Apoptosis researchers gathered on the Island of Hawaii for an AACR Special Conference devoted to apoptosis and cancer. The meeting was organized by John C. Reed of the Burnham Institute and Scott Lowe of the CSHL3 and was designed to explore the latest developments in understanding of apoptosis mechanisms of relevance to tumor biology.

Dysregulation of apoptosis occurs commonly in cancers and has been implicated in many events relevant to the pathogenesis and progression of tumors, including: (a) cell accumulation caused by failure of normal cell turnover mechanisms (programmed cell death); (b) creating a permissive environment for genetic instability and oncogene activation, events that might otherwise kill cells; (c) promoting resistance to immune cell attack; (d) contributing to resistance to the cytotoxic effects of chemotherapy and radiation; (e) allowing for tumor cell survival in the face of hypoxia, a topic of relevance to antiangiogenesis-based therapies for cancer; and (f) fostering tumor metastasis by allowing cells to survive in a detached (suspended) state. Although much information has been gained about the molecules underlying the phenomenon of apoptosis in the past decade, new discoveries continue to teach us that we have much yet to learn, particularly where mechanisms of tumor resistance to apoptosis are concerned and with respect to new strategies for restoring apoptosis sensitivity.

The opening evening of presentations focused on apoptosis from a genomics perspective. Lectures were given by Drs. John Reed (Burnham Institute), Michael Hengartner (University of Zurich), and John Abrahms (University of Texas, Southwestern) describing mechanisms of apoptosis in Homo sapiens, Caenorhabditis elegans, and Drosophila melanogaster, respectively. Dr. Reed organized the human genes relevant to apoptosis according to their domain families, introducing the caspases, CARD, death domain, death effector domain, Bcl-2, and IAP family proteins. Over 100 genes belonging to these protein domain families have been recognized in the human genome, suggesting that a minimum of 0.25% of all recognized human genes is devoted to regulation of programmed cell death. The diversity of human genes implicated in apoptosis regulation is reflective of the complex biology of the multicellular mammalian organisms and suggests that cell type and tissue-specific mechanisms have evolved for achieving exquisitely precise control over cell life and death decisions. This abundant diversity of apoptosis-relevant genes also suggests opportunities for carefully fine tuning therapeutic strategies designed to either enhance or inhibit apoptosis for amelioration of disease. Dr. Reed illustrated a few examples of how improved understanding of apoptosis mechanisms has begun to create opportunities for drug discoveries and new diagnostics for cancer, including: (a) small molecule antagonists of antiapoptotic Bcl-2 family proteins that mimic endogenous inhibitors of these proteins (BH3 peptides); (b) inhibitors of IAP family proteins based on mimics of SMAC and other endogenous antagonists of these antiapoptotic proteins; and (c) new antiapoptotic genes whose expression correlates with risk of relapse after local therapy (surgery or radiotherapy) for early stage solid tumors.

Dr. Hengartner’s presentation focused on his recent work studying the response of germ cells in C. elegans to X-irradiation. Although the somatic cells of the nematode are strikingly resistant to X-irradiation, the germ cells undergo apoptosis in response to genotoxic injury, providing a genetically retraceable model system for understanding mechanisms that link DNA damage to apoptotic responses. In response to X-irradiation, germ cells of C. elegans up-regulate expression of EGL1, a BH3 only proapoptotic protein and antagonist of the Bcl-2 homologue CED9. A screen for worms in which germ cells failed to undergo apoptosis after X-irradiation revealed a requirement for the HUS1 gene. The HUS1 protein accumulates in nuclei after DNA damage, localizing to chromosomes and participating in a multiprotein complex that includes RAD1, RAD9, and possibly RAD17. HUS1-deficient animals failed to up-regulate EGL1 mRNA in their germ cells after X-irradiation. Ablation of EGL1 reduces but does not abolish X-irradiation-induced apoptosis in germ cells of C. elegans, however, suggesting that EGL1 represents only one of the genes relevant to apoptosis induction under these circumstances. Interestingly, the worm homologue of p53 is relevant to this pathway and is required for X-irradiation-induced apoptosis but not cell cycle arrest in germ cells of C. elegans adults. In addition, ablation of p53 in C. elegans does not prevent induction of EGL1 mRNA in germ cells in response to X-irradiation. Thus, the targets of p53 which link DNA damage to apoptosis induction remain to be elucidated.

John Abrahms (University of Texas, Southwestern) reviewed the central role of IAP family proteins in the regulation of cell death in the fruit fly D. melanogaster. IAPs represent a family of antiapoptotic proteins that directly bind and inhibit certain caspase family cell death proteases. The IAPs, in turn, are antagonized by several proapoptotic IAP-binding proteins, including RPR, Grim, and Hid in the fly. The genes encoding these IAP antagonists are all located near each other on the same chromosome. The human genome contains functionally homologous genes, but humans and flies use different strategies for achieving regulation of IAPs by their respective antagonists. In Drosophila, expression of IAP antagonists such as RPR, Grim, and Hid is inducible, placing the bulk of regulation at a transcriptional level. In contrast, regulation of IAP antagonists in humans occurs at a post-transcriptional level, where these proteins are initially sequestered in mitochondria, undergoing release into the cytosolic in response to specific types of apoptotic stimuli. However, Dr. Abrahms also presented evidence suggesting a new facet to cell death induction by IAP antagonists. At least some of the IAP antagonists have an additional function which allows them to target mitochondria, promoting cell death through an undefined mechanism. It could be that Drosophila IAP antagonists such as RPR interact with Bcl-2 family proteins on the surface of mitochondria, promoting changes in mitochondrial membrane permeability which unmask caspase activators, thereby triggering mitochondria-dependent pathways for cell death. Additional work on this interesting facet of IAP antagonist function will likely provide new insights into the similarities and differences in apoptosis regulation in flies versus humans.

