Main

A number of biochemical and cellular changes occur in response to an episode of cerebral hypoxia-ischemia, an event occurring in situations such as birth asphyxia. Understanding the cellular changes that evolve is important for determining accurate prognosis and management of the condition and also for aiding in the development of effective therapies. Advances in noninvasive MR techniques have made it possible to repeatedly monitor in vivo some of the anatomical and biochemical changes that occur after an episode of cerebral hypoxia-ischemia. Traditionally, conventional T1-weighted and T2W imaging sequences have been used to detect the size and location of an ischemic insult well after the event (i.e. many hours or days postischemia) (1,2). DW imaging techniques have in general been more sensitive for detecting early cellular injury, with the image changes being apparent minutes after the start of ischemia, and such results have been obtained primarily from studies in adults (2–8). Investigations in neonates have been less frequent, despite cerebral hypoxia-ischemia remaining one of the major causes of death and disability in infants (9–11). Indeed, it is known that during an episode of hypoxia-ischemia, the tolerance of the brain to an ischemic insult and the mechanisms mediating damage depend on the stage of brain development (11–16). Recently, we demonstrated that areas of hyperintensity occurred essentially simultaneously in both DW and T2W images after the start of hypoxia-ischemia in 1-wk-old rats (17). This finding contrasted with reports of changes in DW images preceding changes in T2W images in adults. It was not clear whether this difference was specific to the model of stroke used, which involved a combination of hypoxia and unilateral cerebral ischemia, or whether it was related to age-dependent differences in the response of the brain to a hypoxic-ischemic insult detectable with MR.

The present study was designed to determine whether MR imaging changes in response to an episode of cerebral hypoxia-ischemia differ between immature and older brains. Note that the rat matures substantially between 2 and 4 wk. At 3 wk of age, the rat is weaned. At 2 wk of age, water content in the brain is decreasing, whereas myelin content and the rate of aerobic glycolysis are increasing (11). Indeed, to obtain a comparable area of infarction without excessive mortality using unilateral carotid artery occlusion, the episode of hypoxia must be reduced from 120 min in 1-wk-old animals to 60 and 30 min in 2- and 4-wk-old animals, respectively (18,19). Also, at 4 wk of age, cerebral myelination is considerable and responses to neuroprotective agents no longer resemble the responses seen at 1 or 2 wk of age (11,18). In the present study, T2W, T2 multiecho and DW MR images were acquired in 2- and 4 wk-old rats before, during, and after an episode of cerebral hypoxia-ischemia that was 60 and 30 min in duration, respectively.

METHODS

General animal preparation. All care and use of animals was according to the Canadian Council for Animal Care guidelines, and the study was reviewed and approved by the local animal care committee. Pregnant Wistar rats were ordered (Charles River, Montreal, Canada) approximately one wk before parturition. Animals were maintained on a 12-h light/12-h dark cycle with free access to food and water. Pups were weaned at 3 wk, and experiments were performed on 2- and 4-wk-old rats. On the day of the experiment, each rat was anesthetized with isoflurane (4% for induction and 2.5% for maintenance) and the right common carotid artery was isolated and ligated with 5-0 silk suture as described previously (17). The incision site was infiltrated with 0.25% bupivicaine and sutured closed. After recovery from anesthesia, the animal was returned to its cage for about 60 min before the start of the MR experiment.

Cerebral hypoxia-ischemia in the magnet. The animals were anesthetized by mask with isoflurane (4% induction, 1.25-1.5% maintenance) in 45% oxygen (one third oxygen and two thirds air). The animal's head was held by an incisor bar and a foam head holder to restrict movement. While in the magnet, the respiration rate was monitored with a wire loop designed to detect movement in the magnet resting on the animal's back. Rectal and chamber temperatures were monitored continuously and rectal temperature was maintained at 37.5-38°C with a circulating water blanket. Rectal temperature near or slightly below 37°C was associated with either no damage or only minor damage in the 2-wk-old animals. Hypoxia was initiated by changing the anesthetic carrier gas to 8% humidified oxygen with a balance of nitrogen. Isoflurane was reduced to 0.75% 5 min after the start of hypoxia and to 0.5% if the respiration rate decreased. The duration of hypoxia was generally 60 min for 2-wk-old and 30 min for 4-wk-old rats. On discontinuing hypoxia, animals were again supplied with 45% oxygen carrying 0.75-1.25% isoflurane.

