HIF and reactive oxygen species regulate oxidative phosphorylation in cancer

© The Author 2008 . Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

HIF AND REACTIVE OXYGEN SPECIES REGULATE OXIDATIVE

PHOSPHORYLATION IN CANCER

Eric HERVOUET1#a, Alena CÍZKOVÁ2,3#, Jocelyne DEMONT1, Alena VOJTÍSKOVÁ2,

Petr PECINA2, Nicole L.W. FRANSSEN - van HAL5, Jaap KEIJER 5, Hélène

SIMONNET1b, Robert IVÁNEK 3,4, Stanislav KMOCH3, Catherine GODINOT1*and

Josef HOUSTEK2

1 Centre de Génétique Moléculaire et Cellulaire, UMR 5534, Centre National de la Recherche

Scientifique –Claude Bernard University of Lyon 1 –Villeurbanne, France

2 Institute of Physiology, Academy of Sciences of the Czech Republic, Videnska 1083, 142

20 Praha, Czech Republic

3 Institute of Inherited Metabolic Disorders, Faculty of Medicine, Charles University, Ke

Karlovu 2, Prague, Czech Republic

4 Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, Videnska

1083, 142 20 Praha, Czech Republic

5 RIKILT - Institute of Food Safety, Wageningen, The Netherlands

# Eric HERVOUET and Alena CÍZKOVÁ contributed equally to this study.

a Present address : Inserm U601, Institute of biology, 9 quai Moncousu, F-44035 Nantes,

France.

b Present address : UMR 5201, CNRS, Université Claude Bernard Lyon 1, F-69373 Lyon,

France.

Carcinogenesis Advance Access published May 29, 2008 by guest on O

ctober 3, 2014http://carcin.oxfordjournals.org/

Dow

nloaded from

http://carcin.oxfordjournals.org/

* Corresponding author: Dr. C. Godinot ; Centre de Génétique Moléculaire et Cellulaire,

UMR 5534, Centre National de la Recherche Scientifique – Université Claude Bernard de

Lyon 1 – 43 Bvd du onze novembre, 69622 Villeurbanne, cedex, France. Telephone: +33-

478364192 ; e-mail: [email protected]

Acknowledgements: This work was supported by the LRCC (“Ligue Régionale Contre le

Cancer”), by the CNRS (UMR 5534, IFR 41), by the INSERM (ATC VIE0208), by the

Claude Bernard University of Lyon 1 (BQR grant), by the MSMT CR (MSM0021620806 and

1M6837805002) and by the GACR (303/07/0781). We also thank EGIDE for granting

subsidies for travels between France and Czech Republic.

Running title: HIF, ROS and mitochondria

by guest on October 3, 2014

http://carcin.oxfordjournals.org/D

ownloaded from


Abstract

A decrease in oxidative phosphorylation (OXPHOS) is characteristic of many cancer

types and, in particular, of clear cell renal carcinoma (CCRC) deficient in vhl (von Hippel

Lindau gene). In the absence of functional pVHL, HIF1-α and HIF2-α subunits are stabilized,

which induces the transcription of many genes including those involved in glycolysis and

ROS metabolism. Transfection of these cells with vhl is known to restore HIF-α subunit

degradation and to reduce glycolytic genes transcription. We show that such transfection with

vhl of 786-0 CCRC (which are devoid of HIF1-α), also increased the content of respiratory

chain subunits. However, the levels of most transcripts encoding OXPHOS subunits were not

modified. Inhibition of HIF2-α synthesis by RNA interference in pVHL-deficient 786-0

CCRC also restored respiratory chain subunit content, and clearly demonstrated a key role of

HIF in OXPHOS regulation. In agreement with these observations, stabilization of HIF-α

subunit by CoCl2 decreased respiratory chain subunit levels in CCRC cells expressing pVHL.

In addition, HIF stimulated ROS production and mitochondrial Mn-SOD content. OXPHOS

subunit content was also decreased by added H2O2. Interestingly, desferrioxamine that also

stabilized HIF did not decrease respiratory chain subunit level. While CoCl2 significantly

stimulates ROS production, desferrioxamine is known to prevent hydroxyl radical production

by inhibiting Fenton reactions. This indicates that the HIF-induced decrease in OXPHOS is at

least in part mediated by hydroxyl radical production.

Keywords: Clear cell renal carcinoma; mitochondrial biogenesis; reactive oxygen species;

Hypoxia-inducible factor; siRNA; microarrays; desferrioxamine; cobalt chloride; hydrogen

peroxide; Mn-superoxide dismutase.



ownloaded from


Introduction

More than fifty years ago, Warburg observed that cancer cells maintained a high

glycolytic rate even in the presence of oxygen. They use glycolysis rather than oxidative

phosphorylation (OXPHOS) to produce most of their ATP [1]. Whereas the stimulation of

glycolysis is explained by an up-regulation of glycolytic genes mainly by stabilization of the

transcription factor, Hypoxia-Inducible Factor (HIF), the down-regulation of oxidative

phosphorylation is not fully understood.

Previous studies from our laboratory and from others have shown that all

mitochondrial respiratory chain activities and subunit contents were lower either in renal

tumors exhibiting a deficiency in von Hippel Lindau tumor suppressor gene, (vhl) than in

normal adjacent tissue [2,3] or in cultured CCRC cells (786-0 cells issued from such tumors)

before being transfected with vhl [4, 5].

