Bid-mediated Mitochondrial Pathway Is Critical to Ischemic Neuronal Apoptosis and Focal Cerebral Ischemia*

We have investigated the role of the BH3-only pro-death Bcl-2 family protein, Bid, in ischemic neuronal death in a murine focal cerebral ischemia model. Wild-type andbid-deficient mice of inbred C57BL/6 background were subjected to 90-min ischemia induced by left middle cerebral artery occlusion followed by 72-h reperfusion. The volume of ischemic infarct was significantly smaller in the bid-deficient brains than in the wild-type brains, suggesting that Bid participated in the ischemic neuronal death. Indeed, following the ischemic treatment there was a significant reduction of apoptosis in the ischemic areas, particularly in the inner infarct border zone (the penumbra), of thebid-deficient brains. In addition, activation of Bid in the wild-type brains could be readily detected at ∼3 h after ischemia, as evidenced by its proteolytic cleavage and translocation to the mitochondria as determined using Western blot analysis and immunofluorescence staining. Correspondingly, mitochondrial release of cytochrome c could be detected around the same time Bid was cleaved in the wild-type brains. However, no significant cytochromec release was detected in the bid-deficient brains until 24 h later. This suggests that, although the mitochondrial apoptosis pathway might be activated by multiple mechanisms during focal cerebral ischemia, Bid is critical to its early activation. This notion was further supported by the finding that caspase-3 activation was severely impaired in thebid-deficient brains, whereas activation of caspase-8 was much less affected. Taken together, these data suggest that Bid is activated early in neuronal ischemia in a caspase-8-dependent fashion and that Bid is perhaps one of the earliest and most potent activators of the mitochondrial apoptosis pathway. Thus, the role of Bid in the induction of ischemic neuronal death may render this molecule an attractive target for future therapeutic intervention.

Ischemic stroke is caused by cerebral vascular compromise. Vascular occlusion or disruption leads to brain ischemia and hypoxia that can result in serious neuronal injury. Although damaged neurons often die from necrosis, a significant amount of neurons may die from apoptosis. Most of the evidence for the role of apoptosis in ischemic neuronal death is obtained with the use of rodent models of global or focal ischemia.
The connection between ischemic neuronal death and caspase activity has been well established. Increased levels of either expression or activities of effector caspase-3 can be detected in ischemic neurons (1)(2)(3). Cleavage of caspase substrates can also be clearly demonstrated (1,4). DNA fragmentation in ischemic brains, usually induced by caspase activities (5,6), has been demonstrated by many investigators, by either gel electrophoresis or TUNEL 1 staining (1,7,8). A strong correlation has been found among caspase-3 activity, TUNEL staining, and appearance of apoptotic neurons (1,9,10). In addition, administration of caspase inhibitors during the induction of ischemia suppresses the activity of effector caspases and decreases the number of TUNEL-positive cells and apoptotic bodies (1,8). Inhibition of caspases also reduces the infarct size significantly in transient focal ischemia (11)(12)(13) and rescues susceptible CA1 neurons in a global ischemia model (1).
One key question is how the apoptotic cell death mechanism is activated during neuronal ischemia. In non-neuronal systems, two apoptotic pathways have been defined (14 -16): the extrinsic pathway, mediated by the death receptor family proteins; and the intrinsic pathway, mediated by the mitochondria. Current evidence suggests that, in neuronal ischemic death, both pathways may be activated.
A key step in the intrinsic pathway, mitochondrial release of cytochrome c, has been reported in ischemic neurons after global or focal ischemia (17,18). Released cytochrome c, together with Apaf-1 and dATP, can activate the initiator caspase, caspase-9. Caspase-9 has been reported to be present in mitochondria and is released concurrently with cytochrome c (19,20). Activation of caspase-9 has recently been found in the degenerating CA1 neurons after global ischemia (21).
The expression of Fas, normally in non-neuronal cells or in neurons at low levels, is significantly up-regulated in neurons during focal ischemia based on in situ hybridization and/or immunohistochemistry studies (22)(23)(24)(25)(26). More importantly, focal ischemia-induced infarction is greatly reduced in mice expressing mutated nonfunctional Fas (Fas lpr mice) (26,27). In vitro studies have also demonstrated a susceptibility of primary neurons to FasL-induced apoptosis (23,27). Correspond-ingly, the initiator caspase-8 downstream of the death receptor was found to be activated in ischemic neurons (28).
