[Frontiers in Bioscience 1, d1-11, January 1, 1996]


Jonathan L. Tilly, Ph.D.

The Vincent Center for Reproductive Biology, Department of Obstetrics and Gynecology, Massachusetts General Hospital/Harvard Medical School,Boston, MA 02114 USA

Received 12/14/95; Accepted 12/29/95; On-line 1/1/96

2. Physiological Cell Death: Mechanisms and Genes

2.1. Introduction.

The concept of apoptosis suggests that at some point in the destiny of any given cell, an irreversible cascade of gene-directed events is set in motion which permits efficient removal of that cell from its neighboring counterparts without a disruption in normal tissue function [reviewed in 1-3]. Many of the features of physiological cell death have been remarkably conserved through evolution such that the existence of a universal pathway for execution of the cell suicide command has been proposed [1-3]. Unlike the random process of pathological cell death (most often referred to as necrosis) which occurs in contiguous tracts of cells following exposure to highly noxious stimuli, apoptosis proceeds in an orderly fashion and in general does not elicit an immune response at the site of cell loss [4,5]. At its most basic level, a cell dying by apoptosis exhibits a marked reduction in cytoplasmic volume coincident with or immediately followed by nuclear pyknosis resulting from high and low molecular weight genomic DNA cleavage. Intracellular organelles and nuclear remnants are then neatly packaged into plasma membrane-bound vesicles, referred to as apoptotic bodies, and phagocytized by neighboring cells or resident macrophages [4,5]. The time course for apoptosis, from the first discernible morphological change to phagocytic removal, can be remarkably short [6]. Thus, much of our knowledge of the features of apoptosis have been derived from in vitro studies of isolated cells as the detection of apoptosis in vivo is hindered by the efficiency and rapidity of phagocytic clean-up.

Physiological cell death occurs in essentially all multicellular organisms. Functions of cell death include embryonic pattern formation [7], development of the male and female reproductive tracts [8,9], establishment of balanced pre- and post-synaptic neuronal junctions in the developing brain [reviewed in 10], removal of autoreactive T-cells from the thymocyte cell repertoire [reviewed in 11], regression of post-lactational breast tissue [12], and a general homeostatic maintenance of proper cell numbers in most tissues [1,2,4]. Of equal importance, disruptions in the normal course of cell death have been linked to many pathological disorders ranging from tumorigenesis (abnormally low cell death rates; [13]) to Alzheimer's disease (excessive cell death; [10]). This has served to fuel the fire for research by a large number of investigators on the underlying events that can be manipulated to regain control of normal cell death in the affected tissue. The search for these clues to life and death has primarily centered on identification of proteins that either promote or prevent activation of the cell suicide pathway in a variety of tissues (Table 1) [reviewed in 2,3,14,15]. The number of genes which encode proteins involved in the regulation of apoptosis has grown at an almost exponential rate, albeit much less is known of their mechanisms of action. For the sake of clarity and brevity in this review, a sampling of the genes identified thus far will be grouped into four classes for subsequent discussion: 1) proteins encoded by members of the bcl-2 gene family; 2) oxidative stress response factors; 3) transcriptional regulators; and 4) cytoplasmic proteases including calpain and members of the interleukin-1ß-converting enzyme (ICE) gene family.

Table 1. Reported Regulators of Physiological Cell Death1
NameRole in Cell DeathMechanism of Action?2Reference
BCL-2InhibitionCellular Redox State31,32,33
Calcium Homeostasis34
Ras Interaction35
BAX Interaction14,18,28,29
BCL-XLInhibitionBAX Interaction19,27,28
BCL-XSActivationBCL-2 Interaction19,27,28
MCL-1InhibitionBAX Interaction20,28
BADActivationBCL-2/BCL-XL Interaction22
BAG-1InhibitionBCL-2 Interaction21
BAKActivationBCL-2/BCL-XL Interaction24,25,26
CED-9InhibitionUnknown (BCL-2 homolog)17
SODInhibitionAnti-Oxidant Enzyme81
GSHPxInhibitionAnti-Oxidant Enzyme31,81
CatalaseInhibitionAnti-Oxidant Enzyme81
p53Activationbcl-2/bax Gene Transcription40,43,44,45
c-mycActivationGene Transcription41,42
ICH-1SInhibitionICH-1L Antagonism49
CED-3ActivationUnknown (ICE homolog)47
1This is only a partial listing of proteins which have been implicated in activating or inhibiting the cell suicide pathway.
2The precise mechanisms of action of the majority of cell death regulators indicated are unknown. The examples provided are derived from studies currently available in the literature, and probably do not represent the full spectrum of biological functions.
3The protein referred to as ICErelII [ref. 53] has also been termed ICH-2 [ref. 51] and TX [ref. 52].

