Chapter 14 Neurodegeneration in the Neonatal Brain
Introduction
Underlying mechanisms of neuronal cell death are common to both immature and mature brains; however, important developmental differences in propinquity to cell cycle, metabolic requirements, connectivity, basal and stimulated expression of cell death proteins, excitatory receptor expression and subunit composition, antioxidant mechanisms, synaptic density, and neurotrophin requirement account for significant differences in injury response between the immature and mature central nervous system (CNS) [Corbett et al., 1993; Martin, 2001, 2002; Ferriero, 2004; Johnston, 2005, 2009; Northington et al., 2005; Zhu et al., 2005; Kostovic and Jovanov-Milosevic, 2006]. In this chapter, we will review information obtained on basic cell death processes from both mature and immature model systems, but highlight findings from studies on the response of the immature brain to injury. While clearly important to this topic, a discussion of the contribution of inflammation to injury, recovery, and developmental neuroplasticity following neonatal brain injury is outside the scope of this chapter. A recent review of this topic is available elsewhere [Vexler and Yenari, 2009] and in Chapter 15.
Types of Cell Death
Cells can die by different processes. These processes have been classified generally into two distinct categories, called necrosis and apoptosis. These forms of cellular degeneration were classified originally as different because they appeared different morphologically under a microscope (Figure 14-1). Necrosis is a lytic destruction of individual cells or groups of cells, while apoptosis (derived from a Greek word for “dropping of leaves from trees”) is an orderly and compartmental dismantling of single cells or groups of cells into consumable components for nearby cells. Apoptosis is an example of programmed cell death (PCD) that is an adenosine triphosphate (ATP)-driven (sometimes gene transcription-requiring) form of cell suicide often committed by demolition enzymes called caspases, but other apoptotic and nonapoptotic, caspase-independent forms of PCD exist [Lockshin and Zakeri, 2002]. Apoptotic PCD is instrumental in developmental organogenesis and histogenesis and adult tissue homeostasis, functioning to eliminate excess cells. Each day in normal humans, estimates reveal that between 50 to 70 billion cells in adults and 20 to 30 billion cells in a child between the ages of 8 and 14 die due to apoptosis [Gilbert, 2006]. Another form of cell degeneration seen first with yeast and then in metazoans has been called autophagy [Klionsky and Emr, 2000]. Autophagy is an intracellular catabolic process that occurs by lysosomal degradation of damaged or expendable organelles. Necrosis and apoptosis both differ morphologically (see Figure 14-1) and mechanistically from autophagy [Klionsky and Emr, 2000; Lockshin and Zakeri, 2002].
Fig. 14-1 Cell death continuum.
After its initial description by Portera-Cailliau et al., the continuum concept, in its original form, organized cell death as a linear spectrum with apoptosis and necrosis at the extremes and different syncretic hybrid forms in between (top) [Martin et al., 1998]. Subsequently we have found that this concept is fully realized in the neonatal rat brain following hypoxia-ischemia (HI), with clearly apoptotic (A) and necrotic (G) cells found intermixed with the hybrid forms (B–F). Cells at the extremes have the well-described structures of necrosis (G – swelling and vacuolation of organelles, loss of cell membrane integrity, maintenance of nuclear membrane integrity, random digestion of chromatin) and apoptosis (A – condensation and darkening of cytoplasm within intact cell membrane, intact organelles until late phases of apoptosis, compaction of chromatin into few uniformly dense and rounded aggregates and loss of nuclear membrane). Cells with irregular chromatin condensation, organized in a “clockface” pattern around an intact nuclear membrane (E, F), may or may not have preservation of the cytoplasmic membrane, and their cytoplasmic organelles are disrupted. This structure has commonly been reported in models of excitotoxicity. Increasing organization of chromatin into regular crescenteric or rounded aggregates, dissolution of the nuclear membrane, and preservation of the cytoplasmic membrane, with or without swelling of cytoplasmic organelles (B–D), is the rule as cell death forms more closely mimic apoptosis. Autophagocytic neurons with large numbers of cytoplasmic vacuoles, partially condensed nuclear chromatin, and preservation of cellular integrity are found after neonatal HI (H). Autophagocytic cell death and apoptosis exist on a continuum. Drugs known to inhibit or promote autophagy can also modulate cell death along the apoptosis–necrosis continuum.
More recently, the morphological and molecular regulatory distinctions between the different forms of cell death have become blurred and uncertain due to observations made on degenerating neurons in vivo and to a new concept that attempts to accommodate these observations. This concept, in its original form, posits that cell death exists as a continuum with necrosis and apoptosis at opposite ends of a spectrum, with hybrid forms of degeneration manifesting in between (see Figure 14-1) [Portera-Cailliau, et al., 1997a, b; Martin et al., 1998; Martin, 1999]. For example, the degeneration of neurons in diseased or damaged human and animal nervous systems is not always strictly necrosis or apoptosis, according to the traditional binary classification of cell death, but also occurs as intermediate or hybrid forms with coexisting morphological and biochemical characteristics that lie in a structural continuum, with necrosis and apoptosis at the two extremes [Portera-Cailliau et al., 1997a, b]. Thus, neuronal cell death can be syncretic. The different processes leading to the putative different forms of cell death can be activated concurrently with graded contributions of the different cell death modes to the degenerative process.
