<p>Normal cells require input from extrinsic growth factors to control proliferation and survival. Recent studies have demonstrated that these same extrinsic signals also regulate cellular metabolism to ensure that metabolism adequately supports the demands of cell function, proliferation, and cell survival. The PI3K/Akt pathway is downstream of many growth factors and can promote both glucose metabolism and cell survival. Aberrant activation of the PI3K/Akt pathway is common in cancer, and its activation can contribute to the growth factor independence that is a hallmark of neoplastic cells. Metabolic demand is high in stimulated and leukemic T cells, and activation of Akt can increase glucose metabolism to meet these requirements. There is great interest in targeting the unique metabolism of cancer cells for cancer therapy, thus making an understanding of the interaction of metabolism and cell death essential. </p><p>Akt is also anti-apoptotic and can inhibit cell death by regulating members of the Bcl-2 family. Interestingly, the ability of Akt to prevent cell death is inextricably linked to its metabolic function. Several recent studies have demonstrated that glucose metabolism can affect Bcl-2 to family members to promote cell survival, but the role of Akt-dependent glucose metabolism in the regulation of Bcl-2 family members is not understood. Using a model of growth factor withdrawal-induced apoptosis, we show that Akt prevents cell death by maintaining glucose metabolism to regulate the Bcl-2 family members Puma and Mcl-1, and demonstrate the importance of this pathway in the survival of stimulated T lymphocytes and leukemia.</p><p>After growth factor withdrawal, active Akt suppressed Puma induction in abundant glucose, but Puma was rapidly upregulated in glucose-deficient conditions and was necessary and sufficient to promote efficient cell death. Importantly, glucose was not uniquely required, as provision of alternative mitochondrial fuels allowed Akt to suppress Puma and maintain survival. This metabolic regulation of Puma was mediated through partially p53-dependent transcriptional induction as well as control of Puma protein stability. </p><p>In addition to inhibiting Puma expression, active Akt prevented the loss of Mcl-1 after growth factor withdrawal by sustaining Mcl-1 protein synthesis in a glucose-dependent manner. Mcl-1 was essential for preventing Bim-induced apoptosis, as Akt could not inhibit Bim induction after growth factor deprivation. Slowing of Mcl-1 synthesis by inhibiting glucose metabolism reversed Mcl-1-mediated resistance of leukemic cells to the Bcl-2 inhibitor ABT-737. Importantly, Akt and glucose-reliant Mcl-1 expression required mTOR-dependent phosphorylation of 4EBP, and treatment with mTOR inhibitors also reversed ABT-737 resistance. </p><p>Together, this study demonstrates that Akt promotes cell survival by preventing metabolic checkpoints that stimulate Puma expression and stability and inhibit Mcl-1 synthesis, advancing our understanding of the links between metabolism and cell death. These studies highlight the importance of cellular metabolism--including a potential role for the alternative sugar fructose--in cancer cell survival that may provide a mechanistic understanding to drive development of metabolism-targeted cancer therapies.</p> / Dissertation

<p>It has been long known that many types of cancers have high metabolic requirements and use reprogrammed metabolism to support cellular activities. The first identified metabolic alteration in cancer cells was elevated glucose uptake, glycolysis activity and lactate production even in the presence of oxygen. This metabolic program, termed aerobic glycolysis or the Warburg effect, provides cells with energy as well as biosynthetic substrates to sustain cell survival and rapid cell proliferation. Cancer metabolism is closely linked to genetic mutations and oncogenic signaling pathways, such as PI3K/Akt, cMyc and HIF pathways. These oncogenic signals can direct metabolic reprogramming while changes in metabolic status can regulate activities of these signaling pathways in turn. In addition to glucose, later studies also found utilization of alternate nutrients in cancer cells, including glutamine and lipids. Glutamine is the second major metabolic fuel and can be converted to various substrates to support cell bioenergetics needs and biosynthetic reactions. Usage of metabolic fuels in cancer cells, however, is variable. While certain cancers display addiction to one type of