Journal of the Neurological Sciences, 113 (1992) 187-197 ~ 1992 Elsevier Science Publishers B.V. All rights reserved 0022-5111X/92/$05.00
187
JNS 03890
Effects of dimethylthiourea on ischemic brain damage in hyperglycemic rats Johan Lundgren a,b, Maj-Lis Smith ~' and Bo K. S i e s j 6 a " Laboratoo' for Experirnental Brain Research, Department of Neurobioh#,~', Experimental Research Center, Unit~ersity of Lurid, Lurid, Sweden, and ~ Department of Paediatrics, Unicersity ttospital, Lund, Sweden (Received 2 December, 1091) (Revised, received 18 May, 1992) (Accepted 8 July, 1992)
Key words: Ischemia; Hyperglycemia; Brain damage; Dimethylthiourca; Seizures; Postischemic seizures; Rats
Summary Hyperglycemia is known to worsen the outcome of transient global or forebrain ischemia. The aggravating effect is believed to be mediated by the additional formation of lactate and of H +. Recent evidence suggests that reactive oxygen species contribute to the damage after brain ischemia. Since acidosis accelerates free radical damage in vitro, we decided to explore if ischemic damage in hyperglycemic subjects is ameliorated by dimethylthiourea (DMTU), an established free radical scavenger. In one series of hyperglycemic rats, we studied whether preischemic administration of DMTU alters the clinical outcome, notably the incidence and frequency of seizures. In two different series, the effect of DMTU on tissue damage was assessed by light microscopy after 15 h of recovery. Longer periods could not be studied since seizures developed. In the first of these serics the animals were anesthetized with isoflurane, and in the second with halothane. The latter anesthesia largly suppressed the "early" postischemic seizures, i.e. those occurring after 1-4 h. Dimethylthiourea treatment altered the clinical outcome after ischemia. Thus, the "late" postiscbemic seizures appeared milder and occurred significantly later than in untreated animals. The fatal outcome was also delayed since treated animals died after 35.5 _+ 8.2 h (mean _+ SD) of recirculation, as compared to 19.8 _+ 3.6 h of recirculation in control animals. However, all DMTU-treated (and control) animals died. In the first morphological series (isoflurane anesthesia) the histopathological analysis was complicated by the occurrence of prefixation seizures; such seizures were recognized in 4/16 animals. When these 4 animals were excluded from the analysis (2 treated and 2 control animals), DMTU pretreatment did not ameliorate the damage, except in the substantia nigra pars reticulata (P < 0.05). In the second series, comprising animals anesthetized with halothane, only one animal out of 16 had "early" seizures, and none showed "late" seizures before death. Among these animals DMTU treatment significantly ameliorated damage to caudoputamen and cingulate cortex (P < (1.01). We conclude that treatment with the free radical scavenger DMTU partly ameliorates ischemic brain damage associated with excessive acidosis, and marginally delays the development of post-ischemic seizures. However, the effects were moderate and could, at least in part, have been caused by nonspecific effects of DMTU. Furthermore, all DMTU-treated animals died. The results thus give little support to the notion that the aggravating effects of acidosis is due to enhancement of free radical production.
Introduction S u b s t a n t i a l evidence exists that the b r a i n level a n d / o r supply of glucose influence the o u t c o m e of
Correspondence to: Johan Lundgren, M.D., Ph.D., Laboratory for Experimental Brain Research, Department of Neurobiology, Experimental Research Center, University Hospital, S-221 85 Lund, Sweden. Tel (46-46) 173552; Fax (46-46) 151480.
c e r e b r a l ischemia. Thus, p r e i s c h e m i c h y p e r g l y c e m i a a g g r a v a t e s b r a i n d a m a g e ( M y e r s a n d Y a m a g u c h i 1977; Siemkowicz and H a n s e n 1978; K a l i m o et al. 1981; R e h n c r o n a et al. 1981; Pulsinelli et al. 1982a; Smith et al. 1988). It alters the d i s t r i b u t i o n and evolution o f b r a i n injury ( I n a m u r a et al. 1988; Smith et al. 1988), e n h a n c e s e d e m a f o r m a t i o n (Pulsinelli et al. 1982b; W a r n e r et al. 1987), and favours fatal p o s t i s c h e m i c seizures ( M y e r s 1979; Siemkowicz 1985; Smith et al.