Caspase-8 and Fas play a role in the activation of Bid, a "BH3-only" pro-apoptosis Bcl-2 family protein (29). Death signals mediated by the death receptors, i.e. Fas and tumor necrosis factor receptor-1, can activate Bid through proteolytic cleavage by caspase-8 (30 -32). The p15 truncated Bid (tBid) is able to translocate to mitochondria and induces cytochrome c release. Bid thus has the capability to communicate death signals from the extrinsic pathway to the intrinsic pathway. The significance of this communication has been demonstrated in a murine model of hepatocyte apoptosis triggered by receptor activation (33)(34)(35).
Although the role of Bid in neuronal death has not been well studied, the involvement of both apoptotic pathways in ischemic neuronal apoptosis prompted us to test the hypothesis that Bid is critical to the development of ischemia-induced neuronal apoptotic death. Bid may accomplish this by relaying death receptor signals to mitochondria, where it induces cytochrome c release, which then activates downstream effector caspases, leading to apoptosis. This hypothesis is tested in the current study. We found that, in a murine focal ischemia model created by middle cerebral artery occlusion, Bid was cleaved and translocated to mitochondria through caspase-8 activation. In the absence of Bid, cytochrome c release in the cortex neurons was reduced and the activation of effector caspase-3 was delayed and reduced. Furthermore, bid-deficient mice demonstrated significant resistance to ischemic neuronal death, as evidenced by minor infarction and reduced apoptosis in the cortex. Thus, our study clearly indicates that Bid is activated and contributes prominently to the development of ischemic neuronal apoptosis.
Mice-Mice deficient in bid were established by gene targeting and had been back-crossed to the C57BL/6 background for 12 generations (33,34). Wild-type and bid-deficient mice used in this study were male littermates ϳ25-35 g in weight. They were maintained as homozygotes in a specific pathogen-free facility on the campus in compliance with the National Institutes of Health and University of Pittsburgh policies.
Evaluation of Mouse Cerebrovascular Anatomy-Variations in the cerebral circulation anatomy are an important factor that affects the lesion sizes in rodent models of cerebral ischemia. In this study, we quantitatively evaluated the cerebrovascular anatomy in bid-deficient mice and their littermates as described previously (36,37). Mice were sacrificed in a CO 2 chamber. Transcardial perfusion fixation was performed immediately after death via the left ventricle with heparinized saline (10 units/ml) followed by warm formalin. Evans Blue, 2%, mixed equally with 20% gelatin in water and kept warm to prevent solidifying, was injected through cannula. The brains were removed and stored in formalin. Ventral and dorsal photographs of the brain were taken with a dissecting microscope for visualization of the middle cerebral artery (MCA) and posterior cerebral artery (PCA) territories. The MCA territory was determined by the localization of anastomoses, and distances between the midline and the anastomoses were measured at coronal planes 2, 4, and 6 mm from the frontal pole. Subsequently, the plasticity of the posterior communicating artery (PcomA) was graded on a qualitative scale of 0 -3 with 0 ϭ no anastomosis between PCA and superior cerebellar artery (SCA), 1 ϭ anastomosis between PCA and SCA in capillary phase, 2 ϭ small truncal PcomA, and 3 ϭ truncal PcomA.
Murine Model of Transient Focal Ischemia-Focal cerebral ischemia was produced by intraluminal occlusion of the left MCA with a nylon monofilament suture (38,39). In brief, mice were anesthetized with 1.5% isoflurane in a 30% O 2 , 70% N 2 O mixture under spontaneous breathing. After the left internal carotid artery (ICA) and external carotid artery (ECA) were isolated under an operating microscope, the ECA was dissected distally. A 5-0 monofilament nylon suture was inserted into the ICA through the ECA stump and gently advanced for ϳ11 mm to block the origin of MCA. After ischemia, blood flow was restored by withdrawing the intraluminal suture from the ICA. Rectal temperature was controlled at 37.0 Ϯ 0.5°C during the procedure via a temperature-regulated heating pad. Mean arterial blood pressure was monitored through a tail cuff, and arterial blood gas was analyzed at 15 min after the onset of ischemia. Changes in regional cerebral blood flow (rCBF) before, during, and after MCA occlusion were evaluated in animals using laser-Doppler flowmetry (38). The animals underwent MCA occlusion for 90 min and then reperfusion for up to 72 h. After recovering from anesthesia, the animals were maintained in an airconditioned room at 20°C.
