Apoptotic Neuronal Death following Deep Hypothermic Circulatory Arrest in Piglets

After approval by the Institutional Animal Care and Use Committee of the Joseph Stokes Jr. Research Institute, 39 piglets aged 5–10 days were studied. Anesthesia was induced with intramuscular ketamine (33 mg/kg) and acepromazine (3.3 mg/kg), followed by tracheal intubation, mechanical ventilation, and intravenous catheter insertion. Anesthesia was maintained with intravenous fentanyl (initial loading dose of 25 μg/kg, then 10 μg · kg⁻¹· h⁻¹) and droperidol (0.25 mg · kg⁻¹· h⁻¹).

A femoral arterial catheter was inserted to monitor arterial pressure (P23XL Transducer; Spectramed, Critical Care Division, Oxnard, CA), glucose concentrations (Surestep; Lifescan, Milpitas, MN), blood gases, pH, and hemoglobin concentrations (I-STAT; iSTAT Company, East Windsor, NJ). Expired carbon dioxide was measured (Normocap; Daytex-Ohmeda Division, Instrumentarium Corporation, Helsinki, Finland) and electrocardiographic monitoring (ECG; Hewlett-Packard, Federal Republic of Germany) was performed. Thermistors (Yellow Springs Instruments, Yellow Springs, OH) were inserted into the cranial epidural space, rectum, and esophagus to monitor brain and core body temperatures.

Animals were divided into eight groups. Controls (group 1) underwent surgical preparation and were killed (n = 8). Group 2 underwent surgical preparation and then hypothermic CPB, followed by 1-day survival (n = 3). Groups 3–8 underwent surgical preparation and then hypothermic CPB and DHCA, followed by 1-h (n = 5), 4-h (n = 6), 8-h (n = 5), 24-h (n = 5), 72-h (n = 4), or 1-week (n = 3) survival. In groups 2–8, physiologic variables were recorded after surgical preparation (before CPB), after CPB cooling, during CPB rewarming, and 2 h after discontinuation of CPB. In group 1, physiologic variables were recorded after surgical preparation.

In survival groups, intramuscular ketamine and acepromazine were administered before killing. Then the scalp was reflected, after which the skull and dura over the left cerebral hemisphere were removed. Intravenous heparin (300 U/kg) was administered. Frontal, parietal, and occipital neocortical tissue samples (200–300 mg) were obtained with a biopsy drill, which aspirated the samples into a liquid nitrogen-cooled reservoir. The frozen tissue was stored in liquid nitrogen until analysis.

After pentobarbital (100 mg/kg) was administered intravenously (euthanasia), chilled saline 0.9% (1 l) and then 4% paraformaldehyde in 0.1 m phosphate-buffered saline (1 l; pH, 7.4) were infused into the carotid artery to fix the brain in situ . After the brain was removed in toto , the right hemibrain was isolated and cut coronally into 2.5-mm blocks. Alternating blocks were prepared for frozen or paraffin-embedded sections. The blocks destined for frozen section were stored at 4°C in 25% sucrose, followed 2 days later by immersion in cold isopentane before storage at −70°C.

The blocks destined for paraffin sectioning were dehydrated in ethanol and xylene (Citadel 2000; Shandon-Lipshaw, Pittsburgh, PA) and embedded in paraffin (Histoembedder 1160; Leica, Deerfield, IL). One 8-μm section was cut (Microtome 2155; Leica) from a paraffin block and stained with hematoxylin and eosin to determine cell damage. One 40-μm section was cut from a frozen block (Jung Biocut BSF-5TC; Leica) with use of tissue-freezing medium (Triangle Biomedical Sciences, Durham, NC) to determine immunohistochemically the location of caspase-3 and -8 activities.

In the 24-h, 72-h, and 1-week survival groups, a neurologic examination was performed before killing. Points were assessed for deficits in level of consciousness (range, 0–25), respiration (0–5), cranial nerve function (0–6), sensory function (0–14), gait (0–25), and behavior (0–20). 11The scores from each category were summed. The minimum score (0) represents no deficits (normal examination findings), whereas the maximum score (95) indicates severe impairment (brain death).