HIF is composed of a constitutively expressed HIF1-β subunit and of oxygen sensitive

HIF1-α or HIF2-α subunits [6]. Under aerobic conditions, the HIF-α subunits are

hydroxylated [7] at the level of conserved proline and asparagine residues. After proline

hydroxylation, the HIF-α subunits are ubiquitinated by the von Hippel Lindau tumor

suppressor protein, pVHL, which targets them for degradation by the proteasome. Under

hypoxia where HIF-α hydroxylation is reduced or in the absence of pVHL where

ubiquitination does not occur, the HIF-α subunits are stabilized and translocated to the

nucleus. They associate with HIF1-β, recruit transcriptional co-activator proteins such as

p300/CPB, and bind to the hypoxia response elements (HRE) present in over 70 known genes,

which enhances their transcription. HIF-α asparagine hydroxylation prevents the binding of

transcriptional co-activators [for review, see 8]. The prolyl hydroxylases, which contain Fe++

in their active site, can be inactivated by iron chelators such as desferrioxamine (DFO).



ownloaded from


Moreover, transition metals such as CoCl2 stabilize HIF by depleting cellular ascorbate,

essential cofactor of prolyl hydroxylases [9]. As early as 1996, Semenza et al [10] showed

that activation of glycolytic genes by HIF1-α increased conversion of glucose to pyruvate and

lactate. In addition, activation of the HIF-targeted kinase, PDK1 inhibiting pyruvate

dehydrogenase suggested that respiration was decreased by substrate limitation [11]. More

recently, Fukuda et al [12] suggested that HIF-1α could up-regulate the COX4-2 isoform [13]

and the LON protease required for COX4-1 isoform degradation. The ensuing COX4-

1/COX4-2 isoform switch could increase the COX activity at low oxygen concentration,

because the COX4-2 isoform imposes higher turnover than the COX4-1 isoform. The switch

could then improve electron flux through the respiratory chain.

Simultaneously to the restoration of respiratory chain protein complexes observed

after vhl transfection in 786-0 cells, their capacity to rely on OXPHOS for their growth was

also increased [4]. However, the induction of COX4-2 could not be observed in these cells

[14]. This result that appears in contradiction with that observed by Fukuda et al [12] can

probably be explained by the fact that these cells express HIF2-α but not HIF1-α [15]. This

would indicate that COX4-2 can only be induced by HIF1-α. Therefore, additional

mechanisms parallel to the COX4 subunit isoform switch should be involved in the down-

regulation of respiratory chain complexes in these cancer cells.

In the present paper, we have studied whether inhibition of HIF synthesis by RNA

interference could restore the level of respiratory chain subunits in the absence of pVHL and

whether stabilization of HIF by compounds such as DFO or CoCl2 that prevents HIF

hydroxylation would decrease their content in the presence of pVHL. The COX4 subunit was

initially chosen as the preferred target. Then, since all mitochondrial respiratory chain

activities and their subunit contents were decreased in these cancer cells, we also studied two

complex III proteins, the core 2 subunit and the 13.4 kDa subunit that were markedly



ownloaded from


modified by the presence or absence of pVHL [4,5]. Complex III is particularly relevant to

this study because it is known as one of the primary site of mitochondrial ROS production

during hypoxia [16] and participates in the production of superoxide anions [17]. Indeed, this

study will show that an enzyme involved in ROS production such as the mitochondrial

manganese superoxide dismutase (Mn-SOD) is markedly induced in vhl-deficient cells and

that increases in ROS production are correlated with the down-regulation of respiratory chain

subunits.

Materials and methods

Cell culture: pVHL-deficient parental 786-0 cells that had been stably transfected

either with a vector expressing wild-type vhl, (VHL+ cells, clones WT8) or with a void vector

(VHL- cells, clone PRC3) [18] were kindly provided by Dr Kaelin. Cells were routinely

grown as described in [4] and used at 70-80% confluence. Cells were counted in the presence

of 0.04% Trypan blue.

Comparative microarray analysis of gene expression after transfection of 786-0 cells

with a vector expressing pVHL or with a void vector: Total RNA was extracted using the

TRIZOL (Invitrogen, Carlsbad, CA), quantified and purity checked by NanoDrop

spectrophotometer (NanoDrop Technologies, Rockland, De) and 1 % TBE/agarose gels.

Quality control was performed on Agilent 2100 bioanalyser - RNA Lab-On-a-Chip (Agilent

technologies, Palo Alto, CA). Total RNA (50 μg) was directly labeled by incorporation of

either Cy3-dCTP or Cy5-dCTP during reverse transcription. cDNA was purified using the

PCR purification protocol (QIAquick PCR purification kit, QIAgen, USA). For each cell line,

three biological replicates were analyzed in duplicate against common reference (RNA pool

aliquots from all the experiments) using an oligonucleotide microarray containing Human

10K Oligo Set (MWG) spotted on Ultra GAPSMT Coated Slides (Corning, NY, USA). The



ownloaded from


microarray preparation, hybridization conditions and data acquisition were performed as

previously described [19, 20]. Image analysis of the resulting TIFF files was done and

expression data were obtained using Gene PixPro software (Axon Instruments, Union City,

CA). Data analysis was performed in R statistic environment (http://www.r-project.org/)

using Linears models for Microarray data with Limma package [21] that is part of

Bioconductor project (http://www.bioconductor.org/). Normalization of gene expression data

was done by Loess and G-quantile function, background was corrected by multiple testing

corrections using Benjamini & Hochberg False discovery rate method (FDR) [22].

RT-PCR : Total RNA was extracted with Trizol™ reagent (Invitrogen™). cDNA was

prepared as before and used for semi quantitative PCR or quantitative PCR (Light cycler

instrument™, in the presence of SYBR-Green™ (Roche diagnostics)) using previously

described primers [4, 14]. Genomic DNA of 786-0 VHL+ cells, used as positive control of

primers, was extracted with the “GeneEluteTM mammalian genomic DNA miniprep kit”

(Sigma) as recommended by the manufacturer.

RNA interference: The siRNA duplex oligonucleotides designed by Sowter et al. [15]

to target HIF2-α were used. They are targeted to nucleotides 1260 to 1280 of the HIF2-

α mRNA (NM001430) and have the following sequences:

sense 5'-CAGCAUCUUUGAUAGCAGUdTdT-3',

antisense 5'-ACUGCUAUCAAAGAUGCUGdTdT-3'.