The in vivo reality of a neuronal cell death continuum was revealed first in neonatal and adult rat models of glutamate receptor excitotoxicity [Portera-Cailliau et al., 1997a, b] and then very nicely in rodent models of neonatal hypoxic-ischemic encephalopathy (HIE) [Nakajima et al., 2000; Northington et al., 2001b, 2007]. The hybrid cells can be distinguished cytopathologically by the progressive compaction of the nuclear chromatin into few, discrete, large, irregularly shaped clumps (see Figure 14-1). This morphology contrasts with the formation of few, uniformly shaped, dense, round masses in classic apoptosis and the formation of numerous, smaller, irregularly shaped chromatin clumps in classic necrosis. The cytoplasmic organelle pathology in hybrid cells has a basic pattern that appears more similar to necrosis than apoptosis, but is lower in amplitude than in necrosis (e.g., mitochondrial swelling). Toxicological studies of cultured cells have shown that stimulus intensity influences the mode of cell death [Lennon et al., 1991; Fernandes and Cotter, 1994; Bonfoco et al., 1995], such that apoptosis can be induced by injurious stimuli of lesser amplitude than insults causing necrosis [Raffray and Cohen, 1997], but the cell death modes were still considered distinct [Bonfoco et al., 1995].
Basic research is uncloaking the molecular mechanisms of cell death [Orrenius et al., 2003; Yuan et al., 2003], and, with this, the distinctiveness of different cell death processes, as well as the potential overlap among different cell death mechanisms. Experimental studies on cell death mechanisms, and particularly the cell death continuum, are important because they could lead to the rational development of molecular mechanism-based therapies for treating neonatal hypoxic-ischemic encephalopathy (HIE). The different categories of cell death are discussed below.
Necrosis
Cell death caused by cytoplasmic swelling, nuclear dissolution (karyolysis), and lysis has been classified traditionally as necrosis [Trump and Berezesky, 1996]. Cell necrosis (sometimes termed oncosis) [Majno and Joris, 1995], results from rapid and severe failure to sustain cellular homeostasis, notably cell volume control [Trump et al., 1965]. The process of necrosis involves damage to the structural and functional integrity of the cell plasma membrane and associated enzymes, e.g., Na+, K+ ATPase, abrupt influx and overload of ions (e.g., Na+ and Ca2+) and H2O, and rapid mitochondrial damage and energetic collapse [Bonfoco et al., 1995; Leist et al., 1997; Martin et al., 2000; Golden et al., 2001]. Metabolic inhibition and oxidative stress from reactive oxygen species (ROS) are major culprits in triggering necrosis. Inhibitory crosstalk between ion pumps causes pronecrotic effects when Na+, K+ ATPase “steals” ATP from the plasma membrane Ca2+ ATPase, resulting in Ca2+ overload [Castro et al., 2006].
The morphology and some biochemical features of classic necrosis in neurons are distinct (see Figure 14-1). The main features are swelling and vacuolation/vesiculation of organelles, destruction of membrane integrity, random digestion of chromatin due to activation of proteases and deoxyribonucleases (DNases), and dissolution of the cell. The overall profile of the moribund cell is maintained generally as it dissolves into the surrounding tissue parenchyma and induces an inflammatory reaction in vivo. In necrosis, dying cells do not bud to form discrete, membrane-bound fragments. The nuclear pyknosis and karyolysis appear as condensation of chromatin into many irregularly shaped, small clumps, sharply contrasting with the formation of few, uniformly dense and regularly shaped chromatin aggregates that occurs in apoptosis. In cells undergoing necrosis, genomic DNA is digested globally because protease that digest histone proteins that protect DNA, and DNases are coactivated to generate many randomly sized fragments seen as a DNA “smear.” These differences in the cytoplasmic changes and condensation and digestion of nuclear chromatin in pure apoptosis and pure necrosis are very diagnostic.
Recent work has shown that cell necrosis might not be as chaotic or random as envisioned originally, but can involve the activation of specific signaling pathways to eventuate in cell death [Proskuryakov et al., 2003]. For example, DNA damage can lead to poly(ADP-ribose) polymerase activation and ATP depletion, energetic failure, and necrosis [Ha and Snyder, 2000]. Other pathways for “programmed” necrosis involve death receptor signaling through receptor interacting protein 1 (RIP1) kinase and mitochondrial permeability transition (Figure 14-2) [Crompton, 1999; Festjens et al., 2006; Hitomi et al., 2008].