188 1988; Lundgren et al. 1990a). The mechanisms by which elevated brain glucose levels exaggerate brain injury seem related to enhanced accumulation of lactate plus H ~, with a further decrease in intra- and extracellular p H (Ljunggren et al. 1974; Rehncrona et al. 1980, 1981; Welsh et al. 1980; Kraig et al. 1987; Chopp et al. 1988; Siesj6 1988; Kraig and Chesler 199(I). Reactive oxygen species (ROS) have been implicated as mediators of ischemic injury to many organs. including the brain (Halliwell 1985; Siesj6 et al. 19891. The best evidence is that pretreatment with a variety of free radical scavengers lessens the damage that develops following ischemia of different types (Abe et al. 1988; Hall et al. 1988: P a t t e t al. 1988; Beckman et al. 1989; Martz et al. 1989; Imaizumi et al. 1990: Martz el al. 1990). In a few cases, formation of ROS in the tissue has also been demonstrated in vivo ( P a t t e t a[. 1988; Bromont et al. 1989; Oliver et al. 1990; Kirsch et al. 1991). However, contradictory, results have been reported and, in general, we do not accurately know the route by which ROS are formed, the conditions under which they play important pathogenic roles, and the mechanisms by which they exert their harmful effects. It has been proposed that the participation of free radicals is likely when ischemia is of long duration, or if it is complicated by hyperglycemia or hyperthermia (Siesj6 et al. 1989; 1990). These are the conditions under which ischemia gives rise to pannecrosis rather than selective neuronal necrosis. Common to these conditions is that the cerebral microvessels arc involved in the damaging process. In contrast, short periods of ischemia under normothermic and normoglycemic conditions leave the microvessels in the brain essentially unaffected (Pulsinelli et al. 1982b; Smith ct al. 1984). Several facts suggest that free radicals are formed in cerebral microvessels. First, xanthine oxidase, which is one of the main generators of ROS (McCord 19851, seems localized to the microvessel fraction (Jarasch et al. 1986). Second, ischemia leads to activation of leucocytes and such activation inw)lves enhanced oxygen consumption and generation of ROS (Babior et al. 1987; Curnutte and Babior 1987). At least in the heart, reperfusion injury, is ameliorated by prior leukocyte depletion (Simpson el al. 1987; Breda ct al. 1989). Third, the edema which is formed after cerebral ischemic insults is at least in part reduced by free radical scavengers ( C h a r et al. 1987; Martz ct al. 1989; Betz 1990). Since excessive acidosis may increase the formation of ROS by at least two different mechanisms, it can be speculated that acidosis-related damage is mediated by free radicals (Siesj6 et al. 1985; Siesj6 1988). The first mechanism is related to release of prooxidant iron (or copper) and chelation to low molecular weight compounds. A decrease in pH favours iron release from intracellular ferritin (Harrison 1977), and from extra-
cellular transferrin (Lestas i%'(~: Aisen. 19791. When chelated by low molecular a.eight compounds, iron in its reduced form (Fe a+) catalyzes the formation ~[ highly reactive - O H through the Habcr-Weiss reaction (Halliwell 1990). The second mechanism rclatcs to the fact that protonation of -O; yiclds the more reactive and more lipid-mluble perhydroxyl radical (.HO~), and since this is an acid with a p K~, of 4.9 its formatiori is favoured by low pH (Gebicki and Bielski 19811. In vitro, lowering of pH in a brain homogenatc triggers enhanced production of frec radicals and increases lipid peroxidation (Barber and Bernheim 1:967; Siesj6 ct al. 1985: Rehncrona ct al. i9891, Using H20:~ production as a reflection of ~eactive oxygen formation (Patt el al. 19881 we failed ~,~ yetiS' such formation after a brief period of ischemia in normoglyccmic animals (Agardh et al. 199t)). tu lhc light oLt our basic hypothesis, this result was not unexpected. However, the results obtained following 15 rain of ischcmia in hyperglycemic animals were also negative, since we were unable to find an increase in H , O , production, or an increase in bleomycin-rcactive frec iron in (?$1: (l,undgren et al. 1991a). Furthermore, we could find no additional reduction in tissm: contents of glutathione or ~z-tocophcrol in hyperglycemic animals subjected to ischemia. It has recently become clear that DMTLi, an established . O t t scavenger, ameliorates ischemic brain injury tbllowing long periods of either transient or permanent ischemia ( P a t t c t al, i988; Martz et a ! ]q~89, 1990; Betz 1990). Since amelioration ol damage by an established free radical scavenger may be a sensitive indicator of such injury we explored the hypothesis that ischemic brain damage in hyperglycemic animals was ameliorated by DMTIJ.
Methods
Animals and experimental groups All experiments were performed on adult male Wistar rats of an S.P.F. strain (M011egaard's Breeding Center, Copenhagen). Two series of experiments were performed, both by the same person but with slightly different anesthesia. In the first series (protocol 1, isoflurane anesthesia), the effect of treatment with DMTU was studied with regard to clinical (Clin DMTU, lz--11) and morphological outcome (Mo D M T U I, n = 8), and compared to no treatment (Clin Control n - - 5 ; Mo Control 1, n = 8). In the second series (protocol 2, halotbane anesthesia), the effect of D M T U on the morphological outcome was investb gated (Mo D M T U 2, n .... 8" Mo Control 2, n = 8 1 . In the first series the weight of lhe animals was 290-360 g, and in the second it was 235-280 g, corresponding to lhe ages of about 91/ and 65 days, respectively. All
189 animals were housed in macrolon cages, and fasted overnight prior to the experiments, but allowed water ad libitum.
Preparation and dose of DMTU Two g D M T U (Janssen Chimica, Geel, Belgium) were freshly mixed in 20 ml 0.9% NaCI, yielding a concentration of 100 rag. ml J. The solution was used within 6 h. Each animal received 750 mg. kg -l (7.5 m l . k g -~) i.p. 45-60 min prior to the induction of ischemia. Control animals in the second series received saline i.p. (7.5 m l . k g I).