2.2. The bcl-2 Gene Family.

Proteins encoded by bcl-2 and related genes are probablythe most well-studied cell death regulatory factors [reviewed in 14,16]. The bcl-2 (B-cell leukemia/lymphoma-2) proto-oncogene was the first negative regulator of apoptosis reported. This gene was originally identified through analysis of the human t(14;18) (q32;q21) chromosomal translocation which predisposes affected individuals to the development of B-cell lymphomas. This translocation juxtaposes the bcl-2 gene with the immunoglobulin G heavy chain locus, leading to deregulated over-expression of bcl-2 [14]. Following reports of this initial correlation between high levels of bcl-2 expression and tumorigenesis, numerous studies have confirmed the death repressor role of the BCL-2 protein in a variety of cell types and under a spectrum of conditions known to trigger cell death [14,16]. These data are more intriguing when taken with the fact that BCL-2 was found to prevent apoptosis without altering proliferation rates, a function not attributed to any other protein at the time. Of additional note, a homolog of bcl-2 (termed ced-9) has been characterized in the nematode, Caenorhabditis elegans. Through elegant gene mutation studies, the protein encoded by the ced-9 gene has been shown to play a fundamental role in cell death inhibition in the worm during development, analogous to the death repressor activity of BCL-2 in vertebrates [17].

The proposed function of BCL-2 as a regulator of apoptosis also served as the catalyst for subsequent investigations by increasing numbers of laboratories to characterize other proteins involved in the cell death cascade. Two of these initial efforts led to simultaneous reports of the first additional members of the bcl-2 gene family, namely the bax (bcl-2-associated-x gene; [18]) and bcl-x (bcl-2-related gene-x; [19]) genes. The BAX protein was isolated via its ability to bind to, and thus co-immunoprecipitate with, BCL-2. Analysis of the function of BAX through gene transfer experiments revealed its actions to be that of a death susceptibility factor, initially believed to act by countering the death repressor activity of BCL-2 [18]. However, it is now thought that BAX directly leads to cell death following the formation of BAX:BAX homodimers, a process that can be disrupted by the presence of BCL-2 [14]. At the same time that bax was reported, another bcl-2 homolog termed bcl-x was identified in chicken lymphoid cells by low stringency hybridization cDNA cloning. Interestingly, the human cDNA was also cloned in this study; however, unlike its avian counterpart, the human bcl-x gene appears to undergo alternative splicing to yield two messenger RNA variants: a long isoform (encoding BCL-XL) which functions like BCL-2 to suppress apoptosis, and a truncated or short isoform (encoding BCL-XS) which can mimic the actions of BAX by antagonizing BCL-2-promoted cellular survival [19]. Subsequent to these findings, several additional members of the bcl-2 gene family have been isolated [20-26], and the proteins encoded by these genes have been shown to interact with each other to direct cell fate [24-28]. Additionally, precise functional domains present within the sequences of BCL-2 and related factors have been characterized and appear prerequisite for protein:protein interaction and the regulation of cell death [29].

The function of any member of the BCL-2 family remains an area of intensive investigation as very little data exist regarding the actions of these cell death regulatory proteins within the cell. The localization of BCL-2 within intracellular membranes, particularly those of the mitochondria, initially suggested that this factor may regulate the reduction-oxidation (redox) state of the cell [30]. Assuming this to be the case, two independent studies subsequently reported that over-expression of bcl-2 protects cells from death induced by reactive oxygen species [31,32]. On the other hand, a follow-up study suggested that BCL-2 may actually function as a pro-oxidant factor, thus conveying its protective effects indirectly through induction of the cell's normal oxidative stress response repertoire of defense enzymes [33]. In any case, it should be noted that the regulation of cellular redox is probably not the only mechanism by which BCL-2 acts, as additional studies have implicated this protein in calcium homeostasis [34] and growth factor-associated signaling events [35].

2.3. Reactive Oxygen Species.

The link of BCL-2 to reactive oxygen species provides an interesting association between this cell death regulator and the second category of genes to be discussed, namely members of the oxidative stress response gene family. Reactive oxygen species are generated in all cells as a consequence of normal metabolic function, or as a result of ligand-activated receptors tied to membrane phospholipid metabolism [reviewed in 36]. Members of this gene family include three forms of superoxide dismutase enzyme (secreted, mitochondrial and cytosolic) which convert superoxide anion radical to peroxide intermediates, as well as catalase and glutathione peroxidase which are responsible for metabolism of peroxides to water. Extensive cellular damage occurs in response to accumulation of reactive oxygen species or their intermediates, and recent data indicate that the end-result of this damage in most cases is the induction of apoptosis [reviewed in 37]. Additionally, disruptions in nuclear DNA integrity (e.g. formation of 8-hydroxydeoxyguanosine moieties) elicited by free radicals provide an important link to the next family of cell death genes to be discussed, the transcriptional regulators of cell death.