Determination of Infarct Volume-At 72 h after MCA occlusion, brains were removed and frozen in 2-methylbutane at Ϫ30°C, covered with mounting medium, and stored at Ϫ80°C. Brains were serially cut (15 m thick) on a cryostat. Nine coronal sections 1 mm apart, starting at 2 and ending at 10 mm from the anterior pole, were cut and stained with cresyl violet to visualize the infarct area. The infarct area was determined with an image analysis system (MCID, St. Catherine's, Ontario, Canada) as previously described (40). The hemispheric infarct area in each section was calculated by subtracting from the ipsilateral ischemic hemisphere the area of normal brain based on the data of the non-ischemic contralateral side. This technique minimizes the effect of edema on measurement of infarct size (41). Infarct volume was then calculated by summing up the infarct areas over all sections and multiplying by the distance between sections (1 mm).
TUNEL Staining-This was performed on fresh-frozen sections as described previously (42). Brain sections were fixed with 4% paraformaldehyde in PBS for 15 min and then permeabilized with 1% Triton X-100 for 30 min, followed by three PBS washes. The sections were then incubated at 37°C for 90 min in the buffer containing 200 mM potassium cacodylate, 4 mM MgCl 2 , 1 mM 2-mercaptoethanol, 100 units/ml terminal deoxynucleotidyltransferase, and 20 nmol/ml biotin-conjugated 16-dUTP (Roche Molecular Biochemicals). After three washes in PBS, the sections were incubated at room temperature for 15 min with fluorescein-avidin D diluted in PBS at 8 g/ml, and examined by fluorescence microscopy at 495/515 nm (excitation/emission).
Immunofluorescence Staining-Coronal sections adjacent to the ones used for cresyl violet staining or TUNEL staining as described above were used. The procedures for immunofluorescent staining were the same as described previously (42). Briefly, the sections were incubated overnight at room temperature with the primary antibody diluted in PBS containing 2% horse serum, 0.2% Triton X-100, 5 mg/ml bovine serum albumin, and 0.2% glycine. After wash in PBS three times, sections were incubated with secondary antibody conjugated with a fluorochrome at room temperature for 1 h. Following the final wash, the sections were placed on coverslips using a fluorescent mounting medium (Vector). Double immunofluorescent staining was conducted as described in Refs. 43 and 44. Briefly, the staining proceeds in a linear fashion. After staining with the first set of primary and secondary antibodies as described above, the second set of primary and secondary antibodies conjugated to a different fluorochrome were applied in the same way. Attention was paid to avoid cross-reaction of the secondary antibodies to the different primary antibodies in the two sets, which were prepared in different species. Microscopic examination was conducted at excitation/emission wavelengths of 550/565 (for red fluorochrome) or 495/515 (for green fluorochrome).
Western Blot Analysis-Protein isolation from brain tissues, based on total cell extracts or subcellular fractionation (cytosolic and mitochondrial), was performed as described previously (44). For SDS-PAGE, protein samples at the indicated amounts were boiled in the loading buffer (100 mM Tris-HCl, 200 mM DTT, 4% SDS, 0.2% bromphenol blue, 20% glycerol) and separated on a 12% SDS-PAGE, followed by transfer to polyvinylidene difluoride membranes. The membrane was blocked with 3% nonfat milk and stained with the primary antibodies for 2 h at the optimal concentrations. After six washes in PBS with 0.2% Tween 20, the horseradish peroxidase-conjugated second antibody was applied, and the blot was developed with ECL reagents (Amersham Biosciences). To control for equal sample loading in each subcellular fraction, immunoblotting was performed using antibodies against the mitochondrial marker cytochrome c oxidase IV (PharMingen) and the cytosolic marker ␣-tubulin (Sigma). Immunoreactivity on each individual lane of the blots was semiquantified by a gel densitometric-scanning program and analyzed with an image analysis system (MCID, St. Catherine's, Ontario, Canada) as previously described (40).
Analysis of Caspase Activities-Brains were quickly removed and tissues were dissected from the two hemispheres separately. Protein extracts were prepared on ice by Dounce homogenization of tissues in a lysis buffer containing 25 mM HEPES, 5 mM EDTA, 1 mM EGTA, 5 mM MgCl 2 , 5 mM DTT, 0.5 mM PMSF, and 10 g/ml each of pepstatin, leupeptin, and aprotinin. Cell lysate was centrifuged at 14,000 rpm for 30 min (4°C), and the supernatant was used for the enzymatic assay. One hundred micrograms of the protein extracts were incubated for 1 h at 30°C with 20 M site-specific tetrapeptide substrates conjugated to aminotrifluromethyl-coumarin (AFC) in a caspase assay buffer (20 mM PIPES, 100 mM NaCl, 10 mM DTT, 1 mM EDTA, 0.1% (w/v) CHAPS, 10% (w/v) sucrose, pH 7.2). The substrates were Ac-IETD-AFC for caspase-8-like activity or Ac-DEVD-AFC for caspase-3-like activity (Biomol). The release of the fluorogenic group AFC was measured by a fluorescence spectrometer (PerkinElmer Life Sciences LS50-B), at 405/ 500 nm (excitation/emission).
