Superoxide anions and hyperoxia inactivate
endothelium-derived
relaxing factor
G. M. RUBANYI
AND P. M. VANHOUTTE
Department
of Physiology and Biophysics, Mayo Clinic and Mayo Foundation,
Rochester, Minnesota 55905
interference with the production, the transit or the action
on smooth muscle of endothelium-derived
relaxing facAm. J. Physiol. 250 (Heart Circ. Physiol. 19): HSZZ-H827, tor(s). The results demonstrate that under bioassay con1986.-Experiments were designedto determine the effects of ditions, when the endothelium-derived
relaxing factor is
oxygen-derived free radicals on the production and biological in prolonged contact with physiological
salt solution,
activity of endothelium-derived relaxing factor or factors resuperoxide
anions
accelerate
its destruction. As a conleasedby acetylcholine. Ringsof canine coronary arteries without endothelium (bioassayrings) were superfusedwith solution sequence, superoxide dismutase (a specific scavenger of
passing through a canine femoral artery with endothelium. superoxide anions; 11) considerably prolongs the biologrelaxing factor(s),
Superoxide dismutasecausedmaximal relaxation of the bioas- ical half-life of endothelium-derived
when the partial pressure of oxygen is lowsay ring when infused upstream, but not downstream, of the particularly
femoral artery; this effect of superoxidedismutasewasinhibited ered.
RUBANYI,
and hyperoxia
G. M.,
AND
inactivate
P. M. VANHOUTTE.Superoxide
endothelium-derived
relaxing
anions
factor.
METHODS
Experiments were performed on femoral and left circumflex coronary arteries taken from mongrel dogs of
either sex (18-28 kg), anesthetized with pentobarbital
sodium (30 mg/kg iv). The blood vessels were studied in
modified Krebs-Ringer bicarbonate solution (control solution) of the following millimolar
composition:
NaCl
118.3, KC1 4.7, CaC12 2.5, MgSO, 1.2, KH2P0,
1.2,
NaHC03 25.0, calcium disodium-EDTA
0.026, and glucose 11.1.
acetylcholine; bioassay;catalase;coronary artery; dog; femoral
The bioassay apparatus designed in earlier work was
artery; superoxidedismutase;vascular smooth muscle
used (15). Side branches of segments (3-3.5 cm long) of
the left and right femoral artery were tied. The endothelium was kept intact as much as possible. The segments
ENDOTHELIAL
CELLS release a diffusible relaxing subwere fixed to stainless steel cannulas (1.5 mm ID) and
stance (or substances) when exposed to acetylcholine (7, placed into an organ chamber maintained
at 37°C and
9, 15, 16). The chemical nature of the endotheliumfilled with 12 ml of aerated (95% 02-5s COZ) control
derived relaxing factor(s) is still unknown. Anoxia (3, 8) solution. The segments were perfused at constant flow
antioxidants
and nonspecific radical scavengers (9, 15) (2 ml/min)
by means of a multichannel
roller pump
inhibit endothelium-dependent
relaxations evoked by (Gilson, Minipuls 2) with control solution maintained at
acetylcholine, suggesting that oxidative mechanisms play
37°C and containing phentolamine
( 10B5M), propranolol
an important
role in the production or the release of (5 x 10B6 M), and indomethacin
(10B5M), to inhibit CYendothelium-derived
relaxing factor(s). Since oxidative
and ,&adrenoceptors and cyclooxygenase, respectively (1,
processes in endothelial
cells can generate oxygen-de17). A stainless steel tube was also placed in the organ
rived free radicals (13), early speculations suggested that
chamber through which control solution was pumped at
the relaxing mediator may be a free radical (6,8). Experthe same rate. A ring of coronary artery, in which the
iments using isolated blood vessels of the rabbit (9, 19) endothelium
had been removed (bioassay ring), was susand the dog (18) ruled out this possibility, since scaven- pended directly below the organ chamber by means of
gers of oxygen-derived free radicals do not eliminate the two stainless steel stirrups passed through its lumen.
endothelium-dependent
relaxations
to acetylcholine.
