the Department of Neurosurgery, Department of Neurology and Neurological Sciences, and the Program in Neurosciences, Stanford University School of Medicine, Stanford, Calif.
Abstract
Background and Purpose— Apoptotic cell death is associated with acute brain injury after subarachnoid hemorrhage (SAH). The Akt/glycogen synthase kinase-3 (GSK3) pathway plays an important role in the cell death/survival pathway after a variety of cell death stimuli. However, its role in acute brain injury after SAH remains unknown.
Methods— We used an endovascular perforation model of SAH in rats. Phospho-Akt and phospho-GSK3 expression was examined by Western blot analysis and immunohistochemistry. Terminal deoxynucleotidyl transferase–mediated uridine 5'-triphosphate-biotin nick end-labeling (TUNEL) and a cell death assay were used for detection of apoptosis. We administered LY294002 to examine the role of the Akt/GSK3 pathway in the phosphoinositide 3-kinase pathway after SAH.
Results— Phosphorylation of Akt and GSK3 was accelerated after SAH. In the cerebral cortex, where acute brain injury was the most severe, phosphorylation of these proteins was observed in the early phase after SAH. Cortical neurons with continuous Akt phosphorylation did not colocalize with TUNEL-positive cells at 24 hours. LY294002 reduced Akt and GSK3 phosphorylation and increased brain injury after SAH.
Conclusions— The present study suggests that the Akt/GSK3 pathway might be involved in neuronal survival in acute brain injury after SAH.
Key Words: apoptosis brain injuries subarachnoid hemorrhage
Introduction
Subarachnoid hemorrhage (SAH) results in a high mortality rate; 15% of patients with SAH die before reaching the hospital and 30% die within 24 hours of onset.1 Patients who survive the initial hemorrhage and overcome vasospasms frequently experience persistent cognitive deficits, psychosocial impairments, and a decrease in quality of life as a result of acute brain injury.2 However, most of the research has focused on the late phase, when vasospasm occurs, whereas the mechanisms of acute brain injury are poorly understood. Recent studies have shown that apoptosis is involved in acute brain injury in experimental SAH.3–5 Moreover, apoptosis has been observed in human patients with SAH.6 These reports suggest that apoptosis might be a therapeutic target for acute brain injury after SAH.
The serine-threonine kinase, Akt, plays an important role in the cell death/survival pathway.7 Growth factors activate Akt and promote cell survival through the phosphoinositide 3-kinase (PI3-K) pathway, which can be inhibited by LY294002.7,8 Phosphorylation at serine-473 is required for Akt activity.8 After phosphorylation, Akt functions through phosphorylation and inhibition of several substrates, including glycogen synthase kinase-3 (GSK3).9 Akt phosphorylates GSK3 on serine-9 to render it inactive, a mechanism by which neurons become resistant to apoptotic stimuli.10 Akt pathways are the focus for clarifying the apoptotic neuronal death machinery in several models of neurodegenerative diseases.11–13 However, the mechanism of cell survival mediated by the Akt pathway after SAH remains unclear. In the present study, we investigated the role of the Akt/GSK3 pathway in acute brain injury after SAH.
Materials and Methods
SAH Rat Model
SAH was produced in male Sprague-Dawley rats (270 to 300 g) by a previously reported method.14 The rats were anesthetized with 2.0% isoflurane in 70% nitrous oxide and 30% oxygen delivered by face mask. Rectal temperature was controlled at 37°C during surgery. The left common carotid artery was isolated, and a 3-0 monofilament nylon suture was inserted through the internal carotid artery to perforate this artery at the bifurcation of the anterior and middle cerebral arteries. SAH was confirmed in each rat at autopsy. Sham-operated rats underwent identical procedures except the perforation.
Physiological Data and Mortality
After anesthesia, the femoral artery was exposed and catheterized with a PE-50 catheter to allow measurement of blood gas values, blood pH, and continuous recording of mean arterial blood pressure. Mortality was calculated 24 hours after SAH.
Drug Injection
To investigate the role of the PI3-K pathway after SAH, we administered a PI3-K inhibitor, LY294002 (Cell Signaling Technology), as described.13 LY294002 (50 mmol/L in 25% dimethyl sulfoxide in phosphate-buffered saline [PBS]) and the vehicle (25% dimethyl sulfoxide in PBS) were injected intracerebroventricularly (10 μL, bregma; 1.4 mm lateral, 0.8 mm posterior, 3.6 mm deep) 30 minutes before SAH.