Statistical Analysis-Statistical significance was determined by analysis of variance and post hoc Fisher's PLSD test.

Mice Deficient in bid Showed Cerebrovascular Anatomy and Hemodynamics after MCA Occlusion Similar to That for the
Wild-type Mice-Mice deficient in bid were created through gene targeting technique (33). These mice had been backcrossed to C57BL/6 background for 12 generations.
The wild-type and bid-deficient mice used in this study were littermates with virtually identical genetic background, which ensures a consistent cerebral vascular structure (37,45). This was further confirmed by the evaluation of cerebral circulation anatomy, which revealed no difference between wild-type and bid-deficient mice in the plasticity of the PcomA and the extent of anastomoses in MCA territory (Table I).
To compare the actual cerebral blood flow in the wild-type and bid-deficient mice during and after MCA occlusion, laser-Doppler flowmetry was performed. The assay indicated that the rCBF was similar between them during both ischemia and reperfusion (Table I). In both types of mice, rCBF was dropped to ϳ12-14% of baseline (pre-ischemia) but returned to ϳ80% of baseline levels within 30 min of reperfusion. All other hemodynamic parameters, such as blood pressure, blood pH, artery O 2 and CO 2 pressure, and core body temperature, were within the same ranges. These results indicate that a direct comparison between the wild-type and bid-deficient mice for the susceptibility to MCA occlusion-induced neuronal death and cerebral infarction is feasible and appropriate in these genetically manipulated mice without the interference from potential vascular dissimilarities (37,45).
bid Deficiency Led to Minor Ischemic Brain Injury-In the initial experiments, mice were subjected to 90 or 120 min of MCAO followed by reperfusion (n ϭ 15-20/group). These experiments confirmed the reproducibility of cortical blood flow changes following MCAO as determined with laser-Doppler flowmetry (Table I). Infarction was reproducibly detected in the MCA territory 1-3 days after MCAO. Because we found that the 90-min duration of MCAO resulted in reproducible infarction and yet relatively low mortality (Ͻ15% in 3 days), we decided to use this injury paradigm in the subsequent studies.
In the wild-type mice, MCA occlusion for 90 min followed by 72 h of reperfusion produced ipsilateral cerebral infarction with an averaging volume of 75 (Ϯ 14.3) mm 3 , as determined based on the loss of cresyl violet staining (Fig. 1a). In contrast, biddeficient mice had a significantly smaller infarct volume (52.8 Ϯ 17.3 mm 3 ), an approximate 30% reduction (p ϭ 0.012) (Fig. 1b). Notably, based on the analysis of infarct size on individual sections, the reduction in infarct size in bid-deficient mice was mainly in the posterior part of the MCA territory (Fig.  1c). These results indicated that Bid played a significant role in the development of ischemic infarction following MCA occlusion.
bid-deficient Brains Demonstrated a Reduction in Apoptotic Neuronal Cell Death following Ischemia-As Bid is an important pro-death Bcl-2 family protein bridging the extrinsic and intrinsic apoptosis pathways, we decided to determine whether bid deficiency could lead to a reduction in apoptotic cell death in the ischemic brains. TUNEL staining was performed to detect apoptosis on brains subjected to MCA occlusion and reperfusion. The brain sections for the staining were adjacent to those used for the measurement of infarct size to ensure a good correlation between the two parameters.
TUNEL-positive cells were readily detectable in the cortices of treated mice (Fig. 2a). Two types of TUNEL-positive cells could be detected in this focal ischemia model as described previously (7). One type of cell had apoptotic morphology with a condensed or fragmented nucleus (Fig. 2a, panels C and G), which was mainly detected in the inner border zone (ischemic penumbra) of the infarction (region A in Fig. 1a). This is the region where apoptotic death most often occurs in this model (39). The other type of TUNEL-positive cells had a diffusive cytoplasm-like staining (Fig. 2a, panel H), which was suggestive of necrosis and could be detected in over 80% of cells in the infarct core (region B in Fig. 1a).