One stirrup was connected to an isometric force transHowever, generating or scavenging oxygen-derived free ducer (Grass FT03C). Changes in isometric tension were
radicals can either augment or inhibit endothelium-derecorded. The assembly of bioassay ring, stirrups, and
force transducer could be moved freely below the organ
pendent relaxations evoked by acetylcholine in canine
coronary arteries (18). This study was designed to deterchamber allowing the preparation to be superfused with
mine whether these effects of the scavengers are due to the perfusate from either of the femoral segments with
H822
0363-6135/86 $1.50 Copyright 0 1986 the American Physiological
Society
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.4 on June 14, 2017
by catalase.Infusion of acetykholine relaxed the bioassayrings
becauseit releaseda labile relaxing factor (or factors) from the
endothelium. When infused below the femoral artery, superoxide dismutaseand, to a lesserextent, catalaseaugmentedthe
relaxations to acetylcholine. Superoxide dismutase, but not
catalase,doubledthe half-life of the endothelium-derivedrelaxing factor(s). This protective effect of the enzyme was augmentedfivefold by lowering the oxygen content of the perfusate
from 95 to 10%. These data demonstrate that: 1) superoxide
anions inactivate the relaxing factor(s) releasedby acetylcholine from endothelial
cells and 2) hyperoxia favors the inactivation of endothelium-derived relaxing factor(s).
OXYGEN-DERIVED
FREE
S.
Before the actual experiment, the bioassay ring was
superfused directly with control solution for 60 min.
During this interval it was stretched in a stepwise manner until the basal tension reached -10 g, the optimal
tension for rings of isolated canine coronary arteries (1,
17). Drugs were infused by means of infusion pumps
(Harvard model 901) at a rate of 0.1 ml/min or less.
Drugs
The following pharmacological
agents were used: acetylcholine hydrochloride
(Sigma); catalase (from bovine
liver; 40,000 U/mg
protein;
Sigma);
indomethacin
(Sigma); meth y 1ene blue (Eastman-Kodak);
l-norepinephrine bitartrate (Sigma); phenidone (Sigma); phentolamine mesylate (Ciba-Geigy); dl-propranolol
(Sigma);
prostaglandin
Fzn (Sigma); superoxide dismutase (from
dog blood; 3,000 U/mg protein; Sigma). All concentrations are expressed as final concentrations
(M or U/ml)
in the superfusate.
Calculations
and Statistics
The biological half-life of the relaxing substance(s)
released by acetylcholine from the endothelium
of perfused femoral arteries was calculated by relating the
decrease in tension in the bioassay ring (expressed as
percent) to the transit time imposed (9, 15). Unless
otherwise noted, each experimental group consisted of at
least six blood vessels taken from different dogs. The
data are shown as means t standard error of the mean
(SEM). Statistical analysis was performed using Student’s t test for paired and unpaired observations. Differences were considered to be statistically
significant
when P < 0.05.
RESULTS
All experiments were performed during contractions
of the bioassay rings caused by prostaglandin
FZCY(4 x
10m6 M) in the presence of indomethacin
(10m5 M).
Site 1
These experiments were performed with a transit time
of 1 s between the femoral artery and the bioassay ring.
AND
EDRF
H823
All drugs were infused upstream of the femoral artery
(Fig. 1).
During direct superfusion of the bioassay ring, the
addition of superoxide dismutase (150 U/ml) to the perfusate caused a significant increase in tension (+20.2 t
5.4%). The same concentration
of the enzyme caused
relaxation (-98.3 t 5.0%; ~2 = 5) during endothelial
superfusion (Fig. 2, upper tracing).
Catalase (860 U/ml) caused comparable, significant
increases in tension during direct or endothelial superfusion (average increase in tension: +10.5 t 2.5 and
+12.6 t 7.2%, respectively; n = 4) (Fig. 2, lower tracing).