Western Blot Analysis
Samples were obtained from the cerebral cortex, caudate putamen, and hippocampus. Whole-cell protein extraction was performed as described.13 Equal amounts of the samples were loaded per lane. The primary antibodies were a 1:2000 dilution of the antibody against Akt and phospho-Akt (serine-473), a 1:1000 dilution of the antibody against GSK3 and phospho-GSK3 (serine-9) (both from Cell Signaling Technology), and a 1:10 000 dilution of the antibody against -actin (Sigma-Aldrich). Western blots were performed with horseradish peroxidase–conjugated immunoglobulin G with the use of enhanced chemiluminescence detection reagents (Amersham International).
Immunohistochemistry
Anesthetized animals were perfused with 10 U/mL heparin saline and subsequently with 4% formaldehyde in PBS. The brains were sectioned at 50 μm with a vibratome. The sections were incubated with the antibody against phospho-Akt (serine-473) (Cell Signaling Technology) and the antibody against phospho-GSK3 (serine-9) (Santa Cruz Biotechnology), each at a 1:50 dilution. Immunohistochemistry was performed with the avidin-biotin technique, followed by methyl green staining.
Immunofluorescent Double-Labeling Staining
The sections were reacted with the antibody against phospho-Akt (serine-473) (Cell Signaling Technology), followed by a Texas Red–conjugated immunoglobulin G antibody (Jackson Immuno-Research) at a 1:200 dilution. The sections were then incubated with the anti–neuron-specific nuclear protein (NeuN) Alexa Fluor–conjugated antibody (Chemicon International) at a 1:200 dilution. The sections were then covered with Vectashield mounting medium with 4',6-diamidino-2-phenylindole (DAPI; Vector Laboratories) and observed with a fluorescence microscope.
In Situ Labeling of DNA Fragmentation
The animals were killed, and the brains were rapidly frozen and sectioned on a cryostat. We performed terminal deoxynucleotidyl transferase–mediated uridine 5'-triphosphate-biotin nick end-labeling (TUNEL) to detect the DNA free 3'-OH ends as described.15 Staining was visualized with 0.025% diaminobenzidine, 0.075% H2O2, and 1% nickel sulfate in PBS, followed by methyl green staining.
Immunofluorescent Double Labeling With Phospho-Akt (Serine-473) Immunohistochemistry and In Situ Staining
The sections were reacted with the anti–phospho-Akt (serine-473) antibody, followed by the Texas Red–conjugated immunoglobulin G antibody (Jackson ImmunoResearch). Then, we performed TUNEL, followed by incubation with fluorescein avidin DCS (Vector Laboratories). Subsequently, the slides were covered with Vectashield mounting medium with DAPI and observed with a fluorescence microscope.
Cell Death Assay
For quantification of DNA fragmentation, which indicates apoptotic cell death, we used a commercial enzyme immunoassay to determine cytoplasmic histone-associated DNA fragments (Roche Molecular Biochemicals). Fresh brain tissue was taken, and extraction of protein from the cytosolic fraction was performed as described.15 A cytosolic volume containing 20 μg of protein was used for the ELISA, according to the manufacturer’s protocol.
Quantification and Statistical Analysis
Comparisons among multiple groups were performed with a one-way ANOVA, followed by a Scheffe post hoc analysis (SigmaStat software; Jandel Corp). Comparisons between 2 groups were achieved with Student’s unpaired t test. The data are expressed as mean±SD, and significance was accepted with P<0.05.
Results
Physiological Data and Mortality
The physiological values just after SAH were as follows: 36.5± 0.7°C body temperature, 7.35±0.06 pH, 143.1±19.4 PaO2, and 40.0±7.5 PaCO2 (values are mean±SD, n=5). There were significant increases in mean arterial blood pressure immediately after SAH (23.7% of baseline values, n=4), which were consistent with a previous report.16 Mean arterial blood pressure approached baseline values within 5 minutes after perforation. Neurological status at 24 hours was severely impaired in the SAH animals compared with the sham-operated animals, as described previously.4,17 The mortality rate at 24 hours was 23.3% (7 of 30 rats).
