www.elsevier.com / locate / bres
Research report
Mild hypothermia decreases the incidence of transient ADC reduction
detected with diffusion MRI and expression of c-fos and hsp70
mRNA during acute focal ischemia in rats
a,e ,
*
a,c a a,b,dAnthony Mancuso
, Nikita Derugin
, Kazushi Hara , Frank R. Sharp
,
a
Philip R. Weinstein
a
Department of Neurological Surgery, University of California at San Francisco and the Department of Veterans Affairs Medical Center,
San Francisco, CA 94121, USA b
Department of Neurology, University of California at San Francisco and the Department of Veterans Affairs Medical Center, San Francisco,
CA94121, USA c
Department of Radiology, University of California at San Francisco and the Department of Veterans Affairs Medical Center, San Francisco,
CA94121, USA d
Department of Neurology, University of Cincinnati, Cincinnati, OH 45267, USA e
Department of Radiology, University of Pennsylvania, Philadelphia, PA 19104, USA Accepted 12 September 2000
Abstract
The effects of mild hypothermia on the apparent diffusion coefficient of water (ADC) and expression of c-fos and hsp70 mRNA were examined during acute focal cerebral ischemia. Young adult rats were subjected to 60-min middle cerebral artery occlusion under either normothermia (37.58C) or hypothermia (338C). Diffusion-weighted echo-planar magnetic resonance imaging was used to monitor changes in ADC throughout the ischemic period. Perfusion MRI with dysprosium contrast was used at the end of the ischemic period to verify that the occlusion was successful. C-fos and hsp70 mRNA expression were examined with in situ hybridization at the end of the ischemic period. The results indicate that the size of the region that exhibited reduced ADC was smaller during hypothermia than during normothermia. Hypothermia also decreased the frequency of occurrence of transient ADC reductions, especially in dorsal aspects of cortex. Expression of both c-fos and hsp70 mRNA were markedly reduced by hypothermia. Transient ADC reduction and c-fos expression are associated with spreading depression, which is believed to contribute to lesion expansion during acute focal ischemia. The results suggest that part of the neuroprotective effect of hypothermia may be due to a reduced incidence of spreading depression.
2000 Elsevier Science B.V. All rights reserved.
Theme: Disorders of the nervous system
Topic: Ischemia
Keywords: Hypothermia; Diffusion MRI; Apparent diffusion coefficient; Focal cerebral ischemia; C-fos; Hsp70
1. Introduction (20 to 308C) has been found to be useful clinically following head trauma and during surgical procedures that Hypothermia is highly effective in reducing injury disrupt the cerebral circulation [23]. However, it is also during cerebral ischemia. Deep to moderate hypothermia associated with numerous adverse affects such as cardiac dysfunction [45,62]. Mild hypothermia (|338C), while less
protective than deep hypothermia [4,24], has gained con-siderable interest recently because it provides significant
*Corresponding author. Department of Radiology / 4283, B1 Stellar- neuroprotection with far fewer complications [36]. The Chance Laboratory, 422 Curie Boulevard, University of Pennsylvania,
effect of mild hypothermia on clinical ischemic stroke has
Philadelphia, PA 19104, USA. Tel.:11-215-898-1805; fax:1
1-215-573-not yet been examined in large clinical trials [21,36].
2113.
E-mail address: [email protected] (A. Mancuso). In animal models of global ischemia, mild hypothermia
can reduce infarct size and the extent of ischemic neuronal occurred in regions exhibiting both transient and perma-damage in non-infarcted regions [8,11,15,17]. For focal nent ADC reduction [47].
ischemia, mild hypothermia can reduce the size of the Hypothermia can affect postischemic expression of both necrotic region [32,35,50] and the extent of induction of c-fos and hsp70 mRNA following temporary occlusion. apoptosis in non-necrotic regions [46]. It can also help to C-fos expression following temporary forebrain ischemia prevent neuronal and endothelial damage associated with is hastened by moderate (308C) hypothermia [33,39]. reperfusion injury, if it is initiated during [28,35] or after Cerebral hsp70 expression can be induced by 8-h sustained ischemia [41]. Applying hypothermia during the acute mild hypothermia without ischemia [18]. Expression fol-phase of prolonged (3 to 4 day) permanent ischemia can lowing brief forebrain ischemia is reduced by moderate reduce the rate of pathophysiological changes, but it hypothermia [16,40]. The effect of hypothermia on c-fos generally does not reduce the final extent of injury [53]. and hsp70 expression during acute ischemia has not been
Hypothermia has many effects on the underlying reported.
pathophysiological changes that occur during cerebral In this work, the effect of intra-ischemic mild (338C) ischemia. It reduces cerebral metabolic rate (CMR) as hypothermia on both transient and persistent ADC reduc-evidenced by decreased glucose and oxygen consumption tion and expression of c-fos and hsp70 mRNA was rates [4], which delays the onset of energy failure and examined. The hypotheses tested were that hypothermia terminal ischemic depolarization. Direct measurement of would reduce the extent of persistent ADC reduction, the
31
ATP levels by P NMR and invasive methods has incidence of transient ADC reduction and c-fos mRNA confirmed that hypothermia slows the rate of energy expression.
depletion [42,63]. However, this does not completely account for its neuroprotective capacity since anesthetics
that diminish CMR to the same extent as hypothermia do 2. Methods
not provide the same degree of neuroprotection [52,61].