During endothelial superfusion, acetylcholine ( 10B6 M)
caused complete relaxation of the prostaglandin
Fzainduced contraction. Catalase pretreatment
(860 U/ml;
10 min) significantly
and reversibly reduced the relaxations evoked by acetylcholine (to -31.7 t 9.0%; n = 5)
and by superoxide dismutase during endothelial superfusion (to -15.2 t 4.3%, n = 3).
Site 2
In this series, acetylcholine (lo-” M) was infused upstream of the femoral artery, while all other drugs were
given downstream from it. The transit time between the
femoral artery and the bioassay ring was varied (Fig. 1).
Superoxide dismutase and catalase. When given at site
2 both superoxide dismutase and catalase caused small
increases in tension (Fig. 3) that were not significantly
affected by varying the transit time.
Acetylcholine. TRANSIT TIME, 8 s. Acetylcholine caused
smaller relaxations (-51.2 t 3.6%) than with a transit
time of 1 s. The response to acetylcholine was augmented
by superoxide dismutase (150 U/ml) when it was added
prior to or during acetylcholine infusion, and, to a lesser
extent, by catalase (860 U/ml) (Fig. 3, upper).
TRANSIT TIME, 30 S. Acetylcholine
caused an increase
in tension that was not significantly affected by catalase;
in the presence of superoxide dismutase (150 U/ml)
acetylcholine caused relaxation (Fig. 3, lower).
In the presence of acetylcholine, but not in its absence,
superoxide dismutase caused dose-dependent, reversible
relaxations (Fig. 4). These relaxations were reversed by
phenidone ( 10B5 M), norepinephrine
( 10d6 M), and methylene blue (low5 M), but not by catalase (860 U/ml) (Fig.
5) .
HALF-LIFE.
Progressive lengthening
(from 1 to 8, 30,
60, or 120 s) of the transit time between the femoral
artery with endothelium
and the bioassay ring caused a
time-dependent
reduction and eventually a reversal of
the relaxations induced by acetylcholine
(Fig. 6); the
half-life of the response to acetylcholine was 8.1 t 1.2 s
(n = 6). Superoxide dismutase (150 U/ml) potentiated
the response to acetylcholine at 8 and 30 s; the enzyme
significantly augmented the half-life (to 15.7 2 1.5 s; P
< 0.05; n = 5) (Fig. 6). Catalase (860 U/ml) did not
significantly
affect the half-life of the relaxations induced by acetylcholine (9.2 t 1.6 s; n = 3).
LOW OXYGEN CONTENT. When the gas mixture aerating the endothelial perfusate was changed from 95% 025% CO2 to 85% N2-10% 02-5% COZ, the contractions of
the bioassay rings to prostaglandin
Fza were augmented
(Fig. 7) and with a transit time of 30 s, acetylcholine
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.4 on June 14, 2017
endothelium
(endothelial
superfusion) or the stainless
steel tube (direct superfusion) (Fig. 1).
In some experiments a stainless steel tube was fixed
to the outflow cannula of the femoral artery with endothelium. With this setup, drugs could be infused separately into each perfusion line either above (site I; allowing contact with the inner surface of the femoral artery)
or below the perfused femoral artery segment (sites 2 and
3). Infusion of drugs at site 2 avoided contact with the
endothelium
of the perfused vessel segment but allowed
interaction with released vasoactive substances in transit. Infusion of drugs at site 3 (-0.5 s transit time)
affected only the superfused bioassay tissue (15). To
increase the transit time, polyethylene tubes of various
lengths were placed into a heat exchanger (37°C). By
connecting the appropriate tube to the outlet of the
femoral artery and positioning the bioassay ring beneath
the corresponding outlet from the heat exchanger, the
transit time could be increased from 1 to 8,30,60, or 120
RADICALS
H824
OXYGEN-DERIVED
FREE RADICALS
AND EDRF
Multi-channel
of artery with endothelium
th
Diffusion
EDRF
Action
muscle cell
Coronary artery ring
without endothelium
For;e
transducer
Endothelium
Direct
SOD i
PG;pa
Catalase
-a-
;
SOD
Catalase
.,/\
I
-i
PG&
FIG. 2. Example of changes in tension in bioassay ring. Effect of
superoxide dismutase (SOD, 150 U/ml; upper tracing) and catalase
(860 U/ml; lower tracing) on contractions to prostaglandin Fz, (PGFs,;
4 x 10e6 M) of a canine coronary artery without endothelium during
direct or endothelial superfusion. Rings relax after switching from
direct to endothelial superfusion due to basal release of relaxing substance(s) from endothelium (16). Superoxide dismutase and catalase
were infused at site 1 (see Fig. 1) during period indicated by horizontal
bars.