Mitochondrial Ca2+ overload, excessive oxidative stress, and decreases in the electrochemical gradient, ADP and ATP, can favor mitochondrial permeability transition, which is defined by disruption of the proton-motive force [Crompton et al., 1998; Crompton, 1999; van Gurp et al., 2003]. This disruption involves the so-called mitochondrial permeability transition pore (mPTP), which functions as a voltage, thiol, and Ca2+ sensor. The mPTP is a large polyprotein transmembrane channel formed at contact sites between the inner mitochondrial membrane and the outer mitochondrial membrane (Figure 14-3). The complete components of the mPTP are still controversial. The primary components of the mPTP are the voltage-dependent anion channel (VDAC; also called porin) in the outer mitochondrial membrane and the adenine nucleotide translocator (ANT) in the inner mitochondrial membrane [Crompton et al., 1998]. The VDAC makes the inner mitochondrial membrane permeable to most small hydrophilic molecules <5 kDa for free exchange of respiratory chain substrates. The ANT mediates the exchange of ADP for ATP. During normal mitochondrial function the intermembrane space separates the outer and inner mitochondrial membranes and the VDAC and the ANT do not interact, or interact only transiently in a state described as “flicker” [Crompton et al., 1998]. When the mPTP is in the open state, it permits influx of solutes of ≤1500 Da and H2O into the matrix, resulting in depolarization of mitochondria and dissipated proton electrochemical gradient. Consequently, the inner mitochondrial membrane loses its integrity and oxidative phosphorylation is uncoupled. When this occurs, oxidation of metabolites by O2 proceeds with electron flux not coupled to proton pumping, resulting in further dissipation of transmembrane proton gradient and ATP production, production of ROS, and large-amplitude mitochondrial swelling triggering necrosis or apoptosis [van Gurp et al., 2003]. Several proteins regulate the mPTP. Cyclophilin D is one of these proteins found in the mitochondrial matrix and it interacts reversibly with the ANT. Inactivation of cyclophilin D can block mitochondrial swelling and cellular necrosis induced by Ca2+ overload and ROS in the adult [Baines et al., 2005; Nakagawa et al., 2005] but cyclophilin D appears to have prosurvival functions in the hypoxia-ischemia (HI)-injured immature brain [Wang et al., 2009]. Another protein that causes mPTP opening is BNIP3, which can integrate into the outer mitochondrial membrane and can trigger necrosis [Vande Velde et al., 2000].
Apoptosis
Apoptosis is a form of PCD because it is carried out by active, intrinsic transcription-dependent [Tata, 1966] or transcription-independent mechanisms involving specific molecules. Apoptosis should not be used as a synonym for PCD because nonapoptotic forms of PCD exist [Schwartz et al., 1993; Amin et al., 2000]. Apoptosis is only one example of PCD. It is critical for the normal growth and differentiation of organ systems in vertebrates and invertebrates [see Jacobson, 1991, regarding Ernst’s discovery of developmental PCD; Glucksmann, 1951; Lockshin and Williams, 1964; Saunders, 1966]. The structure of apoptosis is similar to the type I form of PCD described by Clarke [Clarke, 1990]. In physiological settings in adult tissues, apoptosis is a normal process, occurring continuously in populations of cells that undergo slow proliferation (e.g., liver and adrenal gland) or rapid proliferation (e.g., epithelium of intestinal crypts) [Wyllie et al., 1980; Bursch et al., 1990]. Apoptosis is a normal event in the immune system when lymphocyte clones are deleted after an immune response [Nagata, 1999]. Kerr and colleagues were the first to describe apoptosis in pathological settings [Kerr et al., 1972], but many descriptions were made prior to this time in studies of developing animal systems [Lockshin and Zakeri, 2001].
Classical apoptosis has a distinctive structural appearance (see Figure 14-1). The cell condenses and is dismantled in an organized way into small packages that can be consumed by nearby cells. Nuclear breakdown is orderly. The DNA is digested in a specific pattern of internucleosomal fragments (see Figure 14-1), and the chromatin is packaged into sharply delineated, uniformly dense masses that appear as crescents abutting the nuclear envelope or as smooth, round masses within the nucleus (see Figure 14-1). The execution of apoptosis is linked to Ca2+-activated DNases [Wyllie, 1980], one being DNA fragmentation factor 45 (DFF45) [Liu et al., 1997], which digests genomic DNA at internucleosomal sites only (because proteases that digest histone proteins remain inactivated and the DNA at these sites is protected from DNases) to generate a DNA “ladder” (see Figure 14-1). However, the emergence of the apoptotic nuclear morphology can be independent of the degradation of chromosomal DNA [Sakahira et al., 1999]. Cytoplasmic breakdown is also orderly. The cytoplasm condenses, (as reflected by a darkening of the cell in electron micrographs, see Figure 14-1), and subsequently the cell shrinks in size, while the plasma membrane remains intact. During the course of these events, it is believed that the mitochondria are required for ATP-dependent processes. Subsequently, the nuclear and plasma membranes become convoluted, and then the cell undergoes a process called budding. In this process, the nucleus, containing smooth, uniform masses of condensed chromatin, undergoes fragmentation in association with the condensed cytoplasm, forming cellular debris (called apoptotic bodies) composed of pieces of nucleus surrounded by cytoplasm with closely packed and apparently intact organelles. Apoptotic cells display surface markers (e.g., phosphatidylserine or sugars) for recognition by phagocytic cells. Phagocytosis of cellular debris by adjacent cells is the final phase of apoptosis in vivo.
Variants of classical or nonclassical apoptosis can occur during nervous system development [Pilar and Landmesser, 1976; Clarke, 1990] and also frequently in pathophysiological settings of nervous system injury and disease [Portera-Cailliau et al., 1997a, b; Martin et al., 1998]. Axonal damage (axotomy) and target deprivation in the mature nervous system can induce apoptosis in neurons that is similar structurally, but not identical, to developmental PCD [Martin et al., 1998]. Excitotoxins can induce readily and robustly nonclassical forms of apoptosis in neurons [Portera-Cailliau et al., 1997a, b]. Types of cell death similar to those seen with excitotoxicity occur frequently in pathological cell death resulting from neonatal HI [Martin et al., 2000; Nakajima et al., 2000; Northington et al., 2001b, 2007].