The duplexes were prepared by mixing 14 µl of Lipofectamine™ (Invitrogen) with 80 nM

oligonucleotides in 700 µl DMEM for 30 min at room temperature. 786-0 cells were grown in

DMEM supplemented with glutamine and 10% FCS (Invitrogen), in 25 cm2 dishes up to 60%

confluence before transfection. The cells were rinsed with DMEM to remove any residual

serum before addition of 2 ml DMEM plus the oligonucleotide duplex (final oligonucleotide



ownloaded from


concentrations: 20 nM) in serum-free conditions. Four hours after transfection, 10% FCS was

added. Cells were studied from 24 to 72 h after transfection.

ROS production: Intracellular reactive oxygen species (ROS) levels were estimated

with the fluorescent dye 2',7'-dichlorodihydrofluorescein diacetate (H2DCF-DA) (Molecular

probes, Eugene, OR) which is a nonpolar, nonfluorescent compound that is converted into a

polar nonfluorescent derivative (DCFH) by cellular esterases. DCFH is membrane permeable

and is rapidly oxidized to the highly fluorescent 2',7'-dichlorofluorescein (DCF) in the

presence of intracellular ROS. Cells cultured in 96 well dishes until 70-80% confluence were

incubated in saline buffer (135 mM NaCl, 5 mM KCl, 0.4 mM KH2PO4, 1 mM MgSO4, 20

mM HEPES, pH 7.4, 5.55 mM glucose, 1 mM CaCl2) with 1 μM CM-H2DCFDA. ROS

production was estimated by measuring fluorescence immediately after probe addition and 2 h

later (fluorescence plate reader Victor™, PerkinElmer) DCF : λexc = 485 nm, λem = 535 nm.

Cell number was then estimated in each well by crystal violet staining [23]. The fluorescence

intensity was normalized to the cell number.

Western Blot analysis: Proteins extracted from cell pellets were homogenized in lysis

buffer, quantified using the Bradford reagent, resolved by SDS-PAGE in 6 or 12,5%

acrylamide/bisacrylamide gel and transferred to a BA83 nitrocellulose membrane (Schleicher

and SchüellTM), as described [4]. Staining the gels and membranes either with Coomassie

Blue or Ponceau Red checked equal loading and protein transfer efficiency. Membranes were

blocked with 5% skim-milk in NaCl-phosphate buffer and then incubated overnight at 4°C

with different primary antibodies, washed 5 times with buffer A (NaCl-phosphate buffer

containing 0.1% Tween 20), incubated for 1 h with horseradish peroxidase-conjugated anti-

rabbit IgG (1/10,000) for polyclonal antibodies or anti-mouse IgG (1/5,000) (BioradTM) for

monoclonal antibodies and washed with buffer A. Primary antibodies directed against the

OXPHOS proteins were used as previously described [4]. Polyclonal anti-HIF2-α (1/1,000)



ownloaded from


(Abcam™), anti-actin (1/1,000) (Santa Cruz™), anti-NOX4 (1/1,500) (generous gift of B.

Goldstein) [24], anti-aconitase (1/2,000) [25] and anti-IRP2 (1/1000) (generous gifts of Dr E.

Leibold) [26] were used. For analysis of the content of superoxide dismutases, whole cell

extracts or samples of mitochondria isolated from cultured cells using hypotonic shock were

resolved by SDS-PAGE in 10% gel, blotted to PVDF membrane (0.45 μm pores, Millipore)

and incubated with polyclonal anti SOD1 (cytosolic Cu-Zn superoxide dismutase)

(Calbiochem, No 574597) and polyclonal anti SOD2 (mitochondrial Mn-SOD) (Calbiochem,

No 574596). Amounts of proteins revealed by Western blot analysis were estimated using

either the Molecular Analyst™ or AIDA software.

Iron titration: Hydroxyl radical (•OH) that is mainly responsible for the oxidation of

macromolecules, is formed by the Haber–Weiss reaction as well as by iron catalyzed Fenton

reactions. Therefore, the ability of cells to concentrate iron can enhance oxidative damages.

Intracellular level of Fe2+ was estimated according to Arrigo et al [23].

Statistical analysis: When indicated, data are given as means ± SD with n indicating

the number of experiments. One-way ANOVA and t tests were applied for statistical analysis,

as appropriate, using Prism software™. Values were considered significant when p<0.05*,

p<0.01** or p<0.001***.

Results

Inactivation of HIF2-α by small interfering RNA reversed respiratory chain deficiency

in VHL-deficient cells

A decrease in OXPHOS activities correlating with a decrease in respiratory chain

subunit contents had been previously observed in renal tumors and in cells lacking functional

pVHL [2, 4]. These cells had to rely on glycolysis to sustain their growth. To know whether

this OXPHOS decrease in VHL- cells was directly mediated by accumulation of HIF2-α



ownloaded from


subunits, siRNA targeted to HIF2-α was introduced into pVHL-deficient 786-0 cells that had

been stably transfected either with a void vector, VHL- cells (786-0 PRC3) or with a vector

expressing functional pVHL, VHL+ cells (786-0 WT8). The 786-0 cells are well suited to

answer this question since they are devoid of HIF1-α and the used siRNA treatment prevents

the expression of HIF2-α in these cells [15]. In agreement with these authors, no cell toxicity

was noted after transfection of the siRNA or of Lipofectamine alone. Fig 1A shows that, as

expected, the HIF2-α transcript amount was strongly decreased in the presence of siRNA

specific for HIF2-α in the two cell types. The expression of mRNA coding the constitutively

expressed L32 ribosomal protein, used as a control, appeared at the same cycle N in all the

different conditions indicating that each sample contained similar mRNA amounts. The HIF2-

α signal appeared about six cycles later when comparing cells treated in the presence or

absence of siRNA targeting HIF2-α for VHL+ cells, proving that siRNA had been efficient to

decrease HIF2-α. Moreover, the transfection of the siRNA targeting HIF2-α almost

completely abolished HIF2-α protein content, either in VHL- cells or in VHL+ cells treated

with CoCl2 to prevent the pVHL-induced degradation of HIF2-α (Figure 1B).