Hypothermia also reduces neurotransmitter release dur- 2.1. Animal model ing ischemia, including glutamate [3,12,28,64], aspartate
[7], glycine, gamma-aminobutyric acid [51], and dopamine Male Sprague–Dawley rats weighing 270–325 g were [12]. Glutamate is a potent excitotoxin that can hasten used. They were given free access to food and water prior ischemic depolarization in regions where energy depletion to use in experiments. Anesthesia was induced with 3% is incomplete [57]. In focal ischemia, hypothermia reduces isoflurane in a closed chamber and subsequently main-intra-ischemic increases of glutamate in penumbral [64] tained with 1.5% isoflurane (both in 70 / 30 N O / O )2 2
but not core regions [28]. However, recent results suggest through an intratracheal 16-gauge Teflon tube. A rodent that the reduction in glutamate release does not fully ventilator (Harvard Instruments, South Natick, MA) was account for the neuroprotective effect of hypothermia [66]. used for mechanical ventilation; respiratory adjustments Diffusion-weighted imaging is sensitive to changes in were made to maintain normal arterial blood gas levels. the apparent diffusion coefficient of water (ADC). Re- Just before imaging, pancuronium bromide (0.2 mg in 0.2 duced ADC during ischemia is believed to be, at least in ml of 0.9% saline) was administered intraperitoneally to part, the result of cellular edema, which occurs with induce paralysis and thereby minimize motion artifacts. ischemic depolarization [60]. Such depolarization is pri- Permanent focal ischemia was produced by occluding marily caused by energy failure [6,13,20] or by excitotoxic the MCA as described previously [34]. A 3-0 monofila-injury from neurotransmitters [6]. Reduced ADC also ment nylon suture with a heat-rounded tip was introduced occurs during spreading depression due to abnormal ion into an internal carotid artery (ICA), through the stump of fluxes that produce temporary cellular edema [10,43]. a transected external carotid artery (ECA). It was advanced Spreading depression is believed to be an important 22 mm past the bifurcation of the common carotid artery component of lesion size expansion during focal ischemia (CCA) to block the origin of the MCA. The CCA proximal [27]. Additional phenomena that are not as yet completely to the occluded MCA was left patent. The duration of the understood may also contribute to reduced ADC during occlusion was 60 min.
rectal temperature accurately represented the brain tem-perature [47]. Blood gas samples were taken from a femoral artery catheter before and during the ischemic period. Arterial blood pressure was monitored with the same catheter.
2.2. Magnetic resonance methods
Diffusion-weighted echo-planar images were acquired with an MBEST pulse sequence on a 2-Tesla Omega CSI system (Bruker Medical Systems, Fremont, CA), equipped with 20 gauss / cm self-shielded gradients. A 5.5-cm diam-eter low-pass birdcage coil was used for both pulse transmission and signal reception. Half sine
diffusion-Fig. 1. Regions of interest (ROI) used to evaluate the association of
sensitizing gradients of up to 6 gauss / cm were used to
2 transient ADC reduction and c-fos expression and to determine the
produce b-values of up to 1710 s / mm . The diffusion
frequency of significant transient ADC reduction.
gradients were applied parallel to the long axis of the brain. Other image acquisition parameters were: FOV550
mm, slice thickness52 mm, and acquisition matrix size5 radiographs. A 30% area threshold was assumed for 1283128. The relaxation delay between images was 3.0 s. classifying a region as c-fos-positive.
Images were acquired for a single coronal slice in each rat,
centered at the level of bregma. A series of 8 to 11 images 2.3. In situ hybridization for c-fos and hsp70 mRNA were used to construct each ADC map. To verify that the
occlusion of the MCA was successful, perfusion imaging Frozen brains were cut with a cryostat (Hacker Instru-was performed with dysprosium DTPA–BMA contrast at ments, Fairfield, NJ) into 20-mm-thick coronal sections. the end of the 60-min ischemic period. Additional details The sections were analyzed for expression of c-fos and of the imaging methods were described previously [47]. hsp70 mRNA with modified versions [37] of methods ADC maps were constructed with an IDL (Research described by Schalling [54]. The probe used to detect Systems Inc., Boulder, CO) program written in our labora- hsp70 was a synthetic oligonucleotide that corresponds to tory [47]. The program performed pixel-by-pixel least- highly conserved amino acids (122–129) near the 59-end squares regression to the line: ln S /Ss od5 2b ADC , wheres d of the human hsp70 coding sequence [29]. For c-fos, the
S is the diffusion-gradient attenuated pixel intensity, S iso sequence complementary to bases 142–186 was used [19]. the unattenuated pixel intensity, and b is the attenuation Additional details were described previously [47]. factor [44]. Values of b were calculated from the equation:
2
b5s2Gdg/pd sD 2 d/ 4 , whered g is the gyromagnetic ratio 2.4. Statistical analysis of the hydrogen nucleus, G is the diffusion gradient
magnitude, and d and D are the duration and inter-pulse A single-tailed t-test was used to determine differences delay, respectively, for the diffusion gradients. The pro- in ADC lesion size between the two groups. Differences in gram was also used to compute ADC threshold maps and the frequency of transient ADC reduction between the to interpolate the results to 2563256 resolution. A reduc- normothermic and hypothermic groups in the regions of tion of 20%, relative to the average value of the non- interest of Fig. 1 and in c-fos-positive regions were ischemic hemisphere, was assumed for distinguishing the determined with the Mann–Whitney Rank Sum Test. The ADC lesion from the rest of the brain. Fisher Exact Test was used to determine if c-fos and hsp70 A second IDL program was used to identify regions expression occurred more frequently during normothermia with transient reduction in ADC (TR-ADC). The program than hypothermia. The association of TR-ADC reduction initially masked the region with persistently reduced ADC with c-fos expression was evaluated with the Chi-Squared on all maps in each time series. Persistent reduction in Test. All statistical calculations were performed with ADC (PR-ADC) was determined by averaging the final Sigma Stat 2.03 (SPSS, Chicago, IL). Values presented in three ADC maps and identifying the pixels with reductions Tables 1, 2 and 4 are mean6S.D.
of 20% or more. The program was then used to determine the size of the ADC lesions (as a function of time) within
the five regions of interest (ROI — 1, 2, 3, 4, and 5) 3. Results
shown in Fig. 1. A region was considered positive for
TR-ADC if at least 30% of the region was transiently 3.1. Physiological data included in the ADC lesion. Similarly, the same five
Table 1
a Physiologic blood gas data during pre and intra-ischemic periods
Normothermic group Hypothermic group
MABP pH PaCO2 PaO2 MABP pH PaCO2 PaO2
(mmHg) (mmHg) (mmHg) (mmHg) (mmHg) (mmHg)
Pre-isch 113617 7.4860.06 3066 143614 105613 7.5060.05 3167 158618 Intra-isch 131617 7.4460.06 2967 132616 10364 7.3760.07 42619 144620 a
All values are mean6S.D.
