Hypoxia-induced autophagy cell death or cell survival

Available online at Hypoxia-induced autophagy:cell death or cell survival?Nathalie M Mazure and Jacques Pouysse´gurHypoxia( 3–0.1%oxygen)is capable of rapidly inducing,viathe hypoxia-inducible factor(HIF-1),a cell survival responseengaging autophagy.This process is mediated by the atypicalBH3-only proteins the Bcl-2/E1B19kDa-interacting protein3(BNIP3/BNIP3L(NIX))that are induced by HIF-1.Thesemitochondrial associated BNIP proteins also mediatemitophagy,a metabolic adaptation for survival that is able tocontrol reactive oxygen species(ROS)production and DNAdamage.In contrast,severe hypoxic conditions or anoxia(<0.1%oxygen),where the latter is often confused withphysiological hypoxia,are capable of inducing a HIF-independent autophagic response,generated via an extremenutritional stress response implicating the AMPK-mTOR andunfolded protein response(UPR)pathways.The autophagiccell death that is often observed in these extreme stressconditions should be seen as the outcome of failed adaptation.AddressInstitute of Developmental Biology and Cancer Research,University ofNice,CNRS-UMR6543,Centre Antoine Lacassagne,33Avenue deValombrose,06189Nice,FranceCorresponding author:Mazure,Nathalie M(mazure@unice.fr)andPouysse´gur,Jacques(pouyssseg@unice.fr)Current Opinion in Cell Biology2009,22:1–4This review comes from a themed issue onCell regulationEdited by Guido Kroemer and Eileen White0955-0674/$–see front matter#2009Elsevier Ltd.All rights reserved.DOI10.1016/j.ceb.2009.11.015IntroductionAutophagy plays key physiological cellular functions suchas degradation of long-lived proteins,organelle turnover,adaptation to nutrient depletion,extension of lifespan,cellular development,differentiation,and antiaging.However,autophagy also plays pathophysiological roles.Too much or too little autophagy is linked to:first,pathogen infection,by supplying nutrients to pathogens;second,neurodegeneration,allowing accumulation ofaggregated proteins;third,muscular disorders,causedby the accumulation of autophagosomes that impair cel-lular functions;fourth,liver disease,as a result of exces-sive mitophagy;andfinally,cancer,in which it acts as aprotector against some cancer treatments and most impor-tantly,prevents cell death in a hostile microenvironment[1].Here we will outline the different mechanisms thattrigger autophagy in cells cultured in a rich medium,inwhich oxygen is the only limiting factor,as well as in cellsmore severely restricted in oxygen and nutrients.Autophagy,hypoxia,HIF,and BNIP3/BNIP3LHIF is a key transcription factor that allows rapid adap-tation to and survival in a large range of reduced oxygenconcentrations.By controlling the transcription of hun-dreds of genes HIF drives,or strongly influences,manycellular process such as erythropoiesis,angiogenesis,energy metabolism,pH regulation,cell migration,andtumor invasion.HIF is a heterodimer of an a subunit thatis unstable in the presence of relatively high levels ofoxygen(above5%),and a b subunit(or ARNT)that is notoxygen regulated.While studying the bifunctionalactivity of the transcriptional activation domains(TADs)of HIF-1a,we suggested that the N-TAD and C-TADcould discriminate between the induction of differenthypoxic target genes[2].In this model we proposed thatthe oxygen-sensitive asparaginyl hydroxylase FIH-1(fac-tor inhibiting HIF-1),which inhibits only the C-TAD,does so even at low oxygen values(1%oxygen).Thisaction therefore restricts the expression of the N-TAD/C-TAD-dependent genes but not of genes dependent onlyon the N-TAD.The bnip3gene emerged as one of the N-TAD genes that was fully expressed in moderate hypoxia(3–1%oxygen).Several studies previously characterizedBNIP3,a BH3-only protein,as a proapoptotic protein[3].We were very surprised by this because we never foundfull expression of BNIP3to be