It is generally accepted that reducing the body temperature to 33-34°C is neuroprotective against cerebral ischemic insults without causing many side effects
[9, 28, 29]. Here, we show that, in the Et-1 model, a short mild hypothermic treatment can be delayed up to 1 hour after stroke onset without losing its beneficial effect. Although the reduction in neurological and functional deficit and apoptosis was no longer observed after delaying hypothermia for 2 hours, this study clearly shows a lasting inhibitory effect on activation of glia. Laser-doppler flowmetry experiments in the striatum showed no influence of cooling on the MCA occlusion caused by Et-1. Finally, we show a clear correlation between neurological and functional outcome with apoptosis but not with inflammation.
The STAIR criteria suggest to reproduce data in as many experimental stroke models as possible, especially in models mimicking hospital settings. In the Et-1 model, as reperfusion is only established slowly, it resembles clinical reality closely and thus data could be extrapolated to the clinic
. Besides, due to the short half-life of Et-1 (1.4 to 3.6 minutes)
, infusion of this vasoconstrictor leads to an occlusion (>75%) of the MCA for 20–30 minutes, before a gradual reperfusion
. Afterwards the blood flow is restored. As such, a reproducible infarct with a clear distinction between the core (striatum) and the penumbra (cortex) of the insult is established and it becomes possible to distinguish certain effects solely to the core or penumbral area
[21, 37]. Our study showed that cooling had no influence on the MCAO caused by Et-1. This result indicates that the effects of hypothermia are not mediated by modulation of the CBF.
Since cells in the penumbra undergoing apoptosis are major targets for intervention
, we assessed the number of cells containing active caspase-3. Previously, we demonstrated that apoptosis in the penumbra occurs in the subacute phase and peaks 24 hours after the insult
. Chaitanya et al.
 showed that 3 hours MCAO induces apoptosis, reaching maximal levels at 24 and 72 hours after the insult
. Another study investigated the effect of 20 minutes and 2 hours MCAO on apoptosis. They observed that caspase-3 activity peaks at 24 and 72 hours after reperfusion in the 20 minutes MCAO group, whereas, it only peaks at 24 hours after 2 hours MCAO
. These results are consistent with our data. Previous research in our laboratory also investigated nuclear fragmentation in neuronal cells and showed 24 hours after the insult that mild hypothermia significantly affected apoptotic neuronal cell death in the penumbral region, consistent with the effects observed here on activated caspase-3 expression
. Phanithi et al.
 showed that intra-ischemic mild hypothermia inhibits the caspase-3 expression in the penumbra, at 24 hours, after 1 hour of focal ischemia
. Maier et al.
 confirmed that 1 and 2 hours of intra-ischemic mild hypothermia was effective to reduce apoptosis at 72 hours after the insult
. Similar results on infarct volume have been reported. For instance, Maier et al.
 showed, 3 days after the insult, that 2 hours of hypothermia reduced the infarct volume, even when delayed up to 90 minutes after 2 hours MCAO
. In the same MCAO model, 4 hours of post-ischemic hypothermia (started 4 hours after ischemia onset), could no longer protect the rat brain
[48, 49]. However, a study by Ohta
 showed that the hypothermic treatment could be delayed up to 4 hours after 2 hours MCAO if cooling was prolonged for 48 hours
. However, serious side effects are then to be expected, but were not investigated.
In our study, the number of apoptotic cells correlates well with infarct volume (r2 = 0.96) and with neurological deficit (r2 = 0.99). This observation is in accordance with the general idea that cell death in the penumbra, which is the predominant area that is salvaged by hypothermia, occurs through apoptosis
[19–21, 44]. It is also coherent with our hypothesis that the beneficial effects of short hypothermic treatment on infarct volume and neurological deficit are predominantly mediated by inhibition of apoptosis in the penumbra, at least as established at 24 hours after the insult.
After infusion of Et-1, CD-68 expression increased in the striatum and the cortex of the insult after 24 hours. These results are consistent with previous findings in the Et-1 model
 and other MCAO models
[51–53]. The hypothermic treatment significantly attenuated this increase by approximately 50%, when initiated 20 minutes after the infusion of Et-1. Similar results were obtained in studies investigating the effect of 2 hours of intra-ischemic mild hypothermic treatment after 2 hours MCAO. They showed a reduction of the infarct volume and CD-68 expression at 1, 3 and 7 days after the insult
[15, 16, 22]. Similar results were also observed after MCAO of 8 minutes
. Our study further showed that delaying hypothermia for 1 or 2 hours could still reduce the amount of phagocytic cells in the core of the insult. Surprisingly, this observation did not correlate with decreased neurological or functional deficit. This could mean that a reduction in phagocytosis is not instrumental to the neuroprotective effects of hypothermia. However, it is conceivable that inhibition of microglial activation and reduced infiltration of monocytes within the first 24 hours attenuates the induction of cell death between 24 and 72 hours after the insult.
GFAP is considered the best marker to investigate reactive gliosis
. Stroke induces a massive increase in GFAP expression after a day, in several stroke models ranging from 30 minutes MCAO to the photothrombosis model
[56, 57]. Our study confirms these results in the Et-1 model, in the core as well as in the penumbra. Our major finding is that short hypothermic treatment reduced the expression of GFAP in the striatum and cortex. More specifically, Zoli et al.
 described peak levels in astrogliosis 24 hours after the insult, especially in the penumbral area
. Our results are consistent with these findings and might explain why the effect of hypothermia on GFAP expression was less pronounced compared to that in the core of the insult. Although the strongest effects were found when hypothermia was initiated after 20 minutes, delaying hypothermia up to 2 hours after the administration of Et-1 still reduces GFAP expression. Further research is necessary to clarify whether delaying short hypothermic treatment for 2 hours (or more) provides neuroprotection at later time points after the insult. However, it is also possible that increased GFAP expression after cerebral ischemia is not entirely detrimental. Indeed, GFAP null mice showed more sensitivity to ischemia compared to wild type mice
This study confirms that short hypothermic treatment can be protective when it is applied within a relatively short therapeutic window up to 1 h. It is known that, when the treatment is delayed, long cooling periods seem more protective
[34, 35]. In a balance of risk and benefit, a short duration of hypothermia may be the initial choice. The hypothermia treatment protocol should be tailored to each patient's situation, depending on the time after the insult and whether the patient is in the intra-ischemic or in the post-ischemic period. For instance, shorter cooling periods can be induced when patients arrive very quickly in the hospital after the onset of the insult, while longer cooling periods can be applied to patients that arrive later to the hospital. Also co-morbidities should be taken into account when conducting experimental or clinical studies
. Based on this assumption and the results of the available data, a larger randomized, controlled clinical trial of hypothermia in acute ischemic stroke is warranted.