All other chemicals used were of reagent grade.
Preparation and pretreatment of microsomes. Male Wistar strain rats(TN.O.,Zeist, The Netherlands), 220250 g, were killed by decapitation. Liver microsomes were prepared as previously described [4]. Before use the
microsomes vs~ere thawed, diluted 5-fold with ice-cold buffer A(50 mM sodium phosphate, 0.1 mM EDTA pH
7.4) and, in order to remove potential protective cytosolic contamination (e.g. endogenous GSI~, subsequently
washed twice with buffer A by centrifugation (40 min, 115000 x g at 4 C). Then the pellet was resuspended in
buffer A and the microsomes (final concentration: microsomes derived from 1/g g liver in 1 ml) were
preincubated at 37 C with HNE (1 or 0.1 mIvn or NEM (0.1 mM)for 30 min. The preincubation was
terminated by the addition of 1 volume ice-cold buffer B (50 mM Tris-HCI, 150 mM NaCI, pH 7.4). The
microsomes were immediately centrifuged twice with buffer B (40 min, 115000 x g at 4 C)to wash out HNE or
NEM. The pellet was resuspended in buffer B.In order to stimulate lipid peroxidadon the pretreated microsomes
were incubated with 0.2 mM ascorbate and 10 M FeSOq with or without the addition of 1 mM GSH in buffer
B as described previously [4].
Measurements. Lipid peroxidaUon was measured with the thiobarbituric acid (TBA) assay, as previously
described [4], and expressed as the absorbance at 535 nm vs 600 nm (~A535-600 Non-protein thiol
determination was performed according to Ellman [6]. Protein determinations were made by the method of
Bradford [7], using bovine serum albumin as standard. Vitamin E (a-tocopherol) was assayed after tissue
extraction [8] by the HPLC method described by Driskell et al. [9]. Microsomal GSH transferace activity was
determined by the method of Morgenstem and Depiene[10] using 1-chloro-2,4-dinitrobenzene.

Results and discussion.


Glutathione(GSH)is known to protect liver microsomes against lipid peroxidation[11, 12].
There is increasing evidence [13, 14J that this protection proceeds via the mechanism we
proposed previously [4].
R

2 GSSG

vit. E

labile
factor

RH

vit.

GSH

Radicals in the membrane rapidly react with vitamin E. The thus formed vitamin E radicals
(a-chromanoxyl radicals) in the membrane are regenerated to vitamin E by cytosolic GSH. The
latter reaction, however, proceeds too slowly to give efficient protection against lipid
peroxidation, unless it is catalyzed by a membrane bound heat labile factor [4], which is
probably a protein [15].
In order to study the effect of GSH on lipid peroxidation, rat liver microsomes were
incubated with 10 mM iron (II) and 0.2 mM ascorbate, which resulted in a rapid and extensive
peroxidation of the membranes (iig. 1). Ten M Iron (III) alone does not stimulate lipid
42

~.z
~.i

0.9
0.8
x

0
0
e

0.7

0.6

Q
d

0.5

M
N

f
E

0.4
0.3

o.z
a~
0
U

LV

4V

INCUBATION TIME (MIN)

Fig. 1. Time course of lipid peroxidation(


)and GSH concentration(- - -)in control microsomes.
Microsomes were preincubated for 30 min at 37 C in phosphate buffer A without further additions and
washed twice with fris buffer B. Next 0.2 mM ascorbate () or 0.2 mM ascorbate in combination with 1
mM GSH (+) was added to the pretreated microsomes. (0) is the GSH concentration in the latter
incubation. The reactions were started in buffer B with the addition of Fel+,final concentration 10 M.
Data represent one example out of 5 duplicate experiments.

peroxidation nor does 0.2 mM ascorbate [4]. Ten M Iron (In, however,is capable of inducing
lipid peroxidarion [4]. During the process of lipid peroxidation iron (II) is oxidized to iron (III)
[16]. Since the combinarion of iron (III) and 0.2 mM ascorbate induces lipid peroxidarion and
0.2 mM ascorbate augments the lipid peroxidation induced by 10 mM iron (II)[4], the role of
ascorbate seems to be maintenance of iron in the reduced iron(In form [4, 17]
GSH does not srimulate iron (II) induced lipid peroxidation in heated microsomes [4]. This
indicates that GSH itself does not reduce iron (III) at a sufficiently high rate for the stimulation
of lipid peroxidation. The stimularion of lipid peroxidarion by GSH observed upon incubating
heated microsomes with 10 mM iron (II) and 0.2 mM ascorbate [4] can be explained by a GSHdependentregeneration of ascorbate, that becomes oxidized during lipid peroxidarion. Recently,
GSH has indeed been shown to reduce dehydroascorbate to ascorbate [19].
As seen in fig. 1, GSH protects control rat liver microsomes against 10 mM iron (II) and 0.2
mM ascorbate induced lipid peroxidarion for some rime. After 15 min lipid peroxidation
starts. During this lag time hardly any GSH is consumed (tig. 1). At first sight this is
inconsistent with the mechanism we propose because the vitamin E radicals become reduced at
the expense of GSH. As shown in table I, however,control liver microsomes contain 220 17
nmol vitamin E/ g protein. Since the incubarion system contains approximately 1 mg protein /
ml the concentration vitamin E versus GSH is 1 : SOQO (0.2 nmol vitamin E / ml vs 1 mmol
GSH / ml). If all the vitamin E in the membrane would be converted into a vitamin E radical
which is subsequently regenerated by GSH (one cycle) this would only result in a 0.02 %
reduction of the GSH concentration. Of course more cycles can cause a larger GSH
consumption. T'he total reducrion of the GSH content in this way is maxunally equal to the
43

pretreatment
control
HNE 1 mM
NEM 0.1 mM

nmol vitamin E/ g protein


220 17
213 20
209 11

Table I. Effect of different pretreatrnents on the vitamin E content of the liver microsomes. The
results are
expressed as mean S.D, n = 3.

amount of radicals formed. The small GSH consumption we observed might, to some extent,
be explained by the comperition of ascorbate with GSH for vitamin E radicals [18]. However
,
in a comparable lipid peroxidaiion inducing system, but without ascorbate, it was demonstr
ated
that the protecrive effect of GSH by the membrane-bound factor was also not accompanied with
a substanrial GSH consumption [15]. The GSH consumption that occurs once the process
of
lipid peroxidation has started i.e. after 15 min (fig. 1) is the result of the regeneration of
ascorbate by GSH [4, 19], and of the reaction of GSH with reactive products formed by
lipid
peroxidation, e.g. aldehydes. Within the time span studied, however, the GSH concentra
tion
does not fall under effective scavenge concentration, a GSH concentration of 0.1 mM still
has
an inhibitory effect [4].
Both in vivo and in vitro radical stress reduces the GSH-dependent protection [4]. Since
highly cytotoxic aldehydes, of which HNE is of major importance, are produced by radical
stress, we investigated whether the reduction in the GSH-dependent protection might be caused
by HNE.
The GSH-dependent protection against lipid peroxidation is reduced after 0.1 (fig. 2) or 1
mM (data not shown). HNE pretreatment of the liver microsomes followed by extensive
washing. HNE pretreatment by itself does not result in the production of TBA reactive material.
The HNE pretreatment neither has effect on the vitamin E content of the rnicrosomes (table I)
nor on the extent of the GSH consumption during the GSH-dependent protection (fig. 2). A
direct reacrion between GSH and HNE that is added during the preincubation is not observed
nor expected, since after preincubation of the microsomes with HNE, the excess of HNE
is
removed by extensive washing before the GSH addition. The onset of the GSH consumption is
advanced, apparently because lipid perolcidation starts earlier. The regenerarion of ascorbic acid
by GSH might explain, both part of the GSH consumption and the stimulation of lipid
peroxidation by GSH after HNE pretreatment[4, 19].
The main mechanism by which HNE is supposed to exert its general toxic effects is covalent
binding of HNE to free thiol groups of proteins [5]. To invesrigate whether this process is
also
involved in the reducrion of the GSH-dependent protection, we compared the effect of the
synthetic thiol inactivator N-ethyl maleimide(NEM)with that of HNE. Sunilar to HNE no lipid
peroxidation is increased by preincubating the microsomes with 0.1 mM NEM. NEM
pretreatment also diminishes the GSH-dependent protection (fig. 3) without influencing the
vitamin E concentration (table I) or the extent of the GSH consumption. Because lipid
peroxidation is not inhibited by GSH after NEM pretreatment, the onset of the GSH
consumption is advanced (iig. 3).
Both HNE and NEM were found to inhibit the membrane bound GSH-dependent protectio
n
without influencing the GSH and vitamin E concentrarion. We therefore conclude that the
44

1.3
1.2
1.1
1
0.9
0
0
~

0.8
x
0.7

0.6

0.5

0.4
0.3
0.2
0.1
0
U

LU

4V

OV

INCUBATION TIME (MIN)

Fig. 2. Time course of lipid peroxidation(


)and GSH concentration (- - -) in 0.1 mM HNE
pretreated microsomes. Microsomes were preincubated for 30 min at 37 C in phosphate buffer A with
0.1 mM HNE and washed twice with fris buffer B. The incubation systems were as described in fig. 1.
Data represent one example out of 4 duplicate experiments. Preincubation of the microsomes with 1
instead of 0.1 mM HNE resulted in a comparable reduction of the GSH-dependent protection (n=2).

decrease in GSH-dependent protection caused by HNE or NEM is the result of inacrivarion of


the heat labile factor, previously proposed [4]. Probably this factor contains a sulfhydryl moiety
that is irreversible blocked by HNE and NEM through covalent binding. This sequestration
annihilates the catalytic activity of the membrane bound factor, that reduces the vitamin E radical
to vitamin E in the membrane by GSH.
Apart from the GSH-dependent reduction of vitamin E radicals several other defense
mechanisms against lipid peroxidation have been described. Much attention has been given to
the cytosolic and microsomal bound GSH peroxidases [1]. During the process of lipid
peroxidation lipid hydroperoxides(LOOH)are formed. These peroxides can induce a secondary
radical chain reaction (chapter 3). Protection against this LOOH-dependent lipid peroxidation is
achieved by reducing these lipid peroxides to less reactive alcohols by a GSH peroxidase.
However, the cytosolic GSH peroxidases do not reduce the membrane bound lipid
hydroperoxides at a measurable rate [20], although Ursini et al.[21] reported about a cytosolic
GSH peroxidase exerting peroxidase activity for these peroxides. It has been stated that
inhibirion of microsomal lipid peroxidarion requires the consecutive action of a membrane
bound phospholipase A2, which releases fatty acyl hydroperoxides from peroxidized
phospholipids, and of the cytosolic GSH peroxidases, which reduce these peroxides [22, 23].
Besides the cytosolic GSH peroxidases, a microsomal membrane bound GSH transferase,
exerting peroxidase activity for organic peroxides has been described [10]. Since this
microsomal GSH transferase is located in the membrane, it might directly utilize membrane
bound lipid hydroperoxides and might therefore protect effecrively against lipid peroxidation.
Morgenstern and DePierre [10] suggested that the microsomal GSH-dependent heat labile factor
might be the membrane bound GSH transferase, that protects by its peroxidase activity. More
45

~.:
~.:
~.i
i
o.s
0
0
io

O.f

x
N
t7

O.i
N

O.f
i

0.
0.4
0.?
0.2
0.1
0
INCUBATION TIME (MIN)

)and GSH concentration (- - -) in 0.1 mM NEM


Fig. 3. Time course of lipid peroxidation(
pretreated microsomes. Microsomes were preincubated for 30 min at 37 C in phosphate buffer A with
0.1 mM NEM and washed twice with fris buffer B. The incubation systems were as described in fig. 1.
Data represent one example out of 3 duplicate experiments.

recently, Yonaha and Tampo [24] even suggested that the vitamin E radical produced by
scavenging lipid radicals during peroxidation is regenerated to vitamin E by the membrane
bound GSH transferase in the presence of GSH. As shown in table II, however, we found that
NEM pretreatment of the microsomes acrivates the microsomal GSH transferase. On the other
hand, as described above, NEM treatment reduces the GSH-dependent protecrion. Taken the
data together, we may conclude that the GSH-dependent protection in our model system does
not proceed via the membrane bound GSH transferase, but rather via another heat labile
enzyme. Moreover, in a comparable in vitro system it was demonstrated that hardly any
LOOH-dependent lipid peroxidation occurs [24]. Since the GSH peroxidases can only protect
against LOOH-dependent lipid peroxidation, no major effect of the peroxidases in the in vitro
system is to be effected.
We also invesrigated the contribution of the membrane bound phospholipase A2 to the GSHdependent protection against lipid peroxidation. It has been suggested that the protection of
cytosolic GSH peroxidass is dependent on the microsomal phospholipase A2[21, 22]. This
was demonstrated by the reduction of the cytosolic GSH transferase B-dependent protection

pretreatment
control
NEM 0.1 mM

GSH transferase activity


(nmol /min / mg protein)

87 1
204 10

liver
Table II. Effect of0.1 mM NEM pretreatment on the microsomal GSH vansferase activity of control
microsomes. The results are expressed as mean S.D, n = 3

46

1 .4
1.?
1.2
1.1

0.9
0
0
~o
N
M
~
Q
4

0.8
0.7
O.fi
0.5
0.4
0.3
0.2
0.1
0
U

LV

4V

INCUBATION TIME (MIN)

Fig. 4. Time course of lipid peroxidation in control microsomes. 0.2 mM Ascorbate (Q), 0.2 mM
ascorbate in combination with 1 mM GSH (+), 0.2 mM ascorbate in combination with 0.1 mM
mepacrine (0) or 0.2 mM ascorbate in combination with 1 mM GSH and 0.1 mM mepacrine (0) was
added to the control microsomes,suspended in fris buffer B. The reactions were started with the addirion of
Fel+,final concentration 10 M. Data represent one example out of 3 duplicate experiments.

against in vitro lipid peroxidation, by heating the microsomes or by treating the microsomes
with p-bromoacylphenylbromide, an irreversible phospholipase A2 inhibitor. The protecrion by
the GSH transferace B was stated to proceed via its peroxidase acrivity. However the methods
used to inhibit phospholipase A2 are not selective, e.g. p-bromoacylphenylbromide was shown
to inhibit a wide spectrum of enzymes by reacting with an essential sulfhydryl group [26], so
p-bromoacylphenylbromide might also react with th SH group on the heat labile factor. Of
course heating is also very unspecific, this also inacrivates the heat labile factor. We used 0.1
mM mepacrine to inhibit the microsomal phospholipase A2, although mepacrine also affects
lipid peroxidation (iig. 4). The reduction of lipid peroxidation by mepacrine itself might be
attributed to a small radical scavenging capacity of mepacrine, since after heating the
microsomes the same protection was observed (data not shown). As shown in fig. 4, mepacrine
does not reduce the GSH-dependent protecrion. We therefore conclude that the GSH-dependent
protecrion is not-dependent on a phospholipase A2. The protection by GSH transferase B
observed by Tan et al. [22] might, to some extent, be explained by the detoxication of HNE,
since GSH transferace B utilizes HNE as a substrate [27].
In vivo and in vitro lipid peroxidation has been shown to reduce the protection of liver
microsomes by GSH [4]. Devasagayam [28] demonstrated that different microsomal membrane
preparations of rat liver are not equally sensitive towards oxidative stress. During the
peroxidation of membranes susceptible to lipid peroxidation, HNE and other aldehydes are
formed. These reactive products of lipid peroxidation can diffuse to membranes that are more
resistant to radical stress, and there inactivate the membrane bound GSH-dependent factor. So,
in principle, HNE and probably also other aldehydes can stimulate the process of lipid
peroxidation, that eventually generates more aldehydes.
47

In summary, the reduction in GSH-dependent protection observed after radical stress [4]
might be due to the inactivation of the GSH-dependent protecting factor by HNE formed.
The
chain reaction of microsomal lipid peroxidarion might be initiated after destruction of
the
microsomal GSH-dependent heat labile factor. The data presented here support the notion [4]
that during lipid peroxidation, the progressive membrane damage cannot be stopped anymore
via the membrane bound GSH-dependent factor, neither by a direct nor by an indirect increase
in the GSH level. This is of importance in providing adequate measures against the hepatotox
ic
effect of many xenobiorics, once the process of lipid peroxidation has already started.
References.
1 Oxidative Stress(Bies, H,ed), Academic Press, New York/London (1985).
2 Esterbauer, H.,(1982) in Free Radicals, Lipid Peroxidation and Cancer (McBrien, D.H.0 and Slater,
T.F.,
eds), Academic Press, New York/L,ondon, pp. 101- 128.
3 Bast, A and Haenen, G.R.M.M.(1984) Trends Biochem. Sci. 9,510 - 513.
4. Haenen, G.R.M.M, and Bast, A.(1983)FEBS Lett. 159, 24 - 28.
5 Leurs, R., Rademaker, B., Kramer, K., Timmerman, H. and Bast, A.(1986) Chem.-Biol. Interact. 59, 211
218.
6 Ellman, G.L.(1959) Arch. Biochem. Biophys. 82, 70 - 77.
7 Bradford, M.M.(1976) Anal. Biochem. 72,246 - 254.
8 Rammell, C.G., Cunliffe, B. and Kieboom, A.J.(1983) J. Liquid Chromatogr. 6, 1123 -1130.
9 Driskell, W.J., Nese, J.W., Bryant, C.C. and Bashok, M.M.(1983) J. Chromatogr. 231, 439 - 444.
10 Morgenstem,R. and DePierre, J.W.(1983), Eur. J. Biochem. 134, 591 - 597.
11 Christopherson, B.O.(1968), Biochem J. 106, 512 - 522.
12 Burk, R.F.(1983) Biochem. Pharmacol. 31,601 -602.
13 McCay,P.B., Lai, E.K. Powell, S.R. and Breuggemann, G.(1986) Fed. Proc. FASEB 45, 451.
14 McCay,P.B.(1986) Prot. third Biennial Meeting of the Society for Free Radical Research, Dusseldorf, p.
29.
15 Burk, R.F.(1983) Biochim. Biophys. Acta 757, 21 - 28.
16 Rao, G.H.R., Genazd, J.M., Eaton, J.W. and White (1978)Prostaglandins Med. 1, 55 - 70.
17 Kornbrust, D.J. and Mavis, R.D.(1980) Mol. Pharmacol. 28, 2051 - 2055.
18 Forni, L.G. and Willson, R.L.(1983) in Protective Agents in Cancer (McBrien, D.H.0 and Slater, T.F.,
eds), academic Press, New York/London, pp. 159 - 173.
19 Bast, A, Naenen, G.R.M.M. and Savenije-Chapel E.M.(1987), Arzneim. Forsch. in press.
20 McCay, P.B., Gibson, D.D., Fong, K.-L. and Hombrook, K.R.(1976) Biochim. Biophys. Acta 431, 459 468.
21 Ursini, F., Maiorino, M., Valente, M.,Ferri, L. and Gregolin, C.(1982) Biochim. Biophys. Acta 710, 197
211.
22 Tan, K.H., Meyer, D.J., Belin, J. and Ketterer, B.(1984) Biochem. J. 220,243 - 252.
23 Sevanian, A. Stein, R.A. and Mead,J.F.(1981) Lipids 16, 781 - 789.
24 Yonaha, M. and Tampo, Y.(1986) Chem.Pharm. Bull. 34,4195 - 4201.
25 Davis, H.W., Suzuki, T. and Schenkman, J.B.(1987) Arch. Biochem. Biophys. 252,218 - 228.
26 Kyger, E.M. and Franson, R.C.(1984) Biochem. Biophys. Acta 794, 96 - 103.
27 Alin,P, Danielson, H. and Mannervik, B.(1985)FEBS Lett. 219,267 - 270.
28 Devasagayam, T.P.A.(1986) FEBS left. 199, 203-207.

48

- Chapter 6 Is phospholipase A2 involved in the glutathione-dependent


protection against in vitro microsomal lipid peroxidation?
Introduction.
During the process of lipid peroxidation, membrane lipids are converted into lipid
hycroperoxides(LOOH) by a free radical mechanism. These LOOH are thought to be highly
reactive. They can initiate another radical chain (LOOH-dependent lipid peroxidation) [1].
Glutathione(G5H)pt~otects against lipid peroxidation via a microsomal free radical reductase
[2]. Apart from the protection by this membrane-bound factor, it has been shown that GSH also
protects against lipid peroxidation via a cytosolic factor [1]. Ever since it was discovered that
LOOH are converted to LOH by cytosolic GSH-peroxidase (GSH-px), it was suggested that
GSH-px protects against lipid peroxidation via this mechanism. Indeed, initially it was found
that partially purified GSH-px protected against in vitro peroxidation of rat liver microsomes
that are heated (in order to inactivate the membrane-bound free radical reductase). However,
further purificarion of the GSH-px demonstrated that the observed protection was the result of
contamination of the partially purified GSH-px by another cytosolic factor [1]. Recently,
Gibson et al. [3] isolated this GSH-dependent protective factor and demonsixated that this
cytosolic factor was not identical to one of the GSH-px or GSH-transferaces (GSH-tr).
Moreover, liver cytosol does not contain enzymes that reduce LOON incorpited in
phospholipids at a measurable rate. However, it should be noted that a GSH-px that uses these
incorporated peroxides as substrate has been characterized [1]. Since phospholipase A2(PLAS
releases LOOH from peroxidized phospholipids, and since it is known that cytosolic GSH-px
can reduce LOOH thus released [1], it was tempting to suggest that a cooperation between a
membrane-bound PLA2 and cytosolic GSH-px is needed for the protection against lipid
peroxidation. PLA2 is proposed firstly to release the esterified LOOH - PLA2 preferentially
removes peroxidized fatty acids -and subsequently the GSH-px are thought to detoxify the
released LOON. Tan et al. [4] stated to have produced experimental evidence for the latter
mechanism, because they found a reducrion of the GSH-dependent protection by GSH-tr 1-2 by
inacrivating PLA2, either by the alkylating PLA2 inhibitor p-bromophenacylbromide(BPB)or
by heating the microsomes. In this study the involvement ofPLA2in the glutathione-dependent
protection against lipid peroxidation was re-examined.
Materials and methods.
Reduced glutathione (GSH), thiobarbituric acid, mepacrine and p-bromophenacylbromide (BPB) were
obtained from Sigma (St. Louis, USA). Liver microsomes were prepared from untreated male Wistar rats (200250 g, Harlan Olec C.P.B., Zeist, The Netherlands) as previously described [2]. BPB treatment of the
microsomes(1 mM BPB during 15 min at 37C) was stopped with the addition of an equimolar amount of GSH.
In order [o remove the excess of reagent these microsomes were washed as previously described [5]. Lipid
peroxidation was induced with the combination of 10 M FeSO4 and 0.2 mM ascorbate, it was measured with
the thiobarbituric acid method and expressed as the absorbance at 535 nm vs 600 nm (D535-6001 ~2~ GSH-tr
was isolated according to Zoetemelk [6]. The final amount of GSH-llr 1-2 in the incubations had a GSH-tr
activity towazd 1-chloro-2,4-dinitro-benzene of 3.8 mol / ml /min.

49

/g

~~

,o- - - -o- -

_ --

_~

1.0
~

e/

o.s P%
o

///

~ ~ '_"

~~~_~~,,
~~/.~
~~

lU

2U

3U

4U

5~

Time (min)
Fig. 1. Time course of lipid peroxidation in control microsomes induced by 10 M Fel+ and 0.2 mM ascorbate.
The additions made were: none(0); GSH-tr 1-2(4); 1 mM GSH(~); GSH-tr 1-2 and 1 mM GSH (1). In the
curves depicted with a dotted line 0.1 mM Mepacrine was added.

Results and discussion.


In the present study, it was found that GSH protected against 10 M Fee+ and 0.2 mM
ascorbate induced lipid peroxidation in control rat liver microsomes (fig 1), probably via the
membrane-bound free radical reductase [2]. The potenriation of the GSH-dependent protection
by GSH-tr 1-2, previously observed by Tan et al. [4], was confirmed (fig. 1). However, in the
present study it was shown that methods used by Tan et al. [4] to inhibit PLA2, i.e. by heatpretreatment or by BPB pretreatment of the microsomes, were rather unspecific. By both
methods the GSH-dependent protecrion by the membrane-bound free radical reductase is
inhibited (fig. 2, 3). Since it is known that the free radical reductase contains an essential and
vulnerable thiol group [5], the inacrivation by BPB might be a result of inactivation of this
group. The reactivity of BPB towards thiol groups was also acknowledged by Tan et al. [4],
since the BPB treatment was terminated by the addition of GSH. However the effect of BPB
treatment on the GSH-dependent protection via the microsomal free radical reductase was not
determined by Tan et al. [4]. Therefore, the reducrion of the GSH-dependent protection by BPB
or by heating, as previously reported by Tan et al. [4], is most likely the result of inactivation of
the free radical reductase rather than a selective inhibition of the membrane-bound PLA2. In
support of this statement, we found that mepacrine, another PLA2 inhibitor, did not reduce the
GSH-dependent protection, nor via free radical reductase nor via GSH-tr 1-2 (fig 1). This
indicates that PLAZ is not involved in the protecrion by GSH-tr 1-2.
The mechanism by which GSH-tr 1-2 potentiates the GSH-dependent protection is not
elucidated by the results presented in this study. Although we used a highly purified GSH-tr,
contaminarion of our enzyme preparation by other cytosolic factors, e.g. the factor described by
Gibson et al. [3], might be a possible explanation. However, heating the microsomes abolished
50

.o
/.~.
~~.
~~~
a
~
~~ ~

-~ -o

/ ~
.S
M

,%~
//

t
V

~0
Time (min)

Fig. 2. Time course of lipid peroxidation in BPBpretreated microsomes induced by 10 M Fel+


and 0.2 mM ascorbate. Additions made were:
none () or 1 mM GSH(~).

10

30
20
Time (min)

40

Si

Fig. 3. Time course of lipid peroxidation in heated


microsomes induced by 10 M Fel+ and 0.2 mM
ascorbate. The additions made were: none(p);
GSH-tr 1-2 and 1 mM GSH (1), in the absence
or presence (- - -) of 0.1 mM Mepacrine.
()

the protection by our enzyme preparation (fig 3), while the factor described by Gibson et al.[3]
Brill protected against lipid peroxidation in heated microsomes. Also after hearing of the isolated
GSH-tr 1-2, the protecrion against lipid peroxidation in control microsomes by the enzyme
prepararion was lost (data not shown). Apparently a cooperation between a cytosolic and
membrane-bound factor is involved in the observed protection.
It could have been anticipated that no effective protection against lipid peroxidation was
observed by the GSH-px acrivity of GSH-tr 1-2 in our in vitro model system. The assumption
that the protection proceeds via the detoxication of LOOH,implies that these LOOH play a
pivotal role in process of lipid peroxidarion. However, lipid peroxidation is induced by an
extensive radical stress that is artificially forced upon the membranes. The observed lipid
peroxidation is probably mainly a result of LOON-independent lipid peroxidation. In a
comparable in vitro system it was demonstrated that hardly any LOOH-dependent lipid
peroxidation occurs [5]. Since GSH-px can only protect against LOOH-dependent lipid
peroxidation, no major protective effect is to be expected in an in vitro system were mainly
LOOH-independent lipid peroxidation occurs.
In liver microsomes, tt~ majority of the polyunsaturated fatty acids is bound on the sn-2
position of the phospholipids, although a significant amount is also located e.g. on the sn-1
position. Since the process of lipid peroxidation probably cannot discriminate between
polyunsaturated fatty acids on the sn-1 or sn-2 position, lipids on both posirions will be
peroxidized during oxidative stress. The protective mechanism described by Tan et al. [4] is
focused on the PLA2-mediated removal of LOOH from the sn-2 position of phospholipids. In
principle, the release of LOOH from e.g. the sn-1 position has to be considered too. Actually, in
liver microsomes PLA1 activity is much more prominent than PLA2 activity, but also PLAT
activity is, relarive to the activity of the GSH-tr needed, extremely low. If the LOOH would be ,
as reactive as is implicitly suggested by Tan et al. [4], than PLA activity would be much to low

51

to provide any measurable protecrion.


In conclusion, our results demonstrate that the reported inhibition of the GSH-dependent
protection by the allcylating PLA2 inhibitor BPB or by heating the microsomes is probably
caused by the inactivation of the free radical reductase and not by inactivation of PLA2.
Additionally it was found that mepacrine, another more selective PLAZ inhibitor, did not reduce
the GSH-dependent protecrion by purified GSH-tr 1-2. Moreover, during the process of lipid
peroxidarion also fatty acids located on other places as the sn-2 posirion are peroxidized, and
these lipid hydroperoxides cannot be removed by PLA2. Also, in rat liver microsomes PLA2
activity is probably much to low to remove the peroxidized fatty acids from the phospholipids
efficiently. Finally, in vitro lipid peroxidation probably consists mainly of LOOH-independent
lipid peroxidation, and no protectors can be provided against this type of lipid peroxidarion via
the detoxication of LOOH.Based on these arguments, it is concluded that PLAZ is probably not
involved in the GSH-dependent protecrion against in vitro microsomal lipid peroxidarion.
References.
1 A. Bast and G.R.M.M. Naenen, Glutathione and cytochrome P-450; what is the significance of their
interrelationship? Trends Biochem. Sci. 9(1984)510 - 513 and references herein.
2 G.R.M.M. Naenen and A. Bast, Protection against lipid peroxidation by a microsomal glutathione-dependent
labile factor, FEBS Lett. 159(1983)24 - 28.
3 D.D. Gibson, J. Hawrylko and P.B. McCay, GSH-dependent inhibition of lipid peroxidation: Properties of a
potent cytosolic system which protects cell membranes,Lipids 20(1985)704-711.
4 K.H. Tan, D.J. Meyer, J. Berlin and B. Ketterer, Inhibition of microsomal lipid peroxidation by glutathione
and glutathione transferases B and AA,Biochem. J. 220(1984)243 - 252.
5 G.R.M.M. Naenen, J.N.L. Tai Tin Tsoi, N.P.E. Vermeulen, H. Timmerman and A. Bast, 4-Hydroxy-2,3trans-nonenal stimulates lipid peroxidation by reducing the glutathione-dependent protection, Arch. Biochem.
Biophys.259(1987)449 - 456 and references herein.
6 C. Zoetemelk, Glutathione conjugation and bioactivation of 1-2 dibromo compounds, Ph.D.-thesis, Leiden,
1987.

~~

- Chapter 7 Effect of thiols on lipid peroxidation in rat liver microsomes.


Abstract. The stimulatory or inhibitory effects of various thiol compounds on in vitro lipid
peroxidation by iron-ascorbate in rat liver microsomes were determined. Glutathione had no
measurable pro-o~dant capacity, in contrast, it protected against lipid peroxidarion. N-acetyl lcysteine and ~-meshy:-glutathione hau na effect on in vitro lipid peroxidarion.:-Cysteine
stimulated lipid peroxidarion and also of d-penicillamine and dl-dithiothreitol the pro-oxidant
capacity predominated the anti-oxidant capacity. Cysteamine afforded a pronounced protection
against in vitro lipid peroxidation. In contrast to the labile character of the glutathionedependent protection, the protecrion by cysteamine was not affected by heat-pretreatment of the
liver microsomes or alkylating protein sulfhydryl groups by N-ethyl maleimide. Again in
contrast to glutathione, the protection against in vitro microsomal lipid peroxidation by
cysteamine was not reduced after in vivo lipid peroxidation induced by CC14. This suggests
that even after the process of lipid peroxidation has been started, administration of cysteamine
might still be beneficial.
Introduction.
T'he peroxidation of membrane lipids has been associated with a wide variety of toxicological
phenomena [1]. Fortunately an array of defense mechanisms is effective in the prevenrion of
lipid peroxidation [1]. Amongst these, glutathione (GSH)-dependent processes are known to be
of major importance in the liver [2,3]. The protection which is associated with liver microsomes
is probably due to the reduction of vitamin E radicals at the expense of GSH, a reaction
catalyzed by a microsomal free radical reductase [2,3]. There is evidence that this reductase
contains an essential and vulnerable SH moiety [4].
For quite some time, low molecular weight thiols have been applied to diminish the damage
provoked by oxidative stress [5,6]. The protective effect of thiols can be brought about directly,
e.g. by scavenging radicals, or indirectly, by elevating GSH levels. However, these
compounds can also act as pro-oxidants, e.g, by initiating lipid peroxidarion directly [7,8], or
indirectly, by inhibiting endogenous defense mechanisms against lipid peroxidarion [9]. In the
present study the direct pro- and anti-oxidant effects of various thiols on microsomal lipid
peroxidation were deternuned. Special attenrion was paid to the question whether other thiols
than GSH can protect against lipid peroxidation via the free radical reductase. All thiols studied
have been used in clinical situations as antidotes, and -except dl-dithiothreitol -are structurally
closely related to GSH (fig 1). S-methyl-glutathione was used to determine the role of the free
SH-group in GSH.
Materials and methods.
Chemicals. Glutathione (GSH),1-cysteine, cysteamine, ascorbic acid and FeSOq were obtained from Merck,
Darmstad, F.R.G. dl-dithiothreitol and S-methyl gluta[hione were from Sigma, St-Louis, USA. d-penicillamine
was a gift of Eli Lilly (CTtrecht, the Netherlands) and N-acetyl l-cysteine was a gift of Bristol-Myers BV (Bussum,
the Netherlands). All other chemicals were of analytical grade purity.
Preparation and incubation of microsomes. Uneated male Wistar rats (Harlan Olec C.P.B., Zeist, The

53

~2
O
H-CHZ-CHi

r+Ftc~rrH-cHZox GLUTATHIONE

(.bOH

CHZ
SH
O

cH,--rni-cH-c-ox
~
~H1

N-ACETYL
L-CYST'EINE

SH
O

nazcxoH

L-CYSTEINE

~Hz
SH

NHz-C~Z

CYSTEAMINE

~HZ
SH
O
M-I2-CH-C-OH

D-PENICILLAMINE

CHy-i-CHI
SH

HO

OH

HC-CH
Hz
SH

DL-DITHIOTREITOL

`
iHz
SH

Fig. 1. Thiols examined in this study.


Netherlands),220-250 g, were killed by decapitation. Liver microsomes were prepared
as previously described [2],
and stored at -80 C. Before use the microsomes were thawed, diluted 5-fold with
ice-cold Tris buffer (50 mM
Tris-HCI, 150 mM NaCI, pH 7.4) and, in order to remove potential protective
cytosolic contamination (e.g.
endogenous GSI~, subsequently washed twice with the Tris buffer by centrifugation
(40 min, 115000 x g at 4
C). Finally, the pellet was resuspended in the Tris buffer and the microsomes (final
concentration 1 - 1.5 mg
microsomal protein / ml) were incubated at 37 C.In vitro N-ethyl maleimid
e pretreated microsomes were
prepared as in [4]. Heat-pretreatment was for 60 seconds at 100 C.In vivo, lipid peroxidat
ion was induced by
CC14 as in [2]. In vitro, lipid peroxidaon was induced by Fee+ or a combination
of vitamin C and Fee+. All
in vitro reactions were started with the addition of iron.
Measurements. Lipid peroxidation was measured with the thiobarbituric acid assay,
as previously described
[2], and expressed as the absorbance at 535 nm vs 600 nm
(D535-6001 The results shown in fg 2-4 aze
typical examples of at least two duplicate experiments. Protein determinations were
made according to the method
of Lowry et al. [13], using bovine serum albumin as standard.

Results.
The aim of this study was to deternune the direct stimulatory or inhibitory effect of
various
thiols, shown in iig 1, on lipid peroxidation in rat liver microsomes. Firstly, the
pro-oxidant
capacities of the compounds were studied. It has been suggested that lipid peroxidation
initiated
by thiols proceeds via direct reduction of metal ions like iron, present in the medium
[7]. Fel+
can iniriate lipid peroxidarion [10]. During the initiation iron becomes oxidized to Fe3+
[11]. For
a measurable extent of lipid peroxidation Fe3+has to be reduced to Fee+[10]. We deternu
ned
the pro-oxidant capacities in an incubation system containing the thiol, 10 M Fel+
and liver
54

1.5

1.0

1.0
vi

~ 0.5

1'
15

O.$

a
~.~

10

20

30

40

50

~.~

1.5

i.o C

10

20

30

40

50

8~

2
l6

1.0

8
v~

~ 0.5
~.$

1 6

21
IJ

0.0t
0

]0

20

30

40

Incubation time(min)

~ 0.0f
0
SO

10

20

30

40

50

Incubation time(min)

Fig. 2. Time course of lipid peroxidation in control microsomes (panels A and B)or in heat-pretreated
microsomes (panels C and D)induced by 10 M Fel+(Panels A and C)or by the combination of 10 M
Fel+ and 0.2 mM ascorbate (Panels B and D). The additions made were: None (1); 1 mM GSH (2); 0.5
mM N-acetyl l-cysteine (8); or 1-cysteine in a concentration of 0.05 mM (3); 0.1 mM (4); 0.2 mM (5);
0.5 mM (6); or 1 mM (7). All reactions were started with the addition of 10 M Fel+. Addition of Nacetyl l-cysteine had no effect on the time course of lipid peroxidation induced by 10 M Fel+ in control
or heat-pretreated microsomes and these time courses aze for the shake of clarity not shown.

tnicrosomes (panels A and C of iig. 2-4). Addition of 0.2 mM ascorbate to the system markedly
increased lipid peroxidation (panels B and D of fig. 2-4). It has been shown that the incubarion
system containing 0.2 mM ascorbate and 10 M Fee+ is an optimal in vitro model system for
screening potenrial scavengers against lipid peroxidarion in rat liver microsomes [12]. Heatpretreated microsomes were used.in order to detemune if enzymatic components participate in
the effects of the compounds(panels C and D of fig. 2-4).
Pro-oxidant acrivity.
The pro-oxidant properties of the thiols on lipid peroxidarion in rat liver microsomes were
deternuned in the incubation system containing 10 M Fel+. As shown in fig 2A, 1 mM GSH
protected against lipid peroxidation induced by 10 M Fee+ in rat liver microsomes. The fact
that no pro-oxidant capacity of GSH was observed might be a result of the anti-oxidant capacity
of GSH, which could overshadow its pro-oxidant capacity. But also after destroying the
enzyme-dependent anti-oxidant mechanism of GSH (by heating the microsomes for 60
seconds), no pro-oxidant activity of GSH could be observed (fig 2~).
In contrast to GSH, addition of 1-cysteine in concentrations up to 0.5 mM (final
55

z.o

A
~.5

~:
1.5

3/

13=

1.0
1.0
m
h

0.5

0.5

~~-

1 3.

1~

0.0
U

10

20

30

40

50

00
0

10

20

30

4fJ

50

10

20

30

40

50

t.0

~.5 C
1.5

3~
1.0

1.0
~

2~
0.$

1 23

~.$
I~
0~

~~

]0

20

30

40

Incubation time(min)

50

Incubation time (min)

Fig. 3. Time course of lipid peroxidation in control microsomes (panels A and B) or in heat-pretreated
microsomes (panels C and D)induced by 10 .M Fel+ (Panels A and C)or by the combination of 10 M
Fel+ and 0.2 mM Ascorbate (Panels B and D). The additions made were: None (1); 0.5 mM dldithiothreitol (2); 0.5 mM d-penicillamine (3). All reactions were started with the addition of 10 M
Fel+.

concentration) drastically increased lipid peroxidation induced in control microsomes by 10 M


Fel+ (fig. 2A). Ata 1-cysteine concentration of 1 mM,however,the extent of lipid peroxidarion
was reduced, when compared to the incubarion system containing 0.5 mM 1-cysteine (fig. 2A).
Heating the microsomes had no significant effect on the pro-oxidant activity of 1-cysteine (iig.
2C~.
dl-Dithiothreitol and d-penicillamine, in a final concentration of 0.5 mM,both acted as prooxidants since they stimulated Fel+induced lipid peroxidarion (fig. 3A), though dl-dithiothreitol
was less effective in this regard. Cysteamine, a well known radioprotective agent, did not
stimulate Fel+induced lipid peroxidation in control nor in heat-pretreated microsomes (fig.
4A,4~). Also N-acetyl 1-cysteine (fig. 2A,2~) and S-methyl glutathione (fig. 4A,4~) had no
effect on Fel+induced lipid peroxidation either in control or in heat-pretreated microsomes.
Anti-oxidant acrivity.
The anti-oxidant properties were studied in the in vitro system in which lipid peroxidation
was induced in liver microsomes by 10 M Fee+ in the presence of 0.2 mM ascorbate. As
shown in fig 2B, GSH delayed but did not prevent the occurrence of lipid peroxidation. After a

56

1.5

i.o A

15~

1.0

z-3~

0.5
0.5

a-

13~

00

0.0
10

20

30

40

50

i.s

i.o C
S

10

20

30

40

50

15~

i.o
z--

~ 0.5

3~

0.5

a--

13~

0.0
0

10

20

30

40

50

Incubation time (min)

00
0

10

20

30

40

50

Incubation time (min)

Fig. 4. Time course of lipid peroxidation in control microsomes (panels A and B)or in heat-pretreated
microsomes (panels C and D)induced by 10 M Fel+ (Panels A and C)or by the combination of 10 M
Fel+ and 0.2 mM Ascorbate (Panels B and D). The additions made were: None (1); 1 mM S-methyl
glutathione (5); or cysteamne in a concentration of 0.5 mM (2); 1 mM (3); or 1.5 mM (4). All reactions
were started with [he addition of 10 M Fel+. Addition of S-methyl glutathione had no effect on the time
course of lipid peroxidation induced by 10 M Fel+ in control or heat-pretreated microsomes and these
time courses are for the shake of clarity not shown.

lag time lipid peroxidation started, the final amount of thiobarbituric acid reactive material
produced equalled the amount in the control incubation system without GSH. After heatpretreatment of the microsomes, addirion of GSH offered no protection anymore against Fel+ /
ascorbate induced lipid peroxidarion (fig 2D).
In contrast to its pro-oxidant activity in the Fel+ incubation system, addition of 1-cysteine
resulted in aconcentration-dependent protection against Fee+ / asco bate induced lipid
peroxidation (fig. 2B). Heating the microsomes reduced the lag time produced by 1-cysteine on
Fel+ / ascorbate induced lipid peroxidarion (fig. 2D).
Conversion of pro-oxidant activity in the Fel+ induced lipid peroxidation into anti-oxidant
activity in the Fee+ / ascorbate induced lipid peroxidation was also observed for dl-dithiothreitol
and d-penicillamine. Addition of 0.5 mM dl-dithiothreitol or d-penicillamine protected the
microsomes against Fel+ / ascorbate induced lipid peroxidation (fig. 3B). This protecrion had a
transient character, since after a lag time lipid peroxidation started. In the incubation systems
containing dl-dithiothreitol the extent of lipid peroxidation after 45 minutes was higher than in
the control incubarion system. Heating the microsomes reduced the anti-o~dant properties of dpenicillamine and dl-dithiothreitol (fig 3D).
57

100

~ ~

~ o

75

~ U
~~ ~
JG
a~

b ~
a~~

25

2.5

10

15

20

30

45

Incubation time(min)

Fig. 5. Protection against lipid peroxidakion in liver microsomes hom oil pretreated rats (open bars) or
CCIq. pretreated rats (hatched bars) by 0.5 mM cysteamine. Lipid peroxidation was induced by 0.2 mM
ascorbate and 10 M Fel+. The protection is expressed as the percentage production ( S.D.) of
thobarbituric acid material by the addition of 0.5 mM cysteamine, compared to the control incubation
system without cysteamine(n=3).