associated with cell death.Why would an apoptotic gene be fully induced by HIF inconditions where HIF induces a cell survival response?We thus questioned its function in a hypoxic tumorenvironment and demonstrated,for thefirst time,thatrapid and full induction of BNIP3could instead protectfrom cell death through an autophagic process[4 ].In2007,Tracy et al.showed that hypoxia triggered non-apoptotic cell death in mouse embryonicfibroblasts(MEFs)through BNIP3induction[5 ].The type of celldeath was determined to be autophagic.Moreover,BNIP3transcription was repressed by pRB/E2F,actingthus as a brake in modulating autophagy when cells areengaged in proliferation.As a follow-up to these results,Azad et al.confirmed that hypoxia-induced cell death wasalso autophagic in glioma and breast cancer cells and thatBNIP3was involved in autophagic cell death[6 ].Alongthis line the repression of BNIP3via SIM2-inducedhypoxia-inducible factor(HIF-1)interference has beensuggested to participate in the development of prostatecancer[7].Is BNIP3a procell death or prosurvival hypoxia-induced protein?Hypoxia-induced autophagy,in a BNIP3-de-pendent manner was confirmed later in MEFs [8 ].In addition,these authors reported a protective effect of BNIP3-induced,Beclin1-induced,and Atg5-induced autophagy in hypoxia.Moreover,HIF-1-dependent expression of BNIP3was described to be essential in mitochondrial autophagy,also called mitophagy,as described in MEFs and later in mole rat heart tissue [9](Figure 1).Clearance of damaged mitochondria by mitophagy was necessary to prevent increased levels of reactive oxygen species (ROS)[8 ].The necessity to lower the production of ROS by autophagy was later confirmed [10,11].In agreement with these studies we showed that hypoxia-induced autophagy was observed in numerous normal and cancer cells in hypoxia (1%oxygen)and in severe hypoxia (0.1%oxygen)without triggering cell death even after a week of hypoxic treatment [12 ].We must point out that cells were maintained in rich medium (high glucose)with strict control of the pH,oxygen being the only limiting factor.We then estab-lished that cells lacking expression of two HIF isoforms,HIF-1or HIF-2,elicited a lower autophagic response under hypoxic conditions.Therefore,HIF-1and HIF-2appear to be evenly matched in their capacity to induce autophagy.Moreover,as shown by Zhang et al.in MEFs,we demonstrated that hypoxia-induced autophagy is a survival process as ablation of either Beclin1,Atg5or Atg7,three major actors of autophagy,enhanced cell death in response to hypoxia.BNIP3L,a BH3-only member of the Bcl-2family also called NIX and induced by hypoxia,shows 56%sequence homology with BNIP3.Interestingly,BNIP3L was shown to trigger mitochon-drial autophagy and to be essential in the final steps of erythroid differentiation [13 ].Because of their sim-ilarities,we questioned the putative role of both BNIP3and BNIP3L in hypoxia-induced autophagy in cancer cells.Indeed,we revealed that the expression of BNIP3and BNIP3L is required for hypoxia-induced autophagy.Simple knockdown of BNIP3or BNIP3L was sufficient to partially suppress autophagy.However,the strength of the response was enhanced when dual ablation of BNIP3and BNIP3L was performed.Interestingly,we found that knockdown of BNIP3and BNIP3L increased cell death in hypoxia as did ablation of Beclin1or Atg5in the same cells.Undoubtedly,BNIP3and BNIP3L appear to be two prosurvival proteins,redefined as proautophagic proteins [14],that are required for the optimal induction of autop-hagy in hypoxia.The next step was to understand the mechanism by which BNIP3and BNIP3L are capable of triggering autophagy.Recent observations have described Beclin1as a novel atypical BH3-only protein interacting with Bcl-2or Bcl-X L [15,16 ]and thus con-trolling the balance between apoptosis and autophagy [17 ].As BNIP3and