2.4 Transcriptional Regulators.

Transcriptional regulators represent an interesting family of proteins as these factors, unlike BCL-2, can modulate both mitosis and apoptosis. Two of these proteins in particular, p53 and c-myc, have been directly linked to the induction of apoptosis [38-42]. Of the target genes affected by nuclear accumulation of p53 or c-myc, members of the bcl-2 gene family have emerged as primary targets for transcriptional regulation. Using both in vivo and in vitro approaches, the p53 protein has been reported to bind sequence-specific enhancer regions in the bax gene promoter and repressor elements in the bcl-2 gene [43-45]. Thus, accumulation of p53 in the nucleus of a compromised cell is thought to trigger a disrupted balance between bax (increased) and bcl-2 (decreased) expression, a scenario known to predispose a cell to apoptosis activation [18]. A primary stimulus for stabilization and nuclear translocation of p53 is DNA damage [46], such as that resulting from attack by reactive oxygen species [36]. If this is the case, cellular damage induced by uncontrolled oxidative stress may lead to apoptosis as a consequence of p53-mediated alterations in expression of cell death-associated genes. The role of c-myc in regulation of the expression of bcl-2 and related genes is not as well characterized; however, the bax gene contains four consensus sequences for c-myc binding which suggests that this transcription factor may as well act, at least in part, via altered expression of these target cell death genes [45].

2.5. Cytoplasmic Proteases.

The fourth family of genes, composed primarily of cytoplasmic proteases, are among the most recently identified components of the cell death pathway. Similar to the ced-9/bcl-2 homology discussed earlier, our knowledge of the role of many of these proteases in cell death stemmed from initial studies of the protease homolog in C. elegans, CED-3 [47]. Genetic mutation of the ced-3 gene leads to a persistence of unwanted cells in the worm due to a loss of normal cell death during development. Sequence homology analysis of ced-3 with known vertebrate genes revealed a striking level of similarity with the vertebrate cysteine protease, interleukin-1ß-converting enzyme (ICE; [47]). Following the initial observation that CED-3 and ICE are likely homologs, numerous reports have described a rapidly growing family of ICE-related proteases that possess two important conserved features: a pentapeptide motif (QACRG) which likely serves as the catalytic domain, as well as cleavage specificity at aspartate residues [reviewed in 15]. To date, members of this family include ICE, ICE-and-ced-3-homolog-1 (Ich-1), cysteine-protease-P32 (CPP32), ICE-relative-II (ICErelII; also called Tx or Ich-2), ICErelIII, and Mch-2 [48-54]. Various approaches leading to enhanced or reduced activity of ICE-related proteases in cells have provided compelling evidence that members of this family of cell death regulators play an important, if not prerequisite, role in apoptosis in vertebrate species [15], analogous to that described for the ced-3 gene product in C. elegans [47].

Not all cytoplasmic proteases implicated in cell death, however, belong to the ICE gene family. In many cell types, calpain has also been proposed as a mediator of apoptosis [55]. This calcium-activated protease is ubiquitously expressed in most cells, and exists in two forms within the cell. These two forms are composed of a common large subunit derived from a single gene bound to isoform-specific small subunits encoded by separate genes. Depending upon the level of cytoplasmic calcium required to activate each form of the enzyme, the two forms of calpain can be easily segregated [reviewed in 56]. Interestingly, the regulation of calpain activity does not reside solely at the level of gene expression as most cells also express a specific inhibitor of calpain termed calpastatin [56]. Therefore, reminiscent of the BCL-2:BAX rheostat described earlier, the balance of calpain to calpastatin is likely an important determinant of whether active calpain is available for execution of the cell death command.

The targets of ICE-related proteases and calpain during apoptosis are diverse, but turn out to be very logical when evaluated in hindsight [reviewed in 15]. In addition to activating each other, enzymes encoded by members of the ICE gene family have been reported to cleave proteins involved in DNA repair [57,58] and messenger RNA processing [59], as well as structural components of the nuclear matrix [60]. Along these same lines, targets for calpain include cytoskeletal proteins (actin, fodrin) and gap junction proteins [56,61]. Thus, a scan of this representative list of proteins which are targets for protease attack during cell death reveals a common or underlying theme, namely that all of these proteins are required for the maintenance of normal cell structure, function and homeostasis.

[Table of Contents] [Next Section] [Previous Section]