Cancer: translation of advances in basic science to human therapy

Colloque organisé par Moshe Yaniv du 22 au 28 juillet 1995

Participants

Latifa Bakiri (Institut Pasteur, Paris), Dirk Bohmann (European Molecular Biology Laboratory (EMBL), Heidelberg, Allemagne), Sara Courtneidge (Sugen Inc., Redwood City, USA), Jean Feunteun (Centre National de Recherche Scientifique (CNRS), Institut Gustave Roussy, Villejuif), Marco Giovannini (Institut Curie, Paris), Edward Harlow (Massachusetts General Hospital Cancer Center, Boston, USA), Peter Herrlich (Institut fur Genetik, Karlsruhe, Allemagne), Peter Howley (Harvard Medical School, Boston, USA), Adi Kimchi (Weizmann Institute of Sciences, Rehovot, Israël), Dominique Lallemand (Institut Pasteur, Paris), Davie Lane (University of Dundee, Great Britain), Arnold Levine (Princeton University, USA), Frank McCormick (Onyx Pharmaceuticals, Richmond, USA), Oren Moshe, organisateur (Weizmann Institute of Sciences, Rehovot, Israël), Carol Prives (Columbia University, New York), Gilles Thomas (Institut Curie, Paris), Thea Tlsty (University of San Francisco, USA), Bruno Tocqué (Rhône-Poulenc-Rorer recherche-développement, Vitry-sur-Seine), Delphine Torchard (Institut Gustave Roussy, Villejuif), George Vande Woude (Cancer Research and Development Center, Frederick, USA), Moshe Yaniv, organiser (Institut Pasteur, Paris).

Compte-rendu

The past decade has provided significant advances in our knowledge base of the cellular and molecular mechanisms of cancer. We are now at a point where it is possible to define specific sets of genes and proteins that are involved in the malignant transformation of a normal cell to a cancer cell. Many of the advances have come from studies of genes called oncogenes, and more recently, from studies of tumor suppressor genes. It is now quite clear that specific human cancer arises as a consequence of multiple independent genetic mutations which may result either in the activation of an oncogene or the inactivation of a tumor suppressor gene. Although there are multiple oncogenes (the current number being about 100) and several different tumor suppressor genes, certain unifying principles have begun to emerge. The oncogenes generally appear to be involved in intracellular signaling pathways which positively regulate the proliferative activity of the cell. Thus, they include cell surface receptors, membrane associated kinases, cytoplasmic signaling molecules, and nuclear transcription factors. The normal function of the tumor suppressor genes on the other hand is to negatively regulate cellular proliferation. The functions which have been ascribed to tumor suppressor genes such as p53 and the retinoblastoma tumor suppressor gene (Rb) which often occurs as a result of mutation in human cancer results in the loss of the normal cellular controls of proliferation. Studies over the past few years have revealed that the activities of these oncogenes and tumor suppressor genes appear to be converging on a few critical pathways involved in regulating cellular proliferation. The opportunity therefore looms in the near future for the assay and development of specific diagnostics and therapeutics in oncology based on our expanding knowledge of these pathways and the proteins controlling them.

The goal of this meeting, entitled “Cancer: The translation of advances in basic science to human therapy”, was to bring together investigators from different fields to discuss how the current advances in molecular biology, biochemistry, genetics and cell biology can be applied and translated to the treatment of human disease, particularly cancer. In July 1995 at Les Treilles, 19 leading scientists involved in biomedically-relevant basic and applied research presented the work carried out in their laboratories and how it might be applicable to the development of new therapies. The topics that were discussed included: signal transduction and the cell cycle, viral and cellular oncogenes, novel cellular targets, p53>, new cancer genes, apoptosis.

