Signaling as an overall cellular process is a prerequisite for the life of normal and cancer cells and is regulated by a multiplicity of mechanisms including “cross-talks” among pathways. Cross-talk is often critical in determining the decision of a cell to go toward a specific biological direction, such as proliferation, apoptosis, or differentiation. Both signaling pathways and cross-talk among them involve such biochemical processes as protein-to-protein interactions and well-regulated phosphorylation and dephosphorylation processes. This symposium was focused on mechanisms of the regulation of signaling with special emphasis on regulatory cross-talk among pathways and on changes that occur in cancer cells. The phenomena discussed were also considered with reference to opportunities that may be offered toward the development of potential therapeutic intervention.

Specific topics considered during the symposium were:protein complex assemblies as targets for multiple signals; the role of multiple phosphorylations of target proteins; the cellular responses to extracellular signals and their coordination; the molecular processes underlying the choices cells make to determine their fate; and finally,opportunities for modification of some of the processes discussed toward developing therapeutically useful intervention.

Protein complex assembly was discussed with emphasis on the molecular aspects of the pathogenesis of VHL3 disease and the effects of the complexes formed by the VHL gene product. The coordination of the activities of Ras and small G proteins at the GEF and GAP levels in the formation of protein complexes and how these complexes involved integrated cross-talk between certain signaling pathways was a focus of discussion. The differences between the signaling protein Ras and the transport factor Ran and the ways by which a single switch mechanism can regulate diverse sets of protein-protein interactions and biological functions were also illustrated.

The role of Rb and of the HDAC family of chromatin remodeling enzymes in the regulation of the cell cycle was discussed also in relation to the Rb recruitment of members of the HDAC family and the Rb functional interactions with SWI/SNF. The anabolic processes required for a cell to duplicate itself were discussed with the emphasis on protein synthesis and particularly on the regulation of mRNA-encoding ribosomal proteins and the role of S6 kinase in this regulation. The regulation of the P53 tumor suppressor protein as affected by the MDM2-dependent P53 degradation and the mechanism of resistance of P53 to MDM2 action, the role of cross-talk through p14ARF in the stabilization of P53 activity, and the P53-dependent activation of NF-KB and its function in P53-induced apoptosis were all extensively discussed.

The impact of telomere dysfunction on intestinal carcinogenesis was assessed in the min-mouse and appropriate crosses, and was discussed in relation to the incidence of premalignant lesions and their progression to malignancy. The interactions between RAS,β-catenin, and p53 pathways in human colon cancer were discussed with particular emphasis on the induction of MDM2 and the activation of the p14ARF pathway by Ras and the functional contribution to cyclin D1 expression of theβ-catenin and Ras pathways. The role of small GTPases of the Ras and Rho families in transmitting signals from growth factor receptors to intracellular signaling pathways controlling cell proliferation was discussed also in relation to the activation of CDKs consequent to the action of the ERK/MAP kinase pathway. Other discussions concerned the activity of TGFβ through the formation of a receptor complex that activates Smad proteins, which, upon entry into the nucleus, in turn assemble complexes that regulate transcription relevant to the oncogenic process. The role of products of the human p16/INK4a/ARF locus in tumor suppression and replicative senescence was evaluated specifically through studies in fibroblasts uniquely deficient for p16INK4a.

Cross-talk in T-cells in response to extracellular signals was discussed with the focus on the molecular responses to phytohemagglutinin and interleukin 2 changes in triggering apoptotic mechanisms, and the role of antigen-presentation mechanisms as effecting apoptosis of T-cells during an immune response. The topic of signaling cross-talks operating in the regulation of α/β and γIFN, IFNs was discussed in relation to the functional regulation of IRFs and the STATs, and it was shown that signaling by IFN may depend on IFN α/β receptor components; the role of IRF3 in IFNα/β gene induction was emphasized. The abnormalities of HLA class I antigens in melanoma cells was discussed in relation to the mechanisms of escape of these cells from T cell-based immunotherapy and the related implication for optimal intervention in the treatment of human melanoma. Given the fact that NHL represents a neoplastic counterpart of mature B cells from GC, the role of the BCL-6proto-oncogene in modulating signals involved in GC formation was discussed, as was the regulation and mechanism of expression of this gene and its product. Regulatory mechanisms of the cell cycle,particularly those controlling the action of CDK2 and of the p53 and Rb pathways, were discussed as possible sites for therapeutic interventions. The potential need to target two or more key molecular sites to secure greater therapeutic effects in the face of redundancies in signaling pathways was discussed; examples were the development of inhibitors that block both EGF and Her-2 receptor tyrosine kinases and inhibitors of the HSP-90 molecular chaperone leading to simultaneous depletion of various oncogenic kinases.

The three examples discussed concerned the molecular pathogenesis of VHL, the GTPase exchange factor and the function of protein complexes,and the signaling through nuclear pores.

William G. Kaelin discussed the molecular pathogenesis of the VHL hereditary cancer syndrome. Patients with VHL disease have germ line mutations of the VHL tumor suppressor gene. Tumors develop when the remaining wild-type allele is inactivated. The VHL gene product (pVHL)forms complexes that contain elongin B, elongin C, Cul2, and Rbx1. These complexes resemble Skp1-Cdc53-F-box complexes and were thought to function as E3 ubiquitin ligases. A frequently mutated subdomain of pVHL, β domain, binds directly to a region of HIF implicated in oxygen-dependent protein turnover. pVHL indeed directs the ubiquitination of HIF, dependent on both intact β and α domains. The α domain binds to elongin C, which in turn recruits the other members of the complex. Cells lacking pVHL do not degrade HIF and accumulate high levels of hypoxia-inducible mRNAs, such as the mRNAs encoding VEGF. The overproduction of angiogenic peptides such as VEGF likely contributes to the vascular nature of VHL-associated neoplasms. In cooperation with P. Dervan, polyamides are synthesized that would bind to HIF sites and ultimately decrease the production of VEGF in the brain.

