Signal transduction pathways that regulate cell survival and cell death

Transcription

1 Oncogene (1998) 17, 3207 ± 3213 ã 1998 Stockton Press All rights reserved 0950 ± 9232/98 $ Signal transduction pathways that regulate cell survival and cell death Tomislav Dragovich 3, Charles M Rudin 3 and Craig B Thompson*,1,2,3 1 Gwen Knapp Center for Lupus and Immunology Research, The University of Chicago, Chicago, Illinois 60637, USA; 2 Howard Hughes Medical Institute, The University of Chicago, Chicago, Illinois 60637, USA and 3 Department of Medicine, The University of Chicago, Chicago, Illinois 60637, USA Apoptosis or programmed cell death (PCD) is a physiological process critical for organ development, tissue homeostasis and elimination of defective or potentially dangerous cells in complex organisms. Apoptosis permits cell death without a concomitant in ammatory response in the surrounding tissues. The process of apoptosis depends on the reception of multiple extracellular and intracellular signals, integration and ampli cation of these signals by second messengers and nally, activation of the death e ector proteases. Defects in control of apoptotic pathways may contribute to a variety of diseases including cancer, autoimmune and neurodegenerative conditions and AIDS. While many components of the regulatory network controlling apoptosis have been de ned, the mechanisms of action and patterns of interaction of these factors remain controversial. This article summarizes some of the known aspects of signaling pathways involved in apoptosis. Keywords: signal transduction; bcl-2 proteins; caspases; death receptors Events that activate the central pathway of cell death Programmed cell death or apoptosis can be initiated by a wide variety of stimuli, including developmental signals, cellular stress and disruption of cell cycle. In contrast, execution of apoptosis is remarkably uniform, involving characteristic morphologic and biochemical changes. The morphologic hallmarks of apoptosis include membrane blebbing, volume loss, nuclear condensation and DNA fragmentation (Kerr et al., 1972). A number of the key factors involved in the regulation, coordination and execution of these events have been identi ed. Genetic studies of the nematode Caenorhabditis elegans led to identi cation of three genes essential for control and execution of apoptosis in developing worms (Hengartner and Horvitz, 1994). Mammalian homologs of all three of these genes have been identi ed. The regulation and execution of apoptosis in C. elegans seems deceptively simple. Apoptotic cell death in the worm requires CED-3, a cousin of mammalian caspases and its activator CED-4 that is homologous to mammalian Apaf-1. Additional control of cell death/survival is provided by the anti-apoptotic factor CED-9 which is *Correspondence: CB Thompson a counterpart of mammalian Bcl-2 protein (Yuan et al., 1993). Recently, a novel C. elegans protein, Egl-1, has been found to bind to and inhibit the antiapoptotic function of CED-9 (Conradt and Horvitz, 1998). The molecular basis of apoptotic signaling appears to be more complex in mammalian cells. In mammalian cells, Bcl-2 is one of a large family of CED-9 related proteins that function as either positive or negative regulators of apoptosis. Mammalian counterparts of the nematode protein CED-3 comprise a large family of cysteine proteases, called caspases, which are critical in generation of apoptotic morphology. The multiple members of these gene families may allow for the independent regulation of apoptotic signaling pathways. Based on the established interactions among the known mediators of apoptosis three classic pathways of apoptotic signaling in mammalian cell have emerged (Figure 1). The rst one is initiated by the withdrawal of growth factors and is regulated by Bcl- 2 family of proteins. This pathway results in cytochrome c release from mitochondria, activation of Apaf-1 and triggering of a caspase cascade. The other well-established apoptosis pathway involves signaling by cell surface death receptors such as TNF or Fas, which through adapter molecules can recruit and activate caspases. The third and least well characterized pathway is initiated by DNA damage. Although in part regulated by proteins such as p53 and ATM, how this pathway results in caspase activation is not known. In all three of these pathways of cell death, caspases and Bcl-2 protein family play a key role in regulation and execution of apoptosis. Caspases Multiple mammalian homologs of CED-3 have been identi ed (1 ± 13). They all cleave peptide substrates after aspartic acid residues (Alnemri et al., 1996). Caspase activation initiates a proteolytic degradation that results in apoptotic morphology (Casiano et al., 1996). Speci c inhibitors of caspases prevent cell death, or at least apoptotic morphology, in response to many of the known signals that trigger apoptosis (Miura et al., 1993; Rabizadeh et al., 1993). Overexpression of most of the known caspases triggers apoptosis in cell lines, although only a subset of caspases appears to be involved in physiological apoptosis. Caspase-9 has been most clearly linked to central pathway of apoptotic induction. Apaf-1 in response to cytochrome c binding can form a complex with caspase-9 that initiates apoptotic events in cell free system (Li et al., 1997).