Ed Harlow (MGIH Cancer Center, Charlestown) opened the meeting with a discussion of the cell cycle, noting that the cell’s decision to divide is governed by its ability to integrate the responses to several key signal transduction pathways that respond to extracellular and intracellular cues. When important proteins in these pathways are mutated, cell division occurs at inappropriate times. His laboratory has been examining the RB (Retinoblastoma tumor suppressor gene) pathway, one of the signal transduction networks that is frequently mutated in human cancer. This pathway controls the transcription of a set of genes that are temporally regulated during normal cell cycle progression. When the RB pathway is deregulated, the temporal regulation of these genes is disrupted, presumably leading to some growth advantage for the developing tumor cell. At present, there are four known steps to this pathway. A cell cycle regulated transcription factor E2F is the ultimate target of this pathway. Its transcriptional activity is controlled by interaction with the retinoblastoma tumor suppressor protein, pRB, for whom this pathway is named. When pRB is bound to E2F transcription is inhibited. pRB’s association with E2F is controlled by phosphorylation of pRB by the cyclin D/cdk4 kinase complexes. These kinases are activated in mid-G1 and by phosphorylating pRB, they regulate the synthesis of E2F-mediated transcription events. Upstream control of the cyclin D/cdk4 kinase complexes is regulated at least in part by the interaction with small molecular weight kinase inhibitors such as p16. It appears that this RB pathway is extremely important in human cancer, as some component of this pathway is mutated in almost all cancers. It is also striking to note that never are more than one member of this pathway mutated in the same tumor cell, arguing that once the signaling events are disrupted, further mutations in this pathway do not lead to any advantage to the developing tumor cell. These data argue that the RB pathway is a key regulator of cell proliferation.

Thea Tlsty (University of California, San Francisco) discussed the heterogeneity of cell cycle checkpoint control in human cells. The study of how genomic integrity is regulated is important not only in the formation and progression of a neoplasia, but also in how a tumor responds to therapy. Previous studies identified p53 as a member of a signal transduction pathway that modulates genomic instability specifically gene amplification. The data presented in this meeting identifies several new complementation groups which control gene amplification that are independent of p53. In a separate set of studies, it has been found that the cell cycle checkpoint control that functions in primary human fibroblasts is differentially regulated in primary human epithelial cells. Several types of human epithelial cells have been tested, skin, breast and bronchial, and all were found to have a relaxed checkpoint control relative to fibroblasts. Finally, the difference in checkpoint control was observed when the cells were exposed to a metabolic inhibitor, PALA, as well as high energy gamma radiation.

Sarah Courtneidge (European Molecular Biology Laboratory [EMBL], Heidelberg) presented data from her laboratory on the SRC family of tyrosine kinases and the cell cycle. The SRC family of protein-tyrosine kinases are a group of enzymes that associate with cytoplasmic membranes. Each is tightly regulated by tyrosine phosphorylation, and each can be rendered transforming by mutation of its regulatory tyrosine. There are no published examples of human tumors with mutations in an SRC family kinase. Nevertheless, several cases have been described where their activity is elevated in tumor cells, including a high proportion of breast and colon cancers. Activation may occur by mutation or abnormal expression of the proteins that normally regulate SRC family kinases. Her laboratory has studied the involvement of SRC family kinases in signal transduction processes initiated by activated growth factor receptors. She has found that they are required for DNA synthesis in response to several growth factors, probably resulting in the transcriptional activation of Myc. These findings raise the possibility that SRC family kinases may make good therapeutic targets in tumors driven by the overexpression of these growth factor receptors. Furthermore, the SRC family kinases are required in G2 for fibroblasts to undergo mitosis.

Frank McCormick (ONYX Pharmaceuticals, Richmond, California) discussed novel cancer therapies based on human cancer genetics utilizing small molecular and viral strategies. Ras oncogenes play a direct role in more than 30 % of all human cancers. Pathways controlled by Ras proteins are mis-regulated in these cancers, and offer opportunities for therapeutic intervention. Two effector pathways have been described for Ras proteins, which regulate the cell cytoskeleton. The two pathways are oncogenic when activated independently, and act synergistically for the full transforming effect of oncogenic Ras. The function of the Rac and Rho pathway in Ras transformation does not appear to their effects on cell morphology, since these effects are suppressed in Ras transformed cells. Rather, they send signals through poorly understood pathways to converge with the MAP kinase cascade on transcriptional regulators. His laboratory is currently screening for compounds that block these signaling cascades. In addition, his group is developing a novel approach to cancer therapy based on lytic viruses whose host range is restricted to tumor cells. An adenovirus whose replication is dependent on loss of p53 is the first example of such an approach. This virus has been shown to replicate efficiently in tumor cells, to kill them and to cause complete destruction of human tumors in nude mice.

