Carbon monoxide-mediated brain lipid peroxidation in the rat
STEPHEN R. THOM
Institute for Environmental Medicine, University of Pennsylvania Medical Center,
Philadelphia, Pennsylvania 19104
THOM, STEPHEN R. Carbon monoxide-mediated brain lipid
peroxidation in the rat. J. Appl. Physiol. 68(3): 9974003,
1990.-Clinical and animal data suggest hat the pathogenesis
of CO poisoning extends beyond the inhibition of hemoglobin
function, but no mechanism has been identified. Evidence of
neurological compromise, particularly loss of consciousness, has
been implicated as a marker for increased mortality and mor-
bidity in clinical reports. Experiments were carried out with
rats to assess whether CO exposure may cause brain lipid
peroxidation. With the use of two methods, measurement of
conjugated dienes and thiobarbituric acid reactivity, brain lipid
peroxidation could be documented as a result of exposure to
CO at a concentration sufficient to cause unconsciousness.
Products of lipid peroxidation were increased by 75% over the
base-line values 90 min after CO exposure. Unconsciousness
was associated with a brief period of hypotension, so brief that
in itself it caused no apparent insult. Lipid peroxidation oc-
curred only after the animals were returned to CO-free air, and
there was no direct correlation with the carboxyhemoglobin
level. This work may provide an explanation for a number of
currently poorly understood clinical observations regarding CO
poisoning.
CO poisoning; ischemia-reperfusion phenomenon
CARBON MONOXIDE is recognized as the leading cause of
death by poisoning in the United States, accounting for
approximately half of all fatal poisonings (19). Survivors
of poisoning by CO may experience acute or delayed
neurological sequelae. These sequelae occur in ~12% of
cases (3, 9, 28). Although death is generally the result of
tissue hypoxia, the biochemical mechanism for morbidity
is unknown. CO can impair O2 delivery to tissues by
binding to hemoglobin (7). Inhibition of intracellular
oxidoreductases has been identified; however, there does
not appear to be a marked alteration in 02 extraction or
cellular respiration even with exposure to rather high
concentrations of CO (4, 11, 31). This may be the result
of the relatively low affinity of these proteins for CO (2).
When functions other than respiration are assayed, there
is evidence of CO toxicity that cannot be attributed solely
to carboxyhemoglobin (HbCO) -mediated hypoxia (13,
25). Animal studies indicate that the cardiovascular al-
terations produced by CO poisoning are in part respon-
sible for the central nervous system (CNS) insult (10,
22), and similar insults can be produced when hypoxic
hypoxia is accompanied by an interval of ischemia (21).
The involvement of O2 free radicals and secondary lipid
peroxidation has been suggested (17).
An interval of unconsciousness is among the factors
associated with an increased risk for morbidity and mor-
tality in some clinical reports (3, 9, 28). We have used a
rat model of CO poisoning in which subtle CNS changes
were reported (27) and added a brief exposure to a level
of CO sufficient to cause a period of unconsciousness.
Our hypothesis was that CO poisoning could cause brain
lipid peroxidation. In addition, we assessed the role of
cardiovascular events associated with unconsciousness.
METHODS
Animal Manipulations
Animals. Male Wistar rats, 160-180 g, were purchased
from Simonsen Laboratories, Palo Alto, CA, and from
Charles River Laboratory, Wilmington, MA. They were
given food and water ad libitum.
Cardiovascular monitoring. Measurements were per-
formed on rats with aortic catheters made of PE-50
tubing. Animals were anesthetized with pentobarbital
(0.05 mg/g ip), and the aorta was visualized through an
abdominal incision. The saline-filled catheter was in-
serted, sutured in place, and then tunneled beneath the
skin to the dorsum of the neck. The catheter was flushed
with heparinized saline and capped with a stopcock. The
incision was closed, and the animals were allowed to
recover from surgery for at least 24 h before study. Blood
pressure was monitored by using a Gould Statham phys-
iological pressure transducer and a strip chart recorder.
Cardiac activity was monitored with electrode leads at-
tached by means of stainless steel sutures placed in the
skin.
Animal treatment. Rats were killed by decapitation. In
one series of experiments the brains were dissected free,
immersed in liquid nitrogen within 45 s of death, and
stored at -7OOC. Frozen brains were later weighed, placed
in 10% (wt/vol) cold 0.25 M sucrose-l mM EDTA at pH
7.0, and homogenized in a blender. Samples of the brain
homogenates were taken immediately for assays of prod-
ucts of lipid peroxidation. In a second series of experi-
ments brains were dissected free within 40-50 s, weighed,
and homogenized in cold sucrose-EDTA. Assays for
products of lipid peroxidation were then immediately
begun, which was 3 min after decapitation.
Gas exposure. Animals were placed in a steel chamber
with a 7-liter volume manufactured by Bethlehem Steel
Corporation (model G15-APSP). Throughout the expo-
sure to CO, air was flushed through the chamber at a
rate of -8 liters/min. For the first 40 min of exposure a
flow of pure CO was maintained to achieve a steady
concentration in the chamber of 1,000 ppm. The CO
0161-7567/90 $1.50 Copyright 0 1990 the American Physiological Society 997

998 CO BRAIN LIPID PEROXIDATION
concentration was then increased to 3,000 ppm by in-
creasing the flow rate. Animals were continuously ob-
served during CO exposure, and once unconsciousness
occurred, the CO flow was discontinued. In the standard
model the CO flow was reinstated once the chamber CO
concentration had decreased to 1,000 ppm, and this level
was maintained for the remain .der of a 3-h period. In a
modified procedure, rats were removed from the co
environment immediately after unconsciousness oc-
lipid dry weight was 6.31 t 1.08 (SD) mg/lOO mg brain.
