Historically, cell death has been divided into two generic categories: apoptosis, which requires energy and in which the cell plays an active role, and necrosis, which occurs accidentally, does not require energy consumption and is considered as a passive, uncontrolled cell death program. Among the conceptually opposite cell death forms, apoptosis is the best understood. This death program has been defined as developmentally programmed and ordered cellular response. Apoptosis is initiated by cell rounding and subsequent detachment from the surrounding cells. Chromatin condenses into "crescent-like" forms abutting the inner nuclear membrane. Plasma membrane convolutes and gives rise to characteristic vesicles containing cellular organelles and cytoplasm, known as the "apoptotic bodies." Apoptosis is generally not accompanied by inflammation since macrophages or neighbouring cells engulf the formed apoptotic bodies before the loss of plasma membrane integrity (Kerr et al. 1972). In contrast to apoptosis, necrosis is characterized by disruption of the plasma membrane with a subsequent water influx and leakage of cell content to the surroundings. Cell death by necrosis can elicit an inflammatory response (Edinger and Thompson 2004).
It now seems clear that the above-described apoptosis versus necrosis dichotomy is an artificial division, and that "programmed" cell death is a more complex physiological process than initially thought. The cell can use different mechanisms/ pathways with underlying apoptotic or necrotic features to accomplish its proper
Institut de Neurobiologie de la Méditerranée (INMED)/U29 INSERM, Parc Scientifique de Luminy, 163, avenue de Luminy - BP 13, 13273, Marseille, CEDEX 09, France e-mail: [email protected]
INSERM, U872, Team 19, Centre de Recherche des Cordeliers, Paris, France Université Pierre et Marie Curie-Paris 6, UMRS 872, Paris, France Université Paris Descartes, UMRS 872, Paris, France e-mail: [email protected]
D.G. Fujikawa (ed.), Acute Neuronal Injury: The Role of Excitotoxic 35
Programmed Cell Death Mechanisms, DOI 10.1007/978-0-387-73226-8_3, © Springer Science+Business Media, LLC 2010
demise in a controlled manner (Jaattela 2002; Broker et al. 2005; Jaattela and Tschopp 2003; Okada and Mak 2004; Golstein and Kroemer 2007; Golstein et al. 2003; Gozuacik and Kimchi 2007; Krantic et al. 2005, 2007; Susin et al. 1996, 1999, 2000; Lorenzo and Susin 2004; Barkla and Gibson 1999; Chautan et al. 1999; Holler et al. 2000; Colbourne et al. 1999; Nicotera et al. 1999a, b; Saelens et al. 2005; Vanden Berghe et al. 2004; Aarts et al. 2003; Moubarak et al. 2007; Festjens et al. 2006; Srivastava et al. 2007; Borst and Rottenberg 2004; Bras et al. 2007; Boujrad et al. 2007; Niquet et al. 2006; Li et al. 2007; Han et al. 2007; Fujikawa et al. 2000a).
To better understand the complexity of neuronal cell death linked to acute injury, we will first briefly discuss two of the most broadly used PCD classifications (Bredesen 2007). One of them bears the overall morphology of the dying cell and divides PCDs into three types: type 1 or "classical" apoptosis, type 2 or "autophagic" death, and type 3 or vesicular, nonlysosomal degradation (Schweichel and Merker 1973; Clarke 1990). The second classification is based on the type of chromatin condensation and comprises classical apoptosis, apoptosis-like PCD, and necrosislike PCD (Jaattela and Tschopp 2003). According to this classification, apoptosis is characterized by compact (stage 2) chromatin condensation, whereas apoptosis-like PCD displays partial/peripheral (stage 1) chromatin condensation. In necrosislike PCD chromatin is either noncondensed at all, or only slightly granulated (Jaattela and Tschopp 2003; Leist and Jaattela 2001).