2 3208 Figure 1 Multiple apoptotic signals converge on caspase activation. Pathways discussed in this review include (1) those derived from cell surface receptor engagement, represented here by Fas receptor signaling, (2) those derived from growth factor withdrawal, represented here by the IL-3 receptor, and (3) those derived from nuclear DNA damage Caspase-9 is a critical regulator of at least some forms of apoptosis. Homozygous inactivation of the caspase- 9 gene resulted in embryonic lethality manifested by severe perturbations of the cerebral cortex and forebrain (Hakem et al., 1998; Kuida et al., 1998). Caspase-9 de cient cells were resistant to cell death induced by gamma-irradiation or cytotoxic drugs although the death response to other apoptotic stimuli, including Fas receptor engagement, was preserved. Caspase activation requires proteolytic processing from inactive proenzymes. The processing sites for procaspases are themselves caspase consensus sequences. Thus caspases may cleave and activate their own as well as other caspase pro-enzymes. This may serve as an important mechanism for amplifying caspasedependent apoptotic pathways. For example, active caspase-9 cleaves procaspase-3; active caspase-3 in turn appears to be required for many of the characteristic apoptotic nuclear changes. Downstream caspases cleave and inactivate proteins crucial for the maintenance of cellular cytoskeleton, DNA repair, signal transduction and cell cycle control (Fraser and Evan, 1996; Nagata 1997). Recent studies have uncovered a possible link between cytosolic caspase activation and apoptotic nuclear changes (Enari et al., 1998; Sakahira et al., 1998). A caspase-dependent deoxyribonuclease, CAD, is inactive and localized in the cytosol when bound to the inhibitor ICAD. Caspase-3 can cleave and inactivate ICAD thus releasing CAD that then translocates to the nucleus and initiates chromosomal degradation. While caspase activation typically results in apoptotic morphology, not all programmed cell death pathways are caspase-dependent. Cells which sustain traumatic damage or are deprived of growth-factors can die even in the presence of caspase inhibitors (Rathmell and Thompson, 1998). However, in the presence of such inhibitors the morphology of dying cell lacks many morphologic features of apoptosis, including typical nuclear changes (Hirsch et al., 1997). Outside of such experimental systems, it appears that caspase activation and cellular autodigestion may invariably accompany apoptosis in vivo. The Bcl-2 family A mammalian homolog of the C. elegans ced-9 gene, bcl-2, was identi ed as one of the genes at chromosomal breakpoint of t(14 : 18) in B-cell lymphomas. Overexpression of Bcl-2 protects cells from a variety of apoptotic stimuli, such as growth-factor withdrawal, exposure to chemotherapy agents or

3 toxins, viral infection and inappropriate oncogene expression. A number of Bcl-2 related proteins have been subsequently identi ed (Reed, 1997). Although they all share domains of structural homology, some Bcl-2 family proteins promote apoptosis while the others have anti-apoptotic function. The relative levels of pro- and anti-apoptotic Bcl-2 family proteins have been suggested to function as a `rheostat' regulating the apoptotic threshold of the cell (Yang and Korsmeyer, 1996). The antagonistic function of some of the members of the family is at least in part explained by their ability to form heterodimers. The anti-apoptotic proteins Bcl-x L and Bcl-2 form heterodimers with pro-apoptotic Bax (Oltvai and Korsmeyer, 1994). An excess of Bax promotes cell death, but coexpression of Bcl-x L or Bcl-2 can neutralize this e ect. An alpha-helical domain known as BH3 and shared by multiple members of the Bcl-2 family is critical for the formation of homo- and heterodimers (Kelekar and Thompson, 1998). Some of pro-apoptotic Bcl-2 family members, such as Bad and Bid, only share homology within the BH3 domains (so-called BH3 proteins). Most Bcl-2 family proteins have a conserved C- terminal hydrophobic domain, which is necessary for insertion into the outer mitochondrial, endoplasmic reticular and outer nuclear membranes. Based on this unusual subcellular localization and homology to bacterial