INTRODUCTION: Autophagy is a regulated degradation pathway functioning in both cell survival and cell death. Its role in cancer is controversial since autophagy can be protective or destructive to tumor cells, depending on individual genetic signatures, stage of malignancy and treatment conditions. Hypoxia is a common feature of solid tumors, correlating with poor prognosis and chemoresistance. We have investigated autophagy in hypoxic cancer cells and examined the role of the hypoxia-inducible protein, BNIP3. METHODS: Multiple cancer cell lines were exposed to chronic hypoxia (<1% O2) in the presence or absence of specific inhibitors for autophagy and apoptosis. Cell death was measured by membrane permeability assay, and autophagy was assayed by GFP-LC3 distribution, LC3 processing, electron microscopy, and acidic vacuole formation. BNIP3 was over-expressed by transient transfection, stably induced in a tetracycline-regulated expression system, or knocked down using siRNA. Whole brain morphology, cell proliferation, and hypoxic response were additionally studied in a BNIP3-null mouse model. RESULTS: Autophagic cell death was detected in hypoxic cancer cells, occurring independent of apoptosis through a mechanism involving BNIP3. BNIP3 itself induced autophagic cell death, and loss of BNIP3 protected against hypoxia-induced autophagy and cell death. Loss of BNIP3 also resulted in differential growth and cell cycle progression in vitro, and increased brain cellularity in vivo compared to wild type controls. Potential mediators of resistance to BNIP3-induced cell death were identified using a novel model of BNIP3 resistance. CONCLUSIONS: Taken together, these results support the emerging theory that autophagy represents an alternative cell death pathway that could be targeted in hypoxic and/or apoptosis-resistant tumors. We have specifically identified BNIP3 as a potential target molecule in this pathway. Finally, we have identified a possibly novel role for BNIP3 in brain development and cell cycle regulation. These findings have important clinical applications given the potential for personalized cancer therapy based on individual tumor characteristics including autophagic capacity, hypoxic status, and BNIP3 activity.

The BNIP3 family, a group of death-inducing mitochondrial proteins, includes BNIP3, NIX and BNIP3h. These proteins share structural and functional similarities. BNIP3 causes neuronal cell death in a necrosis-like, caspase-independent manner with mitochondrial dysfunction. We reported that BNIP3 plays a role in delayed neuronal death in stroke models. Over-expression of BNIP3 causes up to 70% neuronal death, while knockdown of BNIP3 only protects 23% neurons from hypoxia. Thus, we hypothesize that other members of the BNIP3 subfamily compensate for the loss of BNIP3. BNIP3 and NIX were highly upregulated in the oxygen and glucose deprivation (OGD)/reoxygenation model, and knockdown of BNIP3 or NIX protected about 20% - 44% of neurons. Knockdown of BNIP3 family reduced neuronal death by 48%. Mitochondrial membrane potential loss, mitochondrial permeability transition pore (MPTP) opening and reactive oxygen species (ROS) production were all significantly attenuated by BNIP3 and/or NIX inhibition. AIF and EndoG were reported involving in caspase-independent cell death in ischemic stroke. We found that AIF was released from mitochondria and translocated into nuclei in neurons after OGD/reoxygenation, while inhibition of BNIP3 blocked AIF and EndoG translocation and prevented neuronal death. Over-expression of BNIP3 and NIX caused AIF translocation and subsequent neuronal death. These data reveal the effects of the BNIP3 family in neuronal death and indicate that AIF and EndoG are two downstream factors in the BNIP3-mediated cell death pathway. Meanwhile, necrostatin-1 (Nec-1), an inhibitor for a caspase-independent necrotic cell death, is able to protect neurons from death in stroke, mechanism of which is unclear. Here, we confirmed that Nec-1 significantly increased survival of neurons in models of stroke in vivo and in vitro. It also attenuated hypoxia or BNIP3-induced mitochondrial dysfunction and prevented mitochondrial release of AIF. Nec-1 did not affect the expression levels of BNIP3 but prevented its integration into mitochondria. These results suggest that Nec-1 protects neurons against ischemia by targeting BNIP3. In summary, this research indicates that the BNIP3 family is one of the regulators of caspase-independent neuronal death in stroke and that Nec-1 is an inhibitor for BNIP3 and a potential therapeutic agent for stroke.