Apart from GSH,of the thiols studied cysteamine displayed the most pronounced protection
against Fel+ / ascorbate induced lipid peroxidtion and was therefore elaborated further. The
time course of this protection differed from the protection afforded by GSH. Addirion of
cysteamine diminished both the rate at which lipid peroxidation developed and the maximal
extent of lipid peroxidation reached (fig. 4B), whereas addition of GSH delayed lipid
peroxidation while the final extent of lipid peroxidation remained the same (fig. 2B). Heatpretreatment of the microsomes (fig. 2D) or pretreating the microsomes with the thiol alkylator
N-ethyl maleimide [4] destroyed the GSH-dependent protective effect, whereas neither heatpretreatment (iig. 4D) nor N-ethyl maleimide pretreatment (not shown) had any effect on the
protecrion afforded by cysteamine. Also in vivo lipid peroxidarion induced by CC14 had no
effect on the cysteamine-dependent protecrion (fig. 5), whereas the GSH-dependent protection
was found to be reduced [2].
N-acetyl l-cysteine slightly srimulated Fel+ / ascorbate-induced lipid peroxidation (fig. 2B,
2~) whereas S-methyl glutathione had no effect(fig. 4B,4D).
Discussion.
In the present study the direct pro- and anti-oxidant properties of various thiols on lipid
peroxidation in rat liver microsomes were determined. Two different in vitro incubation
systems were used. In the fust incubarion system lipid peroxidation was induced by Fel+. In
this system the pro-oxidant activity of the testcompounds was determined. This activity
probably proceeds via the regeneration of Fel+ that is oxidized to Fe3+ during the iniriarion of

58

lipid peroxidation. The anti-oxidant acrivity of the compounds was tested in an incubation
system in which lipid peroxidarion was initiated by the combination of Fel+ and ascorbate. In
order to determine if an enzymatic component is involved in the pro- or anti-oxidant acriviries of
the various thiols, heat-pretreated microsomes were used. .
It was found that GSH protects rat liver microsomes against lipid peroxidarion, i.e. t delays
the lipid peroxidation process for some time (this study, [2]). GSH had no predominant prooxidant capacity, indicating that GSH itself is incapable of reducing trivalent iron at a
measurable rate (cf[14]). Nevertheless pro-oxidant properties of GSH have been reported, but
they have been attributed to metal contaminations in the GSH samples [15], to the enzymatic
breakdown of GSH into more reactive thiols [16] or to an indirect reduction of Fe3+ to Fel+ by
GSH via the regeneration of ascorbate [17].
The protection which is afforded by GSH is supposed to be a result of the reduction of
vitamin E radicals by GSH, a reaction catalyzed by a free radical reductase [2-4]. Heatpretreatment of the microsomes (this study,[2]) or pretreatment of the microsomes with the
thiol alkylator N-ethyl maleimide [4], abolished the GSH protection,demonstrating that the
protection proceeds via a heat labile factor, notably the free radical reductase, which contains at
least one essential SH group. Also radical stress, generated either in vivo or in vitro, reduced
the GSH-dependent protection [2]. A possible explanarion for this reduction is that toxic
aldehydes produced during lipid peroxidation inacrivate the free radical reductase [4]. 4Hydroxy-2,3-trans-nonenal, which is of major importance in this respect, was shown to be
able to inactivate this type of GSH-dependent protection [4].
Besides GSH,several other low moleculaz weight thiols have been shown to be able to play
a protecrive role against oxidarive damage [5,6]. In the present study the direct pro- and antioxidant activities of various sulfhydryl containing compounds that are structurally related to
GSH were compared.
It was found that N-acetyl l-cysteine and 1-cysteine did not protect against Fel+ induced lipid
peroxidation in liver microsomes. This indicates that the protection of these compounds against
lipid peroxidation in vivo via the free radical reductase can only be achieved via an increase of
the GSH concentration, since they are precursors of GSH. N-acetyl l-cysteine srimulated Fel}/
ascorbate-induced lipid peroxidation, whereas it had no effect on Fel+ induced lipid
peroxidarion. The pro-oxidant efect of N-acetyl l-cysteine is similair to the pro-o}cidant effect of
sodium 2-mercaptoethanesulfonate [17] and probably the result of th indirect reducrion of Fe3+
to Fel+via the regenerarion of ascorbate. l-Cysteine acts as a direct pro-oxidant, because it was
found that 1-cysteine srimulated Fel+induced lipid peroxidation. This might be the result of the
generarion of toxic radicals during the autoxidarion of 1-cysteine, which is catalysed by iron
[18,19]. It has been suggested that the thiyl free radical of 1-cysteine is responsible for the
mutagenity of 1-cysteine in the Ames test [20]. Tien et al. [7] demonstrated, however, that 1cysteine does not initiate lipid peroxidation via thiyl radicals generated during autoxidation, but
more likely via a direct reduction of Fe3+ to Fee+.
The stimulation of Fel+ induced microsomal lipid peroxidation by 1-cysteine was maximal at
a 1-cysteine concentration of 0.5 mM. Above this concentration the pro-oxidant activity of 1cystreine was reduced. In liposomes maximal pro-oxidant activity has been observed at a
concentrarion of 0.25 mM [7]. This phenomenon is comparable to the dual role of ascorbate in
lipid peroxidation. In a low concentration, ascorbate is apro-oxidant, whereas higher
concentrations of ascorbate protect against lipid peroxidation [12]. It has been stated that the
59

ratio of Fel+ / Fe3+ is the determining factor for the iniriation of lipid peroxidarion reacrions [21,
22]. High concentrations of ascorbate might inhibit lipid peroxidarion due to a reductive
maintenance of iron exclusively in the Fel+ form, thus preventing the Fee+ to Fe3+conversion
which is needed for inducing lipid peroxidation [12, 21].
In microsomal lipid peroxidation induced by Fee+ / ascorbate, l-cysteine already displayed
anti-oxidant activity in a concentration of 0.2 mM, since a lag time in the onset of lipid
peroxidation was observed (fig. 2). Probably the effects of 1-cysteine and ascorbate on lipid
peroxidation are addirive. So the anti- and pro-oxidant properties of 1-cysteine and ascorbate are
cumulative, although the concentrarion where the pro-oxidant activity is converted into the antioxidant capacity differs; 0.25 mM for ascorbate [12] and 0.5 mM for 1-cysteine. Probably
ascorbate is more potent in reducing iron. Our observations are in agreement with the results of
Searle and Willson [8] who reported that the combination of iron and 1-cysteine srimulated lipid
peroxidation in rat liver microsomes via a nonenzymaric reaction, however,they did not observe
any anti-oxidant property of 1-cysteine. Seaz et al. [18J stated that, due to its pro-oxidant
acrivity, l-cysteine is of little use as a therapeutic agent.
The effects of dl-dithiothreitol and d-penicillamne were comparable to those of 1-cysteine.
Both compounds srimulated Fel+ induced lipid peroxidation, whereas they produced a lag time
in Fel+ / ascorbate induced lipid peroxidation. During the autoxidation of dl-dithiothreitol
superoxide radicals are formed [23, 24], however, at neutral pH -also used in the present study
- the radical production is very low [23]. Moreover, Tien et al. [7] have demonstrated that the
initiation of lipid peroxidation by dl-dithiothreitol does not occur via superoxide formarion but
more likely via a direct reducrion of Fe3+ to Fel+.
Cysteamine protected against Fel+/ascorbate induced lipid peroxidation. Both, the rate of
lipid peroxidation and the final extent of lipid peroxidation were reduced by cysteamine. This
differs from the protecrion provided by GSH. GSH caused a lag rime in the time course of lipid
peroxidation whereas the final extent of lipid peroxidation did not decrease. A possible
explanarion is that - in contrast to the protection by GSH -the protecrion by cysteamine is
brought about by scavenging radicals and not via the free radical reductase. Cysteamine is a well
known radioprotector [25]. It has been stated that the radioprotective action of cysteamine is
probably neither the result of the replacement of endogenous sulfhydryl containing compounds
(like GSH) which might have been inactivated by radiation, nor the reducrion of partial oxygen
pressure, nor the formation of mixed disulfides, but more likely caused by the scavenging of
free radicals by cysteamine [25]. In this respect the scavenging of OH'radicals by cysteamine
might be of major importance [26], although others [27] suggest tht cysteamine does not
protect against radiation damage by OH'radical scavenging. Scavenging OH'radicals is also
thought to be an important protecrive mechanism in lipid peroxidation, however the contribution
of OH'radicals in Fee+ / ascorbate induced lipid peroxidation has seriously been questioned
[12]. Although our data suggest that cysteamine protects against lipid peroxidation by
scavenging free radicals i.e. a mechanism different from that of GSH, the exact mechanism is
not elucidated.
It has been reported that heat-pretreatment the microsomes [2] or pretreating the microsomes
with the thiol alkylating compound N-ethyl maleimide [4] reduced the GSH-dependent
protection via the free radical reductase. In this study it was observed that these pretreatments
did not affect the protection afforded by cysteamine. This points out that cysteamine does not
.i

protect via an enzymatic factor e.g. the heat labile free radical reductase which is operative
during the protection by GSH. Liver microsomes prepared from rats after in vivo lipid
peroxidation did not show a reduced protection against lipid peroxidation by cysteamine. This
suggests that cysteainine might provide protecrion against toxicity induced by lipid peroxidation,
even when the process of lipid peroxidation has already started. In patients, cysteamine
appeared to be still effective against the hepatoxicity of paracetamol at a relatively late time of
administration [5,6]. In contrast, it has been shown that in cultured cells cysteamine might
induce oxidative stress, which is probably due to the generation of hydrogen peroxide during
the autoxidation of cysteamine [28]. However, cysteamine administered in vivo to mice or rats
caused little toxicity [25]. Also in humans no major adverse effects were observed after
administraring a therapeutic dose of cysteamine [5].
In conclusion,in this study it was found that compounds structurally closely related to GSH
and with a free thiol group such as 1-cysteine and N-acetyl l-cysteine, did not provide direct
protection against in vitro lipid peroxidation, indicating that the enzymatic protection against
lipid peroxidation via the free radical reductase is rather specific for GSH. A comparable
specificity exists for other GSH-dependent enzymes, e.g. the selenium containing G5Hperoxidase [29]. Because S-methyl glutathione did not protect against in vitro lipid
peroxidation we conclude that the free thiol group in GSH is essential for its protecting capacity.
Based on the results obtained in this study, the protective effect of a compound like N-acetyl
1-cysteine against lipid peroxidation is not provided by a direct effect of N-acetyl 1-cysteine
against lipid peroxidation, but the protective effect is more likely indirect by elevating GSH
levels. Treatment of oxidative stress occurring under certain patho-physiological conditions
normally proceeds when damage is already manifest. After in vivo lipid peroxidation the
protection by GSH is less effecrive since the free radical reductase is impaired [2], whereas the
anri-oxidant capacity of cysteamine remains the same. This indicates that administration of
cysteamine might provide an alternative for N-acetyl l-cysteine which is usually applied.
References.
1
2
3

5
6
7
8
9
10
11

A. Bast and G.R.M.M. Haenen, Cytochrome P-450 and glutathione, what is the significance of their
interrelarionship in lipid peroxida[ion, Trends Biochem. Sci. 9(1984)510-513.
G.R.M.M. Haenen and A. Bast, Protection against lipid peroxidation by a microsomal glutathione dependentlabile factor, FEBS Lett. 179(1983)24-28.
P.B. McCay, E.K. Lai, D.D. Gibson, J.L. Poyer, S.R. Powell and G. Breuggeman, Biological systems
which surpress lipid peroixdation, in: P.K. Singal (Ed.), Oxygen radicals in the pathology of heart disease,
Kluwer Academic Publishers, Boston, 1987, pp. 13-24.
G.R.M.M. Haenen, J.N.L. Tai Tin Tsoi, N.P.E. Vermeulen, H. Timmerman and A. Bast, 4-Hydroxy-2,3trans-nonenal stimulates microsomal lipid peroxidation by reducing the glutathione-dependent protection,
Arch. Biochem. Biophys. 259(1987)449-456.
L.F. Prescott, J. Park, G.R. Sutherland, I.J. Smith and A.T. Proudfoot, Cysteamine, methonine and
penicillamine in the treatment of paracetamol poisoning, The Lancet(1976) 109-113.
A.L. Harts,Paracetamol-induced acute renal failure, Br. Med. J. 284(1982)825.
M. Tien, J.R. Bucher and S.D. Aust, Thiol-dependent lipid peroxidation, Biochem. Biophys. Res. Commun.
107(1982) 279-285.
A.J.F Searle and R.L. Willson, Stimulation of microsomal lipid peroxidation by iron and cysteine,
Biochem. J. 212(1983) 549-554.
J. Chaudiere, E.C. Wilhelmsen and A.L. Tappel, Mechanism of selenium-glutathione eeroxidase and its
inhibition by mercaptocarboxylic acids and other mer~captans, J. Biol. Chem.259(1984) 1043-1050.
D.J. Kornbrust and R.D. Mavis, Microsomal lipid peroxidation, Mol. Parmacol. 28 (1980)2051-2055.
G.H.R. Rao, J.M. Genard, J.W. Eaton and White, The role of iron in prostaglandin synthesis, ferrous iron
Gil

mediated o~cidation of arachidonic acid,Prostaglandins Med. 1 (1978)55-70.


12 G.R.M.M Naenen, N.P.E. Vermeulen, H. Timmerman and A. Bast, The practical consequence of the proand anti-oxidant properties of vitamin C in vitro, submitted.
13 O.H. Lowry, N.J. Rosenbrough, A.L. Farr and R.J. Randall, Protein measurement with the Folin phenol
reagent, J. Biol. Chem. 1963 (1951) 265-275.
14 D.A. Rowley and B. Halliwell, Superoxide-dependent formation of hydroxyl radicals in the presence of thiol
compounds,FEBS Lett 138(1982) 33-36.
15 W.D. Cash and M. Gardy, Role of metal contaminants in mitochondria) swelling activities of reduced and
oxidized glutathione preparations, J. Biol Chem. 240(1965)3450-3452.
16 E.G. Mimnaugh,Potentiation by reduced glutathione of adriamycin-stimulated lipid peroxidadon in kidney
microsomes, Bi~hem.Ph~*anac. 35(1486)433?334.
17 A. Bast, G.R.M.M Naenen and E.M. Savenije-Chapel, Inhibition of rat hepatic microsomal lipid
peroxidaton by mesna via glutathione, Arzneim. Forsch., 37(1987) 1043-1045.
18 G. Seaz, P.J. Thornalley, H.A.O. Hill, R. Hems and J.V. Bannister, The production of free radicals during
the autoxidaton of cysteine and their effect on isolated rat hepatocytes, Biochim. Biophys. Acta 719(1982)
24-31.
19 A.J.F. Searle and A. Tomasi, Hydroxyl free radical production in iron-cysteine solutions and protection by
zinc, J. Inorg. Biochem. 17 (1982) 161-166.
20 H. Glat, M Protic-Sablj and F. Oesch, Mutagenicity of glutathione and cysteine in the Ames test, Science
220(1983)961-963.
21 J.M. Braughler, L.A. Duncan and R.L. Chase, The involvement of iron in lipid peroxidation, J. Biol. Chem.
261 (1986) 10282-10289.
22 G, Minotti and S.D. Aust, The requirement for iron (IIn in the initiation of lipid peroxidation by iron (II)
and hydrogen peroxide, J. Biol. Chem. 262(1987) 1098-1104.
23 H.P. Misra, Generation of superoxide free radical during the autoxidation of thiols, J. Biol. Chem. 249
(1974)2151-2155.
24 D.O. Lambeth, G.R. Ericson, M.A. Yorek and P.D. Ray, Implications for in vitro studies of the
autoxidation of ferrous ion and the iron-catalyzed autoxidation of dithiothreitol, Biochim. Biophys. Acta 719
(1982)501-508.
25 Z.M. Bacq and P. Alexander, Chemical protection against X-and gamma-rays, in: Fundementals of
radiobiology, 2nd edition,Pergamon Press, Oxford, 1961, pp. 457-483.
26 Y.N. Korystov and F.B. Vexler, Mechanisms of the radioprotecve effect of cysteamine in ,E. coli, Radiat.
Res. 114(1988) 550-555.
27 J.W. Hulsewede and D. Schulte-Frohlinde, Radiation protection of E. coli strains by cysteamine in the
presence of oxygen, Int. J. Radiat. Biol. 50(1986)861-869.
28 Y. Takagy, M.Shilta, T. Terasima and S, Akaboshi, Specificity of radioprotective and cytotoxic effects of
cysteamine in Hela Sg cells: generation of hydrogen peroxide as the mechanism of paradoxial toxicity,
Radiat. Res. 60(1974)292-301.
29 L. Floh, Glutathione eeroxidase, factor fiction, in: Oxygen free radicals in tissue damage, Ciba Foundation
Symposium 65(new series) Exerpta Medica, Amsterdam, 1979, pp. 95-122.

62

- Chapter 8 Interplay between dihyrolipoic acid and glutathione in the


protection against microsomal lipid peroxidation
Abstract. Reduced glutathione(GSH)delays microsomal lipid peroxidation via the reduction of
vitanun E radicals which is catalyzed by a free radical reductase (Naenen, G.R.M.M. et al.
(198i} arch. Biochem. Biaphys. 259, 449-~56}. Li~oic acid exerts its therapeutic effect iri
pathologies in which free radicals are involved. We investigated the interplay between lipoic
acid and glutathione in microsomal rel+(10 M)/ascorbate (0.2 mM) induced lipid
peroxidation. Neither dihydrolipoate nor lipoate (0.5 mM) displayed protection against
microsomal lipid peroxidarion, measured as thiobarbituric acid reactive material. Dihydrolipoate
had even apro-oxidant activity which is probably due to reduction of Fe3+. However,
protection against lipid peroxidation was afforded by the combination of oxidized glutathione
(GSSG)and dihydrolipoate. It is shown that this effect can be ascribed completely to reduction
of GSSG to GSH by dihydrolipoate. This may provide a rationale for the therapeutic
effecriveness of lipoic acid.
Introduction.
Lipoic acid is essential in three different mitochondrial dehydrogenase complexes [1]. Its
therapeutical application is found in treatment of intoxication with the mushroom Amanita
Phalloides, peripheral polyneuropathies and liver cirrhosis. A reduced level of endogenous
lipoic acid has been found in parients with liver cirrhosis, diabetes mellitus, artheriosclerosis
and polyneurites [1]. It is suggested that free radical mediated peroxidation of cell membranes is
a common pathway in the eriology of these pathologies [2]. Reduced glutathione(GSH)plays a
key role in the defense against oxidarive stress in various ways [3]. As a pivotal mechanism in
its acrion as a protective compound, glutathione circumvents lipid peroxidation via the reducrion
of vitamin E radicals. The regeneration of vitamin E from vitamin E radicals is catalyzed by a
glutathione dependent membrane bound vitamin E free radical reductase [4].
The present study was embarked upon to investigate the influence of lipoic acid on the
glutathione dependent protection against hepatic microsomal lipid peroxidation.
Materials and Methods.
Male Wistar rats (Harlan Olec CPB,Zeist, the Netherlands), 200 - 250 g were killed by decapitation. Livers
were removed and homogenized (1 : 2 w/v) in ice-cold phosphate buffer(50 mM,pH 7.4) containing 0.1 mM
EDTA. The homogenate was centrifuged at 10,000 x g (20 min. at 4C). Subsequently the supernatant was
resuspended in the phosphate buffer(2 g liver / ml) and stored at -80C. Before use the microsomes were thawed
and diluted 5-fold with ice-cold Tris-HCl buffer(50 mM,pH 7.4)containing 150 mM KCl and washed twice by
centrifugation at 115,000 x g (40 min.). Finally the pellet was resuspended in the Tris buffer and used.
Incubarions with microsomes (final concentration of 1/g g liver/ml) were performed at 37 C, with shaking air
being freely admitted n the Tris-NCI / KCl buffer. Compounds that were added were dissolved in the Tris buffer,
Ascorbic acid, GSH and dihydrolipoate were neutralized with KOH before addition. After a preincubation for 3
min. the reactions were started by adding a freshly prepared FeSO4 solution. Experiments were performed three
times. Data presented in figures are representative examples.
Lipid peroxidation was assayed by measuring thiobarbituric acid (TBA) reactive material as reported

63

2~
0.5
m

1~`

11

10

20

30

40

.50

INCUBATION TIME(MIN)
Fig. 1. Time course of lipid peroxidation. Hepatic microsomes were incubated without(1)or with (2) 0.5
mM dihydrolipoate. The incubation reactions were started by the addition of 10 M Fee+.
previously [4]. An aliquot of the incubation (0.3 ml) was stopped by mixing with ice-cold TBA-trichloroacetic
acid-HCl-butylhydroxytoluene solution(2 ml)[4]. After heating (15 min. 80C) and centrifugation (15 min.) the
absorbance at 535 nm vs. 600 nm was determined.
GSH and oxidized glutathione (GSSG) were determined according to Hissin and Hilf [5]. The small
absorbance of dihydrolipoate in the GSH assay was corrected for if necessary.
Racemater of dihydrolipoate and lipoate were gifts from Asta-Werke, Frankfurt, FRG. All other chemicals
used were of analytical grade purity.

Results.
Dihydrolipoate stimulates the Fel+induced microsomal lipid peroxidarion, whereas lipoate
has no effect (fig. 1 and note in legend of fig. 1). Ascorbic acid (0.2 mM), added to Fel+ (10
M)strongly srimulates lipid peroxidation (fig. 2). In the Fel+/ascorbate incubation system, 0.5
mM dihydrolipoate gave a short lag time in the production of lipid peroxidation. This initial
protection by dihydrolipoate was followed by a potentiation of lipid peroxidation. The
incubation with GSH in a concentrarion of 1 mM, which contains an equimolar amount of
sulfhydryl moieties as 0.5 mM dihydrolipoate, gave a more extensive lag time. Compared to 0.5
mM dihydrolipoate (fig. 2) GSSG (0.5 mM)or lipoate (0.5 mM)were ineffective. However,
the combination of GSSG (0.5 mM) with dihydrolipoate (0.5 mM) resulted in a more
pronounced protection against lipid peroxidation than with dihydrolipoate alone (fig. 2). The
lipid peroxidarion found in the presence of GSH with either reduced or lipoate did not differ
from the peroxidation seen with GSH alone.
Heating the microsomes in boiling water for 90 s. abolished the protecrive effects of either
GSH alone or of the combination GSSG and dihydrolipoate (fig. 3). Since it is known that the
GSH dependent protection is mediated by a heat labile vitamin E free radical reductase the
experiments as shown in fig. 2 and 3, suggested that the protective effects of the combination of
C~~

2.0
0
0

1.0

11

10

20

30

40

50

INCUBATION TIME(MIN)
Fig. 2. Time course of lipid peroxidation. Hepatic microsomes were incubated with 0.2 mM ascorbate and
(1) no addition,(2)0.5 mM dihydrolipoate,(3) 1 mM GSH,(4)0.5 mM dihydrolipoate and 1 mM GSH,
(5) 0.5 mM dihydrolipoate and 0.5 mM GSSG. Addition of 0.5 mM GSSG alone or 0.5 mM lipoate
alone had no effect and the result was identical to time course (1). Addition of 0.5 mM lipoate and 1 mM
GSH did not differ from the effect obtained with 1 mM GSH alone (3). The latter time courses are not
indicated for sake or clarity. All reactions were started by addition of 10 M Fel+. _

dihydrolipoate with GSSG proceed via GSH. In order to verify this possibility, the formation
of GSH from GSSG by dihydrolipoate was determined. As shown in table 1 we indeed
observed an extensive but incomplete reducrion of GSSG by dihydrolipoate.
The requirement for dihydrolipoate in the inhibition of lipid peroxidarion by GSSG and the
importance of the dihydrolipoate: GSSG ratio was further investigated. Microsomal lipid
peroxidation was measured in the presence of various concentrations of dihydrolipoate and
GSSG with a total concentration of both compounds of 1 mM. As shown in fig. 4B an optimal
protection at the ratio of 1 : 1 after incubation of the microsomes was found. The inhibirion is
time dependent. A longer incubation period reduced the inhibitory effect of the combination
GSSG and dihydrolipoate. The same phenomenon is observed for GSH, longer incubation
results in less inhibition (fig. 4A).

condirion

GSH(mM)

GSSG(mM)

0.5 mM GSSG
0.5 mM dihydrolipoate +
0.5 mM GSSG

0.00 0.01

0.51 0.02

0.42 0.02

0:27 0.02

Table 1. Reduction of GSSG to GSH by dihydrolipoate. Incubations were performed for 45 min. at 37C
in Tris-HCl/KCl buffer (as in material and methods section). Data represent the mean SEM of three
experiments.

65

4~
~~
53

1.5

-~

1 ~*

0
0

,~, 1.0

0.5

11

10

20

30

40

50

INCUBATION TIME(MIN)
Fig. 3. Time course of lipid peroxidation. Heated hepatic microsomes were used. Indications as n fig.2.
Addition of 0.5 mM GSSG alone or of 0.5 mM lipoate alone had no effect and the result was identical to
time course (1). Addition of 0.5 mM lipoate and 1 mM GSH did not differ from the effect obtained with 1
mM GSH alone (3).

Discussion.
Rat liver microsomes contain a GSH dependent heat labile vitamin E free radical reductase
which participates in the protection against lipid peroxidation [4, 6, 7]. Lipoic acid is
particularly used in treatment of pathophysiological conditions like a wide variety of liver
diseases or dysfunctions in which lipid peroxidation appears to be involved. This prompted. us
to study the effect of lipoic acid on lipid peroxidation and to focus on the interplay between
lipoic acid and glutathione in lipid peroxidation.
During the iniriation of the process of lipid peroxidation Fel+converts to Fe3+. Ascorbate
stimulates Fel+ induced lipid peroxidation by the regeneration of Fel+ from Fe3+ [8]. With a
concentrarion 0.2 mM ascorbate an oprimal ratio Fe3+: Fe2+ which is needed for a maximal rate
and extent of lipid peroxidation is apparently achieved [9].
In the present study we found that neither reduced nor lipoate (0.5 mM)alone displayed a
prominent protective effect on mcrosomal lipid peroxidation. Dihydrolipoate enhanced Fel+
srimulated lipid peroxidation (fig. 1) but it produced a short lag tune in Fel+/ascorbate
stimulated lipid peroxidarion (fig. 2). Both effects are probably due to the reduction of Fe3+ by
dihydrolipoate. In the Fel+ srimulated lipid peroxidarion the reduction of Fe3+, which is formed
during the lipid peroxidarion process, apparently provides a more favorable ratio Fe3+: Fe2+.
This explains the pro-oxidant acrivity of dihydrolipoate (iig. 1). In the Fee+/ascorbate stunulated
lipid peroxidation reduction of iron by both reductants ascorbate and dihydrolipoate will result
in an extensive reduction of iron which may give a deviarion from the oprimal Fe3+:Fe2+ratio.
This is consistent with the observed lag time in lipid peroxidarion time course in th presence of
dihydrolipoate in fig. 2. Similarly, the potentiation of Fel+/ascorbate stimulated lipid
..

O ~ I.0

z
o~

1.5 a

H ~

Q 1.~
X

Q
X p

O ~-

~ v
W .?

~ ~ 0.5

~ ~ 0.5

m
Qv
~~
~

~~
:~

0.0
0.0

0.2

0.4

0.6

0.8

1.0

[GSH](mM)
Fig.4A. The dependence of lipid peroxidation
on the GSH concentration. Lipid peroxidation
was stimulated by 0.2 mM ascorbate and 10 M
Fel+ and was determined after 10 min.(1), 30
min.(2) and 45 min.(3) as relative to the control
(without GSI~.

0.0
0.0

0.2

0.4

0.6

0.8

1.0 [GSSG](mM)

1.0

0.8

0.6

0.4

0.2

0.0 ~~Y~olipoate] (mM)

Fig. 4B. The dependence of lipid peroxidation


on the dihydrolipoate: GSSG ratio. Lipid
peroxidation was stimulated by 0.2 mM
ascorbate and 10 M Fel+ and was determined
after 10 min. (1), 30 min. (2) and 45 min. (3)
as relative to the control (without thiol).

peroxidation seen after 15 min. incubation with dihydrolipoate is in agreement with its capacity
as an iron reductant. During the lipid peroxidation ascorbate is consumed [8] and dihydrolipoate
can replace ascorbate as a reluctant.
Protection by the combinarion GSSG and dihydrolipoate appears to proceed via the
formation of GSH (Table 1). Addition of 0.5 mM dihydrolipoate to 0.5 mM GSSG gave only
partial reduction to GSH (approximately 50 a/o). The protecrion given by GSSG/dihydrolipoate
is in excellent agreement with the protection provided by the corresponding GSH concentration
[4]. If the inhibition of lipid peroxidarion by the combination GSSG/dihydrolipoate occurs
uniquely via GSH,it is anticipated that the ratio dihydrolipoate : GSSG is at least 1 for optimal
inhibirion, which was indeed found (fig. 4B).
The protection against oxidative stress in liver microsomes by GSH predominantly occurs
via the vitamin E free radical reductase thus leading to a continuous regeneration of vitamin E
[4]. As shown in the present study part of the protective effect associated with dihydrolipoate on
microsomal lipid peroxidation can be ascribed to reduction of GSSG to GSH.The inhibition of
lipid peroxidation by dihydrolipoate via the formation of GSH from GSSG provides a rationale
for the therapeuric effectiveness of this drug. At the same rime the concurrent intracellular
presence of glutathione and endogenous lipoic acid may well regulate the intracellular level of
lipid peroxidation in normal physiological conditions.
Acknowledgement.
The generous gifts of dihydrolipoate and lipoate by Asta-Werke (Frankfurt, FRG) is
gratefully acknowledged.
References.
1 Ehrenthal, W.and Prellwitz, W.(1986)in : Interdiziplin~re Bestandsaufname der Polyneuropathien
(Neundtirfer, B. and Sailer, D., els.), pp.154-166,Perimed-Fachbuchverl. GmbH,Erlangen.
2 Halliwell, B.and Cutteridge J.M.C.(1985)in: Free radicals in biology. and medicine, Claredon Press, Oxford.
3 Bast, A. and Naenen, G.R.M.M.(1984) Trends Biochem, Sci. 9,510-513.
4 Naenen, G.R.M.M and Bast, A.(1983)Fedn. Eur. Biochem. Soc. Lett. 159, 24-28.
5 Hissin, P.J. and Hilf, R.(1976) Anal. Biochem. 74, 214-226.

6 Haenen, G.R.M.N, Tai Tin Tsoi, J.N.L., Vermeulen, N.P.E., Timmerman, H. and Bast, A.
(1987) Arch.
Biochem. Biophys. 259,449-456.
7 Burk, R.F.(1983) Biochim. Biophys. Acta 757,21-28.
8 Bast, A., Haenen, G.R.M.N and Savenije-Chapel, E.M.(1987) Arzneim. Forsch./Drug Res.
37, 1043-1045.
9 Minotti, G. and Aust, S.D.(1987) 7. Biol. Chem. 262, 1098-1104.

68

Chapter 9 Mechanism of the reaction of ebselen with endogenous thiols


dihydrolipoate is a better cofactor than glutathione
in the peroxidase activity of ebselen

Abstract. The therapeuric effect of ebselen has been linked to its peroxidase activity. In the
present study, the peroxidase activity of ebselen toward H2O2 with the endogenous thiols
glutathione(GSH)and dihydrolipoate(L(SH)2)as cofactors was determined. When GSH was
used, peroxide removal was described by a ter uni ping pong mechanism with Dalziel
coefficients for GSH and H2O2 of 0.165 0.011 mM min and 0.081 0.005 mM min
respectively. When L(SH)2 was used, peroxidase activity was independent of the concentration
L(SH)2, in the concentration range studied..(5 M - 2 mM), and peroxide removal was only
dependent on the concentration of H2O2 and ebselen, the second order rate constant being 12.3
0.8 mM-1 min-1. To elucidate the difference between GSH and L(SH)2, the molecular
mechanism of the peroxidase activity of ebselen was invesrigated, using UV specirophotometry,
HPLC,~~Se NMR and mass spectrometry.
GSH was found to react fast with ebselen to give a selenenyl sulfide, an adduct of GSH to
ebselen. Subsequently the GSH-selenenyl sulfide is converted into the diselenide of ebselen.
Finally the diselenide reacts with a peroxide and ebselen is regenerated. The formation by GSH
of the diselenide from the GSH-selenenyl sulfide of ebselen is slow and linearly dependent on
the concentration free thiol, however, no net consumprion of GSH was observed. Furthermore,
it is likely that a selenol is an intermediate in diselenide formation.
After reaction between ebselen and L(SH)2 the diselenide of ebselen was immediately
detected. The fast formation of the diselenide with L(SH)2 versus the slow formation of the
diselenide with GSH,accounts for our observation that L(SH)2 is a better cofactor than GSH in
the peroxidase activity of ebselen. Our results suggest that the interaction between ebselen and
L(SH)2 might be of major importance in the mechanism by which ebselen exerts its therapeutic
effect.
Introduction.
In the protection against oxidative stress glutathione (GSH)-dependent processes play an
important role (1). Therefore, one of the strategies in the treatment of various pathologies
associated with oxidative stress, is to stimulate the GSH-dependent protection. In clinical
practice this might be accomplished by administrarion of N-acetyl-l-cysteine, a precursor of
glutathione (2). Another approach is to potentiate the GSH-dependent protection by applicarion
of ebselen, a relatively new seleno-organic compound. Ebselen has been shown to be a
promising anti-inflammatory drug. However, the mechanism of its anti-inflammatory action is
still a matter of debate (3). Ebselen directly inhibits lipoxygenase and cyclooxygenase, it
converts leukotriene LTBq into an inactive isomer, and ebselen itself is a potent antioxidant. In
addition, ebselen possesses a GSH-eeroxidase like activity, using hydrogen peroxide and lipid
hydroperoxides as substrate (4-5). In this study we focus on the eeroxidase activity of ebselen.

69

GSSG
Enz-SeH
ROOH
GSH
Enz-SeSG
H2O
ROH
Enz-SeOH
GSH

Fig. 1 Reaction scheme of the endogenous selenium-dependent GSH-eeroxidase. Hydroperoxides(ROOF


are reduced to the corresponding alcohols at the expense of GSH by a ter uni ping gong mechanism.
The selenium-dependent GSH-eeroxidase reduces the increased "peroxide tone" e.g. in
inflammation, by catalyzing the reacrion between GSH and hydroperoxides (7, 8). In its
eeroxidase acrivity, the enzyme first reacts with a hydroperoxide, a reaction in which the selenol
group of the enzyme (Enz-SeH) is converted into a selenenic acid (Enz-SeOH) and the
hydroperoxide into the corresponding alcohol (fig 1). Subsequently, the formed selenenic acid
reacts with GSH under the formation of a selenenyl sulfide (Enz-SeSG). Finally, the selenol
form of the enzyme is regenerated by a reaction with a second thiol (fig 1)(7).
The selenium-dependent GSH-eeroxidase selecrively uses GSH as cofactor (9). When
ebselen is compared with the selenium-dependent GSH-eeroxidase, ebselen resembles the
selenenic acid form of the enzyme. Under in vivo condirions the eeroxidase activity of ebselen
depends on the reducrion of ebselen to its diselenide by thiols (10). Subsequently, the diselenide
reacts with a peroxide under the regenerarion of ebselen (10). It has already been reported that
synthetic thiols such as dithioerythreitol (11) and N-acetyl-L-cysteine (12) can serve as a
substitute for GSH in the eeroxidase activity of ebselen. Beside GSH,dihydrolipoate(L(SH)2)
is an important endogenous SH-containing compound. L(SH)2 is a dithiol with vicinal thiol
groups, that can form an intramolecular disulfide in an energetically favorable five-membered
ring. L(SH)2 plays a key catalytic role in pyruvate and a-ketoglutarate oxidase. Like GSH,
L(SH)2 provides protection against freeradical-mediated injury. Recently, an interplay between
L(SH)2 and GSH in reducing the "peroxide tone" by protecring against lipid peroxidarion has
been demonstrated (13, 14). The aim of this investigation is to deternune whether L(SH)2 is
also accepted by ebselen as cofactor and,furthermore, to study the mechanism of the eeroxidase
activity of ebselen.
Materials and methods.
GSH was obtained from Sigma (St. Louis,U.S.A.). Ebselen (2-phenyl-1,2-benzisoselenazol-3 (2I~-one) and
ebselen derivarives were gifts of Rhne-Poulenc Nattermann (Cologne,F.R.G.). Racemates of dihydrolipoic acid
70

(()-6,8-dimercapto-octanoic acid) and lipoic acid (()-1,2-dithiolane-3-pentanoic acid) were gifts of Asta Pharma
AG (Frankfurt, F.R.G.). All other chemicals were of analytical grade purity.
All incubations were performed at 37C in a 10 mM potassium phosphate buffer, pH 7.4, unless otherwise
stated. Ebselen was dissolved in DMSO. Maximal concentration DMSO was 1%.
Hydrogen peroxide was determined with[he iron-thiocyanate method according to Hildebrandt and Roots (15).
This method is based on the oxidation of Fel+ to Fe3+ by H2O2.The formed Fe3+,complexed to thiocyanate,
is quantified spectrofotometrically at 480 nm.Free thiol groups were assessed according to Ellman (16).
The HPLC analysis of ebselen, GSH-selenenyl sulfide and diselenide of ebselen was basically according to the
methods of Muller et ai.(17) and of Terlinden et a.(18). Samples of 20 l were injected onto a reversed phase
column (Nucleosil C18, Chrompack, Middelburg, The Netherlands) and the products were monitored by IJVabsorption at 313 nm. The mobile phases consisted of mixtures of acetonitrile(A)and 0.1 % H3POq.(B)at the
flow rates of 0.6 ml / min. A mixture of SO %n A and 50 % B was used to elute ebselen (retention time 2.4Q
min); a mixture of 30 % A and 70 % B was used for the GSH-selenenyl sulfide (retention time 3.12 min); a
mixh~re of 70 %o A and 30 % B was used for the diselenide (rekention pme 3.75 min).
All ~~Se NMR spectra were recorded on a Bruker MSL 400 in a tube of 7 mm outer diameter in dime[hyl
formamide(DMF)at room temperature. Chemical shifts are reported relative to dimethylselenide. Ebselen served
as reference, with a chemical shift of959 ppm (10).
Mass spectrometric analysis was carried out on a Finnigan MAT 90 mass spectrometer. Samples were
duectly introduced into the mass spectrometer. Direct chemical ionization was performed, using isobutane as the
reagent gas, an ion source temperature of50 C,at a pressure of 10~ Ton.
In principle, all results are expressed as mean SD (n = 3-5). In most experiments the SD was smaller than
5% and the SD was omitted for the sake of clarity.

Results.
As shown in fig 2, L(SH)2 did not react spontaneously with hydrogen peroxide at a
measurable rate. However,in combination with ebselen, L(SH)2 degraded hydrogen peroxide
in the incubation medium very fast. PZ 25(RP 62373), an analog of ebselen in which selenium

1.0
.-.

~ 0.5
N
u

~ ~

10

15

time(min)
Fig. 2 Reaction of 0.5 mM dihydrolipoate (, ) or 2 mM GSH(O,1) with 1 mM hydrogen peroxide.
In the experiments depicted with the closed symbols 10 M ebselen was added. The reaction was started
by the addition of hydrogen peroxide.

71

F.

~.~

]0

20

40

50
.~1

~1 t

~H2~2](mM)

[ebselen] =10 M

10

15

2~

ot..
m

U
N

~
0

.~1

~ 1~

r;

.~

~`
~.

.1

~-rT

[dihydrolipoate](mM)

.~1

Fig. 3 Peroxidase activity of ebselen with dihydrolipoate as cofactor. Panel A shows the dependence of hydrogen peroxide removal on the
concentration of ebselen, the pseudo fust order rate constant with a initial concentration of hydrogen peroxide of 1 mM'is shown. Panel B
shows the dependence on the concentration of hydrogen peroxide, the pseudo fust order rate constant with a concentration of ebselen of 10
M is shown.Panel C shows the dependence on the concentration of dihydrolipoate, various concentrations of hydrogen perolcide and ebselen
were used and the second order rate constant(k) was determined using the formula: d[H2O2]/tit= -k [H2O2][Ebselen]. The second order rate
constant(k) appeared to be 12.3 0.8 mM-1 min'1.

[ebselen](M)

30

ctS
}.

~ .O 1

O.2

[H2di]= 1 mM

0.4

.~
~

0.6
.~
~ .1

_^

1~

Q5

0.5

04

0.4

w
Z 0.3
Q
m

0.3

O
m 02

02

0.1

pi
280

i
300

i
i
360
340
320
WAVELENGTH (nm)

380

01
280

i
300

i
320

~
340

~
360

'
380

WAVELENGTH (nm)

Fig. 4 Change in the UV-absorption spectrum of 50 M ebselen by the addition of GSH (panel A) or
dihydrolipoate (panel B). The concentrations of GSH (panel A)were: 0 M (1), 12.5 M (2); 25 M (3),
37.5 M (4), 50 M (5) or 60 M (6). The concentrations dihydrolipoate (panel B) were: 0 M (1), 12.5
M (2) or 25 M (3). All spectra were recorded immediately after the addition of the thiol except for
spectrum 6 panel A, which was recorded 10 min after the addition of60 M GSH.

has been replaced by sulfur, proved to have no catalytic activity in the reaction betwen L(SH)2
and hydrogen peroxide (data not shown). It is also shown in fig 2 that GSH - in contrast to
L(SH)2 -reacted spontaneously with hydrogen peroxide, however, GSH was found to be less
potent -compared to L(SH)2 - in hydrogen peroxide removal when ebselen was present. When
L(SH)2 was used, the reaction kinetics could be described by a second order reaction, ui'which
the rate of hydrogen peroxide removal was linearly dependent on both the concentration of
hydrogen peroxide and the concentration of ebselen (rate constant 12.3 0.8 tnM-1 min-1), but
independent of the concentration of L(SH)2 in the concentrarion range investigated,5M - 2
mM (fig 3). Lipoate -the oxidized form of L(SH)2 -was not effective as cofactor in mediating
the peroxidase acrivity of ebselen (data not shown).
To elucidate the cause of the difference between GSH and L(SH)2 in the peroxidase activity"
of ebselen, the reactivity of ebselen towards both thiols was determined. In the UV-spectrum of
ebselen, the absorption maacimum at 324 nm is due to the isoselenazol ring (4). As shown in fig
4A, addirion of 12.5, 25, 37.5 or 50 M GSH to 50 M ebselen resulted in a concentration
dependent reduction of the absorption at 324 nm, indicating that GSH opens the isoselenazol
ring (c 4). T'he isobesric point at 318 nm indicates that a relative stable product or a mixture of
stable products of constant composirion was formed. The UV-spectrum recorded immediately
after the addition of an excess of GSH to the reacrion mixture, did not differ from the spectrum
obtained with an equimolar amount of GSH and ebselen. However, several minutes after the
addition of GSH, a second product appeared, provided that more than 50 M GSH was added
to the reaction mixture containing 50 M ebselen (iig 4A, curve 6). When L(SH)2 instead of
GSH was used, already at relatively low concentrations of L(SH)2 the spectrum, that was
observed with an excess of GSH only after several minutes, appearedunmediately (fig 4B).
In order to find out which products correspond to the observed UV-absorprions, the reacrion

73

.1

~~

40
0
.~
~ 20
U

A
O
U

20

40

60

80

[GSH](IVs
Fig. 5 Formation of the GSH-selenenyl sulfde of ebselen(1)and the diselenide(~)from 50 M ebselen
() and a varying concentration of GSH;(+) is the concentration free sulfhydryl groups. The reaction
mixtures were analyzed 15 sec after GSH was added using HPLC.

mixtures of ebselen and either GSH or L(SH)2 were analyzed by HPLC. A sample of the
reaction mixtures taken immediately after the addition of the thiol was injected onto the HPLCcolumn. In the case of GSH it was found that the GSH-selenenyl sulfide of ebselen (S-(2phenyl carbamoyl benzeneslenenyl)-glutathione) was formed (iig 5). Only with relatively high
concentrations of GSH,some diselenide of ebselen (2,2-diselenobis-(N-phenyl-benzamid)) was
detected. The reaction between GSH and ebselen was fast and complete, one mole of GSH
consuming one mole of ebselen and the major product being the GSH-selenenyl sulfide of
ebselen.
When the ebselen-GSH mixture was incubated for longer periods, the GSH-selenenyl
sulfide gradually converted into the diselenide of ebselen. The rate of diselenide producrion was
strongly dependent on the GSH concentrarion (fig 6). When ebselen was added in excess to
GSH, no free GSH was detected (fig 5) and the diselenide was formed slowly (fig 6). When
GSH was added in excess to ebselen, diselenide formarion was much faster. No extra thiol was
consumed in the diselenide formarion (data not shown). Diselenide formarion followed second
order rate kinerics with respect to the concentration of the GSH-selenenyl sulfide of ebselen,
and the concentration of free GSH. The rate constant of diselenide formation was found to be
6.1 0.4 mM-1 min-1 (fig 6). Addirion of GSSG did not decrease the GSH-mediated diselenide
formation from ebselen (data not shown).
Variation of the pH of the incubation medium showed that the rate of diselenide formation
decreased when the pH declined. Using a pKa of the thiol of GSH of 8.6(19)for calcularion, it
was found that the second order rate constant correlated with the fraction of GSH deprotonated
(fig 7). This indicates that the deprotonated form of GSH is the active species
When L(SH)2 was used in the reaction with ebselen instead of GSH, the diselenide of
ebselen was immediately detected using HPLC. No selenenyl sulfides nor other intermediates
could be detected. The reaction between the dithiol L(SH)2 and ebselen to form the diselenide
was fast and complete, one mole of L(SH)2 consuming two moles of ebselen (fig 8).
In order to further idenrify the products of the reaction between ebselen and L(SH)2, we
74

TJ
u

10

~20

30

5
3

10

1~

2~

,~ 0

~
>,2

4'
~
~ 3

2
6

time(min)

4
8

4
6~ S n ~

3~v

10

2ti

1-0

~
OA

~ 0.2

~ 0.4
~
0

~~ 0.6
.

o.8 c

10

20

30

40

[GSH]free(M)

Fig.6 Time course of the diselenide (panel A)and the selenenyl sulfide (panel B)concentration after the addition of various amounts of GSH
to 50 M ebselen. The concentrations of GSH added were: 50 M (curve 1),60 M (2), 70 M (3),80 M (4),90 M (5), 100 M (6)and
150 M (7). In panel C the dependence of the rate of selenenyl sulfide consumption on the concentration of free GSH is shown.

time(min)

50

60

20
~

~
+~ ..-a
~ ~ 10
a~ ~
+~

n
0.00

0.05

0.1Q

[GS-]
[GS-]+[GSH]
Fig. 7 pH dependence of diselenide formation from the GSH-selenenyl sulfide of ebselen by GSH. At pH
6.0, 7.4 and 7.6 the rate constant for selenenyl sulfide consumption was determinec and plotted against the
rario GSH deprotonated (calculated using a pKa of 8.6).

studied the reaction mixtures with ~~Se-NMR. Natural abundance of ~~Se is 7.6 % of the total
amount of selenium, and ~~Se can be used to monitor reactions of ebselen with NMR.
However, due to low sensitivity of ~~Se-NMR, relarive high concentrations are needed. Since
ebselen is only poorly solvable in aqueous solutions, the NMR experiments were conducted in
DMF,similar to experiments by Fischer and Dereu (10). When L(SH)2 was added to ebselen,
only the diselenide of ebselen could be detected by ~~Se-NMR. No other products or
intermediates were observed (fig 9). Comparable to the HPLC data, this technique also showed
that one mole of L(SH)2 consumed two moles of ebselen,
The formarion of the diselenide in the reacrion between ebselen and L(SH)2, was confirmed
using direct chemical ionization-mass spectrometry. To 1 ml of a solution of 18 mol ebselen in
DMSO,2 ml of a solution of 9 mol L(SH)2 in 10 mM potassium phosphate buffer, pH 7.4,
was added. A sample of this reaction mixture was immediately analyzed, and it was found that
the positive chemical ionization-mass spectrum of the formed product (fg 10) had the same
characteristic as the mass-spectrum obtained from a reference sample of the diselenide (not
shown).The cluster of ions at m/z = 545 - 557 correspond to the calculated pattern of isotopes
of a compound with the same chemical composition as the protonated diselenide
(C26H21N202Se2). T'he base peak of the ion at m/z = 207, which is derived from lipoic acid,
also indicated that L(SH)2is oxidized in a reaction with ebselen.
To answer the question why L(SH)2 is a better cofactor than GSH in the reaction with
ebselen, two types of experiments were conducted. In the first experiment the effect of GSH
addition on the removal of hydrogen peroxide by the combinarion of L(SH)2 and ebselen was
determined. To the incubation medium fust GSH (1 mM)was added, then ebselen (10 M)and
subsequently L(SH)2(0.5 mM). The reaction was started by the addition of hydrogen peroxide
(0.5 mM).It was found that GSH did not reduce the L(SH)2-supported peroxidase activity of
ebselen (data not shown).
In the second experiment, 50 M GSH was added- to 50 M ebselen which resulted in a

.1

`~,

~- 40
0
...
~.
v 20
U
F.
O
U

10

20

30

40

[dihydrolipoateJ (M)

Fig. 8 Formation of the diselenide () from 50 M ebselen () and a varying concentration of


dihydrolipoate;(+)is the concentration of free salihydryl groups. The reaction mixtures were analyzed 15
sec after dihydrolipoate was added using HPLC.

1000

500

ppm
Fig. 9 Reaction between ebselen and dihydrolipoate in Bimethyl formamide monitored using ~~Se-NMR
spectrometry. The concentration of ebselen was 12 mM. In spectrum A the concentration of
dihydrolipoate was 3 mM,in spectrum B the concentration was 6 mM. The signal at 959 ppm is due to
ebselen, the signal at 472 ppm to the diselenide of ebselen.