pore forming molecules, such as the poreforming subunit of diphteria toxin, it is suspected that some members of Bcl-2 protein family function as membrane pores or channels (Muchmore et al., 1996). Indeed, both Bcl-x L and Bcl-2 as well as the proapoptotic Bax were found to have pore-forming capacity in vitro (Minn et al., 1997; Schendel et al., 1997; Antonsson et al., 1997). Recent studies support the central role of Bcl-2 family proteins in the regulation of mitochondrial membrane integrity and cytochrome c release (Yang et al., 1995; Vander Heiden et al., 1997). Overexpression of either Bcl-2 or Bcl-x L inhibits cytochrome c release from mitochondria. In contrast, when transfected in mammalian or yeast cells, Bax overexpression results in dissipation of mitochondrial membrane potential and release of cytochrome c (Rosse et al., 1998; Manon et al., 1997). The molecular mechanism of this regulation has not been determined. Some investigators suggest that Bcl-2 like proteins regulate the opening of what has been referred to as mitochondrial permeability transition pore (PTP) or megachannel. The hypothetical PTP would consist of the adenine nucleotide translocator in the inner mitochondrial membrane, voltage-dependent anion channel in the outer membrane as well as other components (Marzo et al., 1998). Physical association between Bcl-2 family proteins and components of PTP remains speculative although both Bcl-x L and Bcl-2 localize to contact points between the inner and outer mitochondrial membrane and Bax was recently co-puri ed with other components of the pore (Marzo et al., 1998). In addition, some of the antiapoptotic Bcl-2 proteins may function by complexing with and inactivating factors that trigger caspase activation. Both cytochrome c and Apaf-1 have been de ned as co-factors that are critical for the activation of caspase-9 and the subsequent apoptotic cascade. As mentioned before, in C. elegans CED-9 directly interacts with CED-4 preventing it from activating caspase CED-3. In mammalian cells, Bcl-x L has been reported to interact with the mammalian CED-4 homolog Apaf-1 and by that mechanism might prevent Apaf-1 from activating downstream caspases (Hu et al., 1998). These several lines of evidence suggest that Bcl-2 family members function at a critical decision point immediate upstream of an irreversible commitment to cell death. The Bcl-2 family proteins, Apaf-1, and caspases together comprise mammalian equivalent of the C. elegans `apoptosome' consisting of Ced-9, Ced-4 and Ced-3. Loss of mitochondrial transmembrane potential and activation of caspases represent molecular hallmarks of mammalian central apoptotic pathway. Bcl-2 family proteins appear to be critical in regulation of both events. Nevertheless, signi cant gaps in this model remain. For example, although the model integrates the three critical apoptotic regulators de ned in C. elegans, it fails to explain how extramitochondrial Bcl-2 inhibits apoptosis and why it can do so even after cytochrome c is released (Rosse et al., 1998; Li et al., 1997). It does not explain the nding that caspase inhibitors do not inhibit Bax toxicity in mammalian cells suggesting that Bax may induce apoptotic changes independently of the mammalian `apoptosome' (Xiang et al., 1996). Further analysis of the molecular basis of programmed cell death execution will be required to answer these and other important questions. Receptors and second messengers that regulate cell death and survival Homeostasis in mammalian cells is dependent on the continuous integration of survival and death signals from the extracellular environment (Ra, 1992). Extracellular signals that activate apoptotic or survival pathways include peptide growth factors, cytokines and interleukins. Signaling pathways can be triggered by both soluble factors and by direct cell to cell contact. One of the most intensively studied families of receptors involved in apoptotic signaling is that of tumor necrosis factor receptor (TNFR) family (Bazzoni and Beutler, 1996). This is a family of multimeric receptors that upon activation by their respective ligands may initiate either cell death or survival signals. Extracellular domains of the TNFR family members are