Michael Moran discussed GTPase exchange factor signaling and the function of related protein complexes. Ras is a membrane associated GTPase. Ras can have two different conformations as a function of its interactions with GDP and GTP, and in each it is able to associate physically with different sets of cellular proteins. Ras functions normally as an essential molecular switch downstream of a variety of extracellular signals, but it is not a direct target for these signals. Distinct proteins interact directly with Ras to catalyze the exchange of nucleotides and to stimulate its GTPase activity. These GEFs and GAPs appear to be the direct targets of the intracellular signals generated in cells in response to extracellular stimuli. The various Ras-specific GAPs and GEFs are multidomain proteins. The activities of Ras and other small G proteins are coordinated at the level of their GEFs and GAPs, which, in some instances, reside in the same or directly interacting proteins. Considerable cross-talk between signaling pathways may occur through protein complexes that include GEFs and GAPs. The concerted activities of GEFs and GAPs ensure the transient nature of small G protein activation. Other forms of regulation that occur at the level of the Ras GEFs involve protein and lipid phosphorylation and dephosphorylation, proteolytic destruction, and protein complex formation. GRF2 becomes phosphorylated in response to calcium signals. Signals that cause GRF2-mediated activation of Ras,Rac, and MAP kinase pathways also induce the phosphorylation of GRF2,its modification by ubiquitin polymers, and proteosome-dependent destruction.4 The function of Ras GEFs in cell regulation is being addressed; the results suggest a role for GRF2 in the regulation of MAP kinase pathways and in cell cycle regulation. Whereas calmodulin associates tightly with GRF2 in a calcium-dependent manner, this interaction is not required for the GRF2-mediated activation of Ras in response to calcium signals. Activation of Ras by GRF2 is not sufficient for activation of the Raf-Mek-ERK kinase cascade. The emerging model is that GRF2 directs the assembly of protein complexes that coordinate the interaction of Ras and Rac with downstream effectors. In conclusion, Ras GEFs convert second messenger-type signals into protein complexes. These complexes directly regulate Ras, couple Ras with effectors, mediate cross-talk between GTPases, and trigger negative feed-back pathways. Through these diverse interactions, the GEFs are pivotal in the conversion of extracellular signals into cellular responses.

Alfred Wittinghofer discussed signaling across the nuclear pore. Constitutive transport across the pore involves proteins, RNAs, and protein-RNA complexes. A large amount of cellular signaling involves regulated nuclear transport, usually of components of the transcriptional apparatus. Constitutive and regulated transport involves a complex machinery that is regulated in a major part by Ran. The GEF for Ran, RCC1, is exclusively localized in the nucleus, whereas the RanGAP is only found in the cytoplasm. This separation of activities creates a gradient of RanGTP across the nuclear pore that is believed to be the major driving force for the system. Ran in the GTP-bound form interacts tightly with two types of proteins that are Ran effectors: (a) proteins with a Ran-binding domain(RanBD) such as RanBP1 and RanBP2; (b) a family of transport receptors such as exportin and importin with a less-defined Ran-binding motif. The latter are directly involved in binding the import or export target proteins; RanBP1 and RanBP2 interactions with RanGTP help dissociate the transport receptor complexes. The differences between the signaling protein Ras and the transport factor Ran show how a common switch mechanism can be diverted to regulate a diverse set of protein-protein interactions and biological functions.

The specific topics concerned the role of Rb and chromatin remodeling enzymes in cell cycle regulation, the relationship of S6 kinase and ribosome biogenesis to cell growth, and the regulation and function of the P53 tumor suppressor protein. In addition, the increased initiation and decreased progression of carcinomas by telomere dysfunction were also indicated and related to p53 function.

Doug Dean discussed the role of Rb and chromatin remodeling in cell cycle regulation. Rb is functionally inactivated by constitutive hyperphosphorylation in most tumors that do not have mutations in the retinoblastoma gene. Rb binds to the E2F family of cell cycle transcription factors and represses the transcription of cell cycle control genes. Moreover, the Rb-E2F complex that forms at the promoter of cell cycle genes actively represses transcription, and this active repression seems to be important for Rb growth suppression. This repression is dependent upon the ability of Rb to recruit members of the HDAC family of chromatin remodeling enzymes. Evidence was presented that interaction with a second ATPase-dependent chromatin remodeling complex known as SWI/SNF (BRG1) is also important for Rb function. In the C33 cell line, a complex of pRb, HDAC, BRG1, and E2F suppresses cyclin E; cyclin D represses the action of the complex; the hypophosphorylated pRB/BRG1/E2F complex inhibits cyclin A-cdc 2 and is disrupted by cyclin E-dependent hyperphosphorylations of pRb with consequent liberation of E2F. Progressive phosphorylation leads to distinct Rb-chromatin remodeling complexes that appear to regulate the exit of cells from the G1 and S phases of the cell cycle. In contrast to its role with Rb, the SWI/SNF complex has been shown to be associated with the transcriptional activation of other genes.

George Thomas discussed the relationship of S6 kinase and ribosome biogenesis to cell growth. The distinction was considered between growth or increase in cell mass and proliferation or increase in cell number. A cell must first grow in mass before it proliferates. Little is known concerning the molecular mechanisms that control cell growth;they include all of the anabolic processes required for a cell to duplicate itself. The major products being synthesized are components of the translational machinery. A proliferating cell expends 80% of its energy on the biosynthesis of protein synthetic components, most notably ribosomes. This energy is also used in the transport,processing, and assembly of mature ribosomal subunits. The mRNAs encoding ribosomal proteins are controlled at the translational level through S6 kinase, presumably mediated by increased 40S ribosomal protein S6 phosphorylation. Studies in both Drosophilia and in the mouse addressed this issue, which have led to two novel observations concerning the control of cell growth: Drosophilia deficient in the S6 kinase gene(dS6K) exhibit delay in development and a reduction in body size and have smaller cells rather than fewer cells. Thus, the dS6K gene product regulates cell size in a cell-autonomous manner without impinging on cell number. Regulatory mechanisms sensing the state of the translational machinery have been described in bacteria, but none have been reported in higher eukaryotes. In adult fasted mice, livers deficient in the S6 gene grew in response to nutrients, roughly doubling in size, although nascent 40S ribosome biogenesis was abolished. These liver cells failed to complete one to the one-and-a-half rounds of proliferation required for regeneration after partial hepatectomy. Active cyclin D/cdk4 complexes were formed, but the induction of cyclin E and active cyclin/cdk2 complexes was blocked. These results imply that abrogation of 4OS ribosome biogenesis is sensed by a checkpoint that prevents cell cycle progression.