Bruno Tocqué (Centre de Recherches de Rhône Poulenc Rorer, Vitry-sur-Seine, France) discussed oncogene neutralizing binding decoys to control Ras dependent transformation and to induce tumor cell apoptosis. Src homology type 2 (SH2) and type 3 (SH3) domains appear to have an important role in signal transduction pathways initiated by tyrosine kinases. SH2 domains allow proteins with signaling functions to interact with tyrosine kinases and tyrosine-phosphorylated proteins at the plasma membrane, whereas SH3 domains allow a distinct type of interaction through binding to proline-rich sequences. The adaptor protein Grb2 consists of one SH2 domain and two SH3 domains and connects tyrosine kinase receptors to activation of the Ras pathway. Its closely related counterpart, Grb3-3, thought to arise by alternative splicing of Grb2 transcripts, lacks a functional SH2 domain but retains functional SH3 domains. Evidence suggests that Grb3-3 might deliver specific signals causing cells to undergo apoptosis. He discussed the mechanism of Grb3-3 function and its putative involvement in several pathologies. His data strengthens the notion that cells may use alternative splicing as a means to drive either a proliferative or a suicidal program. He reported on the purification and the molecular cloning of a Ras-GTPase Activating Protein, (GAP)-binding protein, G3BP, a cytosolic 68 kDa protein, that co-immunoprecipitates with GAP and that physically associates with the SH3 domain of GAP, which had previously been shown to be essential for Ras signaling. G3BP mRNA is widely expressed as a single mRNA species of 3.3 kb, being more abundant in adult skeletal muscles. The G3BP cDNA revealed that G3BP is a novel protein and shares several features with heterogeneous nuclear RNA binding proteins (hnRNPs): G3BP, a 466 amino acid protein, exhibits ribonucleoprotein motifs (RNP), RNP1 and RNP2, a RG rich domain and acidic sequences. Recombinant G3BP binds effectively to GAP SH3. G3BP co-immunoprecipitates with GAP only when cells are in a proliferating state, suggesting a recruitment of a GAP-G3BP complex when Ras is in its activated conformation. The data support a connection between GAP and RNA metabolism. A Ras GTP initiated regulation of specific mRNAs, such as those of Grb2 and Grb3, may be required for Ras dependent responses.

Moshe Yaniv (Institut Pasteur, Paris) discussed cell cycle and transformation control of the snf/swi human complex. The snf/swi complex was characterized genetically and biochemically as a multiprotein complex that is involved in chromatin remodeling during transcription. The snf2/swi2 protein is considered as the motor of this complex since it contains seven domains typical of ATP-dependent DNA and RNA helicases. The drosophila homologue of this protein, brahma, was cloned as a suppressor of polycomb. lt permitted tire isolation of two human homologues hbrm and brg1, two closely related proteins of a molecular mass close to 200kd. Preparation of antibodies specific to each one of these proteins allowed to follow the fate of these proteins during the cell cycle. While both proteins are nuclear in interphase, both are excluded from the condensed chromatin during metaphase. This is accompanied (or caused) by phosphorylation of both proteins and degradation of the majority of hbrm. Re-entry to G1 is coincident with dephosphorylation of both proteins and denovo synthesis of hbrm.

Examination of Ras transformed mouse cells revealed that brm is down-regulated at the level of transcription by oncogenic ras while brg1 remains unchanged. There is a good correlation between the level of ras and the decrease or total disappearance of brm. Transfection of those cells with an LTR based construct and constitutive synthesis of physiological levels of brm reverses the transformed phenotype (suppression of growth in low serum, agar and nude mice) without blocking cell growth. In a parallel study it was shown that brm is absent in F9 embryonal carcinoma cells but reappears upon differentiation while brg1 is constitutive. Biochemical evidence also suggests that the brm and brg1 containing complexes are distinct. These results clearly demonstrate that brg1 and brm follow different fates during cell cycle, ras transformation and cell differentiation and suggest that brm may be an important regulator of cell growth.