Akt Phosphorylation
Western blot analysis showed that phospho-Akt expression was significantly increased at 1 hour in the cerebral cortex, 6 hours in the caudate putamen, and 24 hours in the hippocampus compared with control samples (Figure 1A). Akt expression was not changed after SAH. An immunohistochemistry study showed that phospho-Akt expression increased and peaked at 1 hour in the cerebral cortex and 6 hours in the caudate putamen, which was consistent with the results of the Western blot analysis (Figure 1B). In the hippocampus, phospho-Akt expression temporarily decreased at 1 and 6 hours and increased thereafter. These results indicate that phosphorylation of Akt is accelerated after SAH, and the timing of this phosphorylation is different among brain regions.
GSK3 Phosphorylation
Because phospho-Akt functions through phosphorylation and inhibition of GSK3, we further examined the phosphorylation of GSK3. Western blot analysis showed that phospho-GSK3 expression was significantly increased at 1, 6, and 24 hours in the cerebral cortex and at 6 hours in the caudate putamen compared with control samples (Figure 2A). In the hippocampus, phospho-GSK3 was significantly increased at 24 hours compared with the samples from 6 hours. Immunoreactivity of GSK3 was not significantly changed. An immunohistochemistry study showed that phospho-GSK3 expression was increased after SAH, and it peaked at 1 hour in the cerebral cortex and at 6 hours in the caudate putamen (Figure 2B). In the hippocampus, phospho-GSK3 expression temporarily decreased at 1 and 6 hours and increased thereafter. These results indicate that phosphorylation of GSK3 is accelerated after SAH, and the timing and distribution of this phosphorylation is similar to Akt phosphorylation.
Neuronal Expression of Phospho-Akt
Double immunofluorescence for phospho-Akt (serine-473) and NeuN demonstrated that phospho-Akt–positive cells colocalized mainly with neurons in the cerebral cortex 1 hour after SAH (Figure 3). This result suggests that phosphorylation of Akt occurred mainly in neurons.
DNA Fragmentation After SAH
DNA fragmentation was analyzed with in situ TUNEL (Figure 4A) and a commercial enzyme immunoassay (cell death assay; Figure 4B) 24 hours after SAH. No TUNEL-positive cells were observed in any regions of the control brains (Figure 4A). Twenty-four hours after SAH, diffuse TUNEL-positive cells were observed in the cerebral cortex, including the interhemispheric region and the caudate putamen. More TUNEL-positive cells were observed in the cerebral cortex than in the caudate putamen, although statistical analysis was not performed. In contrast, no TUNEL-positive cells were seen in the hippocampus. The cell death assay revealed that DNA fragmentation was significantly increased in the cerebral cortex and caudate putamen 24 hours after SAH (Figure 4B). More DNA fragmentation was observed in the cerebral cortex than in the caudate putamen. DNA fragmentation was not significantly increased in the hippocampus. These results indicate that apoptotic cell death was increased in the cerebral cortex and caudate putamen 24 hours after SAH.
Double Labeling With Phospho-Akt (Serine-473) and TUNEL Staining
Twenty-four hours after SAH, diffuse TUNEL-positive cells spread throughout the cerebral cortex, but these cells did not colocalize with phospho-Akt–positive cells (Figure 5A). In contrast, some cortical neurons showed continuous Akt phosphorylation, which did not colocalize with TUNEL-positive cells (Figure 5B). These results suggest that the cellular population of phospho-Akt (serine-473) is different from that of DNA fragmentation after SAH.
Administration of the PI3-K Inhibitor
To investigate the role of Akt/GSK3 signaling as a downstream effector of the PI3-K pathway after SAH, we administered LY294002. Western blot analysis showed that expression of both phospho-Akt and phospho-GSK3 was significantly decreased in the LY294002-treated animals compared with the vehicle-treated animals in the cerebral cortex 24 hours after SAH (Figure 6A). We then examined DNA fragmentation in these animals to investigate the role of the PI3-K pathway in DNA damage after SAH. DNA fragmentation at 24 hours was significantly increased in the LY294002-treated animals compared with the vehicle-treated animals (Figure 6B). These results indicate that inhibition of Akt/GSK3 signaling through the PI3-K pathway increases acute brain injury after SAH.
