• Aucun résultat trouvé

ATR controls cellular adaptation to hypoxia through positive regulation of hypoxia-inducible factor 1 (HIF-1) expression

N/A
N/A
Protected

Academic year: 2021

Partager "ATR controls cellular adaptation to hypoxia through positive regulation of hypoxia-inducible factor 1 (HIF-1) expression"

Copied!
11
0
0

Texte intégral

(1)

HAL Id: hal-01191375

https://hal.archives-ouvertes.fr/hal-01191375

Submitted on 6 Feb 2016

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

positive regulation of hypoxia-inducible factor 1 (HIF-1)

expression

F Fallone, S Britton, L Nieto, Bernard Salles, C Muller

To cite this version:

F Fallone, S Britton, L Nieto, Bernard Salles, C Muller. ATR controls cellular adaptation to

hy-poxia through positive regulation of hyhy-poxia-inducible factor 1 (HIF-1) expression. Oncogene, Nature

Publishing Group, 2012, 32 (37), �10.1038/onc.2012.462�. �hal-01191375�

(2)

ORIGINAL ARTICLE

ATR controls cellular adaptation to hypoxia through positive

regulation of hypoxia-inducible factor 1 (HIF-1) expression

F Fallone1,2, S Britton1,2, L Nieto1,2, B Salles1,2,3and C Muller1,2

Tumor cells adaptation to severe oxygen deprivation (hypoxia) plays a major role in tumor progression. The transcription factor HIF-1 (hypoxia-inducible factor 1), whose a-subunit is stabilized under hypoxic conditions is a key component of this process. Recent studies showed that two members of the phosphoinositide 3-kinase-related kinases (PIKKs) family, ATM (ataxia telangiectasia mutated) and DNA-PK (DNA-dependent protein kinase), regulate the hypoxic-dependent accumulation of HIF-1. These proteins initiate cellular stress responses when DNA damage occurs. In addition, it has been demonstrated that extreme hypoxia induces a replicative stress resulting in regions of single-stranded DNA at stalled replication forks and the activation of ATR (ataxia telangiectasia and Rad3 related protein), another member of the PIKKs family. Here, we show that even less severe hypoxia (0.1% O2) also induces activation of ATR through replicative stress. Importantly, in using either transiently silenced ATR cells, cells

expressing an inactive form of ATR or cells exposed to an ATR inhibitor (CGK733), we demonstrate that hypoxic ATR activation positively regulates the key transcription factor HIF-1 independently of the checkpoint kinase Chk1. We show that ATR kinase activity regulates HIF-1a at the translational level and we find that the elements necessary for the regulation of HIF-1a translation are located within the coding region of HIF-1a mRNA. Finally, by using three independent cellular models, we clearly show that the loss of ATR expression and/or kinase activity results in the decrease of HIF-1 DNA binding under hypoxia and consequently affects protein expression levels of two HIF-1 target genes, GLUT-1 and CAIX. Taken together, our data show a new function for ATR in cellular adaptation to hypoxia through regulation of HIF-1a translation. Our work offers new prospect for cancer therapy using ATR inhibitors with the potential to decrease cellular adaptation in hypoxic tumors.

Oncogene advance online publication, 22 October 2012; doi:10.1038/onc.2012.462 Keywords: ATR; PIKK; hypoxia; HIF-1a regulation; translation; replicative stress

INTRODUCTION

Hypoxic areas are frequently found in human solid tumors as a result of morphologically and functionally inappropriate vascular-ization, irregular blood flow, anemia and high oxygen consump-tion of rapidly proliferating cells.1,2 A large body of clinical

evidence suggests that intra-tumoral hypoxia correlates with the elevated aggressive behavior of cancer cells and their resistance to therapy, leading to poor patient prognoses.3 Tumor cells must

adapt to hypoxic stress through alterations of cellular metabolism and stimulation of neovascularization. A key regulator of the cellular response to oxygen deprivation is the transcription factor hypoxia-inducible factor 1 (HIF-1) whose accumulation results in the induction of a plethora of target genes that collectively confer cellular adaptation to hypoxia.1 HIF-1 is comprised of a labile oxygen-regulated a-subunit mainly targeted for normoxia-dependent degradation by the proteasomal system, whereas its b-subunit is constitutively expressed in most cells.1,4,5 Therefore, HIF-1 activity is dependent upon the limiting expression of a-subunit. Under hypoxia, HIF-1a is stabilized, enters the nucleus and heterodimerizes with HIF-1b. The heterodimer HIF-1 binds responsive elements to transactivate a variety of hypoxia-responsive genes, therefore contributing to the adaptative response to hypoxic conditions.1

HIF-1 is overexpressed in many human cancers and correlates with poor prognosis outcome.6 In most of these cases,

over-expression is due to the stabilization of the HIF-1a protein by hypoxia. However, there is an increasing body of evidence demonstrating that a number of non-hypoxic stimuli such as genetic alterations that activate oncogenes and inactivate tumor suppressor genes or growth factors signaling are also highly capable of turning on this transcription factor.5Accordingly, HIF-1

represents an attractive, yet challenging, target for the development of pharmacological inhibitors.5,7 We and others

have recently demonstrated that the hypoxic accumulation of HIF-1 is regulated by the DNA-dependent protein kinase (DNA-PK) and ataxia telangiectasia mutated (ATM), two members of the phosphoinositide 3-kinase-related kinases (PIKKs) family,8,9that in eukaryotes initiate cellular stress responses when genome integrity, mRNA translation, or nutrient availability is compromised.10 We have shown that DNA-PK, a key player of DNA double-strand breaks (DSBs) repair by the non-homologous end-joining pathway, is activated by mild hypoxia conditions (o1% O2).8Activated DNA-PK

positively regulates nuclear HIF-1a accumulation and subsequent levels of expression of HIF-1 target genes.8 More severe hypoxia (o0.1% O2) also activates ATM,11,12an essential building block of

the DNA DSBs signaling pathway.13It has been recently shown that ATM also increases HIF-1a expression and HIF-1-dependent transcription under hypoxic conditions.9

Interestingly, the mechanism of hypoxia-induced activation of both DNA-PK and ATM is distinct from that of the DNA DSBs and is

1

CNRS, IPBS (Institut de Pharmacologie et de Biologie Structurale), Toulouse, France and2

Universite´ de Toulouse, UPS, IPBS, Toulouse, France. Correspondence: Professor C Muller, Biology of Cancer, IPBS, CNRS UMR5089, 205 route de Narbonne, BP 6418231077 Toulouse Cedex, France.

E-mail: muller@ipbs.fr

3

Current address: TOXALIM, INRA, 180 chemin de Tournefeuille, 31027 Toulouse, France. Received 15 March 2012; revised 20 July 2012; accepted 24 August 2012

Oncogene (2012), 1–10

&2012 Macmillan Publishers Limited All rights reserved 0950-9232/12

(3)

likely to involve modification of higher-order chromatin structure and chromatin-remodeling complexes.8,9,12 ATR (for ataxia telangiectasia and Rad3 related kinase), another member of the PIKKs family, is also involved in DNA damage signaling.14 ATR controls a pathway that is mainly activated by agents that interfere with the functions of replication forks such as hydroxyurea or ultraviolet light15 but also responds to DSBs in an ATM-dependent manner.16–19Hypoxia itself is also known to induce a replication-associated damage response. For example, severe hypoxia induces S-phase arrest resulting in regions of single-stranded DNA (ssDNA) at stalled replication forks and the activation of ATR.20,21 Loss of ATR results in a further loss of viability in S-phase cells under hypoxic conditions.22In accordance

to the emerging role of PIKKs in hypoxic adaptation, our experiments are designed to determine a potential role of ATR in the regulation of HIF-1 expression and function. Here we demonstrate that, under severe hypoxic conditions (0.1% O2), ATR

is activated at early time points and that this cellular stress response favors hypoxia adaptation by regulating HIF-1a protein synthesis. In addition to ATM and DNA-PK, our data emphasize the new and important role of ATR in the regulation of HIF-1 hypoxic accumulation.