George Vande Woude (NCI-Cancer Research and Development Center, Frederick) discussed Mos and Met, and mechanistically how they might be involved in transformation and metastasis. The Mos proto-oncogene product is a regulator of oocyte meiotic maturation in vertebrates. One key substrate of Mos is the kinase which activates MAP kinase (MAPK). While p34cdc2 kinase cycles between meiosis I and II and, unique to M-phase of oocyte maturation, MAPK is constitutively activated. His laboratory has demonstrated that the spindle of somatic cells expressing the Mos/MAPK cascade resembles the anastral spindles found in maturing oocytes. In these cells, the spindle pole is attached to the membrane which is probably responsible for all of the cells in G2-M becoming binucleated. Important for this interpretation is the antithetical result that oocytes deficient in Mos/MAPK, (but not p34cdc2 kinase) have asters and mitotic-like spindles and undergo cytokinesis. These studies may, for the first time, explain parthenogenesis in oocytes. Moreover, Mos/MAPK activation in oocytes can be induced in the absence of p34cdc2 activity. In these oocytes, large openings form in the germinal vesicle adjacent to partially condensed chromatin, and microtubule arrays emanate from these regions. This activity is consistent with the proposal that the Mos/MAPK cascade participates in the formation of the anastral phenotype of the meiotic spindle during oocyte maturation. The expression of the transformed phenotype, which results from constitutive activation of MAPK by certain oncogenes in somatic cells, is therefore due to the inappropriate expression of meiotic phenotypes imposed by MAPK on all stages of the somatic cell cycle.

George Vande Woude also discussed the hepatocyte growth factor/scatter factor (HGF/SF) which is the ligand for the Met proto-oncogene tyrosine kinase growth factor receptor. HGF/SF is preferentially expressed in cells of mesenchymal origin and is an effector of mitogenic, motogenic (scattering) and morphogenic responses, especially in epithelian and carcinoma cells where Met is expressed. HGF/SF is also a potent inducer of angiogenesis and causes a variety of cells expressing Met to become invasive in vitro. HGF/SF induces both mitogenic and invasive behavior in SK LMS-1 cells, a human leiomyosarcoma cell line. His laboratory has studied how HGF/SF:Met participate in the invasive process. HGF/SF is structurally related to the plasminogen activator family and urokinase (uPA) proteolytically activates HGF/SF from uncleaved pro-HGF/SF. The expression of uPA receptor (uPAR) and UPA is induced within eight hours of treatment of SK LMS-1 cells with HGF/SF. This results in uPA becoming cell associated and promoting the activation of plasminogen to cell associated plasmin. Thus, cell associated uPA and plasmin are able to dissolve extracellular matrix components and, when coupled with the motogenic (motility) scattering activity also induced by HGF/SFMet, signaling would allow cells to invade tissue. Therefore, the mitogenic, motogenic, angiogenic and invasive properties of HGF/SF:Met signaling explains many of the phenotypes required for tumor invasion and metastases. This receptor-ligand pair may also play an important role in eliciting these phenotypes in human tumors.

Peter Herrlich (Institut fur Genetik, Karlsruhe) discussed therapy attempts exploiting the molecular roles of a family of adhesion molecules. During invasion and metastasis, tumor cells of diverse origin need to cope with similar micro-environments and to surmount similar obstacles. It is therefore reasonable to assume that they make use of common molecular mechanisms. These mechanisms are also likely to occur in normal life, e.g. at some stage of embryogenesis or even in the adult organism. The adhesion protein family CD44 serves as a good example. Their diversity is generated by alternative splicing from a wealth of 10 to 11 variant exons and, on top, by massive post-translational modifications. Alternative splicing is governed by trans-acting factors with at least some splice variant specificity. Certain CD44 isoforms are preferentially expressed during tumor progression and indicate poor prognosis in Non-Hodgkin lymphoma, colorectal and mammary cancer. The expression pattern in normal cells together with biochemical analyses have guided them to follow some features of their action. The common pattern seems to be their involvement in proliferation control. For instance, lymphocytes from transgenic mice expressing the isoform CD44 v4-v7 under Thy-1 promoter control respond faster and at lower dose to antigen, and bone marrow from these mice repopulates lymphoid organs of irradiated recipients faster than control bone marrow. Interesting sites of expression are Schwann cells, where CD44 seems to be incorporated into the interplay between the tumorigenic neu-oncogene-drive and the tumor suppression by merlin (NF2), and, in the embryo, the apical ectodermal ridge of limb buds, where CD44 v1-v10 represents an essential component in the growth stimulation of the underlying mesenchyme. The variability of the extracellular domain of the CD44 proteins suggests that they act in different micro-environments and it may be this aspect that causes tumor cells to produce specific CD44 proteins such that they proliferate in challenging new micro environmental conditions during metastasis.