77

207

100

553
551 ~

80
549
60
sss
40

198
~

547
~

276
263 ~ 278
r

550 560
540
/
\
473
~~ 553 ~/

20
189

557

245 II I~u' r2 334

\JL/
,7
200

300

400

500

600

m~z
Fig. 10 Mass spectrum of the products of the reaction between ebselen and dihydrolipoate. To 18 mol
ebselen in 1 ml DMSO,2 ml of an aqueous solution of 9 mol dihydrolipoate was added. After the
addition of dihydrolipoate, a sample of the reaction mixture was directly introduced into the mass
spectrometer. In the insert, the calculated mass spectrum of a compound with the chemical composition of
the protonated diselenide is depicted.
rapid formation of the GSH-selenenyl sulfide of ebselen (fig 5). When 15 seconds after the
addition of GSH, L(SH)2 was also added, it was found that the diselenide of ebselen was
immediately formed (fig 11). Apparently, GSH is less effective than L(SH)2 in the generation
of the diselenide from the GSH-selenenyl sulfide of ebselen (fig. 5,6).
In order to elucidate the mechanism of the reaction between L(SH)2 and the GSH-selenenyl
sulfide of ebselen which immediately forms the diselenide, the interacrion of ebselen with the
synthetic thiol, mercaptoethanol (MSH), was invesrigated. Similar to GSH, MSH reacted
rapidly with ebselen, one mole of MSH consuming one mole of ebselen and the major product
being the MSH-selenenyl sulfide of ebselen. When various concentrations of MSH were used
and the concentration of ebselen was kept constant, and the reaction mixture was analyzed
immediately after the addition of MSH by HPLC, the results were identical to those obtained
with GSH as depicted in fig 5, be it that the MSH-selenenyl sulfide instead of the GSHselenenyl sulfide was formed. Diselenide formation followed second order rate kinetics with
respect to the concentrarion of the MSH-selenenyl sulfide of ebselen and the concentration of
free MSH (rate constant 2.2 0.1 mM-1 min-1).
When 50 M GSH was added to 50 M ebselen, all ebselen was immediately converted into
the GSH-selenenyl sulfide. As described above, diselenide formation with equimolar amounts

78

..1
.-.
~ ~
0
..,
cd
~.

~ 2O
0
U

10

20

30

40

[dihydrolipoate](M)
Fig. 11 Formation of the diselenide(~)from the GSH-selenenyl sulfide(~)of ebselen and a varying
concen~ation of dihydrolipoate. The GSH-selenenyl sulfide is generated by the addition of 50M GSH to
50 M ebselen. After 15 seconds, dihydrolipoate in the given concentrations was added. Subsequently after
another 15 seconds, the reaction mixture was analyzed using F3PLC; (+) is the concentration free
sulthydryl groups detected.

of GSH and ebselen was slow. Addition to this incubation mixture of 50M MSH,15 seconds
after the addition of GSH, yielded the MSH-selenenyl sulfide of ebselen and also enhanced
diselenide formation (fig 12).
Comparable results were obtained when MSH (50 M) was added to ebselen (50 M) 15
seconds before GSH(50 M). Approximately half of the MSH-selenenyl sulfide was converted
into the GSH-selenenyl sulfide, and diselenide formarion was faster compared to the incubation
system without MSH. Also when GSH (50 M) and MSH (50 M) were mixed and the
reaction was started by the addition of ebselen (50 M), the compounds detected in the
incubation mixture were the diselenide and the selenenyl sulfides of ebselen with both MSH and
GSH. When a combination of MSH and GSH was used, the concentrations selenenyl sulfide of
ebselen with either GSH or MSH were independent of the way the reaction was started.
Diselenide formation was fastest with 100 M GSH and slowest with 100 M MSH. An
intermediate rate of diselenide formarion was observed when both 50 M GSH and 50 M
MSH were added to the reacrion mixture (iig 12).
Discussion.
Ebselen is a selenium containing heterocyclic compound, which displays anti-inflammatory
activity (6, 8). Like the endogenous selenium-dependent GSH-eeroxidase, ebselen catalyzes the
reaction between GSH and peroxides, and this catalyric activity is probably linked to the
therapeutic effect of ebselen (4-5). In general, GSH is regarded as the cofactor for the
eeroxidase activity of ebselen. Remarkably, in the present study we found that the
endogenously occurring thiol dihydrolipoate (L(SH)2)is a better cofactor than GSH in the
pero~dase acrivity of ebselen (fig 2). In contrast to GSH,L(SH)2 did not react spontaneously
with hydrogen peroxide. However, in the ebselen catalyzed reduction of hydrogen peroxide,
L(SH)2 proved to be far more effective than GSH (fig 2). Our results indicate that with either
79

.~~

~
~

.~~

50
~

_ _

~'

.~

~
a>

25

U
O
U

~
15
~~

f~

E
~
.. _ "

..~
,..

.~ 10

~ E
o ,~

5
~~

C
D
incubation system

Fig 12 Reaction of the combination of GSH and MSH with ebselen. In panel A the concentration GSHselenenyl sulfide of ebselen ([ebselen-GSH adduct]), the concentration MSH-selenenyl sylfide of ebselen
([ebselen-MSH adduct]) and the concentration of diselenide ([diselenide]) aze shown. In all incubation
systems, except E,fust 50 M ebselen was added. In incubation system A 100 M GSH was added; in
incubation system B 100 M MSH was added; in incubation system C 50 M GSH and 15 sec thereafter
SO M MSH was added; in incubation system D 50 M MSH and 15 sec thereafter 50 M GSH was
added. In incubation system E 50 M GSH and 50 M MSH were added, and 15 sec thereafter 50 M
ebselen was added. The reaction mixtures were analyzed with HPLC 15 sec after the last addition. In panel
B the rate of disappearance of the GSH-selenenyl sulfide of ebselen (k ebselen-GSH adduct), the rate of
disappearance of the MSH-selenenyl sulfide of ebselen (k ebselen-MSH adduct) and the rate of diselenide
formation(k diselenide)for the same incubation systems as in panel A are depictec.

GSH or L(SH)2, ebselen is reduced to its diselenide, but the rate of this reaction differs
considerably for bath thiols.
In a recently proposed scheme by Fischer and Dereu (10), the reaction between ebselen and
thiols in the first instance yields a selenenyl sulfide. Subsequently, the selenenyl sulfide slowly
undergoes a cleavage reaction to the diselenide and disulfide. This reaction was supposed to be
an equilibrium, the equilibrium constant depending on the nature of the thiol and the pH of the
medium and the reaction being independent of any selenol intermediate (10).
In accordance to the mechanism proposed by Fischer and Dereu (10), we found that in the
reaction between GSH and ebselen, selenenyl sulfide formarion was fast, while the diselenide
formarion was slow (fig 5). However, we also found that diselenide formation proceeds via a
second order reaction, since the rate of diselenide formation was linearly dependent on the
concentrarion of free GSH and the concentration of selenenyl sulfide (fig 6). The second order
rate kinerics do not fit the equilibrium reaction between the selenenyl sulfide and the diselenide
80

0
N
2 ~ \
i
~ Se
Ebselen

~ /
H2O
2 GSH

kl

ks
\ /
I~
H
~

~I
~

SeOSe

~\
2 ~ /

\ /

Selenenic acid anhydride

~
Se S

Selenenyl sulfide
H2O

k4
xzo2

GSH

k2
GSSC

O
\
~~
~ 5eH
Selenol

~
/

k3

~ ~

~ ~
N
H

N
H
Se

/
\

Se
Diselenide

Fig 13 Proposed reaction scheme for the GSH-mediated peroxidase activity of ebselen. The reaction
between GSH and the GSH-selenenyl sulfide or the reaction between H2O2 and the diselenide is rate
determining in the GSH-supported eeroxidase activity of ebselen. In contrast to the scheme presented by
Fischer and Dereu (10), a selenol intermediate is suggested to be involved.
presented by Fischer and Dereu (10). Addirionally, we did not observe a diminished diselenide
formation with GSH when GSSG was added, which also pleads against the equilibrium
proposed by Fischer and Dereu (10).
Based on the present data, we propose a modified reaction scheme for the GSH eeroxidase
activity of ebselen, as depicted in fig 13. In this mechanism the nucleophile GSH attacks the
selenium atom of ebselen, and substitutes the amide nitrogen under the formation of a selenenyl
sulfide. Subsequently, the sulfur atom of the selenenyl sulfide is attacked by a second GSH
molecule under the formarion of an as yet unidentified selenol and GSSG. Of both subsriturio~
reactions by GSH, the second is relatively slow (second order rate constant 6.1 0.1 mM-1
min-1)and can be rate deternuning in the eeroxidase activity of ebselen. The secnd order
kinetics of this reaction indicate that it involves an SN2-type substiturion reaction in which the
selenol of ebselen is the leaving group. Since no selenol could be detected, it is likely that the
selenol reacts rapidly with a second selenenyl sulfide. By this nucleophilic attack of the selenol
at the selenium atom of the selenenyl sulfide, GSH is released from the selenenyl sulfide and a
diselenide of ebselen is formed.
In accordance to Fischer and Dereu (10),in the conversion of the GSH-selenenyl sulfide into
the diselenide of ebselen the net reacrion is that two selenenyl sulfides form one diselenide and
GSSG. In contrast to the scheme presented by Fischer and Dereu (10), the diselenide is not
81

2 I~

,rr ~ ~ ---~ I

Se O Se
Diltydrolipoate

Selenenic acid anhydride

O
~
~

Se
S

~ ~

~ Se
Ebselen

Dihydrolipoate

o~

xzo

H2O

~
Se

SH

HS `
S
COO-

Selenenyl sulfide

COOSelenenyl sulfide
HZOZ

Lipoate

Lipoate
O
NH ~ ~
~ SeH

\ ~N ~ ~
I /

Selenol

H
Se

Selenenyl
sulfide

Dihydrolipoate

~ ~ N ~ /
\

Se
Diselenide

Fig 14 Proposed reaction scheme for the dihydrolipoa[e-mediated eeroxida


se activity of ebselen. The
reaction between H2O2 and the diselenide is rate determining in the dihydrol
ipoate-supported eeroxidase
activity of ebselen. In the formation of the diselenide from the selenol, only
the reaction of the selenol
with the dihydrolipoate-selenenyl sulfide is depicted. Alternatively, the diseleni
de might also be forrned in
a reaction of the selenol with ebselen, a reaction omitted for shake of clarity .

generated from the selenenyl sulfide by the equilibrium reaction as propos


ed by them, but by
two SN2-type nucleophilic substitutions; %rstly by GSH at the sulfur atom
of the selenenyl
sulfide with the formation of a selenol intermediate, and subsequently
by this selenol at the
selenium atom of a second selenenyl sulfide in which the diselenide of
ebselen and GSSG are
formed. GSH is thus a catalyst in formarion of the diselenide from the selenen
yl sulfide. In both
nucleophilic substiturion reactions, the deprotonated thiol or selenol is most
likely the reacrive
species (fig 7)(19). Finally, the diselenide reacts with hydrogen peroxi
de and ebselen is
regenerated. A selenenic acid anhydride has been reported to be an intermediate
in this reaction
(10). Recently, Maiorino et al.(20) were able to trap a selenol intermediate with
the electrophile
iodoacetamide in the reaction between GSH and ebselen, which.further support
s the modified
reacrion scheme we present (fig 13).
When L(SH)2 was used instead of GSH in the reaction with ebselen,
a rapid diselenide
formation was found (fig 8). That indeed the diselenide was formed,
was cnfirmed using
~~Se-NMR (fig 9) and mass spectrometry (fig 10). Previously, Muller al.
et (17) generated a
mass spectrum of the diselenide using FAB. We applied direct chemical
ionisation, that
appeared to be a more convenient technique for the diselenide. In the direct chemic
al ionization
mass specrium, the cluster of ions with m/z 545 - 557 perfectly matched the theoret
ical isotope
distribution of a compound with the chemical composirion of the protonated
diselenide (fig 10).
Based on these results and in analogy with the reacrion between ebselen and
GSH, we
82

propose a reaction scheme for the L(SH)2-supported peroxidase activity of ebselen (iig. 14) in
which firstly ebselen reacts with the dithiol L(SH)2 to yield a selenenyl sulfide intermediate. In
contrast to the reacrion with GSH,the L(SH)2-selenenyl sulfide rapidly forms a selenol (fig 8)
probably because a free sulfhydryl group is available in the L(SH)2-selenenyl sulfide that
functions as an intramolecular nucleophile. Subsequently, the formed selenol reacts with
ebselen or with the L(SH)2-selenenyl sulfide with the formation of the diselenide of ebselen.
Sunilar to the reaction between ebselen and GSH,the reaction between ebselen and L(SH)2 is
most likely a nucleophilic subsritution by one of the sulfhydryl groups of L(SH)2 at the
selenium atom of ebselen. In principle either one of the sulfhydryl groups of L(SI-~2 may attack
the selenium atom. As describec above, the deprotonated thiol is most likely the nucleophile
involved in this type of substitution reactions. In general it is known that both the extent of
ionization of a thiol, as well as the intrinsic nucleophilicity of the corresponding thiolate anion
deternune the overall reactivity of thiols in this type of nucleophilic reactions. It has been
suggested that the lower the pKa of a thiol, the lower the nucleophilicity of the thiolate, but the
higher the relarive concentration of thiolate is (19, 21, 22). When the effect of the pH on the
reactivity of thiols in SN2 reactions was studied, it was found that of the two opposite effects of
the pKa on the reactivity, the effect of the pKa on the fracrion thiol deprotonated contributes
more, if the pH < pKa (21, 22). This means that the thiol in L(SH)2 with the lower pKa would
be the better nucleophile in our experimental set-up.
Preliminary results indicated that the pKa of both thiol groups in L(SH)2 are virtually
identical, indicating that the nucleophilicity of the respecrive thiolate anions and fraction ionized
of both thiols do not differ very much. We also found that both sulfhydryl groups influence
each other, deprotonarion of one thiol increased the pKa of the other sulfhydryl group. A
comparable ambiguity of the pKa value of other thiols is known. Upon protonarion of its amine,
the pKa of the thiol in cysteine decreased from 10.6 to 8.6 (23). It should also be noted that
sterical hindrance might reduce the nucleophilic reacrivity of the secondary thiol in L(SH)2.
Since the two different L(SH)2-selenenyl sulfides of ebselen were too reactive to be isolated and
analyzed, we were not able to determine the relarive contriburion of each thiol in the reacrion
with ebselen. Moreover, intramolecular conversion of one selenenyl sulfide into the other comparable to the intramolecular conversion of 8-S-acetyldihydroliponamide into 6-Sacetyldihydroliponamide and vice versa (24)-might hamper such an attempt. As a result of
substitution to the selenium atom of ebselen by one of the thiol groups of L(SH)2, the pKa of
the other, free thiol in the L(SH)2-selenenyl sulfide of ebselen will probably drop. If the
presumed relation between pKa and nucleophilicity holds, the free thiol group would be'
rendered into a stronger nucleophile, which would enhance the nucleophilicity of this thiol, and
hence the rate of selenol and subsequent diselenide formation.
The large difference in the rate of diselenide formarion in the reaction between ebselen and
GSH or L(SH)Z accounts for our observation that L(SH)Z is a better cofactor than GSH in the
peroxidase activity of ebselen. When L(SH)2 is used instead of GSH,selenol formarion is not
rate limiting anymore. With L(SH)2 as cofactor, hydrogen peroxide removal is governed by the
second order reaction between the diselenide and hydrogen peroxide (fig 3). In this case the rate
of peroxide removal(V)can be described by the following equarion:
- d[H2O2]
V

dt

= 12.3 103[H2O~ [Ebs] (M-1 miri1)


83

In this equation,[Ebs] is the concentration ebselen added to the incubation mixture.


According to our reaction scheme for the GSH-mediated penoxidase activity of ebselen (fig
13), the rate of peroxide removal with GSH as cofactor can be described by the following
equation (see appendix):
k2[GSH]
2
2 -1
1
~
+ k [Gs~ + k ~ O ]+ k
[Ebs] k
~ k1 [GSH]
3
2
4 2 2
5~ ~
2

Because in diselenide formation, the reaction between ASH and the selenenyl sulfide is rate~, limiring (i.e. kl > 2k2 and k3 > k2[GSH]) and because the reaction between the diselenide and
hydrogen peroxide can be described by a second ordzr reaction (i.e. ks ~ 1~[H2O2]) the
equarion can be simplified to:
1

-1

~ {k2[GSH] + k4[H2O2]}

[Ebs]

In this equation, k2 is half of the overall rate of selenenyl sulfide consumption as calculated in
fig 6, thus k2 = 6.1 0.4 mM-1 min-1, and kq is twice the second order rate constant that
describes. penoxidase activity with ebselen and L(SH)2, thus kq is 24.6 1.6 mM-1 min-1.
Interestingly, this simplified equation is comparable to the Dalziel equation that describes
peroxide removal by the endogenous selenium dependent GSH-penoxidase (fig 1)(25). The
reaction between ebselen.and either substrate, hydroperoxide or GSH,can be rate limiring. The
Km and Vmax for these substrates are not constant, the apparent kinetic coefficients of one
substrate are a function of the concentration of the other substrate. The Dalziel constants of
ebselen for GSH(= 1 /k2)and hydrogen peroxide(= 2/k4)found in this study are respecrively
0.165 0.011 mM min and 0.081 0.(}05 mM min at pH 7.4, 37C. Recently, Maiorino et al.
(20) also reported that the ebselen catalyzed reaction between GSH and peroxides can be
described by a ter uni ping-pong mechanism. They stated that the Dalziel coefficients of ebselen
with hydrogen peroxide as substrate were 0.66 mM min for GSH and 0.015 mM min for
hydrogen peroxide. They also determined the Dalziel coefficients of ebselen with other
hydroperoxides than hydrogen peroxide as substrate. Surprisingly, they found that the Dalziel
coefficient for GSH depended on.the hydroperoxide used, the coefficient varied from 1.4 mM
min to 0.073 mM min (20). In principle, this Dalziel coefficient describes the reacrion between
GSH and ebselen, a reaction that is independent of the hydroperoxide used. For comparison,
the Dalziel coefficient of the selenium dependent GSH-penoxidase for GSH is indeed
independent of the hydroperoxide used (25). The Dalziel coefficient of ebselen for GSH we
found is within the range Maiorino et al.(20)reported.
Our data clearly show that the L(SH)2-mediated formarion of the diselenide of ebselen is
much faster (fig 8) when compared to the corresponding GSH-supported reaction (fig 5, 6). At
first sight, the most likely explanation for this difference is the availability of a second
intramolecular nucleophilic sulfhydryl group in the L(SH)2-selenenyl sulfide of ebselen, in the
vicinity of the electrophilic sulfur atom attached to selenium in the selenenyl sulfide.
Alternatively, L(SH)2 might be a better cofactor than GSH in the penoxidase activity of ebselen
simply because L(SH)2 is a better reductor. The redox potential of the couple L(SH)2 -lipoate is
-0.32 V, versus -0.24 V of the couple GSH-GSSG (19). However, it should be noted that just

84

xs
~

~\ /

~ ~
1

~ SeSG
XSH

2a

~ \ /

~ SeH

2b

xssx

0
GSH

~ \ /

XSH

~ Se-SX

Fig 15 Possible reacrions of thiois (XSij with the GBH-selenenyl sulYide of ebselen. 'Thiols may attack
at the sulfur atom of the GSH-selenenyl sulfide and form direcfly a selenol (reaction 1). Alternatively,
thiols may attack at the selenium atom of the GSH-selenenyl sulfide (reaction 2a). The formed XSHselenenyl sulde can be converted into a selenol by an attack of a thiol at the sulfur of the formed
selenenyl sulfide (reaction 2b). For MSH reactions 1 and 2b proceed slow, while reaction 2a is fast, For
dihydrolipoate reaction 2b is fast, and also selenol formation from the GSH-selenenyl sulfide of ebselen is
fast. The latter effect is probably due to a fast thiol exchange of dihydrolipoate wth GSH in the GSHselenenyl sulfide, comparable to MSH, and subsequent fast selenol formation from the dihydrolipoateselenenyl sulfide of ebselen (reaction 2a + 2b ).

because a compound is a better reductor, this does not imply that redox reacrions with this
reductor go faster compared to reactions with a less potent reductor. For example, in the
spontaneous reduction of hydrogen peroxide, the activation energy of the redox reaction with
L(SH)2 proved to be higher than that with GSH (fig 2).
In vivo both GSH and L(SH)2 are present. GSH might reduce the L(SH)2-mediated activity
of ebselen because ebselen might be entrapped as GSH-selenenyl sulfide. Therefore the effect
of GSH on the L(SH)2-mediated peroxide removal by ebselen was also determined. The
contribution of each thiol might not be determined by the overall reaction rate of diselenide
formation out of ebselen by each thiol, but by the reaction rate of ebselen with each of the thiols,
a reaction that is fast for both thiols. In fact, the pKa of the sulfhydryl group of GSH (8.6-8.9)
(19)is lower than the pKa of both the sulfhydryl groups of L(SH)2(preliminary results indicate
that the pKa for both thiol groups, when the other thiol group is not deprotonated, is 9.4). This
implies that at pH 7.4 more GSA than L(SH)2 is deprotonated and thus the reaction between
ebselen and GSH would be faster than the reaction with L(SH)2. Moreover, in our experiment
GSH was added before L(SH)2, so the formation of the GSH-selenenyl sulfide was favored,
and also during the actual peroxidase acrivity the GSH-selenenyl sulfide might be formed. It
was found that GSH did not reduce the L(SH)2-supported penoxidase activity of ebselen.
-The rate offormation of the diselenide from the GSH-selenenyl sulfide of ebselen by L(SH)2
was also determined. It was found that immediately after the addition of L(SH)2 the diselenide
was formed (fig 11). L(SH)2 was far more effective than GSH in the generation of the
diselenide from the GSH-selenenyl sulfide of ebselen (fig. 5, 6). GSH does not affect the
L(SH)2-mediated penoxidase activity of ebselen, because the GSH-selenenyl sulfide of ebselen
is rapidly converted into the diselenide by L(SH)2. In principle, there are two possible
nucleophilic substitution mechanisms for the reaction between L(SH)2 and the GSH-selenenyl
sulfide of ebselen (fig 15). Firstly, by an attack at the selenium atom, L(SH)2 might substitute
for GSH in the GSH-selenenyl sulde of ebselen (fig 15, reaction 2a). The formed L(SH)2selenenyl sulfide of ebselen is rapidly converted into the selenol intermediate (fig 15, reaction
2b). Secondly, L(SH)2 may attack at the sulfur atom of the GSH-selenenyl sulfide (fig 15,
SS

reaction 1). In that case, the selenol and a mixed disulfide of GSH with L(SH)2 would be
formed. The mixed disulfide is probably rapidly converted into lipoate and GSH (13). Both
nucleophilic substitutions give rise to a selenol, and after a nucleophilic substitution of the
selenol at the selenium of a second GSH-selenenyl sulfide of ebselen, the diselenide is formed.
Nucleophilic attack of L(SH)2 at the GSH-selenenyl sulfide of ebselen might occur either on
the selenium or on the sulfur atom (fig 15). A selenol anion is a better leaving group compared
to a thiolate anion, however, nucleophilic substitution at dicoordinate selenium is much faster
than nucleophilic substitution at dicoordinate sulfur (26). The latter difference between selenium
and sulfur is also illustrated by the fact that the thiols did not react with PZ 25, the sulfur analog
of ebselen, while the reaction of the thiols with ebselen proved to be fast (fig 5). For bisalkylthio-selenides the two opposite factors on the reactivity of selenitam compared to the
reactivity of sulfur appear to cancel each other out. In this case a factor like the steric bulk of the
alkyl groups determines whether the sulfur or the selenium is attacked (26). To determine which
mechanism prevails in the case of L(SH)2 and the GSH-selenenyl sulfide of ebselen,
mercaptoethanol(MSH)was used as a tool.
Like GSH, MSH was found to react rapidly with ebselen to form a selenenyl sulfide.
Diselenide formarion depended on the concentrarion of free MSH and the second order rate
constant for diselenide formation with MSH (2.2 0.1 mM-1 min-1)was lower than with GSH
(6.1 0.4 mM-1 min-1 ). In contrast to the L(SH)2-selenenyl sulfide of ebselen, the MSHselenenyl sulfide of ebselen is stable. Therefore, the reaction between MSH and the GSHselenenyl sulfide of ebselen can be used to elucidate the mechanism of the reaction between
L(SH)2 and the GSH-selenenyl sulde of ebselen.
When ebselen, MSH and GSH were present at equimolar concentrations, the selenenyl
sulfides of both thiols were detected. The ratio of the concentration of the MSH-selenenyl
sulfide versus the concentration of the GSH-selenenyl sulfide proved to be independent of the
sequence of addirion (fig 12). This indicates that virtually immediately both selenenyl sulfides
are in equilibrium. So the nucleophilic substitution on the selenium of the selenenyl sulfide was
fast for both thiols tested. In equilibrium, approximately half of the selenenyl sulfide consisted
of the MSH-selenenyl sulfide of ebselen, indicating that MSH was equally effective as a
nucleophile at the selenium of the GSH-selenenyl sulfide, as GSH at the selenium of the MSHselenenyl sulfide. The fact that GSH was a better nucleophile than MSH -the pKa of the
sulfhydryl group of GSH(pKa = 8.6-8.9) is lower than the pKa of MSH(pKa = 9.6)(19) - is
probably leveled out by the fact that GSH is also a better leaving group than MSH.
With GSH or MSH, diselenide formarion from ebselen was slow with either one of the
thiols alone or with the combination of both. With GSH (100 M) alone, second order rate
constant was higher than with MSH(100 M)alone. Apparently, GSH is a better nucleophile at
the sulfur of the GSH-selenenyl sulfide, than MSH at the sulfur of the MSH-selenenyl sulfide.
The difference between GSH and MSH in diselenide formarion can be explained by reasoning
that GSH was probably the better nucleophile under the incubation conditions used, and that the
sulfur of the GSH-selenenyl sulfide might be a better electrophile. In the reaction of either GSH
or MSH at the sulfur of their corresponding selenenyl sulfide, the leaving group is identical,
namely the selenol.
In the reacrion between MSH and the GSH-selenenyl sulfide of ebselen, exchange of thiols
is fast while diselenide formarion is slow (fig 12). This indicates that MSH reacts primarily at
the selenium of the GSH-selenenyl sulfide of ebselen (fig 15, reaction 2a). When these results
86

are extrapolated to the reacrion between L(SH)2 and the GSH-selenenyl sulfide of ebselen, the
first step is probably the nucleophilic substitution of GSH by L(SH)2 on the selenium (iig 15,
reaction 2a). The formed L(SH)2-selenenyl sulfide of ebselen rapidly forms a selenol (iig 15,
reaction 2b). Diselenide formarion finally proceeds by a nucleophilic substiturion by the selenol
at the selenium of another GSH-selenenyl sulfide. The proposed scheme for the reaction of
L(SH)2 with the GSH-selenenyl sulfide of ebselen implies that L(SH)2 is a better cofactor than
GSH in the peroxidase activity of ebselen, not because L(SH)2 is a better reductor than GSH (in
that case reaction 1, fig 15 would be preferred), but because of the vicinal thiol groups in
L(SH)2 that can form an intramolecular disulfide.
In vivo, more GSH than L(SH)2 is available (19). However, as shown in this study, in the
peroxidase acrivity of ebselen, L(SH)2, at a concentration of 5M, was found to be more
potent than GSH in the mM range (fig 2, 3). Moreover, GSH does not reduce the L(SH)2 mediated peroxidase activity of ebselen. It should be noted however, that in vivo most L(SH)2
is bound in amide linkage to the s-amino group of a lysine residue. In its biological function,
L(SH)2 shuttles between the oxidized and reduced form (27). In the reaction of bound lipoate
with pyruvate, fust pyruvate is cleaved by a carboxylase to yield CO2 and an enzyme bound
"active aldehyde". Subsequently the "arrive aldehyde" is believed to attack the disulfide linkage
of bound lipoic acid in a nucleophilic displacement reaction,followed by a reverse condensarion
(27). It is presumed that the acetyl group is attached to the primary SH group of bound
dihydrolipoic acid (24). The acetyl group is transferred from L(SH)2 to Co-enzyme A, a
reaction catalyzed by lipoic acid transacetylase. Finally L(SH)2 reduces NAD+ to NADH and
lipoate. The latter reaction is catalyzed by a dihydrolipoic dehydrogenase, and it -has been
reported that this reaction is freely reversible with high reaction rates in both directions (27).
The reaction of pyruvate or NADH with membrane bound lipoate might provide reductive
equivalents for the peroxidase activity of ebselen.
Recently, it has been reported that ebselen totally blocks in vivo Sephadex-induced lung
edema in rats (28). Co-administration of GSH reduced the protective effect of ebselen. It has
been stated that this is due to the formation of the GSH-selenenyl sulfide of ebselen. It was
suggested that the formed selenenyl sulfide had a much higher aqueous solubility than the parent
ebselen, and the loss of acrivity of ebselen was ascribed to the enhanced clearance of ebselen, as
selenenyl sulfide, from the animal (28). As demonstrated in this study, L(SH)2 is capable of
converting the GSH-selenenyl sulde of ebselen rapidly into the diselenide (fig 11). Inferred
from the fact that the retenrion rime of the diselenide is longer than that of ebselen with reversed
phase HPLC,the diselenide is probably less hydrophilic than ebselen. Therefore, conversion of
ebselen into the diselenide probably will not enhance renal clearance. As already mentioned,
L(SH)2 is a better cofactor than GSH in the peroxidase activity of ebselen. Additionally,
L(SH)2 might have an important contribution in the therapeuric effect of ebselen by prevenring
the enhanced clearance of ebselen as GSH-selenenyl sulfide.
Acknowledgements.
We like to thank F.J.J. de Kanter and B. van Baar (Department of Organic and Anorganic
Chemistry, Free University, Amsterdam) for their technical assistance in the NMR and MS
experiments.

87

References.

their
1 Bast, A and G.R.M.M. Hamen. Cytochrome P-450 and glutathione: What is the significance of
interrelationship in lipid peroxida[ion. Trends Biochem. Sci. 9:510-513(1984).
2 Haenen, G.R.M.M., N.P.E. Vermeulen, H. Timmerman and A. Bast. Effect of several thiols on lipid
peroxidation in rat liver microsomes. Chem. Biol. Interact. in press (1989).
LTB4 in
3 Kuhl, P., H.O.Borbe, H. Fischer, A. Rtimer and H. Safayhi, Ebselen reduces the formation of
48
31:1029-10
ins
human and porcine leukocytes by isomerisation to its SS-12R-6-trans-isomer, Prostagland
(1986).
4 Wendel, A., M. Fausel, H. Safayhi, G. Tiegs and R. Otter. Activity of PZ 51 in relation to glutathione
peroxidase. Biochem.Pharmacol. 33:3241-3245(1984).
5 Muller, A., C. Cadenas, P. Graf and H. Sies, Glutathione eeroxidase-like activity n vitro and antioxidant
capacity ofPZ 51. Biochem Pharmacol. 33:3235-3239(1984).
of
6 Pamham, M.J. and S. Kindt. Effects of PZ 51 (ebselen) on glutathione eeroxidase and secretory activities
(1984).
mouse macrophages. Biochem. Pharmacol. 33: 3247-3250
mechanism.
7 Floh, L., G. Loschen, W.A. Gunzler and E. Eichele. Glu[athione eeroxidase V, the kinetic
(1971).
Hoppe-Seyler's Z. Physiol. Chem. 353:987-999
pathological
8 Pamham,M.J. and E. Graf. Seleno-organic compounds and the therapy of hydroperoxide-linked
conditions. Biochem. Pharmacol. 36:3095-3102(1987).
(D.W.
9 Floh, L. Glutathione eeroxidase, fact or fiction in: Oxygen Free Radicals and Tissue Damage
(1979).
95-122
,
Amsterdam
Fitzsimons, ed.) Ciba Foundation Symposium 65(new series)Exerpcia Medics,
A selenium 10 Fischer, H. and N. Dereu. Mechanism of the catalytic reduction of hydroperoxides by ebselen:
77 study. Bull. Soc. Chim. Belg. 96:757-768 (1987).
ADP11 Muller, A., H. Gabriel and H. Sies. Protective glutathione-dependent effect of PZ 51 (ebselen) against
(1985).
89
34:1185-11
Pharmacol.
Biochem.
.
hepatocytes
Fe induced lipid peroxidation in isolated
and
12 Cotgreave, I.A., M.S. Sandy, M. Berggren, P.W. Moldus and M.T. Smith. N-Ace[ylcysteine
isolated
in
cytotoxicity
ced
digaat-indu
against
(ebselen)
51
PZ
of
effect
giutathione-dependent protective
hepatocytes. Biochem. Pharmacol. 36:2849-2904(1987).
protection against
13 Bast, A. and G.R.M.M. Hamen. Interplay between lipoic acid and glutathione in the
1
(1988).
9b3:558-56
Acts
Biophys.
Biochim.
n.
microsomal lipid peroxidatio
microsomal lipid
14 Scholich, H., M.E. Murphy and H. Sies. Antioxidant activity of dihydrolipoate against
61
(1989).
1001:256-2
Acts
Biophys.
Biochim.
l.
a-tocophero
on
dependence
peroxidation and its
ependent
15 Hildebrandt, A.G. and I. Roots. Reduced nicodnamide adenine dinucleotide phosphate (NADPI~-d
in liver
reactions
oxidation
function
mixed
during
peroxide
hydrogen
of
breakdown
formation and
microsomes. Arch. Biochem. Biophys. 171:385-397(1975).
16 Ellman, G.L. Tissue sulfhydryl groups. Arch. Biochem. Biophys. 82:70-77(1959).
mation of ebselen in
17.Muller, A, H. Gabriel, H. Sies, R. Terlinden, H. Fischer and A. Romer. Biotransfor
perfused rat liver. Biochem. Pharmacol. 37:1103-1109(1988).
of ebselen in human
18.Terlinden, R., M. Feige and A. Rtimer, Determination of the two major metabolites
plasma by high-performance liquid chromatography. J. Cromatography 430:438-442(1988).
19 Jocelyn,P.C. The Biochemistry of the SH Group. Academic Press, London (1972).
substrate specificity of
20 Maiorino, M., A. Roveri, M. Coassin and F. Ursini. Kinetic mechanism and
glutathione eeroxidase activity of ebselen (PZ 51). Biochem. Pharmacol. 37:2267-2271 (1988).
mide. Biochem.
21 Lindley, H. A study of the kinetics of the reaction between thiol compounds and chloroaceta
J. 74:577-584 (1960).
ethylene oxide, J.
22 Danehy,J.P. and C.J. Noel. The relative nucleophil character of several mercaptans towazd
Am. Chem. Soc. 82:2511-2515 (1960).
groups in L-cysteine
23 Doornbos, D.A. and M.T. Feitsma, The acid strength of the sulfhydryl and ammonium
Weekblad 102:587-598
and D- penicillamine: The deterrnination of the micro acid stability constants,Phann.
(1967).
coli. Formation of 8-S24 Yang, Y.-S., and P.A. Frey. Dihydrolipoyl transacetylase of Escherichia
Acetyldihydrolipoamide. Biochemistry 25:8173-8178(1986).
Physiol.
W.A., H. Vergin, I. Muller and L. Floh, Glutathion-Peroxidase VI. Hoppe-Seyler's Z.
Giinzler,
25
Chem. 353:1001-1004 (1972).
site of nucleophilic attack on
26 Kice, J.L. and H. Slebocka-Tilk. Reactivity of nucleophiles toward and the
bis(alkylthio)selenides. J. Am.Chem. Soc. 104:7123-7130 (1982).
88

27 Reed, L.J. Chemistry and function of lipoic acid, in: Comprehensive Biochemistry Vol. 14 Biological
oxidants(M.Florkin and E.H. Stoltz Eds.) Elsevier, Amsterdam,99-126 (1966).
28 Cotgreave, I.A., U. Johansson, G. Westergren, P.W. Moldus and R. Brattsand. The anti-inflammatory
activity of ebselen but not thiols in experimental alveolids and broncholitis. Agents Actions 24:313-319
(1988).

Appendix.
Derivarion of the equation that describes the GSH-penoxidase activity of ebselen, according
to the mechanism depicted in fig 13. The derivation is based on the method used to describe the
penoxidase activity of the selenium-dependent GSH-penoxidase, according to Floh et al. (6).
The abbreviations used are: A = Ebselen; B = Ebselen-GSH adduct; C = Selenol; D =
Diselenide; E = Selenenic acid anhydride; T =total concentration of ebselen added; V =
penoxidase acriviry.
The reactions are:
(I)

~ B
A + GSH 1

(II)

k
B + GSH ~ C

(III) C + B

k
~ D + GSH

k
~ E
(N) D + H2O2 4

(~

GSSG

+ H2O

k
E ~ 2A + H2O

Based on this reaction scheme, the following equarions can be derived:


d~A~ =
t

2k5[El - ki [GSHI [Al

(1)

d[B]
[GSH][A] - k2[GSH][B] - k3[B][C]
dt kl

(2)

d[C]
[GSH][B] - k3[B][C]
dt k2

(3)

d[E]
[E]
dt k4[H2~2l CDA - k5

4)

The concentrarion ebselen added(T)is:

The penoxidase acriviry(V)is:


- d[H2O2]

In steady state conditions:


89

d[A] d[B] d[C] 2d[D]


dt
dt
dt
dt
gives:
(7)
and
(4)
of
Combination

~~~

2d[E] 0
dt

CEl = k4~2~2~ [D]


5

g)

Combination of(1),(7) and (8) gives:


E _ 2k4 CH2~21
2k 5
[~l
[l
kl [GSH]
k 1[GSH]
[A]

(9)

Combination of(2),(7),(3) and (9) gives:


~~ - k 1 [Al =
2k 2

(1~)

kq CH2O2] [D]
k2[GSH]

Combination of(3) and (7) gives:

3
Subsriturion of(8)-(11)in (5) gives:
T-

2 4 GS~2~ + k4 ~G~2~ + 2 + 2 k4 kH2~2~ j[D] + k 2[kSH]


k
3
5
2~ Hl
1~ ~

(12)

Substitution of(12)in (6) gives:


2 k 4[H2O2]
V - k4 ~2~2]

k 4[H2O2]

+ 2+

2 k4[H2O2]~-1
{T
ks

k 1[GS H] + k2[GSH]

k2[GS H]
}~13~
k
3

Equarion (13)can be simplified to:


2
v

90

T
~ k1 [GSH] + k2[GSH] +k4[H2O2]+ k5 ~ {

k2[GSH]
k3

PaptII Catecholamines in oxidative stress

91

92

- Chapter 10 Modulation of oxidative stress by catecholamines.


Abstract. Catecholamines are able to protect against lipid peroxidarion in rat liver microsomes.
This effect is probably due to anon-enzymatic scavenging of free radicals by the catecholmoiety of the catecholamines. In the protection against lipid peroxidation by the catecholamines,
semiquinone radicals and quinones are formed. These products inacrivated the GSH-dependent
protection against lipid peroxidarion and the calcium sequestration of liver microsomes. This
effect is probably due to arylation of free thiol groups of the enzymes responsible for the GSHdependent protecrion and calcium sequestrarion, namely the free radical reductase and calcium
ATPase. By catecholamines the injury produced by oxidative stress is shifted in first instance
from lipid peroxidation to sulfhydryl arylarion. The implicarions of the effect of catecholamines
on oxidative stress in brain, heart and liver is discussed.
Introduction.
Particularly aerobic life-forms produce continuously free radicals. Fortunately, cells have
several lines of defense against these reacrive molecules, and under normal conditions cells are
able to cope with radical production. However, several situations give rise to an unbalanced
radical production or oxidative stress [1]. It is generally accepted that oxidative stress is
involved in several pathophysiological processes [1]. One of the most important targets for free
radicals appears to be biomembranes. The peroxidation of poly-unsaturated membrane lipids,
known as lipid peroxidation, deteriorates all membrane functions. The barrier funcrion of the
membrane is impaired, the physical state of the membrane is altered, signal transducrion over
the membrane becomes hampered and all membrane-bound enzymes can be inactivated..
The role of catecholamines in oxidative stress is ambiguous. On the one hand it is known
that catecholamines are able to scavenge highly reacrive radicals as for instance the OH'radical
[2]. On the other hand there are some indications that catecholamines are able to induce or
aggravate injury by oxidative stress [3]. For example, (3-adrenoceptor overstimulation is
thought to produce heart injury [4, 5], probably partially mediated by oxidative stress [4-6].
Recently, it has been reported that catecholamines can also enhance OH' producrion [7].
Moreover, during the autoxidation of catecholamines reactive oxygen species [8] and
semiquinone radicals are generated [8-11]. At physiological pH, however, these radicals are
more likely to arise only through catalyzed catecholamine oxidarion reactions[5, 8-11]. Apart
from this, thiol reactive compounds like quinones are formed during catecholamine oxidarion
[9-12].
The effect catecholamines on oxidative stress is especially relevant in ischemia-reperfusion
injury, because at least part of the damage inflicted under this pathological condition is the result
of an enhanced production of free radicals, and at the same time catecholamines are being
released [5, 14]. The aim of this study is to determine the direct effects of catecholamines on
oxidative stress. To this end oxidative stress was forced upon rat liver microsomes. The
modulation by catecholamines of the effects of oxidative stress on these membranes was
determined.

93

Material and methods.


Chemicals. Reduced glutathione (GSI~,(-)isoproterenol HCI,(+)isoproterenol bitartrate, dopamine HCI,
Quint and chromatographically purifed xanthine oxidase(grade IIn were fmm Sigma(St. Louis, MO).Catechol
and a-methyl-dopa were from Janssen (Beerse, Belgium). 4-Methylcatechol was from Fluka(Buchs,Switserland).
4-Methyl-ortho-benzoquinone was synthetized according Willst~uer and Pfannenstiel [13] by the Ag20-mediated
oxidation of4-methylcatechol.
Methods. Male Wistar rats (Harlan Olec C.P.B., Zeist, The Netherlands) of 200-250 g were used. Liver
microsomes were prepared as described previously [15]. Lipid peroxidation was measured with the thiobarbituric
acid assay [15] and expressed as the absorbance at 535 nm versus 600 nm (4A535-600. Oxidative stress was
induced by (i} 50 M xanthine and 55 mU xanthine oxidase,(u) 1 mM H2O2 and 40 nM horseradish eeroxidase
or (iii) 10 M Fel+ and 0.2 mM ascorbate. The microsomes were suspended in a 150 mM Tris-HCI buffer, pH
7.4, containing 150 mM NaCI. In the experiment where the calcium sequestration was measured another buffer
was used(vide infra).
Pretreatrnent of the microsomes with (-)isoproterenol and xanthine-xanthine oxidase consisted of a incubation
of 10 min at 37C with 100 M (-)isoproterenol, 50 M xanthine and 55 mU xanthine-oxidase. The incubation
was terminated by the addition of one volume of ice-cold buffer. The microsomes were immediately centrifuged
twice with the buffer (40 min, 115,000 x g at 4 C). The pellet was resuspended in the buffer, and in order to
induce lipid peroxidation, incubated with 10 M Fel+and 0.2 mM ascorbate.
Electron spin resonance(ESR)spectra were recorded on a Varian E-103,equipped with an aqueos flat cell. It
was indicated when MgC12 was added. Instrumental settings were: modulation amplitude 1 G; time constant 30
sec; scan time 30 min, amplifier power, 25 mW;reciever gain 2.106.
Calcium sequestration was determined by incubating microsomes (1 mg/ml) in 115 mM KCI, 10 mM NaCI,
1 mM KH2POq,20 mM Hepes,4 mM MgC12, pH 7.2 at 37C in the presence of an ATP regenerating system
(10 mM creative phosphate and 10 U/ml creative kinase). Calcium sequestraon was started by the addition of2
mM ATP. Changes in the free calcium concentration were monitored by meas~uing Quint fluorescence(50 M
Quint) on a Perkin Elmer MPF-2A fluorescence spectrophotometer. An excitation wavelength of 339 nm (10
nm slit) and an emission wavelength of 500 nm (10 nm slit) were used. The concentration of free calcium was
within the physiological range (200-350 nivn. Calcium sequestered was quantified by releasing the sequestrated
calcium with ionomycin(2 ivn and adding known quantities of calcium into the cuvette.
Protein content was assayed according to Bradford [16] using bovine serum albumin as standard.

Results.
Oxidative stress induced by xanthine-xanthine oxidase (fig lA), H2O2-horseradish
eeroxidase (fig 1B) or iron-ascorbate (fig 1~) resulted in the peroxidation of rat liver
microsomes. Addition of the catecholamines (-)isoproterenol ,dopamine or a-methyldopa
resulted in aconcentration-dependent decrease of lipid peroxidarion,irrespective of the method
used to induce lipid peroxidation. Both the rate and the final extent of lipid peroxidation were
reduced by adding catecholamines (fig 1). The effect of(+)isoproterenol was identical to that of
(-)isoproterenol. The protection by catecholamines was non-enzymatic, since heating the
microsomes had no effect on the inhibition of lipid peroxidation by the catecholamines. Since
catechol and 4-methylcatechol offered the same protection as the catecholamines, it can be
concluded that the catechol moiety of the catecholamines is responsible for this effect. Of the
experiments performed, only the result obtained with (-)isoproterenol and 4-methylcatehcol in
control microsomes are depicted in figure 1,for sake of clarity.
During the protection; the catecholamines become oxidized. This is inferred from an
enhanced aminochrome production (not shown). During the oxidation of catecholamines,
reacrive products are formed [7-10]. As demonstrated in figure 2, semi-ortho-quinone radicals
are produced. Addition of Mgt+increased the ESR signal, indicating that the observed radical is
indeed a netui-ortho-quinone radical.

94

~~~~ 1A

1~

2,~
.~
~.
3

0.2

~,,~ ~~..-3
~~
4~.
.,
.r-~.~ t

10

30

20

40

50

0.4
B

ir

O
O
~O
M

2~.
~.,~.
.ter
-~
.~.
-''3

0.2

i
~I

.ter
__.r~~
~ ~~ --

11
0
1.5

0
O
~

10

40

50

1'

4-~
,
~4,a

1.0

~~

m
v-~

~.
~r

0.0 -~
0

30

20

10

~~i~

--

~~.
~~f

20

30

40

50

1111e ~1111ri~

Fig. 1. Protection against mcrosomal lipid peroxidation by (-)isoproterenol or 4-methylcatechol. Lipid


peroxdation was induced by xanthine-xanthine oxidase (panel A),by H2O2-horseradish eeroxidase(panel
B)or by Fee+-ascorbate (panel C) as described in the materials and methods secrion. In the experiment
depicted with the dotted lines, 4-methylcatechol was used, in the other experiments (-)isoproterenol was
used. The concentrations of compounds were:0 M (curve 1), 2M (curves 2),4M (curves 3), 10 M
(curves 4)or 20 M (curves 5).

Fig. 2. E.S.R, spectrum of (-)isoproterenol semiquinone radicals generated from 10 mM (-)isoproterenol


by 40 nM horseradish peroxidase and 1 mM H2O2(trace A). In trace B,50 mM MgC12 was also added.
Inlrument settings are described in the materials and methods section.

Catecholamines can offer complete protection against xanthine-xanthine oxidase induced


lipid peroxidation. During this protection, oxidation products of the catecholamines are formed.
To determine the effect of these oxidation products on the membranes, microsomes subjected to
the combinarion of 1(}0 M (-)isoproterenol and xanthine-xanthine oxidase were isolated. No
lipid peroxidation in these membranes was observed. The membranes were washed,in order to
remove the xanthine-oxidase and the catecholamine. Subsequently, in these pretreated
membranes the GSH-dependent protection was assessed. To this end the pretreated liver
microsomes were peroxidized by addirion of the combinarion of iron and ascorbate. In control
membranes, addition of 1 mM GSH induced a lag time in the occurrence of lipid peroxidation
(fig. 3A). In xanthine-xanthine oxidase pretreated membranes, with or without the addirion of
100 M (-)isoproterenol, the GSH-induced lag time was absent (fig. 3B).