nutrient, others are capable of using multiple nutrients. </p><p>The unique metabolic features of cancer cells raise the possibility of targeting metabolism as a novel therapeutic approach for cancer treatment. Using pharmacological inhibitors, previous research has provided corroborating evidence that metabolic stress can impact survival and growth of proliferative cancer cells by regulating cell apoptotic machinery and cell cycle checkpoints. Due to lack of genetic tools and side effects from these inhibitors, however, mechanistic understanding of cell response to metabolic inhibition was limited in these studies. More importantly, how metabolic stress affects cancer progression in a physiological condition has not yet been well investigated. Lastly, current research has not examined metabolic program in indolent cancers and the metabolic requirements and activities in less proliferative cells also remain to be understood.</p><p>This work examines nutrients utilization in B cell derived acute and chronic leukemia (B-ALL and B-CLL). B-ALL is an aggressive form of leukemia. Using cell lines and primary patient samples, we found B-ALL cells primarily used glucose through aerobic glycolysis, similar to other proliferative cancer cells. B-ALL cells were also more sensitive to inhibition of glycolysis than normal B cells. Employing an untargeted metabolomics profiling in combination with isotope labeled glucose tracing approach, we show in a B-ALL model that genetic ablation of glucose transporter Glut1 partially reduced glucose uptake, sufficiently hindered anabolic pathways and promoted catabolic metabolism. This metabolic shift led to sharply curtailed B-ALL proliferation in vitro and reduced leukemic burden in vivo. Furthermore, this partial inhibition of glucose metabolism sensitized B-ALL cells to apoptotic stimuli and non-cytotoxic metabolic inhibition significantly enhanced efficacy of a tyrosine kinase inhibitor to eliminate B-ALL cells in vitro and in vivo. Thus, partial inhibition of glucose metabolism can provide a plausible adjuvant therapy to treat cancers that depend on glycolysis for survival and proliferation. </p><p>In contrast to B-ALL, B-CLL is an indolent form of cancer. Most B-CLL cells exhibited low glucose metabolic activities that were comparable with normal B cells at resting stage. Similar to chronically stimulated and anergic B cells, these B-CLL cells also failed to upregulate glucose metabolism in response to IgM stimulation. We also observed an altered amino acid and acyl-carnitine profile and increased glutaminase mRNA in B-CLL relative to normal B cells, suggesting the capability of using alternate nutrients such as glutamine in these cells. Finally, we explored the possibility of suppressing mitochondria metabolism to induce B-CLL cell death through inhibition of the nuclear hormone receptor and metabolic regulator ERRalpha. ERRalpha is known to regulate mitochondrial metabolism and was expressed higher in B-CLL than normal B cells. ERRalpha inhibition decreased viability of oncogene transformed pro-B cells, suggesting ERRalpha as a potential target for B-CLL treatment.</p><p> Collectively, this work investigates metabolic phenotype in two forms of leukemia derived from B cells. It reveals different metabolic requirements and activities in aggressive and indolent leukemia and explores different approaches to suppress metabolism in these cancers. Findings of this work shed light on how to potentially design metabolic approach to improve cancer treatment.</p> / Dissertation

Le cancer est un problème majeur de santé public, tuant chaque année plusieurs millions de personnes. L’inhibition de la mort cellulaire programmée, l’apoptose, est considérée comme l’un des paramètres principaux impliqués dans son initiation et son développement. La régulation de la voie intrinsèque (mitochondriale) de l’apoptose est régulée par la famille Bcl-2. Jusqu’à maintenant, on pensait que la protéine PUMA, une protéine pro-apoptotique, était principalement exprimée au niveau mitochondrial. Nous avons montré qu’à l’état basal, PUMA était exprimé au niveau du cytosol des lymphocytes B humains. Cependant, suite à un signal apoptotique, PUMA est capable de transloquer du cytosol à la mitochondrie, de façon indépendante des caspases mais dépendante de l’activation de la MAPKinase p38, permettant ainsi son interaction avec les protéines anti-apoptotiques Bcl-2 et Mcl-1 dont l’inhibition conduit à l’apoptose. Les protéines anti-apoptotiques, Mcl-1 notamment, sont