caused a further increase in tension. Superoxide dismutase caused relaxations that decreased as the transit time
increased (30 s: -29.0 + 9.2%, n = 4; 60 s: -10.7 f 1.2%,
n = 3); it reversed the contraction evoked by acetylcholine to a relaxation (Fig. 7). The augmentation,
by the
enzyme of the relaxations induced by acetylcholine at
transit time of 30 s was significantly larger, and started
at lower concentrations
in the presence of 10 than in
95% oxygen (Fig. 4). The half-life was not affected by
the lower oxygen content alone (7.6 f 1.8 s; n = 4), but
the combination of low oxygen content plus superoxide
dismutase prolonged it to 81 + 5 s (P < 0.05; n = 4) (Fig.
6).
Site 3. Superoxide dismutase (150 U/ml) and catalase
(860 U/ml) given immediately upstream of the bioassay
ring caused increases in tension, which were similar in
the absence and presence of acetylcholine (10e6 M, infused at site 1) (Table 1). The contractions evoked by
both enzymes were comparable to those that they caused
during direct superfusion (Figs. 2 and 4) or when injected
at site 2 during endothelial superfusion in the absence of
acetylcholine (Figs. 3 and 5).
DISCUSSION
The present study confirms earlier observations that
acetylcholine relaxes vascular smooth muscle cells by
releasing a labile relaxing factor (or factors) from endothelial cells (2, 3, 6-9, 15, 16). It confirms that free
oxygen radicals generated during univalent reduction of
molecular oxygen are not the endothelium-derived
relaxing factor(s) released by acetylcholine (9,X3,19). Indeed,
in the bioassay-apparatus,
infusion
of superoxide
dismutase and catalase downstream of the source of
endothelial factors augmented rather than abolished the
relaxations of the bioassay rings induced by acetylcholine.
In agreement with observations made on isolated rings
of canine coronary arteries (18), the present experiments
suggest that hydrogen peroxide facilitates the release of
endothelium-derived
relaxing factor(s). This conclusion
is based on the finding that superoxide dismutase evokes
catalase-sensitive endothelium-dependent
relaxation in
the organ chamber (18) and the release of endotheliumderived relaxing factor(s) in the bioassay apparatus. The
data in the bioassay apparatus are consistent with a
continuous generation of small amounts of superoxide
anion, which does not trigger the release of endotheliumderived relaxing factor, except after its accelerated transformation to hydrogen peroxide by superoxide dismutase
(11, 18). The facilitatory role of hydrogen peroxide on
the release of endothelium-derived
relaxing factor(s) by
acetylcholine is demonstrated further by the inhibitory
effect that catalase has an acetylcholine-induced
relax-
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.4 on June 14, 2017
I
FIG. 1. Left: schematic illustration
of 3 phases
(production, diffusion, and action) of endotheliumdependent inhibitory responses to acetylcholine (ACh)
in blood vessels. EDRF, endothelium-derived relaxing
factor. Right: bioassay apparatus used for separate
analysis of 3 phases (see left) of endothelium-dependent relaxations to ACh. Perfused segments of femoral
artery with endothelium serve as “donor” of endothelium-derived relaxing factor, and a ring of canine
coronary artery without endothelium serves as bioassay tissue. In this system, released endothelium-derived relaxing factor is transported to smooth muscle
cells of bioassay ring by perfusate (transit). (Modified
from Ref. 16).
OXYGEN-DERIVED
ACh,
Tram
8sec.
1 O’%vl,
site
FREE RADICALS
ACh
1
ACh
ACh
it time,
Catalase,
ACh,lO’%l,
Transit
3Osec.