Dirk Bohmann (EMBL, Heidelberg) discussed the biochemistry and genetics of Jun-regulation. The transcription factor encoded by the c-jun proto oncogene is a nuclear component of a number of signal transduction pathways which are implicated in the control of cell-growth, differentiation, apoptosis, etc. Defects in Jun regulation may lead to cell transformation and cancer. The control of Jun activity by phosphorylation was discussed and a model presented to integrate different modes of regulation by phosphorylation. MAPK-phosphorylation was shown to be a primary event which controls other regulatory phosphorylations, primarily mediated by CK-II. The negative regulation of Jun activity was investigated. C-jun is a substrate for ubiquitin-dependent degradation in vivo. V-jun (isolated from a chicken transforming retrovirus) escapes this control mechanism because it has lost, through the deletion of the N-terminal delta-domain, an important ubiquitination and degradation

signal. The correlation between Jun-stabilization and transformation was extended to Ras-transformation which causes an increase in c-Jun stability. This effect is mediated by MAPK phosphorylation, indicating that Jun degradation is a signal dependent mechanism. During Drosophila eye development, Jun is required for “sevenless” -induced photoreoeptor différentiation. This process involves the phosphorylation of Jun by the Drosophila ERK-type MAPK “rolled” and a synergistic interaction with the Ets-like transcription factor “pointed”. The Drosophila eye appears to be a powerful genetic system to study the regulation (and de-regulation) of the universal Ras pathway, as well as Jun signaling.

Peter Howley (Harvard Medical School, Boston) discussed the papillomaviruses and human cancer. The human papillomavirus is associated with human cancer and code 2 viral oncoproteins which are expressed in the human cancers. These oncoproteins are referred to as E6 and E7. E6 targets the ubiquitination and degradation of p53, functionally inactivating p53. E7 targets the pocket proteins, including the retinoblastoma tumor suppressor protein. The half-life of p53 in HPV positive cancer cells and in HPV immortalized cell lines is significantly shorter than that seen in primary cells. A molecular explanation for this observation is provided by in vitro studies which have shown that the E6 proteins of the cancer associated HPVs (i.e. HPV-16 and -18) can complex p53 in vitro and promote its ubiquitination and subsequent degradation by the proteasome. Expression of HPV-16 E6 has also been shown to decrease the half-life of p53 in vivo. It is unknown, however, if the complex involving p53 and E6 is specifically required to target p53 to the proteasome or if p53 is normally degraded by the proteasome in the absence of E6. Recent studies have suggested that p53 may normally be degraded by ubiquitin mediated proteolysis in the absence of E6. For example, there are increased levels of p53 protein in cells in which the ubiquitin system is conditionally inactivated. Furthermore, limited ubiquitin dependent degradation of p53 can be observed in vitro in rabbit reticulocyte lysate in the absence of E6. The effect of specific peptide aldehydes which can inhibit the 26S proteasome complex on the steady state levels of p53 and on the cell cycle were examined. The steady state levels of wild type p53 were increased in cells treated with the proteasome inhibitors, and pulse-chase experiments indicated that the increased p53 levels resulted from stabilization of the protein. p53 levels were also increased following inhibitor treatment of cells expressing the E6 protein of the cancer-associated HPVs. These results implicate the proteasome and ubiquitin dependent proteolysis in the degradation of p53 in normal cells as well as in E6 expressing cells. Increased levels of the cyclin:cdk inhibitor p21, a downstream effector of p53 function, were also observed in cells treated with the proteasome inhibitor. The increase in p21 levels was accompanied by an increase in p21 mRNA, and appeared to involve p53 dependent and possibly p53-independent mechanisms. Flow cytometric analysis indicated that proteasome inhibitors affect cell cycle progression in both the G1 and G2 phases.

Arnold Levine (Princeton University) discussed the loss of p53 functions in cancer. The p53 protein is present at high levels in embryonal carcinoma cells, the stem cells of testicular teratocarcinomas. Despite the fact that the protein is present in high levels, its specific activity as a transcription factor for several genes tested (mdm 2, p21-WAF-1) is very low to non-detectable in these cells. Induction of a differentiation pathway in these cells, by using dibutyryl cyclic AMP and retinoic acid, produces endoderm cells containing one tenth the level of p.53 protein with 18 – 22 fold increased activity as a transcription factor. Thus the signals to differentiate these cells increase the specific activity of p53 by 180 – 220 fold. Similarly, treatment of the embryonal carcinoma cells with a DNA damaging agent, such as etoposide, increases the p53 protein levels and dramatically increases the activity of the p53 protein as a transcription factor. When this occurs, the embryonal carcinoma cells undergo a p53 dependent apoptosis. Embryonal carcinoma cells without p53 (from the knockout mouse) fail to undergo apoptosis or undergo apoptosis with a 100 fold lower efficiency. These results help to explain why p53 is found exclusively in the wild-type configuration (no selection for mutation) in embryonal carcinoma cells; i.e. it is inactive in these cells and so there is no selective pressure for mutation. These results also help to explain why treatment of this cancer with a DNA damaging agent, cis-platin, induces apoptosis in the cancer cells and cures 90 – 95% of these cancer cases. This is an example where the molecular studies have helped to explain the unusual nature of this type of cancer.