Type 1 PCD or classical apoptosis, is the best-characterized cell death outcome at both genetic and biochemical levels (Danial and Korsmeyer 2004; Hengartner 2000). This mode of PCD involves the activation of a family of cysteine proteases named caspases (Thornberry and Lazebnik 1998) and it can be triggered via "death receptors" along the "extrinsic pathway" or via mitochondrial "intrinsic pathway" (Green and Reed 1998). The extrinsic pathway is initiated by interaction of death receptors (Fas, TNF-R) with their cognate ligands (FasL, TNF). Death receptors contain death domains (DD) and death effector domains (DED), which, upon ligand binding, engage into homeotypic protein-protein interactions. Receptor assembly into oligomers results in the formation of a Death-Inducing Signaling Complex (DISC) through conformational alterations. Adaptor proteins (e.g. TRADD, RaiDD, FADD) then associate with the relevant receptors via their DD and DED domains. Procaspase-8/-10 are further recruited to the complex and proteolytically cross-activated by the local accumulation of their pro-enzymes in DISC vicinity (Jaattela and Tschopp 2003; Danial and Korsmeyer 2004). The intrinsic pathway consists of two branches. One is triggered by endoplasmic reticulum stress resulting from the accumulation of unfolded or misfolded proteins and is associated with caspase-12 activation (Rao et al. 2002). The second is initiated by mitochondrial outer membrane permeabilization (MOMP), and is regulated by the Bcl-2 (B-cell lymphoma-2) family of proteins (Tsujimoto 2002). By regulating MOMP, these proteins control the release of key death regulatory proteins from mitochondria (Danial and Korsmeyer 2004). Although many aspects of this regulation remain poorly understood, it is clear that multidomain Bcl-2 proapoptotic proteins such as Bax (Bcl-2-associated X protein) and Bak (Bcl-2 homologous antagonist killer) are required for this cell death induction because their double genetic ablation confers resistance to type 1 PCD (Letai et al. 2002;
Scorrano and Korsmeyer 2003). Indeed, Bax or Bak create pores in the mitochondrial outer membrane. These pores allow the release of cytochrome c and other proteins, such as Smac/DIABLO or Omi/HtrA2, into the cytoplasm. The process progresses further by antagonistic interaction of Smac/DIABLO or Omi/HtrA2 with caspase inhibitors, which results in their inhibition and indirectly allows for caspase activation. The latter can also be achieved directly by cytochrome c binding to Apaf-1 into complexes that aggregate to form apoptosomes. This aggregation requires energy from ATP. The apoptosomes activate caspase-9, which then cleaves and activates caspase-3 and caspase-7. These executioner caspases engage a cascade of proteolytic activity that leads to the digestion of structural proteins and DNA degradation. This "caspase-dependent" DNA degradation, which relates to Caspase activated DNase (CAD), leads to a typical "apoptotic" inter-nucleosomal DNA fragmentation (180-200 bp) and chromatin compaction into spherical or crescent masses abutting on the nuclear envelope (Hengartner 2000).
Autophagic type 2 PCD is characterized morphologically by the appearance of autophagic, double membraned vacuoles. These cytoplasmic vesicles contain cellular organelles, such as mitochondria or endoplasmic reticulum (Gozuacik and Kimchi 2007). It is important to distinguish "autophagic death" from autophagy. Indeed, autophagy is an adaptive process involved in survival response to low-nutrient states (in hypoglycemia, after axonal injury, etc), allowing for the catabolism of cellular constituents to produce energy or to remove damaged organelles. However, autophagy is also associated with cell death with either necrotic or apoptotic phenotype [for review, see Nixon (2006)]. The criteria for detection of this type of PCD were until recently exclusively morphological and even the existence of this PCD outcome is not yet generally accepted. It is in particular not clear whether autophagy represents a consequence (cell dies following autophagy) or a cause (cell death requires autophagy) of the cell death process. However, identification of the autophagy-related genes such as Atg5 or Atg6/Becn1 and the fact that their inactivation precludes "autophagic" PCD (Shimizu et al. 2004) strengthened the experimental evidence supporting the existence of this PCD as a specific and distinct cell death modality.
Type 3 PCD has been initially characterized by the presence of swelling organelles followed by the appearance of "empty" spaces in the cytoplasm which merge and make connections with the extra-cellular space (Schweichel and Merker 1973). The plasma membrane is fragmented, but nuclear disintegration is retarded. This type of PCD has subsequently been subdivided into sub-types 3A (i.e. "non-lysosomal disintegration") and 3B (i.e. "cytoplasmic degeneration") (Clarke 1990). Sub-type 3A is characterized by nuclear disintegration whereas sub-type 3B displays karyolysis (Clarke 1990; Beaulaton and Lockshin 1982).