intra-ischemic periods are summarized in Table 1. The the first 20 min of ischemia. Transient expansion of the blood pressure, pH, and PO2 values are all in the normal ADC lesion into dorsal medial cortex occurred repeatedly. physiological range. However, the PCO2 levels before The time-course for the percentage of ROI-1 (see Fig. 1) ischemia for both groups and during ischemia for the included in the ADC lesion is shown in Fig. 3B. Three normothermic group were slightly low. The potential significant transient increases occurred with a duration of significance of this observation will be presented in the approximately 3 min. The maximum extent of the transient
discussion section. increases was slightly more than 50% of ROI-1.
Results very similar to those shown in Fig. 2A were
3.2. Normothermic group observed in two other rats for the normothermic group,
which also showed transient ADC reduction and c-fos Fig. 2 shows typical ADC maps and autoradiographs for expression in ROI-1. Smaller final ADC lesions were rats in both the normothermic and hypothermic groups. observed in the remaining four rats. Two rats had ADC Regions with ADC reductions greater than 20% of the lesions that encompassed all of the basal ganglia and a mean value of the contralateral hemisphere (ADC lesion) small part of cortex. For both of these rats, TR-ADC are shown in color. For the normothermic rat in Fig. 2A, occurred repeatedly in much of cortex. The remaining two ischemia initially produced significant ADC reductions of rats had persistent ADC reduction in only part of the basal 20–45% in the lateral basal ganglia. The region with ganglia; in these rats TR-ADC occurred over a large part reduced ADC spread during the following 57 min and of the MCA territory.
eventually encompassed almost all of the MCA territory. For the rat in Fig. 2A, c-fos mRNA was induced in the The second ADC map in the figure represents the average region where TR-ADC was observed. Hsp70 mRNA was of the final three ADC maps; the colored region for this weakly induced over a large part of the MCA distribution. averaged result was defined as the persistently ischemic It was induced to a much greater extent at the dorsomedial region. On the third ADC map, the persistently ischemic edge of the MCA territory in cortex, where TR-ADC was region is shown in dark gray. By digitally superimposing observed, but not in cingulate cortex.
this region on all of the ADC maps acquired during the C-fos expression was observed in all seven rats of the 60-min ischemic period, regions exhibiting TR-ADC were normothermic group, almost exclusively in regions with readily identified. For example, the ADC maps from 42 to TR-ADC. Three rats exhibited c-fos in only medial frontal 45 min after the onset of ischemia (lower part of Fig. 2A) and cingulate cortex, similar to the results shown in Fig. indicate that TR-ADC occurred in region 1 (defined in Fig. 2A. The remaining four rats exhibited c-fos in a larger 1). The duration of ischemia (min:sec) is shown below region of cortex, which included medial frontal and
each ADC map. cingulate cortex. Two rats with c-fos expression
through-The complete time-course for the ADC lesion size for out cortex also showed c-fos in the lateral basal ganglia. the rat in Fig. 2A (as a percentage of the ischemic The time-course for ADC changes within the c-fos-positive hemisphere) is shown in Fig. 3A. The data indicate that region of two rats is shown in Figs. 4A and B. The much of the expansion of the ADC lesion occurred during ordinate of these graphs represents the percent of the
Table 2
a Average ADC lesion size following 15, 30, 45 and 60 min of ischemia
Time Normothermic Hypothermic P
(min of ischemia) ADC lesion size ADC lesion size
(% of hemisphere) (% of hemisphere)
15 32615 1968 ,0.05
30 38615 19611 ,0.01
45 38619 20610 ,0.01
60 38617 22613 ,0.05
a
Fig. 3. Representative ADC lesion size changes for a normothermic and a hypothermic rat. The top graphs represent the results for the entire hemisphere and the bottom graphs represent the results for ROI-1 (defined in Fig. 1).
c-fos-positive region that exhibited significant ADC reduc-tion. The results in Fig. 4A are for the rat that was shown in Fig. 2A. The graph indicates that three significant peaks in ADC lesion size occurred in the c-fos-positive region. Four significant peaks were observed for the rat repre-sented by Fig. 4B. Two peaks were observed for all of the other rats of the normothermic group except one, which had only one peak.
All seven rats of the normothermic group exhibited hsp70 mRNA expression. Expression was weak in central portions and strong along the periphery of the PR-ADC region. Hsp70 was generally limited to the MCA territory; almost no hsp70 was observed in medial frontal or cingulate cortex where TR-ADC occurred. Strong hsp70 expression in much of the MCA territory was observed in one rat. For this rat, the ADC lesion initially expanded very rapidly to cover a large part of the MCA territory, but subsequently contracted to encompass only about half of the basal ganglia.
3.3. Hypothermic group
ADC, c-fos and hsp70 results for a hypothermic rat are
Fig. 5. Results for a normothermic rat and a hypothermic rat with a large
shown in Fig. 2B. The initial ADC lesion was very small.
expansion in ADC lesion size that reversed spontaneously during early
It expanded gradually to eventually include much of the
ischemia.
basal ganglia and a small portion of cortex. The time-course for the size of the ADC lesion is shown in Fig. 3C.