The extracted lipid was redissolved in sufficient chioro-
form to achieve a concentration of 0.6 mg lipid/O.3 ml
chloroform, and 0.3 ml of the solution was combined
with 0.9 ml methanol for spectrophotometric analysis.
The procedure allowed for complete dissolution of the
lipid residue, and by diluting with methanol, the ultra-
violet light absorption of the blank solution was mini-
mized. This procedure gave more reproducible results in
our hands than resuspension in heptane or methanol,
which has been used by others (15, 26).
curred. The standard model and the modified procedure
are illustrated in Fig. 1. The CO level was monitored by
a pump and detector tube method (Enmet, Ann Arbor,
MI) during early experiments and later by infrared analy-
sis of gas samples with a Warren E. Collins analyzer
model 303.
Ultraviolet absorbance and difference spectra were
recorded by using a dual- beam spectroph .otometer (H i-
tachi, model 3210). To compare the samples in day-to-
day experiments absorbance readings at 233 nm were
taken with fused silica optical cells and a Perkin-Elmer
Lambda spectrophotometer. Absorbance readings from
the various animal groups were related by determining
the dry weight of the total tissue lipid in each sample.
For this analysis lipid was extracted as described above,
except that the chloroform phase was transferred to
aluminum weighing pans and evaporated in a vacuum
oven at 45OC.
Hemorrhagic hypotension. Transient hypotension was
precipitated in some rats by withdrawing blood through
an aortic catheter into a heparinized glass syringe. Blood
pressure was monitored during this process, and blood
withdrawal was stopped when the mean pressure reached
-60 mmHg. This process usually took up to 1 min. After
10 s of hypotension the blood was reinjected into the rat
and the catheter flushed with saline.
Malonaldehyde (MDA). Brain homogenate (0.2 ml) was
placed in 0.2 ml of 8.1% (wt/vol) sodium dodecyl sulfate.
The measurement of thiobarbituric acid (TBA) reactivity
was then carried out according to the method of Ohkawa
et al. (20) as a nonspecific but sensitive indication of the
formation of lipid peroxide. The absorbance readings
were standardized by measuring the protein content in
Analytic Procedures
Conjugated dienes. Immediately after the brain was
homogenized, a 0.8.ml sample (-80 mg brain tissue) was
added-to 3 ml of chloroform-methanol (1:2 vol/vol), and
total tissue lipid was extracted purified. The method
of Bligh and Dyer (1) was followed, except that the final
extraction was performed with 4% NaCl-15 mM EDTA,
as suggested by Kogure et al. (15). The chloroform phase
was transferred to another vial and evaporated under NP.
Among 32 rat brains processed in this fashion the average
each homogenate according to the method of Lowry et
al. (16). The average protein content measured in the 27
brains used in this assay was 10.7 t 1.8 (SD) mg/lOO mg
brain. MDA standard was prepared by using 1,1’,3,3’-
tetraethoxypropane.
Thromboxane &. Frozen brains were homogenized as
described above, except that 0.4 mM indomethacin was
added to inhibit cyclooxygenase activity during process-
ing. The homogenate was then immediately parceled into
aliquots of 0.8 ml, frozen in liquid N2, and stored over-
night at -7OOC. Homogenates were thawed under a
stream of NZ, 40 ~1 were removed for protein analysis,
and the remainder was processed after the method out-
lined by Powell (23). Ethanol (3 ml) was added to the
brain homogenate, vortexed, and left for 5 min. Water
(16.2 ml) was then added to make the solution 15% (vol/
vol) ethanol, and the suspension was centrifuged at 400
g for 10 min. The supernatant was acidified to pH 3.0
and passed through a Sep-Pak Cl8 cartridge (Waters
Associates, Milford, MA). Cartridge contents were eluted
first with 15% (vol/vol) aqueous ethanol (10 ml) and
then petroleum ether (10 ml). Methyl formate (5 ml) was
then passed through the cartridge, collected in polypro-
pylene tubes, and evaporated under a Ng stream. Control
studies l with radioactive thromboxane Bz demonstrated
that 90% of thromboxane added to homogenates could
be recovered by using this method. The amount of throm-
boxane B2 present in the precipitate from the methyl
formate fraction was assessed by using a thromboxane
B2 1251 radioimmunoassay kit (Amersham International,
Amersham, UK). Radioactivity in the sample was meas-
STANDARD MODEL
8 3000
2
L.O.C. = loss of
consciousness
I-
$2 2000
rE
i2
- 1000
- z (Air) 0 w 1.5 hours 1 :
d79 * 0.019 (a) O.lf3 * 0.046 (b)
n=6 n=12 A233/mg lipid * S.D.
MODIFIED MODEL
3000
2000
IO00
(Air)OJ 1 1.5 hours i 2 hours I ,
i
A233/mg 0.150 0.041 1c) 0.2!2 0.077 (d) lipid * *
* S.D. n=4 n=4
FIG. 1. Schematic diagram of CO exposure and air-breathing inter-
vals before animals were killed and lipid extraction in standard and
modified models. Procedures followed are as described in METHODS.
Superscripts a and b, values from Table 1; superscripts b-d, values
significantly greater than a (P < 0.001) but not significantly different
from each other.