Karen Vousden discussed the regulation and function of the p53 tumor suppressor protein. p53 induces cell cycle arrest and apoptosis in cells exposed to various forms of stress. Under normal circumstances,p53 is a short-lived protein that is maintained at very low levels. Activation of a p53 response is mediated by stabilization of the protein and accumulation within the cell. The cellular protein MDM2 is a regulator of p53 stability. MDM2 functions as a ubiquitin ligase for both p53 and itself. The RING finger of MDM2 is necessary for the ubiquitin ligase function, and mutations in this domain stabilize MDM2 and abrogate its ability to degrade p53. The extreme COOH-terminus of p53 is also important for degradation; p73, a p53-related protein, is not degraded after interaction with MDM2. P53 can be activated by diverse signals, and stabilization may occur through resistance to degradation by MDM2. Phosphorylation of p53 can contribute to stabilization under some circumstances; inhibition of MDM2 expression,cytoplasmic localization of p53, and separation of p53 and MDM2 to distinct nuclear locations also play a role. The p14ARF protein can inhibit MDM2-targeted degradation of p53. Deregulated E2F1 expression activates p14ARF, indicating that p14ARF participates in the activation of p53 in response to abnormal proliferative signals driven by oncogene activation. A peptide derived from the NH2terminus of p14ARF can bind MDM2, relocalize MDM2 to the nucleolus, and inhibit the degradation of p53. A cryptic nucleolar localization signal in the COOH-terminus of MDM2 does not function in unstressed cells but is necessary for p14ARF-dependent relocalization. An MDM2 mutant lacking the p14ARF-binding region is predominantly localized to the nucleolus, even in the absence of p14ARF expression. p53 can cause activation of the transcription factor NF-κB, and NF-κB function is required for p53-mediated apoptosis but not for growth arrest. The mechanism of activation of NF-κB by p53 is distinct from that mediated by TNF and involves MEK1 as well as the IkB systems and the activation of pp90rsk. An inhibitor of MEK1 blocked activation of NF-κB by p53 and completely abrogated p53-induced cell death. Inhibition of NF-κB can potentiate apoptosis induced by chemotherapeutic agents in the absence of wild-type p53. However, in tumors that retain wild type p53,inhibition of NF-κB may diminish rather than augment the therapeutic response.

Ronald DePinho discussed the increased initiation and the decreased progression of epithelial carcinomas consequent to telomere dysfunction. In human cells, insufficient levels of telomerase leads to telomere attrition with passage in culture and possibly with aging and tumorigenesis in vivo. Critical reduction in telomere length is not observed in the mouse because of promiscuous telomerase expression and long telomeres. However, telomere attrition in aging telomerase-deficient p53 mutant mice promotes the development of epithelial cancers. In these carcinomas, telomere dysfunction generates a fusion bridge-breakage process that leads to the formation of complex nonreciprocal translocations. Whereas initiation of these cancers was increased, the progression toward advanced stages was attenuated. Using the min-mouse, the impact of telomere dysfunction on intestinal carcinogenesis was assessed by crossing mice mutant for mTR and APC(APCMIN carries a nonsense mutation). Increased genomic instability in mice with short telomeres correlated with an elevated rate of intestinal adenoma progenitor lesions. The incidence of macroscopically visible polyps reached a maximum in mice with moderate telomere dysfunction, whereas severe telomere dysfunction correlated to a decrease in the number of adenomas because of growth inhibition. High rates of mitotic abnormalities, apoptosis, and activated DNA-damage responses correlated with impaired growth of adenomas in late generations of mTR-1 mice. The results imply that telomere shortening can increase the initiation of premalignant lesions in epithelial cells, but that in the presence of intact DNA-damage checkpoints, the progression of these lesions is impaired by critical telomere shortening.

The mechanisms involved in these choices were discussed from four vantage areas, namely, the interactions between Ras, β-catenin, and p53 pathways in human cancer; oncogenic signaling through small GTPases; the relationship of TGFβ to oncogenes; and the products of the p16/INK4a/ARFlocus in tumor suppression and replicative senesence.

Frank McCormick discussed interactions between Ras, β-catenin, and p53 pathways in human cancer. Activation of Ras and β-catenin pathways occurs frequently in many types of cancer. Inactivation of the p53 pathway is also a frequent event. All three pathways are altered in the majority of metastatic colon cancers. Ras activates the Raf-MAP kinase cascade and the PI3 kinase pathway. Cyclin D1 is a transcriptional target of the MAP kinase pathway, and the PI3 kinase affects cyclin D1 stability. Cyclin D1 is also a transcriptional target of the β-catenin pathway in colon cancer cells. Inhibition of the MAP kinase pathway leads to G1 arrest in colon cancer cells; this results from the loss of cyclin D1 and cdk4 and the subsequent redistribution of p27 to cdk2. Inhibition of cdk4 seems more important, because high levels of p27 do not cause growth arrest in these cells despite inhibition of cdk2 activity. High levels of cdk4 activity may therefore make colon cancer cells relatively independent of cdk2.