RESULTS

ATR is activated through replicative stress in hypoxic cells and regulates HIF-1 expression and function

To investigate if ATR could regulate the expression of HIF-1a upon hypoxia, we first evaluated in U2OS cells the effect of ATR

knockdown using two independent small interfering RNAs (siRNAs), achieving a knockdown of around 80% of ATR expression (Figure 1a). Transfected cells were placed in a 21% (normoxia) or 0.1% (hypoxia) oxygen atmosphere for 6 h. As a control of ATR activation, the level of Chk1 (checkpoint kinase 1) phosphorylation on serine 345 was monitored, as this site is a well-known ATR target.23 As previously described for extreme hypoxia (o0.02% O2),21we detected a specific increase of the phosphorylated Chk1

in cells exposed to 0.1% O2, confirming ATR activation in our

experimental conditions. In hypoxic cells, Chk1 phosphorylation was completely abrogated by ATR knockdown compared with control RNAi. Interestingly, ATR silencing almost completely abrogated HIF-1a protein accumulation upon hypoxia (490% decrease for si1 and si2). In U2OS cells, the b-subunit of HIF-1 is induced by hypoxia. This result was rather unexpected since HIF-1b is generally considered as a constitutively express protein that is not regulated by oxygen levels in most cells.1,4,5 The molecular mechanism underlying the induction of HIF-1b in this specific model remains at the present time uncharacterized. Under hypoxic conditions, ATR knockdown was also paralleled by a significantly lower HIF-1b content (around 80% for si1 and 90% for si2). Taken together, these results emphasize that ATR is activated by severe hypoxic conditions and participates to HIF-1 accumulation under hypoxia.

We then investigate if the kinase activity of ATR was required for the observed effect on HIF-1. Since no specific pharmacological ATR inhibitor was available, we used CGK733, a small molecule that inhibits both ATM and ATR kinase activities and blocks their checkpoint signaling pathways.24In parallel, a specific inhibitor of

21% 0.1% O2 Ctrl 1 2 Ctrl 1 2 si ATR 21% 0.1% O2 KU55933 21% 0.1% O2 ATR HIF-1α HIF-1β 1 0.1 0.05 0.03 0 0 1 0.2 0.06 CGK733 - + - - + -- - + - - + - + - - + -- - + - - + KU55933 1 0.83 0.28 0.6 0.6 0.53 1 0.65 0.12 0.1 0.01 0 1 0.85 0.26 0.12 0.13 0.17 1 0.5 0.08 α-Tubulin Chk1-P Chk1 Chk1-P Chk1 Chk1 Chk1-P 0.1% O2 CGK733 si Ctr + + - -- - + + - + - + si Chk1 ATR - + - + - + - + - + 0.1 0.2 1 0.5 1.7 1.1 0.6 0.1 0.9 0.9 1 0.1 0.9 0.6 1.3 0.9 + + + + + -- 2 4 6 12 h WT KD Chk1 HIF-1β α-Tubulin HIF-1α (long exp.) CGK733 HIF-1α HIF-1β α-Tubulin 0.1% O2 HIF1α HIF1β α-Tubulin HIF1α HIF1β α-Tubulin 1 0.2 0.96 0.1 1 0.47 0.97 0.42

Figure 1. Inhibition of ATR expression and activity down-regulates hypoxic accumulation of HIF-1 in a chk1-independent manner. (a) U2OS cells transfected by control (Ctrl) or anti-ATR (1 or 2) siRNAs (si) were placed under normoxic (21% O2) or hypoxic (0.1% O2) conditions for 6 h.

(b) U2OS cells were pre-treated with KU55933 or CGK733 as specified and placed in normoxic or hypoxic conditions for 6 h. (c) HeLa cells were pre-treated with KU55933 or CGK733 when specified and placed in normoxic or hypoxic conditions for 6 h. (d) WT or dominant negative KD ATR was induced in U2OS stable clones and cells were placed in normoxia or incubated for the indicated time under hypoxic conditions. (e) U2OS cells transfected by control (Ctrl) or anti-Chk1 siRNA (si) were treated or not for 1 h with CGK733 and placed under hypoxic conditions for 6 h. Protein levels were analyzed by western blots using the indicated antibodies. The relative levels of HIF-1a or HIF-1b expression compared with controls are indicated, after normalization with a-tubulin expression levels. (a–c) For the quantification of HIF-1a and b, hypoxic cells transfected with control siRNA or untreated cells are used as a control and arbitrarily set at 1. (d) The levels of protein expression after 4 h of hypoxia is used as a control whereas the hypoxic cells transfected with control siRNA are used in (e).

(4)

ATM, KU55933 was used as a control.25Exposure to CGK733, but not KU55933, was able to inhibit Chk1 phosphorylation upon hypoxic conditions confirming the absence of effect of KU55933 on ATR activity (Figure 1b). A large decrease of HIF-1a and HIF-1b hypoxic accumulation was observed in cells treated with CGK733, whereas only a slight effect was observed with KU55933 treatment. Similar results were obtained in HeLa cells (Figure 1c). We then used ATR-inducible cell lines expressing either a catalytically inactive (kinase-dead, KD) mutant or a wild-type (WT) form of the protein.26 From 4 h of hypoxia, the accumulation of HIF-1a was dramatically reduced in ATR KD cells compared with WT cells, confirming the involvement of the kinase activity of ATR in the observed effect (Figure 1d). To investigate if our results could be mediated by the ATR-activated Chk1 kinase, Chk1 expression was knocked-down using RNAi. In conditions where Chk1 was efficiently silenced (490% decrease in Chk1 expression) (Figure 1e), the hypoxic levels of HIF-1a and b proteins were similar to control and the down-regulation of both HIF-1 subunits in the presence of CGK733 was still observed in the Chk1-depleted cells. Therefore, our compelling results highlight that activated ATR is able to regulate hypoxic HIF-1 accumulation in a Chk1-independent manner.

We next performed in HeLa cells time-course experiments to address the kinetics of CGK733-mediated inhibition of HIF-1a protein accumulation. CGK733 decreased the hypoxic-dependent accumulation of HIF-1a protein by about 50% after 60 min of treatment as compared with untreated controls, with almost complete abrogation of its expression observed after 2 h (Figure 2a). The HIF-1b protein levels decreased only after 2 h of treatment (expression levels reduced to about 25% of the untreated control) (Figure 2a) with no further additional effect for longer incubation times (data not shown). We then examined the dose-dependent effect of CGK733 on HIF-1a protein accumula-tion. CGK733 decreased hypoxic accumulation of HIFa protein in a dose-dependent manner with an IC50of 3.1 mM(Figure 2b). Again,

the effect on HIF-1b was less pronounced with an IC50of 18 mM.

Therefore, these results demonstrate that the ATR-dependent inhibition of HIF-1 is an early event occurring during the first hours of incubation in hypoxic conditions with a more rapid and pronounced effect on the a-subunit of the complex.

We then investigated the signal that triggered ATR activation in these hypoxic conditions. Using a nuclear fractionation technique, we have previously demonstrated that hypoxia triggered loading of DNA-PK in a chromatin detergent-resistant fraction that also contained the HIF-1 subunits.8 Similar ATR relocalization in detergent-resistant fractions has been shown to occur after exposure to drugs such as hydroxyurea or aphidicolin that

induce rapid replication arrest.17 In accordance with our

previous work,8 we observed here that activated DNA-PK was

recruited to the extraction-resistant fraction FIV in hypoxic cells as compared with cells grown in normoxia (Figure 3a). It is noteworthy that, in our conditions, the phosphorylated form of the histone variant H2AX (g-H2AX) and the chromatin-associated protein HP1a were only present in the fraction FIV assessing the quality of subfractionation, with a significant increase of g-H2AX expression in hypoxic cells as previously described.8 In hypoxic cells, ATR was redistributed onto the latest fractions (FIII and FIV) along with its partner ATRIP (ATR-interacting protein), reflecting its chromatin loading.17Interestingly, this recruitment in hypoxic cells occurs along with the recruitment of the ssDNA-binding protein RPA (replication protein A). These results are in accordance with several observations that have highlighted the importance of RPA-coated ssDNA in ATR activation.27Finally, the HIF-1 subunits were mainly associated with the extraction-resistant fraction FIV as previously shown.8 These results demonstrate that our experimental conditions lead to an early replication stress that triggers activation of ATR. Analysis of the cell-cycle distribution showed that cells exposure to low oxygen concentration (0.1%) induces at 12 h a cell-cycle arrest marked by a large decrease in G2/M (from 13.5 to 5.6%) and an accumulation of cells in G1-phase (from 55.2 to 69.3%) as compared with cells grown in normoxia (Supplementary Figure 1). CGK733, but by neither KU55933 nor NU7026 (a DNA-PK-specific inhibitor),28,29 significantly affected this cell-cycle arrest (increase in G2/M from 5.6 to 15.3% and decrease in G1/S from 69.3 to 55% in CGK733-treated cells as compared with untreated cells) confirming the existence of a replicative stress in these cells whose signaling is regulated by ATR. To reinforce the hypothesis that replication stress was involved in the ATR-mediated regulation of HIF-1, we used aphidicolin, a known inhibitor of DNA replication.30 Cells were treated with CGK733 in the presence or not of aphidicolin and then cultured under hypoxic conditions. Aphidicolin treatment in combination with CGK733 enhances the decrease of HIF-1 expression observed in the presence of CGK733 alone (Figure 3b). Taken together, using different approaches we demonstrated that ATR is activated by a replicative stress under our hypoxic conditions, and that this activation triggers the control of HIF-1 accumulation.