BNIP3L also possess atypical BH3-domains [18 ],we proposed a model [12 ](Figure 1)in which atypical BH3-domain containing proteins,over-expressed in hypoxia,are determinants in displacing Beclin1from Bcl-2/Beclin1or Bcl-X L /Beclin1complexes.Free Beclin1can thus induce autophagy.Although BNIP3and BNIP3L possess BH3-domains,we propose that their affinities are too low to firmly associate with Bcl-2or Bcl-X L ,a minimal requirement for triggering apop-tosis.We have therefore proposed that the ‘low affinity’BH3-domains of Beclin,BNIP3and BNIP3L,in contrast to those of Bid,Bad,and Puma have evolved to induce autophagy only [12 ].Autophagy,HIF,and the othersBNIP3and BNIP3L are not the only proteins involved in hypoxia-induced autophagy.In physiological conditions,a protein named sequestosome 1(SQSTM1/p62)plays the role of a ‘garbage man’of the cell driving the damaged or unfolded proteins to the autophagic pathway and degradation.Mathew et al .have shown that in autop-hagy-defective tumor cells,p62accumulated thereby triggering a DNA damage response and tumorigenesis2CellregulationFigure1Hypoxia-induced autophagy pathways along oxygen gradientassociated or not with metabolic stress in cells.AMPK,AMP-activated protein kinase;Beclin,Beclin-1;BNIP,Bcl-2/E1B 19kDa-interacting protein;HIF,hypoxia-inducible factor;mTOR,mammalian Target Of Rapamycin;PERK,PKR-like ER kinase;UPR,unfolded protein response.[19 ].It ensues that p62emerged recently in the hypoxic picture as thefirst protein to be degraded through autop-hagy in response to hypoxia but independent of HIF-1 [20].The clearance of p62led to the upregulation of ERK phosphorylation,which may be responsible for cell sur-vival and proliferation of tumor cells.DJ-1,also called CAP1/RS/PARK7,acts as an oncogene with many func-tions including the regulation of androgen receptor-de-pendent transcription,in sensing oxidative stress and in driving Akt-mediated cell survival.Ablation of DJ-1 impaired considerably hypoxia-induced autophagy and p62turnover in U2OS cells;however,its role has not yet been defined[21].Interestingly,inhibition of the PDGFR family of proteins also blocked hypoxia-induced autophagy in BE colorectal carcinoma cells in a HIF-1-dependent manner but in a BNIP3-independent and BNIP3L-independent manner[22].HIF-1target gene expression profiling revealed two groups of genes:genes from group I(vegfc,itgb4,pdgfb,and ptprb)required upregulation of PDGFR family proteins and were required to induce hypoxia-induced autophagy and genes from group II(bnip3,bnip3l,hig2,and pfkfb4)that did not require PDGFR family proteins.The authors suggested that a PDGFR inhibitor would trigger a more dynamic pool of HIF-1a;however,this pool of HIF-1a would be less transcriptionally active giving a decrease in expres-sion of HIF-target genes and thus hypoxia-induced autophagy.Autophagy,hypoxia but not HIFFinally,HIF-1is not always at the center of hypoxia-induced autophagic picture.Severe hypoxic conditions, often accompanied by drastic glucose and amino-acid restriction,will inhibit via HIF-independent mechanisms effecting the mTOR pathway,a central check point for autophagy.Parallel to HIF,an important player in this action is the energy sensor AMP-activated protein kinase (AMPK),which is switched on by metabolic stresses[23]. Another pathway that plays a key role in autophagic induction in drastic hypoxia(0.01%oxygen or anoxia) is the unfolded protein response(UPR)pathway.In these two stresses situations,starvation and accumulation of toxic protein aggregates,autophagy clearly represents a transient survival process that could have a fatal end.One good example of a fatal end on stimulation of autophagy is shown in the work of Turcotte et al.