INTRODUCTION: Autophagy is a regulated degradation pathway functioning in both cell survival and cell death. Its role in cancer is controversial since autophagy can be protective or destructive to tumor cells, depending on individual genetic signatures, stage of malignancy and treatment conditions. Hypoxia is a common feature of solid tumors, correlating with poor prognosis and chemoresistance. We have investigated autophagy in hypoxic cancer cells and examined the role of the hypoxia-inducible protein, BNIP3. METHODS: Multiple cancer cell lines were exposed to chronic hypoxia (<1% O2) in the presence or absence of specific inhibitors for autophagy and apoptosis. Cell death was measured by membrane permeability assay, and autophagy was assayed by GFP-LC3 distribution, LC3 processing, electron microscopy, and acidic vacuole formation. BNIP3 was over-expressed by transient transfection, stably induced in a tetracycline-regulated expression system, or knocked down using siRNA. Whole brain morphology, cell proliferation, and hypoxic response were additionally studied in a BNIP3-null mouse model. RESULTS: Autophagic cell death was detected in hypoxic cancer cells, occurring independent of apoptosis through a mechanism involving BNIP3. BNIP3 itself induced autophagic cell death, and loss of BNIP3 protected against hypoxia-induced autophagy and cell death. Loss of BNIP3 also resulted in differential growth and cell cycle progression in vitro, and increased brain cellularity in vivo compared to wild type controls. Potential mediators of resistance to BNIP3-induced cell death were identified using a novel model of BNIP3 resistance. CONCLUSIONS: Taken together, these results support the emerging theory that autophagy represents an alternative cell death pathway that could be targeted in hypoxic and/or apoptosis-resistant tumors. We have specifically identified BNIP3 as a potential target molecule in this pathway. Finally, we have identified a possibly novel role for BNIP3 in brain development and cell cycle regulation. These findings have important clinical applications given the potential for personalized cancer therapy based on individual tumor characteristics including autophagic capacity, hypoxic status, and BNIP3 activity.

The BNIP3 family, a group of death-inducing mitochondrial proteins, includes BNIP3, NIX and BNIP3h. These proteins share structural and functional similarities. BNIP3 causes neuronal cell death in a necrosis-like, caspase-independent manner with mitochondrial dysfunction. We reported that BNIP3 plays a role in delayed neuronal death in stroke models. Over-expression of BNIP3 causes up to 70% neuronal death, while knockdown of BNIP3 only protects 23% neurons from hypoxia. Thus, we hypothesize that other members of the BNIP3 subfamily compensate for the loss of BNIP3. BNIP3 and NIX were highly upregulated in the oxygen and glucose deprivation (OGD)/reoxygenation model, and knockdown of BNIP3 or NIX protected about 20% - 44% of neurons. Knockdown of BNIP3 family reduced neuronal death by 48%. Mitochondrial membrane potential loss, mitochondrial permeability transition pore (MPTP) opening and reactive oxygen species (ROS) production were all significantly attenuated by BNIP3 and/or NIX inhibition. AIF and EndoG were reported involving in caspase-independent cell death in ischemic stroke. We found that AIF was released from mitochondria and translocated into nuclei in neurons after OGD/reoxygenation, while inhibition of BNIP3 blocked AIF and EndoG translocation and prevented neuronal death. Over-expression of BNIP3 and NIX caused AIF translocation and subsequent neuronal death. These data reveal the effects of the BNIP3 family in neuronal death and indicate that AIF and EndoG are two downstream factors in the BNIP3-mediated cell death pathway. Meanwhile, necrostatin-1 (Nec-1), an inhibitor for a caspase-independent necrotic cell death, is able to protect neurons from death in stroke, mechanism of which is unclear. Here, we confirmed that Nec-1 significantly increased survival of neurons in models of stroke in vivo and in vitro. It also attenuated hypoxia or BNIP3-induced mitochondrial dysfunction and prevented mitochondrial release of AIF. Nec-1 did not affect the expression levels of BNIP3 but prevented its integration into mitochondria. These results suggest that Nec-1 protects neurons against ischemia by targeting BNIP3. In summary, this research indicates that the BNIP3 family is one of the regulators of caspase-independent neuronal death in stroke and that Nec-1 is an inhibitor for BNIP3 and a potential therapeutic agent for stroke.