Discussion
Accumulating evidence suggests that the Akt signaling pathway plays a crucial role in neuronal survival in several models of neurodegenerative diseases. We have reported that Akt phosphorylation at serine-473 was increased after cerebral ischemia,13 traumatic brain injury,12 and spinal cord injury.11 Inhibition of this phosphorylation by LY294002 facilitated neuronal injury.11,13 Moreover, upregulated Akt phosphorylation inhibited neuronal injury against ischemia and spinal cord injury,11,18 suggesting that activation of Akt signaling is neuroprotective.
In the present study, we evaluated the relation between acute brain injury after SAH and Akt/GSK3 signaling. Interestingly, phosphorylation of Akt (serine-473) and GSK3 (serine-9) was correlated with the severity of brain injury caused by SAH. In the cerebral cortex, where acute brain injury was the most severe, phosphorylation of these proteins was accelerated in the early phase after SAH. In contrast, this phosphorylation tended to temporarily decrease in the early phase and to increase during the late phase in the hippocampus, where no acute brain injury was detected. These results suggest that the timing and distribution of Akt and GSK3 phosphorylation might largely depend on the severity of the stress caused by SAH. We reported that the phospho-Akt level was decreased in the ischemic core, whereas it was enhanced in the ischemic penumbra after focal cerebral ischemia.13 In our SAH model, there were no severe mass lesions that corresponded to the ischemic core in focal cerebral ischemia. Instead, all brain regions might receive moderate to mild damage after SAH. This may be why phosphorylation of Akt and GSK3 increased in all regions and why the timing of this phosphorylation was correlated with the severity of damage. Temporary dephosphorylation of Akt after the insult to the hippocampal CA1 subregion was also observed after global cerebral ischemia.19 The exact machinery of this temporary dephosphorylation and its effect on subsequent hippocampal damage remain unclear. Although mechanisms downstream of the Akt/GSK3 pathway are largely unknown, GSK3-mediated phosphorylation of -catenin20 or the tau protein21 seems to enhance neuronal cell death. Further investigation of effectors downstream of the Akt/GSK3 pathway could help clarify the mechanisms of acute brain injury after SAH.
In the present study, we used an established rat model of SAH, in which we perforated the internal carotid artery with a 3-0 monofilament suture.14 This is a noncraniotomy model that theoretically mimics the mechanism of clinical SAH, although its drawbacks are large variations in the severity of bleeding and a high mortality rate.16 This perforation SAH model has been used for the study of acute brain injury in rats.4,22 Apoptotic cell death has been linked to acute brain injury after SAH3–5 and might be the therapeutic target for clinical SAH. After initial bleeding, apoptotic cell death might be caused by acute cerebral ischemia,4 subarachnoid blood toxicity,23 or acute vasospasm.24 Distribution of apoptotic cell death after SAH is controversial.3,4 In the present study, apoptotic cell death assessed by TUNEL was located mainly in the cerebral cortex and the caudate putamen. In contrast, no TUNEL-positive cells were observed in the hippocampus. These results were in accord with the results of the cell death assay, which also detected apoptosis. It has been reported that TUNEL-positive cells were observed in the hippocampus as well as in the cerebral cortex in the same SAH model.3 This discrepancy might be the result of different time points used to assess brain damage after SAH. We humanely killed the animals at 24 hours to assess acute brain injury as described,22 which might not be long enough to assess hippocampal damage. Although intracranial pressure was not specifically evaluated in our study, it rapidly and dramatically increased after SAH in the same perforation model,16,22 resulting in acute global ischemia. Considering the delayed death of hippocampal CA1 neurons after global ischemia, it seems to play a significant role in hippocampal damage after SAH.15 Proximity to the blood clot might be the cause of the different apoptotic patterns between the cerebral cortex and caudate putamen. After SAH, blood is rapidly distributed throughout the brain surface, and the cerebral cortex is covered with a thick blood clot. This subarachnoid blood clot, which has been linked to cell injury and oxidative stress,23 might result in the greater apoptotic cell death in the cerebral cortex compared with the caudate putamen.