[24 ].They studied VHL-deficient renal cell carcinomas cells(RCCs)in which HIF-a is constitutively stable and showed that the small molecules STF-62247kills these cells by indu-cing autophagy.ConclusionsThe above brief description of‘Autophagy and HIF’points out that different types of cells do not respond identically to hypoxia.In fact some investigators have reported hypoxia-induced autophagy cell death whereas others reported,for the same cell types,hypoxia-induced cell survival.These discrepancies were particularly obvious when assigning the role of BNIP3s as prodeath or prosurvival proteins.These contradictory conclusions reflect the differences in defining‘hypoxia’(Figure1). Autophagy induced under conditions of well-controlled hypoxia with no nutrient limitation and appropriate pH control is prosurvival(top half of Figure1).Autophagy induced under severe hypoxia and concomitant to meta-bolic stress(low glucose and low pH)is often associated with cell death(bottom half of Figure1)[14].In solid tumors hypoxic gradients are highly dynamic.In the same tumor,hypoxia-induced autophagy might be like the two-faced Janus,where cell death and survival occur side by side.HIF-dependent and HIF-independent autophagy is dynamically interlinked as the tumor grows.The rapid increase in the number of research publications on autop-hagy and hypoxia shows promise in better understanding the mechanisms involved.The veil is not totally removed but we are starting to obtain a global idea of the autop-hagic picture in a hypoxic microenvironment. AcknowledgementsWe thank MC Brahimi-Horn for critical reading.The laboratory is funded by grants from the Ligue Nationale Contre le Cancer(Equipe labellise´e LNCC),the Association pour la Recherche contre le Cancer(ARC),the Agence Nationale pour la Recherche(ANR),the National Cancer Institute (INCa),and METOXIA(FP7-EU program),the Centre cassagne,the University of Nice,the CNRS and INSERM.References and recommended readingPapers of particular interest,published within the period of review, have been highlighted as:of special interestof outstanding interest1.Shintani T,Klionsky DJ:Autophagy in health and disease:adouble-edged sword.Science2004,306:990-995.2.Dayan F,Roux D,Brahimi-Horn MC,Pouyssegur J,Mazure NM:The oxygen sensor factor-inhibiting hypoxia-inducible factor-1controls expression of distinct genes through thebifunctional transcriptional character of hypoxia-induciblefactor-1alpha.Cancer Res2006,66:3688-3698.3.Willis SN,Adams JM:Life in the balance:how BH3-onlyproteins induce apoptosis.Curr Opin Cell Biol2005,17:617-625. 4.Pouyssegur J,Dayan F,Mazure NM:Hypoxia signalling incancer and approaches to enforce tumour regression.Nature 2006,441:437-443.This review introduces the concept of BNIP3as a prosurvival protein via hypoxia-induced autophagy.5.Tracy K,Dibling BC,Spike BT,Knabb JR,Schumacker P,Macleod KF:BNIP3is an RB/E2F target gene required forhypoxia-induced autophagy.Mol Cell Biol2007,17:6229-6242. This is an elegant study that identifies BNIP3as a key regulator of hypoxia-induced autophagy and suggests a new role for the RB tumor suppressor by‘repressing’BNIP3induction in proliferating cells.6.Azad MB,Chen Y,Henson ES,Cizeau J,McMillan-Ward E,Israels SJ,Gibson SB:Hypoxia induces autophagic cell death in apoptosis-competent cells through a mechanism involvingBNIP3.Autophagy2007,4:195-204.Prolonged hypoxia(<1%O2)induces autophagic cell death in apoptosis-competent cells,through a mechanism involving BNIP3.7.Farrall AL,Whitelaw ML:The HIF1alpha-inducible pro-cell deathgene BNIP3is a novel target of SIM2s repression throughcross-talk on the hypoxia response element.Oncogene2009, 28:3671-3680.Hypoxia-induced autophagy Mazure and Pouysse´gur38.Zhang H,Bosch-Marce M,Shimoda LA,Tan YS,Baek JH,Wesley JB,Gonzalez FJ,Semenza GL:Mitochondrial autophagy is an HIF-1-dependent adaptive metabolic response to hypoxia .J Biol Chem 2008,283:10892-10903.First evidence that mitochondrial autophagy induced by hypoxia is