Abstract Alternative splicing provides a versatile mechanism by which cells can generate proteins with different or even antagonistic properties. Herein we describe a novel splice variant of the hypoxia-inducible death gene Bnip3. Sequence analysis of the new Bnip3 protein revealed an N-terminus that was identical to Bnip3 but contained an Endoplasmic reticulum (ER) retention motif within the C-terminus, therefore we designated the new Bnip3 isoform NIPLET for (Nip-Like ER Target). While Bnip3 was predominately localized to mitochondria and promoted mitochondrial perturbations and cell death, NIPLET was preferentially localized to the ER and opposed the cytotoxic actions of Bnip3. Interestingly, NIPLET suppressed mitochondrial injury from Bnip3 activation and mitochondrial permeability transition pore opening by a mechanism dependent upon the dynamin motor protein Mitofusin-2 (MFN2). Notably, mutations of NIPLET within the critical ER retention motif rendered NIPLET defective for interacting with MFN2 and suppressed necrosis induced by Bnip3 or hypoxia. Hence, our findings reveal a novel signaling pathway that functionally couples ER and mitochondria for cell survival to a mechanism that is mutually dependent and obligatorily linked to a novel BNIP3 protein in cardiac myocytes. / May 2016

Autophagy is a process by which cells re-cycle organelles and macromolecular proteins during cellular stress. Defects in the regulation of autophagy have been associated with various human pathologies including heart failure. In the heart tumor suppressor p53 protein is known to promote apoptotic and autophagic cell death. We found p53 over-expression increased endogenous protein level of the hypoxia-inducible Bcl-2 death gene Bnip3 which leads to loss of mitochondrial membrane potential (ΔΨm). This was accompanied by autophagic flux and cell death. Conversely, loss of function of Bnip3 in cardiac myocytes or Bnip3-/- mouse embryonic fibroblasts prevented mitochondrial targeting of p53 and autophagic cell death. These data provide the first evidence for the dual regulation of autophagic cell death of cardiac myocytes by p53 that is mutually dependent on Bnip3 activation. Hence, our findings may explain how autophagy and cell death are dually regulated during cardiac stress conditions where p53 is activated.

Autophagy is a process by which cells re-cycle organelles and macromolecular proteins during cellular stress. Defects in the regulation of autophagy have been associated with various human pathologies including heart failure. In the heart tumor suppressor p53 protein is known to promote apoptotic and autophagic cell death. We found p53 over-expression increased endogenous protein level of the hypoxia-inducible Bcl-2 death gene Bnip3 which leads to loss of mitochondrial membrane potential (ΔΨm). This was accompanied by autophagic flux and cell death. Conversely, loss of function of Bnip3 in cardiac myocytes or Bnip3-/- mouse embryonic fibroblasts prevented mitochondrial targeting of p53 and autophagic cell death. These data provide the first evidence for the dual regulation of autophagic cell death of cardiac myocytes by p53 that is mutually dependent on Bnip3 activation. Hence, our findings may explain how autophagy and cell death are dually regulated during cardiac stress conditions where p53 is activated.