21% 0.1% Time (min) HIF-1α HIF-1β β-Actin - - + + + + + - 0 20 40 60 90 120 CGK733 1 1 1 0.53 0.2 0.14 1.83 1.77 1.8 1.2 0.4 - - 2.5 5 10 20 21% 0.1% CGK733 (μM) O2 O2 HIF-1α HIF-1β β-Actin 1 0.6 1 0.84 0.7 0.45 0.16 0.13 0.1 1 1

Figure 2. CGK733 inhibits hypoxic accumulation of HIF-1 subunits in a time- and dose-dependent manner. (a) HeLa cells were placed in normoxia or hypoxia during 3 h. Cells placed in hypoxia were then treated or not with CGK733. At indicated times, cells extracts were prepared and subjected to immunoblot assays. (b) HeLa cells were placed in normoxia or treated for 6 h in hypoxic condition in the presence or absence of increasing concentrations of CGK733. Protein levels were analyzed by western blots using the indicated antibodies. The level of HIF-1a and HIF-1b expression in hypoxia in treated samples compared with controls (arbitrarily set at 1) is indicated, after normalization with b-actin expression levels.

0.1% O2 FI FIII FIV ATRIP RPA70 γH2AX HP1α ATR HIF-1β HIF-1α DNAPKcs-P - + - + - + HIF-1α HIF-1β CGK733 - + Aph. - + - + - + - - + + - + - + 21% 0.1% O2 1 0.3 0.97 0.05 1 1 0.8 0.4 1.4 0.7 0.85 0.1 α-Tubulin α-Tubulin

Figure 3. Replicative stress induced by hypoxia is involved in ATR activation and HIF-1-mediated regulation. (a) U2OS cells incubated under normoxic ( ) or hypoxic ( þ ) conditions were fractionated and aliquots of the FI, FIII and FIV were separated on SDS–PAGE gels and blotted for the indicated antibodies. (b) HeLa cells were incubated under normoxic or hypoxic conditions during 6 h in the presence or absence of aphidicolin (Aph) and CGK733. Protein levels were analyzed by western blots using the indicated antibodies. The levels of HIF-1a and HIF-1b expression in treated samples compared with controls (arbitrarily set at 1) are indicated, after normalization with a-tubulin expression levels. For HIF-1a, the hypoxic untreated cells are used as a control whereas untreated normoxic cells are used as a control of HIF-1b expression.

ATR regulates HIF-1 expression F Fallone et al

3

(5)

ATR regulates the translation of HIF-1a under hypoxic conditions Our first set of results demonstrates that ATR activity is involved in the regulation of both HIF-1 subunits with a more pronounced effect on the labile HIF-1a subunit. As HIF-1a protein is degraded mainly through the ubiquitin/proteasome pathway, experiments were performed to evaluate if activated ATR has the ability to interfere with the mechanism(s) responsible for HIF-1a protein stabilization and degradation. In hypoxia, CGK733, as well as KU55933, inhibited the accumulation of HIF-1a protein despite the presence of the proteasome inhibitor MG132, whereas in similar conditions, the difference in HIF-1a accumulation between NU7026-treated cells versus untreated cells were no longer observed in accordance with our recent published data8 (Figure 4a, upper panel). Note that MG132 treatment leads to a clear increase in b-catenin expression as expected.26Similar results were obtained with another proteasome inhibitor, lactacystin (data not shown). In addition, we assessed that siRNAs against ATR were still efficient to inhibit HIF-1a expression in the presence of MG132 (Figure 4a, lower panel). To confirm that ATR does not regulate HIF-1a through its degradation, DFX (desferrioxamine)-treated cells were exposed to the protein synthesis inhibitor,

cycloheximide (CHX) to block further HIF-1a mRNA translation. The kinetics of HIF-1a decay in CGK733-treated and untreated cells was compared by immunoblotting and no differences were observed between these two conditions (Figure 4b, upper panel). Similarly, no differences were observed when the HIF-1a degradation was compared between ATR WT versus KD cells (Figure 4b, lower panel). Taken together, this set of experiments excludes a role for ATR in the control of HIF-1a stability and supports its involvement in the control of HIF-1a biosynthesis.

Under hypoxia no differences in HIF-1a and HIF-1b mRNA levels were observed between ATR-deficient and -proficient cells as seen by using CGK733-treated cells or the ATR WT/KD cellular model (see Supplementary Figure 2). These results strongly suggest that ATR-mediated regulation of HIF-1 expression occurs at the translational level. Consistently, inhibition of ATR expression (using ATR siRNA) and activity (CGK733-treated cells and ATR KD cells) decreased de novo synthesis of HIF-1a, as supported by a twofold reduction of 35S-methionine/cysteine incorporation into HIF-1a in ATR-deficient cells as compared with controls (Figure 4c). Note that in these conditions, ATR depletion or inhibition was without effect on the translation of b-catenin used as a control (Figure 4c). Taken together, these compelling results demonstrate

MG132 - + β-Catenin α-Tubulin HIF-1α NU7026 CGK733 - -KU55933 + - - + - -- - + - - - + -- - - + - - - + α-Tubulin WT KD ATR CHX (h) DFX HIF-1α 1 0.9 0.7 0.54 1 0.8 0.72 0.56 0 0.5 1 1.5 0 0.5 1 1.5 HIF1α 0 0.5 1 1.5 0 0.5 1 1.5 1 0.79 0.5 0.23 1 0.87 0.48 0.19 CHX CHX + CGK α-Tubulin time (h) DFX IP HIF-1α KU55933 0.1% O2 CGK733 - + - -+ IP β-Catenin 0.1% O2 WT KD HIF-1α β-Catenin ATR 1 0.96 0.38 1 0.51 0.1% O2 si Ctrl ATR1 1 0.53 1 1.15 1 1.09 1 1.09 0.86 MG132 - + α-Tubulin HIF-1α 0.1 % O2 NU7026 Ctrl 1 2 Ctrl 1 2 si ATR - -- - - + - - - + 0.1 % O2

Figure 4. Regulation of HIF-1a by ATR occurs at the translational level. (a) Upper panel: HeLa cells were treated in the presence (þ ) or absence ( ) of MG132 (20 mM) for 1 h, then treated or not with the indicated inhibitors at 20 mMand placed in hypoxia during 1 h. Protein levels were

analyzed by western blots using the indicated antibodies. Lower panel: Similar experiments were performed with U2OS cells transfected with control (Ctrl) or two siRNAs directed against ATR (si 1 and 2), (b) Upper panel: HeLa cells were treated for 2 h with DFX (260 mM). After 2 h (T0),

cells were exposed to CHX to block protein biosynthesis in the presence (þ ) or in the absence (  ) of CGK733 and extracts were prepared at the indicated times. The relative levels of HIF-1a compared with T0 (set at 1) for each condition are indicated, after normalization with a-tubulin expression levels. Lower panel: similar experiments were performed with ATR WT and KD cells. Note that a twofold decrease in HIF-1a expression is observed as expected at T0 in KD compared with WT cells. (c) After metabolic labeling as indicated in Materials and methods, total extracts obtained were immunoprecipitated with the indicated antibody. Quantifications of incorporated radioactivity relative to the corresponding control samples are indicated.

(6)

that inhibition of ATR reduces HIF-1 translation without affecting its mRNA expression or proteasome degradation.

Translation can be regulated at multiple levels and several mechanisms have already been described for HIF-1a.31–33 The

PI3K–AKT–mTOR (mammalian target of rapamycin) pathway has been implicated in the regulation of HIF-1a protein at a translational level.34,35mTOR is able to stimulate translation by

alleviating the inhibitory effect of 4E-BP1 on translation initiation,36through the phosphorylation on threonine 37/46. To test the involvement of mTOR, 4E-BP1 T37/46 phosphorylation was monitored in HeLa cells under normoxic or hypoxic conditions using the PI3K inhibitor LY294002 as a control (Supplementary Figure 3). While LY294002 inhibited hypoxia-induced HIF-1a protein accumulation as described37 and the phosphorylation of 4E-BP1 on threonine 37/46, CGK733 inhibited HIF-1a accumulation but did not affect the phosphorylation of 4E-BP1, in both normoxia and hypoxia (Supplementary Figure 3). These results demonstrate that the effect of CGK733 on HIF-1a accumulation is independent of the PI3K–AKT–mTOR pathway.