Through a right-neck incision, the carotid artery and external jugular vein were exposed, through which cannulae (Bio-Medicus; Medtronic, Minneapolis, MN) were advanced to the ascending aorta and right atrium for CPB. Heparin (200 U/kg) was administered intravenously. After arterial cannulation, 10 ml/kg blood was collected for transfusion after CPB.

The CPB circuit used a bubble oxygenator (Bio-2; Baxter Healthcare Corporation, Santa Ana, CA) with an arterial filter (LPE 1440 extracorporeal filter; KOL Bio-Medical Instruments, Chantilly, VA) receiving oxygen at a rate of 1 l/min and a nonpulsatile roller pump (RS 7800; Renal Systems, Minneapolis, MN) flowing at 100 ml · kg⁻¹· h⁻¹. Arterial blood gas management followed α-stat principles. The CPB prime contained pig whole blood, heparin (2,000 U), fentanyl (50 μg), pancuronium (1 mg), calcium chloride (500 mg), dexamethasone (30 mg/kg), cefazolin (25 mg/kg), furosemide (1 mg/kg), and sodium bicarbonate (25 mEq). Plasma-lyte A (Baxter, Deerfield, IL) was added to yield a hematocrit of 25% during CPB.

Hypothermia was induced with CPB and ice bags around the head and body. At a brain temperature of 19°C, rewarming was initiated in the CPB group, whereas DHCA was induced for 90 min in the DHCA groups. To rewarm, CPB pump flow and arterial perfusate were increased gradually to 100 ml · kg⁻¹· h⁻¹and 38°C, respectively, facilitated with a radiant-heating lamp and circulating-water blanket. After 10 min of CPB, the heart was defibrillated as necessary, and mechanical ventilation was resumed.

At a brain temperature of 28°C, mannitol (0.5 g/kg) was administered. CPB was discontinued when all body temperatures were greater than 32°C (approximately 40 min of CPB rewarming).

After CPB, cannulae were removed, protamine (4 mg/kg) was administered intravenously, and the neck incision was sutured closed. The blood collected before CPB was transfused over 1–2 h to maintain a mean arterial pressure >50 mm Hg, after which dextrose (5%) in lactated Ringer's solution was infused intravenously (4 ml · kg⁻¹· h⁻¹). In the 1-, 4-, and 8-h DHCA reperfusion groups, anesthesia was continued until killing. In the 24-h, 72-h, and 1-week survival groups, anesthesia was discontinued, and the trachea was extubated when purposeful movements, airway reflexes, strength, and regular breathing had returned.

Hematoxylin and eosin-stained slides containing frontoparietal neocortex, corresponding to the areas that were biopsied in the opposite hemisphere, were evaluated for neuronal cell death, inflammation, hemorrhage, and infarction. The slides were scored semiquantitatively on a scale of 0–5: 0 = no damage (normal neuronal structure); 1 = rare damage (<1% neurons dead; no inflammation or infarction); 2 = mild (1–5% neurons dead; no inflammation or infarction); 3 = moderate (6–15% neurons dead; no inflammation or infarction); 4 = severe (16–30% neurons dead, inflammation, or infarction [one or all of these]); and 5 = very severe (>30% of all neurons dead; inflammation and infarction). 6Apoptotic cells were defined by the presence of nuclear karyorrhexis and minimal cytoplasmic change, whereas necrotic cells were identified by a pyknotic nucleus or no nucleus, along with a swollen, eosinophilic cytoplasm.

The adenosine triphosphate (ATP) concentrations were measured by the Luciferin-Luciferase reaction. In brief, 1 mg frozen neocortical tissue was dissolved in 100 μl 0.1-N NaOH in methanol and brought to 0.3 ml with H²O. After heating (90°C) for 15 min, 5 μl was added to 100 μl Luciferin-Luciferase solution (Sigma Chemical, St. Louis, MO). Standard ATP concentrations (Sigma Chemical Co) were also prepared. The sample and standard reactions were read in a Luminometer (LKB 1250; Perkin Elmer-Wallac, Turku, Finland).