Among the transcription factors which induce expression of IAP antagonist RPR is p53. During screens for p53-inducible genes, Dr. Abrahms discovered a new IAP antagonist called SKL. This new protein has been recently identified through alternative strategies by several groups and is encoded by a gene adjacent to the chromosomal interval that encodes RPR, Hid, and Grim. The SKL protein contains a conserved N-terminal motif which is responsible for binding IAPs. The SKL gene contains a p53-binding site motif similar to that found in the RPR gene. SKL is among 12 additional genes that are induced by X-irradiation in a p53-dependent manner. Analysis of these other p53-inducible genes is likely to provide new insights into mechanisms of genotoxic stress responses.

A session devoted to inhibition of death included three talks on Bcl-2 family proteins by Drs. Susan Corey (Walter & Elisa Hall Institute), Yoshihide Tsujimoto (Osaka University), and Craig Thompson (University of Pennsylvania), as well as a presentation by Dr. Douglas Green (La Jolla Institute for Allergy & Immunology) on mechanisms of mitochondria-dependent cell death. Dr. Corey’s presentation focused on the BH3-only proapoptotic protein Bim, an antagonist of Bcl-2. She presented evidence from Bim knockout mice suggesting that Bim is a tumor suppressor in vivo. Mating Myc/IgH mice with Bim knockouts results in faster onset of lymphomas. Bim knockout mice also accumulate more T and B lymphocytes, as well as myeloid cells, and have increased levels of circulating immunoglobulins attributable to excessive numbers of plasma cells. Cells from these animals display increased resistance to certain apoptotic stimuli, including microtubule-binding anticancer drugs, which is of interest because Bim associates with microtubules and is released on microtubule disruption to interact with Bcl-2-family proteins on the surface of mitochondria. Of keen interest were the results of experiments in which Bcl-2 knockout mice were bred with Bim knockouts. The lack of Bim complements the lack of Bcl-2, producing nearly normal animals. In contrast to the phenotype of Bcl-2 knockout mice, which are runted, display shortened ears, develop polysystic kidney disease and premature graying of their hair, as well as lymphoid system collapse, the double knockout of Bcl-2 and Bim results in mice which are essentially normal. These data suggest that Bim and Bcl-2 counteract each other in a delicate balance that is important, at least, for T and B lymphocyte, myeloid cells, nephrogenic progenitors, and melanocytes in vivo.

Dr. Tsujimoto (Osaka University) focused most of his remarks on interactions of Bcl-2 and Bcl-XL with VDAC, which resides in the outer membrane of mitochondria. In an in vitro reconstitution system using unilammellar liposomes, Dr. Tsujimoto et al. have obtained evidence that Bcl-XL closes the VDAC channel. In contrast, recombinant Bax protein causes VDAC pore opening, generating channels of sufficient size to allow leakage of cytochrome c. The structure of this Bax/VDAC pore is presently undefined. The first α-helix of the Bcl-XL protein, called the BH4 domain, is responsible for interactions with VDAC, according to Dr. Tsujimoto. Using synthetic peptides that fuse the HIV tat membrane penetration sequence with BH4, his group has demonstrated suppression of apoptosis in vitro and in animal models. The tat-BH4 peptides, e.g., reduce apoptosis in the small intestine after X-irradiation of mice, suggesting possible therapeutic opportunities.

Dr. Thompson (University of Pennsylvania) addressed the connection between mitochondria, metabolism, and the mechanism of the Bcl-2 family genes. He reviewed evidence that on growth factor withdrawal from factor-dependent lymphoid cells, rates of glucose transport into cells decline, along with reductions in the rate of glycolysis, ATP levels, and oxygen consumption. Cells overexpressing Bcl-XL display marked resistance to growth factor deprivation as well as glucose deprivation. These cells suffer less severe reductions in intracellular ATP concentrations and can survive for days to weeks in the absence of glucose. Similarly, a double knockout of genes encoding proapoptotic Bax and Bak results in murine cells which can survive for weeks without nutrients in culture. Dr. Thompson argued that defects in the cell death pathway controlled by Bcl-2 family proteins can allow cancer cells to survive under suboptimal conditions. He proposed that an alternative mechanism for promoting tumor cell growth, thus, would be to simply increase nutrient uptake and proposed that dysregulation of the PTEN/AKT pathway represents one such opportunity for tumor cells. Dysregulation of AKT activity results in increased production of the GLUT-1 glucose transporter and marked increases in the rate of glycolosis. An intriguing idea is that this alteration in metabolic pathways found in some cancer cells might somehow be exploited for therapeutic purposes.

Dr. Douglas Green (La Jolla Institute of Allergy & Immunology) used the Drosophila Snyder cells (S2) in conjunction with siRNA to study the requirements for certain genes during apoptosis induced by UV radiation, etoposide, and other cell death stimuli. Using siRNA to ablate expression of cytochrome c in these cells, Dr. Green provided evidence that, at least in Drosophila cells, cytochrome c is not required for apoptosis. This finding suggests important differences in the apoptosis mechanisms of flies and humans. Dr. Green speculated that determinants exposed on the surface of mitochondria in Drosophila cells may substitute for cytochrome c, allowing for activation of the Apaf homologue Dark and triggering caspase activation.