MDM2 is a major target of the Raf MAP kinase pathway. Induction of mdm2 transcription by activated Raf is p53-independent and occurs through Ets and Ap-1 sites in the P2 promoter. In tumor cells that retain wild type p53, Ras may contribute to suppression of p53 activity through induction of MDM2. The Ras pathway also activates the p14ARF pathway. The Ras seems to induce expression of MDM2 early in G1 after the stimulation of receptor tyrosine kinases. This MDM2 keeps p53 at a basal level and allows progression to S phase. E2F is activated partly through ras-dependent transcription of cyclin D1 and turns on expression of p14ARF. This protein can inhibit MDM2 and allow accumulation of p53. P53 accumulation in response to DNA damage is determined by Ras activity and p14ARFexpression.

C. J. Marshall reported that small GTPases of the Ras and Rho families are involved in transmitting signals from growth factor receptors to intracellular pathways controlling cell proliferation. Signals from Ras and Rho GTPases seem to interact. Some transmembrane receptors have to activate both Ras and Rho to get ERK MAP kinase activation. A second point of interaction is the expression of the cyclin dependent kinase inhibitor p21Wafl. Swiss 3T3 cells transformed by oncogenic Ras have elevated levels of Rho GTP. When Rho is blocked in these cells, p21Wafl is induced and the cells arrest. Thus, Rho activity suppresses a signaling pathway of Ras that is required for p21 expression. These cells fail to respond to the Rho signal that leads to stress fibers arguing that, in turn, Ras suppresses a signal from Rho. Using cells in which components of cell cycle control have been inactivated by homologous recombination, it was shown that the loss of the tumor suppressor pRb reduces the requirement for the ERK MAP kinase pathway for cell cycle reentry. This suggests that the ERK MAP kinase pathway plays a major role in the activation of the cyclin D-dependent kinases that phosphorylate and inactivate pRb105.

Joan Massague outlined the relationship of TGF-β to oncogenesis. The TGF-β family of secretory polypeptides is a major source of signals which have many different effects depending on the type and state of the cell. A simple system has been elucidated recently that mediates many diverse TGF-β responses and involves a family of membrane receptor protein kinases and a family of receptor substrates (the Smad proteins) that enter into the nucleus where they act as transcription factors. The ligand TGF-β assembles a receptor complex that activates Smads, and the Smads assemble multisubunit complexes that regulate transcription. An upcoming Smad complex is met in the nucleus by a set of partner proteins that are specific to a particular cell type in a particular set of conditions. These partners determine the DNA sequences that the Smad complex will bind, the transcriptional coactivator or corepressors it will recruit, the other transcription factors it will cooperate with, and how long all this will last. The mix of Smad partners and regulators present in a given cell at the time of TGF-β stimulation thus decides the outcome of the response and defines the “cellular context.” The TGF-β pathway plays a dual role in tumorigenesis. The ability of TGF-β to inhibit epithelial cell proliferation is diminished or lost in many carcinoma cells. Mechanisms of cell cycle arrest by TGF-β include the inhibition of myc, CDC25A, and CDK4 by P15 and CDK2 by p27 and result in the arrest of G1 progression to S phase. These mechanisms are inactivated by loss-of-function mutations in the TGF-βpathway in colon cancer or by an inhibitory effect of a hyperactive Ras pathway in breast cancer. TGF-β can also exacerbate the malignant phenotype at later stages of tumorgenesis by fostering tumor invasion and metastasis. In breast cancer cells that retain TGF-β signaling components, but which have lost the antimitogenic response, as is the case in cells with a hyperactive Erb-2/EGFR-Ras pathway, TGF-βsignaling has been reprogrammed, leading to gene responses that support the invasive or metastatic behavior of the cell.

Gordon Peters discussed the role of the products of the human INK4α/ARF locus in tumor suppression and replicative senescence. The INK4α/ARF locus on human chromosome 9p21 encodes two structurally distinct proteins designated p16INK4a and p14ARF (p19ARF in the mouse), by exploiting different first exons (1a and 1b) spliced to a common second exon that is translated in alternative reading frames. P16INK4a functions as an inhibitor of Cdk activity by binding directly to Cdk4 and Cdk6, which initiate the phosphorylation and functional inactivation of pRb. Ectopic expression of p16INK4a causes cells to arrest in the G1 phase of the cell cycle in a pRb-dependent manner. P14ARF interacts directly with MDM2. Ectopic expression of ARF, therefore, stabilizes p53 and causes cells to arrest in G1 and G2,accompanied by increased expression of the p53 regulated genes p21 and MDM2. ARF is predominantly a nucleolar protein; its expression is negatively regulated both by p53 and by pRb,the latter through its ability to repress E2F-dependent transcription. Most of the data from human tumors favor p16INK4aas the primary target for inactivating lesions, which range from simple missense mutations to homozygous deletion of the entire locus, and transcriptional silencing promoter methylation. P16INK4a accumulates as cells undergo replicative senescence. It is also induced, along with ARF, in response to oncogene signals via the Ras-Raf-MEK pathway. In fibroblasts from a member of a melanoma-prone family who is homozygous for an intragenic deletion in p16/INK4a, the loss of 19 bp and the resulting frameshifts in exon 2 produce two p16-related fusion proteins, neither of which binds detectably to Cdk4 or Cdk6. One of the products retains the known functions of p14ARF, implying that these fibroblasts are uniquely deficient for p16INK4a.These cells have a finite life span, and their life span can be extended by viral oncoproteins that interfere with p53. Ectopic expression of telomerase and H-Ras results in the outgrowth of morphologically transformed, anchorage-independent cells that retain functional p53. Thus the INK4a/ARF locus seems to operate as a mechanism of defense against oncogenic stimuli that is independent of telomere attrition.

The specific discussions concerned cross-talks in T cells, cross-talks in the regulation of two IFN systems, and HLA class I antigen abnormalities in melanoma cells and their relationships to escape mechanisms and T cell-based immunotherapy.