T2W and DW MR images were obtained before, during, and up to 1 h after hypoxia-ischemia. After the last image, anesthesia was discontinued and the animals were returned to their cages. Twenty-four hours after the ischemic insult, the animal was again anesthetized with 1.25-1.5% isoflurane and T2W and DW images were acquired. This procedure was repeated at 48 h posthypoxia in the 2-wk-old rats.

MR imaging techniques. MR experiments were performed in a 9.4 T/21-cm horizontal bore magnet (Magnex, UK) equipped with an MSLX Bruker console (Germany). T2W and DW images were acquired with a quadrature coil tuned to 400.045 MHz. T2W images were acquired with a multislice single echo sequence (TE = 80 ms; TR = 1200 ms) and DW images with a Stejskal-Tanner spin echo sequence (TE = 30 ms, TR = 1200 ms) (20). Two diffusion-sensitive gradient pulses of 10 ms duration and 85 mT/m amplitude were separated by 25 ms, which resulted in a gradient factor of 1062 s/mm2. The diffusion gradient pulses were applied in the phase-encoding direction (Y). Eight slices 1.5 mm thick were acquired with both sequences. The field of view was 3 cm, and a 256 × 128 data matrix was acquired in both MR imaging experiments. Two averages per phase-encoding step resulted in a total acquisition time of approximately 6 min for each experiment. T2 maps were calculated from data acquired with a multislice, multiecho sequence (4 slices, 6 echoes, TE = 21.6 ms) in animals examined in the latter half of the study.

Histology. After the last MR scan at 24 or 48 h posthypoxia-ischemia, the animal was anesthetized with pentobarbital (80 mg/kg) and perfusion-fixed with either 10% buffered formalin or a mixture of 2% paraformaldehyde and 1% glutaraldehyde. The brain was removed and carefully inspected to identify the extent of gross infarction. The brain was immersed in fixative, and coronal slices were subsequently embedded in paraffin and sectioned. Sections were stained with hematoxylin and eosin, and regions of infarction or selective necrosis were indicated by the presence of pyknotic nuclei; or eosinophilic cells were marked on schematic drawings of the sections and compared directly with MR images of the corresponding slices. Areas of damage were not quantified, because both infarction and selective necrosis were observed frequently, whereas delineation of the latter is somewhat subjective and inaccurate.

Data analysis. Data were grouped according to the following time intervals: prehypoxia, every 10 min during the hypoxia, every 10 min after hypoxia, and 24 and 48 h posthypoxia. The areas of increased intensity in DW and T2W images were measured in coronal slices containing the anterior caudate or midthalamus, using software developed at the Institute for Biodiagnostics. T2 and the intensity within DW images were quantified by taking measurements in both the ipsilateral and contralateral parietal cortex at the levels of the anterior caudate and the midthalamus. The Newman-Keuls test was used for comparison of means obtained from repeated measurements at different time points. The paired t test and the Mann-Whitney U test used to compare means of values obtained from the ipsilateral and contralateral hemispheres. The t test was used to compare data from 2- and 4-wk old animals. Data are presented as mean ± SE and differences were considered significant at p < 0.05.

RESULTS

MR imaging in 2-wk-old rats. MR imaging experiments were performed on fourteen 2-wk-old animals with a mean body weight of 28.0 ± 0.8 g. One of the 2-wk-old animals died during hypoxia, and data from this animal were not included in the quantitative analysis. Inspection of the images showed that before the start of hypoxia, there were no detectable differences between the ipsilateral and contralateral hemispheres in either the DW or T2 images (Fig. 1, A and D). The first MR image change appeared as an area of increased brightness usually 10-20 min after the start of hypoxia (Fig. 1, B and E). The area of hyperintensity in DW images first appeared in the hemisphere to a portion of the contralateral hemisphere in 9 of 13 animals. The onset of the first appearance of hyperintense areas in the T2W images was similar to that in the DW images. Furthermore, the area of hyperintensity in DW and T2W images increased during hypoxia (Fig. 1, C and F). Once hypoxia was terminated, hyperintensity in the DW images diminished and became undetectable within 60 min in the majority of animals (9 of 13) (Fig. 2A). At 24 h posthypoxia, hyperintense areas in the ipsilateral hemisphere reappeared in both T2W and DW images, although these hyperintense areas were usually less extensive than those that occurred during hypoxia (Fig. 2, B and E). The areas of hyperintensity persisted and intensified at 48 h (Fig. 2, C and F).