Operatit~e and experimental procedures The operative and experimental procedures for protocol 1 have been described earlier (Lundgren et al. 1990b). Anesthesia was induced with 3.5% isoflurane (Forene '~, Abbott Laboratories Ltd, England), and 70% N20 in Oz, after which the animal was intubated and connected to a respirator delivering 1.5-2.0% isoflurane. Then the drug was injected to animals intended for D M T U treatment. A central venous catheter was inserted through a neck incision via the right caval vein. A string was placed around each common carotid artery and a tail arterial catheter was inserted for blood pressure measurement, analysis of blood gases and pH (ABL 30, Radiometer A / S , Copenhagen, Denmark), and plasma glucose (Beckman Glucose Analyser 2, Beckman Instruments Inc. Fullerton, CA, USA). Needle electrodes for E E G recording were inserted bitemporally with the reference electrode in the tail, and temperature probes were placed subcutaneously on the skull bone and in the rectum. After the surgical procedures the isoflurane concentration in the delivered gas mixture was decreased to ~ 1.0%, and the animals were allowed a steady state period of 30 rain. During the first 5 min of the steady state period a glucose load of an i.v. solution containing 25% glucose (0.92 g. kg ~ body weight) was given, followed by a slower infusion (0.04g. kg J • rain - l ) of the same solution. Heparin was given (50 IE) prior to the first blood sampling. After the steady state period the glucose infusion and the isoflurane administration were discontinued, and the induction of ischemia was started with central venous exsanguination. When the blood pressure was below 50 mm Hg, bilateral carotid clamping was performed. Blood pressure was kept close to 50 mm Hg with additional withdrawal or reinfusion of shed blood, and ischemia was verified by absence of E E G activity. Ischemia was terminated by removal of the carotid damps, reinfusion of shed blood, and i.v. injection of 0.5 ml of a 0.6 M sodium bicarbonate solution. During the early recirculation period the blood pressure was continuously recorded, and PaO2, PaCO 2 and pH were controlled. When the animal showed strong breathing movements (after 20-30 rain
of recovery) it was extubated, and transferred to a macrolon cage. For the second series of experiments (protocol 2) the experimental conditions were altered so as to reduce the "early" seizures. For that purpose halothane (Halothane '~'~, ISC Chemicals, England) in similar concentrations was used instead of isoflurane for induction of anesthesia and during the surgical procedures. The concentration was decreased to and kept at 0.4-0.5% during the steady state and ischemic period, whereafter it was discontinued. During the steady state period, the muscle-relaxant vecuronium bromide (Norcuron ", Organon Teknika, Boxtel, Holland) was administered at a concentration of 1 rag" ml-~ mixed with the 25% glucose solution. The experimental procedure was approved by the Ethical Committee for Laboratory Animal Experiments at the University of Lund.
Histology Following 14-16 h of recovery the animals were perfusion-fixed with phosphate-buffered formaldehyde (Auer et al. 1984). After stablization for one day at 5°C the brains were dissected and stored in cold fixative. Later the brains were cut in coronal slices, dehydrated in ethanol and xylene, embedded in paraffin, subserially sectioned at 5 p,m, and stained with celestine blue and acid fuchsin. The sections were examined at 40 X, 100 × and 40(I × magnification with light microscopy. The examiner was blinded to treatment. In the parietal cortex, hippocampus CA1 sector and dentate hilar region the number of injured neurons were counted. In the caudoputamen, cingulate cortex and SNPR the percentage of affected area was estimated. In the other investigated brain structures (the lateral reticular, medial ventero-posterior and lateral ventero-posterior thalamic nuclei, the medial geniculate body and the colliculus superior) the damage was scored on a four-graded scale. Grade 0 represents no observable histopathological changes, and grade 1, 2 and 3 1-5%, 6 - 5 0 % and 51-100% of damaged neurons, respectively.
Statistics The differences in physiological parameters between the groups were analyzed with ANOVA analysis of variance followed by Scheffe's test. The differences between treatment and control animals in length of survival and time after ischemia until the first recognized "early" or "late" seizure activity were analyzed with Student's t-test. Since in some animals the exact time point for death and seizures was impossible to establish, the earliest possible time-point was used for the calculation. The differences in morphological scores between groups, using the individual hemisphere's scores or the "hemispheric mean", were analyzed non-parametrically with the Mann-Whitney U-test.
190 Results
Physiological parameters As described above, two different protocols were used in this study, of which the experimental conditions differed with respect to anesthesia and muscle relaxant. In the first series of experiments (protocol 1) the weight of the rats was higher compared to the second series, in which the treated group (Mo D M T U ) also weighed less than control ( P < 0.05). D M T U treatment significanly decreased blood pressure compared to the control groups in the first series of experiments, but not in the second, probably due to the vecuronium bromide treatment in these animals. However, in this series blood pressure was also lower 5 rain after ischemia, when the muscle relaxant effect had disappeared. The p H was decreased after ischemia in the control groups compared to the D M T U treated animals (significant for Clin Control and Mo Control 1), partly due to CO2-retention (significant for Clin Control and Mo Control 1). In the Clin D M T U group the mean plasma glucose was higher after ischemia than in the control group. However, most importantly, plasma glucose levels were similar in all animals before ischemia, and t e m p e r a t u r e did not differ between the groups. Clinical outcome The clinical outcome following 10 min of ischemia under hyperglycemic conditions is illustrated in Fig. 1. In control animals, the lower 5 in the figure, the results were similar to those described in previous reports
(Lundgren et al. 1990a, b). q hus, 1/5 control animals (Clin Control) exhibited repeated "early" seizure activity, whereafter no seizures were observed until the development of the "'late" fatal seizures (onset 14-23 h of recovery), to which all 5 animals succumbed (15-24 h of recovery). All animals treated with D M T U a p p e a r e d more behaviourally depressed than controls, with obvious muscle hypotonia in the recovery period. They preferred to lie down, in a more "flat" position than is usually encountered after the ischemic period, and showed less spontaneous movements. They reacted to external stimuli, like noice and handling, but to a lesser extent than the control animals. Fig. 1 shows that among the 1l animals treated with D M T U for clinical observation (Clin D M T U ) 3 showed transient "early" postischemic seizures, and one animal showed minor signs of seizure activity between 4 and 14 h of recovery. No animal died following the "early" seizures, Out of the 11 animals, 9 showed clinical improvement after 3 - 6 h of recovery and attained a prone position, b u t they were still not moving around as an alert animal would. One of the 11 animals seemed extremely weak, it never moved or showed any convulsive activity, and was lound dead after 36 h of recireulation. In total, 9/11 D M T U - t r e a t e d animals developed "late" seizures with the earliest onset after 2t h of recovery. However, in 4 animals the seizures were mild and difficult to recognize (in Fig. 1 shown as a star within parenthesis). The " l a t e " seizure activity appeared significantly later in D M T U - t r e a t e d rats ( P < 0.0l, Student's t-test). One of the 11 animals showed a different postischemic
I°
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
:
...............