Les mitochondries sont des organites essentiels qui fusionnent et fissionnent en permanence. Cette dynamique est régulée par différentes GTPase, notamment OPA1 qui est impliquée dans la fusion mitochondriale. OPA1 possède également une fonction anti-apoptotique, régulée par BNIP3, un membre pro-apoptotique de la famille Bcl-2. En conditions de stress, comme l'hypoxie, la protéine BNIP3 est induite et inhibe OPA1, ce qui conduit à une fragmentation des mitochondries et à l'apoptose. En plus de sa fonction pro-apoptotique, BNIP3 a récemment été impliquée dans l'autophagie et la mitophagie, une forme d'autophagie sélective des mitochondries qui assure le contrôle qualité de l'organite. Mon travail de thèse a visé à étudier l'implication des protéines BNIP3 et OPA1 dans la balance entre survie par autophagie-mitophagie et mort par apoptose dans les neurones. La protéine BNIP3 a été induite ou surexprimée en utilisant, respectivement, un mimétique de l'hypoxie ou des lentivirus, dans des neurones en culture primaire. Par des analyses biochimiques et de microscopie nous avons étudié les effets de cette induction/surexpression sur la morphologie mitochondriale, l'autophagie-mitophagie et l'apoptose, ainsi que l'impact d'OPA1 sur ces processus. Nous avons montré que l'induction de BNIP3 provoque une séquence d'évènements qui débute par la fragmentation du réseau mitochondrial, est suivie par un processus d'autophagie-mitophagie de survie et se termine par la mort des neurones. Alors que la surexpression d'OPA1 diminue la fragmentation des mitochondries et la mort des neurones, elle n'affecte pas l'autophagie-mitophagie dans nos conditions. Les mutations du gène d'OPA1 sont responsables d'une maladie neurodégénérative, l'Atrophie Optique Dominante Autosomale de type 1 (ADOA1), qui se traduit par une atrophie du nerf optique, pouvant conduire à la cécité, et des atteintes neurologiques extra-oculaires variées. Nous avons démontré dans des modèles in vitro et in vivo de l'ADOA1 que le taux de protéine BNIP3 basal est réduit. Ceci s'accompagne in vitro d'une diminution de l'autophagie-mitophagie qui pourrait contribuer à sensibiliser les neurones haploinsuffisants en protéine OPA1 à différents stress. BNIP3 participerait donc, de concert avec son partenaire OPA1, au contrôle du destin des neurones, favorisant la survie cellulaire grâce à son activité pro-autophagique et mitophagique lors de dommages modérés, mais entraînant la mort quand les dommages sont trop conséquents. La découverte d'une modulation de l'expression de BNIP3 dans des modèles d'ADOA1 permet de proposer un rôle de la protéine dans l'étiologie de la maladie. / Mitochondria are essential organelles that constantly fuse and divide. These dynamic processes are controlled by various GTPases, such as OPA1, which is involved in mitochondrial fusion. OPA1 has also an anti-apoptotic function, which is regulated by BNIP3, a pro-apoptotic member of the Bcl-2 family. Upon stresses, such as hypoxia, BNIP3 is induced and inactivates OPA1, leading to mitochondrial fragmentation and apoptosis. Besides its pro-apoptotic function, BNIP3 was recently shown to be involved in autophagy and mitophagy, a selective autophagy of mitochondria that ensures their quality control. This study aims to decipher the role of BNIP3 and its partner OPA1 in the balance between survival by autophagy-mitophagy and death by apoptosis in neurons. BNIP3 induction or overexpression was achieved, respectively, using a mimetic of hypoxia or lentiviruses, in primarily cultured neurons. Various microscopy and biochemistry approaches were used to analyse the impact of this induction/overexpression on mitochondrial morphology, autophagy-mitophagy and apoptosis. We showed that BNIP3 induction led to a sequence of events that started with mitochondrial network fragmentation, followed by pro-survival autophagic and mitophagic processes, and ended by cell death. OPA1 mutations lead to a neurodegenerative condition: Autosomal Dominant Optic Atrophy type 1 (ADOA1). ADOA1 patients suffer from optic nerve atrophy leading to blindness and various extra-ocular neurological defects. We evidenced in in vitro and in vivo ADOA1 models a lowered BNIP3 basal level. This decrease is associated in vitro with a reduction of autophagy-mitophagy that could lead to increased susceptibility to various stresses. BNIP3 thus controls, together with its partner OPA1, the fate of neurons, favoring cell survival upon moderate damages thanks to its pro-autophagic and mitophagic activity, or leading to cell death upon extensive damages. Having demonstrated that BNIP3 expression is affected in ADOA1 models, we propose a role of the protein in the etiology of the disease.