H825
AND EDRF
site2
SOD,
SOD,
site2
ACh
site1
site2
ACh
time,
b
I
4min
/
Catalase,
site2
SOD,
site2
FIG. 3. Isometric tension recordings during endothelial superfusion of canine coronary arteries without endothelium
contracted by prostaglandin Fza (PGFZ,, 4 X lo-” M). Effects of superoxide dismutase (SOD, 150 U/ml, infused at site
2) and catalase (860 U/ml, infused at site 2) on relaxations induced by acetylcholine (ACh, 10s6 M, infused at site I).
Transit time, 8 s: for 6 experiments relaxations induced by ACh averaged -51.2 k 3.6, -99.4 2 1.2, and -65.2 k 3.1%
in control solution, in presence of SOD dismutase and in that of catalase, respectively. Increases in tension caused by
SOD and catalase averaged +9.2 k 2.3 and +10.5 2 3.1%, respectively (100% = 7.2 k 0.6 g). Transit time, 30 s: ACh
caused contraction (+7.9 k 2.3%; n = 11) in control solution (left), which was not affected by catalase (middle) but
reversed to relaxation (-47.3 k 7.2%; n = 5) by SOD (100% = 6.9 2 0.6 g).
% 02 + ACh
Soperoxide
Dismutase,
U/ml
FIG. 4. Effect of increasing concentrations of superoxide dismutase
(infused at site 2) in presence of acetylcholine [ (ACh; 10e6 M) infused
at site I] during endothelial superfusion of canine coronary arteries
without endothelium contracted with prostaglandin Fga (4 x 10B6 M).
Experiments were performed with a transit time of 30 s and at 2 levels
of oxygenation (10 and 95% 02). For comparison, effects of enzyme
during direct superfusion are also shown. Open triangles, direct superfusion + 95% 02; open squares, direct superfusion + 10% 02; fiZkd
circles, endothelial superfusion + ACh ( 10m6M) + 95% 02; filled squares,
endothelial superfusion + ACh (10m6 M) + 10% 02. Data shown as
means k SE (n = 4) and expressed as % of initial contraction to
prostaglandin Fl,, (100% = 7.1 k 0.5 and 8.4 -t 0.5 g in 95 and 10% 02,
respectively). * Difference between responses obtained at 95 and 10%
O2 is statistically significant (P < 0.05).
ation when infused at site 1 in the bioassay apparatus;
this must be due to an action of the enzyme involving
the production of relaxing factor(s) by the endothelial
cells, since catalase infused at site 2 augments rather
than inhibits the relaxation evoked by acetylcholi ne.
A major finding of the present study is that superoxide
dismutase, infused downstream from the site of release
of endothelium-derived
factor(s), potentiates its effect
and prolongs its half-life. One possibility
is that the
enzyme alters the vasodilator properties of substance(s)
released by endothelial
cells. However, the reversal by
phenidone [inhibitor of the endothelial production of the
relaxing factor(s)], norepinephrine
[inactivator
of the
factor(s) in transit], and methylene blue [inhibitor of the
action of the factor(s) on smooth muscle] of acetylcholine-induced
relaxations both in the absence (8, 9, 15,
16) and in the presence of superoxide dismutase (present
study) indicates that the known characteristics of the
released relaxing factor are not changed by the enzyme.