O
O
~

~.5 A

1.5

1.0

1.0

o.s

o.s

D
1J

o.o

o.o
0

10

20

30

time(min)

40

50

10

20

30

40

50

rime(min)

Fig. 3. Lipid pero~dation in control microsomes (panel A) and microsomes pretreated with 100 M
(-)isoproterenol in combination with xanthine-xanthine oxidase (panel B). Lipid peroxidation was induced
by the combination of 10 M Fel+and 0.2 mM ascorbate. Further additions were; none (),~ or 1 mM
GSH(~).
~.

l.o

i.o

0.5

0.5

d
o.o

o.o
0

10

20

30

40

50

10

time (min)

20

30

40

50

time(min)

Fig. 4. Lipid peroxidation in microsomes pretreated with 100 M 4-methyl-ortho-benzoquinone (panel


A), or microsomes pretreated with 1 mM GSH and 100 M 4-methyl-ortho-benzoquinone (panel B).
Lipid peroxidation was induced by the combination of 10 M Fee+ and 0.2 mM ascorbate. Further
additions were; none (), or 1 mM GSH(~).

In order to determine whether the inactivarion of the GSH-dependent protecrion might be due
to the formation of reactive quinones, 4-methyl-ortho-benzoquinone was synthetized. This
compound served as a test substance for quinones of the catecholamines, since these quinones
are unstable due to intramolecular cyclization [7-10]. It was found that preincubation of the
microsomes with 4-methyl-ortho-benzoquinone inacrivated the GSH-dependent protection (iig.
4A). To investigate whether thiol arylation by 4-methyl-ortho-benzoquinone is responsible for
the observed inacrivation, the effect of addition of GSH was assessed. It was found that the
addition of GSH prevented the reduction of the GSH-dependent protection by 4-methyl-orthobenzoquinone (fig. 4B).
Of the other enzymes located on the~microsomal membrane, the calcium-ATPase was also
investigated. The enzyme contains an essential and vulnerable sulfhydryl group, sinc the
synthetic sulfhydryl-alkylating agent N-ethylmaleimide impaired. the capacity of the enzyme to
sequestrate calcium (table 1). It was found that 4-methyl quinone was equally potent as Nethylmaleunide in inactivaring this enzyme (table 1).

N-ethylmaleimide
concentration(M)

1
2.5
1

0
10-4
10-4
10-3

Cat+-sequestrarion(%)
100 5
53 2
23 1
0

4-methyl-ortho-benzoquinone
concentration(M)
0
1 10-4
2.5. 10-4
1 10-3

Cat+ -sequestration(%)
100 5
50 3
63
not determined

Table 1. Inactivation of the calcium sequestration by N-ethylmaleimide and 4-methyl-ortho-benzoquinone in rat


liver microsomes. The microsomes were incubated for 10 minutes with the compounds. Calcium sequestration is
expressed relative to the sequestration in control membranes S.D.

97

Discussion.
In this study the modulation of oxidative stress by catecholamines was determined. Rat liver
microsomes were subjected to several forms of oxidative stress, namely xanthine-xanthine
oxidase, iron-ascorbate and H2O2-horseradish peroxidase. All these types of oxidative stress
resulted in lipid peroxidation. As demonstrated in this study, the catecholamines
(-)isoproterenol, dopamine and a-methyldopa were able to protect against lipid peroxidarion
induced irrespecrive of the way lipid peroxidation was induced(fig 1). The similar effect of the
two enantiomeres of isoproterenol, indicates that probably no receptor-mediated process is
involved. ~Ioreaver, heating did not impair the effect of the catecholamines on lipid
peroxidation, indicating that the effect was non-enzymatic. Catechol and 4-methylcatechol were
also able to prevent microsomal lipid peroxidation, indicating that the catechol moiety of the
catecholamines is responsible for this effect. A similar conclusion was reached previously by
Kappus et al. [17]. The protection is most likely due to scavenging of the radicals) that induce
lipid peroxidation. It has previously been described that catecholamines are very potent
scavengers of the highly reacrive hydroxyl radical [2]. However, scavenging of the hydroxyl
radical is probably not responsible for the the observed protection against lipid peroxidarion,
since it has been accepted that the hydroxyl radical is not involved in the induction of lipid
peroxidarion [18].
In the protecrion against lipid peroxidation by the catecholamines, the catecholanunes are
oxidized. Products like semiquinone radicals, quinones and aminochromes are formed, as
inferred from ESR experiments (iig 2) and aminochrome formation. These compounds are
probably produced during the scavenging of radicals by the catecholamines. For example,it has
been amply demonstrated that in the reaction of catecholamines with the superoxide radical,
generated by xanthine-xanthine oxidase, semi-ortho-quinone radicals are formed [19]. Semiortho-quinone radicals are stabilized by the addition of divalent metal ions, because a relative
stable metal complex is formed, a phenomenon first described by Eaton [20]. We used this
characteristic to demonstrate that the observed ESR-signal is indeed emanating from a semiortho-quinone radical (fig 2). It has been well described that semi-ortho-quinone radicals and
ortho-quinones are potenrially toxic products, especially due to their ability to arylate protein
thiols [9-12].
Various enzymes contain a free sulfhydryl groups) that is essential for their catalytic
activity. For example, it has been described that the enzyme involved in the GSH-dependent
protection in rat liver microsomes, the free radical reductase, contains an essential and
vulnerable thiol group [21-23]. It was deternuned whether this enzyme is inactivated by the
products formed in the protection by catecholamines against lipid peroxidarion. It was found
that incubation of microsomes with (-)isoproterenol in combination with xanthine-xanthine
oxidase reduced the GSH-dependent protection, despite a complete protection against
xanthine-xanthine oxidase induced lipid peroxidation by the catecholamine. Therefore, it was
concluded that the GSH-dependent vitamin E free radical reductase was inacrivated.
To determine whether this effect might be due to the formation of quinones, 4-methylortho-benzoquinone was syntherized. This compound was used, since the quinones of the
catecholamines are not stable, due to an intramolecular 1,4-Michael addition reaction in which a
leucochrome is formed [8-12]. The electrophilic benzoquinone group reacts with the amine of
its side chain under the formation of afive-membered ring. The reaction of the quinones with
98

thiols is much faster, since thiols are much better nucleophiles than amines. Moreover, at
physiological pH only 0.1 % of the amine of the catecholamines exists in its non-ionized form.
Only this form is capable of undergoing nucleophilic addition reactions to benzoquinones [10].
Therefore, it is likely that these products generated in situ in liver microsomes react with
protein sulfhydryl groups, rather than with the their own amine group. It has been demonstrated
that the reaction of ortho-quinones with the sulfhydryl group of GSH or cysteine proceeds
respectively 1400 and 2100 rimes as fast as the intracyclization reacrion [11]. However, if it
would be attempted to add a synthetized quinone of a catecholamine to a suspension of
microsomes, the quinone would react with the amine of its side chain probably already during
the preparation of a solution of the compound. In 4-methyl-ortho-benzoquinone, no
intramolecular cyclizarion is possible, making it a suitable test compound [11].
It was found that also in microsomes preincubation with 4-methyl-ortho-benzoquinone, the
GSH-dependent protecrion was reduced (iig 4). Addirion of GSH prevented the inactivation of
the GSH-dependent protection by 4-methyl-ortho-benzoquinone. This is probably due to the
fact that the non-protein thiol GSH, instead of the sulfhydryl group on the free radical
reductase, was arylated.
Also other enzymes with essential ihiol groups are located on microsomes. Of these
enzymes, the calcium-ATPase is probably the most frequently investigated one, because of its
pivotal role in free radical mediated cell damage [24]. It was found that 4-methyl-orthobenzoquinone was equally potent as the syntheric sulfhydryl-alkylating agent N-ethylmaleimide
in inactivating this enzyme (table 1). This indicates that the calcium-ATPase can also be
inactivated by products of the catecholamines formed in the protection against lipid
pero~dation.
It has previously been reported that sulfhydryl containing enzymes, not located on the
microsomes are also inactivated by oxidation products of catecholamines [9]. It has been
suggested that the reactive species responsible for the inactivation of the reverse DNA
transcriptase is a semi-ortho-quinone radical and not an ortho-quinone [9]. Although the
ortho-quinone we used had the same effect as in situ generated oxidation products of
catecholamines, we cannot exclude that the semi-ortho-quinone radical instead of the orthoquinone is the product responsible for the enzyme inactivarion by the oxidation products of
catecholamines. Moreover, is was possible to demonstrate the presence of the semi-orthoquinone radical using ESR,indicating that these radicals were indeed generated. However,its is
generally believed that the ortho-quinone and not the semi-ortho-quinone radical is the
arylating oxidarion product of catecholamines [10].
In this study it was observed that catecholamines can scavenge radicals that are able to induce
lipid peroxidarion. However,in this protection reactive products are formed that are tonic due to
their ability to arylate sulfhydryl groups. This observarion very well fits in the concept
introduced by Simic and Hunter [25]. They stated that the selectivity of the effects of radicals
depends on the reactivity of the radical. Highly reacrive radicals react aspeciiically with any
compound, and the nature of the damage merely depends on the place where they are formed, a
phenomenon lrnown as site-specific damage. Less reactive radicals are able to inflict less but
more selecrive injury. These radicals are able to reach a specific target, before they react with
other cellular consrituents (fig 5).
Although the identity of the radicals) that induce lipid peroxidation during oxidative stress is
not lrnown [17], this radical is probably highly reactive. The products formed in the protecrion
99

specificity

low

medium

high

YeLICtlVlly
Fig. 5. Relation between the reactivity of a free radicals and the specificity of their effect(after Smic and
Hunter [25]).
by catecholamines against lipid peroxidation are able to arylate sulthydryl groups. The product
responsible for this effect does not have to be a radical of the catecholamines, also quinones are
able to arylate sulfhydryl groups. The observation that catecholamines inhibit lipid peroxidation
does not mean that the catecholamnes completely protect against oxidative stress. As indicated
in this study, by catecholamines the damage provoked by oxidative stress is in fust instance
shifted from lipid peroxidation to enzyme inactivation by sulfhydryl arylarion. This sulfhydryl
alkylation may indirectly stimulate lipid peroxidation in the liver, because the GSH-dependent
protection against lipid pero~dation is inactivated. Moreover, one of the most important targets
in lipid peroxidatin-induced cell damage, the calcium ATPase,can directly be inacrivated by
arylarion of its sulfhydryl group by reactive products of catecholamines. In addirion, it should
be noted that some catecholamines are cardiotoxic, probably because thay are able to induce
oxidative stress via (3-adrenoceptor hyperstimulation [4,5].
The modulation of oxidative stress by catecholamines is especially important in ischemia
induced damage. During ischemia, xanthine dehydrogenase is converted to xanthine oxidase,
and (hypo)xanthine, the substrate for xanthine oxidase, accumulates as a result of ATP
catabolism. Moreover, neutrophils enter the ischemic tissue and they also are a potenrial source
of superoxide radicals [26]. In addition, catecholamines are released [14]. Koide et al [27]
suggested that circulating catecholamines, released as a response to stress, ameliorate ischemic
brain damage. They speculated that these catecholamines penetrate that blood-brain barrier, that
has been damaged during ischemia. Interacrion with an adrenoceptor in the brain is supposed to
be involved in the protection by the circularing catecholamines [27]. However, it is not
explained by Koide et al.[27] how adrenoceptor-activation by catecholamines in the brain can
protect against ischemic damage. Alternatively, the protecrion by catecholamines might be
caused by inhibition of ischemia-induced lipid pero~darion, however,it should be noted that by
this mechanism sulfhydryl groups in the brain are not protected and probably become more
affected.
~~

Recently, Woolf et al. [2~] stated that the levels of circulating catecholamines are excellent
markers that appear to reflect the extent of brain injury and that may predict the likelihood of
recovery; the higher the level of catecholamines, the more severe the injury, the lover the
change of recovery. Seemingly, the results of Woolf et al. [28] and Kioke et al. [27] do not
concur. It is possible that the release of catecholamines is determined by the extent of the brain
injury, and that despite a higher level of catecholamines, the catecholamine cannot protect
against the more severe injury.
In contrast to the assumed protecrion of the ischemic brain, it has been known for for many
years that catscholamines administered in large doses produce ~ayocard:a? necrosis, although the
molecular mechanism for this cardiotoxicity is not fully understood. It has been suggested that
lipid pero~darion is involved in catecholamine-induced heart damage [6, 29]. Vitamin E, an
endogenous free radical scavenger, was reported to protect against (-)isoproterenol-induced
cardiomyopathy [29]. Singal et al.[29] speculated that the heart damage was due to free radicals
produced during the (aut)oxidation of catecholamines that subsequently induce lipid
peroxidation. However,in this study it is shown that free radicals of catecholamines are not able
to react with poly-unsaturated fatty acids. In contrast, catecholamines protect against lipid
peroxidarion. Therefore, it is more likely that enzyme-inhibition via sulfhydryl arylarion by
oxidation products of catecholamines is involved in catecholamine-induced cardiotoxicity. In
addirion to free radical production, j3-adrenoceptor overstimulation is probably involved in
catecholamine-induced myocardial damage [3, 4]. Interestingly, oxidative stress reduces
(3-adrenoceptor function, which may provide aself-limiting feedback mechanism in the
catecholamine-induced acute cardiotoxicity mediated via ~3-adrenoceptor overstimulation [4].
However, an impaired function of the ~i-adrenoceptor is supposed to contribute to chronic heart
failure [30]. The cardiotoxicity of catecholamines probably also contributes to the reported
correlation between the concentration of circulating catecholamines and the outcome in traumaric
brain injury [28]
Several catecholamines are hepatotoxic. The toxicity of a-methyldopa(which is not exactly a
catecholamine) is suggested to be due to the scavenging of superoxide radicals generated by
cytochrome P-450, a reacrion leading to the production of a reactive semiquinone and quinone
[31]. The final toxicity is produced by covalent binding of the reactive products of a-methyl
dopa [31]. This scheme~pei~fecfly matches with the results obtained in this study.
Recently, it has been reported that catecholamines activate the membrane-bound
glutathione-S-transferace in rat liver microsomes [32]. It was suggested that by a cytochrome
P-450-induced oxidarion of catecholamine, activated oxygen species are formed, and that these
acrivated oxygen species activate the microsomal glutathione-S-transferase, presumably by the
oxidation of the sulfhydryl group of the enzyme [32]. In contrast, we found that arylation of the
sulfhydryl group of the microsomal glutathione S-transferace by ortho-quinones, rather than
sulfhydryl oxidation by active oxygen species is responsible for observed activarion of the
enzyme by catecholamines [Naenen,submitted for publicarion]. It has well been described that
alkylarion of the free sulfhydryl group of the microsomal glutathiane S-transferase stimulates
the acrivity of this enzyme [20-22]. The alternarive mechanism presented by us coincides with
the scheme presented by Dybing et al[31] and the results presented in this study.
In conclusion, catecholamines are able to scavenge radicals that induce lipid peroxidation.
The catecholamines are oxidized in this process, and the products formed are toxic due to their
101

ability to react with free sulfhydryl groups.


References.
1
2
3

4
5
6
7
8
9
10

11
12
13
14
15
16
17

18
19
20
21
22

23

24
i25~
~~
102

H. Sies (ed.), Oxidative Sh~ess, Academic Press, New York-London, 1985.


A. Slivka and G. Cohen, Hydroxyl radical attack on dopamine, J. Biol. Chem. 260(1985) 15466-15472.
L.H. Opie, C. Muller, D. Nathan, P. Danes and W.F. Lubbe, Evidence for the role of cyclic AMP as
second messenger of arrhythnogenic effects ofbeta-stimulation, in: P.Hamet and H. Sands (Eds.) Advances
in cyclic Nucleotide research, volume 12, Pathophysiological aspects od cyclic nucleotides, Raven Press,
New York (1980)63-69.
G.R.M.M. Haenen,P. van Dansfik, N.P.E. Vermeulen, H. Timmerman and A. Bast, The effect of hydrogen
peroxide on (3-adrenc~cegtor function in Lhe heart, Free Rad. Res. Comets.4(1Q88)243-249.
S.L. Jewett, L.J. Eddy and P. Hochstein, Is the autoxidadon of catecholamines involved in ischemiareperfusioninjury? Free Rad. Biol. Med.6(1989) 185-188.
A.A. Noronha-Dutra, E.M. Steen and N. Woolf, The correlation between catecholamine and lipid
pero~ridation induced damage in heart cells, Basic Res. Cazdiol. 80(1984)S1:133-136.
H. Iwahashi, H. Morishita, T,Ishri, R. Sugata and R. Kido, Enhancement by catechols of hydroxyl-radical
formation in the presence of ferric ions and hydrogen peroxide, J. Biochem 105(1989)429-434.
W. Bors, C. Michel, M. Saran and E. Lengfelder, The involvement of oxygen radicals during the
autoxidation of adrenalin, Biochim. Biophys. Ac[a 540(1978) 162-172.
M.M. Wick and G.Fitzgerald, Inhibition of reverse transcriptase by tyrosinase generated quinones related to
levodopa and dopamine, Chem.-Biol. Interactions 38(1981)99-107.
Kalyanaraman, P.I. Premovic and R.C. Sealy, Semiquinone anion radicals from addition of amino acids,
peptides and proteins to quinones derived from oxidation of catechols and catecholamines, J. Biol. Chem.
161(1987)11080-11087.
D.C.S. Tse, R.L. McCreery and R.N. Adams, Potential oxidative pathways of brain catecholamines, J.
Med. Chem 19(1976)37-40
S. Ito, T. Kato and K. Fujita, Covalent binding of catechols to proteins through the sulfhydryl group,
Biochem.Parmacol. 37(1988) 1707-1710
R. Willst~uer and A.Pfannenstiel, Ueber o-Chinon, Chem. Berichte 4(1904)4744-4746
L. Carlsson, T. Abrahamsson, O. Almgren and L. Svensson, Ischemia-induced release of Waradrenaline and
creatine kinase in the isolated, working rat heart, Acta Physiol. Stand. 130(1987)83-93.
G.R.M.M. Naenen and A. Bast, Protection against lipid peroxidation by a microsomal glutathionedependentlabile factor,FEBS Lett 159(1983)24-28.
M.M. Bradford, A rapid and sensitive method for the quantitadon of microgram quantities of protein
utilizing the principle of protein-dye binding, Anal. Biochem. 72(1976)246-254.
H. Kappus, H. Kieczka, M. Scheulen and H. Remmer, Molecular aspects of catechol and pyrogallol
inhibition of liver microsomal lipid peroxidation stimulated by ferrous ion-ADP-complexes or by carbon
tetrachloride, Arch.Pharmacol. 300(1977) 179-187.
A. Bast and M.H.M. Steegs, Hydroxyl radicals are not involved in the NADPH dependent microsomal lipid
peroxidation, Experientia 42(1986)555-556.
H.P. Misra and I. Fridovich, The role of superoxide anion in the autoxidarion of epinephrine and a simple
assay for supero~cide dismutase, J. Biol. Chem.247(1972)3170-3175.
D.R. Eaton, Complexing of metal ions with semiquinones. An electron spin resonance study, Inorganic
Chem. 3(1964) 1268-1271
G.R.M.M Naenen, N.P.E. Vermeulen, J.N.L. Tai Tin Tsoi, N.P.E. Vermeulen, H. Timmerman and A.
Bast, Arch. Biochem. Biophys.259(1987)449-456.
G.R.M.M Naenen, N.P.E. Vermeulen, J.N.L. Tai Tin Tsoi, H.M.N. Ragetli, H. Timmerman and A. Bast,
Activation of the microsomal glutathione-S-transferace and reduction of the glutathione-dependent protection
against lipid peroacidation by acrolein, Biochem.Pharmacol. 37(1988) 1933-1938.
G.R.M.M Naenen, N.P.E, Vermeulen, H. Timmerman and A. Bast, Is phospholipase AZ involved in the
glutathione-dependent protecon against in vitro microsomal lipid peroxidation, in: O. Hayashi, E. Niki,
M. Kondo and T. Yoshikawa (Eds.) Medical, biochemical and chemical aspects of free radicals, Elsevier,
Amsterdam (1989)pp. 1291-1294.
S. Orrenius, Biochemical mechanisms of cytotoxicity, Trends Biochem. Sci. 10(1985)S15-S20.
M.G. Simic and E.P.L. Hunter, Interactions of free radicals and antioxidants, in: O.F. Nygaard and M.G.

Simic (Eds.) Radioprotectors and Anticarcinogens, Academie Press, New York-London (1983)pp. 449-460.
26 J.C. Romson, B.G. Hook, S.L. Kundel, G.D. Abrams, M.A. Schork and B.R. Lucchesi, Reduction of
ultimate extent of ischemic myocazdial injury by neutrophil depletion in the dog, Circulation 67 (1983)
1016-1023.
27 T. Koide, T.W. Wieloch and B.K. Siesj, Circulating catecholamines modulate ischemic brain damage, J.
Cerebral Blood Flow Metabolism 6(1986) 559-565.
28 P.D. Woolf, R.W. Hamill, L.A. Lee, C. Cox and J.V. McDonald, The predictive value of catecholamines
in assessing the outcome in traumatic brain injury, J. Neurosurg. 66(1987)875-882.
29 P.K. Singal, N. Kapur, K.S. Dhillon, R.E. Beamish and N.S. Dhalla, The role of free radicals in
catecholamine-induced cardiomyopathy, Can. J. Phys. Pharmacol.60(1982) 1290-1397.
30 H. Nawrath, Adrenoceptor-mediated changes of excitation and contraction n isolated heart muscle
preparations, J. Cardiovasc.Pharmacol. 14 S3(1989) 1-10
31 E. Dybng, S.D. Nelson, J.R. Mitchell, H.A. Sasame and J.R. Gilette, Oxidation of a-methyldopa and
other catechols by cytochrome P-450-generated superoxide anion: Possible mechanism of methyldopa
hepatitis, Mol. Phannacol. 12(1976)911-920.
32 Y. Aniya and M.W. Anders, Activation of rat liver microsomal glutathione S-transferace by reduced oxygen
species, J. Biol. Chem 264(1989) 1998-2002.

103

~~

- Chapter 11 Activation of the microsomal glutathione S-transferase by


metabolites of a-methyldopa.
Introduction.
Thy glutathione S-transferaces (GSH-tr) are- ascup of is~zymes that catalyze the reacrion of
electrophiles with glutathione (GSH) [1]. By binding to GSH and further metabolism to
mercapturic acids, the excretion of an electrophilz is enhanced due to an increase in water
solubility and due to acrive transport mechanisms for GSH-related conjugates [1]. GSHconjugation is a major route of detoxication of xenobiotics, however, for several compounds
GSH-conjugation enhances toxicity [2].
The GSH-tr also have GSH-peroxidase activity [1]. In this mode of action, the GSH-tr
catalyze the reacrion between GSH and hydroperoxides, thereby converting the hydroperoxide
into the corresponding less reactive alcohol. The GSH-tr form the selenium-independent GSHperoxidases [1]. Like selenium-dependent GSH-peroxidases, GSH-tr accept organic
hydroperoxides as substrate, but in contrast to the selenium-dependent GSH-peroxidases, the
GSH-tr display no catalytic activity towards hydrogen peroxide [1]. Most of the GSH-tr are
located in the cytosol, but a membrane GSH-tr has also been described [3-5]. In the rat liver,
the microsomal GSH-tr only accounts for 2% of the total GSH-tr activity, when esrirnated with
1-chloro-2,4-dinitrobenzene(CDNB)as substrate [3].
It has well been described that the cytochrome P-450 system catalyzes the transformation of
several substrates into reactive metabolites, especially in liver microsomes due to the high
content of this enzyme complex [6]. Due to its location near cytochrome P-450, the contriburion
of the microsomal GSH-tr in the detoxication of eleclrophiles generated by cytochrome P-450
might be more substantial than based on its relarive acrivity [4]. Moreover, the activity of the
microsomal GSH-tr can be enhanced by alkylation of its thiol group [3-5, 7]. It has been
speculated that via alkylarion of its sulfhydryl group the rnicrosomal GSH-tr becomes activated
when it is needed, i.e. when alkylating electrophiles are to be detoxified by GSH-conjugation

[7l
Recently, we have shown that reactive products of catecholamines, like ortho-quinones and
semi ortho-quinone radicals which are formed in a reaction of catecholamines with free
radicals, are able to arylate sulfhydryl groups [8]. Moreover, we found that via the producrion
of these ortho-quinones and semi ortho-quinone radicals the activity of enzymes located on
liver microsomes is altered [8]. As shown by Dybing et al. [9], the cytochrome P-450 system
generates similar reacrive products from catechol-containing compounds. Of these compounds,
a-methyldopa is of physiological importance, because it is used as an anrihypertensive drug and
its application often gives rise to hepatic injury [9]. The toxicity of a-methyldopa is probably
related to the formation of reactive products in a cytochrome P-450 catalyzed reaction [9]. The
aim of this study was to determine whether reacrive products generated from a-methyldopa by
the cytochrome P-450 system are indeed able to acrivate the microsomal GSH-tr.

105

Materials and methods.


Chemicals. a-Methyldopa was from Janssen (Beerse, Belgium). NADPH, xanthine oxidase (grade III),
reduced glutathione (GSH), tyrosinase, horseradish eeroxidase, (-)isoproterenol, 1-chloro-2,4-dinitrobenzene
(CDNB) were &om Sigma (St. Louis, MO).4-Methylcatechol was from Fluks(Bachs, Switserland). 4-Methylortho-quinone was synthetized according to Willst~tter and Pfannensteil [10]. The buffers used were: buffer A =
50 mM sodium diphosphate, 0.1 mM EDTA, pH 8.0 unless otherwise noted; buffer B = 50 mM Tris-HCI, 150
mM NaCI, pH 7.4.
Methods. Microsomes were prepared as described previously [l l], and stored at -80 C. In order to determine
the microsomal GSH-tr activity, the microsomes were thawed and diluted 5-fold with buffer A. To remove
cytosolic contamination, the microsomes were washed twice with buffer A by centrifugation (40 min, 115,000 g
at 4 C). The microsomal pellet was resuspended in buffer A,final concentration microsomes derived hom 2 g
liver in 1 ml buffer. In the experiment were lipid peroxidation was induced, the microsomes were washed and
suspended in buffer B, as described previously [l l]. Concentrations of the compounds used were: NADPH,2
mM; NEM,5 mM; a-methyldoes, 1 mM;4-methylcatechol, 1 mM;(-)isoproterenol, 1 mM; 4-methyl-orthoquinone, 1 mM unless otherwise indicated. All reactions were terminated by the addition of an equimolaz
concentration of GSH, except when it was indicated that GSH was added prior to the addition of the other
compounds. When NEM was added to the incubation medium,it was added after the indicated incubation time of
th microsomes (unless otherwise indicated) and the microsomes were incubated for an additional 30 seconds.
Activity of the enzymes were: 800 pM horseradish eeroxidase; 1 mM H2O2; 5 U/ml tyrosinase; 55 mU/ml
xanthine oxidase; 50 M xanthine. Incubation times are indicated in the respective tables. In order to determine
the microsomal GSH-tr activity, the microsomes had to be diluted 200-500 fold with buffer A,pH 6.5. GSH-tr
activity was determined using CDNB as substrate. Lipid peroxidaon was measured using the thiobarbituric acid
method as described previously [l l], and expressed as the absorbance at 535 nm versus 600 nm (~A535-6001
Protein determinations were made according to the method of Bradford [12], using bovine serum albumin as
standard. The results are expressed as mean standard deviation, n = 3 - 20.

Results.
Liver microsomes of the rat contain a membrane bound GSH-tr that catalyzes the GSHconjugation of 1-chloro-2,4diniirobenzene(CDNB)(table 1). Incubation of the microsomes for
30 seconds with the sulfhydryl alkylating agent N-ethyl maleimide(NEM)activated the GSH-tr
activity towards CDNB (table 1). The incubarion of NEM was ternunated with the addition of
an equimolaz concentration of GSH. GSH added to the rnicrosomes prior to NEM, almost
completely prevented the activation of the microsomal GSH-tr by NEM (table 1). These

NADPH

GSH

NEM

control

+
-

+
+

+
+

57 7
62 1
623
181 11
95 11

+ a-methyldoes

+
+
+

+
-

46 2
131 9
614
1803

GSH-tr
nmol/min/mg

Table 1. Cytochrome P-450-mediated stimulation of the microsomal GSH-tr by a-methyldoes.


Composition of the incubation systems is described in the materials and methods section. Incubation time
was 10 minutes. In the incubation where it was indicated that GSH was used, it was added to the
microsomes before NADPH and a-methyldope. In [he incubation system where NEM was present, it was
added 10 minutes after the reaction was started, and the reaction mixture was incubated for an additiona130
seconds.
106

control

+ a-methyldopa
+ NEM
+ GSH
+(-)isoproterenol
+ 4methylcatechol

H2O2

HRP

GSH-tr

+
-

nmol/min/mg
57 7
515
566

617

+
+
+

+
+
+

937
213 24
523

44 8
9~g

66 it
108 10

Table 2. Horseradish eeroxidase(HPR)-mediated stimulation of the microsomal GSH-tr by a-methyldopa,


(-)isoproterenol and 4-methylcatechol. Composition of the incubation systems is described in the
materials and methods section. Incubation time was 10 minutes. In the incubations where GSH was used,
it was added to the microsomes before NADPH and a-methyldopa. In the incubation system where NEM
was present, it was addec 10 minutes after the reaction was started, and the reacon mixture was incubated
for an additiona130 seconds.

experiments are in agreement with previous observarions that demonstrate the acrivation of the
microsomal GSH-tr via alkylarion of its sulfhydryl group [3-5,7].
As also shown in table 1, a-methyldopa itself is not able to activate the microsomal GSH-tr,
nor is NADPH. However, with the combinarion of NADPH and a-methyldopa the activity of
the microsomal GSH-tr is enhanced (table 1), indicaring that interacrion of a-methyldopa with
the cytochrome P-450 system yields a product that is able to acrivate the microsomal GSH-tr.
NEM was srill able to activate the microsomal GSH-tr of microsomes pretreated with

a-methyldopa and NADPH (table 1), indicating that the activation by a-methyldopa and
NADPH in our experimental set-up was not maximal. The activity of the microsomal GSH-tr
after the addirion of NEM in control microsomes (table 1) was idenrical to the acrivity after NEM
in microsomes pretreated with a-methyldopa and NADPH (table 1). Moreover, GSH was able
to protect against the acrivation of the microsomal GSH-tr by a-methyldopa and NADPH (table
1). These findings indicate that the activarion of the microsomal GSH-tr by the combinarion of
a-methyldopa and NADPH proceeds via the same mechanism as the activation by NEM,i.e.
via a reaction with the free sulfhydryl group of the microsomal GSH-tr.
To invesrigate whether the observed activation is mediated by the formation of a orthoquinone of a-methyldopa, we determined the effect of in situ generated ortho-quinones of
a-methyldopa. The ortho-quinone was generated in two ways.
In the first method, a-methyldopa was oxidized by the combination of horseradish
eeroxidase and hydrogen peroxide. This mediates a one electron oxidation of the catechol
moiety of a-methyldopa. The semi-ortho-quinone radical yields the corresponding orthoquinone in a dismutation reaction. It was found that neither hydrogen peroxide, nor horseradish
eeroxidase, nor the combination of the two were able to activate the microsomal GSH-tr (table
2). However, when a-methyldopa was added in combination with hydrogen peroxide and
horseradish eeroxidase, the microsomal GSH-tr was activated (table 2). G5H was able to
protect against the acrivation produced in this way (table 2). NEM was able to further stimulate
107

Addition

GSH-tr
nmol/min/mg
57 7
69 7
102 8
187 7

control
+ tyrosinase
+ a-methyldopa
+ a-methyldopa + NEM

Table 3. Tyrosinase-mediated stimulation of the microsomal GSH-tr by a-methyldopa. Composition of


the incubation systems is described in the materials and methods section. Incubation dme was 10
minutes. In the incubation system where NEM was present, it was added 10 minutes after the reaction was
started, and the reaction mixture was ircubated for an addidona130 seconds.

the microsomal GSH-ir (table 2) to the same level as observed in control microsomes treated
with NEM alone (table 1). Also the products formed in the horseradish peroxidase-hydrogen
peroxide catalyzed oxidation of(-)isoproterenol and 4-methylcatechol were able to activate the
microsomal GSH-tr, while (-)isoproterenol or 4 methylcatechol alone did not activate the
enzyme (table 2).
The tyrosinase-mediated two electron oxidation of a-methyldopa was the second method
used to produce ortho-quinones in situ. By this method the ortho-quinone is directly
generated. Tyrosinase alone is not able to activate the microsomal GSH-ir, however when
a-methyldopa was also added, the activity of the GSH-tr was enhanced (table 3).
In order to further confirm our hypothesis that the ortho-quinone of a-methyldopa is
responsible for the observed activation of the microsomal GSH-tr, 4-methyl-orthobenzoquinone was synthetized. 4-Methyl-ortho-benzoquinone cannot decompose through an
intramolecular cyclizarion reaction, in contrast to the ortho-quinone of a-methyldopa [8, 13].
Previously, it has been shown that 4-methyl-ortho-benzoquinone was a suitable test compound
for the ortho-quinone of catecholamines and a-methyldopa [8, 13]. In figure 1, it is shown that

addirion

NEM

GSH

GSH-tr

control

nmol/min/mg
57 7
181 11

+ 4-methyl-ortho-benzoquinone

+1
+2
+3

+
-

164 6
56 1
154 5
175 8
172 3

1 NEM added after 4-methyl-ortlw-benzoquinone


2NEM added before 4-methyl-ortho-benzoquinone
3IVEM added together with 4-methyl-ortho-benzoquinone
Table 4. Activation of the microsomal GSH-tr by 4-methyl-ortho-benzoquinone. Composition of the
incubation systems is described in the materials and methods section. After the addition of 4-methylortho-benzoquinone or NEM the incubation time was 30 seconds. The reaction was terminated by the
addition of a concentration of GSH equal Go the concentrations of 4-methyl-ortho-benzoquinone and NEM
together (except when it was indicated that GSH was used, in that experiment GSH was added before 4methyl-ortho-benzoquinone).

108

150
.~
~~ ~
~ ~ 100
~ ~
~

~ ~

50

C~
.001

.O1

.1

10

100

[4-methyl-ortho-benzoquinone](mM)
Figure 1. Concentration-dependent effect of 4-methyl-ortho-benzoquinone on the activity of the
microsomal GSH-tr. Incubation time was 30 seconds. Basal activity of the microsomal GSH-tr was 57
7 nmol/min/mg.

4-methyl-ortho-benzoquinone stimulates the microsomal GSH-tr in a concentration dependent


way. Maximal activation was observed at a concentrarion of 1 mM,above this concentrarion the
stimulatory effect of 4-methyl-ortho-benzoquinone decreased. The activity of the microsomal
GSH-tr induced by 1 mM 4-methyl-ortho-benzoquinone (table 4) was idenrical to the activity
induced by 5 mM NEM (table 1). Neither addirion of5 mM NEM to microsomes treated with 1
mM 4-methyl-ortho-benzoquinone, nor addirion of 1 mM 4-methyl-ortho-benzoquinone to
microsomes treated with 5 mM NEM,nor addition of the combinarion of 5 mM NEM and 1 mM
4-methyl-ortho-benzoquinone at the same time resulted in a greater activity of the microsomal
GSH-tr when compared to the activity in microsomes treated with either 1 mM 4-methyl-orthobenzoquinone or 5 mM NEM alone (table 4). Addition of an equimolar concentration of GSH
prevented the acrivarion of the microsomal GSH-tr by 4-methyl-ortho-benzoquinone (table 4).
These results denote that 4-methyl-ortho-benzoquinone is able to activate the microsomal
GSH-tr via the same mechanism as NEM.
Recently, it has been suggested that reduced oxygen species are able to activate the
microsomal GSH-transferace [14]. As demonstrated above, hydrogen peroxide is not able to
activate the microsomal GSH-tr (table 2). Superoxide anion radicals generated by xanthinexanthine oxidase were also not able to acrivate the GSH-tr (table 5). It has also been speculated
that lipid peroxidation enhances the activity of the microsomal GSH-tr and that via this
mechanism the protection against lipid peroxidation is enhanced because the GSH-proxidase
activity of the microsomal GSH-tr is stunulated [14]. To test this hypothesis, lipid peroxidation
was induced in the liver microsomes by 10 M Fel+ or by the combination of 10 M Fel+and
0.2 mM ascorbate. A moderate level of lipid peroxidation was generated in the experiment
where only iron was used, whereas the combinarion of iron and ascorbate produced a more
pronounced peroxidarion of the membrane lipids (table 5). It was found that by moderate lipid
109

addirion

control
+ xanthine +xanthine oxidase
+ 10 M Fel+
+ 10 M Fel++ 0.2 mM ascorbate

lipid pero~cidarion

GSH-tr

X535-600
0
n.d.
0.050 0.004
1.024 0.058

nmol/min/mg
57 7
60 5
62 5
43 4

n.d.= no[ determined.


Table 5. Effect of superoxide anions and lipid peroxidation on the activity of the microsomal GSH-tr.
Composition of the incubation systems is described in the materials and methods section. Incubation time
was 10 minutes.
peroxidation, the activity of the microsomal GSH-tr was not affected, while after more
substantial lipid peroxidation the activity of the microsomal GSH-transferace was reduced rather
then increased (table 5).
Discussion
In liver microsomes, the cytochrome P-450 containing mixed function oxidase system is
present at a high concentration [6]. Via the cytochrome P-450 system various xenobiorics are
converted into reactive products [6]. Liver microsomes also contain amembrane-bound
glutathione S-transferace (GSH-tr), an enzyme involved in the detoxicarion of various reactive
products [3-5]. The microsomal GSH-tr might be of importance in the detoxication of reactive
cytochrome P-450 products, especially because of its location at the same subcellular site as
cytochrome P-450 system [4]. In the primary aminoacid sequence of the microsomal GSH-tr
one cysteine residue is found [5]. Alkylation of the free sulfhydryl group of this cysteine
enhances the activity of this enzyme [5, 7]. It has been speculated that the activation of the
microsomal GSH-tr by alkylarion of its free sulfhydryl group ensures that the protective
capacity of this enzyme is enhanced when it is needed,i.e. when alkylative reactive products are
to be detoxified by conjugation to GSH [7]. To put the presumed interplay between the
microsomal GSH-tr and the cytochrome P-450 system to the test, reactive products of
a-methyldopa were generated. The effect of these products on the activity of the microsomal
GSH-tr was deternuned.
As shown in this study, a-methyldopa itself is not able to activate the microsomal GSH-tr.
Also NADPH,a cofactor of the cytochrome P-450 complex, was not able to activate the GSHtr. However, addirion of the combinarion of a-methyldopa and NADPH to liver microsomes of
the rat enhanced the catalytic activity of the GSH-tr. The sulfhydryl alkylaring agent N-ethyl
maleimide (NEM) was still able to activate the microsomal GSH-tr in a-methyldopa and
NADPH pretreated microsomes; the final activity was equal to that stimulated idrectly by NEM
in control microsomes. Addition of GSH prevented the activation by NADPH and amethyldopa. These resulfs indicate that a reactive product of a-methyldopa generated by
cytochrome P-450 complex activates the microsomal GSH-tr probably in a similar way as
NEM,i.e. via alkylation of the free sulfhydryl group of the microsomal GSH-tr.
Previously, Dybing et al. [9] studied the activarion of a-methyldopa by the cytochrome
P-450 complex. a-Methyldopa added to rat liver microsomes was activated in such a way that it
110

gave rise to a large amount of covalent binding, provided NADPH and oxygen were present. It
was suggested that a-methyldopa was oxidized by cytochrome P-450-generated superoxide
anions to the corresponding reactive semi ortho-quinone radical or ortho-quinone. These
reactive products were assumed to be responsible for the covalent binding [9]. GSH in a
concentrarion of 1 mM almost completely abolished the binding reacrion, probably by binding to
the reactive product of a-methyldopa [9]. It is well known that ortho-quinones are very
reactive toward sulfhydryl groups [8, 9, 13]. The results of Dybing et al. [9] concur with the
findings of this study, and suggest that the ortho-quinone or semi ortho-quinone radical of
a-methyldopa is responsible for the observed activation of the microsomal GSH-tr by arylarion
of the free sulfhydryl group of the enzyme.
To determine whether the reactive products of a-methyldopa are indeed able to acrivate the
microsomal GSH-tr, these products were generated in situ, either by a H2O2-horseradish
peroxidase-supported one-electron oxidation or by a tyrosinase-supported two-electron
oxidation. It was found that the products formed activated the microsomal GSH-tr. Moreover,
the results obtained with NEM and GSH in these microsomes point to the same mechanism for
the activarion of the ortho-quinone of a-methyldopa and NEM.
The ortho-quinone of a-methyldopa is not stable, due to a rapid intramolecular cyclization
reaction [9, 13]. Ina 1,4-Michael type addition reaction the amine of its side chain reacts with
the ortho-benzoquinone moiety under the formation of a leucochrome. However, the reaction
of the ortho-benzoquinone moiety with free sulfhydryl groups is much faster than the
intracyclization reaction [9, 13]. Therefore, in situ generated ortho-quinone of a-methyldope
would rather react with sulfhydryl groups than with its own amino function. Nevertheless,
synthetic ortho-quinone of a-methyldope cannot serve as atest-compound because the reaction
of the ortho-quinone with its amine function is still too fast. Because of the unstablilty of the
ortho-quinone of a-methyldope and related products, 4-methyl-ortho-benzoquinone is often
used as test-compound [9, 13].
It was found that 4-methyl-ortho-benzoquinone in a concentrarion up to 2 mM activated the
membrane-bound GSH-tr in rat liver microsomes, probably also via the alkylation of free
sulfhydryl groups. The latter sugestion is based on the finding that the activation of the
microsomal GSH-tr by NEM was not additive to that of 4-methyl-ortho-benzoquinone.
Moreover, GSH prevented the effect of both 4-methyl-ortho-benzoquinone and NEM.
Concentrations of 4-methyl-ortho-benzoquinone above 10 mM reduced activity of the
microsomal GSH-tr. Possibly the same mechanism as for the previously reported inhibirion of
the microsomal GSH-tr by para-benzoquinone is involved in this effect of 4-methyl-orthobenzoquinone, viz. by a mixed-type noncomperirive inhibirion with CDNB [15].
The data presented in this study confirm the hypothesis that the microsomal GSH-tr is
activated by reactive products formed by the cytochrome P-450 complex [4]. In the case of
a-methyldope the ortho-quinone or semi ortho-quinone radical of a-methyldope is probably
the product formed that is responsible for the effect. The acrivarion is probably mediated via the
arylarion of the free sulfhydryl group of the microsomal GSH by that reactive product. GSH in
a concentration of 1 mM was able to prevent the activation of the enzyme. In the liver, the
concentration of GSH can amount to approximately 5-10 mM. Therefore, it is probable that
activation in vivo only takes place after a substanrial reduction of the GSH-content of the liver.
In this case the concentrarion of GSH rather than the maximal activity of the GSH-tr can be rate
limiting. Moreover, it has been suggested that the cytosolic GSH-tr are not able to catalyze the
111

reaction of the ortho-quinone of dopamine with GSH [16]. It is not investigated in this study
that the ortho-quinone of a-methyldopa is a substrate of the microsomal GSH-tr, nor that the
activity of the microsomal GSH-tr toward this product is enhanced by the proposed arylation.
Recently, a similar activation of the microsomal GSH-tr was reported by Aniya and Anders
[14]. The catechol-containing compound noradrenaline was shown to activate the microsomal
GSH-tr. NADPH was needed for this acrivarion by noradrenaline. Therefore, Aniya and Anders
[14] suggested that cytochrome P-450 was involved in the activation of the microsomal GSH-tr
by noradrenaline. Catalase and superoxide dismutase were reported to partially protect against
the activation by noradrenaline-NADPH. Therefore, it was suggested that the microsomal
GSH-tr is activated by reduced oxygen species like superoxide anions and hydrogen peroxide
formed by adrenaline oxidarion or the cytochrome P-450 system [14]. These reduced oxygen
species were presumed to activate the microsomal GSH-tr by the oxidation of its sulfhydryl
group [14]. In contrast to this presumed mechanism, we found that neither hydrogen peroxide,
nor superoxide anions were able to activate the microsomal GSH-tr. Previously, Dybing et al.
[9] demonstrated that superoxide dismutase prevented the covalent binding of a-methyldopa to
liver microsomes when NADPH was also present. They suggested that this was due to
scavenging of superoxide radicals produced by cytochrome P-450, which in turn would activate
a-methyldopa by rendering into a reactive electrophile i.e. the ortho-quinone or semi-orthoquinone radical. Identical results were obtained with the catecholamines dopa, dopamine and
adrenaline [9]. Therefore, the activation of the microsomal GSH-tr by noradrenaline reported by
Aniya and Anders [14] is, in contrast to their explanation, rather due to the cytochrome P-450
mediated formation of the ortho-quinone or semi-ortho-quinone radical of noradrenaline than
to the production of reduced oxygen species. The reacrive product of noradrenaline activates the
microsomal GSH-tr by arylation its free sulfhydryl group.
In a preliminary report, Aniya and Anders [17] argued that noradrenaline activates the
membrane-bound GSH-tr in liver microsomes by disulfide bond formation, possibly through a
catecholamine a-receptor. It should be noted that liver microsomes do not contain an
a-adrenoceptor, and therefore the activation of the microsomal GSH-tr in vitro by
noradrenaline cannot be the result of an interaction of noradrenaline with an a-adrenoceptor.
However,in hepatocytes a-adrenoceptors are located on the plasma membrane. Stimulation of
these a-receptors by e.g. noradrenaline, induces an efflux of GSH from the liver cells [18].
Previously, Masukawa and Iwata [19] have shown that deplerion of GSH in the liver in vivo
results in an activation of the microsomal GSH-tr. Possibly, noradrenaline is able to activate the
microsomal GSH-tr indirecfly, via an a-adrenoceptor-mediated reduction of the GSH-content of
the liver. However, the a-adrenoceptor only addresses a small pool of the total amount of GSH
present in the liver [18], and it is likely that the maximal a-adrenoceptor-mediated reducrion of
the GSH-content of the liver is not sufficient for acrivation of the microsomal GSH-tr.
It has well been described that liver microsomes of the rat contain amembrane-bound
enzyme that is involved in the protection of these membranes against lipid peroxidarion by GSH
[7, 11]. It has been suggested that the enzyme involved in this GSH-dependent protection is the
microsomal GSH-tr [3, 20, 21]. This protection would be mediated via the GSH-penoxidase
activity of the enzyme [3, 20, 21]. During lipid peroxidarion reacrive lipid hydroperoxides are
formed. The microsomal GSH-tr is though to prevent against lipid peroxidation by converting
lipid hydroperoxides into their corresponding less reactive alcohols [3, 20, 21]. In this respect it
112

should be noted that the peroxidized fatty acids formed during lipid peroxidation remain
esterified to glycerol. For the detoxication of these organic hydroperoxides by the cytosolic
GSH-tr, the covalent bond between the lipid hydroperoxides and glycerol has to be broken, in a
phospholipase A2-mediated reaction [22]. However, as demonstrated previously, the
combination of phospholipase A2 and the cytosolic GSH-tr are not able to provide effective
defense against in vitro lipid peroxidation [22].
It has been speculated that due to its location the microsomal GSH-tr displays a GSHperoxidase activity toward esteriiied lipid hydroperoxides [23]. Indeed, preliminary results of
Mousialou and Morgenstern demonstrate such an activity of the microsomal GSH-tr [21].
T"nereore, the microsomal GSH-tr would be abie to protect against lipid peroxidation via the
same mechanism as previously proposed for the cytosolic selenium-dependent phospholipid
hydroperoxide GS1I-peroxidase (at that time named PIP),i.e. by directly detoxicating esterified
lipid hydroperoxides without the intermediary action of phospholipase A2[24].
However, there are several indications that the microsomal GSH-tr is not responsible for the
GSH-dependent protecrion against microsomal lipid peroxidaUon. For example, NEM treatment
of the microsomes stimulates the geroxidase activity of the microsomal GSH-tr, but it
inactivates the GSH-dependent protection against lipid peroxidarion [23]. Similaz result were
obtained with acrolein in stead of NEM [7]. Moreover, pretreatment of liver microsomes with
4-methyl-ortho-benzoquinone reduces the GSH-dependent protection [8], whereas the results
presented in the present study show that 4-methyl-ortho-benzoquinone activates the
mcrosomal GSH-tr. Actually, it could have been anticipated that no effective protection against
in vitro lipid peroxidarion by the GSH-peroxidase activity of the microsomal GSH-tr is
observed. By this mechanism only protection against lipid hydroperoxide-dependent lipid
peroxidation is given, while in vitro lipid peroxidation probably consists for the greater part of
lipid hydroperoxide-independent lipid peroxidarion. Previously, we suggested that the GSHdependent protection of rat liver microsomes against lipid peroxidation is more likely due to the
radical reductase-mediated regenerarion of vitamin E that has been converted into a radical[11,
23].
Aniya and Anders [14] speculated that reduced oxygen species that induce lipid peroxidation
would activate the microsomal GSH-tr, and that the resulting increased catalytic activity of the
enzyme may enhance detoxicarion of hydrogen peroxide or lipid hydroperoxides. At first,
however, it should be noted that GSH-tr, including the microsomal GSH-tr, do not accept
hydrogen peroxid as substrate in their GSH-peroxidase activity. Moreover, in this study it was
found that neither hydrogen peroxide nor superoxide anions were able to activate the
microsomal GSH-tr. Since the identity of the free radical that induces lipid peroxidarion is still
unknown, it was lso tested whether in vitro lipid peroxidation would activate the microsomal
GSH-tr. It was found that a moderate level of lipid peroxidation had no detectable effect on the
activity of the enzyme, more pronounced lipid peroxidation even inacrivated the microsomal
GSH-tr. Apparently, lipid peroxidation inhibits the microsomal GSH-tr, a conclusion
previously reached by Harris and Stone [25]. Also the reported activation of the microsomal
GSH-tr by carbon tetrachloride [26] is probably not caused by the carbon tetrachloride induced
lipid peroxidarion as suggested by Aniya and Anders [14], but more likely by alkylarion of the
free sulfhydryl group of the microsomal GSH-tr by a reacrive metabolite of carbon tetrachloride
generated in a cytochrome P-450 mediated reaction (comparable to the mechanism of
a-methyldopa). Also the carbon tetrachloride-mediated reduction of the GSH-content of the
113

liver may contribute to the activation of the microsomal GSH-tr.