Figure 1C shows that, in the absence of siRNA, the amounts of Core 2 (complex III)

and COX 4 (complex IV) subunits, were much lower in VHL- than in VHL+ cells while the

ATPase β subunit content did not significantly change. These data are in agreement with

those previously described [4]. Transfection with the siRNA targeting HIF2-α, increased the

contents of these respiratory chain subunits in VHL- cells. Therefore HIF2-α decreased

respiratory chain subunit amounts and HIF2-α acted downstream pVHL in the regulation of

respiratory chain subunit biogenesis. Within 72 hours of RNAi treatment, the amounts of

cytochrome c oxidase COX4 subunit (Fig 1D) and of complex III Core 2 protein (Fig 1E) had

approximately doubled. Similar data were observed with the 13.4 kDa complex III subunit



ownloaded from


(not shown). Transfection with the siRNA did not change the amount of these respiratory

chain subunits in VHL+ cells in which the HIF2-α subunit was absent, demonstrating the

essential and specific role of HIF2-α to down regulate these respiratory chain subunits.

It should be reminded that the above experiments have been performed with 786-0 cells, that

express HIF2-α but not HIF1- α [15]. In these cells, the COX4 subunit that was titrated witth

the anti-COX-4 antibody must be the COX4-1 isoform. Indeed, we could have never detected

the transcript encoding the COX4-2 subunit isoform by RT-PCR, even in the presence of

CoCl2 that stabilizes HIF-2α in VHL+ cells. This cannot be due to a technical PCR problem

since the COX4-2 primers designed within a single exon could efficiently amplify this gene

when genomic DNA extracted from VHL- or VHL+ cells was used as template [14].

Differential effects of HIF2-α stabilization by CoCl2 or DFO on respiratory chain

subunit content.

To confirm the down-regulation of respiratory chain proteins by HIF2-α, we then

studied whether drugs known to stabilize HIF2-α such as CoCl2 and DFO could also decrease

respiratory chain subunit level in cells expressing pVHL. Total proteins were extracted from

VHL+ cells after incubation for 8 h with 250 μM DFO or CoCl2. HIF2-α and COX 4-1

contents were analyzed by Western blot (Figure 2A, 2B). As expected, a high signal was

detectable with anti HIF2-α antibody after 8 h of incubation with CoCl2 or DFO while

untreated VHL+ cells did not significantly express HIF2-α. As previously shown [14], the

COX 4-1 amount was decreased by CoCl2 (Figure 2A, 2B). However, no significant

difference in COX 4-1 amount could be detected between cells treated or not with DFO. The

contents of Core 2 and 13.4 kDa complex III subunits were also decreased by CoCl2, but were

both not modified by the presence of DFO (not shown). Hence, although the HIF2-α content

was highly induced in the presence of DFO, DFO did not cause the HIF-induced decrease of

respiratory chain subunit levels.



ownloaded from


Effect of pVHL deficiency on the transcriptome of 786-0 cells: microarray analysis of

transcripts expressed in 786-0 cells containing either an active or non functional pVHL.

In order to better understand the mechanisms involved in the regulation of respiratory

chain subunit amounts by the pVHL/HIF system in 786-0 cells, we have studied the influence

of pVHL on mRNA expression by a large scale microarray analysis. The MWG microarray

that was used comprised a set of 10,162 genes. About 6,000 gave a valid signal. At a p value

< 0.0001, twenty seven genes were up-regulated and seventy eight down regulated (Table 1)

when comparing VHL+ to parental VHL-. Supplementary tables 1 and 2 report differences

observed with a p value < 0.05. Expected differences in the expression of vhl or of well-

known HIF-induced genes such as vegf or cyclin D1, for example, confirmed the validity of

the data.

Among differentially expressed genes, none concerned OXPHOS subunits and

OXPHOS-specific assembly factors and differences in the expression of only very few genes

known to be directly involved in mitochondrial metabolism could be put forward. There was

no significant change in pro-mitochondrial regulatory genes, such as PGC1α, NRF1 and

TFAM, and only a slight decrease of p53 in VHL- cells. Therefore these data confirmed that

decreased content of respiratory chain proteins did not likely result from lower expression of

their genes but probably from changes in post-transcriptional regulation. Among the over-

expressed genes, the mitochondrial Mn-SOD was 7.2-fold more expressed in VHL- than in

VHL+ cells. Another gene encoding Lysyl oxidase, LOX4 known as a HIF target [27] was

also increased by a factor of 12. The importance of these changes observed in Mn-SOD and

lysyl oxidase, two enzymes producing H2O2 underscored the role that ROS might play in the

regulation of the metabolism of these cells and ROS metabolism was therefore further

studied.



ownloaded from


Influence of HIF on superoxide dismutases, aconitase, NOX 4 and iron-regulatory

proteins

On the basis of these microarray data, the expression of mitochondrial Mn-SOD (Type

2) and that of cytosolic Cu-Zn-SOD (Type 1) were analyzed in 786-0 cells. Figure 2C shows

that, in agreement with the microarray data, the mitochondrial Mn-SOD was markedly

stimulated when VHL- were compared to VHL+ cells. There was 6-fold specific content of

Mn-SOD in cell lysates and 4-fold in isolated mitochondria of VHL+ cells. On the contrary,

no significant change could be observed in the level of the cytosolic Cu-Zn SOD that also did

not appear among the differentially expressed genes in the microarray analysis. Therefore, the

cells devoid of pVHL have a particularly efficient tool to transform superoxide anions

produced by the mitochondria into H2O2.

H2-DCFDA represents a well established fluorescence probe to monitor mitochondrial

ROS production in intact cells [28]. Although DCFDA is not a good quantitative probe for

measuring H2O2, it detects very well several other oxygen radicals and can be used as a

qualitative marker of the overall oxidative stress [29]. Figure 2D shows that DCF

fluorescence was significantly higher in VHL- cells devoid of active pVHL (hence,

expressing HIF2-α) than in VHL+ cells transfected with functional vhl (hence, inducing

HIF2-α degradation). CoCl2 treatment further increased DCF fluorescence. This technique

could not be used in the presence of DFO that complexes iron since oxidation of DCFH to

DCF is known to occur slowly if at all in the absence of ferrous ion [29]. Therefore DCF

fluorescence is not a valid test for measuring ROS in the presence of DFO.