Superoxide dismutase did not augment the action of the
endothelium-derived
factor on the coronary smooth muscle, to judge from experiments where it was infused at
site 3. Thus the most logical explanation for the potentiating effect of superoxide dismutase is that it prevents
inactivation
of the relaxing factor(s) in transit. The
enzyme may protect endothelium-derived
relaxing factor(s), either by decreasing the concentration
of superoxide anion or by augmenting that of hydrogen peroxide
(11, 12). The latter explanation is unlikely because catalase, infused at site 2, augments rather than inhibits
acetylcholine-induced
responses and does not reverse the
facilitatory action of superoxide dismutase. The lack of
effect of catalase per se on the half-life of endotheliumderived relaxing factor(s) implies that oxidizing free radicals (hydrogen peroxide, free hydroxyl radical) are not
involved in the protection exerted by superoxide dismutase. Accelerated inactivation
of endothelium-derived
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.4 on June 14, 2017
T
OXYGEN-DERIVED
H826
SOD,
I
I
I
150
U/ml,
FREE RADICALS
AND EDRF
site2
ACh,lO%,
+25
1.0
Norepinephrine,
site2
r
I
4min
39
t
;1
Transit
1 Ow6M,
Catalase,
SOD
site2
I(-[PGF2a
QS% 02+
r
60
120
FIG. 6. Effect of increasing transit times on relaxation induced by
acetylcholine (10B6 M, infused at site I) during endothelial superfusion
of canine coronary arteries (without endothelium) contracted with
prostaglandin Fza (4 x 10m6M) at 2 levels of 02 concentration (10 and
95% 02) in superfusate. Experiments were performed in absence and
presence of superoxide dismutase (SOD, 150 U/ml) infused at site 2.
Data shown as means k SE (numbers represent number of observations) and expressed as % of initial contraction to prostaglandin F,.
* Effect of SOD is statistically significant.
site2
95% 02 1
1
PGF2Q
FIG. 5. Isometric tension recording during endothelial superfusion
of canine coronary arteries without endothelium contracted by prostaglandin Fg,(PGFz,, 4 x 10m6M), with a transit time of 8 s. Superoxide
dismutase (SOD, 150 U/ml) was infused at site 2, and acetylcholine
(ACh, 10e6 M) at site 1. Phenidone (10B5 M), norepinephrine (10m6M),
and methylene blue ( 10B5 M) but not catalase (860 U/ml) (all infused
at site 2) reversed inhibitory effect of ACh. In 3 experiments, reversal
averaged 105 k 8.2, 92 + 6.5, and 101 -+ 4.5% for phenidone, norepinephrine, and methylene blue, respectively. Effect of these 3 substances
was statistically significant (P C 0.05).
factor(s) by the superoxide anion would explain the
augmented half-life due to superoxide dismutase. Superoxide anions can reduce oxidized substances (12). Other
reducing (antioxidant)
agents, including cysteine, dithiothreitol, phenidone, and norepinephrine
inactivate endothelium-derived
relaxing factor(s) in transit (9, 15).
Thus it is likely that the inactivation
of endotheliumderived relaxing factor is due to the reducing properties
of the superoxide anion, which can be released from
endothelial cells after its enzymatic intracellular
production (13), or can be generated in salt solutions by photolysis (4). This effect of superoxide anion probably
becomes evident only with prolonged transit time, since
under control conditions superoxide dismutase does not
augment the activity of the endothelium-derived
relaxing
factor(s) at transit times of I (this study) or 4 (9) s. In
rings with endothelium
studied in organ chambers, the
protecting effect of superoxide dismutase is not obvious
(18, 19) because of the presumably short diffusion time
between the endothelial
cells and the smooth muscle in
30
time,sec.
10% 02
SOD.150
pGF2~~y~
T
2QI I
Y
ACh,
10m6M,
U/ml.
site2
,
i
1
I
ACh
sits
1
-
4min
PGF2a
7. Isometric tension recording during endothelial superfusion
(with a transit time of 30 s) of a canine coronary artery (without
endothelium) contracted with prostaglandin FPa (PGFg,, 4 x lo* M).
Switching gas mixture aerating superfusate from 95% O&I% COz (95%
02) to 85% &lo%
02-5% CO2 (10% 02) caused an increase in tension
(mean increase 18.5 * 3.1%; n = 7); such increases in tension were not
observed during direct superfusion (not shown). Acetylcholine (ACh,
IOm6M, infused at site I) caused contraction in absence (upper) but
relaxation in presence (lower) of superoxide dismutase (SOD, 150 U/
ml, given at site 2). Note that in presence of low O2 SOD causes
relaxation.
FIG.
the intact blood vessel wall.