A way to determine what is the mechanism regulating HIF-1a translation is to define the region of HIF-1a mRNA that confers regulation by ATR. In fact, HIF-1a 50-untranslated region (UTR) contains an internal ribosome entry site (IRES) that has been reported to mediate its translation during hypoxia38and 30UTR are known to be targeted by miRNA.39 To define the HIF-1a mRNA

regions conferring regulation by ATR, HIF-1a was transiently expressed under the control of a strong constitutive CMV (cytomegalovirus) promoter either as full-length cDNA, containing 50UTR and 30UTR (50ORF30), or just its open reading frame (ORF). All the constructions used are summarized in Figure 5a. Results show that exposure of cells to the ATR inhibitor CGK733 is able to efficiently down-regulate the CMV-driven expression of both HIF-1a full-length cDNA containing HIF-1a ORF (ORF) as well as 50and 30UTR (50ORF30) (Figure 5b). A similar regulation is observed when HIF-1a ORF is fused to the GFP coding sequence (ORF–GFP). Furthermore, CGK733 exposure was unable to down-regulate the 50UTR or the 30UTR fused to GFP (50UTR–GFP, 30UTR–GFP) (Figure 5c). Finally, siRNA-mediated

depletion of ATR was showing a similar decrease of the expression of HIF-1a ORF–GFP as compared with control siRNA (Figure 5d). Taken together, our results show that ATR expression and kinase activity regulates HIF-1a translation during hypoxia through a region contained in HIF-1a ORF.

ATR kinase activity is necessary for cellular adaptation to hypoxia We then checked if ATR expression and activity are involved in the cell adaptive response to hypoxia by using three different cell models: ATR WT versus KD cells, U2OS treated or not with CGK733 and U2OS transfected with siRNAs targeted against ATR versus control siRNA. Our previous results (see Figure 2a) show that, in hypoxic cells, HIF-1 is mainly present in the detergent-resistant fraction (FIV) of the extracts that is thought to be the chromatin-associated and therefore active form of the complex. We first demonstrated that exposure to CGK733 dramatically reduces the expression of both HIF-1 subunits in the detergent-resistant fraction FIV (Figure 6a). This decrease is observed, but to a lesser extent after treatment with KU55933. The hypoxic expression of HIF-1 in the FIV fraction of ATR KD extracts was largely reduced as compared with similar extracts obtained from WT cells (Figure 6b) and similar results were obtained in ATR-silenced cells as compared with control (Figure 6c). We then evaluated the expression of two HIF-1 target genes involved in the metabolic adaptation to hypoxia, the glucose transporter GLUT-1 and the carbonic anhydrase IX (CAIX), a membrane spanning protein involved in the enzymatic regulation of tumor acid-base balance.40,41By using U2OS treated or not with CGK733, ATR WT versus KD cells and U2OS transfected with targeted siRNAs against ATR versus control siRNA, we clearly show that the mRNA expression levels of GLUT-1 and CAIX were significantly higher in hypoxic ATR-proficient than -deficient cells (Figures 6d–f). Note that, in these conditions, we also observed a decrease in GLUT-1 and CAIX expression in KU55933-treated cells (Figure 6d).

These results were also confirmed at proteins levels. In WT cells, the expression of both GLUT-1 and CAIX was induced after 12 h of hypoxia and maintained thereafter (Figure 6g). This induction was

5’UTR-GFP 3’UTR-GFP 5’ORF3’ ORF ORF-GFP α-Tubulin 3’UTR-GFP 5’UTR-GFP - + - + CGK733 GFP U2OS CGK733

5’ORF3’ ORF ORF-GFP

short exp. HIF-1α HIF-1α–GFP HIF-1β α-Tubulin 1 2 3 4 5 6 7 - - + - + - + Ctr l α-Tubulin GFP ORF-GFP siRNA : HIF-1α–GFP Ctrl Ctrl ATR

Figure 5. Translational regulation of HIF-1a by ATR involves elements located in the ORF region of HIF-1a mRNA. (a) Schematic representation of mRNAs expressed by the plasmids used. The spheres correspond to the mRNA cap. (b) U2OS cells were transiently transfected with HIF-1a full-length cDNA containing the ORF and both UTRs (50ORF30), or with the ORF alone (ORF) or fused with GFP (ORF–GFP). Forty eight hours

after transfection, cells were treated when indicated with CGK733 for 6 h. As control of HIF-1a endogenous expression, untransfected cells were used (Ctrl). (c) U2OS cells were transfected with plasmids containing 50- or 30UTR fused to GFP (50UTR–GFP and 30UTR–GFP). Forty-eight

hours after transfection, cells were treated when indicated with CGK733 for 6 h. GFP expression level was analyzed in whole-cell extracts by immunoblotting. (d) Expression of the HIF-1a ORF fused with GFP in U2OS cells transfected with control (Ctrl) or anti-ATR siRNAs.

ATR regulates HIF-1 expression F Fallone et al

5

(7)

largely decreased in the presence of an inactive form of ATR (KD). Similarly, the expression of these two proteins was reduced in cells treated for 24 h with CGK733 (Figure 6h). Finally, ATR silencing using siRNAs almost completely abrogated GLUT-1 and CAIX accumulation in cells exposed for hypoxia during 12 h (Figure 6i).

Finally, we observed a specific decrease in cell survival upon hypoxia in cells where the ATR expression and activity was inhibited (see Supplementary Figure 4). Therefore, our compelling results highlight that ATR kinase activity is able to regulate HIF-1 accumulation and to contribute to an efficient hypoxic adaptation.

21% 0.1% O2 *** CAIX *** 21% 0.1% O2 ** GLUT-1 ** KU55933 CGK733 Control * 21% 0.1% O2 CAIX WTKD 21% 0.1% O2 * GLUT-1 HP1α HIF-1β HIF-1α 21% 0.1% O2 CGK733 - -KU55933 + - + -- - + - - + FIV 1 0.52 ND α-Tubulin CAIX GLUT-1 21% 0.1% O2 Ctrl 1 2 Ctrl 1 2 siATR GLUT-1 - 2 4 6 12 24 h KD WT KD WT KD WT KD WT KD WT KD CAIX 0.1% O2 WT α-Tubulin ATR WT KD WT KD WT KD WT KD - 2 4 6 h 0.1% O2 HP1α HIF-1β HIF-1α FIV ATR 10.3 4.31.2 2.6 0.8 HP1α HIF-1β HIF-1α 21% 0.1% O2 FIV Ctrl 1 2 Ctrl 1 2 siATR 1 ND ND 21% 0.1% O2 ** CAIX siCtr siATR1 ** GLUT-1 21% 0.1% O2 CAIX 21 GLUT-1 0.1 CGK (μM) - - - 20 - - 2.5 5 10 20 -KU (μM) α-Tubulin O2 (%)

Figure 6. ATR expression and activity controls cell adaptation to hypoxia. Cells were fractionated and aliquots of the fractions IV (FIV) were separated on SDS–PAGE gels and blotted for the indicated antibodies. (a) U2OS cells were placed under normoxic or hypoxic conditions for 2 h. When indicated, the cells were pre-treated (þ ) or not (  ) with 20 mMKU55933 or CGK733. (b) U2OS cells expressing WT or KD ATR were

placed under normoxic or hypoxic conditions for the indicated times. (c) U2OS cells transfected by control (Ctrl) or two different siRNA directed against ATR (siATR 1 or 2) were placed under normoxic (21% O2) or hypoxic (0.1% O2) conditions for 2 h. In these experiments (a–c),

the level of hypoxic HIF-1a expression in ATR-deficient samples compared with controls (set at 1) is indicated, after normalization with the expression levels of the chromatin-associated protein HP1a, ND not detectable (d) U2OS cells were pre-treated in the presence or absence of 20 mMKU55933 and CGK733 then placed under normoxic or hypoxic conditions for 3 h. CAIX and GLUT-1 mRNA expression were measured by qPCR. (e) U2OS cells expressing WT or KD ATR were placed under normoxic or hypoxic conditions for 6 h and CAIX and GLUT-1 mRNA expression were measured by qPCR. (f ) U2OS cells transfected by control (siCtrl) or anti-ATR siRNA (siATR1) were placed under normoxic (21% O2) or hypoxic (0.1% O2) conditions for 3 h. CAIX and GLUT-1 mRNA expression were measured by qPCR. (d–f ) Bars and errors flags represent

the means±s.d. of at least three independent experiments; *statistically significant by Student’s t-test, Po0.05 relative to control **Po0.01 relative to control, ***Po0.001 relative to control. (g) U2OS expressing WT or KD ATR were placed under normoxic or hypoxic conditions for the indicated times and GLUT-1 and CAIX expression levels were analyzed in whole-cell extracts by immunoblotting. (h) U2OS cells were placed under normoxic or hypoxic conditions for 24 h. When indicated, the cells were pre-treated with 20 mMKU55933 (KU) or CGK733 (CGK) at the indicated concentrations. GLUT-1 and CAIX expression levels were analyzed in whole-cell extracts by immunoblotting. (i) U2OS cells transfected by control (Ctrl) or anti-ATR siRNAs (siATR 1 or 2) were placed under normoxic (21% O2) or hypoxic (0.1% O2) conditions for 12 h.