~__..-)-
........
,
. . . .
..................
*t (*)
t
**t
1° 3. 4. 5.
i**t 8
(i
12
~t 1"8
~t
2"4
3"0. . . .
3"6
4"2
4"8
5"4
6"0~
T i m e recirculation Fig. 1. The figure illustrates the clinical outcome after 10 min of ischemia in hyperglycemicrats. Eleven rats received preischemic DMTU, and 5 rats were control animals. (*), *, **, and *** denote suspected but vague, one or a few minor, a couple oL and several seizure :episodes, respectively. *, denotes when the animals died. The lines on each side of a symbol denote the range between which the seizures and/or death occurred.
191 pattern, since the rat survived for a longer period and died without showing seizures between 48 and 60 h of recovery. The survival time was prolonged in DMTUtreated animals. It was 35.5 _+ 8.2 h compared to 19.8 + 3.6 h in control animals ( P < 0.01, Student's t-test). In conclusion, DMTU-treatment delayed the appearence of "late" seizures, lessened the intensity of the seizures in some animals, and hindered the appearence of any recognizable seizure activity in 2/11 animals. Also, survival time improved significantly. However, and most importantly, all 11 DMTU-treated animals died.
,~. . . . . ,.s.. . . . . ,
8~ 80m
co
ft. 60-
E
ca
m
~
o
~
o
o
40
o
20-
Histopathological outcome It should be recalled that the histopathological evaluation was performed after 15 h of recovery. This time was chosen because the seizures that characteristically develop after 15-18 h of recirculation in hyperglycemic animals severely aggravate the morphological outcome, thereby complicating the morphological evaluation (Smith et al. 1988; Lundgren et al. 1992a). In the first series of experiments (protocol 1), when isoflurane was used as the anesthetic, the results showed that some animals developed "early" and one animal "late" seizures prior to perfusion-fixation (see above). The seizures were in some animals very mild or difficult to
,- . . . . tt ..... ,
100-
o o
0 -
Control 1
DMTU 1
Control 2
DMTU 2
Fig. 2. The figure illustrates the percentage damage to caudoputamen after cerebral ischemia complicated by preischemic hyperglycemia in DMTU-treated and control animals. In the first series (left panel) isoflurane and in the second series (right panel) halothane were used as an anesthetic agents. Open symbols represent the left and filled symbols the right hemisphere, respectively. The hemispheres belonging to animals that convulsed prior to perfusion-fixation are shown by triangles, tt denote significant differences ( P < 0.0I) from the control group as calculated from individual hemispheres and animals with seizures (each 2 triangles) excluded (Mann-Whitney U-test).
Fig. 3. Photomicrographs showing damage incurred to caudoputamen after cerebral ischemia complicated by preischemic hyperglycemia in control animals (A) and DMTU-treated animals (B). Note dark shrunken neurons and sponginess in the control animal. Bar = 75 urn.
1~)2
detect. This complicates the morphological cwduation, since some seizures in the DMTU-trcated anin~als could have remained unrecognized. Therefl~rc, the second series was performed with halothanc as thc anesthetic agent (protocol 2). Halothane-anesthetized animals have previously been shown to bc associated with a seizure-free period until the development of "late" seizures (Smith et al. 1988) (compare also Lundgrcn ct al. 199(ta, 1992a). Despite halothanc anesthesia one of the saline-treated control animals showed "early" seizure activity, but the remaining 15/16 subjects exhibited no such activity.
7 .....
n~,
-
--
,
100
o~ x
o
80
o
rlo
60
(,
e-
"6 c ¢J
40
o •
¢o
E m D
# 20
oo
I
recovery.
i
&
Control animals
In both series the control animals showed a similar degree of brain damage in all structures evaluated (Figs. 2-7), and the degree was similar to that recorded in previous studies (Smith et al. 1988; Lundgren et al. 1992a). Thus, both regions showing selective neuronal vulnerability, such as caudoputamen and the dentate hilar region, and regions showing "additive" damage correlating to excessive acidosis, such as the cingulate cortex and SNPR, were severely damaged after 15 h of
N
Control 1
DMTU 1
Control 2
DMTU 2
Fig. 4. ]'he figure illustrates the percentage damage to cingulate cortex. For explanation of groups a n d symbols, see legend for Fig. "~
DMTU-treated animals
In the first series (protocol 1, isoflurane anesthesia) D M T U treatment did not significantly reduce morphological damage in any brain structures evaluated after 15 h of recovery. However, there appeared to be a little less damage in the treated compared to control ani-
Fig. 5, Photomicrographs showing the damage incurred to cingulate cortex in control animals (a) and D M T U - t r e a t e d animals (b), Note the triangular-shaped dark shrunken neurons in the control animals. Bar = 75 t x m
193 ,- . . . .
+ ..... ,
.~. . . . . n.~. . . . . . ,
I00 •
v
80
rv" ft..
z
60
.E
g
o 40
E
2O
0
co Control 1
c~ DMTU 1
co Control 2
c8~ DMTU 2
Fig. 6. T h e f i g u r e i l l u s t r a t e s the p e r c e n t a g e d a m a g e to s u b s t a n t i a n i g r a p a r s r e t i c u l a t a ( S N P R ) . F o r e x p l a n a t i o n o f g r o u p s a n d symbols, see l e g e n d s to Fig. 2. * d e n o t e s s i g n i f i c a n t d i f f e r e n c e s ( P < 0.05) f r o m the c o n t r o l g r o u p as c a l c u l a t e d f r o m i n d i v i d u a l h e m i s p h e r e s a n d a n i m a l s with s e i z u r e s ( e a c h 2 t r i a n g l e s ) e x c l u d e d ( M a n n - W h i t ney U-test).