An ApoAlert Caspase Colorimetric Assay Kit (Clontech, Palo Alto, CA) was used to measure caspase-3 and -8 activities. Frozen neocortical samples (5–10 mg) were homogenized in 0.5 ml cell lysis buffer (4°C) containing 25 mm HEPES (pH, 7.5), 5 mm EDTA, 5 mm MgCl², 5 mm dithiothreitol (DTT), and 10 μg/ml each of pepstatin, leupeptin, aprotinin, and PMSF. Homogenates were centrifuged at 10,000 rpm for 10 min. The supernatant was brought to 50 μl with cell lysis buffer, to which was added 50 μl of 2× reaction buffer, followed by 5 μl ρ-nitroanalide (pNA) substrate. For caspase-3 activity the pNA substrate was DEVD-pNA (1 mm), and for caspase-8 activity it was Ac-IETD-pNA (4 mm). The 1-ml 2× reaction buffer was prepared by adding 5 μl of 1.0 m DTT to 1 ml reaction buffer. Standard pNA concentrations were prepared in lysis buffer. The caspase-3 and caspase-8 reactions were incubated at 37°C for 1 and 2 h, respectively. Sample and standard solutions were measured at 405 nm (DU-640 spectrophotometer; Beckman, Fullerton, CA). Supernatant total protein concentration was determined by Bradford assay (Bio-Rad, Hercules, CA).

The same supernatant was used for Western blot analysis as for biochemical assays. Separate gels were run for each DHCA reperfusion group, containing representative samples in eight wells and molecular standards in two wells. Because of sample concentration and volume considerations, the lanes for the CPB, DHCA 4-h reperfusion, DHCA 8-h reperfusion, DHCA 24-h reperfusion, and DHCA 1-week reperfusion groups were loaded with 0.25 μg protein, and the lanes for the control, DHCA 1-h reperfusion, and DHCA 72-h reperfusion groups were loaded with 0.50 μg protein. These gels were the source of blots against antibodies for caspase-3 and -8, TNF-α, Fas, and Fas-ligand. For cytochrome c , the samples were concentrated (Speedvac Concentrator; Savant Instruments, Farmingdale, NY) and the lanes were loaded with 5 μg protein from each DHCA group.

Protein samples were mixed with Laemmli sample buffer consisting of 1% SDS, 0.1% 2-mercaptoethanol, and 0.1% bromophenol blue in a ratio of 1:1 (vol/vol). After boiling for 3 min, the proteins were separated by 8–16% gradient sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE; Bio-Rad). After electrophoresis (200 V) at room temperature for 30 min, proteins were transferred onto a nitrocellulose membrane (50 V) and stored overnight at 4°C. The blots were then immersed in 2% bovine serum albumin for 30 min at room temperature on a rocker platform. The primary antibody (10 μl) was added and allowed to react overnight at 4°C. The blot was then washed in TBST (50 mm Tris [pH, 7.5], 150 mm NaCl, and 0.05% Tween-20), after which it was immersed in 2% BSA containing the secondary antibody.

After washing with TBST, the blot was immersed in LumiGLO (Cell Signaling, Beverly, MA) for 3 min and rinsed in water. The blot was wrapped in cellophane and exposed to CL-exposure film (Pierce, Rockford, IL) for 15 sec and subsequently developed. The primary antibodies for caspase-3 and caspase-8 (Cell Signaling) were the cleaved fragments at 17–19 kDa (Asp175, rabbit polyclonal immunoglobulin (Ig)G) and 10 kDa (Asp 384, mouse monoclonal IgG1), respectively. The primary antibody for TNF-α (48 kDa) was a purified goat IgG (R&D systems, Minneapolis, MN). The primary antibodies for Fas (48 kDa) and Fas-ligand (37 kDa) were rabbit polyclonal IgG, and that for cytochrome c  (10 kDa) was a goat polyclonal IgG (Santa Cruz Biotechnology, Santa Cruz, CA). The secondary antibodies to the rabbit primary antibodies were a goat antirabbit IgG (diluted 1:1,000; Cell Signaling); donkey antigoat (diluted 1:2,000; Santa Cruz) for the goat primary antibody; and goat antimouse IgG (diluted 1:1,000; Cell Signaling) for the mouse primary antibodies. All secondary antibodies were bound to horseradish peroxidase.