GLUT-1 and CAIX expression levels were analyzed in whole-cell extracts by immunoblotting. 6

(8)

DISCUSSION

The DDR (DNA damage response) induced during hypoxic stress has recently emerged as an important signaling pathway that allows cells to withstand modifications in environmental condi-tions. Recent studies showed that two members of the PIKKs family involved in DNA DSBs signaling and repair, ATM and DNA-PK regulate the hypoxic-dependent accumulation and degrada-tion of HIF-1a.2,8,9,42,43The histone variant H2AX, a target of PIKKs that participates in DDR and repair, has also been identified as a crucial actor for hypoxia-driven neovascularization.44 We show

here that another member of the PIKKs superfamily, ATR, activated by severe hypoxia (0.1% O2) through replicative stress, regulates

the expression of both HIF-1 subunits, therefore contributing to cell adaptation to hypoxia. Interestingly, a functional ATR was required to observe this effect that involves, for the a-subunit of HIF-1, the regulation of its translation. The involvement of ATR kinase activity in this novel regulation pathway raises the possibility of using ATR inhibitors to eradicate hypoxic cells during cancer therapy.

It has been previously demonstrated that the replicative stress induced by extreme hypoxia (from 0.02 too0.1% O2) leads to the

activation of ATR.20,22,45In addition to these reports, our current data show that ATR is activated under less severe hypoxic conditions (0.1% O2) in a similar replication-dependent manner.

Indeed, we have shown that exposure of cells to these hypoxic conditions induces a cell-cycle arrest as well the phosphorylation of Chk1, both events being dependent of activated ATR (see Supplementary Figure 1 and Figures 1a–c). In hypoxic cells, relocalization of ATR, concomitantly with its partner ATRIP and the ssDNA-binding protein RPA, into chromatin-associated fractions outlined the existence of a replicative stress (Figure 3a).

What are the cellular end points of ATR activation in hypoxic cells? Recent study demonstrates that hypoxic activation of ATR is responsible for the activation of the DNA replication initiation checkpoint45but the downstream events, independently of

cell-cycle arrest, remain in fact poorly characterized. We report here for the first time that expression and activity of ATR positively regulates HIF-1 expression under hypoxia in different cellular models suggesting a canonical pathway (see Figure 1). These results were obtained in cells where ATR expression was abrogated as well as in cells expressing an inactive form of the protein. Finally, the use of an ATR inhibitor, CGK733 reproduces this effect. We also showed that the inactivation of ATR through expression of a KD form of the protein or pharmacological inhibition of its kinase activity, as well as inhibition of its expression by siRNA, negatively regulates HIF-1 trans-activating function through expression of two of its target gene GLUT-1 and CAIX (see Figures 6d–i). Inhibition of ATR expression and activity also inhibited cell survival upon hypoxic conditions (Supplementary Figure 4). Interestingly, another work has recently uncovered a potential link between replicative stress and cellular adaptation to hypoxia in the specific model of endothelial cells.44

In fact, this study demonstrated that hypoxia induces replication-associated generation of g-H2AX (the phosphorylated form of the histone variant H2AX) in endothelial cells in vitro and in mice.44 Moreover, endothelial-specific H2AX deletion resulted in reduced hypoxia-driven retina and tumor neovascularization. Therefore, while not directly demonstrating the role of HIF-1 in this process, this work underlined that activation of a replicative stress response (through H2AX phosphorylation) is crucial for hypoxia-driven neovascularization. According to these results, including ours, we propose a model whereas severe hypoxia induces a replicative stress, initiating ATR activation and contributing to an adaptive response, through regulation of HIF-1 expression, to counterbalance oxygen deprivation. The role of ATR in maintaining a suitable level of HIF-1 in cells exposed to severe hypoxia explains why inhibiting its function or expression

contributes to down-regulate HIF-1 expression. Accordingly, in cells exhibiting a high level of replicative stress (such as the hypoxic cells exposed to aphidicolin, a known inhibitor of DNA replication, Figure 3b), the inhibition of ATR is a very efficient approach to inhibit HIF-1 expression. These results are in apparent contradiction with the study of Rapisarda and Melillo46showing that HIF-1a accumulation is not further inhibited when ATR-deficient cells are exposed to UVC, a DNA-damaging agent known to interfere with DNA replication. However, in this study the authors showed that exposure to UVC alone inhibit HIF-1a translation.46 This inhibition is dissociated from the induction of DNA lesions, suggesting the involvement of a specific pathway independent of an intact sensing of DNA damage, in contrast to our study.46Finally, it is interesting to note that no further increase in HIF-1 expression level is observed when hypoxic ATR-proficient cells are exposed to aphidicolin (see Figure 3b). HIF-1 protein level is tightly regulated1,4,5,47 and one might speculate that some

compensatory mechanisms exist limiting the level of expression of HIF-1 to a given range. This hypothesis could explain why increasing the replicative stress above a certain threshold do not contribute to further increase the accumulation of the protein.

Although we show that activation of ATR affects the hypoxic expression of both HIF-1 subunits, the expression of the labile a-subunit appears to be more sensitive to ATR inhibition as emphasized in the time and dose experiments with CGK733 (Figure 2). In contrast to most of the known regulation mechanisms of HIF-1a hypoxic accumulation, which occur at the level of HIF-1a protein degradation, we found that the ATR-dependent regulation of HIF-1a accumulation during hypoxia occurs at the stage of HIF-1a translation. Up-regulation of translation through the PI3K–AKT–mTOR pathway is an important mechanism of HIF-1a-protein accumulation induced by growth factors under aerobic conditions.34,48However, the fact that ATR inhibition was without effect on the amount of phosphorylated 4E-BP1, a downstream target of mTOR does not favor this hypothesis (Figure 5a). Other studies have reported a negative regulation of HIF-1a translation. The translation efficiency of individual genes is thought to be regulated largely by the 50- and 30-untranslated parts of the mRNA. It has been reported that HIF-1a protein translation can persist during hypoxia, owing to an IRES in the 50UTR of its mRNA38

whose activity is regulated by IRES trans-acting factors.32,49We demonstrated here that inhibition of ATR expression and activity was able to efficiently down-regulate the expression of transfected HIF-1a full-length cDNA but also a minimum cDNA just containing the ORF (Figure 5). In contrast, this inhibition fails to negatively regulate the expression of constructs containing either the 50UTR or the 30UTR regions of HIF-1a fused to GFP, ruling out the involvement of the IRES present in the 50UTR of HIF-1 mRNA in the observed effect. MicroRNAs (miRs) have been postulated to influence HIF-1a expression in response to hypoxia50 or in a way independent of hypoxia.51miRs are short non-coding RNAs that bind their target mRNAs, usually at sites within the 30UTR, to destabilize mRNA and/or inhibit its translation. Interestingly, effective miR-binding sites have also been recently identified in ORFs and it is suggested they function with a similar mechanism to miRNA targeting 30UTR sites.52

Therefore, this mechanism remains an attractive hypothesis to explain the observed effect on HIF-1a translation, which warrants further investigation.