Aconitase is an iron-sulfur protein that is an essential mitochondrial enzyme and is

highly susceptible to oxidative damage. Loss of its activity is frequently used as an indicator

of ROS generation. Under conditions of increased hydrogen peroxide concentration, the loss

of activity becomes irreversible as a result of Fenton reaction [30]. Such conditions occur in



ownloaded from


VHL- cells due to Mn-SOD upregulation. Indeed, the aconitase activity was significantly

lower in VHL- than in VHL+ cells (Figure 2E). However, the aconitase content was only

slightly lower in VHL- than in VHL+ cells. This decrease was however below the limit of

significance (Figure 2F). Treating the VHL+ cells with CoCl2 for 8 h also decreased the

aconitase activity (Figure 2E) and did not modify the amount of aconitase (Figure 2F).

Importantly, these findings indicate inevitable increase of hydroxyl radical production in

VHL- cells.

No significant difference could be put forward for NOX4, a NADH oxidase (not

shown) that is known to be involved in hydrogen peroxide metabolism [31]. In parallel, as an

increased level of iron could be responsible for increased ROS production, we compared iron

contents and IRP1 and IRP2 (not shown), that are iron regulatory proteins important in iron

transport in the cell [25, 26]. No significant differences could be detected between VHL- and

VHL+ cells (not shown).

Effects of H2O2 on OXPHOS subunit contents and cell viability

To test the direct effects of H2O2 on OXPHOS subunit contents, VHL+ cells were

incubated with increasing concentrations of H2O2 for 8 h. Figures 3A and 3B show that COX

4-1 and Core 2 protein amounts were both decreased after exposition to H2O2. The decrease

observed for COX 4-1 was more pronounced than for Core 2, which can be at least partly

explained by the slower turnover of Core 2 and 13.4 kDa subunits than that of COX 4 [14].

Importantly, Figure 4A shows that incubation of the cells for 24 h with increasing amounts of

H2O2 was more toxic for cells expressing HIF (VHL-) and susceptible to production of higher

levels of hydroxyl radicals (Figure 2) than for cells in which HIF degradation was induced by

pVHL: at 50 µM H2O2, the cell number had been decreased by more than 50% in VHL- cells

while about 75 % of the VHL+ cells were still alive. DFO was also more toxic for VHL- cells

than for VHL+ cells. However the presence of 50 or 100 µM DFO, an agent inhibiting Fenton



ownloaded from


reaction, diminished the H2O2-induced cell toxicity, especially in the case of VHL- cells

(Figure 4A). Indeed, the DFO-induced protective effect against H2O2 was not very efficient in

VHL+ cells. We have seen above (Figure 2B) that DFO was also able to partly protect VHL-

cells against HIF2-α-induced COX 4-1 subunit degradation.

Figures 4C and 4D show that, as expected, the VHL+ cells always contain more

COX4-1 than the VHL- cells incubated under the same conditions (ccompare lanes 1 to 2, 3

to 4, 5 to 7 or 6 to 8). In addition, when COX4-1 amount is decreased by H2O2 treatment

(compare lane 1 to 3 or 2 to 4), the presence of DFO restored the normal COX4-1 level

(compare lanes 3 to 6 or 4 to 8).

Discussion

Previous studies from our laboratory had shown that, in agreement with observations

made on several cancer types [32, 33], including the early observations of Warburg [1],

OXPHOS is down-regulated in renal cancer cells [2]. In CCRCs, vhl encoding pVHL is one

of the main genes involved in tumorigenesis. When vhl is transfected in vhl-deficient CCRCs,

the cells loose their capacity to develop tumors in nude mice [34] and their OXPHOS subunit

contents are restored [4]. In this paper, the use of small interfering RNA targeted to HIF2-α in

cells devoid of HIF1-α suggested that the presence of HIF was primarily responsible for the

down-regulation of respiratory chain subunit biogenesis. The amount of OXPHOS subunits

was inversely correlated to the expression of HIF-α independently of pVHL. Indeed, in the

VHL- cells, the inhibition of HIF2-α expression by siRNA increased the content of OXPHOS

subunit, as shown for complex III (core 2; 13.4 kDa subunits) or complex IV (COX4-1

subunit). The effect was specific since siRNA targeted to HIF2-α modified neither cell

viability nor OXPHOS subunit content in control VHL+ cells.



ownloaded from


Fukuda et al [12] recently confirmed that transfection of vhl in another line of vhl-

deficient renal cancer cells, RCC4, could also restore COX deficiency. They proposed that

HIF1 could induce COX4-2 subunit after binding to a hypoxia response element present in the

COX 4-2 gene 5'UTR; on the contrary, the COX 4-1 isoform would be degraded probably

because of HIF1-induced increased expression of the LON mitochondrial matrix protease.

However this up-regulation of COX4-2 is not relevant in 786-0 cells used here since no

COX4-2 expression could be detected in these cells. Contrarily to RCC4 cells, the 786-0 cells

do not contain HIF1-α but only HIF2-α while RCC4 cells express both isoforms. The

differential COX4-2 expression adds to the already observed differences between these two

factors. While some of their effects may be inherent to both, many seem to be quite specific

or even antagonizing, such as their differential effect on c-Myc transcriptional activity [35].

The absence of COX4-2 transcript also shows that the putative induction of COX4-2

transcription regulated by a novel oxygen responsive element described by Hüttemann et al

[36] did not occur in these cells, which could be expected since, in culture, the cells were not

submitted to hypoxia and since, in vivo, COX4-2 transcription was seen in lung and liver, but

not in brain, heart, and kidney [13].