The five ADC maps at the bottom of Fig. 2B indicate that One normothermic rat also showed a single TR-ADC that TR-ADC did not occur during early ischemia. For the was fairly large (Fig. 5A). C-fos was induced throughout normothermic group, TR-ADC was most commonly ob- the region that exhibited TR-ADC in Fig. 2C. Two typical served during early ischemia and in ROI-1. The time- time-courses for the ADC lesion size within the c-fos-course for ADC changes in ROI-1 is shown in Fig. 3D; no positive region are shown in Figs. 4C and D. Transient significant TR-ADC was observed in this or any other ADC reductions in the hypothermic group were longer ROI. Also, no c-fos or hsp70 expression was detected. Six than they were in the normothermic group (Figs. 4A and other rats of the hypothermic group exhibited no c-fos or B) and did not recur.
hsp70 and almost no TR-ADC. Three of these rats had Moderate hsp70 mRNA expression was observed in a PR-ADC regions that involved all of the basal ganglia and small band of lateral cortex for the rat in Fig. 2C. The faint part of cortex; one had a PR-ADC region that was limited hsp70 expression normally observed in the persistently to the basal ganglia. Three rats had PR-ADC regions that ischemic region of normothermic rats was not observed in covered less than 10% of the ischemic hemisphere, despite this or any other hypothermic rat. Three rats exhibited exhibiting large perfusion deficits. This contrasts sharply moderate hsp70 expression in a thin band along the dorsal with the findings for the normothermic group, where TR- periphery of the PR-ADC region. One rat exhibited strong ADC, c-fos and hsp70 were observed in all seven rats. hsp70 expression in a large part of cortex within the MCA Four rats of the hypothermic group showed a single distribution, where a single transient reduction in ADC incident of TR-ADC at the beginning of the ischemic occurred during the first 10 min of ischemia.
period. Results for one such rat are shown in Fig. 2C. No
significant ADC lesion was observed in the first map. In 3.4. Statistical comparisons between the normothermic subsequent maps (not shown), a small region with reduced and hypothermic groups
ADC appeared in lateral basal ganglia and lateral cortex.
As shown at the bottom of Fig. 2C, this region expanded Mean ADC lesion sizes for the normothermic and abruptly to include most of the dorsal half of cortex, hypothermic groups following 15, 30, 45 and 60 min of including cingulate cortex. It contracted during the next 5 ischemia are summarized in Table 2. The values for the min and eventually was limited to the basal ganglia at the normothermic group were significantly larger than the end of the ischemic period. The complete time-course for corresponding values for the hypothermic group at all four the size of the region with reduced ADC is shown in Fig. time-points.
Table 3
Association of c-fos mRNA expression and transient ADC reduction a
Region Normothermic (n57) Hypothermic (n511)
c-fos TR-ADC c-fos c-fos TR-ADC c-fos
Regions used for this analysis are shown in the anatomical map in Fig. 1.
c-fos expression, based on the regions of interest shown in was significantly higher than the 1.060.0 for the hypo-Fig. 1. For the normothermic group, all seven rats ex- thermic group (P,0.05). For the hypothermic rats, the hibited both c-fos and TR-ADC in region 1. In regions 2 duration of the single ADC lesion size peak was much and 3, the incidence of c-fos expression was lower, but longer than it was for the normothermic group, generally nearly all c-fos-positive regions exhibited TR-ADC. Little lasting for about 10 min.
c-fos or TR-ADC was observed in regions 4 and 5. Of the All seven rats of the normothermic group exhibited total 35 regions of interest for the normothermic group, moderate hsp70 expression in at least some part of the only five exhibited c-fos without TR-ADC or TR-ADC ischemic hemisphere. Only four of the 11 hypothermic rats without c-fos. Statistically, the association between c-fos exhibited any hsp70, which was significantly less than for and TR-ADC was highly significant (P,0.001). For the the normothermic group (P,0.05).
hypothermic group, the frequency of c-fos expression in regions 1, 2 and 3 was less than it was for the
normother-mic group (on a percentage basis). Of the 11 rats, c-fos was 4. Discussion
observed in only four rats for regions 1, 2 and 3. For the
hypothermic group, only six of the 55 regions of interest The most significant findings of this work were that did not show correspondence between c-fos expression and during acute focal cerebral ischemia, mild hypothermia TR-ADC; the association between these two parameters diminished the incidence of transient ADC reduction and was again highly significant (P,0.001). c-fos mRNA expression. TR-ADC was observed much Table 4 summarizes the frequency of TR-ADC in the more frequently in dorsal cortex for the normothermic regions of interest shown in Fig. 1. For region 1, the mean group than it was in the hypothermic group. Also, c-fos-number of incidents of TR-ADC for the normothermic positive regions of rats in the normothermic group had a group was 1.960.6, which is significantly higher than the higher number of significant peaks in ADC lesion size, 0.560.7 for the hypothermic group (P,0.01). The fre- 2.361.0, than the hypothermic group, 1.060.0. Both of quency of TR-ADC for the normothermic group was also these findings support the second hypothesis of this work, higher in regions 2 and 5, but the differences were not that hypothermia decreases the incidence of transient ADC
significant (P.0.05). reduction. C-fos expression was observed in all seven rats
The frequency of transient ADC reduction within the of the normothermic group. For the hypothermic group, a c-fos-positive regions was much lower in the hypothermic markedly reduced incidence of c-fos expression was group than in the normothermic group. For the c-fos- observed (five of 11 rats), which is consistent with the positive regions of the normothermic group, the average third hypothesis. We also previously reported a very high number of peaks in ADC lesion size was 2.361.0, which incidence of c-fos expression during normothermic MCA occlusion [47]. Both TR-ADC [10,43] and c-fos expression [26] are believed to be associated with spreading
depres-Table 4
sion. Thus, the results of this work suggest that mild
Average frequency of occurrence of transient ADC reduction in the five
a hypothermia reduces the incidence of spreading depression
regions of interest shown in Fig. 1
during acute focal ischemia.