If superoxide anions play a role in the inactivation
of
endothelium-derived
relaxing factor(s) in transit, the
protective action of superoxide dismutase should depend
on the extent of the generation of the free radical (which
is the function of oxygen tension) and on the rate of its
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.4 on June 14, 2017
10w5M,
39
02
95%
site1
OXYGEN-DERIVED
FREE
This
Institute
1. Direct effects of superoxide dismutase
and catalase on coronary arterial smooth muscle
in bioassay system
TABLE
Prostaglandin
4x 10*M
Control
Superoxide
dismutase,
150 U/ml
Catalase,
860 U/ml
Fza,
Prostaglandin
Acetylcholine,
Received
Fza plus
lo4 M
7.5k1.5
+0.69*0.10
3.9kl.l
+0.58&O.
12
+0.79*0.18
+0x&o.
15
Values are means & SE of 4 experiments.
Results
are expressed
in
absolute
changes
in tension
(g); + = augmentation.
Catalase
and
superoxide
dismutase
were infused
at site 3 (see Fig. 1). Prostaglandin
FzO and prostaglandin
FZP plus acetylcholine
were infused
at site 1 (see
Fig. 1).
The
authors
and (Janet
thank
Helen Hendrickson
Beckman
for secretarial
for preparing
assistance.
the
illustra-
AND
work is supported
research
Grants
11 June
H827
EDRF
in part by National
HL-31183,
HL-31547,
1985; accepted
in final
form
Heart,
Lung, and Blood
and HL-34634.
4 December
1985.
REFERENCES
1. COHEN,
R. A., J. T. SHEPHERD,
AND P. M. VANHOUTTE,
Prejunctional and postjunctional
actions of endogenous
norepinephrine
at
the sympathetic
neuroeffector
junction
in canine coronary
arteries.
Circ. Res. 52: 16-25, 1983.
2. DE MEY, J. G., M. CLAEYS, AND P. M. VANHOUTTE.
Endotheliumdependent
inhibitory
effects
of acetylcholine,
adenosine
triphosphate, thrombin
and arachidonic
acid in the canine femoral artery.
J. Pharmacol. Exp. Ther. 222: 166-173, 1982.
3. DE MEY, J. G., AND P. M. VANHOUTTE.
Anoxia
and endothelium
dependent
reactivity
of the canine femoral artery. J. PhysioL. Land.
335: 65-74, 1983.
4. FREEMAN,
B. A., AND J. D. CRAPO. Biology of disease. Free radicals
and tissue injury.
Lab. Invest. 47: 412-426,
1982.
5. FRIDOVICH,
I. Quantitative
aspects of the production
of superoxide
anion radical by milk xanthine
oxidase.
J. Biol. Chem. 245: 40534057, 1970.
R. F, The role of endothelium
in the responses
of
6. FURCHGOTF,
vascular
smooth
muscle
to drugs. Annu. Rev. Pharmacol. Toxicol.
24: 175497,
1984.
7. FURCHGOTT,
R. F., AND J. V. ZAWADZKI.
The obligatory
role of
endothelial
cells in the relaxation
of arterial
smooth
muscle
by
acetylcholine.
Nature Land. 288: 373-376,
1980.
8. FURCHGOTT,
R. F., J. V. ZAWADZKI,
AND P. D. CHERRY.
Role of
endothelium
in the vasodilator
response
to acetylcholine.
In: Vasodilatation, edited by P. M. Vanhoutte
and I. Leusen.
New York:
Raven,
1981, p. 49-66.
9. GRIFFITH,
T. M., D. H. EDWARDS,
M. J. LEWIS,
A. C. NEWBY,
AND A. H. HENDERSON.
The nature
of endothelium-derived
vascular relaxant
factor. Nature Lond. 308: 645-647, 1984.
10. KONTOS,
H. A., E, P, WEI, C. W. CHRISTMAN,
J. E. LEVASSEUR,
J. T. POVLISHOCK,
AND E. F. ELLIS,
Free oxygen
radicals
in
cerebral
vascular
responses.
Physiologist 26: 156-169, 1983.
11. MCCORD,
J. M., AND I. FRIDOVICH.
Superoxide
dismutase.