Several presentations addressed mechanisms of cellular responses to chemotherapy and radiation, a topic of great interest to cancer researchers. Dr. Junying Yuan (Harvard), e.g., addressed the role of the LKB1 kinase as a mediator of p53-dependent cell death in the small intestine. Homozygous mutations inactivate the LKB1 gene in patients with Peutz-Jaeger syndrome. LKB1 associates with microtubules and may participate in a pathway that somehow senses microtubule disruption. Overexpression of active LKB1 kinase induces apoptosis in a p53-dependent manner. Moreover, p53 knockout cells are insensitive to LKB1-induced apoptosis, and gene transfection of p53 into these cells restores sensitivity. Dr. Yuan speculated that p53 induces expression of a gene which somehow modifies LKB1-dependent apoptosis pathways, e.g., p53 might induce expression of a LKB1 substrate. Additional presentations by Drs. Andre Gudkov (University of Illinois), Klaus Debatin (University of Ulm), Waifik El Deiry (University of Pennsylvania), Yuri Labeznik (CSHL), and Tak Mak (Toronto) further addressed mechanisms of p53-dependent pathways of relevance to DNA-damaging anticancer drugs. Dr. El Deiry demonstrated previously that p53 induces expression of the TNF family death receptor, DR5. Using DR5 knockout mice, he showed evidence of tissue-specific requirement for DR5 for apoptosis induced by X-irradiation, e.g., DR5 appears to be required for optimal apoptotic responses in the thymus but not the spleen after X-irradiation. Dr. El Deiry also presented evidence of yet another Bcl-2 family member whose expression is induced by p53. He showed that expression of Bid is up-regulated by 3–5-fold after induction of p53 in a model cell line. Consensus p53 binding sites are found in the Bid-encoding genes of humans and mice, and p53 can bind in vitro to these DNA sites. Continuing with the theme of tissue specificity and p53 target genes, Dr. El Deiry showed that although X-irradiation induces increases in Bid in the red pulp of spleen, the p53 target gene PUMA is induced in the white pulp. Both Bid and PUMA are BH3-only proapoptotic members of the Bcl-2 family. The functional importance of Bid for p53-induced apoptosis was suggested by experiments using murine embryonic fibroblasts cells that are deficient in Bid, revealing resistance to doxorubcin-induced apoptosis, but it remains to be determined whether Bid is critical for apoptosis in vivo. Finally, Dr. El Deiry presented provocative evidence suggesting that p53 may also directly induce transcription of the caspase-6 gene, at least in some types of tumor cells. Thus, the list of p53-dependent targets potentially involved in p53-mediated apoptosis grew significantly at the meeting.

Dr. Tak Mak (Toronto) presented data on Chk2 knockout mice. These mice are viable and do not spontaneously produce tumors but do exhibit increased rates of tumorigenesis when treated with chemical carcinogens, such as 7,12-dimethylbenz(a)anthracene. Dr. Mak presented a model in which DNA damage activates the ATM kinase, in turn activating Chk2. He suggested that both ATM and Chk2 are important for activating p53 for apoptosis induction, with ATM phosphorylating serine 15 and Chk2 phosphorylating serine 20 on the p53 protein. Chk2 knockout mice exhibit resistance to apoptosis induction in certain tissues after X-irradiation, lending further support to the hypothesis that Chk2 plays an important role in p53-induced apoptosis in vivo.

Dr. Yoichi Taya (National Cancer Center Research Institute, Tokyo, Japan) continued the theme of regulation of p53 by phosphorylation, presenting his evidence for a role of phosphorylation at serine 46. Dr. Taya suggested that phosphorylation at serine 46 dictates the difference between p53-induced cell cycle arrest versus apoptosis. He presented evidence that only when p53 is phosphorylated on serine 46 are certain genes induced, among which is the gene encoding p53 apoptosis inducing protein 1. This protein targets mitochondria, inducing depolarization of these organelles and contributing to apoptosis. His group is attempting to purify the serine 46 kinase and has identified a protein complex that appears to contain casein kinase-2α and -2β. Additional work is ongoing to determine whether this kinase complex is critical for some aspects of p53-induced apoptosis.

Additionally on the theme of genotoxic stress responses and mechanisms of chemotherapy-induced apoptosis, Yuri Labeznik (CSHL) used siRNA techniques to generate evidence that caspase-2 may play an important role in apoptosis induction in at least one tumor cell line after treatment with etoposide. An engineered caspase-2 cDNA, which is resistant to the siRNA molecule, was capable of restoring sensitivity to chemotherapeutic drugs in this particular tumor model. Details of the mechanism by which caspase-2 may participate in chemotherapy-induced apoptosis remain unclear.

TNF family death receptors represent important initiators of apoptotic pathways, which are capable in many types of cells of triggering apoptosis through a mitochondria-independent cell death pathway. Several presentations were devoted to this family of cytokine receptors, including presentations by Drs. Marcus Peter (University of Chicago), Ralph Schwall (Genentech), Shigekazu Nagata (Osaka University), and Steven Frisch (Burnham Institute). Dr. Peter presented evidence that the protein c-FLIP can be both an inhibitor and enhancer of Fas-induced apoptosis, suggesting that a delicate balance between the levels of caspase-8 and FLIP determines the ultimate sensitivity of cells to apoptosis induction by death receptors. He showed, using an inducible dimerization system, that dimerization of FLIP can induce caspase-8 activation. Dr. Peter also showed that antisense-mediated reductions in FLIP decrease rather than increase in sensitivity of HeLa cells to Fas-involved apoptosis. These findings suggest that previous models of the role of FLIP in tumor cell biology may have been overly simplified. Dr. Peter also presented interesting data regarding Fas resistance mechanisms in tumors. He showed that although Fas induces caspase-8 activation in some cancer cell lines that overexpress Bcl-2 or Bcl-XL, these cells nevertheless remain viable. Dr. Peter showed that Bcl-XL in collaboration with the death effector domain protein bifunctional apoptosis regulator protein can sequester active caspase-8 on the surface of mitochondria, evidently preventing it from reaching substrates and thereby squelching apoptosis. These findings reveal a postreceptor mechanism for thwarting apoptosis induction by TNF family death receptors.