Region Normothermic Hypothermic P
Hypothermia has been shown to reduce the incidence of
Frequency Frequency
transient recurring DC potential shifts measured with
of TR-ADC of TR-ADC
cortical electrodes during MCA occlusion [14]. Such shifts
1.0 1.960.6 0.560.7 ,0.01
are believed to be caused by transient cellular
depolariza-2.0 1.361.5 0.560.7 .0.05
tion that occurs during spreading depression. Moderate
3.0 0.460.5 0.460.7 .0.05
4.0 0.060.0 0.260.4 – hypothermia reduced the average number of DC potential
5.0 0.761.3 0.460.5 .0.05 shifts from 3.75 (at 378C) to 2.0 (at 308C). These results a
attenuated changes believed to be associated with spread- infarct size. Conversely, increased incidence of spreading ing depression. However, in a recent study done with depression during MCA occlusion, caused by electrical electrode measurement of extracellular potassium in corti- stimulation, increases neuronal damage and infarct size [2]. cal penumbra, hypothermia did not affect the number of The slowing of ADC lesion expansion by hypothermia peaks in potassium concentration during acute focal is- is also consistent with the first hypothesis of this study and chemia [56]. Such peaks are believed to be caused by with observations described in a recent preliminary report spreading depression. The findings of Sick et al. contrast [67]. In addition to reduced incidence of spreading depres-with those of this work and the work of Chen et al. sion, the slowed rate may have been caused directly by However, the potassium measurements of Sick et al. were reduced rates of ATP depletion or excitotoxin release. only made at a single point in penumbra. Further studies, Consistent with both of these mechanisms, DC potential with electrodes at multiple penumbral cites, would be measurements during acute ischemia have shown that useful for clarifying this discrepancy. hypothermia delays the onset of terminal ischemic
depolar-Spreading depression during acute focal ischemia is ization [5].
initiated by high concentrations of interstitial excitatory In addition to being caused by spreading depression, amino acids and / or potassium [27]. During MCA occlu- some of the transient ADC reductions observed in this sion in rats, measurements with dialysis probes have work may have been caused by transient ischemia. Follow-shown that mild hypothermia significantly reduces intersti- ing the initial blood flow reduction with occlusion, an tial glutamate accumulation in cortical regions of the MCA ischemia-induced increase in collateral blood flow may territory. After approximately 1 h of MCA occlusion, have occurred. For example, the results in Figs. 5A and 5B interstitial glutamate levels in dorsolateral parietal cortex may have been caused by this mechanism. This would were 2.5–3 times higher at 378C than they were at 338C most likely occur in rats with only a limited reduction in [3,64]. A more recent study has confirmed that mild blood flow, during the earliest part of the ischemic period. hypothermia (328C) markedly reduces cortical but not Four hypothermic rats and one normothermic rat exhibited striatal glutamate levels [28]. Similar effects were reported a single TR-ADC during early ischemia. Thus, hypo-for interstitial accumulation of glycine [28]. Thus, the thermia may have preserved energy metabolism suffi-reduced incidence of TR-ADC and c-fos expression ob- ciently to allow improved recovery from the brief is-served in this work may have been caused by diminished chemia. However, not enough rats exhibited these patterns levels of interstitial glutamate in cortical regions. to statistically evaluate this hypothesis. We previously During MCA occlusion, mild hypothermia only slightly reported that blood flow reduction in our model of focal reduces steady state extracellular potassium levels in ischemia can be somewhat variable [48]. The variability is penumbral and core regions [55,56]. In addition, mild likely a consequence of leaving the CCA proximal to the hypothermia does not inhibit the initiation of spreading occluded MCA patent. This model is used in our work depression by cortical application of potassium chloride because it is a better model for clinical stroke, where some [59]. These results suggest that the neuroprotective effect residual blood flow in the ischemic region is usually of hypothermia during focal ischemia is not mediated by present. MCA occlusion models with an intraluminal changes in potassium-induced spreading depression. suture and CCA ligation may give more reproducible Reduced incidence of spreading depression during is- lesion sizes, but they may also be too severe to accurately chemia could be an important component of the neuro- represent the pathophysiological changes that occur clini-protective effect of hypothermia, since spreading depres- cally.
sion is believed to increase the size of the necrotic region Whether caused by spreading depression or transient during focal ischemia [9,58]. Blood perfusion levels in a ischemia, an important finding of this work is that during region depolarized by spreading depression increase in both normothermic and hypothermic acute ischemia, c-fos response to the increased energy demand necessary to expression is associated with TR-ADC. This observation is restore ionic gradients [27]. Spreading depression in the consistent with our previous findings for normothermic rats absence of marked perfusion deficit does not produce subjected to 30 or 60 min MCA occlusion [47].
faster expression may have been the result of faster thermia during focal ischemia may be due to a reduced recovery of intracellular signaling with hypothermia. These incidence of spreading depression. Hypothermia also sig-results indicate that induction of c-fos by transient cellular nificantly reduced the ADC lesion size. Hsp70 expression depolarization is not blocked by hypothermia. In our work, was markedly diminished by hypothermia, indicating that the reduced incidence of c-fos expression may have the neuroprotective affect of hypothermia is not mediated resulted from hypothermic attenuation of the stimulus for by expression of this stress protein.
expression, i.e., spreading depression.
The effect of intra-ischemic hypothermia on HSP70
protein expression has been examined following transient Acknowledgements
forebrain ischemia [40]. Normothermic gerbils subjected to
10-min bilateral CCA occlusion exhibited HSP70 protein This work was supported by grants from the National in hippocampus and neocortex 24 h after reperfusion. With Institute for Neurological Disorders and Stroke, NS22022 moderate hypothermia during ischemia (308C), almost no (PRW), NS14543 (FRS), and NS28167 (FRS). We thank HSP70 was observed. Similarly, rats subjected to 8-min Mr Leo Benjamin and Ms Theresa A. Marsh for perform-bilateral CCA and hypotension at 378C showed strong ing the in situ hybridization studies and Dr Michael F. HSP70 protein expression in hippocampus 48 h after Wendland for assisting with MRI data acquisition. ischemia, but very little HSP70 expression when the
intra-ischemic temperature was reduced to 308C [16]. These
results, and the results of this study, indicate that the References
ischemic stimulus for hsp70 expression is diminished by
hypothermia. Hsp70 induction is believed to be controlled [1] J. Ananthan, A.L. Goldberg, R. Voellmy, Abnormal proteins serve as eukaryotic stress signals and trigger the activation of heat shock
by the formation of denatured proteins [1]. Hence,
hypo-genes, Science 232 (1986) 522–524.
thermia may play an important role in reducing the
[2] T. Back, M.D. Ginsberg, W.D. Dietrich, B.D. Watson, Induction of
denaturation of proteins during ischemia. The results of spreading depression in the ischemic hemisphere following ex-this work also suggest that the neuroprotective effect of perimental middle cerebral artery occlusion: effect on infarct
hypothermia is not related to HSP70 protein expression, at morphology, J. Cereb. Blood Flow Metab. 16 (1996) 202–213. [3] C.J. Baker, A.J. Fiore, V.I. Frazzini, T.F. Choudhri, G.P. Zubay, R.A.
least not during acute focal ischemia.