As expected from the above RNAi experiments, HIF stabilization by CoCl2 in VHL+

cells also decreased respiratory chain subunit contents. However, since CoCl2 inhibits the

mitochondrial intermediate peptidase which prevents pre-COX 4 cleavage and hence

cytochrome c oxidase assembly [14], the relationship between CoCl2 effects on HIF

stabilization and on respiratory chain subunit degradation could not be established without

ambiguity. In addition, DFO induced HIF2-α over-expression but, unexpectedly, did not

decrease respiratory chain subunits content suggesting that the presence of HIF2-α was not

sufficient in the presence of DFO.



ownloaded from


Since HIF is a transcription factor involved in the regulation of a large panel of genes,

we looked for differences in gene expression that could explain the HIF-induced OXPHOS

down regulation by using large-scale microarray analysis. Mitochondrial biogenesis is

regulated at the level of transcription by expression of several nuclear respiratory factors such

as NRF1 or NRF2, which is stimulated by PGC-1 or PRC [see 37 for review]. The binding of

NRF1 or NRF2 to promoters or enhancers of numerous OXPHOS subunit genes stimulates

their transcription. Although a decrease in respiratory chain proteins is well documented in

renal cancer [2, 3, 15] as well as in many other cancer types, [32, 33], a decrease in the

expression of these transcription factors has not been reported, neither by large scale

microarray nor SAGE analysis comparing tumoral to normal adjacent tissue [38] or pVHL-

deficient renal cancer cells transfected or not with wild type vhl [39]. Our study confirmed

that the expression of these transcription factors or of transcripts coding for OXPHOS

complex subunits did not correspond to the major differences observed in protein contents

when comparing VHL- to VHL+ cells. Microarray data (Tables 1 and Supplementary data) did

not reveal significant differences between VHL- and VHL+ cells for LON protease or COX 4-

2 gene expression. However, our study shows prominent changes in the expression of genes

involved in ROS metabolism such as the mitochondrial Mn-SOD or the LOX 4 lysyl oxidase.

Importantly, this was paralleled by protein contents of Mn-SOD (Figure 2D), which is located

in the matrix where it dismutates superoxide anions into hydrogen peroxide.

Several reports have implied ROS production in HIF-α stabilization in various types

of cancer cells [see 16, 40, for reviews]. According to Schroedl et al, [41] an increase in ROS

production during hypoxia can control HIF-α activation. Furthermore, Okamoto et al [42]

observed a higher content of ROS-modified nucleotides and proteins in human renal cell

carcinoma showing that RCC constitutively elaborates more ROS than is produced by the

non-tumorous parts of kidneys. Accordingly, our data also showed that the absence of pVHL,



ownloaded from


which results in HIF stabilization is associated with a ROS increase inducing a higher DCF

fluorescence (Figure 2D) and a decreased aconitase activity (Figure 2E). In respiratory chain,

the ROS production originates from one electron leak reacting with oxygen to give primary

oxygen radicals, superoxide anions [43]. The increased ROS production by mitochondrial

respiratory chain is often compensated by up-regulation of intramitochondrial superoxide

dismutase (Mn-SOD) and other scavenging mechanisms. The increase in Mn-SOD, observed

by proteomic analysis of CCRC compared to normal adjacent tissue [44, 45] or in VHL- cells

in our study (Figure 2C) would then cause conversion of intramitochondrially generated

superoxide anions in these cells thus increasing the level of hydrogen peroxide. The Mn-SOD

increase might be a mechanism that the cell developed to survive. Indeed, the VHL- cells are

more susceptible to H2O2 toxicity than VHL+ cells (Figure 4A). Since an increase in H2O2

destroyed respiratory chain subunits (Figure 3A and 3B), there has to be an autoregulation

between levels of HIF, ROS and OXPHOS subunit content. ROS production stabilized HIF

but decreased OXPHOS subunit contents. This can in turn decrease ROS production through

complex I and III downregulation [46].

The mitochondrial Mn-SOD increase we observed in VHL- cells (Figure 2D) leads to

enhanced conversion of superoxide radicals to molecular oxygen and H2O2. H2O2 being

membrane permeable can then be carried out towards other cell compartments and be

converted to water in the presence of catalase or peroxidases present in these compartments.

However, the toxicity vastly increases when H2O2 is transformed into hydroxyl radicals by

Fenton reactions, which exactly seems to occur in VHL- cells. This hypothesis is strongly

supported by experiments with DFO, which prevents Fenton reactions by complexing iron

[29] and hence inhibits the production of hydroxyl radicals. This explains why DFO treatment

did not decrease respiratory chain subunit amounts in VHL+ cells, whereas the other HIF-

stabilizing compound CoCl2 decreased them similarly to what was observed in VHL- cells.



ownloaded from


This shows that ROS enhanced by HIF stabilization are important factors responsible for HIF-

induced respiratory chain down regulation. The fact that DFO protects the VHL- cells against

H2O2-induced degradation of COX 4 subunit (Figure 4C, D) suggests that the ROS species

most important in OXPHOS subunit degradation are the hydroxyl radicals produced by the

Fenton reactions. The scheme shown in Figure 5 summarizes the various ways discussed in

this paper by which HIF and ROS modify energy metabolism.

While the mechanisms by which ROS modulate HIF stability and regulation of target

gene expression have been thoroughly investigated in previous reports, this paper shows that

the ROS increase induced by HIF stabilization appears to be essential to down regulate the

level of respiratory chain subunits. However, the reason why HIF increases ROS production

remains hypothetical. ROS production might be a consequence of hindered electron transfer

through respiratory chain complexes. The alteration of respiratory chain complexes involves

changes in electron flux. Decreasing COX activity decreases the rate of oxygen utilization and

leaves more oxygen available for superoxide anions production through complex III or

complex I. That a decrease in complex III can increase ROS production is more puzzling.

This paradox has already been observed in mitochondrial diseases involving cytochrome b

mutations and showing increase in ROS production. In fact, we cannot also exclude that the

down-regulation was connected with some qualitative changes in the complex thus promoting

electron leak and ROS production.