Cells can die by PCD through mechanisms that are distinct from apoptosis [Jacobson, 1991; Schwartz et al., 1993]. The structure of nonapoptotic PCD is similar to the type II or type III forms of cell death described by Clarke [Clarke, 1990]. Interestingly, there is no internucleosomal fragmentation of genomic DNA in some forms of nonapoptotic PCD [Schwartz et al., 1993; Amin et al., 2000].
Autophagy
Autophagy is a mechanism whereby eukaryotic cells degrade their own cytoplasm and organelles [Klionsky and Emr, 2000]. The degradation of organelles and long-lived proteins is carried out by the lysosomal system. Although autophagy functions as a cell death mechanism, it is primarily a homeostatic nonlethal stress response mechanism for recycling proteins to protect cells from low supplies of nutrients. Autophagy is classified as type II PCD [Clarke, 1990]. A hallmark of autophagic cell death is accumulation of autophagic vacuoles of lysosomal origin. Autophagy has been seen in developmental and pathological conditions. For example, insect metamorphosis involves autophagy [Lockshin and Zakeri, 1994], and developing neurons can use autophagy as a PCD mechanism [Schweichel and Merker, 1973; Xue et al., 1999]. Degeneration of Purkinje neurons in the mouse mutant Lucher appears to be a form of autophagy, thus possibly linking excitotoxic and autophagic cell deaths to constitutive activation of the GluRδ2 glutamate receptor [Yue et al., 2002].
The molecular controls of autophagy appear common in eukaryotic cells from yeast to human, and it is believed that autophagy evolved before apoptosis [Yuan et al., 2003]. However, most of the work has been done on yeast, with detailed work on mammalian cells only beginning [Mizushima et al., 2002]. Double-membrane autophagosomes for sequestration of cytoplasmic components are derived from the endoplasmic reticulum (ER) or the plasma membrane. Tor kinase, phosphatidylinositol 3 (PI3) kinase, a family of cysteine proteases called autophagins, and death-associated proteins function in autophagy [Bursch, 2001; Inbal et al., 2002]. Autophagic and apoptotic cell death pathways crosstalk. The product of the tumor suppressor gene Beclin1 (the human homolog of the yeast autophagy gene APG6) interacts with the anti-apoptosis regulator Bcl-2 [Liang et al., 1998]. Autophagy can block apoptosis by sequestration of mitochondria. If the capacity for autophagy is reduced, stressed cells die by apoptosis, whereas inhibition or blockade of molecules that function in apoptosis can convert the cell death process into autophagy [Ogier-Denis and Codogno, 2003]. Thus, a mechanistic continuum between autophagy and apoptosis exists (see Figure 14-1).
Autophagy appears to have a significant role in neurodegeneration after neonatal HI that may be insult severity-, time-, and region-specific [Lockshin and Zakeri, 1994; Koike et al., 2008; Ginet, et al., 2009]. Genetic deletion of the atg7 gene results in near-complete protection from HI in adult mice [Koike et al., 2008], and pharmacologic inhibition of autophagy with 3-methyladenine up to 4 hours after focal ischemia is neuroprotective in p12 rats [Puyal et al., 2009]. Conversely, induction of autophagy immediately following neonatal global HI in mice appeared to be an endogenous neuroprotective mechanism in other studies [Carloni et al., 2008]. Pre-insult blockade of autophagy with methyladenine inhibited expression of autophagocytic proteins and switched cell death from apoptotic to necrotic; conversely, enhancing early autophagy with pre-insult administration of rapamycin provided neuroprotection in this model [Carloni et al., 2008]. Interestingly, known neuroprotective preconditioning strategies also increased markers of autophagy [Carloni et al., 2008] while providing the expected neuroprotection. Regional differences in induction of autophagocytic proteins, severity of insult, and timing of drug administration might account for these discrepant results.
Molecular and Cellular Regulation of Apoptosis
Apoptosis is a structurally and biochemically organized form of cell death. The basic machinery of apoptosis is conserved in yeast, hydra, nematode, fruitfly, zebrafish, mouse, and human [Ameisen, 2002]. Our current understanding of the molecular mechanisms of apoptosis in mammalian cells is built on studies by Robert Horvitz and colleagues on PCD in the nematode Caenorhabditis elegans [Metzstein et al., 1998]. They pioneered the understanding of the genetic control of developmental cell death by showing that this death is regulated predominantly by three genes (ced-3, ced-4, and ced-9) [Metzstein et al., 1998]. Several families of apoptosis-regulation genes have been identified in mammals, including the Bcl-2 family [Metzstein et al., 1998; Cory and Adams, 2002], the caspase family of cysteine-containing, aspartate-specific proteases [Wolf and Green, 1999], the p53 gene family [Levrero et al., 2000], cell surface death receptors [Nagata, 1999], and other apoptogenic factors, including cytochrome c, apoptosis inducing factor (AIF), and second mitochondrial activator of caspases (Smac) [Liu et al., 1996; Li et al., 1997; Hegde et al., 2002; Klein et al., 2002; Lockshin and Zakeri, 2002]. Moreover, a family of inhibitors of apoptosis proteins (IAPs) actively blocks cell death, and IAPs are inhibited mitochondrial proteases [Hegde et al., 2002]. Specific organelles have been identified as critical for the apoptotic process, including mitochondria and the ER (see Figure 14-3). In seminal work by Li, Wang, and colleagues, it was discovered that the mitochondrion integrates death signals mediated by proteins in the Bcl-2 family and releases molecules residing in the mitochondrial intermembrane space, such as cytochrome c, which complexes with cytoplasmic proteins (e.g., apoptotic protease activating factor-1 [Apaf-1]) to activate caspase proteases, leading to internucleosomal cleavage of DNA [Liu et al., 1996; Li et al., 1997]. The finding that cytochrome c has a function in apoptosis, in addition to its better-known role in oxidative phosphorylation, was astounding, although foreshadowing clues were available. The release of cytochrome c from mitochondria to the cytosol with concomitant reduced oxidative phosphorylation was described as the “cytochrome c effect” in irradiated cancer cells [van Bekkum, 1957]. The ER, which regulates intracellular Ca2+ levels, participates in a loop with mitochondria to modulate mitochondrial permeability transition and cytochrome c release through the actions of Bcl-2 protein family members [Scorrano et al., 2003].