Solomon, Intraischemic hypothermia decreases the release of
gluta-A limitation of this study is that transient gluta-ADC
reduc-mate in the cores of permanent focal cerebral infarcts, Neurosurgery
tion was generally not observed in 100% of the c-fos- 36 (1995) 994–1001.
positive regions. Three possible reasons for this are: (1) [4] F.C. Barone, G.Z. Feuerstein, R.F. White, Brain cooling during
the temporal resolution of the ADC maps was not suffi- transient focal ischemia provides complete neuroprotection, Neuro-sci. Biobehav. Rev. 21 (1997) 31–44.
cient to consistently capture complete waves of transient
[5] R.D. Bart, S. Takaoka, R.D. Pearlstein, F. Dexter, D.S. Warner,
ADC reduction, (2) partial volume effects may have
Interactions between hypothermia and the latency to ischemic
obscured some of the ADC changes, and (3) an ADC depolarization: implications for neuroprotection, Anesthesiology 88 reduction threshold of 20% was too high to detect all (1998) 1266–1273.
occurrences of transient ADC reduction. Nevertheless, the [6] H. Benveniste, L.W. Hedlund, G.A. Johnson, Mechanism of de-tection of acute cerebral ischemia in rats by diffusion-weighted
results demonstrate a statistically significant association of
magnetic resonance microscopy, Stroke 23 (1992) 746–754.
hypothermia with reduced TR-ADC and c-fos expression.
[7] F. Boris-Moller, T. Wieloch, Changes in the extracellular levels of
The low PCO2values during ischemia for the normother- glutamate and aspartate during ischemia and hypoglycemia. Effects mic group may have resulted in slightly reduced cerebral of hypothermia, Exp. Brain Res. 121 (1998) 277–284.
blood flow levels, relative to the hypothermic group. [8] A. Buchan, W.A. Pulsinelli, Hypothermia but not the N-methyl-D -aspartate antagonist, MK-801, attenuates neuronal damage in gerbils
However, because the degree of hypocapnia was very
subjected to transient global ischemia, J. Neurosci. 10 (1990) 311–
slight (|10 mmHg), the extent of reduction in perfusion
316.
was likely very small [65], especially since isoflurane was [9] E. Busch, M.L. Gyngell, M. Eis, M. Hoehn-Berlage, K.A. Hos-used as an anesthetic [68]. The reduced perfusion rates smann, Potassium-induced cortical spreading depressions during
could potentially reduce the incidence of spreading depres- focal cerebral ischemia in rats: contribution to lesion growth assessed by diffusion-weighted NMR and biochemical imaging, J.
sion in the normothermic group. However, we observed
Cereb. Blood Flow Metab. 16 (1996) 1090–1099.
that the incidence of spreading depression was much
[10] E. Busch, M. Hoehn-Berlage, M. Eis, M.L. Gyngell, K.A.
Hos-higher for the normothermic group than it was for the smann, Simultaneous recording of EEG. DC potential and diffusion-hypothermic group. Thus, the slight hypocapnia probably weighted NMR imaging during potassium induced cortical
spread-did impact our primary conclusions. ing depression in rats, NMR Biomed. 8 (1995) 59–64.
[11] R. Busto, W.D. Dietrich, M.Y. Globus, I. Valdes, P. Scheinberg, M.D.
In summary, the observations presented in this work
Ginsberg, Small differences in intraischemic brain temperature
indicate that mild hypothermia reduced the incidence of
critically determine the extent of ischemic neuronal injury, J. Cereb.
TR-ADC and c-fos expression. Both of these phenomena Blood Flow Metab. 7 (1987) 729–738.
are believed to be associated with spreading depression. [12] R. Busto, M.Y. Globus, W.D. Dietrich, E. Martinez, I. Valdes, M.D.
of neurotransmitters and free fatty acids in rat brain, Stroke 20 [32] A. Kader, M.H. Brisman, N. Maraire, J.T. Huh, R.A. Solomon, The effect of mild hypothermia on permanent focal ischemia in the rat, (1989) 904–910.
Neurosurgery 31 (1992) 1056–1060. [13] A.L. Busza, K.L. Allen, M.D. King, B.N. van, S.R. Williams, D.G.
[33] F. Kamme, K. Campbell, T. Wieloch, Biphasic expression of the fos Gadian, Diffusion-weighted imaging studies of cerebral ischemia in
and jun families of transcription factors following transient forebrain gerbils. Potential relevance to energy failure, Stroke 23 (1992)
ischaemia in the rat. Effect of hypothermia, Eur. J. Neurosci. 7 1602–1612.