related, while intracellular domains di er widely both in sequence and in their a nity for various adapter molecules and second messengers. The subfamily of receptors most closely linked to induction of cell death (e.g. Fas, TNFR1 and DR3) do contain a conserved intracellular sequence known as a `death domain'. Through these domains, receptors associate with a number of intracellular signal transduction molecules that themselves contain the `death domain'. For example, the intracellular domain of activated Fas receptor interacts with that of a linker molecule called FADD (Fas-associated death domain-containing protein) (Chinnaiyan et al., 1995). Subsequent transduction of the apoptotic signal is dependent on another domain of FADD, the `death e ector domain', which interacts with additional downstream molecules. One of them is caspase-8, which after activation by FADD can initiate a cascade of caspase activation leading to 3209

4 3210 degradation of intracellular substrates and cell death (Muzio et al., 1996). TNFR1 similarly induces cell death through interaction with a related but di erent adapter molecule called TRADD (TNFR-associated death domain protein) (Hsu et al., 1996). While activation of Fas, TNFR1 and DR3 receptors induces apoptosis by associating with cytoplasmic factors such as FADD, TRADD and RIP, other receptors such as CD40, TNFR2 and CD30 promote cell survival by interacting with a distinct group of second messengers, including the TNF receptorassociated factors or TRAFs (Arch et al., 1998). To date six mammalian TRAF proteins have been identi ed. Some members of the TRAF family negatively regulate apoptosis through induction of survival pathways. For example, recruitment of TRAF2 protein to TNF receptors has been reported to promote survival and activates JNK (c-jun N- terminal kinase) and NF-kB (nuclear factor-kb) (Yeh et al., 1997). A number of recent studies have linked NF-kB activation to the induction of cell survival pathway (Liu et al., 1996; Beg and Baltimore, 1996). Additional regulatory proteins further modify signaling pathways downstream from TRAF recruitment. These include adapter proteins like A20, cellular inhibitor of apoptosis (c-iap), TRAF-interacting protein (TRIP) and TRAF family member associated NF-kB activator (I-TRAF/TANK) (Cheng and Baltimore, 1996; Lee and Choi, 1997; Rothe et al., 1996; Roy et al., 1997). TRIP inhibits NF-kB activation by TRAFs while I-TRAF/TANK augments it depending on the speci c cellular context. c-iap1 and c-iap2 interact with TRAFs and have been reported to interfere with the activation of caspases even downstream of cytochrome c release (Roy et al., 1997). FLICE inhibitory protein or FLIP, which contains two death e ector domains and an inactive caspase domain, binds to FADD and caspase-8 and inhibits death receptor-induced apoptosis (Hu et al., 1997; Thome et al., 1997). This web of interacting proteins leads to very di erent responses in di erent cells and at di erent times. Interaction of CD30 and CD40 receptors with TRAF proteins, depending on cellular context can promote either cell death or cell survival (Arch et al., 1998). It is interesting that signaling through CD30 and CD40 can promote survival of naive lymphocytes, but subsequently increases their sensitivity to death inducing signals from Fas or TNFR1 (Rathmell et al., 1996; Duckett and Thompson, 1997). This may serve as an internal control mechanism to prevent unlimited proliferation of cells that might otherwise result in development of autoimmune disease or neoplasia. Another extracellular signal for induction of apoptosis is deprivation of cells of their anchorage to extracellular matrix or from cell to cell contacts. The induction of apoptotic pathway through disruption of cell anchorage to the substratum is termed anoikis (Aplin et al., 1998). The exact signaling pathway responsible for the induction of apoptosis after cell anchorage deprivation is unknown. One of the two pathways proposed includes tyrosine kinase FAK, PI-3 kinase and Akt. The other, parallel pathway may involve the stress induced kinases including MEKK-1 and JNK (Frisch and Rouslahti, 1997). Overexpression of Bcl-2 in epithelial cells can protect the cells from anoikis. The