Bcl-2 Family of Survival and Death Proteins
The bcl-2 proto-oncogene family is a large group of apoptosis regulatory genes encoding about 25 different proteins, defined by at least one conserved B-cell lymphoma (Bcl) homology domain (BH1–BH4 can be present) in their amino acid sequence that functions in protein–protein interactions [Metzstein et al., 1998; Cory and Adams, 2002]. Some of the protein products of these genes (e.g., Bcl-2, Bcl-xL, and Mcl-1) have all four BH1–BH4 domains and are anti-apoptotic. Other gene products, which are pro-apoptotic, are multidomain proteins possessing BH1–BH3 sequences (e.g., Bax and Bak), or proteins with only the BH3 domain (e.g., Bad, Bid, Bim, Bik, Noxa, and Puma) that contains the critical death domain. Bcl-xL and Bax have α-helices resembling the pore-forming subunit of diphtheria toxin [Muchmore et al., 1996]; thus, Bcl-2 family members appear to function by conformation-induced insertion into the outer mitochondrial membrane to form channels or pores that can regulate the release of apoptogenic factors (see Figure 14-3).
The expression of many of these proteins is regulated developmentally, and the proteins have differential tissue distributions and subcellular localizations. Most of these proteins are found in the CNS, but the relative quantities of pro- and anti-apoptotic family members change over time [Li et al., 1997; Shimohama et al., 1998]. It appears that in mouse brain, the relative abundance of the pro-apoptosis Bcl-2 family members declines markedly after the perinatal period, while the anti-apoptotic Bcl-2 family members exhibit stable expression over time in the brain [Shimohama et al., 1998]. The subcellular distributions of Bax, Bak, and Bad in healthy adult rodent CNS tissue [Martin et al., 2003] are consistent with in vitro studies of non-neuronal cells [Wolter et al., 1997; Nechushtan et al., 2001]. Bax, Bad, and Bcl-2 reside primarily in the cytosol, whereas Bak resides primarily in mitochondria (see Figure 14-3). Bcl-2 family members can form homodimers or heterodimers and higher-order multimers with other family members. Bax forms homodimers or heterodimers with Bak, Bcl-2, or Bcl-xL. When Bax and Bak are present in excess, the anti-apoptotic activity of Bcl-2 and Bcl-xL is antagonized. The formation of Bax homo-oligomers promotes apoptosis, whereas Bax heterodimerization with either Bcl-2 or Bcl-xL neutralizes its pro-apoptotic activity. Neonatal HI enhances the relative pro-apoptosis balance of Bcl-2 family proteins. Marked increases in mitochondrial Bax occur within 24 hours of HI in neonatal rats, during which time there is no change in the relative amount of Bcl-xL [Northington et al., 2001a] and changes in the subcellular distribution of Bax occur rapidly, prior to the activation of downstream apoptosis-effector mechanisms [Lok and Martin, 2002].
Release of cytochrome c from mitochondria may occur through mechanisms that involve the formation of membrane channels comprised of Bax or Bak [Antonsson et al., 1997] and Bax and the VDAC [Shimizu et al., 2000] (see Figure 14-3). Cytochrome c triggers the assembly of the cytoplasmic apoptosome (a protein complex of Apaf1, cytochrome c, and procaspase-9), which is the engine driving caspase-3 activation in mammalian cells [Li et al., 1997]. Bcl-2 and Bcl-xL block the release of cytochrome c [Kluck et al., 1997; Yang et al., 1997] from mitochondria, and thus the activation of caspase-3 [Liu et al., 1996; Li et al., 1997]. The blockade of cytochrome c release from mitochondria by Bcl-2 and Bcl-xL [Liu et al., 1996; Vander Heiden et al., 1997] is caused by inhibition of Bax channel-forming activity in the outer mitochondrial membrane [Antonsson et al., 1997] or by modulation of mitochondrial membrane potential and volume homeostasis [Vander Heiden et al., 1997]. Bcl-xL also has anti-apoptotic activity by interacting with Apaf-1 and caspase-9, and inhibiting the Apaf-1-mediated autocatalytic maturation of caspase-9 [Hu et al., 1998]. Boo can inhibit Bak- and Bik-induced apoptosis (but not Bax-induced cell death), possibly through heterodimerization and by interactions with Apaf-1 and caspase-9 [Song et al., 1999]. Bax and Bak double-knockout cells are completely resistant to mitochondrial cytochrome c release during apoptosis [Wei et al., 2001]. BH3-only proteins, such as Bim, Bid, Puma, and Noxa, appear to induce a conformational change in Bax or they serve as decoys for Bcl-xL, allowing Bax to form pores in the outer mitochondrial membrane [Letai et al., 2002]. The role of the BH3-only proteins in neonatal HI brain injury is unclear. There are no reports of Bid activation in neonatal HI, and in a single study using Bid-deficient mice no protection from neonatal HI brain injury was found [Ness et al., 2006].