In conclusion, a reacrive product of a-methyldopa generated by the cytochrome P-450
complex activates the microsomal GSH-tr. The most likely mechanism of this acrivation is
arylation of the free sulfhydryl group of the microsomal GSH-tr by a reactive product of
a-methyldopa. The product responsible for this arylarion is probably the ortho-quinone or
semi-ortho-quinone radical of a-methyldopa. These findings support the hypothesis that
reactive products formed by the cytochrome P-450 complex are able to activate the microsomal
GSH-tr and possibly in this way enhance their detoxicarion.
References.
1

B. Mannervik and U.H. Danielson, Glutathione transferaces -structure and catalytic activity, CRC CriCc21
Rev. Biochem.23(1988)283-337.
2 P.J. van Bladeren, Formation of toxic metabolites from drugs and other xenobiotics by glutathione
conjugation, Trends Pharmacol. Sci. 9(1989)295-299.
3 J.W. DePierre and R. Morgenstern, Comparison of the distribution of microsomal and cytosolic glutathione
S-transferace activities in different organs of the rat, Biochem. Pharmacol. 32(1983)721-723.
4 R. Morgenstern, C. Guthenberg, B. Mannervik and J.W. DePierre, The amount and nature of glutathione
transferaces in rat liver microsomes determined by immunochemical methods, FEBS Lett. 160 (1983) 264268.
5 R. Morgenstern, J.W. DePiene and H. J~rnvall, Microsomal glutathione transferace, primary structure, J.
Biol. Chem. 260(1985) 13976-13983.
6 A. Bast and G.R.M.M. Haenen, Cytochrome P-450 and glutathione, what is the significance of their
interrelationship in lipid peroxidation? Trends Biol. Sci.9(1984)510-513.
7 G.R.M.M. Naenen, N.P.E. Vermeulen, J.N.L. Tai Tin Tsoi, H.M.N. Ragedi, H. Timmerman and A. Bast,
Activation of the microsomal GSH-S-transferace and reduction of the glutathione dependent protection
against lipid peroxdation by acrolein, Biochem.Pharmacol. 37(1988) 1933-1938.
8 Modulation of o~cidative stress by catecholamines, previous chapter
9 E. Dybing, S.D. Nelson, J.R. Mitchell, H.A. Sasame amd J.R. Gilette, Oxidation of a-methyldopa and other
catechols by cytochrome P-450-generated superoxide anion: possible mechanism of methyldopa hepatitis,
Mol. Pharmacol. 12(1976)911-920.
10 R. Willsttter and A.Pfannestiel, Ueber o-Chinon, Chem. Berichte 4(1904)4744-4746.
11 G.R.M.M Naenen and A. Bast,Protection against lipid peroxidation by a microsomal glutathione-dependent
heat labile factor,FEBS Lett. 159(1983)24-28.
12 M. Bradford, A rapid and sensitive method for the quantization of microgram quantities of protein utilizing
the principle ofprotein-dye binding, Anal. Biochem.72(197 246-254.
13 D.S.E. Tse, R.L. McCreery and R.N. Adams,Potential oacidative pathways of brain catecholamines, J. Med.
Chem. 19(1976) 1707-1710.
14 Y. Aniya and M.W. Anders, Activation of rat liver microsomal glutathione S-transferace by reduced oxygen
species, J. Biol. Chem. 264 (1989) 1998-2002.
15 P.J. Dierickx, Interaction of 1,4-benzoquinone and 2,4-dichlorophenoxyacetic acid with microsomal
glutathione Iransferase from rat liver, Arch. Internat. Physiol. Biochim. 96(1988) 1-5.
16 M. Miranda, C. di Ilio, A. Bonfigli, A. Arcadi, G. Pitari, S. Dupre, G. Federici and G. del Boccio, A study
on the in vitro interaction between tyrosinase and glutathione S-transferace, Biochim. Biophys. Acta 913
(1987)386-394.
17 Y. Aniya and M.W. Anders, Noradrenaline activation of hepatic microsomal glutathione S-transferace,
Proceedings Eur. Workshop on Drug Metabol., Konstanz FRG (1988) abstract II-404-P3.
18 P. Graf and H. Sies, Hepatic glutathione release upon decreases of extracellular calcium concentration,
Biochem.Pharmacol. 35(1986)2832-2833.
19 T. Masukawa and H.Iwata,Possible regulation mechanism of microsomal glu[athione S-transferace activity
in rat liver, Biochem. Phannacol. 35 (1986)435-438.
20 R. Morgenstern and J.W. DePiene, Membrane-bound glutathone Iransferase, Biochem. Soc. Transact. 15
(1987)719-721.

114

21 E. Mousialou and R. Morgenstren, Studies on [he glutathione dependent inhibition of lipid peroxidation,
Proceedings Eur. Workshop on Drug Metabol., Konstanz FRG (1988) abstract 3-15.
22 G.R.M.M. Naenen, N.P.E. Vermeulen, H. Timmerman and A. Bast, Is phospholipase Az involved in the
glutathione-dependent protection against in vitro microsomal lipid peroxidation? in: O. Hayaishi, E. Niki,
M. Kondo and T. Yoshikawa (Eds.) Medical, biochemical and chemical aspects of free radicals, Elsevier,
Amsterdam, 1989, 1291-1294.
23 G.R.M.M. Naenen, J.N.L. Tai Tin Tsoi, N.P.E. Vermeulen, H. Timmerman and A. Bast, 4-Hydroxy-2,3trans-nonenal stimulates microsomal lipid peroxidation by reducing the glutathione-dependent protection,
Arch. Biochem. Biophys. 259(1987)449-156.
24 F. Ursini, M. Maiorini, M. Valente, L. Ferri and C. Gregolin, Purification &om pig liver of a protein which
protects liposomes and bomembranes from peroxidative degradation and exhibits giutathione eeroxidase
activity on phosphatidylcholine hydroperoxides, Biochim. Biophys. Acts 710(1982) 197-211.
25 C.M. Harris and W.L. Stone, The effects of in vitro lipid peroxidadon on the activity of rat liver microsomal
glutathione S-transferace from rat supplemented or deficient in anrioxidants,Life Sci. 42(1988)41~-420.
26 B. Botti, M.T. Moslen, N.M. Trieff and E.S. Reynolds, Transient decrease of liver cytosolic glutathione Stransferaseactivities in rats given 1,2-dibromcethane or CC14,Chem.-Biol. Interace 42(1982) 259-270.

115

116

- Chapter 12 ~3Adrenoceptor agonists do not reduce hydrogen peroxide


production from superoxide radicals.
Abstract. Recently, it has been reported that ~3-adrenoceptor agonists containing a catechol
moiety reduce hydrogen peroxide production by alveolar macroFhages, an effect suggested to
be due to scavenging by the catecholamines of superoxide, the precursor of hydrogen peroxide.
However, catechelamines interfere wth the method used in that study to determine hydrogen
pero~de formation. We re-examined the obtained results using two independent methods for
measuring hydrogen peroxide that are not affected by catecholamines. We found that
catecholamines, in a concentration up to 10-5 M,had no effect on hydrogen peroxide formation
out of superoxide. It was furthermore established that reduction of hydrogen peroxide formation
by scavenging of superoxide by catecholamines is not very likely.
Introduction.
Free radicals in general and more specifically reacrive oxygen species play a role in the
etiology of lung injury (e.g. Borm et al., 1986, Kramer et al., 1987). Appropriate
pharmacological treatment of lung disease may therefore aim at a reducrion of oxidative stress in
lung tissue. Macrophages form an important source of oxygen radicals. Normally, oxygen
radicals produced by phagocytic cells like macrophages form a primary defense against
microorganisms. However, excessive reactive oxygen formation by macrophages can lead to
tissue damage as well. In fact, adult respiratory distress syndrome has been suggested to result
from macrophage-induced tissue damage. Therefore pharmacological tools to decrease the
oxidative burst of alveolar macrophages are of importance in the prevention of lung tissue
damage.
Henricks et al. (1986, 1987, 1988) reported on the effects of(3-adrenoceptor agonists on
macrophage function. Initially, the observation that isoprenaline (a non-selective
(3-adrenoceptor agonist) and dabutamine (a dil-adrenoceptar agonist) inhibited the hydrogen
peroxide formation, whereas salbutamol (a (32-adrenoceptor agonist) had no effect on the
amount of hydrogen peroxide measured in a macrophage suspension, seemed to point to a
(31-adrenoceptor mediated phenomenon (Henricks et al., 1986). However, an additional report
(Henricks et al., 1987) showed that (3-adrenoceptor antagonists did not preclude the inhibitory
effect of isoprenaline on hydrogen peroxide formation. The effect of isoprenaline on
macrophage hydrogen peroxide production which was attributed to a (31-mediated response
became more enigmatical when radio ligand binding experiments showed only
(32-adrenoceptors on guinea pig alveolar macrophages(Henricks et al. 1987).
We found that the observed effects of isoprenaline can simply be ascribed to interference of
isoprenalne with the hydrogen peroxide assay which was used (Goeptar et al., 1987, 1988). In
their fust studies, Henricks et al. (1986, 1987) used the horseradish peroxidase mediated
oxidation of phenol red by hydrogen peroxide to quantify hydrogen peroxide formation. The
amount of oxidized phenol red is measured spectrophotometrically, The (3-adrenoceptor agonist

117

(-)-isoprenaline, that contains a catechol moiety, reduces the amount of phenol red oxidized by
hydrogen peroxide, while the (3adrenoceptor agonist ()-salbutamol, that has no catechol
group, has no effect(Goeptar et al., 1987, 1988). These experiments were performed at pH 7.
Recently, Henricks et al. (1988) verified these results. At neutral pH, the reducrion by
(3-agonists of the amount of hydrogen peroxide detected in an alveolaz macrophage suspension
was similar to the effect of(3-agonists on the hydrogen peroxide mediated phenol red oxidation.
Additionally, it was stated that at pH 6, ~iagonists like (-)-isoprenaline had "almost no effect
on the detection of hydrogen peroxide" by the horseradish peroxidase method (Henricks et al.,
1988). Based on results obtained at this pH, it was concluded that ~iadrenoceptor agonists
containing a catechol moiety decrease hydrogen peroxide praluced either by macrophages or by
the combination of hypoxanthine-xanthine oxidase via sca:~enging of superoxide radicals
(Henricks et al., 1988). So, by a small reduction of the pH -from 7 to 6 -the phenol red assay
would become selecrive and the effect of catecholamines would change from an interference of
the assay to scavenging of superoxide. However, in contrast to the statement that ~3-agonists
had almost no effect on the detecrion of hydrogen peroxide, it was reported in the same paper
that the addirion of(3adrenoceptor agonists resulted in "a decrease of 0-20 % in the amount of
hydrogen peroxide that was detected at pH 6" (Henricks et al., 1988). Since apparently the
phenol red method is also at pH 6 susceptible to disturbances, we re-examined the obtained
results using two independent methods for measuring hydrogen peroxide, that are not affected
by (3adrenoceptor agonists.
Materials and methods.
Chemicals. Chromatographically purified xanthine oxidase (grade III), hypoxanthine, catalase and (-)isoprenaline hydrochloride were obtained from Sigma, St. Louis, USA.()-Dobutamine hydrochloride was a gift
of Lilly Industries Ltd, Basingstoke,England, LTK.
Hydrogen peroxide determinations. Hydrogen peroxide production by the combination of xanthine oxidase
and hypoxanthine was determined using two independent methods that proved not to be affected by the
catecholamines in the concentrations used. The incubations were performed in an air-saturated 10 mM.phosphate
buffer pH 6, containing 140 mM NaCI and 5.5 mM glucose, at 37 C,for 5 min. In the first method, hydrogen
peroxide was assessed with the iron-thiocyanate method according to Hildebrandt and Roots (1975). In the second
method, the time course of oxygen pressure was recorded with a Clazk oxygen electrode connected to a Yellow
Springs Instrument Co. model 5300 biological oxygen monitor. By the addition of catalase, hydrogen peroxide
production was quantified. The water used in his study was obtained from a Millipore QT"' column.

Results.
The effect of (3-adrenoceptor agonists on the formation of hydrogen peroxide out of
superoxide anion radicals (02-) at pH 6 was determined. Superoxide anion radicals were
generated by the combination of0.5 mM hypoxanthine and 0.01 units / ml of xanthine oxidase.
Hydrogen peroxide is formed by the spontaneous dismutation of superoxide anion radicals.
Two molecules superoxide produce one molecule hydrogen peroxide.
'- + 2 H+ ~ +
2 02
OZ H2O2
In the first method for measuring hydrogen peroxide, it was determined with the ironthiocyanate method according to Hildebrandt and Roots (1975). Hydrogen peroxide oxidizes
Fel+ to Fe3+, and Fe3+ chelated by SCN- is spectrophotometrically determined.
(-)-Isoprenaline and ()-dobutamine, in a concentration up to 10-5 M,proved not to influence
this assay (data not shown). It was found that after the addition of 10-5 M (-)-isoprenaline or
10-5 M ()-dobutamine, hydrogen peroxide production by hypoxanthine-xanthine oxidase was
118

A
T

x.o.

pO2

vi

i,

~
caT.

B
~

x.o.

,~

T
CAT.

time
Fig 1. Time course of oxygen pressure (p02). Addition of 0.06 U / ml xanthine oxidase (X.O.) to a
solution containing 0.5 mM hypoxanthine resulted in a gradual consumption of oxygen. Addition of 5.5
U / ml catalase(CAT.)increased oxygen tension and reduced the rate of oxygen consumption (trace A). In
trace B,10-5 M (-)-isoprenaline was added.

respectively 1Q0 5 %(mean S.D,, n=4) and 100 5 %(n=2) of the control value. Also of
other concentrarions of these catecholamines in the range of 10-~ to 10-5 M, no effect was
observed (data not shown).
In the second method, oxygen consumption was monitored using a Clark oxygen electrode.
T'he combination of hypoxanthine-xanthine oxidase resulted in a decline of the oxygen pressure
(fig. 1), that was linear within the time span studied. Addition of catalase increased the oxygen
tension and reduced the rate of oxygen consumption (iig. 1). The increase of oxygen tension by
catalase was dFternuned by extrapolating the oxygen consumprion after the addirion of catalase
to the time at which catalase was added. The rise was defined as the difference between the
actual oxygen tension at the moment at which catalase was added and the extrapolated oxygen
tension. It was found that catalase regenerated approximately half of the oxygen consumed(60
7 %, n=4). The rate of oxygen consumption after the addition of catalase was half of the
initial oxygen consumprion (51 5 %,n=4}. Catalase catalyzes the decomposition of hydrogen
peroxide in oxygen and water
2H2O2~2H2O +02
regenerates
half of the oxygen that is entrapped as H2O2. Since the
reaction
catalase
By this
above reported effects were observed by the addirion of catalase, it can be concluded that
pracrically all the oxygen consumed by hypoxanthine-xanthine o~dase is converted to hydrogen
peroxide.
Addirion of(-)-isoprenaline in a concentration of 10-~ to 10-5 M had no effect on the rate of
oxygen consumption either before (10-5 M (-)-isoprenaline: 101 5 % of the control value,
n=4) or after (10-5 M (-)-isoprenaline: 101 6 % of the control value, n=4) the addirion of
catalase or on the rise of oxygen tension by catalase(10-5 M (-)-isoprenaline: 101 5 % of the
control value, n=4).

119

Discussion.
In aerobic organisms oxygen is normally reduced to water by the enzymes of the respiratory
chain. However, there are several enzymes that cause only a partial reduction of oxygen. An
example of such an enzyme is the NADPH-oxidase, an enzyme that is located on the membrane
of phagocytic cells like macrophages. Upon stimulation of the macrophage, the one electron
reduction of oxygen to superoxide is induced in a NADPH-oxidase catalyzed reaction. The
formed superoxide can dismutate and in this way hydrogen peroxide is produced. The
producrion of partially reduced oxygen species contributes to the bacteriocidal action of
macrophages. However, in their toxicity the partially reduced oxygen species are not
discriminatory, beside the bacteria the surrounding tissue is vulnerable as well. It has been
suggested that excessive stimulation of alveolar macrophages is involved in several lung
diseases (Boren et al., 1986, Kramer et al., 1987).
Recently, it has been reported that catecholamines reduce hydrogen peroxide formation from
macrophages or from the combination of hypoxanthine and xanthine oxidase (Henricks, 1988).
We (Goeptar et al., 1987, 1988) pointed out that, no receptor mechanism is involved in the
inhibition of hydrogen peroxide production in macrophage suspensions. This has been
confirmed, and alternatively, it was stated that catecholamines reduce hydrogen peroxide
formarion by scavenging of superoxide (Henricks et al., 1988). Since catecholamines interfered
with the method to determine hydrogen peroxide production (Henricks et al., 1988), we reexamined the reported effect using two independent methods for measuring hydrogen peroxide
that were not affected by catecholamines. We observed that catecholamines in concentration up
to 10-5 M do not reduce the hydrogen pero}cide production by scavenging superoxide radicals.
In order to explain the putative scavenging effect of catecholamines it was put forth that
"The superoxide anion ... has the ability to abstract a proton from the catechol moiety of
catecholamines", that finally results in a decreased hydrogen peroxide production (Henricks et
al., 1988), an effect however we were not able to detect (vide supra). To support the theory of
proton abstracrion, reference to Nanni et al.(1980) was made. In this respect it should be noted
that the experiments described by Nanni et al. (1980) were conducted in deoxygenated
dimethylformamide. In this aprotic solution, superoxide anion radicals function as a Br~nsted
base and in this medium the first step in the oxidarion of catecholamines by superoxide anion
radicals is a proton transfer (Nanni et al., 1980). Henricks et al.(1988) extrapolated the results
of Nanni et al. (1980)to an air-saturated aqueous solution that was buffered at pH 6. However,
in this amphiprotic medium it is unlikely that the addition of catecholamines decreases the
percentage of superoxide ionized, since the pH was buffered. Moreover, it remains obscure
why protonation of the superoxide anion will result in a diminished hydrogen peroxide
producrion. Nanni et al. (1980) reported that in the aproric medium the addition of substrates
like catechols to superoxide anions and the subsequent protonation of superoxide anions notably
provokes hydrogen peroxide production. In the first reaction, superoxide anion abstracts a
proton from the substrate to give the substrate anion and the dismutation products of
superoxide, hydrogen peroxide and oxygen. In turn, the substrate anion is oxidized by oxygen
in a multistep process to yield substrate oxidarion products and again hydrogen peroxide (Nanni
et al. 1980).
A comparable mechanism for the reaction of superoxide radicals with catecholamines has
been described in aqueous solutions (e.g. Schenkman et al., 1978). In this reaction scheme one
120

HO ~ I

CH ~
`2
~CH2

202 -p / I
.~ ~

CH

O
~2

~CH2

semiquinone R

catecholuiine R

CH

/ ~CHZ

quinone
~
HO / I
HO ~

R
R
CH
i
N,CHZ

leucochrome R
02

H2~2
R'
~

'p \

CH

R'
~

~ ~CH2 ~ ~
N+
R

aminochrome

CH

/~~~CHZ
N
R

~2

~Z r~ /

~1

CH

p~ I:
N Hz
semiquinone R

Fig. 2. Reaction between catecholamines and superoxide (after Schenlanan et al., 1978).

molecule of hydrogen peroxide is formed by one molecule of superoxide (fig. 2), in contrast to
the spontaneous dismutation of superoxide in which two superoxide molecules are needed for
one hydrogen peroxide. Moreover, the semiquinone radical that is formed hereby is capable to
react with oxygen, a reaction in which oxygen accepts an electron of the semiquinone and
superoxide is regenerated (fig. 2). The formed quinone can cyclisize to give a leucochrome.
This leucochrome is -comparable to the catecholamine -capable to react with superoxide, which
finally results in the regeneration of superoxide and the formation of aminochrome and
hydrogen peroxide (fig. 2). Actually superoxide is a classical catalyst in the oxidarion of
catecholamines in which aminochrome and hydrogen peroxide are formed. A reaction of
superoxide radicals with catecholamines would result in an increase rather than a reduction of
hydrogen peroxide production.
Also the concentrarion of the catecholamines that already produced. an effect, indicates that no
reduction of hydrogen peroxide formation by scavenging superoxide radicals takes place. At a
concentration of 10-6 M,prenalterol (notably not a catecholamine) was found to reduce more
than 50 % of the phenol red oxidation by phorbol myristate acetate-stimulated macrophages
(Henricks et al., 1988). Control hydrogen peroxide production was 19.3 1.7 10-6 M. In
order to give a 50 %reduction of the hydrogen peroxide formarion, 10-6 M prenalterol had to
scavenge approximately 20 10-6 M superoxide radicals. This means that each scavenger
molecule had to react with twenty molecules of superoxide, which is very improbable.
In conclusion, catecholamines, in concentrations up to 10-5 M, do not reduce hydrogen
peroxide formation from superoxide radicals.

121

References.
Borm, P.J.A., Bast, A., Wouters, E.F.M., Slangen, J.J., Swaen, G.M.H. and De Boorder, Tj. Red blood cell
anti-oxidant parameters in silicosis, Int. Arch. Occup. Hlth. 58, 235-244 (1986).
Gceptar, A.R., Naenen, G.R.M.M., Timmerman,H,and Bast, A. The effects of ~i-adrenergic receptor agonists on
the H2O2 formation in alveolar macrophage suspensions are not mediated by R-adrenoceptors, Pharmaceut.
Weekblad Sci. Ed. 9, 340(1987).
Gceptar, A.R., Naenen, G.R.M.M., Timmerman,H.and Bast, A. The effects of fi-adrenergic receptor agonists on
the H2O2formation in alveolar macrophage suspensions are not mediated by (3-receptors, Agents Actions, 25,
375-377(1988).
Henricks, P.A.J., Van Esch, B. and Nijkamp, F.P. ~i-agonists can depress oxidative metabolism of alveolar
macrophages, Agents Actions 19, 353-354(1986).
Henricks, P.A.J., Van Esch, B., Van Oosterhout, A.J.M. and Nijkamp, F.P. ~3-adrenergic agonists diminish
hydrogen peroxide release of guinea pig alveolar macrophages, X~ Internat. Congress ofPharmacol., Sydney,
Abstract 0259(1987).
Henricks, P.A.J., Van Esch, B., Van Oosterhout, A.J.M. and Nijkamp, F.P. Specific and non-specific effects of
(3-adrenoceptor agonists on guinea pig alveolar macrophage function, Eur. J. Pharmacol. 152, 321-330(1988).
Hildebrandt, A.G. and Roots,I. Reduced nicotinamide Binucleotide phosphate(NADPI~-dependent formation and
breakdown of hydrogen peroxide during mixed function oxidation reactions in liver microsomes, Arch.
Biochem. Biophys 171, 385-397 (1975).
Kramer, K., Doelman,C.J.A., Timmerman,H. and Bast, A. A disbalance between beta-adrenergic and muscarinic
responses caused by hydrogen peroxidae in rat airways in vitro, Biochem. Biophys. Res. Comms. 145, 357362(1987).
Nanni, E.J. jr., Stallings, M.D. and Sawyer, D.T. Dces superoxide ion oxidize catechol, a-tocopherol and
ascorbic acid by direct electron transfer? J. Amer. Chem. Soc. 102,4481-4485 (1980).
Schenkman, J.B., Jansson, I., Powis, G. and Kappus, H. Active oxygen in liver microsomes: Mechanism of
epinephrine oxidation, Mol. Pharmacol. 15,428-438 (1978).

122

PaYt III Oxidative stress and receptorfunction

123

124

- Chapter 13 Quantification of agonist-receptor interaction.


Introduction.
Living organisms have to maintain their internal environment within narrow limits. In order
to produce a homeostasis, they have to gather nutrients, dispose of waste products and regulate
oxygen tension, pH and temperature. In doing so, the organisms have to adapt to changes in the
external environment. In mulri-celiuiar organisms, the ce11s are differentiated and fulfill
specialized tasks. For maintenance of a homeostasis, the different activiries of the cells have to
be coordinated. For this purpose, mulri-cellular organisms like mammals are equipped with
sensors that gather information, and transmit signals to the central nervous system. There the
information is integrated, and output signals to effector organs are generated, if necessary. Also
local auto-regulatory mechanisms exist. By these central and local mechanisms adequate acrion
is taken to preserve a homeostasis of the internal environment.
The two most important ways for the transmission of information proceed via nerves or
hormones. Although there are several differences between these ways,the principle method for
transmission is identical. One cell excretes a chemical substance -called hormone or
neurotransmitter -that interacts with acomponent -called receptor - of another cell. By this
interacrion the acrivity of the other cell is modified. Despite that hormones are normally excreted
in the blood and reach all organs, the effect of a hormone can be focused on a certain organ or
even on a certain type of cells of an organ. Apparently a hormone (=messenger) is capable to
recognize the addressed cell. This is caused by the occurrence of receptors for the hormone
exclusively on the target cells. Receptors are proteins that have a high affinity for a certain
hormone. Each hormone has its own receptor. It is also possible that different (sub)types of a
receptor for one hormone exist. A neurotransmitter released by nerve terminals, also interacts
with a receptor of the target cells) for the transmission of information. However,in this case
organ selecrivity is primarily the result of the confined space the neurotransmitter is released in.
Binding of a hormone or neurotransmitter to its receptor is not sufficient for signal
transduction. By the interacrion between receptor and hormone or neurotransmitter, achange of
the receptor has to be induced, that initiates a cascade of reactions what finally results in the
response. An example of such a cascade is given in figure 1. During the reacrions that are
triggered by receptor activation the input signal, that is produced by binding of an hormone or
neurotransmitter to the receptor, is transduced over the cell membrane to the cell interior, since
often the receptors are located on the cell surface. Moreover, the input signal is modulated and
amplified so that a relatively small number of receptors is able to generate a physiological
relevant effect [Arins et al., 1979].
The pharmacological effect produced by numerous drugs is the result of a direct interaction
with a receptor. A discrimination can be made between agonists and antagonists. Both
agonists as well as antagonists bind to the receptor. Agonists mimic the action of the
endogenous neurotransmitter or hormone, since they are capable to activate the receptor.
Antagonists are not able to acrivate the receptor, they prevent binding and subsequent activation
of the receptor by its endogenous ligands. By a pharmacological intervention with agonists or

125

noradr+neline_~ receprorcomplex
-adrenoceptor
Adenylate
cyclase

iNacnve
Adenylate
cyclase
ACTIVE

ATP

c-AMP
Protein
kinase
~NncnvE
Protein
kinase
ACTIVE
Regulatory
proteins
Regulatory
proteins
PHOSPHORYLATED

O
Cnchanels
Positive isotropic
effect

Fig. 1. Schematic representation of the cascade of biochemical reactions involved in the (3-adrenoceptormediatedpositive isotropic effect of the neurotransmitter noradrenaline in the heart. By each reaction the
input signal is amplified.

antagonists, an unbalanced physiological coordinarion of organ activity might be corrected. In


this way it can be attempted to prevent that the composition of the internal environment exceeds
its physiological limits, with as extreme consequence funcrion loss and cell death.
In order to study the pharmacological effect of several drugs, drug-receptor interacrion has to
be quantified. Drug action in an intact animal is too complex for the precise characterization of
the effect of drug on a certain organ. Therefore, the pharmacodynamic phase of drug action is
normally quantified in vitro, thus eliminating the pharmaceurical and phannacokinetic phases of
drug action and feed-back regulation mechanisms of the animal. For agonists in vitro
quantification usually is performed by constructing adose-response curve of the drug on an
isolated tissue preparation that contains the receptor the agonist interacts with. The response of
the agonist to various concentrations of the drug is recorded. The effect of the drug is usually
denoted by the concentrarion of the drug that produces half maximal effect(D2) or the pD2
(-log D2) and by the maximal effect of the drug that is related to the maximal effect of another
drug (intrinsic activity, aE). However parameters as pD2 and intrinsic activity (aE) are not
truly drug-dependent parameters since they also depend on the rissue used to construct the dose126

response curve, despite the fact that the same receptor mediates the response (table 1). Actually,
the action of a drug in isolated tissue is deternuned by the affinity of the drug for the receptor,
by the magnitude of activarion of the receptor by the drug and by the way that the tissue
translates receptor acrivarion into an effect.
Stephenson was the first to expound this concept, and he tried to unravel the different steps
in the pharmacodynamical phase of drug action [Stephenson, 1956]. Fust the drug has to bind
to the receptor. Usually this is a reversible process, and in the simplest case this is the reversible
binding of the drug(A)on the receptor (R),fornung adrug-receptor complex (AR).
A+RAAR
In equilibrium, the drug-receptor interacton can be described by the dissociation constant
~A)
[A].[R]
KA = ~~~

Thefraction of the receptors that are occupied by the drug (y) depend on the concentration of
the drug.
[AR]
[A]
y [AR]+[R] [A]+ KA
Binding of the drug (A) to the receptor (R) might activate R. According to the theory of
Stephenson, receptor activation produces a certain input stimulus (S). The magnitude of the
srimulus induced by a certain drug-receptor complex depends on the drug used. Antagonists - in
contrast to agonists -are not able to activate a receptor and consequently antagonists do not
induce an input stimulus when they bind to a receptor. Moreover, the input stimulus generated
by binding of a full agonist is greater compared to that of partial agonists (the terms full and
partial agonist will be reviewed later on). Stephenson introduced the term efficacy (e) as the
term that expresses the extent of receptor activation by binding of a certain drug to a particular
receptor. In the concept of Stephenson it is assumed that the contribution of a single receptordrug complex to the input stimulus does not dependent on the amount of receptors occupied.
This means that the input stimulus(S) that is produced when a drug binds to a receptor depends
on the efficacy of the drug (e) and the fraction of the receptors that are occupied by the drug (y)
[Stephenson, 1956].
S=ey
Finally, the stimulus that is generated by binding of an agonst to the receptor has to be
translated into a physiological response. The prcess that is responsible for the posirive
inotropic effect of the neurotransmitter noradrenaline in the heart is depicted in figure 1. The
process between receptor activation and effect can be described by deternuning how each
reaction in the cascade transforms its own input stimulus. When these functions are combined,
it is known how the srimulus is transformed into an effect. However, this approach is too
complicated to be practically applicable. Stephenson tackled this problem by studying the
stimulus-effect transfer as a whole. The stimulus(S)is transferred via an unknownfunction (fl
into a response (E =the response relative to the maximal receptor mediated response), in
formula:
E = f(S)
127

In the concept of Stephenson, no attempt is made to obtain a mathematical relation between


the input stimulus and the final effect, but "an empirical relarion can be obtained experimentally"
[Stephenson, 1956]. The relation between srimulus and effect is usually not identical in different
tissues that contain a certain receptor, while the affinity of the drug for the receptor(KA)and the
magnitude of receptor activation of the receptor by the drug is exclusively deternuned by a drugreceptor interaction; thus affinity and receptor acrivation are independent of the tissue used
[Stephenson, 1956; Kenakin, 1987]. Despite the fact that the dissociation constant(Kq) and
efficacy (e) are better parameters than pD2 and intrinsic (aE) to describe drug-receptor
interaction, they are rarely used for this purpose. This is probably because they are difficult to
determine. Here, we discuss some of the methods that can be used to determine the affinity and
efficacy. Moreover, other terms used to quantify drug-receptor interaction will be reviewed
also. The receptor mechanism presented is, in principle, limited to the function of the
~3-adrenoceptor. Although, it is attempted to produce a more general receptor mechanism, some
of the assumptions (implicitly) made, may not hold for other receptor types.
Material and methods.
Male wistar rats (200-250 g, C.P.B. Hazlan Olec Zeist, the Netherlands) were killed by decapitation.
Thyroxine (T4)treatment was 100 g T4/100 g for 7 days. The isolated left and right atria and right ventricle
strip were mounted in water jacketed organ baths. The dose-response curve of (-)isoproterenol in these organs
were constructed as described in Naenen et al.(1989).
Heart membranes were preparec from heart left atria according to the method previously described (Naenen et
al. 1988). The membranes were taken up in buffer A (50 mM Tris-HC1,140 mM NaCI and 5 mM MgC12, pH
7.4 at 37 C)and stored in liquid nitrogen until use. The receptor binding assay were performec with membranes
suspended in buffer A(pH 7.4 at 37 ~C)as described previously(Naenen et al. 1988).
The chemicals used were (-)-timolol maleate, cumene hydroperoxide(CHP)and guanylylimidodiphosphate
(GppNHp)(Sigma).(-)-[l25]-iodocyanopindolol(ICYP) was obtained from New England Nuclear and the cyclicAMP assay kit(code TRK.432) was obtained from Amersham. All other chemicals were of reagent grade.

Results and discussion.


Receptor-mediated,effect of(-)isoproterenol in the left atrium.
Addition of (-)isoproterenol to an electrically stimulated left atrium results in a
concentration-dependent increase of the contracrion force. It has well been documented that this
effect of(-)isoproterenol is mediated by activation of a (3-adrenoceptor located on the left atrium.
The concentration of(-)isoproterenol that produces half maximal effect(D2)in the left atrium of
the rat is 3 10-9 M. Using a radioligand binding study, it was found that the dissociation
constant(Kp)of (-)isoproterenol for the (3-adrenoceptor is 3 10-~ M in membranes prepared
of the left atrium of the rat. This value was identical to the dissociation constant of
(-)isoproterenol previously obtained [Abrahamson, 1986].
In order to calculate the efficacy (e)of(-)isoproterenol, we have to make use of the deimition
of Stephenson that the value of the stimulus (S) is unity at half maximal receptor-mediated
response (S = 1 when E = 0.5)[Stephenson, 1956]. Maximal receptor-mediated response is
obtained with the most powerful agonist(s), the full agonist(s). With a full agonist (but not a
partial agonist) half maximal receptor-mediated response is obtained at the pD2, the
concentration agonist is then D2. When the dissociation constant(Kp)is known,the efficacy of
the full agonist(A)can be calculated:
e=

D2 + KA
D
2

128

()terbutaline

(-)isoproterenol
PD2

orb

as

left atrium

8.5 0.2

110 30

right ventricle

7.6 0.1

13 3

PD2

ocE

ocs

5.2 0.2 0.81 0.07 3.1 0.8 0.028 0.006

n.d.

n.d.

n.d.

n.d.

Table 1. The pD2,, aE,efficacy (e)and as of the positive inotropic effect of(-)isoproterenol and ()terbutaline in
the isolated left atrium and the isolated right ventricle of control rats. n.d. is not determined.

Since (-)isoproterenol is a full agonist, it is possible to calculate the efficacy of(-)isoproterenol


with this formula, the efficacy of(-)isoproterenol in this rissue is 110(table 1). Semingly, the
pI32 is determined with greater accuracy compared to the efficacy (table 1). However, the lower
relarive error of the pD2 compared to that of the efficacy is due to the fact that the pD2 is a
logarithmic term. By the method we applied to calculate the efficacy of (-)isoproterenol, the
variation of this parameter is equal to that of the D2. It has been suggested to express receptor
activation as the logarithm of the efficacy, since a logarithmic transformation stabilizes the
variance and normalizes the distriburion of this variable [Abrahamson, 1986]. The standard
deviation of the logarithm of the efficacy is idenrical to that of the pD2, since we took no
account of the standard deviation of the dissociation constant, and D2 is much smaller than the
dissociarion constant.
In the dose response curve we observed that (-)isoproterenol at a concentration of 3 10-g M
already produced a maximal response, while it can be calculated that at this concentration only
10% of the receptors is occupied by (-)isoproterenol. Addition of more (-)isoproterenol does not
result in a greater effect, despite the fact that by increasing the concentrarion of(-)isoproterenol
the fraction receptors occupied (y)increases and hence also the input stimulus(S)increases.
This phenomenon, known as receptor reserve or spare receptors, can be explained by looking
at the nature of the stimulus-effect transfer (f(S)) [Arins, 1979; Kenskin, 1987]. With a
gradual increase of the input stimulus, one of the enzymes in the cascade of biochemical
reactions that processes the stimulus into the effect will be the first to reach saturation, although
the input srimulus can Brill be far from its maximal level. A further enhancement of the input
stimulus, although giving an increase in the steps preceding the limiting one, will not result in a
further increase in the final effect because the saturated reaction functions as a bottle neck
[Arins et al., 1979].
Receptor-mediated effect of(-)isoproterenol in the right ventricle strip.
Also in an isolated right ventricle strip, (-)isoproterenol produces a positive inotropic
response by (3-adrenoceptor activation, the pD2 is 7.6 (table 1). In this tissue 3 10-~ M
(-)isoproterenol was needed to produce maximal effect. This indicates that(-)isoproterenol has
to bind to 50% of the receptors in order to saturate the bottle neck reaction in the right ventricle
strip, compared to 10% in the left atrium. It should be noted that, although the same sequence of
reactions mediate the positive inotropic effect of(-)isoproterenol in both rissues, this does not
mean that the same biochemical reaction becomes the limiring one in the (-)isoproterenol
mediated response in both rissues. The difference in pD2 indicates that (-)isoproterenol has to
occupy a higher fraction of receptors in the ventricle strip, compared to the left atrium, to
produce half maximal effect, this difference is reflected in the different efficacies of
129

(-)isoproterenol in both prepararions. This does not imply that (-)isoproterenol does not acrivate
the ~3-adrenoceptor in the right ventricle as good. as the (3-adrenoceptor in the left atrium, but it is
caused by the calibration point used in the stimulus-effect transfer (S = 1 when E = 0.5). The
difference in efficacy of(-)isoproterenol in both tissues denotes that each (3-adrenoceptor on the
right ventricle has to be stimulated more, compared to the (3-adrenoceptors located on the left
atrium, to produce half maximal effect and hence a stimulus of 1.
As mentioned above,in the left atrium the maximal receptor-mediated effect is obtained when
merely 10% of the receptors are occupied with (-)isoproterenol. This means that(-)isoproterenol
in this rissue has 90% receptor resrve for the ~3-adrenoceptor,in the right ventricle strip there is
50% reserve. Receptor reserve is not a good parameter for the quanriiication of agonist-receptor
interaction. The amount of reserve of a certain receptor depends on the rissue used, and also on
the sensitivity used to record the effect, that is the degree in which enhancement of the effect by
addition of more agonist can be detected on the recorder. The point in the dose-response curve
used for calibrarion of the stimulus is better chosen,for the concentration of a full agonist that
produces half meimal receptor-mediated effect can usually be determined with the smallest
relative deviation.
Moreover, the term receptor reserve or spaze receptors suggests that some receptors do not
contribute to the effect. This is however not true. Usually the binding of a drug to a receptor is
an equilibrium reaction. The dissociation constant is the quotient of two rate constants. When
the concentration of a drug is equal to its dissociarion constant, this means that the rate of
association of a drug with a receptor is identical to the dissociation rate of the drug-receptorcomplex. Therefore, under this condition, at each time 50 % of the receptors are occupied by the
drug, on the average. However, at different rimes not the same group of receptors is occupied.
When a drug occupies 50% of the receptors this means that during a certain period a receptor is
occupied, on the average, during half the time. By analogy, a receptor reserve of an agonist in a
tissue of 50% does not mean that only half of the receptors aze involved in producing the
maximal effect, but it means that for generating the maximal effect all receptors need to have
only half maximal activation by that agonist, i.e. during a certain period the receptor is activated
by that agonist, on the average, for half the time. In the simplest concept, there are only two
possibilities for the receptor, either the agonist is bound to the receptor and the receptor can
become activated which results in a stimulus, or the receptor is free and no stimulus is
produced. Increasing the concentration of the agonist increases the period that the receptor is
occupied during a certain rime span, and hence it increases the stimulus produced by the
receptor.
Receptor-mediated effect of(+)terbutaline in the left atrium.
Like (-)isoproterenol,()terbutaline induces a ~3-adrenoceptor mediated positive inotropic
effect in the isolated left atrium (table 1). In the left atrium of control rats, the pD2 of
()terbutaline is lower than the pD2 of(-)isoproterenol. Moreover,()terbutaline is not able to
induce the maximal effect induced by (-)isoproterenol, the intrinsic activity (aE) of
()terbutaline is 0.81 (table l). This makes ()terbutaline a partial agonist. According the
theory of Stephenson, the effect generated in a certain rissue is a function of the input stimulus
(E = f(S))[Stephenson, 1956]. This implies that when two agonists produce the same effect,
this is generated by an identical stimulus. In the right atrium of control rats, the maximal
stimulus induced by ()terbutaline is not high enough to induce the maximal ~3-adrenoceptor130

mediated chronotropic effect, or in other words, the maximal activarion of all receptors in the
right atrium by ()terbutaline is not sufficient to saturate a reaction in the cascade of reactions
that transfer the stimulus into the effect. In contrast, with (-)isoproterenol maximal
~3-adrenoceptor-mediated effect in this tissue is already obtained when a small fraction of the
receptors is occupied, submaxima) activation of the receptors by (-)isoproterenol already
saturates the limiting reaction. The observation that (-)isoproterenol is more effective than
()terbutaline in ~3-adrenoceptor activation is expressed in the greater efficacy of
(-)isoproterenol.
Determination ofthe efficacy ofa partial agonist
It is possible to calculate the efficacy of a partial agonist -comparable to the method
described for a full agonist - by combining the results of a functional study with the dissociation
constant(K~,) obtained in a radioligand binding study. When applying this metha,it should be
noted that half maximal receptor mediated effect is not produced at the pD2 of a parrial agonist.
A problem is encountered when the efficacy of a partial agonist is not high enough to produce
half ma7cimal receptor mediated effect, i.e. when the efficacy of the partial agonist is smaller
than 1. However, alternative methods are available for the determination of the efficacy of
partial agonists.
In the method employed by Stephenson, the dualistic character of partial agonists is used
[Stephenson, 1956]. Partial agonists are able to activate the receptor, however, they are less
effecrive in activaring the receptor than full agonists [Arins, 1954]. When a partial agonist
occupies a receptor, this produces a stimulus, but when the partial agonist replaces a full agonist
on the receptor, the net effect of receptor binding by the partial agonist is a reduction of the
srimulus. In this case the partial agonist funcrions as a partial antagonist. When the effect of a
partial agonist on the dose-response curve of a full agonist is detemuned, the efficacy of the
partial agonist can be calculated [Stephenson, 1956].
Another method to determine the efficacy of a partial agonist makes use of the assumption
that two agonists at equi-effective concentrations produce the same srimulus. Of a partial and a
full agonist the concentrations at which they produce an identical effect are determined. When
the inverted values of these concentrarions are plotted against each other, a straight line has to be
obtained. The intercept and the slope of the line can be used to calculate the affinity of the partial
agonist for the receptor, and the efficacy of the partial agonist relative to the efficacy of the full
agonist[Mackay, 1966]. This method has gained much attention recently. It has been called the
"null method",in order to distinguish it from the "operational method". The difference between
these two methods is that in the "null method" no assumptions are made about the relation
between the stimulus and effect[Mackay, 1988]. In the "operational model" it is assumed that
the stimulus-effect curve can be fitted by a rectangular hyperbola [Leff, 1988; Mackay, 1988].
In principle, the "operational model" is restricted in application by this assumption. For
example, in tissues where the stimulus has to reach a "threshold value" before initiating a
biological response, the "operational model" cannot be used. Therefore, -the "null method"
should be preferred [Mackay, 1988]. However, it has been shown that for several receptor
systems, including the (3-adrenoceptor, the experimental data very well fit the "operarional
model"[Leff, 1988].
It is also possible to determine the efficacy (ep) and dissociarion constant(Kp) of a partial
131

~.o

P~= 8.0

U
0.$

W
t`
~.~
-1~

-
log [F]

KF= 1 10-5 M

i.o
U
`~~+ ~.$
W

0.~

10

Stimulus
~

-0p = 2 -

~ ~ i
KP= 1' 10-~ M

i.o
pI.~=7.5

~o

5. 10~~

1. 10-6

0.5

LP]

aE= 0.67

W
0.0
-10

-8

-6

log [P]
Fig. 2. Determination of the dissociation constant and efficacy of a partial agonist(P) with the use of a
stimulus-effect relation. In order to obtain astimulus-effect relation, the dose-response curve of a full
agonist (F) has to be transformed. Hereto the dissociation constant of the full agonist(KF)has to be
known. With this dissociation constant, the stimulus produce by each concentration of the full agonist
can be calculated. The stimulus-effect relation is produced when the stimulus generated by a certain
concen~ation of the full agonist is plotted against the effect that is obtained with this concentration of the
full agonist. By using this stimulus-effect relation, the dose-response curve of the partial agonist can be
transformed into adose-smulus curve for the partial agonist The maximal stimulus in the dose-stimulus
curve of the partial agonist is the efficacy of the partial agonist (ep), the concentration of agonist that
generates a stimulus equal to half the efficacy is equal to the dissociation constant of the partial agonist
(KP).

agonist(P)for a certain receptor by constructing the stimulus-effect relation(E =f(S))(fig 2).