souvent surexprimées dans les tumeurs. Mcl-1 est une protéine à courte demi-vie, rapidement dégradée par le protéasome. Cette dégradation dépend de son ubiquitination réalisée par des E3 ligases (E3). Quelques E3 et une déubiquitinase (DUB), hydrolysant les chaînes d’ubiquitine, régulant l’expression de Mcl-1 ont été décrites. Cependant, ces protéines sont soit très peu exprimées, soit leur inhibition n’a pas d’impact sur l’expression de Mcl-1 dans notre modèle. Nous avons donc entrepris de caractériser de nouvelles E3 et DUB régulant l’ubiquitination de Mcl-1. Après une immunoprécipitation de Mcl-1 dans nos cellules, suivie d’une analyse par spectrométrie de masse, nous avons identifié la DUB USP14. Lorsque son expression est diminuée, l’expression et la stabilité de Mcl-1 augmentent de façon sélective. Nos résultats pourraient contribuer à une approche à double-tranchant dans le traitement du cancer, en retirant les freins à l’apoptose via une diminution de l’expression de Mcl-1 d’une part et en l’activant via PUMA de l’autre. / Cancer is a major public health issue, killing millions of people worldwide each year. The inhibition of apoptosis, a programmed cell death, in its onset and development has been well documented, making it one of the hallmarks of cancer. The regulation of the intrinsic (mitochondrial) pathway of apoptosis is regulated by the Bcl-2 (B cell lymphoma-2) family. Up until now, PUMA, a pro-apoptotic protein, was thought to be mainly expressed at the mitochondria, based on experiments where it had been overexpressed. We showed that endogenous PUMA is mainly expressed in the cytosol of activated or resting B cells. However, upon apoptotic stress, PUMA was able to translocate from the cytosol to the mitochondria, in a caspase-independent but p38-dependent manner, allowing PUMA to bind and inhibit the anti-apoptotic proteins Bcl-2 and Mcl-1, and thereby leading to cell death. The anti-apoptotic proteins, especially Mcl-1, are often overexpressed in tumors. Mcl-1 is a protein with a short half-life, degraded rapidly by the proteasome. This degradation is ubiquitin-dependent, requiring E3 ligases (E3). A handful of E3s and one deubiquitinase (DUB), that hydrolyses the ubiquitin chains, have been reported to regulate Mcl-1 expression. However, they were either very poorly expressed or their inhibition had no impact on Mcl-1 expression in our model. We thus undertook to characterize new E3s and DUBs mediating Mcl-1 ubiquitination. After an immunoprecipitation of Mcl-1 in our cells, followed by a mass spectrometry analysis, we identified the DUB USP14. When knockdown, Mcl-1 expression was selectively increased and its stability enhanced. Our results could help build “double-edge” therapies, removing the breaks on apoptosis on one hand via Mcl-1 downregulation while activating it on the other via PUMA translocation.

No description available.

<p>Cancer occurs when cells acquire a number of mutations that trigger uncontrolled cell growth. The normal cellular response to this dysregulation of growth is the activation of programmed cell death. While focus in cancer research has been mainly concentrated in the mechanism of programmed cell death at the mitochondria, endoplasmic reticulum is slowly emerging as an essential platform for this regulatory mechanism.</p> <p>Bcl-2 is the founding member of the Bcl-2 family of protein, which contributes to the regulation of cell death at the mitochondria and at the endoplasmic reticulum. Previously in our lab, we have shown using MCF-7 cells stably expressing Bcl-2 targeted to the endoplasmic reticulum; they were protected from estrogen deprivation induced cell death. Thus the regulatory mechanism of Bcl-2 at the endoplasmic reticulum represents an interesting avenue to improve current cancer therapeutics.</p> <p>Two approaches were utilized to identify and characterize Bcl-2 and its interacting partners at the endoplasmic reticulum. Using an affinity tag fused to Bcl-2 that has been engineered to target the endoplasmic reticulum, tandem affinity purification was utilized to identify novel Bcl-2 interacting partners when estrogen receptor positive cells are treated with estrogen deprivation. Using fluorescent protein fused to the proteins of interest, Fluorescent Lifetime Imaging Measurement (FLIM) was used to characterize the interactions of Bcl-2 and its known interacting partner at the endoplasmic reticulum. The findings of this thesis verify the applications of the two aforementioned methods in the study of Bcl-2 interacting proteins at the endoplasmic reticulum.</p> / Master of Science (MSc)