(see Figs. 6 and 7) was significantly decreased in DMTU-treated subjects (P
187
JNS 03890
Effects of dimethylthiourea on ischemic brain damage in hyperglycemic rats Johan Lundgren a,b, Maj-Lis Smith ~' and Bo K. S i e s j 6 a " Laboratoo' for Experirnental Brain Research, Department of Neurobioh#,~', Experimental Research Center, Unit~ersity of Lurid, Lurid, Sweden, and ~ Department of Paediatrics, Unicersity ttospital, Lund, Sweden (Received 2 December, 1091) (Revised, received 18 May, 1992) (Accepted 8 July, 1992)
Key words: Ischemia; Hyperglycemia; Brain damage; Dimethylthiourca; Seizures; Postischemic seizures; Rats
Summary Hyperglycemia is known to worsen the outcome of transient global or forebrain ischemia. The aggravating effect is believed to be mediated by the additional formation of lactate and of H +. Recent evidence suggests that reactive oxygen species contribute to the damage after brain ischemia. Since acidosis accelerates free radical damage in vitro, we decided to explore if ischemic damage in hyperglycemic subjects is ameliorated by dimethylthiourea (DMTU), an established free radical scavenger. In one series of hyperglycemic rats, we studied whether preischemic administration of DMTU alters the clinical outcome, notably the incidence and frequency of seizures. In two different series, the effect of DMTU on tissue damage was assessed by light microscopy after 15 h of recovery. Longer periods could not be studied since seizures developed. In the first of these serics the animals were anesthetized with isoflurane, and in the second with halothane. The latter anesthesia largly suppressed the "early" postischemic seizures, i.e. those occurring after 1-4 h. Dimethylthiourea treatment altered the clinical outcome after ischemia. Thus, the "late" postiscbemic seizures appeared milder and occurred significantly later than in untreated animals. The fatal outcome was also delayed since treated animals died after 35.5 _+ 8.2 h (mean _+ SD) of recirculation, as compared to 19.8 _+ 3.6 h of recirculation in control animals. However, all DMTU-treated (and control) animals died. In the first morphological series (isoflurane anesthesia) the histopathological analysis was complicated by the occurrence of prefixation seizures; such seizures were recognized in 4/16 animals. When these 4 animals were excluded from the analysis (2 treated and 2 control animals), DMTU pretreatment did not ameliorate the damage, except in the substantia nigra pars reticulata (P < 0.05). In the second series, comprising animals anesthetized with halothane, only one animal out of 16 had "early" seizures, and none showed "late" seizures before death. Among these animals DMTU treatment significantly ameliorated damage to caudoputamen and cingulate cortex (P < (1.01). We conclude that treatment with the free radical scavenger DMTU partly ameliorates ischemic brain damage associated with excessive acidosis, and marginally delays the development of post-ischemic seizures. However, the effects were moderate and could, at least in part, have been caused by nonspecific effects of DMTU. Furthermore, all DMTU-treated animals died. The results thus give little support to the notion that the aggravating effects of acidosis is due to enhancement of free radical production.
Introduction S u b s t a n t i a l evidence exists that the b r a i n level a n d / o r supply of glucose influence the o u t c o m e of
Correspondence to: Johan Lundgren, M.D., Ph.D., Laboratory for Experimental Brain Research, Department of Neurobiology, Experimental Research Center, University Hospital, S-221 85 Lund, Sweden. Tel (46-46) 173552; Fax (46-46) 151480.
c e r e b r a l ischemia. Thus, p r e i s c h e m i c h y p e r g l y c e m i a a g g r a v a t e s b r a i n d a m a g e ( M y e r s a n d Y a m a g u c h i 1977; Siemkowicz and H a n s e n 1978; K a l i m o et al. 1981; R e h n c r o n a et al. 1981; Pulsinelli et al. 1982a; Smith et al. 1988). It alters the d i s t r i b u t i o n and evolution o f b r a i n injury ( I n a m u r a et al. 1988; Smith et al. 1988), e n h a n c e s e d e m a f o r m a t i o n (Pulsinelli et al. 1982b; W a r n e r et al. 1987), and favours fatal p o s t i s c h e m i c seizures ( M y e r s 1979; Siemkowicz 1985; Smith et al.