physiological role of anoikis is unclear although disregulation of this process may be important for the ability of malignant cells to invade and metastasize elsewhere. Nuclear factors ivolved in apoptosis In addition to extracellular signals, cells can undergo apoptosis in response to internal signals including genetic abnormality. Defective or inappropriate cell cycle progression caused by a variety of genotoxic injuries can result in cell cycle arrest, and subsequent induction of apoptosis. The tumor suppressor gene p53 has been clearly linked to these pathways leading to its designation as the `guardian of the genome'. DNA strand breaks induce rapid p53 upregulation but the exact mechanism of this triggering event remains unknown. The upregulation of p53 is mostly posttranscriptional, involving both, an increase in translation and prolonged half-life. p53 is a sequence-speci c DNA-binding protein, which activates transcription of genes including Gadd45, mdm2, p21 and bax (Cox and Lane, 1995; Wyllie, 1997). Increase in bax transcription may be in part responsible for p53-induced apoptosis following DNA damage (Miyashita and Reed, 1995). However, p53 can also activate apoptosis by transcription-independent mechanisms (Caelles et al., 1994). What is more, there are p53-independent pathways that result in cellular apoptosis following DNA damage. T cell lymphoma lines derived from p53 knockout mice died rapidly after exposure to irradiation or to drugs that promote DNA strand breaks (Strasser et al., 1994). Death occurred by apoptosis and was inhibitable by Bcl-2 suggesting that Bcl-2 acts downstream of p53 signal and pointing out that there are additional, parallel and p53-independent pathways regulating DNA damage induced apoptosis in mammalian cells. Another gene implicated in DNA damage induced apoptosis is ATM (Morgan and Kastan, 1997). This gene is defective in patients with ataxia-telangiectasia disorder, who are prone to lymphoid malignancies, cerebellar degeneration and are extremely sensitive to ionizing radiation. Cells de cient in ATM do not upregulate p53 following irradiation damage, suggesting that ATM plays a role in the induction of p53. The downstream p53-dependent apoptotic pathway, however appears to be intact in ATM-de cient cells. Interactions between the molecules that sense DNA damage and molecules involved in cytosolic activation of caspases have not been established. Such connections must exist as cytosol-nuclear translocation of the nuclease CAD requires cytosolic, caspase-dependent inactivation of ICAD. How `inside-out' or nuclearcytoplasmic signal transduction is achieved remains to be elucidated. Cross-talk between the pathways Di erent pathways involved in regulation of cell proliferation and cell death may have di erent signi cance in various cells or within the same cell at di erent stages of development. We now have some evidence of the cross-talk between the components of

5 di erent signaling pathways within the cell (Figure 2). Bid, a pro-apoptotic Bcl-2 family protein, is cleaved by caspase-8 in response to Fas signaling (Li et al., 1998; Luo et al., 1998). Following cleavage, the C-terminal fragment of Bid binds to mitochondria inducing the release of cytochrome c, which in turn activates Apaf-1 and initiates procaspase-9 processing. Thus, this crosstalk connects two of the classic apoptotic pathways depicted in Figure 1; that of Fas receptor mediated activation of caspase-8 with that of IL-3 withdrawalinduced mitochondrial release of cytochrome c and activation of caspase-9. Another example of the cross-talk between di erent signaling pathways that regulate cell proliferation and cell death involves Bad, a pro-apoptotic Bcl-2 family protein. Bad executes its pro-apoptotic function by binding to anti-apoptotic proteins Bcl-2 and Bcl-x L. Following growth factor receptor engagement, the growth factor dependent serine/threonine kinase Akt phosphorylates Bad (Figure 2). The phosphorylated Bad is sequestered by cytosolic protein thus increasing the levels of free Bcl-2 and Bcl-x L (Zha et al., 1996). In this example, a growth factor initiated, survival-promoting signal interferes with a Bcl-2/Bclx L -dependent regulation of the cellular apoptotic