Although ROS are important in stabilization of HIF that maintains the transcription of

genes involved in tumor development, a ROS excess is toxic for the cancer cell and ROS

level should be precisely regulated. Similarly, Burdon et al [47] had shown that a slight H2O2

increase conferred a proliferative advantage to cancer cells while higher H2O2 concentrations

were toxic. Since one of the main sources of ROS is provided by the respiratory chain

complex I and complex III [46, 48, 49], the decrease in OXPHOS may be a crucial step in the



ownloaded from


development of tumor cells to avoid excessive ROS toxicity. Direct ROS increase after H2O2

addition to VHL- cells also had a deadly effect to the cells. In conclusion, tumor cells must

finely tune the level of ROS up-regulating them to stabilize HIF and preventing their excess

to avoid cell death. The ROS-induced mechanism contributing to OXPHOS down-regulation

is probably one of many parallel mechanisms that take place in cancer cell proliferation

during hypoxia adaptation.

References

[1] Warburg,O. (1956) On respiratory impairment in cancer cells. Science, 124, 269-70.

[2] Simonnet,H., Alazard,N., Pfeiffer,K., Gallou,C., Beroud,C., Demont,J., Bouvier,R.,

Schagger,H. and Godinot,C. (2002) Low mitochondrial respiratory chain content correlates

with tumor aggressiveness in renal cell carcinoma. Carcinogenesis, 23, 759-68.

[3] Meierhofer,D., Mayr,J.A., Foetschel,U., Berger,A., Fink,K., Schmeller,N., Hacker,G.W.,

Hauser-Kronberger,C., Kofler,B., and Sperl, W. (2004) Decrease of mitochondrial DNA

content and energy metabolism in renal cell carcinoma. Carcinogenesis, 25, 1005-10.

[4] Hervouet,E., Demont,J., Pecina,P., Vojtiskova,A., Houstek,J., Simonnet,H. and

Godinot,C. (2005) A new role for the von Hippel-Lindau tumor suppressor protein,

stimulation of mitochondrial oxidative phosphorylation complex biogenesis. Carcinogenesis,

26, 531-9.

[5] Craven,R.A., Stanley,A.J., Hanrahan,S., Dods,J., Unwin,R., Totty,N., Harnden,P.,

Eardley,I., Selby,P.J., and Banks,R.E. (2006) Proteomic analysis of primary cell lines

identifies protein changes present in renal cell carcinoma. Proteomics, 6, 3880-993.



ownloaded from


[6] Wang,G.L., Jiang,B.H., Rue,E.A. and Semenza,G.L. (1995) Hypoxia-inducible factor 1 is

a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc. Natl. Acad.

Sci. U S A, 92, 5510-4.

[7] Jaakkola,P., Mole,D.R., Tian,Y.M., Wilson,M.I., Gielbert,J., Gaskell,S.J.,

Kriegsheim,A.V., Hebestreit,H.F., Mukherji,M., Schofield,C.J., Maxwell,P.H., Pugh,C.W.

and Ratcliffe,P.J. (2001) Targeting of HIF-alpha to the von Hippel-Lindau ubiquitylation

complex by O2-regulated prolyl hydroxylation. Science, 292, 468-72.

[8] Semenza,G.L. (2004) Hydroxylation of HIF-1, oxygen sensing at the molecular level.

Physiology, 19, 176-82.

[9] Salnikow,K., Donald,S.P., Bruick,R.K., Zhitkovich,A., Phang,J.M. and Kasprzak,K.S.

(2004) Depletion of intracellular ascorbate by the carcinogenic metals nickel and cobalt

results in the induction of hypoxic stress. J. Biol. Chem., 279, 40337-44.

[10] Semenza,G.L., Jiang,B.H., Leung,S.W., Passantino,R., Concordet,J.P., Maire,P.,

Giallongo,A.(1996) Hypoxia response elements in the aldolase A, enolase 1, and lactate

dehydrogenase A gene promoters contain essential binding sites for hypoxia-inducible factor

1. J. Biol. Chem., 271, 32529-37.

[11] Kim,J.W., Tchernyshyov,I., Semenza,G.L. and Dang,C.V. (2006) HIF1--mediated

expression of pyruvate dehydrogenase kinase, a metabolic switch required for cellular

adaptation to hypoxia. Cell Metab., 3, 177-85.

[12] Fukuda,R., Zhang,H., Kim,J.W., Shimoda,L., Dang,C.V. and Semenza, G.L. (2007) HIF-

1 regulates cytochrome oxidase subunits to optimize efficiency of respiration in hypoxic cells.

Cell 129, 111-22.

[13] Hüttemann M, Kadenbach B, Grossman LI. (2001) Mammalian subunit IV

isoforms of cytochrome c oxidase. Gene, 267, 111-23.



ownloaded from


[14] Hervouet,E., Pecina,P., Demont,J., Vojtiskova,A., Simonnet,H., Houstek,J. and

Godinot,C. (2006) Inhibition of cytochrome c oxidase subunit 4 precursor processing by the

hypoxia mimic cobalt chloride. Biochem. Biophys. Res. Commun., 344, 1086-93.

[15] Sowter,H.M., Raval,R.R., Moore,J.W., Ratcliffe,P.J. and Harris,A.L. (2003) Predominant

role of hypoxia-inducible transcription factor (HIF)-1alpha versus HIF2-alpha in regulation of

the transcriptional response to hypoxia. Cancer Res., 63, 6130-4.

[16] Guzy,R.D. and Schumacker,P.T. (2005) Oxygen sensing by mitochondria at complex

III, the paradox of increased reactive oxygen species during hypoxia. Experimental

physiology, 91, 807-819.

[17] Yu,C.A., Xia,D., Kim,H., Deisenhofer,J., Kachurin,A.M., Zhang,L., Deng,K.P., Yu,L.

(1998) Three-dimensional structure and functions of bovine heart mitochondrial cytochrome

bc1 complex. Biofactors, 8, 187-9.

[18] Iliopoulos,O., Kibel,A., Gray,S. and Kaelin,W.G. Jr. (1995) Tumour suppression by the

human von Hippel-Lindau gene product. Nat. Med., 1, 822-6.