Although many studies have focused on how Bcl-2 family members regulate mitochondrial apoptosis, it is now evident that Bcl-2 proteins localize to the ER and also translocate to the nucleus [Lithgow et al., 1994; Zhu et al., 2009]. This finding is relevant to neonatal HIE and excitotoxicity, where ER abnormalities may be particularly important to pathogenesis [Portera-Cailliau et al., 1997a, b; Martin et al., 2000; Puka-Sundvall et al., 2000]. The ER functions to fold proteins, and when this capacity is compromised, an unfolded protein response (UPR) is engaged. The UPR can lead to a return to homeostasis or to cell death. Bak and Bax also operate in the ER and function in the activation of ER-specific caspase-12 [Zong et al., 2003]. Cells lacking Bax and Bak are resistant to ER stress-induced apoptosis [Wei et al., 2001]. Translocation and accumulation of Bcl-2 in the nucleus occurs during a period when the total amounts of Bcl-2 decrease via calpain-dependent mechanisms following neonatal HI in mice [Hu et al., 1998; Zhu et al., 2009]. Overexpression of Bcl-2 and Bcl-xL can block ER stress-induced apoptosis [Murakami et al., 2007], but the function of nuclear Bcl-2 is not known.
Protein phosphorylation regulates the functions of some Bcl-2 family members. Bcl-2 loses its anti-apoptotic activity following serine phosphorylation, possibly because its antioxidant function is inactivated [Haldar et al., 1995]. Bcl-2 phosphorylation at serine 24 in the BH4 domain precedes caspase-3 cleavage following cerebral HI in neonatal rat [Hallin et al., 2006]. In addition to interacting with homologous proteins, Bcl-2 can associate with nonhomologous proteins, including the protein kinase Raf-1 [Wang et al., 1996]. Bcl-2 is thought to target Raf-1 to mitochondrial membranes, allowing this kinase to phosphorylate Bad at serine residues. The phosphatidylinositol 3-kinase (PI3-K)–Akt pathway also regulates the function of Bad [Datta et al., 1997; del Peso et al., 1997] and caspase-9 [Cardone et al., 1998] through phosphorylation. In the presence of trophic factors, Bad is phosphorylated. Phosphorylated Bad is sequestered in the cytosol by interacting with soluble protein 14-3-3 and, when bound to protein 14-3-3, Bad is unable to interact with Bcl-2 and Bcl-xL, thereby promoting survival [Zha et al., 1996]. Conversely, when Bad is dephosphorylated by calcineurin [Wang et al., 1999], it dissociates from protein 14-3-3 in the cytosol and translocates to the mitochondria, where it exerts pro-apoptotic activity. Nonphosphorylated Bad heterodimerizes with membrane-associated Bcl-2 and Bcl-xL, thereby displacing Bax from Bax-Bcl-2 and Bax-Bcl-xL dimers and promoting cell death [Yang et al., 1995]. The phosphorylation status of Bad helps regulate glucokinase activity, thereby linking glucose metabolism to apoptosis [Danial et al., 2003].
Caspases: Cell Demolition Proteases
Caspases (cysteinyl aspartate-specific proteinases) are cysteine proteases that have a near-absolute substrate requirement for aspartate in the P1 position of the peptide bond. Fourteen members have been identified [Wolf and Green, 1999]. Caspases exist as constitutively expressed inactive proenzymes (30–50 kDa) in healthy cells. The protein contains three domains: an amino-terminal prodomain, a large subunit (approximately 20 kDa), and a small subunit (approximately 10 kDa). The inactive proenzymes are present at detectable levels under basal conditions, but at least one study suggests that transcription and increased expression of the proenzyme occur following neonatal hypoxia [Delivoria-Papadopoulos et al., 2008]. By far the greatest control of caspase activity occurs through regulated proteolysis of the proenzyme with “initiator” caspases activating “executioner” caspases, although some caspase proenzymes (e.g., caspase-9) have low activity without processing [Stennicke et al., 1999]. Other caspase family members function in inflammation by processing cytokines [Wolf and Green, 1999]. The prodomain of initiator caspases contains amino acid sequences that are caspase recruitment domains (CARD) or death effector domains (DED), which enable the caspases to interact with other molecules that regulate their activation. Activation of caspases involves proteolytic processing between domains, and then association of large and small subunits to form a heterodimer with both subunits contributing to the catalytic site. Two heterodimers associate to form a tetramer that has two catalytic sites that function independently. Active caspases have many target proteins [Schwartz and Milligan, 1996] that are cleaved during regulated and organized cell death. Caspases cleave nuclear proteins (e.g., DNases, poly(ADP) ribose polymerase, DNA-dependent protein kinase, heteronuclear ribonucleoproteins, transcription factors or lamins), cytoskeletal proteins (e.g., actin and fodrin), and cytosolic proteins (e.g., other caspases, protein kinases, Bid).