<p>Bid is a member of the Bcl-2 family of proteins, which are known as the regula- tors of apoptosis. Bid recruits Bax, another Bcl-2 family protein, which forms large oligomers that permeabilize the mitochonrdial outer membrane during apoptosis. In this thesis, Bid complexes embedded in a mitochondria-like supported lipid bilayer were investigated using single molecule fluorescence techniques. The bilayer, con- taining a lipophilic tracer, was formed on a mica surface and ATTO647 labelled Bid was added to it. For experiments where the effect of Bax on Bid complexes was investigated, a wild type Bax or a HiLyte488 labelled Bax was added as well. The protein-bilayer sample was imaged using total internal reflection fluorescence (TIRF). The formation of a fluid bilayer was confirmed by the observation of the lateral diffusion of DiD. Single particle tracking of the lipid molecules was used to measure the diffusion coefficent of DiD which was determined to be 2.2μm2 /s. The TIRF images also revealed two populations of Bid complexes, immobile and mobile. The diffusion coefficient of the observed Bid complexes was determined to be about three times slower than that of DiD (0.8±0.5μm2 /s). This provides evidence that mobile Bid is embedded in the bilayer. Image analysis of immobile Bid complexes showed a step-wise decrease in the fluorescence intensity due to photobleaching. The oligomeric distribution of the immobile Bid complexes was determined from the num- ber of steps, which corresponds to the number of particles in each complex. From these distributions it was concluded that the imaged immobile Bid existed mainly as monomers. However dimer and trimer complexes of Bid were also observed. The detected oligomeric distribution was not affected by the presence of either wild type Bax or Hilyte488 Bax. However Bid was imaged for the first time participating in Bax complexes. The acquired results somewhat differ from what had been observed in confocal imaging of the same samples, where mostly larger Bid complexes (dimers and up) were detected. We attribute the difference to the superior sensitivity of the TIRF method presented here.</p> / Master of Science (MSc)

The primary focus of this research is the mechanisms of cell death in head and neck squamous cell carcinoma (HNSCC) treatment. These cancers typically originate in squamous cells that line the moist mucosal surfaces of head and neck. HNSCC is commonly treated with a platinum based agent, cisplatin. While the drug does offer strong antitumor effects, its prolonged use often results in tumor-acquired resistance, which limits treatment effectiveness. We have shown that cisplatin treatment induces the expression of a pro-apoptotic BCL-2 family member Noxa, which then initiates caspase- dependent apoptosis through its binding and sequestration of pro-survival protein MCL-1 for its inactivation. Without Noxa induction, cell death is significantly reduced when treating HNSCCs with cisplatin. The objectives of this study are (1) to determine the molecular mechanisms by which Noxa induces cell death in HNSCC cells; (2) to determine the molecular mechanisms of cisplatin-resistance in isogenic HNSCC cell lines. We observed an increase of apoptosis by ectopic expression of Noxa in all HNSCC cell lines tested, but not in immortalized human normal oral keratinocytes (NOK), suggesting that Noxa overexpression is sufficient to induce tumor-specific cell death. Noxa-induced cell death was mediated by BAX and BAK activation. BAK activation was mediated through Noxa binding to MCL-1, but not BCL-XL. Cisplatin- resistant cells induced less Noxa and apoptosis, supporting that Noxa induction is prerequisite for apoptosis induced by cisplatin. Taken together, Noxa induces tumor- specific cell death in HNSCC cells primarily through BAX and BAK activation, which suggests the therapeutic potential of this protein.