Ingo Schmitz discussed cross-talks in T cell CD95, which is a member of the TNF receptor superfamily and induces apoptosis upon receptor oligomerization. T cells switch their sensitivity toward CD95-mediated apoptosis during an immune response. Freshly isolated resting peripheral human T cells express low levels of CD95 and are resistant to CD95-induced apoptosis. After activation with phytohemagglutinin, these T cells (d1 T cells) up-regulate CD95,yet are still resistant. After prolonged culture in interleukin 2 these cells become sensitive to CD95-mediated apoptosis. This process represents the downphase of the immune response. Only a few T cells survive as memory cells. The molecules MORT1/FADD and capase-8 assemble in DISC. DISC formation was highly reduced in d1 T cells, which resemble the CD95 type II phenotype. In addition, d1 T cells express considerable amounts of c-FLIPSHORTand Bcl-x, two antiapoptotic proteins. Resistance to CD95-mediated apoptosis of d1 T cells can be explained by the formation of low amounts of DISC and depends on mitochondrial function blocked by Bcl-x. Efficient activation of antigen-specific T cells requires costimulatory signals. Reexposure to antigen and CD28 costimulation reduces activationinduced cell death and increases the number of T cells with effector functions. Restimulation via TCR/CD3 protects T cells from CD95-mediated apoptosis. Activation-induced cell death is mediated predominantly by CD95 and its cognate ligand, CD95L. Costimulation prevents CD95L expression and reduces the activity of the CD95 death-inducing signaling complex and procaspase-8 activation. In parallel, costimulation strongly increases expression of the c-FLIPSHORT and Bcl-x.

Tadatsugu Taniguchi showed the role of signaling cross-talk in the regulation of two IFN systems. Two classes of cytokines, IFN-α/βand IFN-γ play central roles in the innate immune response against viral infections. In addition, IFN-γ is known to play a central role in the adaptive immune response mediated by inflammatory CD4 T cells. Two families of transcription factors are critically involved in the IFN response: the IRFs, and the STATS. IRF-1 and IRF-2 were originally discovered as regulators of the IFN system, and seven additional members have been identified. IFN-α/β and IFN-γ are known to transmit signals via distinct receptor complexes (IFNAR and IFNGR), and it is not known whether IFNAR and IFNGR share any functional aspects in the processes. These receptor molecules are clustered in caveola of the cell membrane. A novel form of signaling cross-talk was found in which signaling by IFN-γ depends on IFN-α/β receptor components. The IFN-α/β receptor component, IFNAR1, provides a niche for the efficient assembly of the IFN-γ-activated transcription factors,STAT1, STAT2, and IRF-9 (ISGF3g). This cross-talk is contingent on a physical association between IFNAR1 and IFNGR2.

IRF-3 undergoes phosphorylation and translocation from cytoplasm to nucleus upon viral infection. The activated IRF-3 binds to the IFN-βgene promoter, together with coactivators CDP/p300. Gene disruption study revealed the essential role of the IRF 3 in both early and late phases of IFN-α/β gene induction. The early phase is largely dependent on IRF-3. The residual induction is dependent on the IFN-α/β signaling pathway and the expression of IRF-7. IRF-3 cooperates in the late phase, with IRF-7 for full procurement of the normal mRNA induction profile for the IFN-α gene family.

Soldano Ferrone discussed HLA class I antigen abnormalities in melanoma cells and their relationships to tumor escape mechanisms and T cell-based immunotherapy. Molecular mechanisms found to underlie abnormal HLA class I phenotypes in the malignant transformation of human melanocytes include: (a) mutations that inhibit translation of β2m genes in most cases and transcription of β2m genes only rarely;(b) loss of one copy of chromosome 6, which carries the MHC region; and (c) defects in the transcription of HLA class I genes and point mutations in HLA class I heavy chains. HLA class I antigen down-regulation is associated with poor prognosis in all types of melanoma except ocular melanoma. In some cell lines, the loss of HLA-I can be counteracted by IFN-γ and other cytokines. Injection of malignant cells with tumor-associated, antigen-specific, HLA class I antigen-restricted CTL into severe combined immunodeficient mice has resulted in the expansion of malignant cell populations with HLA class I antigen loss or down-regulation. If these findings in an animal model system reflect what happens in patients, the success of T cell-based immunotherapy might be counteracted in most patients by the selection of melanoma cells with abnormal HLA class I phenotypes.

These were discussed with reference to cross-talks between signaling pathways in B cell development and lymphomagenesis; the cell cycle and novel cancer therapeutics; and overcoming, with therapeutic agents,cross-talks between signaling in tumors.

Riccardo Dalla-Favera discussed the first of the above topics. NHL represents the transformed counterparts of mature B cells from GC within peripheral lymphoid organs. One key regulator of GC formation is the BCL-6 proto-oncogene, originally identified by virtue of its involvement in chromosomal translocations associated with B cells NHL. BCL-6 acts by modulating various signals that are involved in GC formation. BCL-6 expression is down-regulated at the transcriptional level by CD40 and also by ubiquitination and degradation after phosphorylation via immunoglobulin receptor-dependent MAPK activation. BCL-6 modulates STAT-6-dependent IL-4 signaling via direct binding to the STAT-6 DNA binding site and NFkB signaling when binding to DNA adjacent to NFkB binding sites. The function of BCL-6 as a transcriptional repressor is regulated through direct acetylation by p300.

BCL-6 expression is down-regulated in cells exiting the GC and differentiating into memory cells and plasma cells. In ∼40% of diffuse large cell lymphoma and ∼10% of follicular lymphoma,chromosomal translocations prevent down-regulation and lead to BCL-6 constitutive expression. All GC-derived B-NHL expresses BCL-6. On the basis of recent evidence that BCL-6 expression prevents apoptosis in B cells, BCL-6 provides attractive targets for attacking B-NHL.

