Professional Documents
Culture Documents
FOR
SOME IMPLICATIONS
~~~ CATECHOLAMINE TOXICITY
HAEN
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
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.
Aan Myriam
Contents
Chapter 1
Introduction.
PartI
Chapter 2
in: M.A. Boogaerts (Ed.)Proceedings third Benelux workshop on free radicals in biology
and medicine,in press.
Chapter 3
29
Chapter 4
35
Chapter 5
41
Chapter 6
49
Chapter 7
53
Chapter 8
63
Chapter 9
69
Part II
91
93
Submitted.
105
Submitted.
117
Part III
123
125
145
151
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.
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
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
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
o .
o'
~ ~
~ I
C16H33
II
I
O
~
/
III
C16H33
C16H33
~
~
O
C16H33
~
~
IV
C16H33
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 +.
,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
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
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).
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
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
6
7
8
9
10
11
12
13
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
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
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
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
33
34
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
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.
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
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
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
~.:
~.:
~.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)
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
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
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
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.
.o
/.~.
~~.
~~~
a
~
~~ ~
-~ -o
/ ~
.S
M
,%~
//
t
V
~0
Time (min)
10
30
20
Time (min)
40
Si
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
~~
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
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
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+.
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
00
0
10
20
30
40
50
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
62
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
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
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
(()-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)
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
\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 ~ /
\ /
~
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
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
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
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
(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)
~~~
2d[E] 0
dt
g)
(9)
(1~)
kq CH2O2] [D]
k2[GSH]
3
Subsriturion of(8)-(11)in (5) gives:
T-
(12)
k 4[H2O2]
+ 2+
2 k4[H2O2]~-1
{T
ks
k 1[GS H] + k2[GSH]
k2[GS H]
}~13~
k
3
90
T
~ k1 [GSH] + k2[GSH] +k4[H2O2]+ k5 ~ {
k2[GSH]
k3
91
92
93
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~
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)
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
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
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
i25~
~~
102
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
~~
[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
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
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
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
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
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.
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
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
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
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
123
124
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.
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
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
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.
(-)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).
(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]
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
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
[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
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
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
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
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
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
552
584
1428
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.
~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
26 1
24 1
3.4 0.5
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)
8.50.2
100
11030
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
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
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.
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
8.00.1
100
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
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:~
[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*
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
control
10M
1(}0 M
1mM
383 15
41023
398 6
21415~
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
10 mM NaF
5 2
63
52
45 12
4914
437
100 M GppNHp
397
414
347
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
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
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
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
~.. ~ ~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