Gels were also run without primary antibody to confirm that no bands were observed. The films were scanned and optical densities of the bands were determined with use of Scion Imaging Software (Scion Corporation, Frederick, MD).

Free-floating, 40-μm frozen sections were immersed in 50 mm Tris (pH, 7.5) with 0.1% Triton X-100. Sections were initially treated with 1% H²O²for 30 min to inactivate endogenous peroxidase and then 3 ml of 2% bovine serum albumin for 30 min. After the addition of primary antibody to caspase-3 or caspase-8 (1:500 dilution) at 4°C overnight, sections were washed with 3 ml 50 mm Tris (pH, 7.5) with 0.1% Triton X-100, followed by exposure to secondary antibody (1:500 dilution) for 1 h. Three-milliliter avidin-biotin peroxidase complex (Santa Cruz) was then added for 1 h. Afterward, 75 μl diaminobenzidine solution (Santa Cruz) was added for 5–10 min until color was apparent.

Tissues were then washed with water to stop the reaction. They were brushed on poly-L-Lysine-coated slides and left to dry at room temperature. Tissue sections were immersed in a 1:10 dilution of hematoxylin for 30 sec as a counterstain and dehydrated with increasing concentrations of ethanol and then xylene for 5 min.

Data are presented as mean ± SD. Groups were compared by analysis of variance (ANOVA). For significant F values, multiple means were compared with Tukey test. A P  value of less than 0.05 was considered statistically significant

Table 1displays the average and SD physiologic data for all groups combined during the study. There were no significant differences in physiologic data between the groups. CPB cooling, rewarming, and total CPB durations were 28 ± 9 min, 41 ± 10 min, and 70 ± 17 min, respectively. Neurologic impairment was apparent at 1 and 2 days postoperatively after DHCA (fig. 1). Signs of neurologic impairment manifested mainly as gait disturbance and diminished alertness. Neurologic impairment was not observed in CPB-only groups.

In neocortex, damaged neurons were not observed in the control group, rarely in the CPB and DHCA 1-h, 4-h, and 1-week reperfusion groups, and often in the DHCA 8-h, 1-day, and 3-day reperfusion groups (fig. 2). Although neuronal death was widespread at 8–72 h after DHCA, cortical tissue ATP concentrations remained similar to those in the control group (table 2). Although not quantitated, damaged neurons were usually apoptotic in appearance (fragmented, rounded, dense chromatin with minimal cytoplasmic change) and less often necrotic (nuclear pyknosis with a swollen, eosinophilic cytoplasm) (fig. 3). Damaged neurons usually appeared in clusters located in the superficial (layers 2 and 3) neocortex.

Caspase-3 activity was significantly increased 8 h to 3 days after DHCA, in comparison with that in the control and CPB groups (fig. 4). Caspase-3 active fragments were present on Western blots as early as 4 h after DHCA and as long as 3 days after DHCA (fig. 5) and were not present in control, CPB, or other DHCA groups. Curiously, caspase-3 fragments were biphasic, as the fragments were not observed at 8 h after DHCA. Caspase-8 activity was significantly increased only at 8 h after DHCA. On Western blots, caspase-8 active fragments were present at 8 h after DHCA, as well as 1 day after DHCA and 1 day after CPB only (fig. 5). Caspase-3 and -8 activity appeared in neurons located mainly in superficial neocortex, although activity was also present in glial cells in cortical white matter and basal ganglia (figs. 6 and 7).