Ali Fattaey discussed the cell cycle as related to the development of novel cancer therapeutics. In eukaryotic cells, cell division cycle events or transitions are accomplished through the timely activation and inactivation of the CDKs, which in turn phosphorylate their appropriate substrates at various phases of the cycle. These cells must delay mitosis until the complete duplication of the genome has occurred. Mammalian cells treated with DNA-damaging agents pause at either of two checkpoints in the cell cycle to allow for DNA repair:the G1-S transition in part by p53-mediated inhibition of Cdk2 kinase, or the G2-M transition mediated by regulated inactivation of Cdc2 kinase. Human cancer cells often lack G1 control through the loss of p53 and RB. This difference between normal and human cancer cell cycles may be exploited to develop novel therapies for the treatment of human cancers. For example, inhibition of the enzymes involved in the inactivation of Cdc2 ought to abrogate the G2checkpoint, allowing cells to progress into the M phase in the presence of damaged DNA with catastrophic consequences. Inhibitors of CDK are represented by olomucin derivatives, propanolol, or flavopyridol; 2.3 d-pyridopyridines inhibit cyclin D/cdk 4 specifically. Another new agent, PD183812, is a specific ATP competitive inhibitor of cyclin D/cdk4 and this action results in G1 arrest consequent to loss of Rb phosphorylation, as also reflected by the exclusive growth inhibition of Rb +/+ cells. After DNA damage by agents such as etoposide, cis-platinum, 1,3-bis(2-chloroethyl)-1-nitrosourea,doxorubicin or UV or X-ray, cells arrest in G2;disregulation of Cdc2 kinase at G2-M would be required to achieve the catastrophic progression into M alluded to above.

Paul Workman discussed overcoming cross-talk-related signaling in tumors with therapeutic agents. Initial reservations about the new sets of molecular targets for cancer therapy centered around such issues as:(a) the difficulty in designing agents with sufficient target selectivity; (b) action at any one molecular target being insufficient for strong anticancer activity; (c) the likelihood that significant side effects would be seen; and(d) resistance attributable to redundancies and cross-talks. Many of these early concerns have receded as emerging agents proved to exhibit selective molecular and cellular effects and to display activity in animal models and then in patients. Examples include the anti-Her-2 antibody herceptin, the EGF receptor tyrosine kinase inhibitor Iressa and related agents, the bcr-abl inhibitor STI571, and various inhibitors of VEGF receptor tyrosine kinase and of ras farnesyl transferase. Preeminent is the potential need to target two or more key molecular abnormalities to obtain greater clinical activity. One obvious way to overcome cross-talk-related signaling is to target drugs at loci that are relatively downstream in key oncogenic pathways. For example, CDK inhibitors have the potential to be more effective than agents active at upstream loci, but this may also result in greater side effects. Inhibitors of histone acetylases and deacetylases may also preempty redundancy by directly affecting gene expression. A second way to overcome redundancy is to use selective inhibitors in appropriate combinations or by use of a single drug molecule that simultaneously targets more than one site. Examples of this are kinase inhibitors that block both EGF and Her-2 receptor tyrosine kinases, or SU6668, which in addition to blocking VEGF receptor tyrosine kinase also inhibits FGF and PDGF receptor tyrosine kinases. Another means to overcome signaling redundancy is by inhibiting the ATPase function. Inhibition of hsp 90, for example by the geldanamycin analogue 17AAG, leads to the simultaneous depletion of various oncogenic kinases: both cytostasis and apoptosis can be induced in different tumor cells.

In addition to the oral presentations summarized above, several posters were part of the meeting, as indicated below. Galina Selivanova showed that a synthetic peptide (peptide 46), derived from the p53 COOH-terminal domain (spanning residues 361–382) can restore the growth suppressor function of at least two hot spot p53 mutants,R27[3H] and R248O. Examination of the effect of peptide 46 on different hot spot p53 mutants in extracts from human tumor cells revealed that p53 mutants could be reactivated for specific DNA binding. Peptide 46 also induced growth suppression in tumor cells of a different origin. Peptide 46 seems to bind preferentially to the core domain of mutant p53. Selective targeting of tumors expressing mutant p53 could be achieved while leaving normal cells with wild-type p53 unaffected.

Silvia Miotti reported that FR forms complex with cytoplasmic signaling molecules in ovarian cancer cells. FR is a glycosylphosphatidylinositol-anchored protein that provides a high affinity route for folate internalization. This protein becomes selectively overexpressed in human ovarian carcinomas of nonmucinous origin. Using the ovary carcinoma cell line IGROV1, molecules that might physically and functionally associate with FR were examined. FR is present in low density microdomains together with subunits of heterotrimeric G proteins and the src-family nonreceptor tyrosine kinase lyn. Complexes of FR, lyn, Gai-3 and Gb coimmunoprecipitate; in vitro kinase assay of the immunoprecipitates revealed stimulation of the phosphorylation of lyn and the Gai-3 subunit. Transfection of IGROV1 cells with caveolin 1 resulted in reduced FR expression and decreased phophorylation of lyn and the other molecules present in FR-immunoprecipitates together with a partial reversion of the transformed phenotype.

Anna Rapallo showed that Ki-ras activation in vitro affects G1 and G2-M cell cycle times, inhibits apoptosis, and induces DNA aneuplody. The role of Ki-ras activation (a GÆC transversion in codon 12 with substitution of arginine for glycine) was investigated using control and permanently transfected NIH3T3 mouse fibroblasts. Apoptosis was induced by starvation and by the nonsteroidal anti-inflammatory drug sulindac sulfide. Codon 12 GÆC Ki-ras-transfected cells showed a prolongation of the G1 cell cycle phase of ∼50%, a reduction of the G2-M transit time of 30%, and a decrease of the cell loss factor of∼90%. Apoptotic cells induced by starvation were about 10%in control and <1% in Ki-ras transfected cells. Sulindac sulfide also failed to induce apoptosis in the transfected cells,whereas it was highly apoptotic in the control cells.

Giovanni Martinelli showed that the novel type of chimeric bcr-abl transcript e8-int-a2 found in a philadelphia chromosome-positive chronic myeloid leukemia patient retains the dbl-and cdc24-like domains but not the ph-like domain of the bcr gene. The in-frame e8-int-a2 BCR-ABL transcript, translated into a Mr 197,500 BCR-ABL protein of 1804 amino acid residues named P200 BCR-ABL. This novel BCR-ABL transcript retains the DBL homology and the recently recognized CDC24 homology domains, but not the pleckstrin homology.