Continuing the theme of resistance to TNF family death receptors, Ralph Schwall (Genentech) showed that a colon cancer cell line lacking expression of Bax and Bak displays resistance to TRAIL-induced apoptosis. These and other data suggest that some types of cells require mitochondrial participation in the cell death pathways induced by TNF family death receptors and raised concerns about whether TRAIL will be effective against chemoresistant cancers that have developed defects in the mitochondrial pathway. Nevertheless, TRAIL has antitumor activity in animal models and displays synergy with cytotoxic anticancer drugs in vivo. With regards to underlying mechanisms of TRAIL synergy with cytotoxic drugs, Dr. Schwall showed that pretreatment of a Bax-deficient colon cancer cell line with camptothecin induced up-regulation of Bak and DR5, thereby sensitizing these cells to TRAIL. Thus, clinical trials with this biological agent are eagerly awaited.

Steven Frisch (Burnham Institute) focused on mechanisms of anoikis, the process by which epithelial cells undergo apoptosis on detachment from extracellular matrix. He showed that cell detachment plays an important role in controlling the subcellular location of Fas-associated death domain, an adapter protein essential for signaling by TNF family death receptors. Dr. Frisch presented evidence that Fas-associated death domain is located in the nucleus of attached cells, undergoing translocation into the cytosol on detachment. Efforts are under way to elucidate the mechanisms responsible, but protein phosphorylation may play a role. These interesting findings are relevant to the mechanisms of tumor metastasis and warrant further exploration.

IAP family proteins are conserved throughout metazoan evolution and contain at least one copy of the BIR domain. Some members of this protein family have been shown to directly bind caspases, thereby suppressing apoptosis. Overexpression of certain IAP family genes has been documented recently in human cancers, peaking interest in this family of antiapoptotic genes. Several speakers addressed the molecular mechanisms by which IAPs function as caspase inhibitors, as well as explorations of endogenous IAP antagonists, including Drs. Guy Salvesen (Burnham Institute), Emad Alnemri (Thomas Jefferson University), Julian Downward (Imperial Cancer Research Institute), John Abrahms (University of Texas, Southwestern), Herman Steller (Rockefeller University), and Klaus Michael Debatin (Ulm University).

Dr. Salvesen reviewed the three-dimensional crystal structure of the XIAP (BIR2)-caspase-3 complex. The structure reveals the inhibitory mechanism by which XIAP suppresses caspase-3 and may have important implications for generating strategies that might allow disassociation of IAPs from their caspase targets, thereby promoting apoptosis of tumor cells.

Dr. Steller (Rockefeller University) presented evidence that a Ras-dependent pathway in the fly suppresses the proapoptotic effects of IAP antagonist Hid in Drosophila. He presented evidence that the Ras-pathway triggers activity of kinases that phosphorylate and inactivate the Hid protein. It remains to be determined whether analogous proteins exist in mammalian cells. Genetic screens in the fly also revealed mutant versions of the DIAP1, which are resistant to suppression by various IAP antagonists, including RPR, Hid, and Grim. Most of these mutations map to either the BIR1 or BIR2 domains. Interestingly, some mutant versions of DIAP1 display resistance against RPR and Grim but not Hid and vice versa. These and other data prompted Dr. Steller to speculate that the interactions among IAPs and their antagonists may be more complex than originally appreciated. He introduced data about the important role of ubiquitination in controlling the levels of IAPs. In the fly, DIAP1 and DIAP2 contain RING domains, which are speculated to interact with ubiquitin-conjugating enzymes, by analogy to their human counterparts. Dr. Steller presented evidence that RPR, Grim, or Hid interactions with IAPs may influence their functions as E3 ligases. Thus, these IAP antagonists may trigger ubiquitination and destruction of IAPs. Alternatively, the E3 ubiquitin ligase activity of IAPs may be important for their functions as antiapoptotic proteins, allowing them to not only bind but also induce degradation of active caspases. Regardless of the specific biochemical mechanism, genetic screens in the fly in the Steller Lab have revealed genetic modifiers of the DIAP pathway, which include the E2 protein, ubiquitin-conjugating enzyme-D1.

Dr. Alnemri (Thomas Jefferson University) reviewed his laboratory’s efforts to identify IAP antagonists in mammalian cells. Using proteonomics approaches, they have identified three proteins that can interact with the XIAP. Two of these have recently been identified, including SMAC (Diablo) and OMI (HTRA2). Dr. Alnemri reviewed the insights which have come from solving three-dimensional structures of IAP BIR domains complexed to SMAC or OMI, presenting for the first time the crystal structure of OMI in a complex with the BIR3 domain of XIAP. The IAP antagonists share a conserved tetrapeptide motif in their NH2 terminus, which binds to a crevice on the BIR domain, apparently competing with caspases for interaction at this site. Mutations within this IAP binding abrogate the effects of OMI on IAP-mediated suppression of caspases. Interestingly, both SMAC and OMI are sequestered in mitochondria. Their internal IAP-binding motif is revealed on cleavage by mitochondrial proteases during import of these proteins into these organelles. These structural studies reveal a clear path forward with respect to drug discovery and hint that small molecule antagonists of IAPs which mimic the tetrapeptide structure found in SMAC and OMI may be forthcoming.