La macro-autophagie est une voie catabolique conservée dans l’évolution permettant la dégradation des organites endommagés ou vieux, des protéines à longue durée de vie ou agrégées et des portions du cytosol pour le recyclage métabolique afin de maintenir l'homéostasie cellulaire. L'absence d'autophagie est fréquemment observée dans de nombreuses pathologies incluant les cancers et les maladies neurodégénératives. Beclin 1, un suppresseur de tumeur,est une protéine clé dans la régulation de l’autophagie et participe à la nucléation de l’autophagosome. Beclin 1 est une protéine “BH3-only” pouvant interagir avec le site de fixation au domaine BH3 présent dans la protéine Bcl-2 et ses homologues. Cette interaction inhibe l’autophagie. Certains agents pharmacologiques tels qu’ABT737, appelés«BH3 mimetics», occupent le site de fixation du domaine BH3 de façon compétitive pour perturber l'interaction inhibitrice entre Beclin 1 et Bcl-2/Bcl-XL. Ceci permet à Beclin 1 de maintenir l’activité classe IIIphosphatidylinositol-3-kinase de Vps34 pour la formation du phagophore. L'autophagie est un processus finement régulé par de nombreux complexes protéiques. Les senseurs de la charge énergétique comme l’AMP-dependant kinase(AMPK), la cible mammalienne de la rapamycine (mTOR), la Sirtuin1 (SIRT1) ou les voies d’intégration du stress telles que celles impliquant l'inhibiteur des kinases NF-κB (IκBα) (IKK) et le suppresseur de tumeur p53, ont tous un impact majeur dans la régulation de l'autophagie. Dans de nombreux paradigmes de stimulation autophagique, ils semblent tous agir en amont de la dissociation Beclin 1-Bcl-2. Nos résultats révèlent qu’ABT737 stimule plusieurs voies pro-autophagiques pour obtenir une efficacité optimale. Ces résultats placent la SIRT1, AMPK / mTOR, HDM2et IKK en aval de la dissociation du complexe Beclin 1-Bcl-2. Cette étude démontre que les BH3-mimetics activent des voies multiples de stimulation de l’autophagie, peut-être en raison du degré élevé de connectivité qui existe entre les complexes protéiques de régulation de l’autophagie. Cela signifie qu’un effet spécifique sur l’interactome de Beclin 1 peut affecter d'autres voies dans le réseau du contrôle autophagique. Ces voies ne semblent pas suivre une hiérarchie linéaire, mais doivent être plutôt interconnectées dans un circuit complexe dans lequel la stimulation de l'autophagie par des déclencheurs physiologiques (tels que la carence en nutriments ou le stress des organites) induit un ensemble de changements intimement liés et impliqués dans une boucle de régulation positive qui constituerait un ensemble indissociable composant l’«autophagy switch». / Macro-autophagy is a conserved catabolic pathway that culminates in the degradation of old/damaged organelles,long-lived/aggregated proteins and portions of the cytosol for metabolic recycling to maintain cellular homeostasis.The absence of autophagy is frequently observed in many pathologies including cancers and neurodegenerative diseases. Beclin 1, a bona fide tumour suppressor, is the key autophagy regulatory protein that participates in autophagosome nucleation. Infect, Beclin 1 is a BH3-only protein that can interacts with the BH3 receptor domain contained within Bcl-2 and its homologues. This interaction functions as a inhibitory check on autophagy. Some pharmacological agents such as ABT737, referred to as ‘BH3 mimetics’, occupy the BH3-binding grooves to competitively disrupt the inhibitory interaction between Beclin 1 and Bcl-2/Bcl-XL allowing Beclin 1 to maintain the class III phosphatidylinositol-3-kinase activity of Vps34 for the phagophore formation. Autophagy is a complex process that is regulated by multiple protein complexes beyond that organized around Beclin 1. The energy sensors including AMP-dependent kinase (AMPK), mammalian target of rapamycin (mTOR), Sirtuin1 (SIRT1) as well as stress-integrating pathways such as those involving the inhibitor of NF-κB (IκB) kinases (IKK) and the tumour suppressor protein p53, all have a major impact on the regulation of autophagy. In many paradigms of autophagic stimulation, they all seem to act upstream of the dissociation of Beclin 1-Bcl-2. Our results reveal that ABT737stimulate multiple pro-autophagic pathways to be optimally efficient. These results place SIRT1, AMPK/mTOR,HDM2 and IKK downstream of the dissociation of the Beclin 1-Bcl-2 complex. This study advocates that BH3mimetics trigger multiple autophagy-stimulatory pathways maybe due to the high degree of connectivity that exists among autophagy-regulatory protein complexes meaning that a specific effect on the Beclin 1-interactome might affect other nodes in the autophagy-controlling network. These pathways cannot follow a linear hierarchy and rather must be interconnected in a complex circuitry, in which stimulation of autophagy by physiological triggers (such as starvation or organelle stress) induce an ensemble of intimately linked changes that are coupled to each other in positive feed forward loops constituting an indissociable ensemble that composes the “autophagic switch”.