Daniel D’Orazio discussed three patients with FAP, in whom correlations between genotype, phenotype, and the involvement of the Wnt pathway were examined. In contrast to the classic truncating APC-positive polyps, those derived from the APC-negative FAP patients maintained the APC/β-catenin protein interaction. The truncating APC-positive polyps displayed elevated c-myc mRNA levels,whereas levels were not increased in the missense APC-positive, and the APC-negative polyps were not increased. In the APC-negative FAP, there was no activation of the Wnt pathway. The missense APC-positive FAP revealed a more complex situation with the indication of Wnt pathway activation but no c-myc overexpression.

Caterina La Porta showed that the overexpression of nPKCδ in BL6 murine melanoma cells was not associated with changes in the level of mRNA of TGFβ1 but with increases of the release of this cytokine into the plasma of BL6T metastasised animals. TGFβs are secreted from cells as latent complexes containing the cytokine and its propeptide,LAP. Activated TGFβ has a very short life in plasma, whereas the half-life of the latent form is significantly longer.

Tsukasa Shibuel outlined the characterization of Noxa, a new target gene of p53 encoding a BH3-only member of the Bcl-2 family. p53 and IRF-1 cooperate in tumor suppression. Using a mRNA differential display method, target genes of IRF-1, p53, or both were found to show increased transcription in response to DNA damage. A gene termed Noxa was identified; its cDNA encodes a 103-amino acid protein containing two Bcl-2 homology (BH3)motifs of the Bcl-2 family of proteins. When ectopically expressed, Noxa underwent BH3 motif-dependent localization to mitochondria, and induced cytochrome c release and caspase-9 activation. Both murine and human Noxa genes contain one putative p53-recognition sequence, and these promoters are activated by the coexpressed p53. Ectopic expression of p53 in human Saos2 cells resulted in the increased level of hNoxa mRNA, and this overexpression brought about apoptosis. These results collectively suggest a potential role of Noxa in p53-induced apoptosis.

G. Sava outlined the effects of the Ruthenium complex NAMI-A on cell cycle and on tumor cell invasion and host interactions. NAMI-A is active against lung metastasis of solid tumors. It caused increases of cells in G2-M, depending on the length of drug exposure; these effects were completely reversed by 48 h after treatment. NAMI-A significantly reduced the laminin-induced tumor cell transition from the low- to the high-proliferating phenotype and increased the number of cells expressing CD54 and E-cadherin adhesion molecules. It inhibited serum-stimulated DNA synthesis along with the expression of PCNA and cdc2 in parallel to increased caspase-3 activity and DNA fragmentation. NAMI-A was able to down-regulate MAPK/ERK activity both in serum- and in phorbol ester-stimulated cells. A caspase-3 inhibitor prevented the appearance of DNA laddering induced by NAMI-A or by a selective ERK-inhibitor in both serum- and phorbol ester-stimulated cells. A reduction of cell proliferation was observed after 144 h of exposure of tumor cells to NAMI-A, whereas shorter treatments up to 72 h were inactive. In cocultures of metGM cells and lymphocytes from CBA mice, NAMI-A increased by 64% the lymphocytes in the S phase, and lymphocytes of NAMI-A-treated samples were found to be closely bound to metGM cells. NAMI-A had no toxicity on in vitro primary cultures of either B or T lymphocytes.

K. Lang indicated the involvement of protein kinases and calcium in the regulation of T24 bladder carcinoma cell migration within a three-dimensional collagen lattice. The migratory activity of these T24 cells was examined using time-lapse videomicroscopy and compared with the results obtained with six colon-carcinoma cell lines. The spontaneous migratory activity of the colon carcinoma cell lines varied from <10% locomoting cells to 20% locomoting cells. Migration was inducible by PKC activating phorbol esters to 80% locomoting cells and by the catecholamines epinephrine and norepinephrine to 40%and 60% locomoting cells. Spontaneous locomotory activity of cells of the bladder carcinoma cell line T24 consistently reached 60%locomoting cells, and phorbol esters did not elevate locomotory activity further. The spontaneous locomotion of T24 bladder cells was PKC- and protein tyrosine kinase-dependent. Both types of kinases are present in focal adhesions and are involved in the dynamic reorganization of the actin-cytoskeleton during migration. Protein tyrosine kinases and PKC act directly on actin-associated proteins and influence the calcium cycling within the cell. Thapsigargin inhibits uptake of cytosolic calcium into the endoplasmic reticulum and ionomycin increases the calcium influx from the extracellular environment. Both agents led to an increase in cytosolic calcium and reduced the migratory activity of the T24 bladder carcinoma cells. Norepinephrine and epinephrine also reduced locomotory activity.

A. Migliaccio showed that steroid hormones modulate DNA synthesis by nongenomic action. Estradiol activates the src/ras/MAP-kinase pathway in MCF-7 breast cancer cells. This activation is inhibited by antiestrogens. Estradiol stimulates c-src and MAP-kinase activity in Cos-7 cells made hormone-responsive by transfection with the estrogen receptor-a (ER-a) cDNA. Progestins also stimulate the same pathway in T47D mammary cancer cells. Progestins, in addition to the progesterone receptor (PR) seem to require the presence of ligand-free ER to stimulate the src/ras/MAP-kinase pathway. In human prostate cancer cells, both androgen and estrogen (β form)receptors, in the presence of cognate ligands, interact directly with the c-src-inducing formation of a ternary complex. Microinjection of MCF-7 and T47D cells with cDNA of catalytically inactive src or anti-ras antibody proves that src and ras are required for estadiol and progestin-dependent progression of cells through the cycle. Steroid stimulation of target cells induces Ras-Raf-1 association, connecting active Ras with downstream serine/threonine kinases; geldanamycin disrupts this association and prevents both Raf-1 activation and steroid-induced DNA synthesis. The selective MEK 1 inhibitor, PD 98059,inhibits estradiol, progestin, and androgen stimulation of Erk-2 and abolishes the steroid-dependent S-phase entry. Fibroblasts transfected with transcriptionally inactive mutant receptors are stimulated to enter the S phase upon estradiol treatment at the same extent as wild-type receptor transfected cells.