Compared with the wild-type ischemic brains, a significantly lower number of TUNEL-positive cells were detected in the bid-deficient brains (Fig. 2a), particularly in the penumbral region, where apoptosis dominated (Fig. 2a, panels A-D). The numbers of total TUNEL-positive cells and apoptotic TUNELpositive cells were both significantly reduced in the bid-deficient brains (Fig. 2b). Even in the ischemic core region, where cells died mainly via necrosis in this model, the number of the apoptotic TUNEL-positive cells, which were less dominant, was significantly reduced in the bid-null mice, although the numbers of total TUNEL-positive cells between the two types of mice were not significantly different (Fig. 2, a (panels E-H) and c). The latter was a result of the fact that most TUNEL-positive cells in the ischemic core region were necrotic. Taken together, these results indicate that bid deficiency selectively reduced the apoptotic component of neuronal death in brains subjected to the treatment of ischemia and reperfusion.
Bid Was Activated in the Brain after Ischemia-Studies in the non-neuronal systems have shown that Bid can be activated following Fas/tumor necrosis factor receptor-1 death receptor engagement to yield the active tBid, which translocates to the mitochondria and induces cytochrome c release (30 -34). Activation of Fas in neurons and its involvement in the ischemic neuronal death has been reported (22)(23)(24)(25)27). Because the a Values shown are mean (S.E.) in millimeters (n ϭ 7/group). b A coronal plane is specified based on its distance from the frontal pole in millimeters. c PcomA (posterior communicating artery) plasticity scoring criteria: 0, no anastomoses between PCA and SCA; 1, anastomoses between PCA and SCA in capillary phase; 2, small truncal anastomoses between PCA and SCA; 3, truncal anastomoses between PCA and SCA.
ischemic infarct size and apoptosis were significantly decreased in the brain of bid-deficient mice, we speculated that Bid had been activated in neurons after ischemia. To confirm this conjecture, we examined the status of the Bid molecule by Western blot analysis of the subcellular fractions from ischemic brains and by immunofluorescence staining on sections from the same brains.
Indeed, Bid was cleaved and translocated following ischemia. The p15 truncated Bid could be detected in the mitochondrial fractions of the ischemic brains (Fig. 3). The translocation in ischemic neurons could be further confirmed by immunofluorescence staining (Fig. 4). The immunofluorescence signals of Bid were very low in non-ischemic cortex (Fig. 4A), and showed a diffusive pattern, consistent with its cytosolic distribution in normal cells (29). However, the signals became much stronger following the ischemic insult, and also converted to a punctuated pattern, suggesting mitochondria localization (Fig. 4, B and C). The subcellular localization was further confirmed by double-label staining with an anti-cytochrome c oxidase subunit IV antibody (Fig. 4E). Activation of Bid could be detected ϳ3 h after reperfusion and was most evident at ϳ6 h by both Western blot (Fig. 3) and immunofluorescence (Fig. 4) analyses, suggesting that Bid activation was an early apoptotic event following the ischemic insults. Furthermore, Bid was activated in neurons and was associated with apoptotic cell death, as demonstrated by double-label staining with anti-NeuN (a neuronal marker) (Fig. 4F) or TUNEL. In the latter, many cells in the infarct border zone that showed the punctuated pattern in bid immunofluorescence were also positive for TUNEL as determined 24 h after reperfusion started. Taken together, these data strongly support the notion that Bid contributes signifi-cantly to ischemic neuronal death.
Cytochrome c Release and Caspase-3 Activation Were Significantly Attenuated in bid-deficient Mice in Response to Ischemia-As the bid-deficient mice exhibited significantly higher resistance to ischemic injury and reduced neuronal apoptosis, it is possible that the ischemia-triggered mitochondrial apoptotic responses were attenuated in these mice compared with their wild-type littermates. The results presented in Figs. 5 and 6 support this hypothesis. As determined by Western blot analysis, cytochrome c release into the cytosolic fraction could be readily detected in the wild-type brains ϳ3 h after reperfusion and was continuously detectable at least through the first 24 h (Fig. 5). In contrast, cytochrome c release was significantly attenuated in bid-deficient mice during the early stage of ischemic injury (3-12 h after reperfusion). Cytochrome c release was found in bid-deficient brains ϳ24 h after reperfusion. These results, together with the observation that Bid was activated early on (Figs. 3 and 4), suggest that Bid is critical to the early activation of the mitochondria apoptosis pathway, although this pathway could be activated later in a Bid-independent fashion.