In addition to the a-subunit of HIF-1, our work shows that ATR is also involved in the regulation of the b-subunit. ATR silencing using siRNA was associated with a significant lower content of HIF-1b expression in whole extracts (Figure 1a) as well as in the chromatin-loaded fraction (Figure 6c). Similar decrease in HIF-b expression was observed in the WT/KD model (see Figures 1d and 6b). This effect on HIF-1b expression is reproduced but to a lesser extent when pharmacological inhibitors are used (see also below, the results obtained with VE-821). In fact, in the presence of ATR regulates HIF-1 expression

F Fallone et al

7

(9)

CGK733, the inhibition of HIF-1b expression was observed only at latter time points (from 2 h of hypoxia) and for higher doses of inhibitor (20 mM) compared with the conditions necessary to observe an inhibition of HIF-1a expression (from 1 h of hypoxia and 2.5 mMCGK733, respectively) (see Figure 2). By opposition to HIF-1a, very few data have been reported regarding the regulation of HIF-1b expression that is generally regarded as a constitutively expressed protein. A recent study performed by the group of Rocha and co-workers53demonstrated that NF-kB controls HIF-1b transcriptional levels in response to cytokine stimulation in both human and mouse cells. However, as shown in Supplementary Figure 2, we did not observe any difference upon hypoxia in HIF-1b mRNA levels between ATR-deficient and -proficient cells. Rather, preliminary experiments performed in our laboratory favor a reduction in HIF-1b translation when ATR expression and/or activity were inhibited (Supplementary Figure 5). While the molecular mechanism remains to be deciphered, these results highlight the pharmacological promises of the use of ATR inhibitors to counteract cancer cells hypoxic adaptation. To our knowledge this ability to decrease both HIF-1 subunits appears unique among the already identified HIF-1 inhibitors.7

Therefore, our work opens new pharmacological opportunities to target hypoxic cells within the tumors. Due to replicative stress existing in tumor cells, disruption of ATR function through depletion or KD protein expression decreases cancer cell survival even in the absence of DNA-damaging agents.54This crucial role for ATR led to a growing interest in ATR as a target for anticancer drugs notably in tumors with high levels of replicative stress,14 which is clearly the case in hypoxic areas. Interestingly, during the revision process of our manuscript, a new and highly selective inhibitor of ATR, VE-821, became commercially available.54 As shown in Supplementary Figure 6, this inhibitor, at doses that specifically inhibit ATR kinase activity in cellulo,54 decreases the expression of both HIF-1 subunits as well as HIF-1 target genes, GLUT-1 and CAIX. We examined the dose-dependent effect of VE-821 on HIF-1 protein accumulation in two independent cellular models, HeLa and U2OS. VE-821 decreased hypoxic accumulation of HIF-1a protein in a dose-dependent manner with an IC50of 0. 3

and 3 mM for HeLa and U2OS cells, respectively. Again, as for CGK733-treated cells, the effect of VE-821 on HIF-1b expression was less pronounced with an IC50of 2.5 and 4.5 mMfor HeLa and

U2OS cells, respectively. At molecular levels, while VE-821 effect on HIF-1a expression was still observed in the presence of MG132, ruling out an effect on the protein degradation, a clear decrease of HIF-1a biosynthesis was observed (Supplementary Figure 6). Similarly, exposure to VE-821 decreases HIF-1b biosynthesis (Supplementary Figure 6). These results are very encouraging and confirmed the pharmacological promises of our work.

To conclude, we have demonstrated for the first time the role of ATR in activating an adaptive response to counterbalance oxygen deprivation through regulation of HIF-1 expression. Our work supports ATR inhibition as a new prospect for cancer therapy with the potential to decrease cellular adaptation in hypoxic tumors.

MATERIALS AND METHODS Cell culture and treatments

Human cell lines U2OS (osteosarcoma) and HeLa (cervical adenocarci-noma) were grown in DMEM (Dulbecco’s modified Eagle’s medium; Invitrogen, Illkirch, France) supplemented with 10% fetal calf serum and antibiotics. U2OS stable clones allowing expression upon doxycycline addition of either WT ATR (WT, clone GW33) or dominant negative KD ATR (KD, clone GK41), kindly provided by Dr P Nghiem (University of

Washington, Seattle, USA), were grown as previously described.55 ATR

WT or KD expression was induced with 2 mg/ml doxycycline for 48 h before experiments. All cells were grown in a humidified atmosphere at 37 1C with

5% CO2and experiments were done with non-confluent, exponentially

growing cells at low passages. When indicated, normoxic and hypoxic cells

were preincubated for 1 h in presence of CGK733 (20 mM, unless indicated),

NU7026, LY294002, KU55933, MG132 (20 mM, respectively) or 5 mg/ml

aphidicolin. All drugs are from Sigma-Aldrich (Saint-Quentin Fallavier, France), except KU55933 and aphidicolin, respectively, from Calbiochem (Darmstadt, Germany) and Alexis Biochemicals (Enzo Life Sciences, Villeurbanne, France). For hypoxia treatment, 60–70% confluent cells were

cultured in a hypoxia chamber (InVIVO2 400, Ruskinn) flushed with 0.1% O2

and 5% CO2with balance of 95% nitrogen at 37 1C. Control cells were

placed in a 5% CO2and 95% air incubator (20% O2) at 37 1C. At the end of

the experiments, cells were washed twice with ice-cold phosphate-buffered saline, collected by scraping in normoxia or in the hypoxia chamber and placed in ice for centrifugation.

SiRNA-mediated knockdown

Transfections of U2OS with siRNAs were performed with lipofectamine 2000 reagent (Invitrogen) according to the manufacturer’s instructions.

Briefly, 1 105

cells per sample were transfected in six-well plates with 5 ml

of lipofectamine 2000 and siRNAs at a 30 and 40 nMfinal concentrations,

respectively, for ATR and Chk1 in 1 ml of serum and antibiotics free medium. After 4 h, medium was replaced with complete medium. Total knockdown times were of 72 and 48 h, respectively, for ATR and Chk1. For ATR silencing, HP Validated siRNAs provided from Qiagen (Hilden, Germany) were used (siRNA 1: SI02660231, siRNA 2: SI02664347). siRNA

directed against Chk1 was from Sigma-Aldrich (target sequence: 50

-UCGUGAGCGUUUGUUGAAC-30). Control siRNAs containing non-specific

sequences was provided by Qiagen (SI1022076). Cell extracts and western immunoblotting analysis

Whole-cell extracts and western blots were performed as previously

described.56 Membranes were incubated with primary antibodies

(antibodies are listed in Supplementary Table 1), washed and then incubated with the appropriate horseradish peroxidase-linked anti-mouse or anti-rabbit secondary antibody (Jackson Immunoresearch Laboratories, Suffolk, UK). To isolate nuclear soluble or chromatin subfractions from cells,

a detergent-based extraction procedure was used as reported.8Briefly, cell

fractionation was carried out by four consecutive extractions. The supernatant was collected at each step and labeled as fraction FI, FII, FIII,

FIV. Pellets of about 2 106

cells were first resuspended for 5 min on ice in

200 ml of extraction buffer (50 mMHepes, pH 7.5, 150 mMNaCl, 1 mMEDTA)

containing 0.2% NP40 supplemented with protease inhibitor (20 mg/ml phenylmethylsulfonyl fluoride, 1 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/

ml pepstatin) and phosphatase inhibitors (50 mM NaF, cocktail inhibitor

phosphatase I and II, all from Sigma). Following centrifugation at 1000 g for 5 min, the supernatants were collected (FI), the same extraction was realized and the supernatant were collected (FII) and pooled with the previous one. Therefore, the two-pooled first fractions were named FI. The pellets were further resuspended for 40 min on ice in 200 ml of extraction buffer containing 0.5% NP40 supplemented with protease inhibitor and phosphatase inhibitors. Pellets were clarified by centrifugation at 16 000 g for 15 min and the supernatants were collected (FIII). Insoluble FIV fractions were resuspended in a Tris pH 6.8 buffer supplemented with 2% SDS (sodium dodecyl sulfate), 10% glycerol and heated for 10 min at 100 1C. Equal aliquots of each fraction, derived from equivalent cell numbers, were separated by SDS–PAGE (SDS–polyacrylamide gel electrophoresis).

For measuring the expression levels of HIF-1a or HIF-1b in western blots experiments, the ImageJ software (NIH, Bethesda, MD, USA) was used. A loading control was used in each experiment for normalization (a-tubulin or b-actin for whole-cell extracts, HP1a for the chromatin-associated fraction FIV).

Determination of HIF-1a degradation

HIF-1a accumulation was induced with the hypoxia-mimetic agent DFX

(260 mM) during 2 h. Cells were then treated with 100 mg/ml CHX in the

presence or absence of CGK733 for the indicated time. Extracts were prepared at various time points (up to 1. 5 h) following the addition of CHX, as indicated in Results. HIF-1a steady-state levels were determined by western blot analysis. Similar experiments were performed using the ATR-inducible cell lines, ATR WT and ATR KD.

Determination of protein biosynthesis

Cells were grown in methionine/cysteine-free DMEM for 2 h. Cells were placed in hypoxia chamber for 2 h and then labeled for 1 additional hour in

hypoxia with methionine/cysteine-free medium containing 35

(10)

methionine/cysteine (EasyTag; Perkin-Elmer, Maanstraat, Germany) at a final concentration of 100 mCi/ml. When inhibitors were used, cells were pre-treated for 1 h before hypoxia. Radiolabel incorporation was then monitored by immunoprecipitation of total cell lysates as previously

described46using 5 mg of anti-HIF-1a (Novus Biochemicals, Cambridge, UK),

anti-HIF-1b and anti-b-catenin antibodies (see Supplementary Table 1). Immunoprecipitated proteins were separated by SDS–PAGE. Visualization and quantification were performed using Phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA).