In this regard, Dr. Klaus Michael Debatin (Ulm University) presented proof-of-concept data in which a heptapeptide representing the IAP-binding NH2 terminus of SMAC was used to promote apoptosis of cancer cells in vitro, demonstrating synergy with cytotoxic anticancer drugs. In an in vivo orthotopic model of glioma, Dr. Debatin showed that a tat-SMAC peptide collaborated with TRAIL in inducing potent tumor regression.

Dr. Julian Downward Imperial Cancer Research Fund, London, England (ICFR) added weight to the evidence of an important role for certain IAP antagonists in apoptosis induction. Using siRNA to down-regulate OMI expression, he showed that OMI but not SMAC plays an important role in UV irradiation-induced apoptosis in the U205 tumor cell line. Interestingly, OMI is a serine protease, and it has been speculated that release of this protease from mitochondria into the cytosol may contribute to caspase-independent cell death. Additional work on this alternative function for OMI is required before firm conclusions can be drawn.

NF-κB family transcription factors have emerged as important contributors to the antiapoptotic state of some cancers. This family of transcription factors has been shown to directly induce the expression of several antiapoptotic genes, including Bcl-XL, BFL1, and cIAP2. Several presentations were devoted to pathways of relevance to NF-κB induction, including presentations by Drs. Gabriel Nuñez (University Michigan) and Michael Karin (UCSD). Dr. Nuñez’s remarks focused on the proteins Nod1 and Nod2. These proteins contain a CARD domain followed by a nucleotide-binding domain (NACHT) and several leucine rich repeats (LRRs). The CARD domains of these proteins interact with Cardiak (RIP2), an adapter protein that communicates directly with the IKK complex. Dr. Nuñez reviewed recent evidence that hereditary mutations in Nod2 are associated with Crohn’s disease, an inflammatory bowel disease. Mutations often affect the LRR region, which is speculated to represent a ligand-binding domain (LBD) for bacteria-derived molecules, such as LPS. Dr. Tschopp (University of Lausanne) talked about a similar protein, NALP1 (NAC), which also contains a CARD, NACHT, and LRRs but in addition carries a newly recognized domain called the PAAD (PYK, PYRIN; DAPIN) domain. Dr. Tschopp suggested that the CARD domain of this protein interacts with the CARD of procaspase-5, whereas the PAAD domain interacts with the adapter protein ASC, connecting it to procaspase-1 via a CARD–CARD interaction. He presented evidence of a large multiprotein complex that is induced in macrophages in response to LPS, which he termed the “inflammosome.” Thus, structurally similar molecules may be involved in activation either of NF-κB or caspases. Interestingly, the caspases activated by NALP-1 (NAC) are involved in inflammation through their ability to cleave and activate proinflammatory cytokines, such as prointerleukin-1β and prointerleukin-18, rather than participating in apoptosis. Dr. Tschopp also reviewed the recent evidence that hereditary mutations in the PAAD family protein known as Cryopyrin are causitory in the autoinflammatory disease syndromes of familial cold urticaria syndrome and Muckle-Wells syndrome. Mutations associated with these inflammatory disorders affect the NACHT domain, and Dr. Tschopp suggested that such mutations allow for spontaneous oligomerization of the Cryopyrin protein, triggering caspase-1 activation.

Dr. Tak Mak continued the theme of exploring the interface between inflammation and cancer in his work on Bcl-10, an adapter protein that links certain CARD family proteins to the IKK complex, promoting NF-κB induction through unclear mechanisms. The Bcl-10 protein interacts with MALT, also known as para-caspase. Chromosomal translocations involving either the Bcl-10 gene on chromosome 1 or the MALT gene on chromosome 14 have been identified in gastric lymphomas associated with chronic Helicobacter pylori infection. Dr. Mak suggested a model in which stimulation of B cells by Helicobacter pylori antigens sets the stage for chromosomal rearrangements, which then trigger activation of Bcl-10 or para-caspase, leading to constitutive activation of NF-κB and thereby emulating some of the survival signals which these cells receive through their antigen receptors. Dr. Mak presented evidence that Bcl-10 knockout mice have a defect in T- and B-cell antigen receptor signaling into the NF-κB pathway. T cells from these animals, e.g., fail to proliferate in response to anti-CD3 and anti-CD28 antibodies and also exhibit defects in production of interleukin-2 and expression of interleukin-2 receptors. Alternative pathways for NF-κB induction remain intact in Bcl-10 knockout cells, including LPS, TNF, and interleukin-1. Thus, the Bcl-10/para-caspase pathway may be important in immune responses, as well as lymphogenesis, making it an attractive pharmaceutical target for autoimmune diseases and lymphoid malignancies.

Dr. Michael Karin (UCSD) reviewed the evidence that elevated levels of NF-κB are found in several types of tumors. He introduced both the classical NF-κB pathway and an alternative pathway in which the protein p100 is inducibly cleaved to produce a p52 protein that can partner with Rel-B to activate NF-κB target genes. Although TNFα, interleukin-1, and LPS induce the classic pathway, this alternative pathway involving inducible p100 cleavage appears to be activated by lymphotoxin-β, B-LyS (BAFF), and possibly T- and B-cell antigen receptors. There was speculation at the meeting that perhaps MALT (para-caspase) could play a role in this inducible cleavage of p100, although no direct evidence is available to date. It is intriguing, however, that Drosophila appears to use a caspase called DRONC to link TOLL-like receptors involved in innate immunity to a NF-κB activation pathway. Dr. Karin also presented evidence that induction of NF-κB may be insufficient for protection from apoptosis. In one model system in vitro, e.g., parallel activation of p38 MAP kinase was found to be essential for NF-κB-mediated induction of antiapoptotic genes. Thus, combinatorial pathway activation may play an important role in dictating whether NF-κB provides effective survival signals and may also tip the balance between the effects of NF-κB on expression of inflammatory versus antiapoptotic genes.