Consistent with these observations, activation of the effector caspases was also significantly diminished in the bid-deficient ischemic brains, as indicated by reduced caspase-3 processing (Fig. 6a) and reduced Ac-DEVD-AFC cleavage activity (Fig.  6b). Immunofluorescent staining using an antibody against the active caspase-3 subunit (p17) also showed that much fewer neurons contained the activated caspase-3 in the bid-deficient brains than in the wild-type brains (Fig. 6c). These results indicated that the Bid-mediated mitochondrial apoptotic pathway is important in the development of ischemic neuronal FIG. 1. bid deficiency leads to minor ischemic brain injury. a, cresyl violet histology was performed on brains from wild-type (bid ϩ/ϩ) or bid-deficient mice (bid Ϫ/Ϫ) following 90 min of ischemia and 72 h of reperfusion. Shown are representative coronal sections through anterior mid-caudate and anterior hippocampus levels (top to bottom panels). Regions A and B represent the penumbra and the ischemic core, respectively, where the representative TUNEL staining is performed (see Fig. 2). b, infarct areas from nine coronal sections of wild-type (bid ϩ/ϩ) or bid-deficient (bid Ϫ/Ϫ) ischemic brains (n ϭ 8) were integrated to generate the infarct volume. Data are mean Ϯ S.D. (p ϭ 0.012). c, coronal sections were prepared along the anterior-posterior axis of the wild-type (bid ϩ/ϩ) or bid-deficient (bid Ϫ/Ϫ) brains after the 90-min ischemia followed by a 72-h reperfusion. Infarct areas are compared between the two genotypes in each section. Data are mean Ϯ S.D. (n ϭ 8/group) (*, p Ͻ 0.05 versus the wild-type mice). death and that diminishment of this pathway may contribute to the attenuated ischemic brain injury in bid-deficient mice, as shown in this study.
Finally, the substrate-cleavage activity for Ac-IETD-AFC (for caspase-8-like activity) was measured after ischemia and compared between wild-type and bid-deficient mice (Fig. 6d).
The data indicate an equivalent increase in caspase-8-like proteolytic activity in wild-type and bid-deficient brains at 3-6 h after ischemia. This result is consistent with the notion that, during the early stages of brain injury, Bid activation constitutes a step that is upstream of caspase-3 activation but downstream of caspase-8 activation and thus could bridge the extrinsic apoptotic signals to the intrinsic pathway during ischemic neuronal death. DISCUSSION Most of the evidence supporting a role of apoptosis in neuronal cell death comes from studies using animal models of global or focal cerebral ischemia. In the focal ischemia models, which create a condition relevant to human stroke, a transient ischemia followed by reperfusion is often associated with massive induction of apoptosis-like cell death (46,47). In the current study, we demonstrated that a pro-death Bcl-2 family protein, Bid, prominently participated in the development of focal ischemic brain injury, whereas deletion of this molecule by gene targeting provided significant neural protection. At the gross level, bid-deficient mice had a 30% reduction in infarct size. Consistent with this gross protection was the reduction in the amounts of apoptotic cells, particularly in the penumbral region, in the ischemic bid-deficient brains. Furthermore, mice deficient in bid showed significantly attenuated cytochrome c release and caspase-3 activation in response to ischemia. Taken together, these results support our hypothesis that Bid is an important mediator of ischemic neuronal death in the brain, which may function to activate the mitochondria apoptosis signaling pathway in degenerating neurons.
The marked difference between the bid-deficient mice and their wild-type littermates in susceptibility to ischemic injury is likely attributable to the deletion of the Bid-mediated death signaling pathway per se, rather than the potential difference in cerebrovascular anatomy. The bid-deficient mice and their wild-type littermates used in this study were virtually of the same genetic background. These mice had been back-crossed to C57BL/6 for 12 generations, which would likely eliminate any strain differences in response to ischemic injury as previously noted in rodents (45,48). Indeed, we quantitatively evaluated the distribution of anastomoses and the plasticity of PcomA in the brain and found no significant difference between the two types of mice (Table I). Furthermore, the bid-deficient mice and their wild-type littermates showed the same changes in regional cerebral blood flow during and after ischemia (Table II). These observations demonstrate that the neuroprotective effect by the bid gene deletion is independent of changes in cerebral blood flow or other physiological parameters.