con-trolled by HIF1-dependent induction of BNIP3.9.Band M,Joel A,Hernandez A,Avivi A:Hypoxia-induced BNIP3expression and mitophagy:in vivo comparison of the rat and the hypoxia-tolerant mole rat,Spalax ehrenbergi .FASEB J 2009,23:2327-2335.10.Rouschop KM,Ramaekers CH,Schaaf MB,Keulers TG,Savelkouls KG,Lambin P,Koritzinsky M,Wouters BG:Autophagy is required during cycling hypoxia to lower production of reactive oxygen species .Radiother Oncol 2009,92:411-416.11.Rouschop KM,Wouters BG:Regulation of autophagy throughmultiple independent hypoxic signaling pathways .Curr Mol Med 2009,9:417-424.12. Bellot G,Garcia-Medina R,Gounon P,Chiche J,Roux D,Pouyssegur J,Mazure NM:Hypoxia-induced autophagy ismediated through HIF-induction of BNIP3and BNIP3L via their BH3-domains .Mol Cell Biol 2009,29:2570-2581.New insight into the mechanism of BNIP3and BNIP3L,renamed proau-tophagic proteins,in hypoxia-induced autophagy through their intrinsic low affinity BH3domain.13. Sandoval H,Thiagarajan P,Dasgupta SK,Schumacher A,Prchal JT,Chen M,Wang J:Essential role for Nix in autophagicmaturation of erythroid cells .Nature 2008,454:232-235.Mouse KO of BNIP3L (Nix)demonstrates its key role for the clearance of mitochondria via autophagy in erythroid cells,a process involving loss of mitochondrial membrane potential.14.Mazure NM,Pouyssegur J:Atypical BH3-domains of BNIP3andBNIP3L lead to autophagy in hypoxia .Autophagy 2009,5:868-869.15.Maiuri MC,Criollo A,Tasdemir E,Vicencio JM,Tajeddine N,Hickman JA,Geneste O,Kroemer G:BH3-only proteins and BH3mimetics induce autophagy by competitively disrupting the interaction between Beclin 1and Bcl-2/Bcl-X(L).Autophagy 2007,3:374-376.16. Maiuri MC,Le Toumelin G,Criollo A,Rain JC,Gautier F,Juin P,Tasdemir E,Pierron G,Troulinaki K,Tavernarakis N et al.:Functional and physical interaction between Bcl-X(L)and a BH3-like domain in Beclin-1.EMBO J 2007,26:2527-2539.This paper revealed for the first time a novel autophagy-stimulatory function of BH3-only proteins beyond their established role as apoptotic inducers.17. Pattingre S,Tassa A,Qu X,Garuti R,Liang XH,Mizushima N,Packer M,Schneider MD,Levine B:Bcl-2antiapoptotic proteinsinhibit Beclin 1-dependent autophagy .Cell 2005,122:927-939.This study demonstrates that the antiapoptotic protein,Bcl-2,Bcl-X(L),besides controlling apoptosis also keep in check autophagy by interact-ing with Beclin1.18. Certo M,Del Gaizo Moore V,Nishino M,Wei G,Korsmeyer S,Armstrong SA,Letai A:Mitochondria primed by death signalsdetermine cellular addiction to antiapoptotic BCL-2family members .Cancer Cell 2006,9:351-365.This article gives a BH3affinity profile for BH3-only proteins that corre-lates with apoptotic response.19. Mathew R,Karp CM,Beaudoin B,Vuong N,Chen G,Chen HY,Bray K,Reddy A,Bhanot G,Gelinas C et al.:Autophagysuppresses tumorigenesis through elimination of p62.Cell 2009,137:1062-1075.p62/SQSTM1upregulation observed in defective autophagy contributes directly to tumorigenesis.20.Pursiheimo JP,Rantanen K,Heikkinen PT,Johansen T,Jaakkola PM:Hypoxia-activated autophagy accelerates degradation of SQSTM1/p62.Oncogene 2009,28:334-344.21.Vasseur S,Afzal S,Tardivel-Lacombe J,Park DS,Iovanna JL,Mak TW:DJ-1/PARK7is an important mediator of hypoxia-induced cellular responses .Proc Natl Acad Sci U S A 2009,106:1111-1116.22.Wilkinson S,O’Prey J,Fricker M,Ryan KM:Hypoxia-selectivemacroautophagy and cell survival signaled by autocrine PDGFR activity .Genes Dev 2009,23:1283-1288.23.Hardie DG,Sakamoto K:AMPK:a key sensor of fuel and energystatus in skeletal muscle .Physiology (Bethesda)2006,21:48-60.24. Turcotte S,Chan DA,Sutphin PD,Hay MP,Denny WA,Giaccia AJ:A molecule targeting VHL-deficient renal cell carcinoma thatinduces autophagy .Cancer Cell 2008,14:90-102.A novel approach using small molecule that selectively induces cell death in VHL-deficient cells.4Cellregulation。