Figure 1
figure 1

MR images of coronal slices through the thalamus from a 2-wk-old rat immediately before and during hypoxia-ischemia. Alternating DW images (upper images, A-C) and T2W images (lower panel, D-F) were acquired. An area of hyperintensity in the DW images was distributed in the entire ipsilateral (right) hemisphere and a part of the contralateral (left) cortex 10 min after the start of hypoxia (B). In contrast, the hyperintense area in T2W images was restricted to the ipsilateral hemisphere (D), and this brightness diminished later in hypoxia-ischemia (F).

Full size image
Figure 2
figure 2

MR images in the same 2-wk-old rat at 1, 24, and 48 h after hypoxia-ischemia. These DW (upper panel, A-C) and T2W images (lower panel, D-F) show that the area of hyperintensity become undetectable after 1 h of recovery (A and D) and reappear in the ipsilateral (right) hemisphere 24 h later (B and E).

Full size image

The area of increased intensity in DW and T2 image was analyzed quantitatively in coronal slices through the midthalamus or anterior striatum (Fig. 3). Significant increases in the area of hyperintensity occurred by 10-20 min into hypoxiaischemia. In sections through the midthalamus, the area of DW hyperintensity occupied approximately 60% of the total brain area. The maximal area of hyperintensity in the T2W images was substantially less, reading approximately 30% (p < 0.003). Similar results were obtained from a quantitative analysis of coronal slices through the anterior caudate.

Figure 3
figure 3

Areas of hyperintensity in MR images of 2-wk-old rats during and after hypoxia-ischemia. The areas were measured in coronal slices at the anterior caudate level and the midthalamus level. Data are presented as a percentage of the entire brain area of that slice (mean ± SE) at time points either minutes shortly after recovery (Rec) from hypoxia-ischemia or hours post-hypoxia-ischemia (Post). *p < 0.05 and **p < 0.01; significant difference between values in DW and T2W image. +p < 0.05 and ++p < 0.01; different from analogous prehypoxia value (0 min).

Full size image

MR imaging 4-wk-old rats. MR imaging experiments were performed in fifteen 4-wk-old animals with a mean body weight of 92.6 ± 2.7 g. Four of these animals died during hypoxia, and their data were not included in the quantitative analysis. Before the start of the hypoxia, there was no ipsilateral-contralateral differences on the MR images (Fig. 4, A and E). In the DW images, areas of hyperintensity appeared in the ipsilateral hemisphere approximately 10 min after the start of hypoxia and remained ipsilateral except in 2 of 11 rats that exhibited contralateral changes (e.g. Fig 4b). The T2W images showed no evidence of increased signal intensity throughout hypoxia in nine of the animals. (e.g. Fig. 4F), and only minor changes in the other two animals. The areas of hyperintensity in DW images faded rapidly to small areas of hyperintensity after the termination of hypoxia (e.g. Fig. 4C) and became undetectable by 30 min posthypoxia in eight of the 11 animals. Twenty-four hours posthypoxia, we detected hyperintense areas in both DW and T2 images (e.g. Fig. 4, D and H). Two animals were also examined 4 and 6 h after the termination of hypoxia, and their images showed small areas of hyperintensity in both DW and T2W images in the ipsilateral cortex and thalamus.

Figure 4
figure 4

DW images (upper panel, A-D) and T2W images (lower panel, E-H) of brain slices through the thalamus from a 4-wk-old rat before, during, and after hypoxia. A bright area in the hemisphere ipsilateral to the occlusion was observed during hypoxia-ischemia in DW but not T2W images (B and F).