Carole Prives (Columbia University, New York) discussed the regulation of the structure and function of wild-type and mutant forms of the p53 tumor suppressor protein. When wild type p53 protein is induced in cells, the response is either cell cycle arrest or apoptosis. The p53 gene is mutated in over 50% of most types of tumors. The mutations within the p53 protein are located within the centrally located DNA binding domain. Frequently mutant p53 protein is highly expressed in tumor cells. Such mutants have altered or defective DNA binding properties. An important goal would be to understand the ways p53 is regulated in order to develop ways to destroy mutant p53 expressing cells. Her laboratory has characterized the roles of sequences and sites within the N- and C-termini of the p53 protein in regulating the function of the central specific DNA binding domain. These studies indicated that phosphorylation of the cyclin dependent kinase site within the C-terminus stimulates and alters the DNA binding capability of wild-type p53. Her laboratory demonstrated that certain DNA structures akin to those that arise during damage and repair of DNA in cells are capable of stimulating specific DNA binding by p53. Mutant forms of p53 are capable of binding to DNA, and, in some cases, activating transcription at sub-physiological temperatures but not at 37° C, normal body temperature. She and her co-workers have identified a means to stabilize DNA binding by mutant forms of p53 at physiological temperature. This may eventually be developed for therapeutic purposes with the aim of inducing mutant forms of p53 in tumors to function like the wild-type form leading to cell arrest or death.

David Lane (Dundee University, Great Britain) continued the discussion on the regulation of p53 function. The discovery of the specific loss of suppressor gene functions in many human tumor cells has provided a great focus and challenge for the development of novel treatments for human cancer. How might these lost functions be replaced or exploited to provide specific therapies that would be more effective and less damaging to the normal cells of the patients? While the simple replacement of the missing gene using the approach of “gene therapy” is very attractive, enormous practical difficulties remain in effective manufacture and targeting of these agents. His group has instead taken an approach directed towards the invention and discovery of small molecules able to restore or replace gene function. The model System they have chosen is the p53 tumor suppressor gene pathway, since loss of the pathway by genetic mutation is found in half of all human cancer and is particularly prevalent in those very common cancers such as lung, breast and colon cancer where novel therapeutics could have most impact. Three systems were found in which short synthetic peptides carry out the desired effect or function in vitro. Their success suggests that the search for small organic molecules that mimic these peptides may be fruitful.

Moshe Oren (Weizmann Institute, Rehovot, Israel) discussed p53 and apoptosis. Many attempts have been made to identify the normal functions of the p53 protein, in order to understand how its activity may interfere with the development of cancer. These studies have revealed that one of the important functions of p53 is to promote the death of cells which are potentially harmful to the body, and particularly cells carrying genetic alterations which may potentially lead to the conversion of a normal cell into a cancer cell. Most notably, p53 is strongly activated in response to various types of DNA damage. While some cells respond to the activation of p53 by undergoing a process of apoptosis (programmed cell death), others fail to do so. His work tries to understand what factors determine whether a cell will undergo apoptosis in response to p53 activation, and what molecular mechanisms mediate the induction of cell death by p53. The importance of this knowledge is due to the fact that the success of current cancer therapies is greatly dependent on their ability to induce apoptosis, which may be affected by the status of p53 in the tumor.
A variety of studies have indicated an inverse correlation between the ability of the cell to exit the cell cycle and its tendency to die in response to p53 activation. This has led to the suggestion that a stable growth arrest may protect cells from p53-mediated apoptosis, and that such arrest may be dependent on the normal function of the RB/E2F pathway. To test this notion, an apoptosis assay was developed in HeLa cells, which are detective for both p53 and RB function due to the expression of HPV E6 and E7, as discussed by Peter Howley. Transfection of active p53 into HeLa cells resulted in extensive apoptosis. This response was greatly inhibited, however, by co-transfection with a gene encoding active RB. A mutant RB, defective in its ability to bind E2F, was unable to protect cells from p53-mediated apoptosis.
In order to identify the mechanisms responsible for p53-mediated apoptosis, a series of p53 deletion mutants were tested in the HeLa cells assay. The p53 protein is a sequence-specific DNA binding protein and transcription activator, and these properties are believed to be responsible for many, if not all, of its biological activities. Surprisingly, a truncated p53 protein containing only the N terminal half of the protein was a good inducer of apoptosis, even though it was totally incapable of sequence-specific DNA binding and transactivation. Moreover, overexpression of the Mdm-2 oncoprotein, which completely repressed the ability of p53 to transactivate specific genes in HeLa cells, did not block the ability of p53 to induce apoptosis. Hence, at least in some cell types, the induction of cell death by p53 does not require specific DNA binding and transcriptional activation. These data suggest the existence of a potentially novel mechanism for p53-mediated cell killing.