(1995) 2007–2016. [14] Q. Chen, M. Chopp, G. Bodzin, H. Chen, Temperature modulation
[34] H. Karibe, J. Chen, G.J. Zarow, S.H. Graham, P.R. Weinstein, of cerebral depolarization during focal cerebral ischemia in rats:
Delayed induction of mild hypothermia to reduce infarct volume correlation with ischemic injury, J. Cereb. Blood Flow Metab. 13
after temporary middle cerebral artery occlusion in rats, J. Neuro-(1993) 389–394.
surg. 80 (1994) 112–119. [15] M. Chopp, H. Chen, M.O. Dereski, J.H. Garcia, Mild hypothermic
[35] H. Karibe, G.J. Zarow, S.H. Graham, P.R. Weinstein, Mild intrais-intervention after graded ischemic stress in rats, Stroke 22 (1991)
chemic hypothermia reduces postischemic hyperperfusion, delayed 37–43.
postischemic hypoperfusion, blood–brain barrier disruption, brain [16] M. Chopp, Y. Li, M.O. Dereski, S.R. Levine, Y. Yoshida, J.H.
edema, and neuronal damage volume after temporary focal cerebral Garcia, Hypothermia reduces 72-kDa heat-shock protein induction
ischemia in rats, J. Cereb. Blood Flow Metab. 14 (1994) 620–627. in rat brain after transient forebrain ischemia, Stroke 23 (1992)
[36] K. Kataoka, H. Yanase, Mild hypothermia–a revived countermea-104–107.
sure against ischemic neuronal damages, Neurosci. Res. 32 (1998) [17] G.L. Clifton, W.C. Taft, R.E. Blair, S.C. Choi, R.J. DeLorenzo,
103–117. Conditions for pharmacologic evaluation in the gerbil model of
[37] H. Kinouchi, F.R. Sharp, M.P. Hill, J. Koistinaho, S.M. Sagar, P.H. forebrain ischemia, Stroke 20 (1989) 1545–1552. Chan, Induction of 70-kDa heat shock protein and hsp70 mRNA [18] K.E. Cullen, K.D. Sarge, Characterization of hypothermia-induced following transient focal cerebral ischemia in the rat, J. Cereb.
cellular stress response in mouse tissues, J. Biol. Chem. 272 (1997) Blood Flow Metab. 13 (1993) 105–115.
1742–1746. [38] K. Kogure, H. Kato, Altered gene expression in cerebral ischemia, [19] T. Curran, J.I. Morgan, Fos: an immediate-early transcription factor Stroke 24 (1993) 2121–2127.
in neurons, J. Neurobiol. 26 (1995) 403–412. [39] K. Kumar, X. Wu, A.T. Evans, Expression of c-fos and fos-B [20] C. Decanniere, S. Eleff, D. Davis, P.C. van Zijl, Correlation of rapid proteins following transient forebrain ischemia: effect of
hypo-changes in the average water diffusion constant and the con- thermia, Mol. Brain Res. 42 (1996) 337–343.
centrations of lactate and ATP breakdown products during global [40] K. Kumar, X. Wu, A.T. Evans, F. Marcoux, The effect of hypo-ischemia in cat brain, Magn. Reson. Med. 34 (1995) 343–352. thermia on induction of heat shock protein (HSP)-72 in ischemic [21] G. Devuyst, J. Bogousslavsky, Clinical trial update: neuroprotection brain, Metab. Brain Dis. 10 (1995) 283–291.
against acute ischaemic stroke, Curr. Opin. Neurol. 12 (1999) [41] A.R. Laptook, R.J. Corbett, R. Sterett, D.K. Burns, D. Garcia, G.
73–79. Tollefsbol, Modest hypothermia provides partial neuroprotection
[22] R. Gill, P. Andine, L. Hillered, L. Persson, H. Hagberg, The effect when used for immediate resuscitation after brain ischemia, Pediat-of MK-801 on cortical spreading depression in the penumbral zone ric Res. 42 (1997) 17–23.
following focal ischaemia in the rat, J. Cereb. Blood Flow Metab. 12 [42] A.R. Laptook, R.J. Corbett, R. Sterett, D. Garcia, G. Tollefsbol, (1992) 371–379. Quantitative relationship between brain temperature and energy [23] M.D. Ginsberg, L.L. Sternau, M.Y. Globus, W.D. Dietrich, R. Busto, utilization rate measured in vivo using 31P and 1H magnetic
Therapeutic modulation of brain temperature: relevance to ischemic resonance spectroscopy, Pediatric Res. 38 (1995) 919–925. brain injury, Cerebrovasc. Brain Met. Rev. 4 (1992) 189–225. [43] L.L. Latour, Y. Hasegawa, J.E. Formato, M. Fisher, C.H. Sotak, [24] Y. Goto, N.F. Kassell, K. Hiramatsu, S.W. Soleau, K.S. Lee, Effects Spreading waves of decreased diffusion coefficient after cortical of intraischemic hypothermia on cerebral damage in a model of stimulation in the rat brain, Magn. Reson. Med. 32 (1994) 189–198. reversible focal ischemia, Neurosurgery 32 (1993) 980–984. [44] D. Le Bihan, E. Breton, D. Lallemand, P. Grenier, E. Cabanis, M. [25] N.G. Harris, E. Zilkha, J. Houseman, M.R. Symms, T.P. Ob- Laval-Jeantet, MR imaging of intravoxel incoherent motions: appli-renovitch, S.R. Williams, The relationship between the apparent cation to diffusion and perfusion in neurologic disorders, Radiology diffusion coefficient measured by magnetic resonance imaging, 161 (1986) 401–407.
anoxic depolarization, and glutamate efflux during experimental [45] J. Maher, V. Hachinski, Hypothermia as a potential treatment for cerebral ischemia, J. Cereb. Blood Flow Metab. 20 (2000) 28–36. cerebral ischemia, Cerebrovasc. Brain Met. Rev. 5 (1993) 277–300. [26] D.G. Herrera, H.A. Robertson, Activation of c-fos in the brain, Prog. [46] C.M. Maier, K. Ahern, M.L. Cheng, J.E. Lee, M.A. Yenari, G.K. Neurobiol. 50 (1996) 83–107. Steinberg, Optimal depth and duration of mild hypothermia in a [27] K.A. Hossmann, Periinfarct depolarizations, Cerebrovasc. Brain focal model of transient cerebral ischemia: effects on neurologic Metab. Rev. 8 (1996) 195–208. outcome, infarct size, apoptosis, and inflammation, Stroke 29 (1998) [28] F.P. Huang, L.F. Zhou, G.Y. Yang, Effects of mild hypothermia on 2171–2180.