188 1988; Lundgren et al. 1990a). The mechanisms by which elevated brain glucose levels exaggerate brain injury seem related to enhanced accumulation of lactate plus H ~, with a further decrease in intra- and extracellular p H (Ljunggren et al. 1974; Rehncrona et al. 1980, 1981; Welsh et al. 1980; Kraig et al. 1987; Chopp et al. 1988; Siesj6 1988; Kraig and Chesler 199(I). Reactive oxygen species (ROS) have been implicated as mediators of ischemic injury to many organs. including the brain (Halliwell 1985; Siesj6 et al. 19891. The best evidence is that pretreatment with a variety of free radical scavengers lessens the damage that develops following ischemia of different types (Abe et al. 1988; Hall et al. 1988: P a t t e t al. 1988; Beckman et al. 1989; Martz et al. 1989; Imaizumi et al. 1990: Martz el al. 1990). In a few cases, formation of ROS in the tissue has also been demonstrated in vivo ( P a t t e t a[. 1988; Bromont et al. 1989; Oliver et al. 1990; Kirsch et al. 1991). However, contradictory, results have been reported and, in general, we do not accurately know the route by which ROS are formed, the conditions under which they play important pathogenic roles, and the mechanisms by which they exert their harmful effects. It has been proposed that the participation of free radicals is likely when ischemia is of long duration, or if it is complicated by hyperglycemia or hyperthermia (Siesj6 et al. 1989; 1990). These are the conditions under which ischemia gives rise to pannecrosis rather than selective neuronal necrosis. Common to these conditions is that the cerebral microvessels arc involved in the damaging process. In contrast, short periods of ischemia under normothermic and normoglycemic conditions leave the microvessels in the brain essentially unaffected (Pulsinelli et al. 1982b; Smith ct al. 1984). Several facts suggest that free radicals are formed in cerebral microvessels. First, xanthine oxidase, which is one of the main generators of ROS (McCord 19851, seems localized to the microvessel fraction (Jarasch et al. 1986). Second, ischemia leads to activation of leucocytes and such activation inw)lves enhanced oxygen consumption and generation of ROS (Babior et al. 1987; Curnutte and Babior 1987). At least in the heart, reperfusion injury, is ameliorated by prior leukocyte depletion (Simpson el al. 1987; Breda ct al. 1989). Third, the edema which is formed after cerebral ischemic insults is at least in part reduced by free radical scavengers ( C h a r et al. 1987; Martz ct al. 1989; Betz 1990). Since excessive acidosis may increase the formation of ROS by at least two different mechanisms, it can be speculated that acidosis-related damage is mediated by free radicals (Siesj6 et al. 1985; Siesj6 1988). The first mechanism is related to release of prooxidant iron (or copper) and chelation to low molecular weight compounds. A decrease in pH favours iron release from intracellular ferritin (Harrison 1977), and from extra-
cellular transferrin (Lestas i%'(~: Aisen. 19791. When chelated by low molecular a.eight compounds, iron in its reduced form (Fe a+) catalyzes the formation ~[ highly reactive - O H through the Habcr-Weiss reaction (Halliwell 1990). The second mechanism rclatcs to the fact that protonation of -O; yiclds the more reactive and more lipid-mluble perhydroxyl radical (.HO~), and since this is an acid with a p K~, of 4.9 its formatiori is favoured by low pH (Gebicki and Bielski 19811. In vitro, lowering of pH in a brain homogenatc triggers enhanced production of frec radicals and increases lipid peroxidation (Barber and Bernheim 1:967; Siesj6 ct al. 1985: Rehncrona ct al. i9891, Using H20:~ production as a reflection of ~eactive oxygen formation (Patt el al. 19881 we failed ~,~ yetiS' such formation after a brief period of ischemia in normoglyccmic animals (Agardh et al. 199t)). tu lhc light oLt our basic hypothesis, this result was not unexpected. However, the results obtained following 15 rain of ischcmia in hyperglycemic animals were also negative, since we were unable to find an increase in H , O , production, or an increase in bleomycin-rcactive frec iron in (?$1: (l,undgren et al. 1991a). Furthermore, we could find no additional reduction in tissm: contents of glutathione or ~z-tocophcrol in hyperglycemic animals subjected to ischemia. It has recently become clear that DMTLi, an established . O t t scavenger, ameliorates ischemic brain injury tbllowing long periods of either transient or permanent ischemia ( P a t t c t al, i988; Martz et a ! ]q~89, 1990; Betz 1990). Since amelioration ol damage by an established free radical scavenger may be a sensitive indicator of such injury we explored the hypothesis that ischemic brain damage in hyperglycemic animals was ameliorated by DMTIJ.
Methods
Animals and experimental groups All experiments were performed on adult male Wistar rats of an S.P.F. strain (M011egaard's Breeding Center, Copenhagen). Two series of experiments were performed, both by the same person but with slightly different anesthesia. In the first series (protocol 1, isoflurane anesthesia), the effect of treatment with DMTU was studied with regard to clinical (Clin DMTU, lz--11) and morphological outcome (Mo D M T U I, n = 8), and compared to no treatment (Clin Control n - - 5 ; Mo Control 1, n = 8). In the second series (protocol 2, halotbane anesthesia), the effect of D M T U on the morphological outcome was investb gated (Mo D M T U 2, n .... 8" Mo Control 2, n = 8 1 . In the first series the weight of lhe animals was 290-360 g, and in the second it was 235-280 g, corresponding to lhe ages of about 91/ and 65 days, respectively. All
189 animals were housed in macrolon cages, and fasted overnight prior to the experiments, but allowed water ad libitum.
Preparation and dose of DMTU Two g D M T U (Janssen Chimica, Geel, Belgium) were freshly mixed in 20 ml 0.9% NaCI, yielding a concentration of 100 rag. ml J. The solution was used within 6 h. Each animal received 750 mg. kg -l (7.5 m l . k g -~) i.p. 45-60 min prior to the induction of ischemia. Control animals in the second series received saline i.p. (7.5 m l . k g I).