Adi Kimchi (Weizmann Institute, Rehovot, Israel) discussed “death associated proteins” or DAP genes, which are novel positive mediators of programmed cell death. Programmed cell death can be triggered by external stimuli among which diffusible cytokines may play an important role. Her laboratory used a functional approach of gene cloning for the rescue of these regulated Systems. The rescue was based on positive growth selection of cells after transfection with antisense cDNA libraries. In response to interferon-gamma, HeLa cells undergo a type of cell death that has cytological characteristics of apoptosis. Three novel DAP genes and a known protease were isolated from treated HeLa cells by this strategy. The anti-sense RNA-mediated inactivation of these genes attenuated the process of cell death without interfering with early responses to the cytokine. The full length cDNAs were isolated and the structural/functional properties of the proteins were further analyzed. It was found that DAP-1, codes for a basic, proline rich 15 KD protein that carries a potential SH3-binding motif. Another gene, DAP-kinase, turned out to be a structurally unique calmodulin-dependent serine/theonine kinase. This kinase (160 KD) is located in the cytoskeleton, and carries a few additional functional motifs, including 8 ankyrin repeats in the middle of the molecule and the “death domain” at the carboxy terminal part of the protein. These motifs may form stable complexes with other effector proteins along the death pathways. The intracellular localization of DAP-kinase was correlated with the first hallmarks of cell death that involve loss of stress fibers and cytoskeletal disruptions.
DAP-3 turned out to code for a 46KD protein that carries a potential ATP/GTP binding motif. A fourth gene was identified as cathepsin-D aspartyl protease. Analysis of its expression during cell death revealed that the processing of this protease was altered during cell death, resulting in the accumulation of high levels of active precursor forms, usually found in pre-lysosomal cellular compartments. Cathepsin D therefore represents another important protease with a basic function during programmed cell death in addition to the ICE family of proteases.
The three DAP genes are widely expressed in many cells and tissues. They mediate other scenarios of cell death, such as the Fas/APO-1-induced apoptosis. The overexpression of these genes is toxic, and cells die very rapidly while displaying the set of hallmarks typical to apoptosis. Altogether, these data suggest that a set of genes that function as positive mediators of apoptosis has been identified. They have the potential to be used in gene therapy of cancer, as well as to serve as diagnostic and prognostic markers. Specific inhibitors of these genes will be hopefully used in preventing accelerated cell death in neurodegenerative diseases.

Gilles Thomas (Institut Curie, Paris) discussed positional cloning of genes involved in human tumors, focusing on an oncogene and a tumor suppressor gene from the 22q12 region involved in human tumors. The possible occurrence of an oncogenic event in the 22q12 region was first indicated by a recurrent t(11;22) chromosome translocation found in Ewing sarcoma (ES). Translocation of the same region to a variety of different chromosomes were subsequently observed not only in ES but also in malignant melanoma of soft parts (MMSP), in intra-abdominal desmoplastic tumors (IDT) and in extraskeletal myxoid chondrosarcomas (EMC). Independently, germ line alterations associated with the tumor predisposing disease neurofibromatosis type 2 (NF2) was shown by genetic linkage to reside in 22q12. In order to contribute to the identification of the 22q12 genes which may be involved in these processes, a physical map of part of the q12 region was performed and a contiguous region of about 1.5 megabase was cloned. Several new genes were identified. One of them, called EWS, encodes a protein with an N-terminal domain which shares distant homology with the C-terminal domain of the large subunit of eukaryotic RNA polymerase II. The C-terminal half of EWS contains an RNA binding homologous domain. In the tumors where the 22q12 region is translocated, EWS forms a hybrid gene. In all cases, the encoded chimeric protein links the N terminal domain of EWS to the DNA binding domain of a transcription factors (FLI1, ERG or ETV1 that are all members of the ets family in ES; ATF1 in MMSP, WT1 in IDT; TEC a new orphan nuclear receptor in EMC). These new oncogenes are believed to act by directly interfering with transcription regulation. Another gene, was shown to be the site of germ line mutation in NF2 patients and of somatic mutation in schwannoma and meningioma, two tumors known to be predisposed by NF2. Analysis of the NF2 gene mutation spectrum shows that the mutations predominantly cause the synthesis of a truncated protein. Combined with the monitoring of the chromosome 22 allelic status in tumors, these data indicated that NF2 was a tumor suppressor gene which was inactivated by a two hit process during tumorigenesis. Its product, called schwannomin or merlin, is located at the membrane and homology analysis suggests that is may serve as linker between cytoskeleton and membrane proteins.