Cytochrome c  increased significantly 1 h after DHCA (fig. 8). Fas increased significantly 4 h after DHCA; there was also a tendency for increased Fas (P = 0.07) at 1 h after DHCA (fig. 9). Fas-ligand was not increased significantly (fig. 9). On Western blots, TNF-α was not detected in the control, CPB, or DHCA groups.

This study reveals that DHCA neuronal death involves activation of the apoptotic cascade. Cell death occurred primarily in the superficial neocortex with morphologic features consistent with apoptosis. Despite neuronal death, cortical tissue ATP concentrations were preserved. According to assay and Western blot results, caspase-3 activity coincided with cell death at 8–72 h after DHCA. Caspase-8 activation occurred at 8 h, Fas expression at 4 h, and cytochrome c  release at 1 h after DHCA, showing initiation of the apoptotic cascade well before cell death was apparent. Apoptotic-appearing neurons showed caspase-3 and -8 active fragments in the neocortex. These enzyme activities and cell death were not observed in control and CPB groups. Our findings support the implication that caspase inhibitors and other suppressors of apoptosis might afford neuroprotection for DHCA in neonatal heart surgery.

Our model simulates neonatal heart surgery with CPB and DHCA. 6CPB is used to induce hypothermia and support circulation, as well as effect resuscitation and rewarming after DHCA. Although hypothermia confers neurologic protection, this protection is incomplete. After DHCA, neurons and oligodendrocytes are selectively vulnerable to die in the neocortex and less so in the hippocampus, striatum, and cerebellum. 3–6Selective vulnerability is not a result of uneven brain cooling, as brain temperature gradients are not observed during DHCA in our model. 12Previous work has shown that cell death occurs as early as 6 h after DHCA, peaking at 24–72 h, whereas neurologic function improves despite ongoing cell death. 6 

The difference in time course between histology and function may result from the cells that are injured but not dead interfering with neurologic function, as opposed to dead cells, which can be compensated for by the other live cells. Neurologic deficits include gait disturbance and diminished alertness, consistent with the location of cell death in the neocortex. Cell death in this model appears to be apoptotic from morphological analysis. 6 

The current study showed morphologic, histologic, and neurologic outcomes similar to those in prior investigations but yielded biochemical and molecular evidence of apoptosis. Specifically, the apoptotic cascade is initiated as soon as 1 h after reperfusion, well before cell death appears, and continues for 72 h after DHCA. Cell death and the apoptotic cascade both diminished by 1 week after DHCA.

Certain regions of the brain have been shown to be selectively vulnerable to injury. 3–6These regions appear to differ between neonates and adults, as well as between hypothermic and normothermic ischemia. Bottiger et al.  3studied neuronal death after normothermic global ischemia in adult rats and found neurons in the hippocampus, cerebellum, striatum, thalamus, amygdala, and neocortex to be vulnerable. Laptook et al.  4studied global ischemia in normothermic and hypothermic piglets. They found that neurons in the neocortex, thalamus, basal ganglia, cerebellum, and brain stem were vulnerable after normothermic ischemia and that hypothermia does not confer neuroprotection equally to all brain regions.

In a DHCA puppy model, Mujsce et al.  5observed that the neocortex, basal ganglia, and amygdala were selectively vulnerable and that hypothermic protection was not equal in all regions. In our DHCA model, selective vulnerability was observed in similar locations in so far as the neocortex, hippocampus, striatum, and cerebellum were concerned. 6Our findings differed, however, in the amygdala, thalamus, and brain stem. These differences may represent differences in animal species and models of ischemia.

Evidence is accumulating for apoptosis in neuronal death following global ischemic insult. 7–8In apoptosis, cell death results from a regulated process involving the activation of specific genes, receptors, and proteins that ultimately cause DNA fragmentation and the destruction of cell structure. The apoptotic cascade depends on the activation of caspases, a family of cysteine proteases specific for substrates at the aspartate residue. Caspase-3 is known as the executioner caspase, because it directly activates a nuclease (ICAD) that cleaves DNA strands and destroys the cell cytoskeleton.