An
enzymatic
function
for erythrocuprein
(hemocuprein).
J. BioL.
Chem. 244,6049-6055,1969.
12. MCCORD,
J. M., AND I, FRIDOVICH.
Production
of 0; in photolyzed
water
demonstrated
through
the use of superoxide
dismutase.
Photochem. Photobiol. 17: 115421,
1973.
13. ROSEN,
G. M,, AND B. A. FREEMAN.
Detection
of superoxide
generated
by endothelial
cells. Proc. Natl. Acad. Sci. USA 81: 72697273,1984.
14. ROSENBLUM,
W. I. Effects
of free radical generation
on mouse pial
arterioles:
probable
role of hydroxyl
radicals.
Am. J. Physiol. 245
(Heart Circ. Physiol. 14): H139-H142,
1983.
15. RUBANYI,
G, M., R. R. LORENZ,
AND P. M. VAN~OUTTE.
Bioasay
of endothelium-derived
relaxing
factor.
Inactivation
by catecholamines. Am. J. Physiol. 249 (Heart Circ. Physiol. 18): H95-H101,
1985 .
16. RUBANYI,
G. M,, AND P. M. VANHOUTTE.
Hypoxia
releases vasoconstrictor
substance(s)
from the coronary
endothelium.
J. Physiol.
Lmd.
364: 45-56, 1985.
G. M., AND P. M. VANHOUTTE.
Inhibitors
of prostaglan17. RUBANYI,
din synthesis
augment
@-adrenergic
responsiveness
in canine coronary arteries.
Circ. Res. 56: 117-125, 1985.
G. M., AND P. M. VANHOUTTE.
Oxygen-derived
free
18. RUBANYI,
radicals,
endothelium
and responsiveness
of vascular
smooth muscle. Am. J. Physiol. 250 (Heart Circ. PhysioZ. 19): H815-H821,1986.
19. SILIN, J. P., J. A. STRULOWITZ,
M. S. WOLIN,
AND F. L. BELLONI.
Absence
of a role for superoxide
anion,
hydrogen
peroxide
and
hydroxyl
radical
in endothelium-mediated
relaxation
of rabbit
aorta. Blood Vessels 22: 65-73,
1985.
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.4 on June 14, 2017
elimination
[which is determined by the concentration
of the scavenger (5)]. In fact, the protective action of the
enzyme was augmented by lowering the oxygen tension,
and it was concentration-dependent.
Reducing the oxygen content caused an increase in tension of the bioassay
ring only during endothelial
superfusion; this can be
attributed
to the release of vasconstrictor
substances
from the hypoxic endothelial
cells (3, 16). Although
reduction in oxygen content per se did not affect the
half-life of the endothelium-dependent
relaxing factor(s)
released by acetylcholine, it reversed the contraction of
the bioassay ring caused by superoxide dismutase, infused at site 2 in the absence of acetylcholine,
into
relaxation. This must reflect augmented protection by
superoxide dismutase, at lower oxygen contents, of endothelium-derived
factor released under basal conditions
(9, 15). Likewise, the marked prolongation
of the effect
of acetylcholine in the combined presence of lower oxygen content and superoxide dismutase can only be explained by reduced inactivation of endothelium-derived
relaxing factor. Since a reduction in oxygen content per
se did not affect the half-life of endothelium-derived
relaxing factor(s) and since, under control conditions,
maximally effective concentrations
of superoxide dismutase restored the acetylcholine-induced
relaxation
only by -50% at a transit time of 30 s, the inactivation
of the endothelium-derived
factor by the higher oxygen
content may not be due only to superoxide anions, which
should be scavenged completely by the dose of superoxide
dismutase used (S), but also to other oxygen-derived
reducing radicals, which are not (or less) sensitive to the
enzyme. One likely possibility is that the hyperoxic gas
mixture used generates the formation of 0: as occurs in
other biological systems (5). Thus the present data suggest that the short half-life of endothelium-derived
relaxing factor, observed in this and previous studies (9,
15) may be the consequence of the artificial environment
imposed.
tions
RADICALS