The dose-response curve of a full agonist(F) can be transformed into astimulus-effect curve,
when the dissociation constant(KF) of the full agonist for the receptor is known. Sinc the
dose-response curve gives the concentration agonist that produces half maximal receptormediated effect (D2), the efficacy of the full agonist (eF)can be determined. For each
concentration used in the dose-response curve, the fracrion receptor occupied by the full agonist
132

(yF)can be calculated, and hence also the stimulus(S = eF yF)generated by that concentration
is known. The stimulus-effect curve is obtained when the effect produced by a certain
concentration of the full agonist is plotted against the srimulus produces by that concentration
(fig 2). Using this stimulus-effect curve, the dose-response curve of the partial agonist can be
transformed into adose-stimulus curve,for the stimulus-effect curve is idenrical for all agonists
(provided that the same tissue was used and the effect is mediated via the same receptor). The
dose-stimulus curve obtained is a direct reflection of binding of the partial agonist (P) on the
receptor, and in the simplest model the dose-stimulus function is:

LP]

S =eP ~P~ +KP


With this method. both the efficacy(ep)and the dissociation constant(Kp)of the partial agonist
can be determined. Actually only the intrinsic activity(aE)of the partial agonist(P)is needed to
determine the efficacy of the partial agonist(ep) using this method. Of course no srimulus effect
relation is needed to detemune the efficacy of an agonist with an intrinsic activity(aE) of 0.5;
the efficacy of this partial agonist is according to the definition of Stephenson 1. The stimuluseffect relation constructed with a full agonist with a high efficacy is independent of the
dissociarion constant(KF) of the full agonist (F)for the receptor. The prerequisite for this is
that only a small fraction of the receptors has to be occupied for generation of the maximal
effect, or in other words there has to be a high percentage of "spare" receptors. This implies that
the concentration of agonist needed for maximal effect and hence also the concentrations
employed in the dose-response curve are much lower than the dissociation constant ([F]
KF). This means that the efficacy of the full agonist is:
ep = KF /D2
the full agonist (yF)in the concentrations used to
by
The fracrion of receptors occupied
construct the dose-response curve can be calculated using the formula:
yg = LF~ / KF
Under these condirions the stimulus only depends on the concentration of full agonist that
produces half meimal effect(D2)and the concentration employed in the dose-response curve
~~~)~
S = eF yF = LF7 /D2
Under this condition, the shape of the stimulus effect curve is identical to the shape of the doseresponse curve of the full agonist. A radioligand binding study is only needed to verify the
assumprion that there is a high percentage of "spare" receptors.
The method described in figure 2for determination of the efficacy and affinity of the partial
agonist is in fact a variation of the "null method" described by Mackay (vide supra). No
assumption has to be made about the shape of the srimulus-effect relation. In the "null method"
according to Mackay, the data are transformed in order to produce a linear relationship. The
advantage of such adata-transformation is that it is very easy to fit the data with linear
regression. In the method described in ligure 2, no straight line is produced in the dose-stimulus
curve of the partial agonist. The dose-stimulus relation is a direct reflection of binding of the
partial agonist to the receptor, and can be evaluated by computer programs designed to fit
receptor-binding studies using non-linear regression [Munsn and Rodbard, 1980]. In our
opinion, the method described in figure 2 should be preferred over the method described by

133

ModelI
AR** high
KA
A+R E
~AR~ AR*
\~

AR

medium

none

.~
.~
~

v
a~

ModelII

Fig. 3. Two models that explain the difference in efficacy between different agonists. In both models the
agonist (A) first binds to the receptor (R), yielding the agonist-receptor complex (AR). In model I
agonists with different efficacy induce different conformatial changes of the receptor. These different
conformations of the receptor are not equally effective in triggering the biochemical reactions following
receptor activation that finally produce an effect (either AR, no activation; AR*, medium activation; or
AR**, maximal activation; other conformations are of course also possible). In model II there are only
two conformations of the receptor in the agonist receptor complex, the receptor is either activated (AR*)
or not activated (AR). The equilibrium between the two states of the receptor after agonist binding is
expressed in the efficacy of the agonist.

Mackay for partial agonists, because by linearisation incorrect parameters can be obtained
[Munson and Rolbard, 1980].
The "null methods" described in fig 2 or according to Mackay, are also suitable for two
receptor models, this in contrast to the "operational model" of Leff. It has well been described
that ()terbutaline has a higher affinity for (32-adrenoceptors than for (31-adrenoceptors.
Moreover, the positive isotropic effect in the heart is mediated by both X31- and
X32-adrenoceptors. Therefore, we anricipated that the dose-stimulus curve of()terbutaline in the
left atrium had to be fitted by a two receptor model. However, it was found that the data very
well fitted a one receptor model. Despite that, on a theorerical basis, it was expected that the data
obtained with ()terbutaline could not be analyzed with the "operational model",in pracrice this
model very well fitted the data.
The meaning ofa difference inefficacy.
The efficacies of()terbutaline and (-)isoproterenol for the (3-adrenoceptor in the left atrium
appear to be different. What does this mean? When the efficacy of an agonist in a certain organ
is twice that of another, is this because the compound with the higher efficacy is capable to
activate the receptor twice as much when it binds to the receptor compared to acrivation by the
compound with the lower efficacy, or is this because a twofold higher fraction of receptors
becomes activated through binding of the receptors by the compound with the higher efficacy,
relarive to the fracrion receptors acrivated by the other compound (fig. 3)? In the fust mechanism
there are different conformarions of the receptor with different activities. Binding of agonists
with different efficacies induces different confo~narial changes of the receptor. In the
134

(-)isoproterenol

control
thyro~ne

pD2

ocE

8.5 0.2
9.2 0.2

1
1

()terbutaline
as

110 30 1
510 140 1

PD2

ocE

ocs

5.2 0.2 0.81 0.07 3.1 0.8 0.028 0.006


1
5.8 t 0.2
13 3 0.026 f 0.007

Table 2. The pD2,aE,efficacy(e)and as of the positive inotropic effect of(-)isoproterenol and ()terbutaline in
the isolated left atrium of control rats or of thyroxine pretreated rats.

conformation of the receptor induced by the cc~mpoun~ with ~ higher efficacy, the receptor is
more prone to activate the cascade of biochemical reacrions that finally produce the effect(model
I, iig 3). In the second mechanism them arp just two conforna~ions of the receptor, tie receptor
is either inactive or fully active. When a drug binds to the receptor, the total period in which the
receptor is in the active conformation during a certain time span is expressed. in the efficacy of
that drug. The longer that period, the higher the efficacy(model II, fig. 3).
Recently, Kenakin presented the %rst model(model I, fig 3) as a general explanarion for a
difference in efficacy between two agonists [Kenskin, 1989]. Two agonists have a different
efficacy because they induce different conformarion changes of the receptor. He also argued that
it is possible that the optimal conformation of the receptor for an interaction with one transducer
protein, is not the optimal conformation of the receptor for a interaction another transducer
protein. Therefore, the relative potencies of two agonist that interact with the same type of
receptor but on different tissues might be different, because other transducer proteins are
involved in the production of the effect in the different rissues. This may give rise to an incorrect
receptor classification [Kenskin, 1989]. However, there are some indications that for the
(3-adrenoceptor system in the heart, the second mechanism in which the receptor is either active
or inactive is the better model(model II, fig. 3). In this model the agonist(A)and the receptor
(R)interact yielding the drug-receptor complex (AR). Binding is described by the dissociation
constant(KA). After binding the receptor in the complex can be activated by the agonist(AR*).
The efficacy of the agonist(e)describes the equilibrium between the inactivated receptor(AR)
and the activated receptor(AR*)
KA
e
A+R ~> AR SAR*
In this model no effect of different transducer proteins on the relarive agonist potency is
expected. Receptor acrivarion is most likely a result of an agonist induced conformarion change
in the three-dimensional structure of the receptor protein. For the ~3-adrenoceptor this might be
caused by reduction of a disulfide bridge in the receptor. It has recently been suggested that the
reducing equivalents for such a redox reaction are provided by the agonist that binds the
receptor. The efficacy of the agonist is determined by its redox potenrial[Wong et al., 197].In
fact, Arins already discussed that an interaction of an agonist with a disulfide group of the
receptor might be one of the general mechanisms of receptor acrivation [Arins, 1964].
Effect of thyroxine pretreatment on the receptor-mediated response of(-)isoproterenol and
(+)terbutaline.
It is possible to manipulate the efficacy of an agonist. Treatment with thyroxine enhances
~3-adrenoceptor mediated response. In the isolated left atrium of thyroxine pretreated rats the
135

maximal effect induced by ()terbutaline is identical to the maximal effect produced by


(-)isoproterenol (table 2). This makes ()terbutaline a full agonist in this preparation, while in
the left atrium of control rats ()terbutaline is a partial agonist (table 1). Also the response to
(-)isoproterenol is changed by thyroxine pretreatment, less (-)isoproterenol is needed to produce
half maximal effect. Thyroxine treatment increases the efficacies of both (-)isoproterenol and
()terbutaline (table 2).
Intrinsic efficacy.
Apparently the term efficacy as introduced by Stephenson does not only depend on the rissue
used, but it is also depends on the pretreatment of the animals. It has been argued that the
difference in efficacy of the same agonist for the same receptor in different tissues is due to a
difference in receptor density in the tissues. Therefore Furchgott coined the term intrinsic
efficacy (e), a parameter that relates the efficacy to receptor density ([Rt])[Furchgott, 1966].

[Rtl
By the intrinsic efficacy of a certain agonist the average of the maximal stimulus per receptor by
that agonist (i.e. when the receptor is constantly occupied by that agonist) is given. It is
generally been accepted that the intrinsic efficacy is a truly drug-dependent parameter that does
not depend on species, type of tissue or type of response [Kenskin, 1987].
One of the problems with intrinsic efficacy is that it is not possible to deternune receptor
density with a functional study, a radioligand binding study has to be performed hereto.
However,in a radioligand binding study also receptors that are not coupled the reaction cascade
after receptor activarion contribute to the receptor number. Aside from that, where has the
receptor number obtained by a radioligand binding study to be related to? It is evident that
receptor concentration cannot be calculated using the volume of the organ bath, as is performed
for the concentration of the agonist. In that case receptor density would vary with the volume of
the organ bath, and with the size of the tissue prepararion. Receptor density refers to the "tissue
concentration" of the receptor [Kenskin, 1987]. Usually receptor density is deternuned by
means of a radioligand binding study on a rissue homogenate or on a membrane preparation and
it is expressed as mol receptor per g protein. Receptor density depends on the preparation used
in the radioligand binding study, if a rissue homogenate is usd many non-receptor proteins are
present, and if a membrane preparation is used the degree of purification - i.e. the degree of
contamination of the membrane prepararion with e.g. cytosolic non-receptor proteins -affects
receptor density. Most likely, membranes of different tissues do not contain a fixed amount of a
certain protein. It is well known that different proteins do not react equally effective with the
color reagent in the standard methods used for protein determination. Moreover, receptor
density will also depend on the method used for protein determination, since there is a great
variation in the sensitivity of most proteins between the two most popular methods used. In our
opinion there is no suitable tissue parameter to relate receptor number to, and therefore the
intrinsic efficacy will be tissue dependent. Apart from that, several examples will be given that
illustrate that even though if there would be a convenient reference, the relarive efficacy would
still depend on species, type of tissue or type of response.
For example, when the reaction cascade between receptor activation and effect in the left
atrium is modified by a phosphodiesterase inhibitor, the intrinsic efficacy of ~i-adrenoceptor
136

agonists does not remain constant. By blocking phosphodiesterase, the concentration of second
messenger (cyclic AMP)produced by activation of(3-adrenoceptors is increased, and a smaller
fraction of the receptors has to be occupied by (-)isoproterenol in the left atrium to produce half
malcimal effect. A phosphodiesterase inhibitor makes the efficacy of (-)isoproterenol higher,
while receptor density remains constant and hence also the intrinsic efficacy changes.
Additionally, it has well been described that in assessing the activity of an agonist in isolated
rissue, the pD2 is dependent on the conditions used to construct a dose response curve. For
example, th pD2 of (-)isoproterenol in the isolated trachea decreases when an increasing
amount of carbachol is present in the incubation medium [Buckner and Saini, 1975], a
phenomenon described as physiological antagonism. Since (-)isoproterenol remains a full
agonist in all experiments, this means that the D2 and hence the efficacy depend on the
experimental conditions, and since in all experiments the same tissue is used (so [Rt] is
constant) the variation of the efficacy is reflected in a variation of the intrinsic efficacy. A
comparable variation of the pD2 can be obtained by mechanically decreasing the baseline
tension. This variation in the pD2 might be overcome by never using any baseline tension, but
often the organ itself produces an opposite force to the receptor mediated response. Moreover,
for measuring the activity of (3-agonists in lung tissue, in practice always a certain
precontraction, e.g. with a fixed concentration of muscarinergic agonist, is used. This in order
to produce a greater absolute response that makes it possible to asses the effect of the (3-agonist
with greater accuracy. Additionally, it has been demonstrated that the method used to produce a
physiological antagonism also affects the pD2 of ~i-agonists in the isolated trachea [Van
Amsterdam et al., 1989].
One of the standard methods in receptor studies makes use of irreversible antagonists. It is
assumed that an irreversible antagonist only inactivates a fraction of the receptors and has no
effect on the remaining receptors or on the biochemical processes that lay between receptor
activation and effect. The dose-response curves constructed with an agonist before and after
treatment of an organ with an irreversible agonist are compared. Provided that there is still
receptor reserve for a full agonist in the tissue after receptor inactivarion by the irreversible
antagonist, the absolute value ofthe half minimal receptor-mediated response is not affected by
receptor inacrivation. However, the concentration of the full agonist needed to produce this
effect(D2)increases after treatment with the irreversible antagonist, while the affinity of the full
agonist for the receptor remains the same [Venter, 1979]. In a tissue with receptor reserve, the
reduction of efficacy is proportional to the fraction receptors inactivated and hence the intrinsic
efficacy remains constant. Under these conditions, the intrinsic efficacy does not depend on
treatment with the irreversible antagonist.
In rissue without receptor reserve, receptor inactivation leads to a reduction of the absolute
value of the maximal receptor-mediated response. In this tissue, receptor inactivation is not
reflected in a proportional reduction of the efficacy. Theoretically, it is even possible that the
efficacy increases through receptor inactivarion. In the case that there is a linear relation between
the stimulus and effect, receptor inactivation will not affect the pD2 of a full agonist and hence
the efficacy remains constant while the reduction of the intrinsic efficacy is proportional to the
fracrion receptors inacrivated. Apparently, under this conditions the efficacy rather then the
intrinsic efficacy is independent of the tissue used; irrespective of the fraction of receptor
inactivated by the irreversible antagonist used, the efficacy of the full agonist remains two.

137

Effect

[iso]
(M)

sec(%)

Hertz(%)
1'1

1.10 -11

0.244(0)

4.09(0)

3.10 -11

0.237(6)

4.23(3)

1.10 -10

0.214(24)

4.67(13)

~~

3.10 -10

0.189(44)

5.29(28)

~ SO

1.10 "9

0.161 (67)

6.22(49)

3.10 -9

0.139(84)

7.20(72)

1.10 -g

0.128(93)

7.82(87)

3.10 -8

0.122(97)

8.18(95)

1.10 -~

0.120(99)

8.31 (98)

3.10 -~

0.119(100)

8.40(100)

seconds

...._......~.............

Hertz

r~

-11

-9

-7

log[(-)isoproterenol]

Fig. 4. Effect of dunension on the dose-response curve of(-)isoproterenol in right atrium of control rats.
The positive inotropic effect is expressed either as an increase in the frequency of the contractions (Hertz)
or as the reduction of the time between two succeeding contraction (seconds). The pD2, efficacy and
intrinsic activity (as) aze given in table 3.

Beside treatment with an irreversible agonist, several physiological processes are known to
affect the efficacy of an agonist. For example, desensitisation -the uncoupling of a receptor
with its second messenger system -reduces the efficacy in tissue with receptor reserve while
[R~]remains constant and hence the intrinsic efficacy will also be reduced.
In principle, efficacy is dimensionless [Furchgott, 1966]. However, as shown in figure 4
and table 3, the efficacy and the intrinsic efficacy depend on the dimension in which the effect
is expressed. The positive inotropic effect of (-)isoproterenol in the right atrium can be
expressed either as an increase in the frequency of contraction or as a reducrion of the rime
between two succeeding contractions. This change does not alter the shape of the dose response
curve, but it changes the pD2 of(-)isoproterenol. It is evident that the change in dimension has
no effect on the affinity of the agonist for the (3-adrenoceptor, nor on the density of (3adrenoceptors. This implies that both the efficacy and the intrinsic efficacy depend on the
dimension used to express the effect in. This variation can be overcome by always expressing
the effect in the same dimension, but what uniform dimension should be used when we want to
compare a chronotropic effect to a inotropic effect? In the example given in figure 4, the
funcrion that describes the transfer of the stimulus in an effect (f(S)) is independent of the
dimensions used to express the effect. This does not have to be true for other dimensions. So
beside the efficacy and the intrinsic efficacy, f(S) also depends on the dimension used to
express the effect in. A similar difference in efficacy, intrinsic efficacy and f(S) can be expected
when different types of response mediated via the same receptor system are compared. The
flexibility of the stimulus-effect curve limits the application of the "operational model" of Leff,
were the shape of the srimulus-effect curve is restricted to a regtangular hyperbola.
The assumption that the intrinsic efficacy is exclusively drug-dependent and does not depend
on species, type of tissue or type of response [Kenskin, 1987], implies that the srimulus
138

effect in Hertz

(-)isoproterenol

effect in seconds

pD2

orb

as

PB2.

aE

as

9.0

300

9.4

750

Table 3. The pD2, aE, efficacy (e) and aS of the posive chronotropic effect of (-)isoproterenol in the isolated
right atrium of control rats. The effect is expressed either as an increase in the frequency of the contracrions (in
Hertz) or as the reduction of the time between two succeeding contractions (in seconds).
produced by a certain agonist s~ia a certain receptor is only dependent on the conc~ntsation of
receptors occupied by this agonist([AR]) and the intrinsic efficacy of the agonist (~)for the
receptor since
S=ey=ERty=E[AR]
This also means that in order to obtain half maximal effect(S = 1) via a certain receptor, the
agonist has to occupy the same concentration of receptors (namely[AR]= 1/~) irrespective of
the tissue used. We have given several examples that demonstrate that the intrinsic efficacy is
not constant, even when the same tissue is used. The assumption that the stimulus only depends
on the intrinsic efficacy of the agonist and the concentration of the receptors occupied is an
oversimplificarion, comparable to the oversimplificarion in the original theory of Arins that the
relarive effect produced by a certain agonist would only dependent on the intrinsic activity(aE)
of the drug and the fraction receptors occupied(E = aE y)[Arins, 1954].
Apparently, the intrinsic efficacy is not a truly drug-receptor dependent parameter. This is
not caused by a change in the interaction of the agonist with the receptor in the examples given.
The magnitude of receptor activation by a certain agonist is assumed to be identical in every
situarion provided that the same agonist and receptor is used. The variation of the intrinsic
efficacy is due to the effect the performed manipulations have on the calibration point used to
determine the efficacy, the concentration of agonist that produces half maximal receptor
mediated response(S = 1 when E = 0.5).
Intrinsic activity.
Beside efficacy and intrinsic efficacy a modified deiinirion of the term intrinsic activity (as)
has been introduced to describe drug-receptor interacrion. Originally, the intrinsic activity (aE)
of an agonist was defined as the maximal effect produced by this agonist relative to th highest
effect produced by an agonist via the same receptor [Arins, 1954]. Normally this term is used
to describe the effect of an agonist in a functional experiment; full agonists have an intrinsic
activity(aE) of 1, while partial agonists have an intrinsic activity (aE) between 0 and 1.
According to the modified definirion of Arins: "T'he intrinsic acrivity is inversely proportional
to the number of receptors that has to be occupied to induce a certain srimulus" in a certain tissue
preparation [Arins, 1964]. Intrinsic efficacy(as)in its modified definition is comparable to the
term efficacy as defined by Stephenson. However, the calibration point used for intrinsic
activity (as)is different form that of efficacy. In deternuning the intrinsic activity (as) of an
agonist, the receptor activation by that agonist is related to receptor activation by the agonist
with the highest efficacy. The relation between the efficacy and the intrinsic acriviry (as) of a
certain drug is:
as = e%max
139

in which ems is the highest efficacy of the agonists tested on the tissue. It is very confusing
that the use of intrinsic activity(aE)that originates from its original definition is still frequently
applied. In order to discriminate between the two definitions of intrinsic efficacy, intrinsic
activity in its original definition is depicted as aE, while the modified definition of intrinsic
efficacy will be presented as as. These notations have previously been used for this purpose
[Kenskin, 1985].
As demonstrated by the data given in table 2, thyroxine pretreatment alters the efficacy of
both ()terbutaline and (-)isoproterenol. However, the rario of both efficacies, expressed in
intrinsic activity (as), is independent of the pretreatment. Also the other manipulations that
change the efficacy or intrinsic efficacy have no effect on intrinsic activity (a~}, Apparently,
intrinsic activity (as) is a parameter exclusively determined by an agonist-receptor interaction,
and therefore it is theoretically usefial for the classification of drugs and drug receptors.
Intrinsic acrivity (as)is a drug constant because it is a relative parameter, the efficacy of a
drug is divided by the efficacy of another drug. By this division, the influence of the tissue on
the efficacy is eliminated. However, this division also introduces a greater relative error in the
intrinsic activity (as), because the efficacy of the agonist with the highest efficacy (ems)cannot
be determined without experimental inaccuracy. It should be noted that the influence of this
inaccuracy on the intrinsic activity (as) is not as great as expected in fust instance. This is
because in the determination of the intrinsic activity (as), the efficacies of both agonists have to
be obtained in the same preparation. The variance in the efficacy of an agonist is partly due to a
biological variation in the rissue dependent srimulus-effect relation. By relaring the efficacy of an
agonist to the efficacy of another agonist determined in the same preparation, the biological
variation of the tissue-dependent processes is also eliminated..
Relative efficacy.
Beside all the parameters that already exist for descriprion of the interacrion between agonist
and receptor, Kenskin introduced the term relative efficacy (erel)[Kenakin, 1985]. The relative
efficacy of a drug is the efficacy (e)of that drug related to the efficacy of the agonist with the
highest efficacy (ems):
ere1=%mom
Kenskin also reviewed intrinsic activity(as), but the relation given between intrinsic activity
(as)and (intrinsic) efficacy (equation 17,[Kenskin, 1985]) is incorrect. This might partly be
due to an incorrect calibration point for the srimulus; Kenskin argued that the stimulus is unity at
50% receptor occupancy (S = 1 when y = 0.5) [Kenskin, 1985], however, according to the
convention adopted by Stephenson the stimulus is unity at half maximal receptor mediated effect
(S = 1 when E = 0.5) [Stephenson, 1956]. In fact, relative efficacy is identical to intrinsic
acriviry(as)in its modified definition. Probably because Kenskin did not realize the similarity
between intrinsic activity(as) and relative efficacy, he introduced relative efficacy as a new
parameter.
Tissue-independent parametersfor the descriprion ofagonist-receptor interaction.
In describing drug-receptor interacrion, several parameters have been introduced. In the
action of an agonist, first the compound has to bind to the receptor and secondly the agonist has
to activate the receptor. For the binding of the drug to the receptor dissociation constant(KA) is
a truly drug dependent parameter. Nevertheless there are several practical problems in
140

deternuning this parameter, a subject that has been discussed previously [Kenakin, 1987].
Moreover, binding of an agonist to the receptor induces a conformation change of the receptor.
In the model we proposed (iig. 3, model II) the agonist cannot dissociate from the agonist
receptor complex when the receptor has been acrivated. This implies that also the efficacy of an
agonist contributes to the Kp obtained in a radioligand binding study. For receptor activarion by
an agonist, efficacy and intrinsic efficacy are no truly drug-dependent parameters. The modified
definition of intrinsic activity(as) is an exclusively drug-dependent parameter that describes
receptor activation by an agonist. The use of a parameter like intrinsic efficacy is only of limited
value, especially because of the problems faced in determining receptor density. The
disadvantage of efficacy is that this term depends on the organ used and also on the dimension
used to express the effect in. The intrinsic acrivity (aS) does nit have these limitarions and
therefore it might be preferred to express receptor activation of an agonist in its intrinsic activity
(as).
Also terms like partial and full agonist are of limited value. Pretreatment with thyroxine
makes a full agonist of the partial agonist ()terbutaline (table 4). In hyperthyriodie usually a
~3-adrenoceptor antagonist is used to prevent excessive stimulation of the sensitized
~3-adrenoceptor system. It has to be questioned whether a ~i-blocker with a relatively high
intrinsic sympaticomimetic activity like pindolol can be used for this purpose. In the extreme
case, the partial agonist pindolol is, due to the enhanced endogenous production of thyroxine,
rendered into a full agonist. Also the discrimination between agonists and antagonist is not as
clear-cut as often is suggested. The ~3-adrenoceptor antagonist propranolol, that is thought to
have no intrinsic sympaticomimetic activity, is a partial agonist in a tissue with an
extraordinarily efficient amplifying system [Hermsmeyer et al., 1982].
Organ selective drug action.
For the introduction of organ selectivity by a drug, it is often attempted to search for
receptors or receptor subtypes that are only available on the target organ, thus mimicking the
organ selectivity of hormones. In this approach, primarily the difference in affinity of the drug
between the type of receptor on the target organ and the other types of receptors on the other
organs is used to reach selectivity. However, also the difference in intrinsic activity (as) of the
agonist for the different type of receptors may attribute to tissue selectivity. Organ selectivity of
an agonist achieved in this way is, beside a difference in affinity of the agonist for the different
receptors, caused by a difference in magnitude of activation by the agonist of the different
receptors, and by the difference in efficiency in which the input srimulus generated by receptor
acrivation is transformed into an effect in the different tissues. So both drug-receptor interaction
(KA and as) as well as tissue dependent processes (E = f(S)) determine organ selectivity
produced via the srimularion of different receptor types.
Theorerically, it is also possible to generate an organ selecrive response by addressing a
receptor located on several organs, by malting use of a difference in the cascade of reactions that
process the input stimulus into an effect. As shown in this study, in the isolated left atrium a
lower concentration of (-)isoproterenol is needed, compared the concentration of
(-)isoproterenol needed in the right ventricle, to produce a certain percentual effect, although the
same receptor and agonist are involved. This is caused by the fact that the reaction in the cascade
of reactions following receptor activation that functions as bottle neck, is saturated in the left

141

atrium when a smaller fraction of the receptors is occupied, compared to the fraction of
receptors needed hereto in the right ventricle. Possibly the organ with the higher pD2 for
(-)isoproterenol has a higher receptor density, but it is also possible that that the higher pD2 is
caused by a higher "tissue concentration" of an enzyme transducing the stimulus by one of the
biochemical reactions preceding the bottle neck reaction. Often it is incorrectly assumed that a
higher pD2 can only be caused by higher receptor density [Furchgott, 1966].
When a difference in pD2 of an agonist for a certain receptor type in different organs is used
to introduce organ selectivity, it is stated that the organ with the higher pD2 can be addressed
selectively [Kenskin, 1987]. However, it should be noted that the effect of the agonist is
expressed relative to the maximal receptor mediated effect in that organ. It is usually tacitly
assumed that when the relative effects in different organs are identical, this implies that the
physiological relevance of these effects are equal. However, this does not have to be true. This
is best illustrated by the result the dimension used to express the effect in, has on the relative
response. When a certain effect is expressed in different dimensions, the effect relarive to the
maximal response may depend on the dimension used (fig. 4). Moreover, a certain percentual
increase in contraction force of a ventricle is probably of greater physiological importance than a
similar percentual increase in contraction force of an atrium,just because the ventricle is a more
powerful muscle. To compare the effects of an agonist on different organs, the effects have to
be related to their physiological importance. By introducing organ selectivity via the same
receptor, all reacrions involved in the transformation of the input stimulus in the effect are
important. The pD2 is determined by the biochemical reactions preceding the bottle neck
reaction, the absolute value of the maximal receptor mediated effect is determined by the
biochemical reactions after the limiting one. The magnitude of activation of a certain receptor
after binding of a certain agonist is assumed to be identical in all rissues.
This paper has reviewed some of the parameters used to express drug-receptor interaction. It
is demonstrated that parameters like pD2 and aE are of limited value. It is preferred to express
drug-receptor interacrion by the affinity of the drug for the receptor(Kp) and by the efficacy or
as,although it is not easy to asses these parameters. Several methods suitable for determination
of the efficacy and as have been given. Hopefully, the use of these parameters will lead to a
more uniform descriprion of drug-receptor interaction, which makes a comparison between
different studies possible. We adapted the receptor theory described here, in order to quantify
the effect of oxidative stress on ~3-adrenoceptor function.
References.
Abrahamson, T.(1986)The dil-and i2-adrenoceptor stimulatory effects of alprenolol, oxprenolol and pindolol: a
study in the isolated right atrium and uterus of the rat, Br. J. Pharmacol. 87,657-664.
Arins, E.J. (1954) Affinity and intrinsic activity in the theory of competitive inhibition, Arch. Int.
Pharmacodyn. Ther.99, 32-49.
Arins, E.J.(1964) Molecular pharmacology, volume 1, Academic press, New York -London.
Arins, E.J. , A.J. Beld, J.F. Rodrigues de Miranda and A.M. Simonis (1979) The pharmacon-receptor-effector
concept, in: R.D. O'Brien (Ed.) T'he receptors, a comprehensive treatise, Plenum press, New York, pp. 33-91.
Buckner, C.K., and R.K. Saini(1975) On the use offuncponal antagonism to estimate dissociation constants for
beta adrenergic receptor agonists in isolated guinea-pig trachea, J. Pharrracol. Exp. Ther. 194,565-574.
Furchgott, R.F.(1966) The use of ~i-haloalkylamines in the differentiation of receptors and in the determination
of dissociation constants of receptor-agonist complexes, in: N.J. Harper and A.B. Simmonds (Eds.) Advances
in drug research, volume 3, Academic press,London -New York, pp. 21-55.
142

Haenen, G.R.M.M,P. v Dansfik, N.P.E. Vermeulen, H. Timmerman and A. Bast(1988) The effect of hydrogen
peroxide on ~3-adrenoceptor function in the heart,Free Rad. Res. Comms.4,243-249.
Haenen, G.R.M.M, H.J.M. Plug, N.P.E. Vermeulen, H. Timmerman and A. Bast (1989) Contribution of 4hydroxy-2,3-trans-nonenal to the reduction of(3-adrenoceptor function by oxidative stress. Life Sci, 45, 7176.
Hermsmeyer, K., R. Mason,S.H. Griffen and P. Becker(1982)Rat cazdiac muscle cell automaticity responses to
a- and R-adrenergic agonists and antagonists, Circulation Res. 51,532-537.
Kenskin, T.P.(1985) The quantification of the relative efficacy of agonists, J. Pharrnacol. Meth. 13, 281-208.
Kenskin,T.P.(1987)Pharmacological analysis of drug-receptor interaction, Raven press, New York.
Kenskin, T.P.(1989) Challenges for receptor theory as a tool for drug and drug receptor classification, Trends
Pharmacol. Sci. 10, 18-22.
Leff,P.(1988) Analysis of agonist action using the operational model, Trends PharmacoL Sci. 9, 39.5-398.
Mackay, D. (1966) A new method for the analysis of drug-receptor interaction, in: N.J. Harper and A.B.
Simmonds(Eds.) Advances in drug research, volume 3, Academic press,London -New York, pp. 1-19.
Mackay,D.(1988),Concentration-response curves and receptor classification: null method or operational model?
~(
Trends Pharmacol. Sci. 9,202-205.
Anal.
s~lstems,
ligand-binding
for
approach
Munson,P.7., D. Rolbard (1980) Ligand: A versatile computerized
Biochem. 107, 220-239.
Stephenson, R.P.(1956) A modification of receptor theory, Brit. J. Pharmacol. 11, 379-393.
Van Amsterdam, R.G.M., H. Meurs, F. Brouwer, J.B. Postema, A. Timmermans and J. Zaagsma(1989)Role of
phosphoinosi[ide metabolism in functional antagonism of airway smooth muscle contraction by (iadrenoceptor agonists, Eur. J. Pharmacol. 172, 175-183.
Venter, J.C.(1979) High efficiency coupling between beta-adrenergic receptors and cardiac contractility: Duect
evidence for "spare" beta-adrenergic receptors, Mol.Pharmacol. 16,429-440.
Wong, A S.M. Hwang, H.-Y Cheng and S.T. Crooke (1987) Structure-activity relationships of R-adrenergic
receptor-coupled adenylate cyclase: Implications of a redox mechanism for the action of agonists at (3adrenergic receptors, Mol.Pharmacol. 31, 368-376.

143

144

- Chapter 14 The effect of hydrogen peroxide on ~i-adrenoceptor


function in the heart.
Abstract. A membrane preparation of calf heart left ventricle has been used to study the effect
of radical stress on the (3-adrenoceptor complex. To this end the membranes were incubated for
3d minutes with several concentrations of hydrogen peroxide. This resulted in a dose dependent
peroxidation of the membrane lipids. Preincubation with hydrogen peroxide in the concenirarion
range 10-~ - 10-3 M caused an increase in specific (-)-[125I]-Iodocyanopindolol binding,
possibly due to a decrease in membrane fluidity as a result of lipid peroxidation, thus making
the receptor protein more accessible. Higher concentrations H2O2 reduced the specific
~_~_~i25I]-Iodocyanopindolol binding, which is most likely the effect of deterioration of the
receptor protein by the more pronounced radical stress induced by these higher concentrations.
Also adenylate cyclase acrivity was affected by radical stress. Basal cyclic-AMP production and
cyclic-AMP producrion induced by NaF(10-2 M)or guanylylimidodiphosphate (10-4 M),was
suppressed after pretreatment with concentrarions of H2O2 above 10-4 M. This indicates a
higher sensitivity of the adenylate cyclase toward radical stress when compared to the receptor
protein. Our results show that radical stress can perturb (3-adrenoceptor function considerably in
the heart.
Introduction.
The cardiotoxicity of several catecholamines is well known. The most obvious explanarion
for this toxicity is (3-adrenoceptor hyperstimulation. The strong positive chronotropic and
inotropic effect of catecholamines may lead to hypoxia, depletion of high-energy phosphate
stores and abnormal calcium accumulation in the myocardium. Apart from that, it has been
stated that radical mechanisms play a role in catecholamine induced myocardial necrosis [1].
During the (aut)oxidarion of catecholamines superoxide anion radicals, hydrogen peroxide,
hydroxyl radicals and semiquinone radicals are generated. This hypothesis is supported by the
protecrion which is afforded by vitamin E and the membrane stabilizing agent zinc, against
isoproterenol induced cardiomyopathy [1] and the inability of propranolol to protect against the
toxic effect of isoproterenol on cultured cardiac muscle cells [2], whereas ascorbate protects [3].
Both mechanisms, receptor hyperstimulation and autoxidation, are interrelated. Hypoxia
induced by excessive (3-adrenoceptor stimulation induces an increase in the radical production.
In heart rissue, low oxygen tension gives conversion of xanthine dehydrogenase into xanthine
oxidase by a Cat+ dependent proteolysis [4]. Xanthine oxidase utilizes molecular oxygen as an
electron acceptor to catalyze the producrion of superoxide anion radicals and hydrogen peroxide.
Other sources of radicals induced by hypoxia originate from the accumulation of neutrophils[5]
and an excessive formarion of reduced electron carriers in mitochondria in the heart [6]. On the
other hand radical processes influence several receptor systems [7-10].
The aim of the present study was to evaluate the effect of radical stress on the (3-adrenoceptor
system in the heart. To this end we preincubated membranes of the calf heart left ventricle with

145

several concentrations hydrogen peroxide which resulted in a dose dependent radical stress.
After washing the membranes,(3-adrenoceptor density was measured by means of a receptor
binding assay using (-)-[125I]iodocyanopindolol (ICYP). Adenylate cyclase activity was
measured by a RIA procedure.
Materials and methods.
Heart membranes were prepared from calf heart left ventricles according to the method of IJzerrnan [l l].1fie
membranes were taken up in buffer A (50 mM Tris-HC1,140 mM NaCI and 5 mM MgC12, pH ?.4 at 37 C)and
stored in liquid nitrogen until use. Lipid peroxidation was measured with the thiobazbituric acid method,
as
previously described [12]. Lipid peroxidation was expressed as nmoles malondialdehyde equivalents, using
an
extinction coefficient of 1.56.105 M-lcm-1.Protein determinations were madc ~y the method of Bradford
[13]
using bovine serum albumin as standard.
Heart membranes (final concentration 1.5 mg protein / ml) were incub~t~d zt 37 ~C, with shaking air b.,ing
freely admitted in buffer A(pH 7.4 at 37 C)which contains sodium azide (final concentration 1 mIvn in order
to
inhibit endogenous catalase, and several concentrations hydrogen peroxide(0 - 0.1 Ivn for 30 min. The incubation
was terminated by the addition of one volume of ice-cold buffer A(pH 7.4 at 37 C)and centrifugation for 3 min
in an eppendorf centrifuge(15000 x g)at4C. The precipitated membranes were resuspended in 1 ml fresh buffer
and and centrifugation again for 3 min (15000 x g) at 4 C. This washing procedure was repeated twice. The
receptor binding and adenylate cyclase activity assay were performed with membranes suspended in buffer A(pH
7.4 at 37 C).
All receptor binding assays were performed in triplicate in a final volume of 350 ml using the 125I labeled
(3-adrenoceptor antagonist (-)-Iodocyanopindolol (ICYP). Due to the lipophilic character of ICYP
all
~i-adrenoceptors, e.g. also those in inside-out vesicles, are determined. For the saturation experiments membranes
were incubated for 60 min with several ICYP concentrations(0-700 pIvn. In the other experiments (3-adrenoceptor
density Bm~(fmol / mg protein) was determined with a single ICYP concentration (150 - 250 pIvn. The binding
reaction was terminated by the addition of 3 ml of ice-cold buffer A(pH 7.4 at0 C),followed by rapid filtration
through Whatman G/FC filters. Each filter was washed with an additional2 x 3 ml buffer. The radioactivity of
[he filter was assessed in an Auto-Gamma scintillation spectrometer (Packard, USA). In all experiments nonspecific binding was determined with 10-6 M (-)-timolol and Titer binding was determined by omission of the
membranes. Binding data from the saturation curves were evaluated using the computer program LIGAND on a
Zenith Z-110 microcomputer.
Cyclic-AMP production was performed in buffer A (pH 7.4 at 37C) using several stimulators, indicated in
figure 4. Phosphodiesterase was inhibited by the addition of methylisobutybcanthine (final concentration 5 10-5
1Vn in the incubation medium. All reactions were carried out at 37 C for 30 min, and terminated by heating at
95 ~C for 3 min. The cyclic-AMP content was determined by a RIA
The chemicals used were (-)-pmolol maleate and guanylylimidodiphosphate(GppNHp)(Sigma); hydrogen
peroacide(perhydrol) and sodium azide (Merck).(-)-[1251]ICYP was obtained from New England Nuclear and the
cyclic-AMP assay kit(code TRK.432) was obtained from Amersham. All other chemicals were of reagent grade.

Results.
Incubarion of heart membranes with hydrogen peroxide resulted in a dose dependent increase
in lipid peroxidation (table 1). Preincubation of heart membranes with 10-~ - 10-3 M hydrogen
peroxide increased the amount of specific ICYP binding (fig. 1). Incubation with higher
concentrarion of hydrogen pero~cide
1 . 10-5 M
1 . 10~ M
1.10-3 M
1.10-2M
5 10-2 M

nmol malondialdehyde/ mg protein


< 0.02
0.29 0.05
0.640.06
1.100.10
2.55 0.15

Table 1. Lipid peroxidation in membranes of the left ventricle of the calf heart induced by incubation with
hydrogen peroacide during 30 min. The results are expressed as mean S.E. of four separate experiments.
146

.y

~ c 1 42.2
~ ~~
L
O
~
d
~

~
~
L
~

V
o
c
a~
L

~
~
~
~

$~ $

**

~a+
~~~
+
\~I

71.1

-9

-7
-5
-3
Log [H2O2] (M)

-t

Fig. 1. Concentration dependent effect of hydrogen peroxide pretreatment on R-adrenoceptor density.


(3-adrenoceptor density was determined by ICYP binding. The concentration ICYP used was 150-250 pM.
Each data point is the mean value of a separate experiment performed in triplicate.

concenirarions hydrogen peroxide decreased ~3-adrenoceptor density. The same phenomenon


was observed in the saturation curves shown in fig. 2 and 3. Preincubation with 10-5 M
hydrogen peroxide for 30 min resulted in an increase in (3-adrenoceptor density of
approximately 15 %, whereas preincubation with 10-1 M H2O2 decreased the amount of
(3-adrenoceptors approximately 40 %. The pretreatment had no effect on the non-specific
binding of ICYP,determined. with 10-6 M (-)-timolol. Further analysis of the data revealed that

zoo
.~
m

c
0

~
r

100

Of
0

200

400

600

800

[ICYP] pM

Fig. 2. Total (x), non-specific (0) and specific (+) binding of an increasing concentration of ICYP on
control heart membranes. Non-specific ICYP binding was determined with the addition of 10-6 M
(-)-timolol. Specific ICYP binding was calculated from the difference between total and specific binding.
Kg = 54.4 2.3 pM and B,,,~ = 141.4 2,~ fmol/mg protein.
147

200
.~

a
m
E
v
i

~~0

E
~xxx
may+
~~

i ~

/
x
~ ~
+

-~

~~

200

400

600

800

(ICYP] pM
Fig. 3. Total(x),non-specific(0)and specific(+)binding of an increasing concentration of ICYP
on heart
membranes pretreated with 10-5 M H2O2. Kd = 61.5 5.4 pM and
Bmax = 162.9 5.5 fmol/mg protein,
or on heazt membranes pretreated with 10-1 M H2O2. K.~ = 62.5 5.8 pM and
Bmax = g6.9 3.7
fmol/mg protein.

pretreatment had no effect on the affinity ofICYP for the receptor since no significant change
of
the Kd was observed. Analysis of the data using more than one binding site were
not
significantly better.
The (3-adrenoceptor complex consists of at least 3 proteins; receptor, NS-protein and
adenylate cyclase. The ICYP binding data only provide information about the receptor protein.
In order to get more insight about the effect of oxidative stress on the complete (3-adrenoceptor

Basal
~

GppNHp

NaF

+
100

~
~
0

~/~yt
t

+++~~
+
$

+ +~
+

+~+
+

50
Fz
W

V
DG
W

f
t

{,
i~

0
-8

-6

-4 -2

-8

-6

-4

-2

-8

-6

-4

-2

LOG [H20]

Fig. 4. Concentration dependent effect of hydrogen perolcide pretreatrnent on basal cyclic-AMP production
and on cyclic-AMP production stimulated by NaF(10-2 Ivn or GppNHp(10~ lvn. The control values
are
( S.E.; n=3) basal 21.0 1.2; NaF 603 2.0 and GppNHp 41.1 1.2 pmol
cyclic-AMP/mg
protein/min. Each data point represents a separate experiment.
148

system, we also deternuned the effect of hydrogen peroxide preincubation on cyclic-AMP


production. As shown in fig. 4 incubation of untreated membranes with GppNHp or NaF
resulted, as expected, in a 2 - 3 fold increase in adenylate cyclase activity. Pretreatment with
hydrogen peroxide resulted in a dose dependent reduction of adenylate cyclase activity
independent of the stimulator (fig. 4).
Discussion.
Both ~i-adrenoceptor hyperstimulation and excessive radical formarion are thought to be
involved in the cardiotoxicity induced by several catecholamines. The aim of the present sttidy is
to evaluate the effect of radical stress on the ~3-adrenoceptor system in the heart. Incubation of
heart membranes with hydrogen peroxide vas shown to result a dose dependent lipid
peroxidation. A direct reaction of hydrogen peroxide with unsaturated fatty acids occurs only
very slow. Traces of transition metals (present in the buffer or bound to the membrane)rapidly
react with hydrogen peroxide to form more reactive oxygen species which can induce lipid
pero~dation.
The oxidarive stress induced by hydrogen peroxide markedly affects the ~3-adrenergic
receptor binding protein (fig. 1). Neurotransmitter receptors are thought to be buried in or to
float on the lipid bilayer membrane. The position of a diffusible protein in the membrane is
determined by the interactions between the amino acid residues with the external aqueous and
lipid domains. The two main forces controlling this aze an electrostaric and a hydrophobic
interaction. The lipid fluidity determines the magnitude of the hydrophobic interaction. An
increase in rigidity of the phospholipid bilayer by lipid peroxidation[14] can result in a vertical
displacement of the equilibrium position of the receptor protein toward the aqueous domain
[15]. Thus the receptor becomes more accessible and this can explain the increase in ICYP
binding we observed at low hydrogen peroxide concentrarions.
Increasing the hydrogen peroxide concentration resulted in a more pronounced radical stress.
Primary oxygen radicals, their secondary lipid radical intermediates and aldehydes produced
during lipid peroxidarion can deteriorate membrane bound proteins [16] including the receptor
protein. Receptor proteins of other systems are also affected by free radicals. Ascorbic acid
induced lipid peroxidation, inhibits dopamine antagonist binding in neostriatal membrane
preparations [7] and in rat cortical membranes [8]. Oxygen radicals generated by the xanthinexanthine oxidase system and hydrogen peroxide reduce muscarinic antagonist binding at
sarcolemmal receptors from the canine heart [9]. Furthermore hydrogen peroxide reduces
synaptic GABA receptor binding of[3H]-Muscimol, whereas xanthine-xanthine oxidase
treatment results in a profound stimulation of the binding of[~H]-Muscimol [10].
The adenylate cyclase data (fig. 4) show a 2 - 3 fold stimulation by NaF and GppNHp.
Preincubarion with relative low concentrations of hydrogen peroxide resulted in a reduced
adenylate cyclase activity, independent of the srimulator. These data suggest that oxidarive stress
impairs the adenylate cyclase unit of the ~3-adrenoceptor complex. The reducrion at a relative low
concentration hydrogen peroxide is possibly due to a higher sensitivity of the adenylate cyclase
toward radical stress in comparison to the ~i-receptor protein.
High concentrations of several catecholamines produce myocardial cell necrosis.
f3-adrenoceptor hyperstimularion and excessive radical formation are thought to be involved in
this necrosis. Our results demonstrate that radical stress can perturb (3-adrenoceptor functioning
149

in the heart. After the onset of the catecholamine induced cardiotoJcicity, the radicals
formed can
reduce the toxic effects mediated by (3-adrenoceptor hyperstimulation. On the other hand
there is
evidence that, once the process of lipid peroxidarion has started, the protection against
radical
stress is reduced [12]. So in treating cardiotoxicity, reduction of the radical stress might
be more
profitable then (3-adrenoceptor blockade.
References.
1 P.K. Singal, N. Kapur, K.S. Dhillon, R.E. Beamish and N.S. Dhalla, Can. J. Physiol.
Pharmacol., 1982,
60, 1390.
2 E. Severin, S. Sartore and S. Schiaffino, Experientia, 1977, 33, 1489.
3 K. Ramos and D. Acosta, Toxicology, 1983, 26, 81.
4 J.M. McCord and R.S. Roy, Can. 7. Physiol. Pharmacol., 192,60, 1346.
5 S.W. Werns, M.J. Shea and B.R. Lucchesi, J. Free Rad. Biol. Med., 1985, 1, 103.
6 C. Guarnieri, C. Muscari, C. Ventura and I. Mavelli, Free Rad. Res. Comms, 1985,
1, 123.
7 R.E. Heikkila, F.S. Cabbat and L. Manzino, J. Neurochem., 1982, 38, 1000.
8 S.F. Muakkassah-Kelly, J.W. Andresen, J.C. Shih and P. Hochstein, J. Neurochem.,
1983, 41, 1429.
9 R.C. Arora and M.L. Hess, Biochem. Biophys. Res. Comms., 1985, 130, 133.
10 Y. Yoneda, K. Kuriyama and M. Takahashi, Brain Res., 1985, 333, 111
11 A.P. Uzerman, The beta-adrenoceptor complex,Ph.D. Thesis, Amsterdam, 1985, p
53.
12 G.R.M.M. Hamen and A. Bast, FEBS Lett., 1983, 159, 24.
13 M.M. Bradford, Anal. Biochem., 1976, 72, 246.
14 G.E. Dobretsov, T.A. Borschevskaya, V.A. Petrov and Y.A. Vladimirov,FEBS Lett.,
1977,84, 125.
15 I. Yuli,S. Icerpi, P. Luly and M. Shinitzky, Exprientia, 1982, 38, 114.
16 S.P. Wolff, A. Garner and R.T. Dean, Trends Biochem. Sci, 1986, 11, 27.