RNA extraction and quantitative PCR (qPCR)

Total RNA, reverse transcription and real time PCR were performed as

previously described.42 GLUT-1 oligonucleotides used were as follow:

sense, 50-GCGGAATTCAATGCTGATGAT-30, antisense, 50-CAGTTTCGAGAAG

CCCATGAG-30. For CAIX, we used Assays-on-Demand Gene Expression

probes (Applied Biosystems, Courtaboeuf, France). Plasmids and DNA constructions

All PCR were performed using Phusion polymerase (New England Biolabs, Beverly, MA, USA). DNA constructions were validated by DNA sequencing. A list of primers and restriction sites used for cloning is provided in Supplementary Table 2. HIF-1a ORF–GFP expressing plasmid was generated by subcloning the BamHI–HIF-1a–NotI digestion fragment from pcDNA3-HA-HIF-1a (Addgene plasmid 18949) into a modified pEGFP-C1

(with a NotI site between BglII and BamHI). 50UTR-ORF-30UTR HIF-1a

expressing plasmid was generated by combining by PCR HIF-1a cDNA from IMAGE clone 3842146, amplified using primers FL-5-F and FL-R, with

the missing 50UTR part obtained from U2OS genomic DNA, amplified using

primers FL-F and FL-5-R. The resulting PCR products were fused by PCR

using FL-F and FL-R primers and cloned into pEGFP-C1. 50UTR–GFP plasmid

was generated by combining by PCR the 50UTR amplified from HIF-1a

full-length cDNA using primers 5-F and 5-R and GFP amplified from pEGFP-N1 using 5-GFP-R and 5-GFP-R primers. The resulting PCR products were fused by PCR using primers 5-F and 5-GFP-R and cloned into pEGFP-N1. GFP–

30UTR plasmid was generated by amplifying by PCR HIF-1a 30UTR using 3-F

and 3-R primers and cloned into pEGFP-C1. To generate a plasmid with HIF-1a ORF only, GFP was replaced in pEGFP-C1 by HIF-1a ORF amplified with ORF-F and ORF-R primers.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

ACKNOWLEDGEMENTS

We thank Catherine Botanch, Emilie Mirey and Christine Bordier for their help, Dr Nathalie Mazure and Dr Delphine Larrieu for critical comments on the manuscript and Aymeric Poizot for English rereading. We acknowledge Dr P Calsou, Dr Y Martineau and Dr S Pyronnet for technical advices and reagents. This work was supported by the ‘Ligue Nationale Contre le Cancer’ (Equipe Labe´lise´e, BS), the Cance´ropole GSO (BS and CM), the Association contre le Cancer, ARC (2008-1102, CM) and the Radioprotection Committee of EDF (CM).

REFERENCES

1 Yee Koh M, Spivak-Kroizman TR, Powis G. HIF-1 regulation: not so easy come, easy go. Trends Biochem Sci 2008; 33: 526–534.

2 Majmundar AJ, Wong WJ, Simon MC. Hypoxia-inducible factors and the response to hypoxic stress. Mol Cell 2010; 40: 294–309.

3 Monti E, Gariboldi MB. HIF-1 as a target for cancer chemotherapy, chemosensi-tization and chemoprevention. Curr Mol Pharmacol 2011; 4: 62–77.

4 Brahimi-Horn MC, Pouyssegur J. HIF at a glance. J Cell Sci 2009; 122: 1055–1057. 5 Semenza GL. Evaluation of HIF-1 inhibitors as anticancer agents. Drug Discov

Today 2007; 12: 853–859.

6 Brahimi-Horn MC, Pouyssegur J. Harnessing the hypoxia-inducible factor in cancer and ischemic disease. Biochem Pharmacol 2007; 73: 450–457.

7 Onnis B, Rapisarda A, Melillo G. Development of HIF-1 inhibitors for cancer therapy. J Cell Mol Med 2009; 13: 2780–2786.

8 Bouquet F, Ousset M, Biard D, Fallone F, Dauvillier S, Frit P et al. A DNA-dependent stress response involving DNA-PK occurs in hypoxic cells and contributes to cellular adaptation to hypoxia. J Cell Sci 2011; 124: 1943–1951.

9 Cam H, Easton JB, High A, Houghton PJ. mTORC1 signaling under hypoxic con-ditions is controlled by ATM-dependent phosphorylation of HIF-1alpha. Mol Cell 2010; 40: 509–520.

10 Abraham RT. mTOR as a positive regulator of tumor cell responses to hypoxia. Curr Top Microbiol Immunol 2004; 279: 299–319.

11 Gibson SL, Bindra RS, Glazer PM. Hypoxia-induced phosphorylation of Chk2 in an ataxia telangiectasia mutated-dependent manner. Cancer Res 2005; 65: 10734–10741. 12 Bencokova Z, Kaufmann MR, Pires IM, Lecane PS, Giaccia AJ, Hammond EM ATM activation and signaling under hypoxic conditions. Mol Cell Biol 2009; 29: 526–537.

13 Czornak K, Chughtai S, Chrzanowska KH. Mystery of DNA repair: the role of the MRN complex and ATM kinase in DNA damage repair. J Appl Genet 2008; 49: 383–396. 14 Lopez-Contreras AJ, Fernandez-Capetillo O. The ATR barrier to replication-born

DNA damage. DNA Repair (Amst) 2010; 9: 1249–1255.

15 Zou L, Elledge SJ. Sensing DNA damage through ATRIP recognition of RPA-ssDNA complexes. Science 2003; 300: 1542–1548.

16 Adams KE, Medhurst AL, Dart DA, Lakin ND. Recruitment of ATR to sites of ionising radiation-induced DNA damage requires ATM and components of the MRN protein complex. Oncogene 2006; 25: 3894–3904.

17 Cuadrado M, Martinez-Pastor B, Murga M, Toledo LI, Gutierrez-Martinez P, Lopez E et al. ATM regulates ATR chromatin loading in response to DNA double-strand breaks. J Exp Med 2006; 203: 297–303.

18 Jazayeri A, Falck J, Lukas C, Bartek J, Smith GC, Lukas J et al. ATM- and cell cycle-dependent regulation of ATR in response to DNA double-strand breaks. Nat Cell Biol 2006; 8: 37–45.

19 Myers JS, Cortez D. Rapid activation of ATR by ionizing radiation requires ATM and Mre11. J Biol Chem 2006; 281: 9346–9350.

20 Hammond EM, Dorie MJ, Giaccia AJ. ATR/ATM targets are phosphorylated by ATR in response to hypoxia and ATM in response to reoxygenation. J Biol Chem 2003; 278: 12207–12213.

21 Hammond EM, Green SL, Giaccia AJ. Comparison of hypoxia-induced replication arrest with hydroxyurea and aphidicolin-induced arrest. Mutat Res 2003; 532: 205–213.

22 Hammond EM, Dorie MJ, Giaccia AJ. Inhibition of ATR leads to increased sensi-tivity to hypoxia/reoxygenation. Cancer Res 2004; 64: 6556–6562.

23 Paulsen RD, Cimprich KA. The ATR pathway: fine-tuning the fork. DNA Repair (Amst) 2007; 6: 953–966.

24 Crescenzi E, Palumbo G, de Boer J, Brady HJ. Ataxia telangiectasia mutated and p21CIP1 modulate cell survival of drug-induced senescent tumor cells: implica-tions for chemotherapy. Clin Cancer Res 2008; 14: 1877–1887.

25 Pereg Y, Shkedy D, de Graaf P, Meulmeester E, Edelson-Averbukh M, Salek M et al. Phosphorylation of Hdmx mediates its Hdm2- and ATM-dependent degradation in response to DNA damage. Proc Natl Acad Sci USA 2005; 102: 5056–5061. 26 Aberle H, Bauer A, Stappert J, Kispert A, Kemler R. beta-catenin is a target for the

ubiquitin-proteasome pathway. EMBO J 1997; 16: 3797–3804.

27 Lovejoy CA, Cortez D. Common mechanisms of PIKK regulation. DNA Repair (Amst) 2009; 8: 1004–1008.

28 Veuger SJ, Curtin NJ, Richardson CJ, Smith GC, Durkacz BW. Radiosensitization and DNA repair inhibition by the combined use of novel inhibitors of DNA-dependent protein kinase and poly(ADP-ribose) polymerase-1. Cancer Res 2003; 63: 6008–6015. 29 Willmore E, de Caux S, Sunter NJ, Tilby MJ, Jackson GH, Austin CA et al. A novel DNA-dependent protein kinase inhibitor, NU7026, potentiates the cytotoxicity of topoisomerase II poisons used in the treatment of leukemia. Blood 2004; 103: 4659–4665.