Jean Feunteun (Institut Gustave-Roussy, Villejuif) discussed genetic predisposition to breast and ovarian cancer. The lessons from genetic epidemiology have led to the concept that the distribution of breast and ovary cancer in the general population includes a small fraction of strongly determined hereditary cases within a vast majority of sporadic cases. 5 to 10% of breast cancers and ovary cancers may represent inherited cases that can be attributed to the inheritance of highly penetrant autosomal dominant susceptibility genes. After a brief review of the genes already identified, the talk focused on p53 and BRCA1 genes. Germline p53 mutations are associated with high risk of cancer for individuals belonging to multiple cancer syndrome families that include various childhood cancers and breast cancer in adult females. Germline mutations in the BRCA1 gene account for a large fraction of inherited predisposition to breast and ovarian cancers. The BRCA1 gene is subjected to loss of heterozygosity in most breast and ovarian tumors and therefore is likely to carry a tumor suppressor activity. However, unlike other members of the tumor suppressor family, it is rarely a target for somatic mutation in sporadic breast or ovarian cancers. Mutation carriers may develop either breast or ovarian cancers pointing to an allelic heterogeneity.

The meeting concluded with a general discussion on the new opportunities for cancer therapeutics. It was felt that there were a number of areas where recent advances could be beneficial in oncology:

  1. the detection of individuals predisposed to cancer;
  2. early detection of cancer;
  3. better diagnosis in defining the best management alternatives;
  4. better prognosis;
  5. better treatment.

Some frustration was expressed at the oncology community in general, because oncogene probes have been available for a dozen years or so, yet they are still not used systematically. Another discussion centered on the need to create a genetic database of tumors to identify which genes are mutated, amplified, or deleted in human cancer. In the database, it will be important to have tumor progression data to correlate with the genotype. It was felt that in the next few years it will be important to urge the major cancer centers and funding agencies to demand that cancer genotyping be carried out.

Contributions

  • Latifa Bakiri – c Jun induced apoptosis
  • Dirk Bohmann – Biochemistry and genetics of Jun-regulation
  • Sara Courtneidge – SRC family tyrosine kinases and the cell cycle
  • Jean Feunteun – Genetic predisposition to breast and ovarian cancer
  • Edward Harlow – The retinoblastoma protein
  • Peter Herrlich – Therapy attempts exploiting the molecular roles of a family of adhesion molecules
  • Peter Howley – The papillomaviruses and human cancer
  • Adi Kimchi – DAP genes, novel positive mediators of programmed cell death
  • Dominique Lallemand – The API complex during G0/G1 progression
  • David Lane – Regulation of p53 function
  • Arnold Levine – The loss of p53 functions in cancer
  • Frank McCormick – Novel cancer therapies based on human cancer genetics: small molecular and viral strategies
  • Moshe Oren – Regulation of the structure and function of wild type and mutant forms of the p53 tumor suppressor protein
  • Carol Prives – p53 and apoptosis
  • Gilles Thomas – Positional cloning of genes involved in human tumors
  • Thea Tlsty – Heterogeneity of cell cycle checkpoint control in human cells
  • Bruno Tocqué – Oncogene neutralizing binding decoys to control Ras dependent transformation and to induce tumor apoptosis
  • George Vande Woude – Mos and Met: genes for understanding transformation and metastasis
  • Moshe Yaniv – Cell cycle and transformation control of the snf/swi human complex