150

- Chapter 15 Reduction of ~3-adrenoceptor function by oxidative stress


in the heart.
Introduction
Free radicals are produced in vivo even under normal conditions. Fortunately, cells are
equipped with a defense system that provides protecrion against potential noxious radicals. In a
healthy state, the production of radicals and the protection against radicals is balanced.
However, in several pathophysiological conditions (e.g. ischemia-reperfusion) free radical
production is enhanced, and radicals may overwhelm the defense system thus disturbing a well
preservd and delicate balance. Such a disbalance between radical formation and protecrion has
been described as oxidative stress (Sies 1985).
One of the mechanisms by which heart function is controlled proceeds via the
(3-adrenoceptor complex that is located on heart membranes. The inactivation of membrane
proteins, either directly via radical attack on the protein or indirecfly via perturbarion of the lipid
membrane, is one of the possible routes by which oxidarive stress inactivates cell functions.
Previously, we have demonstrated that in isolated cardiac membranes,(3-adrenoceptor function
was reduced by oxidative stress (Naenen et al. 1988). Moreover, it has amply been
demonstrated that in the ischemic heart in vivo, ~3-adrenoceptor funcrion is impaired. The aim
of this study was to determine the effect of oxidative stress on (3-adrenoceptor function in heart
membranes and in a more intact system, viz. the isolated left atrium and right ventricle of the rat.
Material and methods.
Male wistar rats (200-250 g, C.P.B. Harlan Olec Zeist, the Netherlands) were killed by decapitation. The
isolated left atria and right ventricle strip were mounted in water jacketed organ baths. The dose-response curve of
(-)isoproterenol in these organs were constructec as described in Naenen et al.(1989).
Heart membranes were prepared from heart left atria according to the method previously described(Naenen et
al. 1988). The membranes were taken up in buffer A (50 mM Tris-HC1,140 mM NaCI and 5 mM MgC12, pH
7.4 at 37 C) and stored in liquid nitrogen until use. The receptor binding and adenylate cyclase activity assay
were performed with membranes suspended in buffer A(pH 7.4 at 37 C) as described previously (Naenen et al.
1988). Protein determinations were made by the method of Bradford (1976) using Bodine serum albumin as
standard.
The chemicals used were (-)-timolol maleate, cumene hydroperoxide(CHP)and guanylylimidodiphosphate
(GppNHp)(Sigma).(-)-[125I]-iodocyanopindolol(ICYP) was obtained from New England Nuclear and the cyclicAMP assay kit(code TRK.432) was obtained from Amersham. All other chemicals were of reagent grade.

Results.
The effect of oxidative stress on (3-adrenoceptor function in the heart was determined.
Oxidative stress was induced by the addition of the synthetic organic hydroperoxide cumene
hydroperoxide (CHP). We assessed the effect of oxidative stress on (3-adrenoceptor function in
isolated calf membranes and in isolated organs. In the calf ventricle membranes the parameters
studied were (3-adrenoceptor density and c-AMP formation. In the isolated organs we also
determined the effect of oxidative stress on the ~3-adrenoceptor mediated response. The organs
used were (i) the isolated left atrium of the rat (ii) the isolated right ventricle of the rat.

151

before oxidarive stress

after oxidative stress

552

584

(3-adrenoceptor density (fmol/mg protein)


B
1235

1428

c-AMP producrion (pmol/min/mg protein)


basal
21 1
10-2 M NaF
60 2

8 2
21 3

ICYP(pM)
Kp

Table 1. The effect of oxidative stress, induced by 0.1 mM cumene hydroperoxide, on the dissociation
constant of ICYP for the ~i-adrenoceptor(K~,(3-adrenoceptor density and on c-AMP production n calf
ventricle membranes
1 Effect oxidarive stress on the (3-adrenoceptor system of calf ventricle membranes.
Using the radioligand ICYP, it was found that calf ventricle membranes possess a
(3-adrenoceptor with a affinity for ICYP of 55 pM and a density of 123 fmol/mg protein.
Oxidative stress was induced by cumene hydroperoxide (CHP). The membranes were exposed
at 37 C to 0.1 mM CHP. After 20 minutes the membranes were washed, in order to remove
CHP. Subsequently, a radioligand binding study on these membranes was performed. It was
found that CHP pretreatment increased receptor density, whereas it had no effect on the affinity
ofICYP for the receptor (table 1).
Also the effect of oxidative stress on c-AMP formation was determined. It was found that
0.1 mM CHP reduced c-AMP production, basal and induced by NaF(10 2 M)to approximately
35 Io of the control value (table 1).
2 Effect oxidative stress on ~3-adrenoceptor mediated response in isolated organs.
In order to asses the effect of oxidative stress on a ~3-adrenoceptor-mediated effect of the
heart, the isolated left atrium and right ventricle of the rat were used. It has well been
documented that the positive inotropic effect of agonists like (-)isoproterenol and ()terbutaline
in these organs is mediated via activation of the ~3-adrenoceptor.
2.1 The isolated left atrium of the rat.
In the isolated, field stunulated left atrium, oxidative stress was induced. by adding CHP in a
final concentrarion of 0.1 mM to the Krebs buffer in the organ bath for 20 minutes. Addirion of
0.1 mM CHP induced a gradual increase in the contraction force. After approximately 15
minutes, the baseline tension rose and the contraction force reduced until the atria finally
stopped contracting (iig 1). At that time the incubation with CHP was terminated and the atria
were washed for 1 hour in order to remove CHP. During this washing period, the atria
recommenced contracting. ~3-adrenoceptor response was determined by constructing a dose
response curve with (-)isoproterenol or ()terbutaline. The dose response curve obtained after
oxidative stress was compared to the dose response curve obtained before o~dative stress. Also
(3-adrenoceptor density and c-AMP formation in membranes of control atria and of atria
subjected to oxidative stress were determined. Moreover, in order to further assess which
component of the (3-adrenoceptor complex was affected by oxidarive stress, the positive
inotropic effect of dibutyryl c-AMP and forskolin before and after oxidative stress were
measured.
152

Figure 1. Typical effect of cumene hydroperoxide(CHP)on the contraction force of field stimulted left
atria. At the pme point indicated by the arrow 0.1 mM CHP was added.

2.1.1 Effect of oxidative stress on the positive inotropic response to (-)isoproterenol.


In the positive inotropic effect of (-)isoproterenol a cascade of biochemical events is
involved. First of all, the compound has to bind to the receptor. Binding of an agonist to the
receptor activates the receptor, and by receptor acrivation a sequence of reactions is triggered
that finally produces the positive inotropic effect. Areceptor-mediated effect can be divided into
three processes.
The first process is binding of the agonist(A)to the receptor (R). In the simplest case this is
the reversible interaction of the agonist with the receptor.
A+RAAR
Binding of the agonist is described by the dissociarion constant of the agonist(KA).
[A].[R]
KA = ~~~

The second process in areceptor-mediated effect is acrivation of the receptor. It is assumed


that receptor activation produces a stimulus (S). The stimulus depends on the fraction of
receptors that is occupied (y)and the agonist used. Some compounds are not able to acrivate the
receptor after binding (antagonists) while receptor activarion by full agonists is higher than that
of partial agonists. The term that deternunes the stimulus produced by occupation of a certain
fraction of receptor by an agonist is the efficacy of that agonist.
S=ey
The fraction receptors occupied by the agonist is determined by the concentration agonist([A])
and the affinity cnstant of the agonist(Kp).
LAl
[AR]
y [AR]+[R] [A]+ KA

~1~

The third and last process in areceptor-mediated response is the transfer of the stimulus (S)
in the effect(E =the effect related to the maximal receptor mediated effect). We applied the
theory of Stephenson (1956) that provides no theorerical relarion between stimulus (S) and
effect(E), but "an empirical relation has to be obtained experimentally".
E = f(S)
153

ICYP(per
Kp
(-)isoproterenol(10-~ Ivn
Kp

before oxidative stress

after oxidative stress

26 1

24 1

3.4 0.5

~i-adrenoceptor density (fmol/mg protein)


B~
383 15

3.0 0.2

375 25

Table 2. The dissociation constant (KA) of ICYP and (-)isoproterenol for the (3-adrenoceptor and
(3-adrenoceptor density in membranes prepared from rat left atria before or after oxidative stress induced by
0.1 mM cumene hydroperoxide.

The effect oxidarive stress had on all three processes involved in the posirive inotropic effect of
(-)isoproterenol was quanriiied.
The binding of (-)isoproterenol to the ~3-adrenoceptor was determined using a radioligand
binding study. Membranes of control atria and atria exposed to 0.1 mM CHP were prepared. It
was found that there was no difference in (3-adrenoceptor density and in affinity of ICYP for the
receptor between both membranes. Also the affinity of(-)isoproterenol for the (3-adrenoceptor
was not altered by oxidative stress (table 2). Therefore, it was concluded that oxidative stress
had no effect on the affinity of the agonist for the receptor.
In order to determine the effect of oxidarive stress on the stimulus produced after receptor
binding by (-)isoproterenol (i.e. the efficacy of (-)isoproterenol), we compared the doseresponse curves'consiructed before and after oxidarive stress. It was found that preincubarion of
the atria with 0.1 mM CHP increased the concentration of (-)isoproterenol that produced half
maximal effect(D2). The pD2(=-log D2)decreased from 8.5 to 8.0 (table 3). Also the maximal
effect obtained with (-)isoproterenol was reduced; only 50% of the maximal effect obtained
before oxidative stress could be reached (table 3).
We applied the calibrarion point introduced by Stephenson for calculation of the efficacy.
According to Stephenson, the srimulus is unity at half maximal receptor mediated effect(when E
= 0.5 than S = 1). For a full agonist (but not for a partial agonist) half maximal receptor
mediated effect is obtained at D2. Therefore, the efficacy can be calculated for a full agonist
using the following equarion:
e = D2D KA

(2)

before oxidarive stress


(-)isoproterenol
pD2
Mac Effect(%)
e
as

8.50.2
100
11030
1

after oxidative stress


8.00.2
50 8
3410
1

Table 3. The pD2, maximal effect, efficacy (e) and intrinsic activity (aS) of the positive inotropic effect
of(-)isoproterenol in the isolated left atrium before or after oxidative stress induced by 0.1 mM cumene
hydroperoxide.
154

We assumed that the affinity of (-)isoproterenol for the (3-adrenoceptor (Kp)in the functional
study is identical to that obtained in the radioligand binding study on membranes of the atria,
i.e. 3 10-~ M. Using this method we found that the efficacy of (-)isoproterenol was
respecrively 110 and 34 in control atria and atria subjected to o~dative stress (table 3).
The third and last part in an agonist mediated response is the translarion of the stimulus,
generated by receptor activation, into an effect. In the his receptor theory, Clark assumed that
the magnitude of the receptor-mediated effect is linearly dependent on the fraction receptors
occupied by an agonist,irrespecrive of the agonist used (Kenakin 1987). Since it was observed
that the meimal effect obtained with several agonists was not identical, Arins et al. (1954)
modified the receptor theory. "lhey introduced intrinsic activity (aE), a term that originally
directly related the effect to the fracrion receptors occupied. Because different agonists can have
different intrinsic activities, it is possible that a different effect is induced by two agonists that
occupy the same fraction of receptors. However, in his original theory Arins assumed that
there was a linear relation between fraction receptor occupied and effect(E = aE y). It was
observed that this relation usually is not linear, which made Stephenson (1956) to modify the
receptor theory again. The major advantage of the theory of Stephenson is that no attempt is
made to produce a mathematical model for the relation between the fraction receptors occupied
and the effect. The occuparion of a receptor by an agonist produces a srimulus (S = e y), and
via an unknown function this stimulus is transformed into an effect(E = f(S)). This function is
only dependent on the tissue an the type of receptor used. So for different agonists the stimuluseffect transfer is identical, provided the same tissue was used and the effect is mediated via the
same type of receptor(Stephenson, 1956).
In order to obtain the stimulus-effect relation in the left atrium for j3-adrenoceptors, we
combined the radioligand binding study with the dose-response curve of(-)isoproterenol. Again
assuming that the affinity for (-)isoproterenol in the funcrional study was identical to that
obtained in the radioligand binding study, we calculated the fraction receptors occupied by each
concentration of (-)isoproterenol used in the construcrion of the dose response curve.- The
stimulus generated by those fractions of receptors occupied could also be calculated, since the
efficacy of (-)isoproterenol is known. Moreover, the effect produced by each concentrarion of
(-)isoproterenol was determined in the functional study. Therefore we were able to convert the
dose-response curve for (-)isoproterenol into astimulus-effect relation. In figure 2, the
stimulus-effect relations obtained by this procedure of left atria before and after oxidative stress
are superimposed, showing a complete overlap. Apparently, oxidative stress did not alter the
stimulus-effect relation. It should be noted, however, that the effect is related to the maximal
receptor-mediated response. As indicated above, the maximal receptor-mediated effect is
reduced 50 % by oxidative stress.
In conclusion, oxidative stress had no effect on the affinity of (-)isoproterenol for the
~3-adrenoceptor, it reduced the efficacy of(-)isoproterenol from 110 to 31 and it had no effect on
the shape of the stimulus-effect relarion, although maximal receptor mediated effect was reduced
by 50%.
2.1.2 Effect of oxidative stress on the positive inotropic response to ()terbutaline.
Beside (-)isoproterenol, also ()terbutaline was used to quanrify the effect of oxidative stress
on the (3-adrenoceptor-mediated response. The advantage of()terbutaline over (-)isoproterenol
is that ()terbutaline is a partial agonist. As will be demonstrated, it is possible to quantify
155

1.0
+:~

+*'
,
i~.
#~}+:f

U
N

~~
~~+

0.5

~:
~+
.~.i

OHO

.O1

.1

10

100

Stimulus
Figure 2. Stimulus-effect relation in field stimulated left atria for the p-adrenoceptor system, constructed
with (-)isoproterenol in control atria () or atria pretreated with 0:1 mM cumene hydroperoxide (+).

receptor activation of a partial agonist with a functional study, without using a radioligand
binding study.
In the dose response curve in control atria, it was found that with ()terbutaline only 81 % of
the meimal (-)isoproterenol induced effect could be reached, indicating that ()terbutaline is
indeed a partial agonist in this tissue. The pD2 of()terbutaline was found to be 5.2 in control
atria. After oxidative stress, the pD2 of()terbutaline was reduced to 4.6 and only 64 % of the
maximal (-)isoproterenol induced effect could be reached with ()terbutaline (table 4). It should
be noted that oxidative stress reduced the maximal(-)isoproterenol stimulated effect with 50 %.
In order to quantify the interaction of()terbutaline with the (3-adrenoceptor, we have to use
the stimulus-effect relation for the ~i-adrenoceptor in this tissue obtained with (-)isoproterenol.
The stimulus-effect relation is supposed to be a truly tissue-dependent function, and both
(-)isoproterenol and ()terbutaline make use of the same function to produce the final positive
inotropic effect in the isolated left atrium.
To obtain this stimulus-effect relation we transformed the dose-response curve for
(-)isoproterenol using the affinity constant of(-)isoproterenol obtained in a radioligand binding
study (KA = 3 10-~ M). Seemingly, a radioligand binding study is needed to produce the
srimulus effect relation. However,in constructing the dose-response curve it was observed that
(-)isoproterenol at a concentrarion of 3 10-g M already produced the maximal (3-adrenoceptormediated effect. This concentration is much smaller that the dissociation constant of
(-)isoproterenol for the receptor(3 10-~ M,table 2). Therefore equation (1) that describes the
fraction of receptors occupied can be simplified for these concentrations of(-)isoproterenol:
[A]
Y= KA
156

before oxidative stress


()terbutaline
pD2
M~ Effect(%)
e
as

after o~dative stress


4.60.2
64 6
1.240.03
0.034 0.010

5.20.2
81 7
3.10.8
0.028 0.006

Table 4. The pD2, maximal effect, efficacy (e) and intrinsic activity (as)of the positive inolropic effect
of ()terbutaline in the isolated left atrium before or after oxidative stress induced by 0.1 mM cumene
hydroperoxide. The maximal effect is expressed relative to the maximal (-)isoproterenol mediated
response.

Also the D2 of (-)isoproterenol is much smaller than the dissociation constant, so formula (2)
used to calculate the efficacy can be simplified to:
KA
e=
DZ
This indicates that the stimulus induced by the concentrations (-)isoproterenol employed in the
dose-response curve is independent of the dissociation constant:

S=e ,Y _[Al
D2
Therefore, the stimulus-effect relation constructed with the dose-response curve of
(-)isoproterenol is independent of the dissociation constant of (-)isoproterenol obtained in the
radioligand binding study. The only prerequisite is that the dissociation constant of the full
agonist used to construct the stimulus-effect relation is much greater than the concentrarion of
the full agonist that is needed to produce maximal receptor-mediated effect. The radioligand
binding study is merely needed to verify this assumption.
Using this srimulus-effect relation, it is possible to transfer the dose-response curve of
()terbutaline into adose-stimulus curve. It can be determined which stimulus has to be
produced by ()terbutaline to induce the effect observed in the dose-response curve, using the
stimulus-effect relation. Hereby it is assumed that in order to produce an identical effect,
()terbutaline has to produce the same stimulus as (-)isoproterenol. The dose-stimulus curve
obtained is a direct reflecrion of binding of the partial agonist to the receptor. The dose-srimulus
curve could be fitted by a one receptor model using the formula:
[A]
S e [A]+KA
It was found that the dissociation constant of ()terbutaline for the (3-adrenoceptor was 2.1
0.4 10 5 M, and that the efficacy of ()terbutaline in this tissue was 3.1 before oxidative
stress. Oxidative stress did not affect the affinity of()terbutaline but it reduced the efficacy of
()terbutaline to 1.24 (table 4). No data of a radioligand binding study had to be used to obtain
these results.
2.1.3 Effect of oxidative stress on c-AMP production.
In the positive inotropic effect of (-)isoproterenol a cascade of biochemical reactions is
involved. After receptor activation, the receptor-agonist complex couples to a GS-protein. The
GSprotein becomes activated, and dissociates from the agonist-receptor complex. Subsequently

157

before oxidative stress


c-AMP production (pmol/min/mg protein)
basal
52
10-2 M NaF
45 12
10~ M GppNHp
39 7

after oxidarive stress


2 2
17 5
14 5

Table 5. c-AMP production in membranes prepared from rat left atria before or after oxidative stress
induced by 0.1 mM cumene hydroperoxide.

the activated GSprotein stimulates adenylate cyclase, an enzyme that catalyzes the formation of
c-AMP. c-AMP impels a protein 1Qnase which eventually leads to the pharmacological response.
In order to deternune which component of the (3-adrenoceptor system was res~ionsiUle for
the reduction of the efficacy of (-)isoproterenol and ()terbutaline, c-AMP production in
membranes prepared form control and CHP pretreated atria was determined. As already
indicated in table 2, oxidative stress had no effect on (3-adrenoceptor density. However,c-AMP
production, basal and induced by NaF(10-2 M)or GppNHp (10-4 M), was reduced (table 5).
2.1.4 Effect of oxidative stress on the positive inotropic response to forskolin and dibutyryl
c-AMP.
Also in order to deternune which component of the ~i-adrenoceptor in the isolated left atrium
is affected by oxidative stress, the posirive inotropic effect of dibutyryl c-AMP and forskolin
before and after oxidative stress were measured. It was found that the maximal effect of
forskolin and dibutyryl c-AMP was reduced to the same extent as that of(-)isoproterenol, but in
contrast to the pD2 of (-)isoproterenol the pD2's of forskolin and dibutyryl c-AMP were not
affected by oxidative stress (table 6).
2.2 The isolated right ventricle of the rat.
The same procedure as in the isolated left atrium was used in the field stimulated right
ventricle to induce oxidative stress; 0.1 mM CHP for 20 minutes. The direct effect of CHP on
the contraction of the right ventricle were comparable to that of the left atrium (data not shown).
To assess the effect of oxidative stress, the response of the ventricle before and after incubarion
with 0.1 mM CHP was determined. It was found that the applied oxidarive stress reduced the
pD2 of (-)isoproterenol from 8.0 to 7.5 (table 7). The maximal effect obtained with
(-)isoproterenol was reduced to 52% of that before oxidative stress.

before oxidarive stress


forskolin
pD2
dibutyryl c-AMP
pD2

after oxidarive stress

6.10.1

6.00.2

3.70.2

3.70.2

Table 6. The pD2 of the positive inotropic effect of forskolin and dibutyryl c-AMP in the isolated left
atrium before and after oxidative stress induced by 0.1 mM cumene hydroperoxide.

158

before oxidative stress


(-)isoproterenol
pD2
Max Effect(%)

8.00.1
100

after oxidative stress


7.50.1
52 1

Table 7. The pD2 and maximal effect of the positive inotropic effect of(-)isoproterenol in the isolated
right ventricle strip before or after oxidative sirens induced by 0.1 mM cumene hydroperoxide.

Discussion.
T'he effect of oxidative stress on (3-adrenoceptor funcrion in the rat heart was determined. To
this end calf ventricle membranes, the isolate.~i left atria and right ventricles of the rat veere
exposed to 0.1 mM cumene hydroperoxide(CHP)for 20 minutes.
In the calf membranes, CHP increased ~3-adrenoceptor density, while c-AMP production
was reduced. Previously, we reported that oxidative stress induced by hydrogen peroxide in
ventricle membranes had a similar effect (Naenen et al. 1988). At that time we explained the
increase in receptor density by the effect o~dative stress has on the physical state of the
membrane. Oxidative stress reduces membrane fluidity, and a reduction of membrane fluidity
was reported to increases receptor density. However,it should be noted that receptor density is
related to the amount of protein. Possibly, oxidative stress modifies the membrane proteins in
such a way that they react less efficiently with the color reagent used to quantify the amount of
protein. This might also contribute to the observed increase in ~i-adrenoceptor density.
Nevertheless, the conclusion that c-AMP formarion is much more vulnerable towards oxidarive
stress induced by hydrogen peroxide than the (3-adrenoceptor itself still holds, and is consumed
in the experiment with CHP in this study.
The effect of oxidative stress on (3-adrenoceptor function in the isolated left atrium and right
ventricle was assessed in a functional study. By the ~3-adrenoceptor agonists (-)isoproterenol or
()terbutaline, an increase in the contraction force s induced. In the positive inotropic response
to these agonists firstly the agonist has to bind to the receptor. Subsequently, the receptor in the
agonist-receptor complex becomes activated. Receptor acrivarion initiates a cascade of reactions
that imally produces the effect.
One of the features of such a cascade is that the stimulus generated by receptor activation is
amplified (Arins et al. 1979). In this way a relatively small amount of receptors can mediate a
physiological important effect. As shown in figure 2, the transformation of the input signal is
not linear in the case of the (3-adrenoceptor-mediated increase of the contracrion force in the left
atrium. Moreover, inferred from the same srimulus-effect relarion, above a srimulus of 10 an
increase of the srimulus will not result anymore in an increase of the contraction force. Since
(-)isoproterenol has an efficacy in control atria of 110, this indicates that (-)isoproterenol only
has to occupy 9 Io of the receptors in order to produce maximal effect in control atria.
Increasing this fraction above 9 %increases the stimulus but does not increase the effect. This
phenomenon has been given the name of "spare receptors" or "receptor reserve" (Kenakin
1987). Apparently, in control atria there is a receptor reserve of 91% for (-)isoproterenol.
Despite its name,receptor reserve does not mean that only a part of the receptors is involved in
the production of the maximal effect, but all receptors only need partial activation in time to
reach malcimal organ effect; there is a spare capacity in each receptor.
159

The phenomenon of receptor reserve can be explained by looking at the cascade of reaction
that accomplishes the stimulus-effect transfer. With a gradual increase of the initial
stimulus,
one of the enzymes in the reaction sequence will be the first to reach saturation, although
the
initial stimulus can still be far from its ma~cimal level. A further enhancement of
the initial
stimulus, although giving rise to an increase in the reactions preceding the limiting one, will
not
result in an increase in the smal effect because the saturated reacrion functions as a bottle
neck.
In a funcrional study, the pD2 of a full agonist is detemuned by receptor density,
the affinity of
the agonist for the receptor, the potency of that agonist to activate the receptor and
by the
reactions preceding the bottle neck reaction. The mammal effect is determined by the bottle
neck
reacrion and the reacrions succeeding it.
In this study, it was found that oxidative stress recuced the maximal effect obtained
by
(-)isoproterenol in the left atrium. As inferred form the stimulus-effect relation
obtained after
CHP treatment (fig. 2), in order to produce maximal receptor mediated effect after
oxidarive
stress a srimulus of approximately 10 is needed. The maximal stimulus that could be
generated
with (-)isoproterenol (i.e. the efficacy of (-)isoproterenol) was 34. Therefore,
it can be
concluded that (-)isoproterenol was still able to saturate an enzyme in the reaction cascade
and
thus (-)isoproterenol remained a full agonist after oxidative stress. Apparently, oxidativ
e stress
reduces the maximal inotropic effect that can be mediated by (3-adrenoceptors in the left atrium.
This can be explained by a partial inactivation of an enzymes) that is situated in the
cascade
after the saturable enzyme, or/and by a partial inactivation of the saturable enzyme
itself.
Recently, we reported that sulfhydryl alkylating agents that are produced during
oxidative
stress, like 4-hydroxy-2,3-trans-nonenal, reduce the maximal response to (-)isoproterenol
without affecting the pD2. Possibly the production of these compounds is involved in
the
reduction of the ma~cimal inotropic effect by oxidative stress.
Beside a reduction of the maximal effect of (-)isoproterenol, oxidative stress induced
by
CHP reduced the pD2 of(-)isoproterenol. Since oxidative stress had no effect on the
affinity of
(-)isoproterenol for the receptor, nor on ~3-adrenoceptor density, this pD2 shift is probably
caused by a partial inactivation of an enzymes) situated before the saturable enzyme in
the
reaction cascade that performs the stimulus-effect transfer. It is also possible that after
oxidarive
stress another enzyme that is involved in the stimulus-effect transfer becomes the limiting one.
The pD2 is not a good parameter for quantification drug-agonist interaction. Therefor
e, the
efficacy of(-)isoproterenol before and after oxidative stress was determined. It was found
that
oxidative stress reduced the efficacy of (-)isoproterenol from 110 to 31. Using a radiolig
and
binding study, it was found that oxidarive stress had no effect on (3-adrenoceptor density.
Since
the efficacy was reduced by oxidative stress, and (3-adrenoceptor density was not altered,
the
efficacy per receptor (i.e. the intrinsic efficacy) of (-)isoproterenol is reduced by oxidativ
e
stress. Apparently, intrinsic efficacy is not a truly drug pazameter, in contrast to the suggesri
on
Kenaln (1987) who stated that it is a term which is constant in all situations.
One of the disadvantages of the experiments with (-)isoproterenol is that we had to combine
a funcrional study with a binding study. Therefore, we extended our studies by using a partial
agonist, because it is possible to quanrify receptor interaction of a partial agonist only with
a
functional study, as described in the results section. Of several partial agonists
tested,
()terbutaline appeared to be the best suited, because it produced a positive inotropic effect
that
was not too high to determine the efficacy of the partial agonist accurately, nor too small
to
160

measure the positive inotropic response to the partial agonist accurately. Hereby it should be
kept in mind that after oxidative stress, the maximal effect of the parrial agonist was reduced to
an even greater extent as the maximal effect of the full agonist. The disadvantage of
()terbutaline is that it is a better agonist on X32-adrenoceptors than on j31-receptors.
(-)Isoproterenol does not discriminate between (31- and (32-adrenoceptors (Kaumann et al.
1989). The positive effect on the heart is mainly mediated via (31-adrenoceptors, although there
is also a substantial contribution of j32-adrenoceptors. Nevertheless,in the stimulus-effect curve
of ()terbutaline in the left atrium we constructed, that is a direct reflection of binding of
()terbutaline to the receptor, we did not observe a two receptor model.
The reduction of the efficacy of()terbutaline was identical to that of(-)isoproterenol. Since
~3-adrenoceptor density was not altered, the intrinsic efficacy of()terbutaline was also rerluceci,
again indicaring that nor efficacy nor intrinsic efficacy are truly drug-constants.
Alternative parameters have been introduced to describe drug-agonist interaction. To
accommodate his theory to the observation that the magnitude of areceptor-mediated effect
usually is not linear proportional to the fraction receptors occupied, Arins(1964) modified his
receptor theory. He adopted the receptor theory first expounded by Stephenson (1956) and
incorporated astimulus-effect transduction in his concept. In accordance to the theory of
Stephenson, the relation between stimulus and effect is undefined in the modified theory of
Arins. Moreover, he redefined intrinsic acrivity: "The intrinsic efficacy of a drug is inversely
proportional to the number of receptors that has to be occupied to induce a certain srimulus and,
therefore, to bring about a certain effect" in a certain tissue (Arins 1964). To discriminate
between the two definitions of intrinsic activity, intrinsic acrivity in it first definition is depicted
as aE, whereas in its revised definition it is depicted as as; notations that have been used for
this purpose previously (Kenskin 1987). Intrinsic acrivity (as)in his revised deiinirion is not
identical to Stephenson's efficacy, although this has been suggested (Furchgott 1966). The
relation between intrinsic activity (as) and efficacy is that the intrinsic activity (as)of a drug
represents the efficacy of that drug (e) relative to the efficacy of the agonist with the highest
efficacy(ems.
as = e%max
In our opinion intrinsic activity (as) provides a truly drug-dependent parameter, that is
independent of tissue used, type of response and dimension used to express the effect in. This
is also illustrated by the results obtained in this study. Oxidative stress has no effect on the
intrinsic acrivity (as)of()terbutaline relative to (-)isoproterenol, whereas efficacy and intrinsic
efficacy are reduced. It should be noted that intrinsic activity (as) is identical to the more
recently introduced term relative efficacy.
In order to deternune which component of the (3-adrenoceptor is inactivated by oxidative
stress, c-AMP formation in membranes prepared from control or CHP pretreated atria was
determined. It was found that c-AMP production was reduced to the same extent as the
efficacies of (-)isoproterenol and ()terbutaline. Since no spare receptors exist for c-AMP
producrion, the reducrion of the efficacies is probably attributed to a reducet c-AMP production.
Additionally, it was found that the maximal positive inotropic response of the atria to
dibutyryl c-AMP or forskolin was reduced by oxidative stress to the same extent as with
(-)isoproterenol, but in contrast to (-)isoproterenol the pD2 of dibutyryl c-AMP or forskolin was
not affected by oxidative stress. Forskolin directly stimulates adenylate cyclase, while dibutyryl
161

c-AMP mimics the effect of endogenous prauced c-AMP.1fiis indicate that


s
the reduction of
the efficacy of (-)isoproterenol by oxidative stress is not a result of a diminis
hed c-AMP
responsiveness, nor of an impairment of adenylate cyclase. Therefore,it is temptin
g to speculate
that oxidative stress reduces the function of the GS-protein.
It is well known that oxidative stress may induce lipid peroxidation. During
lipid
peroxidation, the physical state of the membrane is altered, the fluidity of
the membrane is
reduced. Moreover, during lipid peroxidation sulfhydryl-reactive aldehydes are formed.
The
inacrivation of membrane-bound enzymes by oxidative stress may proceed directly
by an attack
of free radicals on the enzyme or indirectly by inducing lipid peroxidation.
The effects of lipid
peroxidarion can be me.~iated either by a reduction of membrane fludity or by
the production of
reactive aldehydes. 4-Hydroxy-2,3-trans-nonenal(HNE)is frequently used
to deternune the
effec~ of ehese reactive aldehydes.
We found that HNE decreases the maximal (-)isoproterenol-induced inotropic
response in
isolated left atria (Haenen et al. 1989). Therefore, we speculated that the reductio
n of the
maximal response to (3-adrenoceptor stimulation by oxidative stress could
be mediated by the
production of HNE The reduction of c-AMP formation by oxidative stress
is most likely not
caused by HNE formation.
The proteins involved in the producrion of c-AMP are located on the membrane.
It is well
known that the ~i-adrenoceptor signal transduction over the membrane is
depressed by a
reduction in membrane fluidity (Peters, 1988). Since it is known that
lipid peroxidation
decreases membrane fluidity, the reduced c-AMP formation after oxidative stress
may be caused
by the effect of lipid peroxidarion on the physical state of the membrane. Lee et al.(1982)
stated
that by lipid peroxidation the physical state of the membrane may be altered to such
an extent
that adenylate cyclase dissociates from the membrane. Peroxidation of the membra
ne was
promoted as suitable method to solubilize adenylate cyclase without detergen
ts (Lee et al.,
1982). These results indicate that neither free radicals that induce lipid peroxidation,
nor reactive
aldehydes that are formed during lipid peroxidarion directly inactivate adenylate cyclase.
This
coincides with the finding in this study that after oxidative stress adenylate cyclase
itself is not
inacrivated, but more likely the coupling between adenylate cyclase and receptor
is impaired.
Apparently, the effect oxidative stress has on membrane fluidity via the induction
of lipid
pero~cidation can explain the reduction of c-AMP formation by oxidative stress in the
left atrium.
However, we cannot exclude that free radicals produced during oxidative stress
directly inhibit
c-AMP-formation, e.g. by inactivating the GS protein.
Also in the isolated right ventricle, the effect of oxidarive stress on (3-adrenoceptor
function
was deternuned. It was found that the reducrion of maximal (-)isoproterenol-mediate
d effect by
oxidative stress was identical to that in the left atrium. Moreover, we found that
oxidarive stress
reduced the pD2 of (-)isoproterenol in the right ventricle from 8.0 to 7.5 We perform
ed no
radioligand binding study on membranes from control ventricles and from ventricl
es exposed to
CHP. Therefore, we were not able to calculate the efficacy of(-)isoproterenol before
and after
oxidative stress. However, when it is assumed that the dissociation constant of(-)isopr
oterenol
is smaller than the D2(e.g. when - as tentatively expected by us -the dissociation
constant of
(-)isoproterenol for the ~3-adrenoceptor in the right ventricle is comparable to that
in membranes
of the left atrium) the reduction of the efficacy of (-)isoproterenol in the right ventricl
e by
oxidative stress is identical to that in the left atrium. Apparently, the effect of oxidativ
e stress on
(3-adrenoceptor function in the right ventricle is similar to that in the left atrium;
it reduces both
162

the efficacies of R-agonists and the meimal ~3-adrenoceptor-mediated effect.


In living cells there is a delicate balance between free radical production and the protection
against these highly reactive molecules. An unbalanced radical production is involved in the
etiology of various pathologies (Bies, 1985). Of these pathologies, ischemia-reperfusion injury
in the heart is one of the most acute life threatening. The involvement of free radicals in this type
of injury has amply been demonstrated (Kloner 1988).
The effect ofischemia-reperfusion injury on (3-adrenoceptor funcrion in the heart has gained
very much attention. The parameters studied were ~3-adrenoceptor density and c-AMP
formation. It has been reported that ischemia increases(Devos et al. 1985, Maisel et al. 19$5,
Mukherjee et al. 1979, Mukherjee et al. 1982, Ohyagani et al. 1988, Strasser et al. 1988, Varner
et al. 1988) or has nc effect (Freissmuth et al. 197, ~arliner et al. 1959) on (3-adienoceptor
number. In all these studies it was found that the affinity of radiolabels or isoproterenol(when
tested) was not affected by ischemia. T'he fraction of high affinity for isoproterenol was reduced
(Devos et al. 1985, Freissmuth et al. 1987) or not affected (Strasser et al. 1988). However, the
concentration of isoproterenol needed to induce half maximal c-AMP formarion was not altered
by ischemia(Devos et al. 1985, Freissmuth et al. 1987).
Several explanations have been given for the increase of(3-adrenoceptor density by ischemia.
Maisel et al. (1985, 1986, 1987) argued that during ischemia (3-adrenoceptors were
redistributed from the intracellular vesicles to the membrane. Strasser et al. (1988), and Drimal
et al. (1987) reached a similar conclusion. This upregulation of(3-adrenoceptors by ischemia
might be a result of the decrease in tissue catecholamine level induced by ischemia. However,
the increase in (3-adrenoceptor number did not correlate with the extent of catecholamine
depletion (Mukherjee et al. 1979). Alternatively, it was argued that redistribution of
(3-adrenoceptors was due to an impaired internatilization, an effect described to ATP depletion
(Buja et al. 1985, Strasser et al. 1988). In contrast, Mukherjee et al. (1979) rejected the
mechanism of receptor translocarion from a cytosolic pool to the membrane as a possible
mechanism for the ischemia-induced increase in ~3-adrenoceptor density, even before Maisel et
al.(1985)formulated their hypothesis.
Maisel et al.(1986) also reported that propranolol treatment leads to a similar redistribution
of (3-adrenoceptors as ischemia. After propranolol treatment ischemia did not further alter
receptor density. In contrast, Karliner et al. (1989) reported that neither ischemia, nor the
combination of ischemia and propranolol treatment increased ~i-adrenoceptor density,whereas
propranolol treatment increased (3-adrenoceptor density in the subepicardium but(3-adrenoceptor
density in the subendocardium was not affected by propranolol treatment. In addition,
Ohayanigi et al.(1988)reported that ischemia increased ~3-adrenoceptor density, while neither
propranolol treatment nor the combination of propranolol treatment and ischemia had an effect
on (3-adrenoceptor density.
An increased protein synthesis is not a satisfactory explanation for the increase in receptor
number, because the relative short period of ischemia that already has an effect(Mukherjee et al
1979, Winter et al. 1988). Alternatively, the observed effect may be an artificial one, based
upon alteration in membrane protein content, however, this did not appear to be a valid
explanation (Mukherjee.et al. 1979)
It is tempting to speculate that the effect of ischemia on (3-adrenoceptor is mediated via
oxidative stress. Ischemia gives rise to oxidative stress. As demonstrated previously, oxidative
163

stress affects ~i-adrenoceptor density (Naenen et al. 19$8). A moderate level of oxidative stress
increase ~3-adrenoceptor density, whereas after a more pronounced level of oxdarive stress
(3-adrenoceptor density returns to its control value, or it is reduced. Possibly these effects are
mediated the reduction of membrane fluidity caused by oxidative stress-induced lipid
peroxidation (Naenen et al. 1988). If the reported effects of ischemia are induced by an
alteration of the physical state of the membrane, this also might explain the conflicting results
obtained. In the various studies, the extent of ischemia is probably not exactly the same. T'he
difference in the extent of ischemia might be responsible for the different effects of ischemia on
(3-adrenoceptor density reported. The fact that in the various studies different animals were
used, also contributes to the divergence of the reported effects.
There is also no consensus with regazd to the effect of ischemia on c-AMP formarion. A
distinction should be made between studies where the c-AMP content of the heart during
ischemia is measured, and studies where membranes of the ischemic heart are isolated and the
ability of these membranes to catalyze the formation of c-AMP from ATP is studied. In general,
it is found that immediately after ischemia there is a rise in the c-AMP content of the ischemic
zone of the heart (Corr et al. 1978, Ohyagani et al. 1988, Opie et al. 1982, Podzuweit et al.
1978, Rabinowitz et al. 1975, Wollenberg et al. 1969). However, gradually the c-AMP content
of the heart decreases and reaches the control value (Corr et al. 1978, Rabinowitz et al. 1975).
In isolated membranes it is usually found that isoproterenol-induced c-AMP formation is
reduced by ischemia (Devos et al. 1985, Freissmuth et al. 1987, Karliner et al 1989, Krause
and England 1982, Vatner et al. 1988, Will-Shahab et al. 1985), an observation that coincides
with the reduction of the fraction of(3-adrenoceptors that bind isoproterenol in the high affinity
(Devos et a1. 1985, Freissmuth et al 1987). This reduction was more pronounced after a
prolonged period of ischemia. (Will Shahab et al. 1985). However, an increase in adenylate
cyclase activity (Maisel et al. 1987, Strasser et al. 1988) and an increase in high affinity
receptors density (Strasser et al. 1988) as a result of ischemia have also been reported.
At first sight, an increase in c-AMP content and an inactivation of adenylate cyclase acrivity
induced by ischemia do not concur. However, it has to be taken into consideration that as a
result of ischemia endogenous catecholamines are released, and these catecholamines may
induce c-AMP formation (Opie et al. 1982, Rabinowitz et al. 1975). When ischemia
continuous, the reduction of adenylate cyclase activity by ischemia becomes more pronounced
and the catecholamines may become depleted. This may be responsible for the reduction of the
c-AMP content after prolonged ischemia.
Will-Shahab et al. (1987) reported that after moderate ischemia the reduction of adenylate
cyclase activity was reversible, while a more pronounced ischemia caused irreversible damage
to adenylate cyclase. They argued that the reversible damage was due to an enhanced calcium
content of the heart, whereas the irreversible damage was caused by oxidative stress (WillShahab et al. 197). The results presented in the present study support this suggestion.
However, according to Will-Shahab et al. (1987) adenylate cyclase is the target of oxidative
stress, whereas the results presented in this study indicate that the coupling protein rather than
adenylate cyclase itself is affected by oxidarive stress, Possibly, the irreversible inactivation is
actually mediated by a reducrion of membrane fluidity.
It has generally been accepted that ischemia reduces maximal (3-adrenoceptor-mediated
response (Weiss et al. 1984). Several explanations have been given for this phenomenon. It has
been suggested that ischemia-induced acidosisis or ATP depletion is responsible for the reduced
164

maximal response. Matthews et al.(1981) and Krause and England (1982)indicated that there
has to be a different cause. Based on the results obtained in this study, this effect of ischemia
might also be mediated by oxidative stress, and possibly by the formation of reacrive aldehydes
like HNE. It has been reported that hypoxia impairs the activity of c-AMP-dependent protein
kinase towards the contractile proteins ((England and Krause 1987). Possibly the reactive
aldehydes inacrivate the same protein.
When the data reported in the literature and the results obtained in this study are combined,
the following hypothesis emerges. During ischemia endogenous catecholamines are released.
These catechalamines interact with the (3-adrenoceptor and stimulate adenylate cyclase. The
excessive c-AMP level produced in this way is cardiotoxic and may cause e.g. arrhytmias. The
increase in c~AMP formation is not a secondary phenomenon iii relatio~i to injury, because the
elevated c-AMP level is observedunmediately before the development of arrhytmias(Blaiklock
et al. 1978). In addition it has been reported that catecholamines may induce lipid peroxidation
in the heart(Noronha-Dutra et al. 1984, Ohta et al. 1988, Sushama Kumari and Menon 1988).
It has been suggested that the free radicals of catecholamines are able to induce lipid
peroxidation (Singal et al. 1982). However, previously we have shown that catecholamines
protect against lipid peroxidation (Naenen et al 1989). The direct effect of catecholamines on
oxidative stress is that they tend to shift the effect form lipid peroxidarion to sulfhydryl arylation
(Naenen et al. 1989). We found that sulfhydryl arylation by catecholamines does not lead to a
reduced ~3-adrenoceptor function in the heart (unpublished results). Since it was found that the
cardiotoxic effects of catecholamines can be mimicked by analogues of c-AMP (Opie at al.
1982), probably the elevated c-AMP formation is responsible for the catecholamine-induced
lipid peroxidation in the heart. In this respect it should be noted that by lipid peroxidation the
catecholamine-induced c-AMP formarion is reduced. This might provide a feedback mechanism
that protects the heart against R-adrenoceptor hyperstirnularion.
Beside (3-adrenoceptor hyperstimulation, also ischemia gives rise to oxidative stress via other
mechanisms.(3-adrenoceptor bloclng agents aze able to protect against excessive stimulation of
the (3-adrenoceptor by endogenous catecholamines. In principle, no protection is provided
against the other ways by which ischemia leads to heart injury by blocking the (3-adrenoceptor.
The results presented in this and a previous study (Naenen et al. 1988)indicate that oxidative
stress reduces the toxic effects by ~3-adrenoceptor hyperstimulation by reducing c-AMP
formation. In the mean while, the protection against oxidative stress may become reduced.
Therefore, it might be more profitable to protect against oxidative stress than to prevent
j3-adrenoceptor stimulation during ischemia. In fact, the antioxidant capacity of (3-blocking
agents (Mak and Weglicki 1989) probably also contributes to the therapeutic effect of these
compounds in ischemia.
References
Arins, E.J.(1954) Arch. Int. Pharmacodyn. Ther. 99, 32-49.
Arins,E.7.(1964) Molecular pharmacology, volume 1, Academic press, New York -London.
Arins, E.J. , A.J. Beld, J.F. Rodrigues de Miranda and A.M. Simonis (1979) in: R.D. O'Brien (Ed.) The
receptors, a comprehensive treatise, Plenum press, New York, pp. 33-91.
R.G. Blaiklock, E.M. Hush and D. Lehr (1978) J. Mol. Cell. Cardiol. 10,499-509.
L.M. Buja, K.H. Muntz, T. Rosenbaum,Z. Haghani, D.K. Buja, A. Sen, K.R. Chien and J.T. Willerson (1985)
Circ. Res. 27, 640-645.
M.M.Bradford (1976) Anal. Biochem. 72,246-254.
165