Above-described types 2 and 3 PCD occur without pronounced nuclear chroma-tin condensation. This nuclear feature is used as the main criterion to classify a given type of cell death as necrosis-like PCD (Jaattela and Tschopp 2003; Leist and Jaattela 2001). Apoptosis-like PCD can be considered as an intermediate morphological phenotype. Indeed, chromatin condensation, which accompanies apoptosis-like PCD, is not as pronounced as in apoptosis but it is more prominent than in necrosis-like PCD.
Apoptosis-like PCD occurs through a caspase-independent mitochondrial route. Apoptosis-inducing factor (AIF) is currently considered the major apoptosis-like PCD effector (Krantic et al. 2007; Susin et al. 1996, 1997, 1999, 2000; Boujrad et al. 2007; Lorenzo and Susin 2007; Dawson and Dawson 2004; Yu et al. 2002; Hong et al. 2004). Upon MOMP, AIF is released from the intermembrane mitochondrial space. The kinetics of AIF release is, however, slower than that of cytochrome c release (Munoz-Pinedo et al. 2006) and caspase-dependent apoptosis appears consistently as more rapid cellular response to death-inducing insult. It is not clear yet whether apoptosis-like PCD represents a secondary cell death outcome, expressed only when apoptosis is inhibited. Indeed, it is still unknown whether AIF release can occur without previous cytochrome c release or under conditions in which released cytochrome c concentration is kept low by mitochondrial recovery. Such a mechanism has been suggested to occur in terminally differentiated rabbit aortic smooth muscle cells at least in vitro (Seye et al. 2004). In this light, the reversibility of cytochrome c release is of particular importance in neurons since these long-lived postmitotic cells can survive cytochrome c release step, at least during a limited time period (Martinou et al. 1999).
After release from the intermembrane mitochondrial space, AIF translocates to the nucleus, leading to large-scale DNA degradation into 50-200 kb fragments (Cregan et al. 2004). Given the absence of an intrinsic endonuclease activity (Mate et al. 2002; Ye et al. 2002), the DNA-degrading capacity of AIF relies on the recruitment of downstream nucleases, such as cyclophilin A (cyp A) (Cande et al. 2004) or endonuclease G (EndoG) (Wang et al. 2002; Bajt et al. 2006; Whiteman et al. 2007).
Necrosis-like PCD occurs by yet poorly understood molecular mechanisms, but it is usually independent of caspase activation, with some exceptions [see Edinger and Thompson (2004), Sperandio et al. (2000), Boise and Collins (2001), Meurette et al. (2007), and Zong and Thompson (2006)]. Note that, although necrosis-like PCD (Edinger and Thompson 2004; Zong and Thompson 2006) as well as programmed necrosis (Moubarak et al. 2007; Boujrad et al. 2007) are morphologically indistinguishable from necrosis, they both differ from that nonregulated, accidental process described by Kerr, Wyllie, and Curie in 1972 (Kerr et al. 1972). These morphological resemblances include principally the fact that in accidental necrosis (Edinger and Thompson 2004), as in necrosis-like PCD (Jaattela and Tschopp 2003) and programmed necrosis [as described recently by us (Moubarak et al. 2007; Boujrad et al. 2007) and others (Vande Velde et al. 2000; Hirt et al. 2000; Gharibyan et al. 2007)], the integrity of the plasma membrane is lost early. Nevertheless, it should be stressed that this loss of plasma membrane integrity occurs in an orderly and similar way in all cases, although its kinetics is certainly different between the programmed and accidental modalities of necrosis. Thus, plasma membrane becomes permeable (as measured by propidium iodide, PI, diffusion) relatively early after initiation of the death process (Chen et al. 2001a; Liu and Schnellmann 2003) whereas leakage of cytoplasmic proteins, such as lactate dehydrogenase (LDH), indicates the terminal phase of permeabilization (Nishimura and Lemasters 2001) in all known death modalities with a necrotic phenotype.
The difference between the molecular mass of PI (0.67 kDa) and LDH (140 kDa) suggests that the size of the membrane pores increases as the death process progresses. If proceeding slowly, such process of membrane permeabilization appears compatible with the programmed (or controlled) character of certain necrotic death outcomes. For the sake of clarity, in the forthcoming text we will use exclusively the term of programmed necrosis to point out the PCD outcome occurring with a necrosis-like morphological phenotype through activation of identified effectors/ proteins (Moubarak et al. 2007).
Was this article helpful?