Full size image

Quantitative analysis of the areas of hyperintensity in the DW images demonstrated significant areas of the increased brightness by 10 min after the start of hypoxia-ischemia (Fig. 5). The area of DW hyperintensity peaked at 30-35% of the total brain area toward the end of hypoxia-ischemia. In contrast, areas of hyperintensity in the T2W images were minimal (p < 0.01). Both were significantly less than the areas of hyperintensity observed in 2-wk-old animals (p < 0.003 for DW and T2W images, respectively).

Figure 5
figure 5

Mean areas of hyperintensity in MR images of 4-wk-old rats during and after hypoxia-ischemia. The areas were measured in coronal slices at the anterior caudate level (Ant; upper panel) and the midthalamus level (Mid; lower panel). Data are presented as a percentage of the entire brain area of that slice (mean ± SE) for each time point, either during hypoxia-ischemia, minutes after recovery from hypoxia-ischemia, or hours post-hypoxia-ischemia (Post). *p < 0.05 and **p < 0.01; significant difference between values in DW and T2W images. +p < 0.05 and ++p < 0.01; different from analogous prehypoxia value (0 min).

Full size image

Hypoxic-ischemic changes in T2 and DW intensity ratios. Quantitative T2 data were obtained from five of the 2-wk-old animals and nine of the 4-wk-old animals (Fig. 6). In 2-wk-old animals, T2 in the parietal cortex of the level of the midthalamus increased by 10% during hypoxia and then returned to the prehypoxic level soon after reoxygenation and normalization of blood flow. T2 was markedly increased again 24 h posthypoxia. In contrast, in 4-wk-old rats, T2 did not change significantly during hypoxia (p < 0.02, different from 2-wk-olds), whereas a marked increase was observed 24 h posthypoxia. The age-dependent decrease in the initial prehypoxia T2 in 2-wk-old compared with 4-wk-old animals is qualitatively similar to the developmental decrease in T2 reported in dog and monkey, reflecting ontogenic decreases in brain water content (21).

Figure 6
figure 6

Mean T2 in 2-wk-old (a and b) and 4-wk-old (c and d) rats from images containing the anterior caudate (Ant, a and c) and midthalamus (Mid, b and d). Mean value ± SE in the ipsilateral (Ipsi) and contralateral (Cont) hemispheres were measured. There was a marked ipsilateral increase in T2 during hypoxia-ischemia in 2-wk-old but not 4-wk-old animals. Minutes into recovery from hypoxia-ischemia (Rec), there was a return in T2 toward prehypoxic values, with an increase again hours post-hypoxia-ischemia (Post). *p < 0.05 and **p < 0.01; ipsilateral different from contralateral. +p < 0.05 and ++p < 0.01; different from analogous prehypoxia value (0 min).

Full size image

DW images were also assessed quantitatively by measuring the ipsilateral: contralateral intensity ratio within the parietal cortex in animals for which we had T2 values (Fig. 7). The pattern of change over time of the DW intensity ratios was equivalent in 2- and 4-wk-old animals. Image intensity ratios were increased during hypoxia, declined after hypoxia, and increased again 24 h posthypoxia in both 2- and 4-wk-old animals (p < 0.4 for all time points, 2- versus 4-wk-old groups).

Figure 7
figure 7

Mean DW intensity ratios at various time points in 2-wk-old (upper panel) and 4-wk-old rats (lower panel). The mean ratio of the intensity within the brightest area in the ipsilateral (I) parietal cortex to that in the contralateral (C) hemisphere is presented for coronal slices of the anterior caudate (Ant) and the midthalamus (Mid) (mean ± SE). Data were analyzed at various time points during and after recovery (Rec) from hypoxia-ischemia and at various hours post-hypoxia (Post). The changes of the university ratios during and after hypoxia-ischemia display a similar pattern in 2- and 4-wk-old rats. *p < 0.05 and **p < 0.01; different from analogous prehypoxia value (0 min).