The double immunofluorescent study showed that phospho-Akt–positive cells colocalized with NeuN-positive cells, suggesting that phospho-Akt might be associated with the fate of neurons after SAH. TUNEL-positive cells spread throughout the cerebral cortex 24 hours after SAH. Even at 24 hours, some cells continued to be phospho-Akt–positive but did not colocalize with TUNEL-positive cells. Moreover, LY294002 prevented phosphorylation of both Akt and GSK3 and increased DNA damage. These results indicate that inhibition of the Akt/GSK3 pathway via the PI3-K pathway increases neuronal injury, whereas continuous activation of the Akt/GSK3 pathway mediates neuronal survival after SAH. In the double-immunofluorescent study, there were TUNEL-negative and phospho-Akt–negative cells at 24 hours, which might imply the existence of Akt-independent mechanisms of protection against SAH. However, our results suggest that at least activation of the survival Akt/GSK3 pathway could be one of the therapeutic targets for acute brain injury after SAH in a clinical situation.
Acknowledgments
We thank Liza Reola, Bernard Calagui, and Trisha Crandall for technical assistance; Cheryl Christensen for editorial assistance; and Elizabeth Hoyte for figure preparation.
Sources of Funding
This study was supported by National Institutes of Health grants P50 NS14543, R01 NS25372, R01 NS36147, and R01 NS38653 and an American Heart Association Bugher Foundation Award.
Disclosures
None.
References
Broderick JP, Brott T, Tomsick T, Miller R, Huster G. Intracerebral hemorrhage more than twice as common as subarachnoid hemorrhage. J Neurosurg. 1993; 78: 188–191.
Hütter BO, Kreitschmann-Andermahr I, Gilsbach JM. Health-related quality of life after aneurysmal subarachnoid hemorrhage: impacts of bleeding severity, computerized tomography findings, surgery, vasospasm, and neurological grade. J Neurosurg. 2001; 94: 241–251.
Prunell GF, Svendgaard N-A, Alkass K, Mathiesen T. Delayed cell death related to acute cerebral blood flow changes following subarachnoid hemorrhage in the rat brain. J Neurosurg. 2005; 102: 1046–1054.
Park S, Yamaguchi M, Zhou C, Calvert JW, Tang J, Zhang JH. Neurovascular protection reduces early brain injury after subarachnoid hemorrhage. Stroke. 2004; 35: 2412–2417.
Matz PG, Fujimura M, Chan PH. Subarachnoid hemolysate produces DNA fragmentation in a pattern similar to apoptosis in mouse brain. Brain Res. 2000; 858: 312–319.
Nau R, Haase S, Bunkowski S, Brück W. Neuronal apoptosis in the dentate gyrus in humans with subarachnoid hemorrhage and cerebral hypoxia. Brain Pathol. 2002; 12: 329–336.
Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y, Greenberg ME. Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell. 1997; 91: 231–241.
Alessi DR, Andjelkovic M, Caudwell B, Cron P, Morrice N, Cohen P, Hemmings BA. Mechanism of activation of protein kinase B by insulin and IGF-1. EMBO J. 1996; 15: 6541–6551.
Cross DAE, Alessi DR, Cohen P, Andjelkovich M, Hemmings BA. Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B [Letter]. Nature. 1995; 378: 785–789.
Hetman M, Cavanaugh JE, Kimelman D, Xia Z. Role of glycogen synthase kinase-3 in neuronal apoptosis induced by trophic withdrawal. J Neurosci. 2000; 20: 2567–2574.
Yu F, Sugawara T, Maier CM, Hsieh LB, Chan PH. Akt/Bad signaling and motor neuron survival after spinal cord injury. Neurobiol Dis. 2005; 20: 491–499.
Noshita N, Lewen A, Sugawara T, Chan PH. Akt phosphorylation and neuronal survival after traumatic brain injury in mice. Neurobiol Dis. 2002; 9: 294–304.
Noshita N, Lewen A, Sugawara T, Chan PH. Evidence of phosphorylation of Akt and neuronal survival after transient focal cerebral ischemia in mice. J Cereb Blood Flow Metab. 2001; 21: 1442–1450.
Bederson JB, Germano IM, Guarino L. Cortical blood flow and cerebral perfusion pressure in a new noncraniotomy model of subarachnoid hemorrhage in the rat. Stroke. 1995; 26: 1086–1091.