L’autophagie est un mécanisme de l’immunité innée contre le VIH-1 déclenché très rapidement dans les cellules T CD4 par les protéines d’enveloppe du virus (Env). Dans les cellules T CD4 appelées « bystander », c’est-à-dire où l’infection est abortive, l’autophagie entraine l’apoptose. Dans les cellules T CD4 infectées de façon productive, l’autophagie est inhibée, empêchant ainsi à la fois la dégradation du virus et/ou des protéines virales et l’apoptose médiée par Env. Le but de cette étude de doctorat est de comprendre comment l’autophagie est bloquée dans les cellules T CD4 infectées par le VIH-1. Durant la première année de ma thèse, j’ai contribué au travail montrant que Vif nouvellement synthétisé bloque l’autophagie dans les lymphocytes T CD4 infectés. Les données ont été publiées en Janvier 2015 dans AIDS sous le titre «HIV-1 viral infectivity factor interacts with microtubule-associated protein light chain 3 and inhibits autophagy». Ensuite, j’ai montré que Vpr contribue aussi à l’inhibition de l’autophagie. En effet, Vpr ectopique diminue grandement le nombre des autophagosomes présents dans les cellules T CD4 quand l’autophagie est induite par un inhibiteur de mTOR. De même, Vpr incorporé dans les virions diminue le nombre des autophagosomes présents dans les cellules T CD4 très rapidement après leur infection productive (4h et 8h). Dans le but de définir le mécanisme par lequel la protéine Vpr entraine ce blocage, j’ai fait des expériences de GST pull-down et j’ai identifié que Vpr interagit avec BNIP3, une protéine pro-autophagique. De plus, le niveau d’expression de BNIP3 est augmenté dans les cellules T CD4 après le contact avec Env, suggérant que BNIP3 pourrait être impliqué dans l’induction de l’autophagie médiée par l’Env. Vpr co-localise avec BNIP3 et Vpr incorporé dans les virions induit une diminution drastique du niveau d’expression de BNIP3 après 8 heures d’infection. En conclusion, j’ai montré qu’au moins deux protéines du VIH-1 sont séquentiellement impliquées dans l’inhibition de l’autophagie dans les cellules T CD4 infectées par le VIH-1, Vpr qui contrôle l’autophagie durant la phase précoce de l’infection, et Vif néo-synthétisé qui inhibe l’autophagie après la transcription du génome viral. La compréhension complète des mécanismes par lesquels le VIH-1 inhibe l’autophagie devrait à terme permettre l’élaboration de nouvelles stratégies thérapeutiques pour lutter contre ce virus. / Autophagy is a potent anti-HIV-1 mechanism. It is triggered in CD4 T cells by the viral envelope (Env) upon HIV-1 entry. In bystander CD4 T cells, autophagy leads to apoptosis. In productively infected CD4 T cells, autophagy is inhibited, preventing thus HIV-1 virophagy and Env-mediated apoptosis. The aim of my PhD study was to understand how autophagy is blocked in the HIV-1-infected CD4 T cells. During the first year of my thesis, I contributed to the work demonstrating that Vif neosynthesized blocks autophagy in the infected CD4 T lymphocytes. The data were published in January 2015 in AIDS under the title “HIV-1 viral infectivity factor interacts with microtubule-associated protein light chain 3 and inhibits autophagy”. Then, I demonstrated that Vpr contributes also to the inhibition of autophagy. Indeed, both ectopic expression of Vpr and Vpr incorporated into the virions decrease the number of autophagosomes in CD4 T cells when autophagy is induced by an inhibitor of mTOR and Env, respectively. To define the mechanism by which HIV-1 Vpr inhibits autophagy, I performed GST pull-down experiments and identified that Vpr interacts with BNIP3, a pro-autophagic protein. Importantly, BNIP3 expression level is increased in CD4 T cells upon Env contact, suggesting that BNIP3 could be responsible for the Env-mediated induction of autophagy. Furthermore, Vpr co-localizes with BNIP3 and viral incorporated Vpr decreases BNIP3 levels after 8 hours of infection. In conclusion, I demonstrated that at least two HIV-1 proteins are sequentially involved in the inhibition of autophagy in HIV-1-infected CD4 T cells, Vpr that controls autophagy during the early phase of infection, and then neo-synthesized Vif that inhibits autophagy after HIV-1 genome transcription. The complete understanding of the mechanisms by which HIV-1 inhibits autophagy should lead to the rising of new molecular strategies to fight against this virus.

PARP-1 has been identified as a major player in apoptotic pathways. Its excessive activation causes mitochondrial dysfunction, permeability, and AIF mitochondrion-to-nucleus translocation. It has been suggested that PARP-1 interacts indirectly with Bnip3, a mitochondrial pro-apoptotic factor. However, the mechanistic linkage is still not well understood. Our lab has shown that cytosolic/nuclear NAD+ depletion is a hallmark for PARP-1 over activation and inhibition of sirtuin activity. Specifically in my project, we think that PARP-1 induced- NAD+ depletion and sirtuin inhibition causes hyperacetylation of the α subunit of the transcription factor HIF-1 allowing increased HIF-1 binding to Bnip3 upstream promoter, and increased Bnip3 expression. Indeed, our PARP-1 Knock out neurons, MNNG and PJ34 treatment, chromatin immunoprecipitation, and HIF-1α loss of function studies strongly confirmed the necessity of HIF-1 to increase Bnip3 expression in hypoxia. Overall, our research suggests a role for HIF-1 in increasing PARP-1 dependent Bnip3 expression in hypoxic models. / May 2016