The data presented here are in concert with a recently published report (49), which showed that neuronal cultures derived from bid-deficient mice were resistant to apoptosis induced by oxygen/glucose deprivation, an in vitro experimental condition simulating ischemia. In addition, in the same study, an up to 67% reduction in infarct size was observed in bid-deficient mice subjected to transient focal cerebral ischemia, as compared with the wild-type controls (49). This level of protection was substantially greater than what we observed in our model system (30% reduction in infarction). This difference could be in part attributed to some minor differences in the genetic background of the bid-deficient mice used in the two different studies, with mice being back-crossed to C57BL/6 for 12 generations (inbred level) in our colony, but only for 7-8 generations in the colony used by the other study. Alternatively, the differences in the severity of ischemic insults employed by the two studies could be a major contributing factor that accounts for the different levels of neuroprotection. Although we employed a standard injury paradigm (90 min of ischemia and 72 h of reperfusion) that resulted in infarction of nearly the entire middle cerebral artery territory, a milder injury paradigm (30 min of ischemia and 48 h of reperfusion) that produced much smaller infarction was adopted by the other study. Indeed, the discrepancy in the level of protection shown in the two independent studies is consistent with the notion that the apoptotic mechanism may play a larger role in neuronal death after a milder ischemic insult than after a severe (or prolonged) ischemic insult (see Ref. 50 for review). In severe ischemic injury, in which necrosis is prominent, the contributions from Bid and the apoptosis execution machinery to the ultimate outcome of injured tissues may be overwhelmed by other mechanisms of cell death, such as oxidative stress, mitochondrial dysfunction, and energy failure. Nevertheless, the two studies clearly demonstrate that Bid activation is an early event in ischemic brain injury and complementarily support a role for Bid in mediating ischemic neuronal death.
The precise mechanism through which Bid is proteolytically activated in neurons after ischemia is unclear at the present time. Several intracellular molecules have been proposed to be the activators of Bid, including caspase-8, granzyme B, or caspase-3 (30 -32, 51). The Fas/FADD/caspase-8 pathway appears to be the most efficient mechanism for Bid cleavage in various cell types and could be the major pathway for Bid cleavage in the ischemic brain. There is abundant evidence that the expression of the Fas receptor is remarkably increased and that its interacting protein FADD is activated in the brain after cerebral ischemia (22-25, 52, 53). This study (Fig. 6) and others (28,49,54) have also provided evidence that caspase-8 is activated during the early phase of ischemic brain injury. Furthermore, in Fas lpr mice expressing dysfunctional Fas, ischemic infarction is markedly decreased as compared with wildtype mice (26,27). Although the above studies did not characterize the changes in caspase-8 and Bid activation in the Fas lpr mice in response to ischemia, it may be speculated that Fas deficiency would lead to the failure of caspase-8 and Bid activation in neurons; this warrants further study for verification.
Among the other mechanisms that potentially induce Bid activation, the proteolysis of Bid by caspase-3 may be an important feedback loop for the amplification of mitochondrial cytochrome c release (55). Because cytochrome c release and activation of caspase-3 are prominent in ischemic neuronal death, a role for caspase-3 as a direct activator of Bid in ischemic neurons should be considered. However, in a separate study, we found that inhibition of caspase-3/7 activities attenuated the production of the 15-kDa tBid in the brain at a later time point of 24 h, but not at the early time point of 6 -12 h, after focal ischemia (data not shown), suggesting that caspase-3 is not responsible for the activation of Bid during the early stage of ischemic injury. The finding that ischemia-induced activation of caspase-3, but not caspase-8, was significantly decreased in bid-deficient mice at 6 h after ischemia indicates that the activation of Bid is downstream of the caspase-8 activity but upstream of the caspase-3 activity.
Consistent with the prominent role of Bid in inducing cytochrome c release upon its translocation to the mitochondria, the data from this study show that ischemia-induced cytochrome c release was markedly attenuated in bid-deficient mice as compared with their wild-type littermates (Fig. 5). Notably, however, neither the inhibition of cytochrome c release nor the neuroprotection against the ischemic infarction was complete in the brain at the later stage of the injury development. These results suggest that mechanisms independent of the actions of Bid must be involved in the neurodegenerating processes after ischemia. Indeed, it has been suggested that several other pro-death events, such as Bax activation (44) and mitochondrial overproduction of ROS (18), may also be important triggers of cytochrome c release in ischemic neuronal death. These death signals could operate synergistically with the activated Bid in activating the death execution machinery.