The subject of transcriptional regulation of apoptosis continued with presentations by Drs. Barbara Osborne (University of Massachusetts) on Notch, Stavros Manolagas (University of Arkansas) on steroid hormone receptors, and Eric Baehrecke (University of Maryland) on steroid receptors. Dr. Osborne reviewed mechanisms of Notch signaling, showing how these transmembrane receptors become cleaved by proteases and releasing a cytosolic domain that translocates into the nucleus and regulates transcription in collaboration with other proteins. She presented evidence of deregulated expression of certain Notch family members in human cancers. In T lymphocytes, Notch-1 plays a role in suppressing apoptosis induced by NUR77 (TR3), an orphan member of the retinoid/steroid family of transcription factors. Antisense and other types of experiments have revealed a critical need for NUR77 in activation-induced apoptosis of T cells. In some systems, NUR77 induces apoptosis through transcriptional mechanisms, but recent reports indicate NUR77 can also have extranuclear functions as a promoter of mitochondria-dependent apoptosis, which can be distinguished from its role as a transcription factor. Investigations are under way in the Osborne lab to delineate the mechanisms by which NUR77 and Notch family proteins regulate apoptosis pathways.

Dr. Stavros Manolagas (University of Arkansas) focused his remarks on nonclassical mechanisms of signaling by estrogen receptors (ERs) and androgen receptors (ARs) in the cytoplasm. ER and AR can form complexes with Src and Shc, activating a protein kinase pathway involving MAP kinase kinase, extracellular signal-regulated kinases, and inducing ELK-1 activation. He showed that some small molecule ligands of ERα can induce the kinase pathway without activating transcription. The LBD of the steroid receptors seems to be sufficient for this kinase-inducing activity. Targeting the LBD to membranes potently activates the MAP kinase kinase/extracellular signal-regulated kinase pathway, whereas targeting the LBD to the nucleus does not. Dr. Manolagas demonstrated, interestingly, that a variety of small molecule ligands of estrogen and ARs can promote the kinase pathway in a gender-independent manner with certain androgens stimulating ERα and conversely estrogens stimulating AR. He calls these small molecule ligands “activators of nongenotrophic estrogen-like signaling” (ANGELS). Data from a variety of in vivo studies were presented by Dr. Manolagas addressing tissue-specific effects of synthetic hormone ligands on survival of osteoblasts and osteoclasts in bone, in the absence of concomitant effects on uterus, seminal vesicles, and other sex hormone-dependent tissue. His data on novel activators of nongenotrophic estrogen-like signaling suggested possibilities for selectively preserving bone density, while avoiding trophic effects on cancer prone tissues.

Dr. Eric Baehrecke (University of Maryland) presented data on steroid regulation of programmed cell death during Drosophila morphogenesis. This elegant system has provided a model for genetic analysis of a steroid-dependent pathway triggering programmed cell death. The steroid hormone ecdysone interacts with a specific member of the retinoid/steroid of transcription factors, controlling a genetic program that results in cell death in the salivary gland during pupation. Dr. Baehrecke reviewed evidence for a role in the E74 and E93 genes in salivary gland dysfunction induced by ecdysone. Interestingly, morphological analysis of the dying cells revealed a role for autophagy, where lysosomes fuse with other organelles and digest cellular constituents. Dr. Baehrecke provided evidence of a caspase-independent pathway, through studies in which baculovirus p35 was expressed in salivary gland cells but found to be insufficient for rescue from ecdysone-induced cell death. Using DNA microarrays, Dr. Baehrecke has obtained evidence that expression of several proapoptotic genes is induced by ecdysone in the salivary gland of developing flies, including RPR, Hid, DBorg-2, DARK, DRONC, drICE, as well as APG-9, a homologue of a yeast gene required for production of autophagic vacuoles. Although the relevance to tumor biology of this autophagic pathway for cell death remains unclear, it has been reported previously that Bcl-2 interacts in mammalian cells with Beclin, a protein implicated in regulation of autophagy. Future studies of this interesting cell destruction mechanism appear likely to provide new insights into caspase-independent pathways for cell death.

Continuing the theme of transcription factors regulating apoptosis, Dr. Thomas Look (Dana-Farber Cancer Institute) presented data on SLUG, a homologue of the C. elegans CES-1, which has been implicated in transcriptional regulation of a proapoptotic BH3-only member of the Bcl-2 family called EGL-1. In C. elegans, CES-1 (a zinc-finger transcription factor) suppresses expression of EGL-1, reducing apoptosis in a particular type of cell in this organism. Expression of CES-1 is suppressed by CES-2 (a b-Zip family transcription factor) in the worm. Dr. Look demonstrated that an analogous system appears to exist in mammalian cells, in which HLF (a b-Zip family protein) suppresses expression of SLUG, resulting in an antiapoptotic state. Chromosomal translocations involving HLF in combination with E2F occur in pre-B-cell acute lymphocytic leukemia, dysregulating this pathway and suppressing apoptosis. Furthermore, SLUG(−/−) mice are more sensitive to X-irradiation than wild-type littermates. A search by Dr. Look for the relevant targets of the SLUG protein revealed a role for BAX. Dr. Look demonstrated that SLUG-deficient cells (derived from knockout mice) contain higher and more sustained levels of BAX mRNA after X-irradiation, compared with wild-type cells. Expression of several other proapoptotic Bcl family members, including Harakiri, Bcl-xL/Bcl-2 associated death promoter, Bcl-2 interacting mediator of cell death, Bik-like killer protein, and BH3 interacting domain death agonist, was not changed. Further functional analysis of the role of BAX in apoptosis regulation by SLUG is under way.