Operatit~e and experimental procedures The operative and experimental procedures for protocol 1 have been described earlier (Lundgren et al. 1990b). Anesthesia was induced with 3.5% isoflurane (Forene '~, Abbott Laboratories Ltd, England), and 70% N20 in Oz, after which the animal was intubated and connected to a respirator delivering 1.5-2.0% isoflurane. Then the drug was injected to animals intended for D M T U treatment. A central venous catheter was inserted through a neck incision via the right caval vein. A string was placed around each common carotid artery and a tail arterial catheter was inserted for blood pressure measurement, analysis of blood gases and pH (ABL 30, Radiometer A / S , Copenhagen, Denmark), and plasma glucose (Beckman Glucose Analyser 2, Beckman Instruments Inc. Fullerton, CA, USA). Needle electrodes for E E G recording were inserted bitemporally with the reference electrode in the tail, and temperature probes were placed subcutaneously on the skull bone and in the rectum. After the surgical procedures the isoflurane concentration in the delivered gas mixture was decreased to ~ 1.0%, and the animals were allowed a steady state period of 30 rain. During the first 5 min of the steady state period a glucose load of an i.v. solution containing 25% glucose (0.92 g. kg ~ body weight) was given, followed by a slower infusion (0.04g. kg J • rain - l ) of the same solution. Heparin was given (50 IE) prior to the first blood sampling. After the steady state period the glucose infusion and the isoflurane administration were discontinued, and the induction of ischemia was started with central venous exsanguination. When the blood pressure was below 50 mm Hg, bilateral carotid clamping was performed. Blood pressure was kept close to 50 mm Hg with additional withdrawal or reinfusion of shed blood, and ischemia was verified by absence of E E G activity. Ischemia was terminated by removal of the carotid damps, reinfusion of shed blood, and i.v. injection of 0.5 ml of a 0.6 M sodium bicarbonate solution. During the early recirculation period the blood pressure was continuously recorded, and PaO2, PaCO 2 and pH were controlled. When the animal showed strong breathing movements (after 20-30 rain
of recovery) it was extubated, and transferred to a macrolon cage. For the second series of experiments (protocol 2) the experimental conditions were altered so as to reduce the "early" seizures. For that purpose halothane (Halothane '~'~, ISC Chemicals, England) in similar concentrations was used instead of isoflurane for induction of anesthesia and during the surgical procedures. The concentration was decreased to and kept at 0.4-0.5% during the steady state and ischemic period, whereafter it was discontinued. During the steady state period, the muscle-relaxant vecuronium bromide (Norcuron ", Organon Teknika, Boxtel, Holland) was administered at a concentration of 1 rag" ml-~ mixed with the 25% glucose solution. The experimental procedure was approved by the Ethical Committee for Laboratory Animal Experiments at the University of Lund.
Histology Following 14-16 h of recovery the animals were perfusion-fixed with phosphate-buffered formaldehyde (Auer et al. 1984). After stablization for one day at 5°C the brains were dissected and stored in cold fixative. Later the brains were cut in coronal slices, dehydrated in ethanol and xylene, embedded in paraffin, subserially sectioned at 5 p,m, and stained with celestine blue and acid fuchsin. The sections were examined at 40 X, 100 × and 40(I × magnification with light microscopy. The examiner was blinded to treatment. In the parietal cortex, hippocampus CA1 sector and dentate hilar region the number of injured neurons were counted. In the caudoputamen, cingulate cortex and SNPR the percentage of affected area was estimated. In the other investigated brain structures (the lateral reticular, medial ventero-posterior and lateral ventero-posterior thalamic nuclei, the medial geniculate body and the colliculus superior) the damage was scored on a four-graded scale. Grade 0 represents no observable histopathological changes, and grade 1, 2 and 3 1-5%, 6 - 5 0 % and 51-100% of damaged neurons, respectively.
Statistics The differences in physiological parameters between the groups were analyzed with ANOVA analysis of variance followed by Scheffe's test. The differences between treatment and control animals in length of survival and time after ischemia until the first recognized "early" or "late" seizure activity were analyzed with Student's t-test. Since in some animals the exact time point for death and seizures was impossible to establish, the earliest possible time-point was used for the calculation. The differences in morphological scores between groups, using the individual hemisphere's scores or the "hemispheric mean", were analyzed non-parametrically with the Mann-Whitney U-test.
190 Results
Physiological parameters As described above, two different protocols were used in this study, of which the experimental conditions differed with respect to anesthesia and muscle relaxant. In the first series of experiments (protocol 1) the weight of the rats was higher compared to the second series, in which the treated group (Mo D M T U ) also weighed less than control ( P < 0.05). D M T U treatment significanly decreased blood pressure compared to the control groups in the first series of experiments, but not in the second, probably due to the vecuronium bromide treatment in these animals. However, in this series blood pressure was also lower 5 rain after ischemia, when the muscle relaxant effect had disappeared. The p H was decreased after ischemia in the control groups compared to the D M T U treated animals (significant for Clin Control and Mo Control 1), partly due to CO2-retention (significant for Clin Control and Mo Control 1). In the Clin D M T U group the mean plasma glucose was higher after ischemia than in the control group. However, most importantly, plasma glucose levels were similar in all animals before ischemia, and t e m p e r a t u r e did not differ between the groups. Clinical outcome The clinical outcome following 10 min of ischemia under hyperglycemic conditions is illustrated in Fig. 1. In control animals, the lower 5 in the figure, the results were similar to those described in previous reports
(Lundgren et al. 1990a, b). q hus, 1/5 control animals (Clin Control) exhibited repeated "early" seizure activity, whereafter no seizures were observed until the development of the "'late" fatal seizures (onset 14-23 h of recovery), to which all 5 animals succumbed (15-24 h of recovery). All animals treated with D M T U a p p e a r e d more behaviourally depressed than controls, with obvious muscle hypotonia in the recovery period. They preferred to lie down, in a more "flat" position than is usually encountered after the ischemic period, and showed less spontaneous movements. They reacted to external stimuli, like noice and handling, but to a lesser extent than the control animals. Fig. 1 shows that among the 1l animals treated with D M T U for clinical observation (Clin D M T U ) 3 showed transient "early" postischemic seizures, and one animal showed minor signs of seizure activity between 4 and 14 h of recovery. No animal died following the "early" seizures, Out of the 11 animals, 9 showed clinical improvement after 3 - 6 h of recovery and attained a prone position, b u t they were still not moving around as an alert animal would. One of the 11 animals seemed extremely weak, it never moved or showed any convulsive activity, and was lound dead after 36 h of recireulation. In total, 9/11 D M T U - t r e a t e d animals developed "late" seizures with the earliest onset after 2t h of recovery. However, in 4 animals the seizures were mild and difficult to recognize (in Fig. 1 shown as a star within parenthesis). The " l a t e " seizure activity appeared significantly later in D M T U - t r e a t e d rats ( P < 0.0l, Student's t-test). One of the 11 animals showed a different postischemic
I°
2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
:
...............
~__..-)-
........