the release of regional glutamate and glycine during extended [47] A. Mancuso, N. Derugin, Y. Ono, K. Hara, F.R. Sharp, P.R. transient focal cerebral ischemia in rats, Neurochem. Res. 23 (1998) Weinstein, Transient MRI-detected water apparent diffusion
coeffi-991–996. cient reduction correlates with c-fos mRNA but not hsp70 mRNA
[29] C. Hunt, R.I. Morimoto, Conserved features of eukaryotic hsp70 induction during focal cerebral ischemia in rats, Brain Res. 839 genes revealed by comparison with the nucleotide sequence of (1999) 7–22.
human hsp70, Proc. Natl. Acad. Sci. USA 82 (1985) 6455–6459. [48] A. Mancuso, H. Karibe, W.D. Rooney, G.J. Zarow, S.H. Graham, M.W. Weiner, P.R. Weinstein, Correlation of early reduction in the [30] T. Iijima, G. Mies, K.A. Hossmann, Repeated negative DC
deflec-apparent diffusion coefficient of water with blood flow reduction tions in rat cortex following middle cerebral artery occlusion are
during middle cerebral artery occlusion in rats, Magn. Res. Med. 34 abolished by MK-801: effect on volume of ischemic injury, J.
(1995) 368–377. Cereb. Blood Flow Metab. 12 (1992) 727–733.
[49] S.M. Massa, R.A. Swanson, F.R. Sharp, The stress gene response in [31] Q. Jiang, M. Chopp, Z.G. Zhang, J.A. Helpern, R.J. Ordidge, J.
brain, Cerebrovasc. Brain Metab. Rev. 8 (1996) 95–158. Ewing, P. Jiang, B.A. Marchese, The effect of hypothermia on
[50] D.J. Moyer, F.A. Welsh, E.L. Zager, Spontaneous cerebral hypo-transient focal ischemia in rat brain evaluated by diffusion- and
thermia diminishes focal infarction in rat brain, Stroke 23 (1992) perfusion-weighted NMR imaging, J. Cereb. Blood Flow Metab. 14
[51] K. Nakashima, M.M. Todd, Effects of hypothermia on the rate of magnetic resonance imaging to the study of experimental cerebral excitatory amino acid release after ischemic depolarization, Stroke ischaemia, Cerebrovasc. Brain Metab.Rev. 6 (1994) 180–210. 27 (1996) 913–918. [61] M. Verhaegen, P.A. Iaizzo, M.M. Todd, A comparison of the effects [52] K. Nakashima, M.M. Todd, D.S. Warner, The relation between of hypothermia, pentobarbital, and isoflurane on cerebral energy cerebral metabolic rate and ischemic depolarization. A comparison stores at the time of ischemic depolarization, Anesthesiology 82 of the effects of hypothermia, pentobarbital, and isoflurane, Anes- (1995) 1209–1215.
thesiology 82 (1995) 1199–1208. [62] C.T. Wass, W.L. Lanier, Hypothermia-associated protection from [53] T.R. Ridenour, D.S. Warner, M.M. Todd, A.C. McAllister, Mild ischemic brain injury: implications for patient management, Int.
hypothermia reduces infarct size resulting from temporary but not Anesthesiol. Clin. 34 (1996) 95–111.
permanent focal ischemia in rats, Stroke 23 (1992) 733–738. [63] G.D. Williams, B.J. Dardzinski, A.R. Buckalew, M.B. Smith, [54] M. Schalling, A. Dagerlind, S. Brene, H. Hallman, M. Djurfeldt, H. Modest hypothermia preserves cerebral energy metabolism during Persson, L. Terenius, M. Goldstein, D. Schlesinger, T. Hokfelt, hypoxia–ischemia and correlates with brain damage: a 31P nuclear Coexistence and gene expression of phenylethanolamine N- magnetic resonance study in unanesthetized neonatal rats, Pediatric methyltransferase, tyrosine hydroxylase, and neuropeptide tyrosine Res. 42 (1997) 700–708.
in the rat and bovine adrenal gland: effects of reserpine, Proc. Natl.
[64] C.J. Winfree, C.J. Baker, E.S.J. Connolly, A.J. Fiore, R.A. Solomon, Acad. Sci. USA 85 (1988) 8306–8310.
Mild hypothermia reduces penumbral glutamate levels in the rat [55] T.J. Sick, R. Tang, M.A. Perez-Pinzon, Cerebral blood flow does not
permanent focal cerebral ischemia model, Neurosurgery 38 (1996) mediate the effect of brain temperature on recovery of extracellular
1216–1222. potassium ion activity after transient focal ischemia in the rat, Brain
[65] T. Yamaguchi, F. Regli, A.G. Waltz, Effect of PaCO2 on hyperemia Res. 821 (1999) 400–406.
and ischemia in experimental cerebral infarction, Stroke 2 (1971) [56] T.J. Sick, G. Xu, M.A. Perez-Pinzon, Mild hypothermia improves
139–147. recovery of cortical extracellular potassium ion activity and
ex-[66] H. Yamamoto, A. Mitani, Y. Cui, S. Takechi, J. Irita, T. Suga, T. citability after middle cerebral artery occlusion in the rat, Stroke 30
Arai, K. Kataoka, Neuroprotective effect of mild hypothermia (1999) 2416–2421.
cannot be explained in terms of a reduction of glutamate release [57] B.K. Siesjo, Pathophysiology and treatment of focal cerebral
during ischemia, Neuroscience 91 (1999) 501–509. ischemia. Part I: Pathophysiology, J. Neurosurg. 77 (1992) 169–
[67] M.A. Yenari, D. Onley, M. Hedehus, A. de Crespigny, G.H. Sun, 184.
G.K. Steinberg, M.E. Moseley, Diffusion-weighted MRI of focal [58] K. Takano, L.L. Latour, J.E. Formato, R.A. Carano, K.G. Helmer, Y.
cerebral ischemia and cortical spreading depression under conditions Hasegawa, C.H. Sotak, M. Fisher, The role of spreading depression
of mild hypothermia, Proc. Int. Soc. Magn. Reson. Med. 6 (1998) in focal ischemia evaluated by diffusion mapping, Ann. Neurol. 39
1181. (1996) 308–318.