The relation of oncogenes and tumor suppressor genes to the core cell death machinery was a topic of considerable emphasis at the meeting, including presentations by Drs. Jean Wang (UCSD) on the RB protein, Gerard Evan (University of California, San Francisco) on c-Myc, Joseph Nevins (Duke University) on E2F, Pier Paolo Pandolfi (Memorial Sloan Kettering Cancer Center) on PML, and Scott Lowe (CSHL) on p53 and p16 INK4a. Gerard Evan (University of California, San Francisco) discussed the role of the c-Myc oncogene in apoptosis regulation and tumorigenesis. The c-Myc protein is known for its activity as both a stimulator of cell proliferation and an inducer of apoptosis. Effective tumorigenesis by c-Myc typically requires collaboration with antiapoptotic proteins that nullify its cell death promoting activity, while leaving its proliferative functions intact. Dr. Evan presented data from Myc/ER transgenic mice, which express a conditional form of the Myc protein whose activity is dependent on binding of small molecule drug Tamoxifen to the ER portion of the fusion protein. Expression of Myc/ER in vivo in pancreatic islets promotes cell proliferation and apoptosis. Dr. Evan demonstrated that loss of p53 or ARF nullifies the apoptotic function of Myc, promoting tumorigenesis. Of significance, expression of Myc in a p53- or ARF-null background is sufficient to produce tumors which are invasive, angiogenic, and genetically unstable. These data imply that even as little as two genetic lesions can be sufficient to provide tumors with the full panoply of attributes needed for aggressive behavior. Data were presented that overexpression of Bcl-XL in pancreatic islets together with Myc/ER also generates adenomas, where Bcl-XL suppresses very potently the proapoptotic effects of c-Myc. In this animal model, turning off Myc by withdrawal of Tamoxifen is sufficient to cause complete regression of adenomas. These findings demonstrate the reversibility of neoplastic process and show that continuous activity of Myc is required to drive neoplastic cell expansion in vivo.

Dr. Joseph Nevins (Duke University) continued with the theme of mechanisms of Myc-induced apoptosis, focusing on the role of E2F family transcription factors. Previous studies have demonstrated reduced apoptosis in response to Myc expression in cells which lack E2F-1, after serum withdrawal. Dr. Nevins demonstrated, in contrast, that E2F-3 knockout cells undergo apoptosis normally in response to Myc. Similarly, although overexpression of either E2F-1 or E2F-3 in cultured cells can drive S phase entry, only E2F-1 is capable of promoting apoptosis. These findings provide evidence that only selected members of the E2F family, of which there are seven in the human genome, induce apoptosis. E2F-1 induces increases in p53 protein. Dr. Nevins reviewed data which have demonstrated a pathway in which E2F-1 induces expression of p19ARF, in turn suppressing expression of MDM2, which then permits accumulation of p53. He presented evidence that DNA-damaging agents induce elevations in E2F-1 protein levels without concomitant increases in E2F-1 mRNA. This effect is caused by a prolongation of the half-life of the protein and is selective for E2F-1, in as much as levels of E2F-2 and E2F-3 do not change in response to DNA damage. Elevations in E2F-1 after genotoxic injury require ATM, based on experiments with cell lines lacking this important kinase involved in a pathway notifying cells of damaged DNA. Dr. Nevins showed that E2F-1 is a substrate for both ATM and the closely related kinase ATM and Rad3-related protein kinase, whereas E2F-2, E2F-3, and other E2F family members are not. Phosphorylation of E2F-1 at serine 31 reduced ubiquitination and turnover of E2F protein, resulting in its accumulation and induction of p19ARF. The specific target genes responsible for E2F-1-induced apoptosis may include APAF-1, which was demonstrated recently to contain E2F-1-binding sites in its promoter and to be directly induced by this transcription factor.

Dr. Jean Wang (UCSD) addressed the role of the RB tumor suppressor gene in tumorigenisis and apoptosis regulation. She presented evidence that RB plays a role in suppressing both cell proliferation and apoptosis. Cells from RB knockout mice display increased apoptosis in response to certain cell death stimuli. Dr. Wang demonstrated that the RB protein is cleaved by caspases, generating an apparent dominant-negative form of the RB protein that promotes apoptosis. Using gene knock-in technology, Dr. Wang demonstrated that mice carrying a noncleavable RB mutant are protected from apoptosis induction by LPS, TNF, and axotomy-induced neuronal apoptosis but are not resistant to X-irradiation or doxorubicin. These findings suggest a specific role for RB in regulating certain pathways for apoptosis, with mechanisms remaining unclear at this time.