,
. . . .
..................
*t (*)
t
**t
1° 3. 4. 5.
i**t 8
(i
12
~t 1"8
~t
2"4
3"0. . . .
3"6
4"2
4"8
5"4
6"0~
T i m e recirculation Fig. 1. The figure illustrates the clinical outcome after 10 min of ischemia in hyperglycemicrats. Eleven rats received preischemic DMTU, and 5 rats were control animals. (*), *, **, and *** denote suspected but vague, one or a few minor, a couple oL and several seizure :episodes, respectively. *, denotes when the animals died. The lines on each side of a symbol denote the range between which the seizures and/or death occurred.
191 pattern, since the rat survived for a longer period and died without showing seizures between 48 and 60 h of recovery. The survival time was prolonged in DMTUtreated animals. It was 35.5 _+ 8.2 h compared to 19.8 + 3.6 h in control animals ( P < 0.01, Student's t-test). In conclusion, DMTU-treatment delayed the appearence of "late" seizures, lessened the intensity of the seizures in some animals, and hindered the appearence of any recognizable seizure activity in 2/11 animals. Also, survival time improved significantly. However, and most importantly, all 11 DMTU-treated animals died.
,~. . . . . ,.s.. . . . . ,
8~ 80m
co
ft. 60-
E
ca
m
~
o
~
o
o
40
o
20-
Histopathological outcome It should be recalled that the histopathological evaluation was performed after 15 h of recovery. This time was chosen because the seizures that characteristically develop after 15-18 h of recirculation in hyperglycemic animals severely aggravate the morphological outcome, thereby complicating the morphological evaluation (Smith et al. 1988; Lundgren et al. 1992a). In the first series of experiments (protocol 1), when isoflurane was used as the anesthetic, the results showed that some animals developed "early" and one animal "late" seizures prior to perfusion-fixation (see above). The seizures were in some animals very mild or difficult to
,- . . . . tt ..... ,
100-
o o
0 -
Control 1
DMTU 1
Control 2
DMTU 2
Fig. 2. The figure illustrates the percentage damage to caudoputamen after cerebral ischemia complicated by preischemic hyperglycemia in DMTU-treated and control animals. In the first series (left panel) isoflurane and in the second series (right panel) halothane were used as an anesthetic agents. Open symbols represent the left and filled symbols the right hemisphere, respectively. The hemispheres belonging to animals that convulsed prior to perfusion-fixation are shown by triangles, tt denote significant differences ( P < 0.0I) from the control group as calculated from individual hemispheres and animals with seizures (each 2 triangles) excluded (Mann-Whitney U-test).
Fig. 3. Photomicrographs showing damage incurred to caudoputamen after cerebral ischemia complicated by preischemic hyperglycemia in control animals (A) and DMTU-treated animals (B). Note dark shrunken neurons and sponginess in the control animal. Bar = 75 urn.
1~)2
detect. This complicates the morphological cwduation, since some seizures in the DMTU-trcated anin~als could have remained unrecognized. Therefl~rc, the second series was performed with halothanc as thc anesthetic agent (protocol 2). Halothane-anesthetized animals have previously been shown to bc associated with a seizure-free period until the development of "late" seizures (Smith et al. 1988) (compare also Lundgrcn ct al. 199(ta, 1992a). Despite halothanc anesthesia one of the saline-treated control animals showed "early" seizure activity, but the remaining 15/16 subjects exhibited no such activity.
7 .....
n~,
-
--
,
100
o~ x
o
80
o
rlo
60
(,
e-
"6 c ¢J
40
o •
¢o
E m D
# 20
oo
I
recovery.
i
&
Control animals
In both series the control animals showed a similar degree of brain damage in all structures evaluated (Figs. 2-7), and the degree was similar to that recorded in previous studies (Smith et al. 1988; Lundgren et al. 1992a). Thus, both regions showing selective neuronal vulnerability, such as caudoputamen and the dentate hilar region, and regions showing "additive" damage correlating to excessive acidosis, such as the cingulate cortex and SNPR, were severely damaged after 15 h of
N
Control 1
DMTU 1
Control 2
DMTU 2
Fig. 4. ]'he figure illustrates the percentage damage to cingulate cortex. For explanation of groups a n d symbols, see legend for Fig. "~
DMTU-treated animals
In the first series (protocol 1, isoflurane anesthesia) D M T U treatment did not significantly reduce morphological damage in any brain structures evaluated after 15 h of recovery. However, there appeared to be a little less damage in the treated compared to control ani-
Fig. 5, Photomicrographs showing the damage incurred to cingulate cortex in control animals (a) and D M T U - t r e a t e d animals (b), Note the triangular-shaped dark shrunken neurons in the control animals. Bar = 75 t x m
193 ,- . . . .
+ ..... ,
.~. . . . . n.~. . . . . . ,
I00 •
v
80
rv" ft..
z
60
.E
g
o 40
E
2O
0
co Control 1
c~ DMTU 1
co Control 2
c8~ DMTU 2
Fig. 6. T h e f i g u r e i l l u s t r a t e s the p e r c e n t a g e d a m a g e to s u b s t a n t i a n i g r a p a r s r e t i c u l a t a ( S N P R ) . F o r e x p l a n a t i o n o f g r o u p s a n d symbols, see l e g e n d s to Fig. 2. * d e n o t e s s i g n i f i c a n t d i f f e r e n c e s ( P < 0.05) f r o m the c o n t r o l g r o u p as c a l c u l a t e d f r o m i n d i v i d u a l h e m i s p h e r e s a n d a n i m a l s with s e i z u r e s ( e a c h 2 t r i a n g l e s ) e x c l u d e d ( M a n n - W h i t ney U-test).
(see Figs. 6 and 7) was significantly decreased in DMTU-treated subjects (P