Full size image

Histology. Inspection of the hematoxylin and eosin-stained brain sections from the 2-wk-old animals revealed areas of necrosis in the striatum, cortex, hippocampus, and thalamus within the hemisphere ipsilateral to the occlusion in all animals. Areas of pan-necrosis occurred in 6 or 13 animals, and areas of laminar and selective neuronal necrosis occurred also in 12 of 13 animals. The distribution of areas of hyperintensity on the final T2W images corresponded well with regions of pan-necrosis or laminar necrosis. Of the nine animals that had bilateral MR intensity changes during hypoxia with recovery afterward, only three demonstrated regions of necrosis in the contralateral cingulate cortex and/or hippocampus. Corresponding regions of increased intensity in these areas were also detectable in the final MR images. In only two animals, areas of brain with eosinophilic neurons in the striatum and cortex were not accompanied by MR imaging changes.

In the 4-wk-old rats killed 24 h after hypoxia-ischemia, ipsilateral infarction was apparent in 8 of 9 animals, and areas of laminar or selective neuronal necrosis were also apparent in 5 of 9 animals. Combined areas of infarction and necrosis corresponded well to the areas of hyperintensity observed on the final T2W and DW images. The animal killed at 4 h posthypoxia had dark shrunken and eosinophilic neurons within the cortex and thalamus and slight MR changes in these same regions. The animal killed at 6 h posthypoxia had cells with pyknotic nuclei and areas of MR hyperintensity in the ipsilateral cortex, hippocampus, thalamus, and striatum. Eosinophilic or selective necrotic cells were also detected in the contralateral cerebral cortex or hippocampus in 5 of 11 animals. Corresponding MR changes in the T2W images were apparent in these regions, except for four animals with eosinophilic cells. Overall, the final area of damage was less than the extent of the increased DW intensity changes seen during hypoxia for the 2-wk-old but not the 4-wk-old animals.

DISCUSSION

This study demonstrates that during and after a transient episode of cerebral hypoxia-ischemia, there are characteristic changes in MR images, some of which are dependent on the age of the animal. The effects of transient hypoxia-ischemia on DW images were qualitatively similar, irrespective of age. Marked increases in DW intensity occurred ipsilateral to the occlusion during hypoxia-ischemia. This occurrence was followed by substantial recovery and then secondary increases in intensity post-hypoxia-ischemia in both age groups. In contrast, the changes in T2W images differed substantially between the age groups. During cerebral hypoxia-ischemia there was an increase in T2 values in 2-wk-old but not 4-wk-old rats. Similar to DW, delayed or secondary increases in T2 occurred hours later.

Onset of DW imaging changes. The onset of hyperintense changes in the DW images after the initiation of hypoxia-ischemia appears to be dependent on age. The delay of 10-20 min posthypoxia before there are imaging changes in 2- and 4-wk-olds is much more rapid than the average 45-min delay observed in 1-wk-old rats (p < 0.001) (17). Furthermore, DW imaging changes occur only minutes after the onset of hypoxia-ischemia in adults rats (5,22). The relative delay in onset of DW hyperintensity changes in 1-wk-old animals may reflect a lower rate of oxidative metabolism and a longer time before depletion of ATP in neonates compared with older animals (23–25). This possibility is supported by the fact that the onset of DW changes has been related to ATP reductions, cell depolarization, and an alteration in ion balance, with a shift of water from the extracellular to intracellular space (2,26).

Distribution and reversibility of the imaging changes. In 2-wk-old animals, the area of DW hyperintensity during hypoxia-ischemia exceeded the subsequent area of imaging changes 24-48 h later, whereas the latter was associated with regions of infarction and selective cell death. DW changes exceeding areas of subsequent infarction have been observed previously in 1-wk-old rats and in some studies of transient cerebral ischemia in adults (5,17,27). Such a response was not apparent in our 4-wk-old animals, in which the maximal areas of DW hyperintensity changes were similar to the imaging changes 24 h posthypoxia. Indeed, areas of DW hyperintensity were greater in 2- than 4-wk-old animals, and this difference could not be attributed to the longer episode of hypoxia-ischemia, because it was apparent even 20 min after the start of hypoxia.