threshold (Gajewski and Thompson, 1996). It is expected that more such examples of cross-talk will be uncovered as we continue to untangle the web of apoptotic signaling pathways. Conclusion Our understanding of the complex processes that determine cell fate has advanced over the past decade yet many unanswered questions remain. The complexity and redundancy involved in regulation of apoptosis re ects its central importance for the growth and development of complex organisms. The signal transduction of apoptotic stimuli resembles a complex information network which includes systems for reception, processing of information, ampli cation, feedback, and response integration. The apoptotic threshold varies widely with the cell type and its stage of development and di erentiation. As the cells exist in a dynamic equilibrium, their threshold for apoptosis may be constantly changing. Alterations in external milieu are sensed through the cell surface receptors as well as multiple cytoplasmic and nuclear factors. These multiple signals are integrated through a complex 3211 Figure 2 Cross-talk between positive and negative apoptotic signaling pathways. Positive and negative apoptotic signals can be integrated at the level of the mitochondrial outer membrane by interactions involving the Bcl-2 family of apoptotic regulators

6 3212 network of positive and negative regulators of apoptosis that ultimately determine cell fate. Bcl-2 family members, both by their ability to interact with other mediators of apoptosis and their subcellular localization, appear to function at the point of convergence of multiple apoptotic signals. Disturbances in cellular metabolism, whether imposed by starvation, infection, toxins or by DNA damage ultimately result in mitochondrial membrane release of cytochrome c and activation of downstream caspases. Based on recent studies, mitochondrial membrane disruption appears to be a `point of no return' for the cell, an ultimate commitment to cell death. Bcl-2 family members and caspases appear to be critical factors involved in regulating the outcome of these events. Recent advances in apoptosis research have identi ed new molecular players that regulate this fundamental and complex biological process. However, there are large gaps in our ability to understand the intricate network of apoptotic signal transduction. While there is considerable information concerning apoptotic signaling pathways that involve cell surface receptors that can induce caspase activation, much less is known about Bcl-2 family proteins and their promiscuous interactions among themselves, with caspases and with components of the mitochondria. A deeper understanding of the physiological importance of numerous molecules involved in apoptotic signaling pathways requires further investigation. In addition, we need to learn more about how DNA damage initiates and regulates the execution of programmed cell death. There is every reason to believe that new and exciting advances in apoptosis research will continue to come at a rapid pace, constantly changing the eld and making reviews like this outdated by the time they reach the press. References Alnemri ES, Livingston DJ, Nicholson DW, Salvesen G, Thornberry NA, Wong WW and Juan J. (1996). Cell, 87, 171. Antonsson B, Conti F, Ciavatta Am, Montessuit S, Lewis S, Martinou I, Bernasconi L, Bernard A, Mermod JJ, Mazzei G, Maundrell K, Gambale F, Sadoul R and Martinou JC. (1997). Science, 277, 370 ± 372. Aplin AE, Howe A, Alahari SK and Juliano RL. (1998). Pharmacol. Rev., 50, 199 ± 251. Arch RH, Gedrich RW and Thompson CB. (1998). Genes & Dev., 12, 2821 ± Bazzoni F and Beutler B. (1996). N. Engl. J. Med., 334, 1717 ± Beg AA and Baltimore D. (1996). Science, 274, 782 ± 784. Caelles C, Helmberg A and Karin M. (1994). Nature, 370, 220 ± 223. Casiano CA, Martin SJ, Green DR and Tan EM. (1996). J. Exp. Med., 184, 765 ± 770. Cheng G and Baltimore D. (1996). Genes Dev., 10, 963 ± 973. Chinnaiyan AM, O'Rourke K, Tewari M and Dixit VM. (1995). Cell, 81, 505 ± 512. Conradt B and Horvitz HR. (1998). Cell, 93, 519 ± 529. Cox LS and Lane DP. (1995). Bioessays, 17, 501 ± 508. Duckett CS and Thompson CB. (1998). Genes Dev., 11, 2810 ± Enari M, Sakahira H, Yokoyama