As suggested by the temporal profile of Bid activation in ischemic brains, the Bid-dependent signaling pathway and its downstream cell death execution events seem to contribute to neuronal injury at the reperfusion phase, rather than at the ischemic phase, after MCA occlusion. Bid cleavage and its mitochondrial translocation, cytochrome c release, and caspase-3 activation were readily detectable ϳ3-6 h after the initiation of reperfusion (Figs. 3, 5, and 6), but none of these changes were observed at the end of MCA occlusion (0 min of reperfusion) or 30 min into the reperfusion (data not shown). These observations are consistent with many previous studies showing that the induction of neuronal apoptosis is generally a delayed event after the ischemia/reperfusion type of brain injury (see Ref. 50 for review). This delay in apoptosis induction after cerebral ischemia has been attributable to several factors, including the restoration of cellular energy levels during the reperfusion phase, which may be essential for the energyrequiring steps in the execution of the apoptotic program (56,57), and the markedly increased oxidative stress after the initiation of reperfusion, which is an established mechanism underlying the activation of apoptotic cascades after cerebral ischemia (54,58).
Although the precise pro-apoptotic actions of Bid on mitochondria have yet to be completely resolved, it is likely that Bid may activate both caspase-dependent and caspase-independent apoptosis execution mechanisms. The activation of caspase-dependent pathway mediated by Bid is mainly caused by the mitochondrial release of cytochrome c and, perhaps, Smac and HtrA2/Omi as well, which activate downstream caspases and inactivate inhibitor-of-apoptosis proteins (30 -35, 60, 61). The activation of caspase-independent apoptosis pathways could be caused by mitochondrial release of apoptosis inducing factor and endonuclease G, and also to mitochondrial dysfunction including mitochondria depolarization, generation of reactive oxygen species, and opening of permeability transition pore (15,16,59,60). At this moment, the relative contribution of these Bid-mediated, mitochondria-dependent cell-killing pathways to the final outcome in our model is not known. However, it will be a logical objective for future investigations. Nevertheless, it is conceivable that blocking apoptosis signals at the mitochondrial level, such as by inhibiting the function of Bid, would probably be a more effective therapeutic approach than FIG. 6. bid-deficient mice manifested significantly reduced caspase-3 activation. a, Western blot for caspase-3 activation. Brain extracts were prepared from the cortex of wild-type (bid ϩ/ϩ) or biddeficient (bid Ϫ/Ϫ) mice after a 90-min ischemia or sham operation (S) at different reperfusion times. Proteins were separated by a 12% SDS-PAGE and immunoblotted with an anti-caspase-3 antibody. Activation of caspase-3 was indicated by the appearance of the 17-kDa subunit. b, caspase-3-like activities in the extracts were measured using Ac-DEVD-AFC as the substrate. The bid-deficient brains (bid Ϫ/Ϫ, gray bar) manifested a significantly lower activity than the wild-type brains (bid ϩ/ϩ, black bar) (*, p Ͻ 0.05 versus the wild-type animals). c, immunofluorescence staining of the activated caspase-3 (p17) in wild-type (a) and bid-deficient (b) ischemic brains at 24 h after ischemia. The caspase-3 p17 positive cells are much decreased in the bid-deficient brains. d, activation of caspase-8. Wild-type (bid ϩ/ϩ, black bar) and bid-deficient (bid Ϫ/Ϫ, gray bar) mice were subjected to a 90-min transient MCAO, and the ipsilateral cortices were removed at 3, 6, and 24 h later, together with the sham-operated brains (sham). Caspase-8like activities were assessed using Ac-IETD-AFC and the relative changes over the sham control samples are shown. *, p Ͻ 0.05 versus wild-type samples. at levels downstream of mitochondrial damage, such as by using caspase inhibitors.
In conclusion, although multiple molecules and pathways may participate in the activation of neuronal apoptosis and contribute to the ultimate cell death after cerebral ischemia, this study demonstrates the significant contribution of a specific pro-death molecule, Bid, to this process. Because of its strategic position in the apoptosis signaling pathways, Bid could be an ideal target for future therapeutic intervention.

TABLE II
There are no differences between wild-type and bid-null mice in hemodynamics during and after the transient ischemia treatment through the middle cerebral artery occlusion Hemodynamic parameters measured are: rCBF (percentage of base-line level); BP (blood pressure; mmHg); pO 2 (artery O 2 pressure; mmHg); pCO 2 (artery CO 2 pressure; mmHg). Data are obtained during and after a 90-min MCA occlusion and expressed as the percentages of pre-ischemia base-line levels. Data are mean (S.D.) from 8 wild-type mice and 7 bid-null mice. a Measurements were taken during and after MCA occlusion.