Pier Paolo Pandolfi (Memorial Sloan Kettering Cancer Center) addressed the role of the tumor suppressor PML in apoptosis regulation. The PML protein resides within subdomains of the nucleus called PODs. PML somehow promotes apoptosis in a general way, although the mechanisms remain unclear. Dr. Pandolfi presented evidence that defects in the expression or structure of the PML protein are very common in human cancers, including many solid tumors. PML also becomes deregulated through chromosomal translocations with the retinoic acid receptor-α gene in acute promyelomonocytic leukemias. The fusion of PML with retinoic acid receptor-α disrupts POD formation, resulting in an apoptosis-resistant state. Dr. Pendolfi demonstrated that SUMOlyation of PML is required its proper targeting to PODs and for recruitment for other proteins into these structures, including the proapoptotic protein DAXX. He showed that PML recruits p53 into PODs after genotoxic injury and proposed the formation of a ternary complex containing p53, PML, and CREB binding protein, which results in acetylation of p53, rendering p53 competent to transactivate target genes. Dr. Pandolfi indicated that PML also interacts with p63 and p73 (members of the p53 family) and similarly enhances their transactivation functions. However, PML may also regulate apoptosis through additional mechanisms which do not require p53 family proteins, as evidenced by its ability to affect translocation of the proapoptotic DAXX protein to PODs. DAXX is a putative transcriptional transrepressor which somehow enhances apoptosis induction by Fas/TNF family death receptors and which interacts physically with PML. Genetic ablation of the PML gene results in increased resistance of cells to Fas-induced apoptosis, although the specific mechanisms remain to be clarified. Regardless, these studies of PML reveal an interesting link between nuclear structures and the regulation of apoptosis pathways.

Dr. Scott Lowe (CSHL) concluded the conference with a presentation on the role of defective apoptosis mechanisms and chemoresistance in cancer. Using a variety of elegant in vivo models, Dr. Lowe provided evidence that suppression of apoptosis by overexpression of Bcl-2 has a profound influence on in vivo resistance of lymphoma cells to antineoplastic drugs. Dr. Lowe used the Eμ-Myc transgenic mouse model of B-cell lymphoma to explore the effects of alterations in various tumor suppressor genes and proto-oncogenes on responses to chemotherapy in vivo. As expected, overexpressing Bcl-2 in Eμ-Myc lymphomas by retroviral transduction suppressed apoptosis induction after treatment with DNA-damaging drug cyclophosphamide, correlating with failure of involved lymph nodes to shrink. However, despite the paucity of cell death, Dr. Lowe provided evidence that these Myc/Bcl-2 tumors do not generally progress, apparently because of DNA damage-induced replicative senescence. Studies with genes of the INK4a locus suggested a role for both the p16 cyclin-kinase inhibitor and p19ARF regulator of the Mdm2/p53 pathway in postchemotherapy-induced senescence. Thus, even if apoptosis is suppressed, anticancer drugs may induce replicative senescence, thereby preventing further tumor growth. Homozygous disruption in INK4a genes (p16/p19) or p53 was shown to ablate the senescence response, resulting in failure of chemotherapy in this transgenic mouse model.

Overall, the meeting showed that tremendous progress has been made in recent years toward understanding the molecular basis of apoptosis regulation in normalcy and its dysregulation in cancer. Importantly, from an understanding of these basic mechanisms, a variety of new strategies for combating cancer is beginning to emerge, e.g., the discovery of endogenous antagonists of antiapoptotic Bcl-2 and IAP family proteins and elucidation of peptidyl-motifs in these antagonists that are sufficient for overcoming apoptosis suppression reveals a path forward for production of nonpeptidyl small molecule antagonists that could be used for treatment of cancer. Small molecule inhibitors of the NF-κB-activating kinases (IKKs) could also be useful for restoring apoptosis sensitivity in cancers, as well as perhaps drugs that antagonize para-caspase, if evidence continues to mount implicating this novel protease in NF-κB induction in some types of tumors. Biologicals for the treatment of cancer, such as TRAIL, also appear to be quite promising, suggesting the possibility of tapping into mitochondria-independent apoptosis pathways and thereby triggering alternative cell death pathways in chemo-resistance cancers. In addition to these new strategies for inducing apoptosis of cancer cells, it should not be forgotten that already, several agents are in human clinical trials which touch on some of the same mechanisms mentioned above, including: (a) Bcl-2 antisense oligonucleotides currently in Phase III trials for chemorefractory malignancies; (b) proteasome-inhibiting drugs that (among other things) prevent IkB degradation and thereby thwart NF-κB activation; and (c) p53 and E1a gene therapies, which trigger expression of apoptosis-inducing genes and sensitize tumor cells to radiation. Prospects, therefore, seem bright for applying information about apoptosis mechanisms for improving treatment of cancer in the near future.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This conference, organized by the American Association for Cancer Research, was held February 13–17, 2002, in Waikoloa, Hawaii.

3

The abbreviations used are: CSHL, Cold Spring Harbor Laboratory; VDAC, voltage-dependent anion channel; IAP, inhibitor of apoptosis; SKL, Sickle; RPR, Reaper; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand; TNF, tumor necrosis factor; DIAP, Drosophila inhibitor of apoptosis protein; INK4a, inhibitor of CDK4; DAXX, death-associated protein; PML, promyelocytic leukemia protein; PODs, PML oncogenic domains; BAX, Bcl associated X protein; EGL1, egg-laying defective 1 protein; OMI, mitochondrial serine protease HtrA2/Omi; BIR, baculovirus IAP repeat; FLIP, FLICE inhibitory proteins; ATM, atxia telangiectasia mutated; SMAC, second mitrochondria-derived activator of caspase; CARD, caspase activation and recruitment domain; UCSD, University of California, San Diego; NF-κB, nuclear factor-κB; NACHT, nucleotide-binding domain; IKK, IκB kinase; LRR, leucine-rich repeat; AR, androgen receptor; ER, estrogen receptor; LPS, lipopolysaccharide; ARF, ADP ribosylation factor; RB, retinoblastoma; LBD, ligand-binding domain; POD, PML oncogenic domain; XIAP, X-linked inhibitor of apoptosis protein; HLF, hepatic leukemia factor.