Sugawara T, Fujimura M, Morita-Fujimura Y, Kawase M, Chan PH. Mitochondrial release of cytochrome c corresponds to the selective vulnerability of hippocampal CA1 neurons in rats after transient global cerebral ischemia. J Neurosci. 1999; 19: 39: 1–6.
Prunell GF, Mathiesen T, Diemer NH, Svendgaard N-A. Experimental subarachnoid hemorrhage: subarachnoid blood volume, mortality rate, neuronal death, cerebral blood flow, and perfusion pressure in three different rat models. Neurosurgery. 2003; 52: 165–175.
Ostrowski RP, Colohan ART, Zhang JH. Mechanisms of hyperbaric oxygen-induced neuroprotection in a rat model of subarachnoid hemorrhage. J Cereb Blood Flow Metab. 2005; 25: 554–571.
Noshita N, Sugawara T, Lewen A, Hayashi T, Chan PH. Copper-zinc superoxide dismutase affects Akt activation after transient focal cerebral ischemia in mice. Stroke. 2003; 34: 1513–1518.
Yano S, Morioka M, Fukunaga K, Kawano T, Hara T, Kai Y, Hamada J-i, Miyamoto E, Ushio Y. Activation of Akt/protein kinase B contributes to induction of ischemic tolerance in the CA1 subfield of gerbil hippocampus. J Cereb Blood Flow Metab. 2001; 21: 351–360.
Zhang Z, Hartmann H, Do VM, Abramowski D, Sturchler-Pierrat C, Staufenbiel M, Sommer B, van de Wetering M, Clevers H, Saftig P, De Strooper B, He X, Yankner BA. Destabilization of -catenin by mutations in presenilin-1 potentiates neuronal apoptosis. Nature. 1998; 395: 698–702.
Takashima A, Honda T, Yasutake K, Michel G, Murayama O, Murayama M, Ishiguro K, Yamaguchi H. Activation of tau protein kinase I/glycogen synthase kinase-3 by amyloid peptide (25–35) enhances phosphorylation of tau in hippocampal neurons. Neurosci Res. 1998; 31: 317–323.
Kusaka G, Ishikawa M, Nanda A, Granger DN, Zhang JH. Signaling pathways for early brain injury after subarachnoid hemorrhage. J Cereb Blood Flow Metab. 2004; 24: 916–925.
Matz PG, Copin J-C, Chan PH. Cell death after exposure to subarachnoid hemolysate correlates inversely with expression of CuZn-superoxide dismutase. Stroke. 2000; 31: 2450–2458.
Bederson JB, Levy AL, Ding WH, Kahn R, DiPerna CA, Jenkins AL III, Vallabhajosyula P. Acute vasoconstriction after subarachnoid hemorrhage. Neurosurgery. 1998; 42: 352–360.
Lack of JunD Promotes Pressure Overload–Induced Apoptosis, Hypertrophic Growth, and Angiogenesis in the Heart
Epidermal Growth Factor Receptor Blockade Mediates Smooth Muscle Cell Apoptosis and Improves Survival in Rats With Pulmonary Hypertension
Fasudil, a Rho-Kinase Inhibitor, Attenuates Angiotensin II–Induced Abdominal Aortic Aneurysm in Apolipoprotein E–Deficient Mice by Inhibiting Apoptosis and Proteolysis
Functional Role of Phosphodiesterase 3 in Cardiomyocyte Apoptosis
Phosphodiesterase-5 Inhibition With Sildenafil Attenuates Cardiomyocyte Apoptosis and Left Ventricular Dysfunction in a Chronic Model of Doxorubicin Cardiotoxicity
Doxazosin Induces Apoptosis in Cardiomyocytes Cultured In Vitro by a Mechanism That Is Independent of 1-Adrenergic Blockade
Vasospasm and p53-Induced Apoptosis in an Experimental Model of Subarachnoid Hemorrhage
Metalloporphyrin-Based Superoxide Dismutase Mimic Attenuates the Nuclear Translocation of Apoptosis-Inducing Factor and the Subsequent DNA Fragmentation After Permanent Focal Cerebral Ischemia in Mice
In Tuberculous Pleural Effusions, Activated Neutrophils Undergo Apoptosis and Acquire a Dendritic CellLike Phenotype
Morphine Potentiates HIV-1 gp120Induced Neuronal Apoptosis