30 Wan C, Kulkarni A, Wang YH. ATR preferentially interacts with common fragile site FRA3B and the binding requires its kinase activity in response to aphidicolin treatment. Mutat Res 2010; 686: 39–46.

31 Laughner E, Taghavi P, Chiles K, Mahon PC, Semenza GL. HER2 (neu) signaling increases the rate of hypoxia-inducible factor 1alpha (HIF-1alpha) synthesis: novel mechanism for HIF-1-mediated vascular endothelial growth factor expression. Mol Cell Biol 2001; 21: 3995–4004.

32 Galban S, Kuwano Y, Pullmann Jr R, Martindale JL, Kim HH, Lal A et al. RNA-binding proteins HuR and PTB promote the translation of hypoxia-inducible factor 1alpha. Mol Cell Biol 2008; 28: 93–107.

33 Lou JJ, Chua YL, Chew EH, Gao J, Bushell M, Hagen T. Inhibition of hypoxia-inducible factor-1alpha (HIF-1alpha) protein synthesis by DNA damage inducing agents. PloS one 2010; 5: e10522.

34 Zhong H, Chiles K, Feldser D, Laughner E, Hanrahan C, Georgescu MM et al. Modulation of hypoxia-inducible factor 1alpha expression by the epidermal growth factor/phosphatidylinositol 3-kinase/PTEN/AKT/FRAP pathway in human prostate cancer cells: implications for tumor angiogenesis and therapeutics. Cancer Res 2000; 60: 1541–1545.

35 Hudson CC, Liu M, Chiang GG, Otterness DM, Loomis DC, Kaper F et al. Regulation of hypoxia-inducible factor 1alpha expression and function by the mammalian target of rapamycin. Mol Cell Biol 2002; 22: 7004–7014.

36 Gingras AC, Gygi SP, Raught B, Polakiewicz RD, Abraham RT, Hoekstra MF et al. Regulation of 4E-BP1 phosphorylation: a novel two-step mechanism. Genes Dev 1999; 13: 1422–1437.

ATR regulates HIF-1 expression F Fallone et al

9

(11)

37 Chen C, Pore N, Behrooz A, Ismail-Beigi F, Maity A. Regulation of glut1 mRNA by hypoxia-inducible factor-1. Interaction between H-ras and hypoxia. J Biol Chem 2001; 276: 9519–9525.

38 Lang KJ, Kappel A, Goodall GJ. Hypoxia-inducible factor-1alpha mRNA contains an internal ribosome entry site that allows efficient translation during normoxia and hypoxia. Mol Biol Cell 2002; 13: 1792–1801.

39 Lei Z, Li B, Yang Z, Fang H, Zhang GM, Feng ZH et al. Regulation of HIF-1alpha and VEGF by miR-20b tunes tumor cells to adapt to the alteration of oxygen con-centration. PloS One 2009; 4: e7629.

40 Airley RE, Loncaster J, Raleigh JA, Harris AL, Davidson SE, Hunter RD et al. GLUT-1 and CAIX as intrinsic markers of hypoxia in carcinoma of the cervix: relationship to pimonidazole binding. Int J Cancer 2003; 104: 85–91.

41 Semenza GL. Targeting HIF-1 for cancer therapy. Nat Rev Cancer 2003; 3: 721–732. 42 Ousset M, Bouquet F, Fallone F, Biard D, Dray C, Valet P et al. Loss of ATM positively regulates the expression of hypoxia inducible factor 1 (HIF-1) through oxidative stress: Role in the physiopathology of the disease. Cell Cycle 2010; 9: 2814–2822. 43 Kang MJ, Jung SM, Kim MJ, Bae JH, Kim HB, Kim JY et al. DNA-dependent protein

kinase is involved in heat shock protein-mediated accumulation of hypoxia-inducible factor-1alpha in hypoxic preconditioned HepG2 cells. FEBS J 2008; 275: 5969–5981.

44 Economopoulou M, Langer HF, Celeste A, Orlova VV, Choi EY, Ma M et al. Histone H2AX is integral to hypoxia-driven neovascularization. Nat Med 2009; 15: 553–558. 45 Martin L, Rainey M, Santocanale C, Gardner LB. Hypoxic activation of ATR and the

suppression of the initiation of DNA replication through cdc6 degradation. Oncogene 2011; 31: 4076–4084.

46 Rapisarda A, Melillo G. UVC inhibits HIF-1alpha protein translation by a DNA damage-and topoisomerase I-independent pathway. Oncogene 2007; 26: 6875–6884.

47 Semenza GL. Hypoxia-inducible factors in physiology and medicine. Cell 2012; 148: 399–408.

48 Zundel W, Schindler C, Haas-Kogan D, Koong A, Kaper F, Chen E et al. Loss of PTEN facilitates HIF-1-mediated gene expression. Genes Dev 2000; 14: 391–396. 49 Masuda K, Abdelmohsen K, Gorospe M. RNA-binding proteins implicated in the

hypoxic response. J Cell Mol Med 2009; 13: 2759–2769.

50 Dolt KS, Mishra MK, Karar J, Baig MA, Ahmed Z, Pasha MA. cDNA cloning, gene organization and variant specific expression of HIF-1 alpha in high altitude yak (Bos grunniens). Gene 2007; 386: 73–80.

51 Wang C, Song B, Song W, Liu J, Sun A, Wu D et al. Underexpressed microRNA-199b-5p targets hypoxia-inducible factor-1alpha in hepatocellular carcinoma and predicts prognosis of hepatocellular carcinoma patients. J Gastroenterol Hepatol 2011;26: 1630–1637.

52 Moretti F, Thermann R, Hentze MW. Mechanism of translational regulation by miR-2 from sites in the 50untranslated region or the open reading frame. RNA

2010; 16: 2493–2502.

53 van Uden P, Kenneth NS, Webster R, Muller HA, Mudie S, Rocha S. Evolutionary conserved regulation of HIF-1beta by NF-kappaB. Plos Genet 2011; 7: e1001285. 54 Reaper PM, Griffiths MR, Long JM, Charrier JD, Maccormick S, Charlton PA et al.

Selective killing of ATM- or p53-deficient cancer cells through inhibition of ATR. Nat Chem Biol 2011; 7: 428–430.

55 Nghiem P, Park PK, Kim Y, Vaziri C, Schreiber SL. ATR inhibition selectively sen-sitizes G1 checkpoint-deficient cells to lethal premature chromatin condensation. Proc Natl Acad Sci USA 2001; 98: 9092–9097.

56 Monferran S, Paupert J, Dauvillier S, Salles B, Muller C. The membrane form of the DNA repair protein Ku interacts at the cell surface with metalloproteinase 9. EMBO J 2004; 23: 3758–3768.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc) 10

Figure

Figure 1. Inhibition of ATR expression and activity down-regulates hypoxic accumulation of HIF-1 in a chk1-independent manner
Figure 2. CGK733 inhibits hypoxic accumulation of HIF-1 subunits in a time- and dose-dependent manner
Figure 4. Regulation of HIF-1a by ATR occurs at the translational level. (a) Upper panel: HeLa cells were treated in the presence ( þ ) or absence (  ) of MG132 (20 m M ) for 1 h, then treated or not with the indicated inhibitors at 20 m M and placed in hy
Figure 5. Translational regulation of HIF-1a by ATR involves elements located in the ORF region of HIF-1a mRNA
+2

Références

Documents relatifs

Clustering and Sørensen dissimilarity analysis based solely on the subset of metavirome sequences shared with metagenomes revealed the same summer and autumn and winter and

More than 30 years ago Lucy [34] proposed on the basis of molecular models that the conformation of the multiple double bonds in the PUFA chains of arachidonic acid-

Nous utilisons le cadre du clustering de documents pour évaluer l’espace de concepts induit par notre hypothèse, et nous comparons ses performances avec l’espace de concepts induit par

This paper proposes a new step-wise integrated approach that could potentially assist municipal professionals in developing optimized renewal plans that identify the most

- دعاست ةيلمع ةنمقرلا ىلع ظفح ةنايصو تاطوطخملا ةيبرعلا كلذو اهنيزختب ىلع صارقلأا ةزنتكملا ( CD-ROMs )، يلاتلابو مهاست يف ةدايز لخد كملا تابت نع قيرط عيب

/ La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur. Access

Moreover, it does not to account for several main aspects of this assembly: (i) the rapid turnover of ParB between clusters, (ii) the absence of accumulation of ParB in front of

The authors say that being critical is especially important in dealing with controversial social issues for the media can shape controversy in terms of who gets heard, and