In cell models of apoptosis using human non-neuronal cell lines, activation of caspase-3 occurs when caspase-9 proenzyme (also known as Apaf-3) is bound by Apaf-1, which then oligomerizes in a process initiated by cytochrome c (identified as Apaf-2) and either ATP or dATP [Li et al., 1997] (see Figure 14-3). Cytosolic ATP or dATP is a required co-factor for cytochrome c-induced caspase activation. Apaf-1, a 130 kDa cytoplasmic protein, serves as a docking protein for procaspase-9 (Apaf-3) and cytochrome c [Li et al., 1997]. Apaf-1 becomes activated when ATP is bound and hydrolyzed, with the hydrolysis of ATP and the binding of cytochrome c promoting Apaf-1 oligomerization [Zou et al., 1999]. This oligomeric complex recruits and mediates the autocatalytic activation of procaspase-9 (forming the apoptosome), which disassociates from the complex and becomes available to activate caspase-3. In the newborn brain, HI in rat pups and hypoxia in piglets induced caspase 9-mediated caspase 3 activation, which is modulated by neuronal and inducible nitric oxide synthase (nNOS and iNOS) derived-NO and by neuronal nuclear Ca2+ influx [Zhu et al., 2004a; Delivoria-Papadopoulos et al., 2008]. These effects appear to occur at both transcriptional and post-translational levels. Once activated, caspase-3 cleaves multiple proteins, including a protein with DNase activity (i.e., DFF-45). DFF-45 cleavage activates a process leading to the internucleosomal fragmentation of genomic DNA [Liu et al., 1997].
So far, three caspase-related signaling pathways have been identified that can lead to apoptosis [Liu et al., 1996, 1997; Liu et al., 1997, 1998], but crosstalk among these pathways is possible. The intrinsic mitochondria-mediated pathway is controlled by Bcl-2 family proteins. It is regulated by cytochrome c release from mitochondria, promoting the activation of caspase-9 through Apaf-1 and then caspase-3 activation. The extrinsic death receptor pathway involves the activation of cell-surface death receptors, including Fas and tumor necrosis factor (TNF) receptor, leading to the formation of the death-inducible signaling complex (DISC) and caspase-8 activation, which in turn cleaves and activates downstream caspases such as caspase-3, 6, and 7. Caspase-8 can also cleave Bid, leading to the translocation, oligomerization, and insertion of Bax or Bak into the mitochondrial membrane. Another pathway involves the activation of caspase-2 by DNA damage or ER stress as a premitochondrial signal [Robertson et al., 2002].
Caspases are also critical regulators of nondeath functions in cells, notably some maturational processes. These nonapoptotic functions include modulation of synaptic plasticity via involvement in long-term potentiation [Gulyaeva, 2003] and cleavage of AMPA receptor subunits [Lu et al., 2002], and normal differentiation and migration of neurons to the olfactory bulb [Yan et al., 2001]. Transient caspase inactivation for the purpose of neuroprotection following neonatal HI may interfere with ongoing, normal nondeath-related functions of caspases. It is unknown whether inhibition of caspase activity in the acute setting of injury impairs similar functions or interferes with axonal sprouting, and whether there are long-term changes in brain structure or function as a result of acute caspase inhibition. Because of the potential importance of the nonapoptotic functions of caspase-3, it might be more appropriate to block the formation of “stress-induced”, cleaved caspase-3 following injury with selective caspase-8 and 9 inhibitors that would not interfere with basal levels of caspase-3 activity. However, the nonapoptotic functions of caspase-8 and 9 are unknown, and the effects of acute caspase-8 and 9 inhibition on the developing brain, other than neuroprotection, are similarly unknown.
Caspases seem to be involved in the evolution of neonatal brain injury caused by HI, although this may vary between models. Caspase-3 cleavage and activation occur in brain after HI and trauma in neonatal rodents [Hu et al., 2000; Blomgren et al., 2001; Northington et al., 2001a; Felderhoff-Mueser et al., 2002], and after hypoxia in neonatal piglets [Delivoria-Papadopoulos et al., 2008]. The extent of caspase-3 cleavage and activation following brain injury is clearly greater in developing rodents compared to adults [Hu et al., 2000; Zhu et al., 2005]. This principle is replicated in immature and mature neuronal culture systems [Lesuisse and Martin, 2002]. Cerebroventricular injection of a pan-caspase inhibitor or intraperitoneal injection of a serine protease inhibitor 3 hours after neonatal HI in rat has neuroprotective effects [Cheng et al., 1998; Feng and LeBlanc, 2003]. Subsequent studies have shown 30–50 percent decreases in HI-induced tissue loss in neonatal rat brain at 15 days after the insult with non-selective inhibitors of caspase-8 and caspase-9 [Feng et al., 2003a, b]. However, the lack of enzyme specificity of caspase inhibitor drugs prevents unambiguous identification of caspases in mediating brain injury in most studies. The class of irreversible tetrapeptide caspase inhibitors covalently coupled to chloromethylketone, fluoromethylketone, or aldehydes efficiently inhibits other classes of cysteine proteases like calpains [Rozman-Pungercar et al., 2003]. Calpains, Ca2+-activated, neutral, cytosolic cysteine proteases, are highly activated following neonatal HI in rats [Ostwald et al., 1993; Blomgren et al., 2001]. MDL28170, a drug that inhibits calpains and caspase-3, exerts neuroprotective actions in the neonatal rat brain by decreasing necrosis and apoptosis [Kawamura et al., 2005] and thus may be a particularly valuable tool for the treatment of neonatal HI. Cathepsins, cysteine proteases concentrated in the lysosomal compartment, are also likely to be activated based on electron microscopy evidence of lysosomal and vacuolar changes found following neonatal HI [Martin et al., 2000]. More potent, selective, and reversible nonpeptide caspase-3 inhibitors have been developed [Han et al., 2005] and used to protect against brain injury following neonatal HI in rats [Han et al., 2002], but the protective effects were more modest compared to initial reports with nonselective pan-caspase inhibition [Cheng et al., 1998].