La progression tumorale est la conséquence de multiples altérations génotypiques et phénotypiques; L’une d’entre elles, nécessaire à la formation de métastases, est l’acquisition d’une résistance à l’anoïkis, forme d’apoptose induite par la perte d’attachement à la matrice extra-cellulaire. Afin d’étudier l’anoïkis, nous avons utilisé deux lignées colorectales humaines isogéniques : la lignée SW480 dérivée de la tumeur primaire et sensible à l’anoïkis et la lignée SW620 dérivée d’une métastase ganglionnaire de cette même tumeur et résistante à l’anoïkis. Nous avons établi que dans les cellules SW480, l’anoïkis est une forme d’apoptose intrinsèque c'est-à-dire débutant à la mitochondrie et donc sous le contrôle des protéines de la famille Bcl-2. Parmi celles-ci, nous avons trouvé que seule la protéine proapoptotique à BH3-seul Bim était régulée différemment dans les deux lignées : tandis que son expression augmente de manière très significative dans les cellules SW480 cultivées en suspension, elle n’augmente que très peu dans les cellules SW620. De manière très intéressante, et malgré cette différence, les deux lignées se sont montrées être sensibles au BH3-mimétique ABT-737 mais seulement lorsqu’elles sont cultivées en suspension. Ces résultats indiquent que, qu’elles soient sensibles ou non à l’anoïkis, les cellules détachées sont « prédisposées à la mort » et que des composés semblables à l’ABT-737 tels que le Navitoclax ou l’ABT-199 pourraient avoir des propriétés anti-métastiques dans les tumeurs solides. Dans la seconde partie de ce travail, nous montrons que la protéine transmembranaire CDCP1 ( CUB Domain Containing Protein 1) semble nécessaire mais non suffisante pour protéger ces cellules contre l’anoïkis. CDCP1 est un substrat de Src mais sa phosphorylation sur tyrosine n’est pas impliquée dans cette protection. Finalement, nous avons identifié deux nouvelles protéines qui interagissent avec CDCP1 : l’ITGB4 et l’EphA2 / Tumour progression is the consequence of multiple genotypic and phenotypic alterations. One of these, necessary for the formation of metastasis, is acquisition of a resistance to anoïkis, a form of apoptosis triggered by loss of attachment to the extra-cellular matrix. In order to study anoïkis, we used two isogenic human colon cell lines : SW480 cells derived from the primary tumour and sensitive to anoïkis, and SW620 cells derived from a lymph node metastasis in the same patient which are resistant to anoikis. We found that anoikis signaling in SW480 cells is a form of intrinsic apoptosis thus starting at the mitochondria and under the control of Bcl-2 family proteins. Among the members of this family, the BH3-only proapoptotic protein Bim was the only one that we found to be differentially regulated between the two cell lines: whereas Bim expression augments strongly during the culture in suspension of SW480, it only slightly does so in SW620 cells. Most interestingly, despite this difference, both cell lines turned to be sensisitive to the BH3-mimetic ABT-737 but only when they are in suspension. This shows that, whether or not they a sensitive to anoikis, detached colon cancer cells are “primed for death” and thus that ABT-737 related compounds such as Navitoclax or ABT-199 might have anti-metastatic properties in solid tumours. In the second part of this work, we show that the transmembrane protein CDCP1 (CUB Domain Containing Protein 1) appears be necessary but not sufficient to protect these cells against anoïkis. CDCP1 is a Src substrate but its tyrosine phosphorylation is not involved in this protection. Finally, we have identified two new proteins interacting with CDCP1: ITGB4 and EphA2.

Les protéines de la famille Bcl-2 sont des régulateurs-clés de l’apoptose (mort cellulaire), qui agissent en contrôlant la perméabilisation de la membrane mitochondriale externe par un processus encore mal connu. La dérégulation des membres de cette famille est souvent associée à la progression tumorale et à la résistance à la chimiothérapie. Notre projet a cherché à éclaircir le mode d’action de ces protéines en se focalisant sur les déterminants structuraux régissant leur interaction avec les membranes biologiques. Les connaissances glanées ont permis (i) de mieux comprendre les déterminants à l’origine de la divergence évolutive entre membres pro- et anti-apoptotiques de la famille Bcl-2 ; (ii) d’ouvrir la voie à la conception de ‘poropeptides’ conçus sur le modèle des hélices d’insertion membranaire des protéines Bcl-2, et qui pourraient être utilisés pour induire l’apoptose de cellules tumorales ou des cellules endothéliales entourant les tumeurs. / Bcl-2 family proteins, which include pro- and antiapoptotic members, positively or negatively regulate mitochondrial outer membrane permeabilization, i.e. a critical step in apoptosis. Over-expression of pro-survival members is associated with tumor progression and may be responsible for chemotherapy resistance. Detailed understanding of the precise mechanisms by which Bcl-2 family members control apoptosis is therefore of considerable therapeutic interest. The overall aim of our project was to delineate a structure-function relationship of Bcl-2 family proteins with emphasis on their membrane-active domains. This analysis has provided a basis (i) to elucidate the molecular mechanisms by which different Bcl-2 proteins evolved opposite functions ; (i) to develop a new generation of pore-forming peptides targeting the mitochondrial outer membrane that may be used to kill neoplastic or tumor endothelial cells.