Jadwiga Chroboczek described the construction and use of delivery vectors for anti-cancer treatment. Two nonviral delivery agents were developed. The synthetic adenovirus subviral particles, dodecahedra,which, like the native virus, enter cells efficiently via the endocytotic pathway and accumulates at the nuclear membrane and a peptidic vector comprising the 20 N-terminal amino acids from the adenovirus fiber protein which very efficiently enters the nucleus. These delivery agents can bring into cancer cell toxins such as saporin or certain anticancer agents.

At the end of the symposium, Ed Harlow offered some concluding thoughts. Already at this time, more than 10 signaling pathways have been discovered, and it is logical to predict that there must be a measure of integration among them. Measurements of integrating cross-talks should best be made proximal to target molecules without diffusion of effects. One should guard against problems related to experimental overexpression of genes/protein product: every protein has an intracellular “home,” and overexpression of proteins may push them to other-than-normal sites and cause artifacts based on inappropriate physical interactions. In studies of cross-talk, one should compare different cell types before making generalizations, and one should use very complex readouts to spread information on how different actions interact in a cell. It is likely that information from multiple laboratories will need to be combined and integrated to achieve reasonably rapid progress in this area. Finally, much better inhibitors are needed to be used as probes in the evaluation of critical cross-talk.

The program committee consisted of the cochairs, Dr. Riccardo Dalla Favera (Columbia University, New York, NY), Dr. Giulio Draetta(European Institute of Oncology, Milan, Italy), David Livingston(Dana-Farber Cancer Center, Boston, MA), and Frank McCormick(University of California, San Francisco, CA). In addition to the program committee members, invited participants included: John Cleveland (St. Jude Children’s Research Hospital, Memphis, TN), Doug Dean (Washington University School of Medicine, St. Louis, MO), Ronald De Pinho (Dana-Farber Cancer Institute, Boston, MA), Gerald Evan(University of California, San Francisco, CA), Pier Paolo Di Fiore(European Institute of Oncology, Milan, Italy), Ali Fattaey (Onyx Pharmaceuticals, Richmond, CA), Soldano Ferrone (Roswell Park Cancer Institute, Buffalo, NY), William Kaelin (Dana-Farber Cancer Institute,Boston, MA), Sabine Kirchhoff (German Cancer Research Center,Heidelberg, Germany), Richard Klausner (National Cancer Institute,Bethesda, MD), Chris Marshall (Cancer Research Center for Cell &Molecular Biology, London, United Kingdom), Joan Massague (Memorial Sloan Kettering Cancer Center, New York, NY), Michael Moran (Mount Sinai Hospital, Toronto, Ontario, Canada), Moshe Oren (Weizmann Institute of Science, Rehovot, Israel), Gordon Peters (Imperial Cancer Research Fund, London, United Kingdom), Andrew Simpson (Ludwig Institute for Cancer Research, Sao Paulo, Brazil), George Thomas (Friedrich Miescher Institute, Basel, Switzerland), Tada Taniguchi (University of Tokyo, Tokyo, Japan), Karen Vousden (National Cancer Institute, Frederick, MD), Alfred Wittinghofer (Max-Planck Institute, Dortmund, Germany), Paul Workman (The Institute of Cancer Research, Belmont, United Kingdom).

The posters were presented by Galina Selivanova (Karolinska Institute, Stockholm, Sweden), Silvia Miotti (National Cancer Institute, Milan, Italy), Anna Rapallo (National Cancer Institute,Genoa, Italy), Giovanni Martinelli (Institute of Hematology and Medical Oncology “Seragnoli,” University of Bologna), Daniel D’Orazio(University Hospital, Basel, Switzerland), Caterina La Porta(Department of General Physiology and Biochemistry, University of Milan, Milan, Italy), Tsukasa Shibuel (Department of Immunology,University of Tokyo, Tokyo, Japan), G. Sava (Department of Biomedical Sciences, University of Trieste, Trieste, Italy), K. Lang (Institute for Immunology, Witten/Herdecke University, Witten, Germany), A. Migliaccio (Institute of General Pathology and Oncology, II University of Naples, Naples, Italy), Jadwiga Chroboczek (Institute of Structural Biology, Grenoble, France).

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 symposium was held June 1–3, 2000, in Trento, Italy. The symposium was cochaired by Drs. Enrico Mihich(Roswell Park Cancer Institute, Buffalo, NY) and Ed Harlow(Massachusetts General Hospital Cancer Center, Charlestown, MA).

3

The abbreviations used are: VHL, von Hippel-Lindau (hereditary cancer syndrome); APC, adenomatous polyposis coli; CDK, cyclin dependent kinase; TNF, tumor necrosis factor; CD95(APO/FAS), member of TNF receptor superfamily; DISC, death-inducing signal complex; FAP, familial adenomatous polyposis; GAP, GTPase activating protein; GEF, guanine nucleotide exchange factor; GC,germinal center; HDAC, histone deacetylase; HIF, hypoxia-inducible factor; IRF, interferon regulatory factor; NHL, non-Hodgkins lymphoma;Ran, Ras-related GTP-binding protein; Rb, retinoblastoma suppressor gene; SWI/SNF, (BRG1) ATPase-dependent chromatin remodeling complex;min-mouse, mouse with multiple intestinal neoplasms consequent to an APC mutation; STAT, signal transducers and activators of transcription; RanBP, Ran-binding protein; VEGF, vascular endothelial growth factor; FR, folate receptor; PKC, protein kinase C.

4

M. Morin, unpublished observations.