The extensive distribution of DW imaging changes in the 2-wk-old animals indicates that at least some of the cellular and extracellular changes detectable on DW images are reversible and not associated with permanent cell injury. Delayed cell death or secondary cell injury also appears to occur in some of the regions, which transiently reverse their imaging changes but then are found to contain areas of selective neuronal necrosis or eosinophilic cells 24 to 48 h later. Thus, simply a change in intensity of DW images cannot predict permanent cell damage, and a reversal of these changes after ischemia cannot predict permanent recovery. It may be that values exceeding a threshold decrease in the apparent diffusion coefficient for water will be able to better predict outcome after a hypoxic-ischemic episode (28,29). Alternatively, it may be that a knowledge of the apparent diffusion coefficient (ADC) and its duration of reduction will be required to predict whether ischemic tissue is irreversibly injured (30). Our studies indicate that such determinants will also be dependent on the age of the subject.

The area of T2W hyperintensity during hypoxia-ischemia was generally less than that of DW hyperintensity but similar to that of areas of T2W hyperintensity 24-48 h later. This finding indicates that an increase in T2 either during or after hypoxia-ischemia is prognostic of poor outcome for that tissue and thus is consistent with the concept that areas of increased T2 correlate well with areas of infarction (17). After an episode of transient hypoxia-ischemia, there may be an interval for which neither T2W nor DW imaging provides a good indication of the ultimate outcome of the tissue, because both can recover temporarily. Tissues that do not recover go on to infarction and are generally located in regions with severe levels of ischemia (17,29,31).

Delayed secondary changes. Several studies of cerebral hypoxia-ischemia in immature animals have demonstrated a recovery of cell function after hypoxia-ischemia, with a secondary deterioration hours later. For example, in 1 and 2-wk-old rats, there is a normalization of pH and a partial or full recovery of high-energy phosphate metabolism and a subsequent decrease again by 15-25 h (32,33). In fetal sheep, there is a transient increase in cortical impedance, reflecting a loss of cellular ionic homeostasis during cerebral ischemia, a recovery postischemia, and a delayed increase again starting approximately 17 h post-ischemia (34). Thus, the recovery of the DW and T2 imaging changes observed in our animals likely corresponds to a recovery in ion homeostasis associated with a normalization of ATP soon after the cessation of hypoxia-ischemia. This transient recovery did not appear to differ among different ages, and similar results with temporary cerebral ischemia have been obtained in adult rats (35,36).

Changes in T2W imaging during hypoxia are age dependent. There was an early increase in intensity during hypoxia in the T2W images of 2-wk-old rats similar to that obtained previously in 1-wk-old rats (17). An early increase was not seen in 4-wk-old rats. This finding is similar to the 3-9 h delay for T2W compared with DW changes in adult rats, cats, or humans after middle cerebral artery occlusion (2,4,7,22,37). The biochemical or anatomical differences that account for differing changes in T2 during hypoxia-ischemia remain speculative. Conventionally, it is believed that the T2 of water reflects primarily the free water content of the brain, that is, the unbound water located primarily in the extracellular rather than intracellular space. An increase in brain water content or a vasogenic edema is generally considered to be the main cause of an increase in T2 (2,38,39). However, changes in protein concentration, which influence the amount of free water, are also considered to influence T2 (39–41). Thus, the MR results suggest that there is either an absolute increase in brain water or an alteration in brain protein occurring during hypoxia-ischemia in young and not older animals. For example, it is possible that there are ontogenic differences in blood-brain barrier function or systems involved in regulating brain water, such as Na+K+-ATPase (1,42–44). Alternatively, aquaporin 4, a water channel protein found in the membrane of glial cells bordering the subarachnoid space, ventricles, and blood vessels, may be more abundant or differentially regulated in the plasma membrane of immature and mature brain (45–47).

To summarize, T2W and DW imaging changes occurring during and after transient cerebral hypoxia-ischemia differ in immature and older animals (Table 1). DW changes during hypoxia-ischemia are more extensive in 2-wk-old than 4-wk-old animals, and there are cellular responses to hypoxia-ischemia that are detectable with T2W imaging techniques in immature but not older brain. Understanding the biochemical or anatomic correlates of these imaging changes will be important for the application of these MR imaging techniques to the diagnosis and prognosis of hypoxic-ischemic brain damage. Additional clinical investigations are indicated to establish whether there are similar differences in human infants.

Table 1 Summary of changes in MR images 1-, 2-, and 4-week-old rats subjected to cerebral hypoxia-ischemia
Full size table