Two principal pathways for the activation of caspase-3 include Fas/FADD/caspase-8 (the extrinsic pathway) and cytochrome c /Apaf-1/caspase-9 (the intrinsic pathway). However, these signaling pathways interact with each other to amplify caspase-3 activity and apoptotic cell death.

In the extrinsic pathway, TNF-α and Fas-ligand are known to bind to cell surface receptors, which recruit the adaptor molecule Fas-associated death domain (FADD), and the zymogen form of caspase-8. After recruitment, procaspase-8 is proteolytically cleaved to generate two distinct subunits: a large (p18) fragment and a small (p10) fragment. 13These subunits assemble into a heterotetramer to activate caspase-3 by cleaving it into 17–21-kDa and 10–13-kDa subunits. 14Our enzymatic assays measured caspase-8 and caspase-3 cleavage activity, and Western blot analysis detected the active fragments of caspase-3 and -8.

In the intrinsic pathway, the mitochondrion increases membrane permeability and releases cytochrome c  into the cytosol in response to ischemia. 15–17Cytochrome c  forms a complex with Apaf-1, which in turn activates caspase-9 to activate caspase-3. Our Western blot analyses detected cytochrome c  in cytosolic fractions of brain tissue homogenates.

In normothermic neonatal hypoxia-ischemia models, evidence of apoptosis includes the presence of damaged neurons with apoptotic structure and TUNEL positivity, as well as caspase activity and the expression of proapoptotic and antiapoptotic factors, such as bcl-2, bcl-x, and bax. 18–23Even more convincing evidence of apoptosis is the effectiveness of caspase inhibitors in affording neuroprotection after ischemia. 24During normal brain development, neurons in the neocortex are known to selectively die by apoptosis. 25–26Comparison of adult and neonatal models of hypoxia-ischemia suggests that apoptosis may be more prevalent in the immature brain. 27–29 

Our study gives biochemical and molecular evidence of apoptosis in another neonatal hypoxia-ischemia model. Because enzymatic assays are not entirely specific for caspase-3 or caspase-8 enzymatic function, we used immunoblot analyses to confirm the presence of active fragments specific for each caspase. Our results implicated both the extrinsic and intrinsic mechanisms of apoptosis, as both caspase-8 activation and cytochrome c  release were observed. The trigger for caspase-8 appeared to be Fas and not TNF-α. However, further experiments are necessary to functionally tie Fas to caspase-8 induction and exclude insensitivity of the blot for TNF-α. To confirm that caspase activity was located in the cells that were damaged, we observed active caspase-3 and -8 fragments in neurons in the superficial neocortex.

In our study, there were some discrepancies between the findings of the enzymatic assay and Western analysis. For caspase-8, enzyme activity was increased at 8 h reperfusion, whereas immunoblot activity was increased 8–24 h after reperfusion. For caspase-3, assays showed activity increased from 8 to 72 h after reperfusion, whereas blots showed biphasic active fragments with increases at 4 h and 24–72 h. The discrepancy between immunoblot and enzyme assay findings may arise from the lack of complete specificity of the enzyme assay and/or from differences in assay methods. The biphasic expression of caspase-3 fragments suggests the possibility of two pathways for activation, perhaps the intrinsic and extrinsic pathways. Cytochrome c , in the intrinsic pathway, was released at 1 h, which might account for the early caspase-3 increase, whereas caspase-8, in the extrinsic pathway, was active later, which might have played a role in the late caspase-3 increase.

It is also important to note the lack of significant expression of Fas-ligand, despite the clear presence of Fas active fragments at 4 h after DHCA. Because Fas-ligand is subject to rapid proteolytic cleavage, its expression is difficult to detect 30; our finding may reflect differences in the stability of Fas and Fas-ligand.

In summary, following DHCA, induction of apoptosis in the neocortex occurs within a few hours of reperfusion and continues for several days. Increased Fas, cytochrome c , and caspase concentrations, coupled with normal brain ATP concentrations and apoptotic histologic appearance, are consistent with the occurrence of apoptotic cell death.