H, Okawa K, Iwamatsu A and Nagata S. (1998). Nature, 391, 43 ± 50. Fraser A and Evan G. (1996). Cell, 85, 781 ± 784. Frisch SM and Ruoslahti E. (1997). Curr. Opin. Cell Biol., 9, 701 ± 706. Gajewski TF and Thompson CB. (1996). Cell, 87, 589 ± 592. Hakem R, Hakem A, Duncan GS, Henderson JT, Woo M, Soengas MS, Elia A, de-la-pompa JL, Kagi D, Khoo W, Potter J, Yoshida R, Kaufman SA, Lowe SW, Penninger JM and Mak TW. (1998). Cell, 94, 339 ± 352. Hengartner MO and Horvitz HR. (1994). Cell, 76, 665 ± 676. Hirsch T, Marchetti P, Susin SA, Dallaporta B, Zamzami N, Marzo I, Geuskens M and Kroemer G. (1997). Oncogene, 15, 1573 ± Hsu H, Shu H-B, Pan MB and Goeddel DV. (1996). Cell, 84, 299 ± 308. Hu S, Vincenz C, Buller M and Dixit VM. (1997). J. Biol. Chem., 272, Hu Y, Benedict MA, Wu D, Inohara N and Nunez G. (1998). Proc. Natl. Acad. Sci. USA, 95, 4386 ± Kastan MB. (1997). N. Engl. J. Med., 333, 662 ± 663. Kelekar A and Thompson CB. (1998). Trends. Cell Biol., 8, 324 ± 330. Kerr JFR, Wyllie AH and Currie AR. (1972). J. Cancer, 26, 239 ± 257. Kuida K, Haydar TF, Kuan CY, Gu Y, Taya C, Karasuyama H, Su MS, Rakic P and Flavell RA. (1998). Cell, 94, 325 ± 337. Lee SY and Choi Y. (1997). J. Exp. Med., 185, 1275 ± Li H, Zhu H, Xu Cj and Yuan J. (1998). Cell, 94, 491 ± 501. Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Alnemri ES and Wang X. (1997). Cell, 91, 479 ± 489. Liu Z, Hsu H, Goeddel DV and Karin M. (1996). Cell, 87, 555 ± 576. Luo X, Budihardjo I, Zou H, Claughter C and Wang X. (1998). Cell, 94, 481 ± 490. Manon S, Chaudhuri B and Guerin M. (1997). FEBS Lett, 415, 29 ± 32. Marzo I, Brenner C, Zamzami N, Susin SA, Beutner G, Brdiczka D, Remy R, Xie ZH, Reed JC and Kroemer G. (1998). J. Exp. Med., 187, 1261 ± Minn AJ, Velez P, Schendel SL, Liang H, Muchmore SW, Fesik SW, Fill M and Thompson CB. (1997). Nature, 385, 353 ± 357. Miura M, Zhu H, Rotello R, Hartweig EA and Yuan J. (1993). Cell, 75, 653 ± 660. MiyashitaT and Reed JC. (1997). Blood, 81, 151 ± 157. Morgan SE and Kastan MB. (1997). Adv. Cancer Res., 71, 1±25. Muchmore SW, Sattler M, Liang H, Meadows RP, Harlan JE, Yoon HS, Nettesheim D, Chang BS, Thompson CB, Wong SL, Ng SL and Fesik SW. (1996). Nature, 381, 335 ± 341. Muzio M, Chinnaiyan AM, Kischkel FC, O'Rourke K, Shevchenko A, Ni J, Sca di C, Bretz JD, Zhang M, Gentz R, Mann M, Krammer PH, Peter ME and Dixit VM. (1996). Cell, 85, 817 ± 827. Nagata S. (1997). Cell, 88, 355 ± 365. Oltvai ZN and Korsmeyer SJ. (1994). Cell, 79, 189 ± 192. Ra MC. (1992). Nature, 356, 397 ± 400. Rabizedeh S, LaCount DJ, Friesen PD and Bredsen DE. (1993). J. Neurochem., 61, 2318 ± Rathmel JC and Thompson CB. (1998). Ann. Rev. Immunol., in press.. Rathmell JC, Townsed SE, Xu JC, Flavell RA and Godnow CC. (1996). Cell, 87, 319 ± 329. Reed JC. (1997). Nature, 387, 773 ± 776. Rosse T, Olivier R, Monney L, Rager M, Conus S, Fellay I, Jansen B and Borner C. (1998). Nature, 391, 496 ± 499.

7 Rothe M, Xiong J, Shu B, Williamson K, Goddard A and Goeddel DV. (1996). Proc. Natl. Acad. Sci. USA, 93, 8241 ± Roy N, Deveraux QL, Takanishi R, Salvesen GS and Reed JC. (1997). EMBO J, 16, 6914 ± Sakahira H, Enari M and Nagata S. (1998). Nature, 391, 96 ± 99. Schendel SL, Xie Z, Montal MO, Matsuyama S, Montal M and Reed JC. (1997). Proc. Natl. Acad. Sci. USA, 94, 5113 ± Strasser A, Harris AW, Jacks T and Cory S. (1994). Cell, 79, 329 ± 339. Thome M, Schneider P, Ho man K, Fickensher H, Meinl E, Neipel F, Mattmann C, Burns K, Bodmer JL, Schroter M, Sca di C, Krammer PH, Peter ME and Tschopp J. (1997). Nature, 386, 517. Vander Heiden MG, Chandel NS, Williamson EK, Schumaker PT and Thompson CB. (1997). Cell, 91, 627 ± 637. Wyllie A. (1997). Nature, 389, 237 ± 238. Xiang J, Chao D and Korsmeyer SJ. (1996). Proc. Natl. Acad.Sci.USA,93, ± Yang E and Korsmeyer SJ. (1996). Blood, 88, 386 ± 401. Yang E, Zha J, Jockel J, Boise LH, Thompson CB and Korsmeyer SJ. (1995). 80, 285 ± 291. Yeh WC, Shahinian A, Speiser D, Kraunus J, Billia F, Wakeham A, de-la-pompa JL, Ferrick D, Hum B, Iscove N, Ohashi P, Rothe M, Goeddel DV and Mak TW. (1997). Immunity, 7, 715 ± 725. Yuan J, Shaham S, Ledoux S, Ellis HM and Horvitz HR. (1993). Cell, 75, 641 ± 652. Zha JP, Harada H, Yang E, Jockel J and Korsmeyer SJ. (1996). Cell, 87, 619 ±