P.B. Corr, F.X. Witkowski and B.E. Sobel (!978) J. Clin. Invest. 61, 109-119.
C. Devos, P. Robberecht, P. Nokin, M. Waelbrceck, M. Clinet, J.C. Camus, P. Beaufort, P. Schoenfeld and J.
Cristophe (1985) Arch. Pharmacol. 331,71-75
J. Drietal, V. Knezl, D. Magna and K. Strizova (1987) Gen. Physiol. Biophys. 6, 583-591.
P.J. England and E.G. Krause(19 7)Biomeel Biochim. Acta 46,369-380.
M. Freissmuth, W. Schutz, M. Weindlmayer-Gtittel, M. Zimpfer and C.K. Spiss (1987) J. Cardiovasc.
Pharmacol. 10, 568-574.
R.F. Furchgott (1966) in: N.J. Harper and A.B. Simmonds (Eds.) Advances in drug research, volume 3,
Academic press, London -New York,pp. 21-55.
G.R.M.M. Haenen, P. v. Dansfik, N.P.E. Vermeulen, H. Timmerman and A. Bast (1988) Free Rad. Res.
Comets. 4, 243-249.
G.R.M.M. Haenen, H.J.M. Ylug, N.P.E. Vermeulen, H. Timmerman and A. Bast(1989) Life Sci. 45,71-76.
J. Karliner, M.B. Stevens, N. Honbo and I.E. Hofman (1989) J. Clin. Invest 83, 474-481.
A.J. Kaumann, H.Lemoine, U.Schwederski-Menke a,~d B. Ehle (1989) Arch. Pharmacol. 339,99-112.
T.P. Kenskin(1987)Pharmacological analysis of drug-receptor interaction, Raven press, New York.
R.A. HIoner(1988)Free Radical Biol. Meel. 4,5-7.
E.G. Krause and P.J. England (1982) J. Mol. Cell. Cardiol. 14,611-613.
E. Lee, A. Baba, A. Ohta and H. Iwata (1982) Biochim. Biophys. Acts 689, 370-374.
A.S. Maisel, H.J. Motulsky and P.A. Insel(1985) Science 230, 183-186.
A.S. Maisel, H.J. Motulsky and P.A. Insel (1986) Circ. Res. 60, 108-112.
A.S. Maisel, H.J. Motulsky, M.G. Ziegler and P.A. Insel (1987) Am. J. Physiol. 253, H1159-H1166.
I.T. Mak and W.B. Weglicki (1988) Circ. Res. 63, 262-266.
P.M. Matthews, G.K. Radda and D.J. Taylor (1981) Biochem. Soc. Transact. 9, 236-237.
K. Mori(1975) Nagoya J. Meel Sci. 39,9-14.
A. Mukherjee, T.M. Wong, L.M. Buja, R.J. Lefkowitz and J.T. Willerson (1979) J. Clin. Invest. 64, 14231428.
A. Mukherjee, L.R. Bush, K.E. McCoy, R.J. Duke H. Hagler L.M. Buja and J.T. Willerson (1982) Circ. Res.
50,735-741.
A.A. Noronha-Dutra,E.M. Steen and N.Woolf, Basic Res. Cardiol. 80,S133-S136.
H. Ohta, J. Azuma, N. Awata, T. Hamaguchi, Y. Tanaka, A. Sawamura,S. Kishimoto and N. Sperelakis (1988)
Cazdiovasc. Res. 22,407-413.
L. Opie, F. Thandroyen, C. Muller and P. Daries (1982) in: R.A. Riemersma and M.F. Oliver (Eds)
Catecholamines in the non-ischaemic and schaemic myocardium, Elsevier, Amsterdam, pp. 203-222.
M.Ohyagani, Y. Matsumori and T. Iwasaki(1988)J. Cardiovasc. Pharmacol. 11, 107-114.
R. Peters (1988) FEBS leu 243, 1-7.
T. Podzuweit, A.J. Dalby, G.W. Cherry and L.H. Opie(1978) J. Mol. Cell. Cardiol. 10, 81-94.
B. Rabinowitz, W.W. Parmley, M. Kligerman, J. Norman, S. Fujimura, S. Chiba and 7.M. Matloff (1975)
Chest 68, 69-74.
H. Sies(Ed.)(1985)Oxidative Stress, Academic Press, New York,London.
P.K. Singal, N. Kapur, K.S. Dhillon, R.E. Beamish and N.S. Dhalla (1982) Can. J. Phys. Pharmacol. 60,
1290-1397.
R.P.Stephenson (1956) Brit. J. Pharmacol. 11, 379-393.
R.H. Strasser, J. Krimmer and R. Marquetant,(1988)J. Cardiovasc. Pharmacol. 12 S15-S24.
S. Bushsets Kumari and V.P. Menon (1988)7. Biosci. 13, 257-262.
D.E. Varner, D.R. Knight, Y.T. Shen, J.X. Thomas, C.J. Homcy and S.F. Vatner (1988) 7. Mol. Cell. Cardiol.
20, 75-82.
J. Weiss, G.S. Couper, B. Hiltbrand and K.I. Shine (1984) Am. J. Physiol. 247, H760-H767.
L. Will-Shahab, E.G. Krause, S. Bartel, W.Schulze and I. Kintner(1985)J. Cazdiovasc. Pharmacol. 7 S23-S27.
L. Will-Shahab, I. Schimke, A. Haberland and I. Kutiner(1987)Biomeel. Biochim. Ac[a 46 S427-5432.
R.J.D. Winter, P.M. Inkpen, K.E.J. Dickinson, R.M. Rudd and P.S. Sever (1988) Cardiovasc. Res. 22, 159162.
A. Wollenberg, E.G. Krause and G. Heier(1969) Biochem. Biophys. Res. Comets 36,-664-670.

i:~

- Chapter 16 Contribution of 4-hydroxy-2,3-trans-nonenal to the reduction


of ~3-adrenoceptor function in the heart by oxidative stress
Abstract. Oxidative stress reduces adenylate cyclase acrivity and also the maximal response to
(3-adrenoceptor stimulation in the rat heart, while (3-adrenoceptor density is not affected or
increased. Since free sulfhydryl groups are essential to ~i-adrenoceptor funcrion and the
sulfhydryl reactive substance 4-hydroxy-2,3-trans-nonenal(HNE)is responsible for part of the
effects of oxidative stress, the effect of HNE on (3-adrenoceptor function in field stimulated left
atria of the rat was determined. To this end field stimulated atria were incu6at~d with 10 M,
100 M and 1 mM HNE for 25 min. After removing the excess of HNE,(3-adrenoceptor
function was determined by measuring the posirive inotropic response to (-)-isoproterenol. It
was found that 10 M HNE had no effect on (3-adrenoceptor funcrion, whereas 100 M HNE
reduced the maximal effect to (-)-isoproterenol without affecring the pD2(-log ECgp). At these
concentrations, HNE had no effect on either (3-adrenoceptor density or on c-AMP producrion.
After 1 mM HNE, the atria stopped contracting. Since the effects of the synthetic thiol
inacrivator N-ethyl maleimide were similar to those of HNE,it was concluded that the reducrion
of R-adrenoceptor function by HNE is probably the result of alkylation of free sulfhydryl
groups. Our results indicate that the reduction of adenylate cyclase activity by oxidative stress is
not mediated by the producrion of HNE, however oxidative stress and HNE both reduce the
maximal response to (3-adrenoceptor srimularion.
Introduction.
The process of lipid peroxidarion is a result of oxidative stress, a disbalance between the
producrion of free radicals and the protection against free radicals [1]. This process has been
associated with several patho-physiological phenomena varying from aging to cancer. In the
heart free radicals are becoming to be recognized as mediators of damage provoked by
hyperoxia, adriamycin or ischemia. During lipid peroxidarion polyunsaturated membrane lipids
are converted into lipid hydroperoxides, which subsequently lead to the formarion of reacrive
aldehydes.
There are several mechanisms by which the deleterious effects of lipid peroxidarion can be
brought about. A direct attack of free radicals on DNA or vital enzymes, and alteration of
membrane fluidity due to the peroxidarion of the polyunsaturated membrane lipids attribute to
the injury inflicted by lipid peroxidation. Additionally, it has been demonstrated that the reactive
aldehydes that are formed upon lipid peroxidation contribute to the tissue damage by lipid
peroxidarion. In this respect 4-hydroxy-2,3-trans-nonenal (HNE)is thought to be of major
importance, because it is produced in relative large amounts and because it is highly reacrive.
The effects of HNE appear to be dependent on the reaction of HNE with proteins [2,3]. The
aldehyde binds as a Michael acceptor compound to cysteine residues of enzymes and this
sequestrarion frequently leads to an impairment of the catalytic activity of the enzymes.
Recently it has been demonstrated that during ischemia ~i-adrenoceptor funcrion is reduced
167

[4,5]. We have shown that oxidative stress induces a comparable reduction of (3-adrenoceptor
funcrion as ischemia [6,7]. Since it is well lrnown that sulfhydryl groups play a essenrial role in
(3-adrenoceptor function [3,8] and that during oxidative stress sulfhydryl reactive substances
like HNE are formed, we investigated whether HNE might be responsible for the effects of
oxidative stress on ~i-adrenoceptor function in an isolated left atrium.
Material and methods.
(-)-Isoproterenol hydrochloride and 5'-guanylylimidodiphosphate(GppNHp) were purchased hom Sigma(S~
Louis, USA). 4-Hydroxy-2,3-trans-nonenal (HNE) was synthesized according to Leurs et al. [3]. All other
chemicals were ofreagent grade.
Male Wistar rats (200-250 g, C.P.B. Harlan Olec, Zeist, The Netherlands) were killed by decapitation. T'he
hearts were rapidly excised. The isolated left atria were mounted in water jacketed organ baths, thermostatted at
37C, containing a Krebs buffer gassed with a mixture of 95% OZ and 5% CO2; pH = 7.4. The composition of
the Krebs buffer was (mIvn: NaCI (117.5), KCl (5.6), MgSO4 (1.18), CaC12 (2.5), NaH2PO4 (1.28), NaHCO3
(25) and glucose (5.5). The left atria were stimulated with a Grass field stimulator at a frequency of 5 Hz,
stimulus strength was 1.5 times the threshold voltage. Resting tension was adjusted to 0.5 g. After an
equilibrium period of 1 hour, the dose dependent contractions induced by (-)-isoproterenol were isometricly
recorded. Subsequently the atria were incubated with 10 M, 100 M or 1 mM HNE for 25 min. After
thoroughly washing, again a dose response curve of (-)-soproterenol was constructed. The effects of (-)isoproterenol are expressed as the pD2(-log of agonist that produces half maximal effect) and the maximal
response after FATE treatment was compared to the maximal response before HNE ireatrnent. Receptor density and
adenylate cyclase activity were assessed as described previously [6].

Results.
The effect of HNE on ~3-adrenoceptor function in field stimulated left atria of the rat was
determined. To this end the atria were incubated with various concentrarions HNE for a period

0-5 M HNE

W
U

a
0
w
z
o
F
U
C

a
F
z
0
U

10-3 M HNE

10'3 M NEM
5 m n

TIME

Fig. 1. Typical effects of HNE and NEM on the contraction force of field stimulated left atria. At the
time-point indicated by the arrow 10 M,100 M,1 mM HNE or 1 mM NEM was added.

168

Concentration HNE
control
10 M
100 M
1 mM

pD2

Max. Effect(%)

8.62 0.13
8.62 0.16
8.49 0.06
N.D. *

100
106 9
47 9~
0*

Different from control value (P c 0.05).


*~ Not determined since the atria stopped contracting.

Table I. The effect of 4-hydroxy-2,3-trans-nonenal(HNE)on the positive inotropic response of rat left
atria induced by (-)-isoproterenol. Results are expressed as mean SD,n = 4.

of 25 min. Addition of 10 M HNE had no effect on the basal contraction of the atria. At a
concentration of 100 M,HNE increased baseline tension after approximately 20 minuts.
Addition of HNE in a concentration of 1 mM resulted in a transient increase, followed by a
decrease in the contraction force. Moreover, the baseline tension rose. Approximately 15
minutes after the addition of 1 mM HNE, the atria did not react to the electrical stimularion
anymore. In figure 1, typical examples of the effects of the various concentrarions HNE are
shown.
In order to determine the effect of HNE on j3-adrenoceptor function, a dose response curve
for (-)-isoproterenol was constructed after ~~VE treatment and subsequent washing of the atria.
This curve was compared to a dose response curve for (-)-isoproterenol determined in the
atrium before HNE treatment. Addition of(-)-isoproterenol to field stimulated left atria resulted
in a dose dependent increase of the contraction force of the atria upon stimulation of the
~i-adrenoceptor. The pD2(-log of agonist that produces half maximal effectof(-)-isoproterenol
in the atria before HNE treatment was 8.62 0.13. It was found that HNE had no effect on the
pD2 of (-)-isoproterenol (table I). Incubation with 10 M HNE had also no effect on the
maximal response induced by (-)-isoproterenol. However, after 100 M HNE the maximal
response was reduced to 47 9 % of the control value. After treatment with 1 mM HNE no
contracrion of the left atria could be generated anymore and therefore no dose response curve for
(-)-isoproterenol could be constructed..
In order to deternune which component of the R-adrenoceptor complex was responsible for

Concentrarion HNE

~3-adrenoceptor density (fmol/mg protein)

control
10M
1(}0 M
1mM

383 15
41023
398 6
21415~

* Different from control value (P < 0.05).

Table II. The effect of 4-hydroxy-2,3-trans-nonenal(HNE) on (3-adrenoceptor density in rat left atria.
Results are expressed as mean SD,n = 3.

169

Concentrarion HNE

control
10M
100M

c-AMP producrion (pmol/mg/min)


basal

10 mM NaF

5 2
63
52

45 12
4914
437

100 M GppNHp
397
414
347

Results aze expressed as mean SD,n = 3.

Table III.
The effect of4-hydroxy-2,3-trans-nonenal(HNE)on c-AMP production in rat left atria.
the reduction of(3-adrenoceptor function by HNE, we deternuned (3-adrenoceptor
density and
c-AMP producrion in membranes prepared from either control or HNE treated
atria. It was
found that neither 10 M nor 100 M HNE had an effect on ~i-adrenoceptor density,
whereas 1
mM HNE reduced (3-adrenoceptor density to 50 % (table In. Moreover, neither
10 M nor
100 M HNE had any effect on c-AMP production, either basal or induced by NaF
or 5'gyanylylimidodiphosphate (GppNHp)(table III). Based on these observations
it is concluded
that the reduction of the ma}cimal response to (-)-isoproterenol by HNE is not mediate
d by
inactivation of the ~i-adrenoceptor protein, the coupling protein or adenylate cyclase.
The mechanism by which HNE is supposed to exert its toxic effects is covalen binding
t
of
HNE of free thiol groups of proteins. To invesrigate whether this process is also involve
d in the
reduction of ~i-adrenoceptor function, we compared the effects of the synthetic thiol
inactivator
N-ethyl maleimide(NEM)to those of HNE.In addirion, the effects of the disulfide
reductant,
dithiothreitol, were deternuned. It was found that the direct effects of NEM on
the field
stimulated left atria were similar to those of HNE (fig. 1). Also the effects of NEM
on
~3-adrenoceptor funcrion were compazable to those of HNE. NEM treatmen did
t
not result in a
pD2 shift for (-)-isoproterenol, but it reduced the mximal response to (-)-isoprotereno
l in a
dose dependent manner (table IV). In contrast to the effects of both HNE
and NEM,
dithiothreitol reduced the pD2 of(-)-isoproterenol without affecting the maxunal
response (table
IV).

pD2
control
Concentration NEM
10 M
100 M
Concentrarion DTT
1 mM
3mM

Max. Effect(%)

8.50 0.15

100

8.53 0.06
8.50 0.21

102 4
39 13*

7.86 0.15*
7.360.15*

109 9
101 5

* Different from control value (P < 0.05). Results aze expressed as mean t SD, n =
3.

Table IV. The effect of N-ethyl maleimide (NEIv~ and dithiothreitol (DTT) on the positive inotropic
response of rat left atria induced by (-)-isoproterenol.
170

Discussion.
The peroxidation of polyunsaturated membrane lipids induces a wide array of cellular
dysfunctions. It has been suggested that lipid peroxidation is involved in the cazdiac injury
provoked by adriamycin, catecholamines and ischemia. To the diverse deleterious effects of
lipid peroxidation, HNE is supposed to contribute a substantial part. HNE inactivates enzymes
via sequestration of essential sulfhydryl groups located on these macromolecules [2,3].
It has recently been demonstrated that (3-adrenoceptor density is not affected [4] or becomes
increased [5] during myocardial ischemia, while adenylate cyclase activity is decreased [4,5].
Previously we have reporte Thai oxidative stress has a biphasic effect on (3-adrenoceptor
density in heart membranes [6]. At a moderate oxidative stress (3-adrenoceptor density was
increased, whereas a more pronounced oxidative stress reduced (3-adrenoceptor density. In
contrast to (3-adrenoceptor density, adenylate cyclase was inactivated already at a relative
moderate oxidative stress [6]. In the isolated left atrium c-AMP production was reduced by
oxidative stress, while ~3-adrenoceptor density was not affected [7]. Moreover, the maximal
inotropic response to (3-adrenoceptor stimulation was reduced by oxidative stress, an effect that
could not be attributed to the impaired adenylate cyclase acrivity [7]. Since it is known that
ischemia induces oxidative stress [9] and the effects of ischemia and oxidative stress on
~3-adrenoceptor function are comparable, it is tempting to suggest that the effects ofischemia on
(3-adrenoceptor function are mediated by oxidarive stress.
It has well been described that sulfhydryl groups play a pivotal role in (3-adrenoceptor
funcrion [8]. Adenylate cyclase activity was shown to be more susceptible to sulfhydryl reacrive
compounds than (3-adrenoceptor density [8]. In addirion, it has been reported that in lung
membranes HNE inactivates ~3-adrenoceptors. At a concentration of 1 mM, HNE reduced
~i-adrenoceptor density to approximately 50%[3]. Adenylate cyclase acrivity is more sensirive
to HNE than (3-adrenoceptor density. Both basal and glucagon stimulated c-AMP production
were inhibited in plasma liver membranes by HNE,even at micromolar concentrarions, whereas
fluoride stimulated activity was increased by HNE [10]. The present study was designed to
determine the effect of HNE on (3-adrenoceptor function in an intact organ, i.e. the isolated left
atrium in order to evaluate the possible contribution of HNE to the effects of oxidarive stress on
(3-adrenoceptor function.
It was found that the addition of HNE to field stimulated left atria resulted in a reduction of
the contraction force, followed by a raise of the basal tension. At a concentration of 1 mM,
HNE finally caused a complete contraction failure. Similar effects were obtained with
4-hydroxy-2,3-trans-pentenal on field stimulated rat diaphragm [11]. It is also interesting to
note that the effects of HNE are comparable to the effects of oxygen radicals generated by
xanthine oxidase /xanthine in an isolated heart [12]. Possibly these effects of the oxygen
radicals are mediated by the production of HNE, however, oxygen radicals might also react
directly with sulfhydryl groups.
Moreover, in the present study we found that HNE induced a done-dependent reducrion of
~3-adrenoceptor function in the isolated left atrium. At a concentrarion of 100 M,HNE reduced
the maximal response of the isolated atrium to 50 % of the control value, however, at that
concentration HNE had no effect on ~3-adrenoceptor density nor on c-AMP formation.
Apparently none of the proteins involved in c-AMP production - i.e. (3-adrenoceptor protein,
N-protein and adenylate cyclase -are affected by 100 M HNE in the isolated left atrium. The
171

observed reduced reactivity of HNE toward adenylate cyclase and ~3-adrenoceptor, compared to
membrane prepazarions [3,10],can be explained by the potent cytosolic defense system against
HNE that is still present in an intact isolated left atrium, whereas this is absent in membrane
prepararions. We have reported that in isolated left atria oxidative stress reduces c-AMP
formarion while ~3-adrenoceptor density is not reduced [7]. Apparently, the impairment of
adenylate cyclase activity by oxidarive stress is not induced by the formation of HNE. In
addition we found that both oxidative stress [7] and HNE reduce the maximal response to
~3-adrenoceptor stimulation in the rat heart, suggesring that this effect of oxidative stress is
mediated by SINE.
The fact that HNE did not reduce the pD2for (-)-isoproterenol already indicates that HNE
has no effect on (3-adrenoceptor density nor on c-AMP formarion. For the positive inotropic
response to (-)-isoproterenol in the left atrium "spare receptors" exist [13]. This phenomenon is
also referred to by the term "receptor reserve". It means that only part of the receptors have to be
occupied by (-)-isoproterenol in order to induce maximal response. This is caused by saturarion
of a reaction in the cascade of biochemical events leading from receptor occupation to the
eventual response at subm~imal receptor occupation [13]. Reduction of(3-adrenoceptor density
would primarily result in an increase of the concentrarion of(-)-isoproterenol needed to induce
maximal effect, and not in a reduction of the maximal effect. The same as for (3-adrenoceptor
density holds true for c-AMP production, since c-AMP production is situated in the cascade
before the saturable reaction. Apparently HNE reacts with a component that is situated after the
saturable reaction. Recently, it has been shown that protein kinases, enzymes that are also
involved in (3-adrenoceptor function, are vulnerable toward HNE [14]. Possibly the inactivation
of these enzymes by HNE is responsible for the reduction of the maximal effect.
At fust sight, the concentrations of HNE needed to produce an effect seem rather high.
However, in evaluating the effects of HNE not only the concentration of HNE has to he
considered, but also the localization of the target enzymes relarive to the site of HNE formation.
Since HNE originates from the peroxidation of membrane lipids, concentrations of this
aldehyde in the membrane can reach 4 - 10 mM in vivo [15]. Moreover, because lipid
peroxidation might be restricted to limited membrane sites, local concentrations can be even
higher [16]. In experiments with externally added HNE the in vivo situation cannot be
mimicked. A remarkable difference between in situ generated HNE and externally added HNE
on the efficiency of the inactivation of membrane bound enzymes has been reported [16].
Because the enzymes involved in (3-adrenoceptor function are all located in or near the
membrane, the reduction of ~3-adrenoceptor function by 10 M to 1 mM HNE observed in the
present study is probably of physiological relevance.
Since the effects of the synthetic thiol inacrivator N-ethyl maleimide are similar to those of
HNE,the effects of HNE re most likely mediated by sequestration of sulfhydryl groups. It is
also interesting to note the effect of dithiothreitol. This disulfide reducing agent decreased the
pD2 of (-)-isoproterenol without affecting th maximal response. Although the pD2 shift
suggests that the action of dithiothreitol is the result of a comperitive receptor interaction, the
effect of dithiothreitol is probably non-competirive. It has well been described that dithiothreitol
treatment reduces (3-adrenoceptor density [17], and in this way dithiothreitol reduces the
receptor reserve for (-)-isoproterenol. The opposite effects of HNE and DTT on the
responsiveness to (-)-isoproterenol also indicate that the effects of HNE are not mediated by a
reduction of(3-adrenoceptor density.
172

In conclusion, our results demonstrate that HNE, in a concentration up to 100 M, has no


effect on either (3-adrenoceptor density or c-AMP production. This indicates that the reported
reduction of adenylate cyclase activity by oxidarive stress is not mediated by the formation of
HNE. Nevertheless, this study demonstrates that HNE reduces the maximal response to
(3-adrenoceptor stimularion, an effect that is also observed after oxidarive stress, indicaring that
this effect of oxidative stress might be mediated by the production of HNE. This reducrion
probably proceeds via alkylation of an essential sulfhydryl group of one or more of the enzymes
involved in the ~i-adrenoceptor mediated response other than (3-adrenoceptor, N-protein or
adenylate cyclase.
References
1 H. Sies Ed., Oxidative Stress, Academic Press, New York/L.ondon (1985)
2 G.R.M.M. Naenen, J.N.L. Tai Tin Tsoi, N.P.E. Vermeulen, H. Timmerman and A. Bast, Arch. Biochem.
Biophys. 259 449-456 (1987)
3 R. Leurs, B. Rademaker, K. Kramer, H. Timmerman and A. Bast, Chem.-Biol. Interact 59 211-218(1986)
4 M.Freissmuth, W.Schutz, M. Weindlmayer-Gifttel, M.Zimpfer and C.K. Spins, J. Cardiovasc. Pharmacol.
10 568-574 (1987)
5 D.E. Vatner, D.R. Knight, Y.T. Shen, J.X. Thomas, C.J. Homcy and S.F. Vatner, J. Mol. Cell. Cardiol.
20 75-82(1988)
6 G.R.M.M. Naenen, P. van Dansfik, N.P.E. Vermeulen, H. Timmerman and A. Bast, Free Rad. Res.
Commun. 4 243-249 (1988)
7 G.R.M.M. Naenen, N.P.E. Vermeulen, H.Timmerman and A. Bast, in Medical, Biochemical and Chemical
Aspects of Free Radicals, Elsevier, Amsterdam,in press (1989)
8 7.M. Stadel and R.J. Letkowitz, Mol. Pharmacol. 16 709-718 (1979)
9 P.S. Rao, M.V. Cohen and H.S. Mueller, J. Mol. Cell. Cardiol. 15 713-716 (1983)
10 L. Paradisi, C. Panagini, M.Payola, G. Barrera and M.U. Dianzini, Chem.-Biol. Interact. 53 209-217(1985)
11 W.Burkl, H. Klingenberg, E. Schauenstein and M.Taufer, Wien. Klin. Wochenschr.80 238-240(1968)
12 M. Gupta and P.K. Singal, Biochem. Pharmacol. 36 3774-3777(1987)
13 T.P. Kenskin,Pharmacological Analysis of Drug-Receptor Interaction,Raven Press, New York(1987)
14 G. Poli, E. Albano, M.U. Dianzani, E. Melloni, S. Pontremoli, U.M. Marinari, M.A. Pronzato and D.
Cottalasso, Biochem. Biophys. Res. Comm. 153 591-597 (1988)
15 A. Benedetti, M. Comporti, R. Fulceri and H. Esterbauer, Biochim. Biophys. Ac[a 792 172-181 (1984)
16 J.F. Koster, R.G. Slee, A. Montfoort, J. Lang and H. Esterbauer, Free Rad. Res. Commun. 1 273-287
(1986)
17 G. Vauquelin, S. Bottari, L. Kanazek and A.D. Strosberg, J. Biol. Chem. 254 4462-4469 (1979)

173

174

Summary
Oxidative stress is involved in the etiology of various patho-physiological processes.
Biomembranes appear to be among of the most important targets for oxidative stress, since
oxidative stress is often accompanied by lipid peroxidation. During the process of lipid
peroxidation poly-unsaturated membrane lipids are perolcidized in a radical reacrion. As a result
of this process the physical state of the membrane is altered, and sulfhydryl reactive compounds
(like 4-hydroxy-2,3-trans-nonenal) are formed (chapter 2).
In this thesis, the interplay between thiols and oxidarive stress is further explored. In part I
the effect of endogenous and exogenous administered thiols on oxidative stress is examined.
Part II deals with the modulation of oxidarive stress by catecholamines. In part III, the effect of
oxidative stress on (3-adrenoceptor function is determined.
In part I of this thesis the protective effect of the thiol glutathione against in vitro
microsomal lipid peroxidation is described. During the process of lipid peroxidation, lipid
hydropero~des are produced. It is often suggested that the protecrive effect of glutathione is due
to the detoxication of these lipid hydropero~des, a process catalyzed by one of the glutathioneperoxidases possibly in cooperation with phospholipase A2. As shown in chapters 3,4 and 5,
the protective effect of glutathione is probably not mediated by one of the glutathioneperoxidases. Also phospholipase A2 appeared not to be involved in the glutathione-dependent
protection (chapter 5,6). Alternarively, we speculated that the protection by glutathione is due to
the regeneration of vitamin E that has been oxidized during the scavenging of radicals. The
regeneration of vitamin E by glutathione would be mediated via amembrane-bound free radical
reductase, inirially called a heat labile factor (chapters 4, 5). It was demonstrated that this free
radical reductase contains an essential thiol group itself(chapter 5).
In clinical pracrice, several low molecular weight thiols are used to protect against oxidative
stress. In this thesis it was shown that these thiols do not protect directly against lipid
peroxidation via the free radical reductase (chapter 7). In addition, it was found that the
endogenous dithiol dihydrolipoate may indirectly protect against lipid peroxidation via the free
radical reductase by the reduction of o~dized glutathione to reduced gliitathione (chapter 8).
In chapter 9, the mechanism of the reaction of ebselen with endogenous thiols was
described. Ebselen is a seleno-organic compound that mimics the endogenous glutathioneperoxidases. It was found that probably a selenol intermediate is formed during the peroxidase
reaction of ebselen. Moreover, dihydrolipoate proved to be a better cofactor than glutathione in
the pero~dase acrivity of ebselen.
In part II of this thesis, the direct modulation of oxidative stress by catecholamines was
determined'(chapters 10, 11 and 12). It was found that catecholamines are able to protect against
lipid peroxidation, probably by scavenging radicals (chapter 10). However, during this
scavenging reacrive products are formed. These reactive products are able to react with thiols
located on several enzymes, e.g. the membrane-bound free radical reductase, the calciumATPase and the microsomal glutathione S-transferace. As a result of this reaction, the activity of
the enzymes is altered (chapters 10, 11).
In part III of this thesis, the effect of oxidative stress on ~i-adrenoceptor function in the heart
was deternuned (chapters 14, 15 and 16). It was found that oxidarive stress reduced the efficacy
175

of(3-adrenoceptor agonists and the maximal ~i-adrenoceptor-mediated response (chapter 15).


The reduction of the efficacy is probably due to an impaired coupling between the receptor and
adenylate cyclase, which might be caused by an inacrivation of the coupling protein or by a
reduction of membrane fluidity (chapters 14, 15). The reduction of the maximal effect might be
the result of the production of thiol reactive products (like 4-hydroxy-2,3-trans-nonenal) during
o~darive stress (chapter 16). It is suggested that the reduction of(3-adrenoceptor function in the
heart by oxidative stress is a protecrive feed-back mechanism that limits the damage provoked
by excessive stimularion of the ~3-adrenoceptor.

S11T1e11V att111g

Radicalen zijn moleculen met een ongepaard electron. Vanwege dit ongepaarde electron zijn
deze moleculen vaak erg reacrief. De radicalen kunnen tal van celbestandelen beschadigen, zoals
DNA,eiwitten, suikers en vetten. De belangrijkste radicalen die in aerobe organismen worden
gevormd zijn zuurstofradicalen. Gelukkig zijn deze organismen goed beschermd tegen zuurstofradicalen. Onder sommige omstandigheden schiet deze bescherming echter te kort. Deze
toestand wordt aangeduid met de term "oxidatieve stress".
Oxidatieve stress ligt ten grondslag aan tal van patho-fysiologische processen, zoals kanker,
onstekingsreakries, hartziekten en veroudering. Biomembranen blijken n van de belangrijkste
aangrijpingspunten van oxidatieve stress te zijn, doordat oxidatieve stress vaak gepaard gaat met
membraanschade, te weten "lipide penoxidatie". Tijdens het proces van lipide peroxidarie
worden (meervoudig) onverzadigde vetzuren van de membraan omgezet in vetzuurhydroperoxiden door een radicaalreactie. Als gevolg van dit proces verandert de fysische
conditie van de membraan, en worden thiol reacrive stoffen (zoals 4-hydroxy-2,3-transnonenal) gevormd (hoofdstuk 2).
Zwavel is een essenrieel element in levende organismen. Zwavel is voor het merendeel
aanwezig als thiol of als disulfide. Thiol en disulfide groepen zijn essenrieel voor de werking
van tal van eiwitten (enzymen), zoals de eiwitten die betrokken zijn bij de werking van de (3adrenoceptor op het hart. Er bestaat een wisselwerking tussen thiolen en oxidarieve stress.
Enerzijds zijn thiolen belangrijk bij de beschernung tegen oxidatieve stress, b.v. het thiol
glutathion (7-glu-cys-gly) beschermt tegen lipide peroxidarie. Anderzijds is schade aan thiol
groepen n van de belangrijkste mechanismen waardoor schade door oxidarieve stress tot stand
komt.
In dit proefschrift wordt de relatie tussen thiolen en oxidatieve stress nader beschreven. Het
weergegeven onderzoek handelt over de bescherming van thiolen tegen lipide penoxidatie (deel
I), de modularie van oxidatieve stress door catecholaminen (deel In en het effect van oxidatieve
stress op (3-adrenoceptor functie (deel III).
In aan deel I ten grondslag liggend onderzoek werd de beschermende werking van het thiol
glutathion tegen in vitro lipide penoxidatie bestudeerd. Tijdens het proces van lipide peroxidarie
worden vetzuur hydroperoxiden gevormd. Veelal wordt aangenomen dat de beschermende
werking van glutathione het gevolg is van het onschadelijk maken van deze vetzuur
hydroperoxiden, een proces waarbij glutathion-penoxidasen en phospholipase AZ betrokken
zijn. Echter, uit hoofdstuk 3, 4 en 5 blijkt dat de beschermende werking van glutathion
176

waarschijnlijk niet wordt gemedieerd door n van de glutathion-peroxidasen. Tevens blijkt


phospholipase AZ niet te zijn betrokken bij de glutathion-afhankelijke bescherming (hoofdstuk
5, 6). Als alternatieve verklaring geven we in dit proefschrift dat de beschermende werking van
glutathion het gevolg is van een samenwerlQng tussen glutathion en vitamine E. Vitamine E
beschermt tegen lipide peroxidatie door radicalen onschadelijk te maken. Tijdens deze
bescherming wordt vitamine E verbruikt. Glutathion zou in staat zijn om vitamine E, dat is
verbruikt tijdens het onschadelijk maken van radicalen, terug om te zetten in het oorspronkelijke
vitamine E. Hierdoor potenrieert glutathion de werking van vitamine E,en vormt de combinatie
van glutathion en vitamine E een meer efficiente bescherming tegen lipide peroxidatie. De
regeneratie van vitamine E door glutathion wordt gemedieerd door een enzym, namelijk een
vrije radicaal reductase, wat in eerste instanrie een hitte labiele factor werd genoemd (hoofdstuk
4, 5). Het vrije radicaal reductase bevat zelf ook een essenrile thiol groep (hoofdstuk 5).
In de kliniek worden verschillende thiolen toegepast om te beschermen tegen oxidatieve
stress. Wij veronderstelden dat de werking van deze stoffen wellicht zou kunnen worden
veroorzaakt door eenzelfde samenwerking met vitamine E als beschreven voor glutathion. Het
bleek echter dat deze thiolen niet direct beschermen tegen lipide peroxidatie via het vrije radicaal
reductase (hoofdstuk 7). Daarnaast werd gevonden dat het endogene dithiol dihydrolipoaat wel
indirect via het vrije radical reductase kan beschermen tegen lipide peroxidarie, namelijk door de
reducrie van geoxideerd glutathion tot gereduceerd glutathion (hoofdstuk 8).
In hoofdstuk 9 wordt het onderzoek naar het mechanisme van de reactie van ebseleen met
endogene thiolen beschreven. Ebseleen is een organische selenium verbinding die de werking
van endogene glutathion-peroxidasen nabootst. Het bleek dat een selenol intermediair wordt
gevormd tijdens de peroxidase activiteit van ebseleen. Bovendien bleek dihydrolipoaat een
betere cofactor dan glutathion te zijn in de eeroxidase activiteit van ebseleen.
Tijdens sommige vormen van oxidatieve stress komen relatief grote hoeveelheden
catecholaminen vrij. Algemeen wordt aangenomen dat deze catecholaminen bijdragen tot de
schade die wordt veroorzaakt door oxidatieve stress. Hierbij kan onderscheid worden gemaakt
tussen een directe bijdrage van catecholaminen en een bijdrage die wordt veroorzaakt door
overmatige stimulatie van receptoren, zoals de (3-adrenoceptor in het hart.
In deel II van dit proefschrift beschreven ondersoek werd de directe modulatie van oxidatieve
stress door catecholaminen bepaald (hoofdstuk 10, 11 en 12). Tijdens dit onderzoek bleek dat
catecholaminen in staat zijn om te beschermen tegen lipide eeroxidatie, waarschijnlijk door
radicalen weg te vangen (hoofdstuk 10). Echter, tijdens het onschadelijk maken van deze
radicalen worden reactieve producten gevormd. Deze reacrive producten zijn in staat om met
thiolen van enzymen te reageren, b.v. met de thiolen op het vrije radicaal reductase, het calcium
ATPase en het microsomale glutathion S-transferase. Als gevolg van deze reactie verandert de
acriviteit van deze enzymen (hoofdstuk 10, 11).
In deel III van het proefschrift wordt het effect van oxidatieve stress op de werking van de
(3-adrenoceptor in het hart behandeld (hoofdstuk 14, 15 en 16). Tijdens het onderzoek werd
gevonden dat door oxidatieve stress de werking van het (3-adrenoceptor complex afneemt; de
mate waarin (3-adrenerge agonisten een effect veroorzaken (= de efficacy van (3-adrenerge
agonisten) en het maximaal(3-adrenoceptor-gemedieerde effect vernunderen (hoofdstuk-15). De
gevonden effecten waren niet terug te voeren op een directe inacrivatie van de receptor zelf door
oxidarieve stress. De eiwitten, die betrokken zijn bij het doorgeven van het signaal wat
177

geproduceerd door receptor activatie (= stimulus), bleken wel te


worden aangetast. De
vernundering van de efficacy is waarschijnlijk het gevolg van een vernund
erde koppeling tussen
de receptor en adenylaat cyclase, de eerste stap in de srimulus transductie.
Dit zou kunnen zijn
veroorzaakt door inacrivatie van het koppelingseiwit, of door een afname van
de vloeibaarheid
van de membraan waarin deze stimulus overdracht plaatsvindt(hoofds
tuk 14, 15). De reductie
van het meimaal effect is mogelijk het gevolg van de productie van sulfhyd
ryl reactive stoffen
(zoals 4-hydroxy-2,3-trans-nonenal) tijdens oxidatieve stress. Deze sulfhydryl
reactieve stoffen
zouden een eiwit dat is betrokken bij het doorgeven van de stimulus inactiv
eren, blijkbaar
doordat dit eiwit een essenrile thiol group bevat (hoofdstuk 16). In dit
proefschrift wordt
gesuggereerd dat de afname van de werking van de ~3-adrenoceptor in het
hart als gevolg van
oxidatieve stress een beschermend terugkoppelingsmechanisme is,
wat de schade die
veroorzaakt kan worden door overmatige stimulatie van de ~3-adrenerge recepto
r beperkt.

178

Curriculum vitae
Guido Haenen was born in 1959 on June 17, at Axel, The Netherlands. He graduated from
high school(VWO)in 1977. In the same year he entered the Faculty of Pharmacy of the State
University Utrecht, where he acquired his B.Sc. in Pharmacy in 1980. After studying
Pharmacy with as principle subject Pharmacology, he received his M.Sc. in Pharmacy in 1984
with the judicium cum laude. During this period he joined the department of Pharmacology for 7
months as a research assistent. In 1985 he received his qualification to pracrice Pharmacy.
From June 195 he worked as a scientific research assistant at the Department of
Pharmacochemistry of the Free University at Amsterdam, where he carried out the
investigarions described in this thesis. At the present time he is research associate at the
Department ofPharmacochemistry.

179

bist of publications
1. G.R.M.M. Haenen and A. Bast.
Protection against microsomal lipid peroxidation by a glutathione dependent heat labile
factor.
Fedn. Eur. Biochem. Soc. Lett. (1983) 159, 24 - 28.
2. G.R.M.M. Haenen,P. van Dansfik, N.P.E. Vermeulen, H. Timmerman and A. Bast.
The effect of hydrogen Heroxide nn (3-adrenoceptor funcrion in the heart.
J. Free Rad. Res. Comm.(1988) 4, 243 - 249.
3. G.R.M.M. Haenen, J.N.L. Tai Tin Tsoi, N.P.E. Vermeulen, H. Timmerman and A. Bast.
4-Hydroxy-2,3-trans-nonenal stimulates lipid peroxidation by reducing the glutathione
dependent protection.
Arch. Biochem. Biophys. (1987) 259, 449 - 456.
4. G.R.M.M. Haenen, J.N.L. Tai Tin Tsoi, H.M.N. Ragetli, N.P.E. Vermeulen, H.
Timmerman and A. Bast.
Activation of the microsomal glutathiofie-S-transferace and reduction of the glutathionedependent protecrion against lipid peroxidation by acrolein.
Biochem. Pharmacol.(1988) 37, 1933 - 1938.
5. G.R.M.M. Haenen, N.P.E. Vermeulen, H. Timmerman and A. Bast.
Is phospholipase A2involved in the glutathione dependent protecrion against in vitro lipid
pero~dation?
in: O. Hayaishi, E. Niki, M. Kondo and T. Yoshikawa (Eds.) Medical, biomedical and
chemical aspects of free radicals, Elsevier, Amsterdam (1989) 1291- 1294.
6. G.R.M.M. Haenen, N.P.E. Vermeulen, H. Timmerman and A. Bast.
Reducrion of(3-adrenoceptor function by oxidative stress.
in: O. Hayaishi, E. Niki, M. Kondo and T. Yoshikawa (Eds.) Medical, biomedical and
chemical aspects of free radicals, Elsevier, Amsterdam (1989) 571 - 574.
7. G.R.M.M. Haenen, N.P.E. Vermeulen, H, Timmerman and A. Bast.
Effect of thiols on microsomal lipid peroxidarion.
Chem. Bol. Interact., in press.
8. G.R.M.M. Haenen, H. Plug, N.P.E. Vermeulen, H. Timmerman and A. Bast.
Contribution of 4-hydroxy-2,3-trans-nonenal to the reduction of beta-adrenoceptor function
by oxidative stress.
Life Sci.(1989) 45, 71 - 76.
9. G.R.M.M. Haenen and A. Bast.
Interplay between dihydrolipoate, glutathion, vitamin E and vitamin C in the protection
against o~dative stress.
in: Neue biochemische, pharmakologische and klinische Erkenntnisse zur Thiocts~ure, in
press.
10.G.R.M.M. Haenen, B.M. de Rooij, H. Timmerman and A. Bast.
(3-Adrenoceptor agonists do not reduce hydrogen peroxide production from superoxide
radicals.
Arch. Int. Pharmacodyn. Ther., in press.
180

11.G.R.M.M. Haenen and A. Bast.


Lipid peroxida.tion; modulation by endogenous and exogenous antioxidants.
in: M.A. Boogaerts (Ed) proceedings third benelux workshop on free radicals in biology
and medicine, in press.
12.G.R.M.M. Haenen, B.M. de Rooij, N.P.E. Vermeulen and A. Bast.
Mechanism of the reaction of ebselen with endogenous thiols.
Mol. Pharmacol., in press.
13.A. Bast and G.R.M.M. Haenen.
Glutathione and Cytochroom P-450: What is the significance of their interrelationship in
Lipid pero~dation?
Trends Biochem. Sci.(1984) 9, 510 - 514.
14.R.H.M. Julicher, L. Sterrenberg, G.R.M.M. Haenen, A. Bast and J. Noordhoek.
Sex differences in the cellular defence system against free radicals from oxygen or drug
metabolites in rat.
Arch. Toxicol.(1984) 56, 83 - 86.
15.A. Bast, G.R.M.M. Haenen and E.M. Savenije-Chapel.
Inhibition of rat microsomal lipid peroxidation by sodium 2-mercaptosulfonate (mesna)
proceeds via glutathione.
Arzneim. Forsch.(1987) 37, 1043 - 1045.
16.R.H.M. Julicher, L. Sterrenberg, G.R.M.M. Haenen, A. Bast and J. Noordhoek.
The effect of chronic adriamycin treatment on heart, kidney and liver tissue of male and
female rat: on the role of oxidative stress in adriamycin toxicity.
Arch. Toxicol.(1988) 61, 275 - 281.
17..R. Goeptar, G.R.M.M. Haenen, H. Timmerman and A. Bast.
The effects of beta-adrenergic receptor agonists on the H2O2 formation in alveolar
macrophage suspensions are not mediated by beta-adrenoceptors.
Agents Actions (1988)25, 375 - 377
18.A. Bast and G.R.M.M. Haenen.
Interplay between lipoic acid and glutathion in the protection against microsomal lipid
peroxidation.
Biochim. Biophys. Acta (1988) 963, 558 - 561.
19.A. Bast and G.R.M.M. Haenen.
Regulation of lipid peroxidation by glutathione and lipoic acid: Involvement of liver
microsomal vitamin E free radical reductase.
in: I. Emerit, L. Packer and C. Auclair(Eds) Antioxidant therapy and preventive medicine,
Plenum Press, in press.
20.A. Bast and G.R.M.M. Haenen.
Cytochrome P-450 and vitamin E free radical reductase: Formation of and protection aginst
free radicals.
in: A. Castes de Paulet, L. Douste Blazy and R. Paoletti (Eds) Free radicals, lipoproteins
and membrane lipids, Plenum Press, in press.
21.F.B. Pruijn, G.R.M.M. Haenen and A. Bast.
Interplay between vitamin E, glutathione and dihydrolipoic acid in protection against lipid
peroxidation.
Fat, in press.
181

~.. ~ ~J
Dit proefschrift zou niet tot stand zijn gekomen zonder de hulp van velen.
Vandaar dat ik
graag iedereen wil bedanken die heeft meegeholpen.
Allereerst bedank ik de driekoppige, hooggeleerde leiding van de vakgroe
p farmacochemie,
prof. dr. A. Bast; prof. dr. H. Timmerman en prof. dr. N.P.E. Vermeulen
voor de deskundige
leiding die ik tijdens mijn promotie periode heb ontvangen. Speciale dank
ben ik verschuldigd
aan beide promotoren, prof. dr. Aalt Bast en prof. dr. Nico Vermeu
len. Ik hoop dat het
enthousiasme van hun begeleiding enigszins terug te vinden is in mijn
proefschrift.
Ook de hoofd- en bijvak studenten die elk op hun eigen wijze hebben meegeho
lpen, dank ik
voor hun bijdragen: Paul van Dansfik, Lennie Huberts, Hans Plug,
Kees Vermunt, Henri
"Hank" Ragetli, Jacintha Tai Tin Tsoi, Frank Jansen, Ben de Rooij,
Sandra Ciere, Harold
Bastiaans en (niet te vergeten) Meta Veerman. Alle andere leden
van de vakgroep
farmacochemie bedank ik voor de prettige contacten en samenwerlng tijdens
mijn promotie.
Daarnaast ben ik erkentelijk voor de hulp die ik heb mogen ontvangen
van pro dr. S. Balt
("derde lezer"), dr. Frans de Kanter en dr. Ben van Baar (vakgroep organis
che en anorganische
chemie), de technische diensten en de teken- en fotoafdeling.
Tenslotte wil ik ook hen bedanken die buiten de directe sfeer van
het onderzoek bij de
totstandkoming van dit proefschrift waren betrokken.

182