Not all forms of apoptotic cell death are caspase-dependent [Beresford et al., 2001; Fan et al., 2003]. The serine protease granzyme A (GrA) mediates a caspase-independent apoptotic pathway [Beresford et al., 2001]. GrA is delivered to target cells through Ca2+-dependent, perforin-generated pores, and activates a DNase (GrA-DNase, non-metastasis factor 23, NM23) that is sequestered in the cytoplasm. NM23 activity is inhibited by the SET complex, which is located in the ER and comprised of the nucleosome assembly protein SET, an inhibitor of protein phosphatase 2A, apurinic endonuclease-1, and a high-mobility group protein (a nonhistone DNA-binding protein that induces alterations in DNA architecture). GrA cleaves components of the SET complex to release activated NM23, which translocates to the nucleus to induce single-strand DNA nicks and cell death that can be apoptotic or nonapoptotic [Fan et al., 2003]. To date, there are no studies of this pathway in neonatal HI.
Inhibitor of Apoptosis Protein Family
The activity of pro-apoptotic proteins must be placed in check to prevent unwanted apoptosis in normal cells. Apoptosis is blocked by the IAP family in mammalian cells [Deveraux et al., 1998; LaCasse et al., 1998; Holcik, 2002]. This family includes X chromosome-linked IAP (XIAP), IAP1, IAP2, neuronal apoptosis inhibitory protein (NAIP), Survivin, Livin, and Apollon. These proteins are characterized by 1–3 baculoviral IAP repeat domains consisting of a zinc finger domain of approximately 70–80 amino acids [Holcik, 2002]. Apollon is a huge (530 kDa) protein that also has a ubiquitin-conjugating enzyme domain. The main identified anti-apoptotic function of IAPs is the suppression of caspase activity [Deveraux et al., 1998]. Procaspase-9 and procaspase-3 are major targets of several IAPs. IAPs reversibly interact directly with caspases to block substrate cleavage. Apollon also ubiquitylates and facilitates proteosomal degradation of active caspase-9 and Smac [Hao et al., 2004]. However, IAPs do not prevent caspase-8-induced proteolytic activation of procaspase-3. IAPs can also block apoptosis by reciprocal interactions with the nuclear transcription factor NF-κβ [LaCasse et al., 1998].
Scant information is available on IAPs in the nervous system and in neonatal brain injury. Survivin is essential for nervous system development in mouse because conditional deletion of survivin gene in neuronal precursor cells causes death shortly after birth, and reduced brain size and severe multifocal degeneration [Jiang et al., 2005]. NAIP is expressed throughout the CNS in neurons [Xu et al., 1997]. XIAP is highly enriched in mouse spinal motor neurons [Martin et al., 2007]. The importance of the IAP gene family in human pediatric neurodegeneration is underscored by the finding that NAIP is partially deleted in a significant proportion of children with spinal muscular atrophy [Roy et al., 1995]. Studies in transgenic XIAP mice indicate that XIAP plays an important role in regulating caspase activity after neonatal HI. Overexpression of XIAP virtually abolished activation of both caspase-9 and 3, and provided a 40 percent reduction in tissue loss in forebrain following neonatal HI in mouse [Wang et al., 2004].
Proteins exist that inhibit mammalian IAPs. A murine mitochondrial protein called Smac and its human ortholog, DIABLO (for direct IAP-binding protein with low pI), inactivate the anti-apoptotic actions of IAPs and thus exert pro-apoptotic actions [Du et al., 2000; Verhagen et al., 2000]. These IAP inhibitors are 23 kDa mitochondrial proteins (derived from 29 kDa precursor proteins processed in the mitochondria) that are released from the intermembrane space and sequester IAPs. High temperature requirement protein A2 (HtrA2), also called Omi, is another mitochondrial serine protease that exerts pro-apoptotic activity by inhibiting IAPs [Suzuki et al., 2004]. HtrA2/Omi functions as a homotrimeric protein that cleaves IAPs irreversibly, thus facilitating caspase activity. The intrinsic mitochondrial-mediated cell death pathway is regulated by Smac and HtrA2/Omi, and both of these proteins co-localize with XIAP in injured brain neurons after neonatal HI in mouse [Wang et al., 2004].