[19] Pellis,L., Franssen-van Hal,N.L., Burema,J. and Keiger,J. (2003) The intraclass

correlation coefficient applied for evaluation of data correction, labeling methods and rectal

biopsy sampling in DNA microarray experiments. Physiol. Genomics, 16, 99-106.

[20] Wang,P., Renes,J., Bouwman,F., Bunschoten,A., Mariman,E. and Keijer, J. (2007)

Absence of an adipogenic effect of rosiglitazone on mature 3T3-L1 adipocytes, increase of

lipid catabolism and reduction of adipokine expression. Diabetologia, 50, 654–65.

[21] Smyth,G.K., Michaud,J.and Scott,H.S. (2005) Use of within-array replicate spots for

assessing differential expression in microarray experiments. Bioinformatics, 21, 2067-75.

[22] Benjamini,Y.and Hochberg,Y. (1995) Controlling the false discovery rate, A practical

and powerful approach to multiple testing. J. Roy. Stat. Soc. B., 57, 289-300.



ownloaded from


[23] Arrigo,A.P., Firdaus,W.J., Mellier,G., Moulin,M.,Paul,C., Diaz-latoud,C. and Kretz-

Remy,C. (2005) Cytotoxic effects induced by oxidative stress in cultured mammalian cells

and protection provided by Hsp27 expression. Methods, 35, 126-38.

[24] Mahadev,K., Motoshima,H., Wu,X., Ruddy,J.M., Arnold,R.S., Cheng,G., Lambeth,J.D.

and Goldstein,B.J. (2004) The NAD(P)H oxidase homolog Nox4 modulates insulin-

stimulated generation of H2O2 and plays an integral role in insulin signal transduction. Mol.

Cell. Biol., 24, 1844-54.

[25] Schneider,B.D. and Leibold,E.A. (2003) Effects of iron regulatory protein regulation on

iron homeostasis during hypoxia. Blood, 102, 3404-11.

[26] Hanson,E.S., Rawlins,M.L.and Leibold, E.A. (2003) Oxygen and iron regulation of iron

regulatory protein 2. J. Biol. Chem., 278, 40337-42.

[27] Erler,J.T., Bennewith,K.L., Nicolau,M., Dornhofer,N., Kong,C., Le,Q.T., Chi,J.T.,

Jeffrey, S.S.and Giaccia,A.J. (2006) Lysyl oxidase is essential for hypoxia-induced

metastasis. Nature, 440, 1222-6.

[28] McLennan,H.R., and Degli Esposti,M. (2000) The contribution of mitochondrial

respiratory complexes to the production of reactive oxygen species. J. Bioenerg. Biomembr.,

32, 153-62.

[29] Tarpey,M.M., Wink,D.A., Grisham, M.B. (2004) Methods for detection of reactive

metabolites of oxygen and nitrogen: in vitro and in vivo considerations. Am. J. Physiol. Regul.

Integr. Comp. Physiol., 286, R431-44.

[30] Yan,L.-J., Levinedagger,R.L. and Sohal,R.S. (1997) Oxidative damage during aging

targets mitochondrial aconitase. Proc.Natl. Acad. Sci. USA, 94, 11168-11172.

[31] Mahadev,K., Motoshima,H., Wu,X., Ruddy,J.M., Arnold,R.S., Cheng,G., Lambeth,J.D.

and Goldstein,B.J. (2004) The NAD(P)H oxidase homolog Nox4 modulates insulin-



ownloaded from


stimulated generation of H2O2 and plays an integral role in insulin signal transduction. Mol.

Cell. Biol., 24, 1844-54.

[32] Pedersen,P.L., Mathupala,S., Rempel,A., Geschwind,J.F. and Ko,Y.H. (2002)

Mitochondrial bound type II hexokinase, a key player in the growth and survival of many

cancers and an ideal prospect for therapeutic intervention. Biochim. Biophys. Acta, 1555, 14-

20.

[33] Cuezva,J.M., Krajewska,M., de Heredia,M.L., Krajewski,S., Santamaria,G., Kim,H.,

Zapata,J.M., Marusawa,H., Chamorro,M. And Reed, J.C. (2002) The bioenergetic signature

of cancer, a marker of tumor progression. Cancer Res., 62, 6674-81.

[34] Semenza,G. L. (2003) Targeting HIF1- for cancer therapy. Nat. Rev. Cancer, 3, 721–32.

[35] Gordan,J.D., Bertout,J.A., Hu,C.J., Diehl,J.A. and Simon,M.C. (2007) HIF-2alpha

promotes hypoxic cell proliferation by enhancing c-myc transcriptional activity. Cancer Cell,

11, 335-47.

[36] Hüttemann,M., Lee,I., Liu,J., and Grossman,L.I. (2007) Transcription of mammalian

cytochrome c oxidase subunit IV-2 is controlled by a novel conserved oxygen responsive

element. FEBS J., 274, 5737-48.

[37] Scarpulla,R.C. (2002) Nuclear activators and co-activators in mammalian mitochondrial

biogenesis. Biochim. Biophys. Acta, 1576, 1-14.

[38] Boer,J.M., Huber,W.K., Sultmann,H., Wilmer,F., von Heydebreck,A., Haas,S., Korn,B.,

Gunawan,B., Vente,A., Fuzesi,L., Vingron,M. and Poustka,A. (2001) Identification and

classification of differentially expressed genes in renal cell carcinoma by expression profiling

on a global human 31,500-element cDNA array. Genome Res., 11, 1861-70.

[39] Galban,S., Fan,J., Martindale,J.L., Cheadle,C., Hoffman,B., Woods,MP., Temeles,G.,

Brieger,J., Decker,J. and Gorospe,M. (2003) von Hippel-Lindau protein-mediated repression



ownloaded from


of tumor necrosis factor alpha translation revealed through use of cDNA arrays. Mol. Cell.

Biol., 23, 2316-28.

[40] Pouyssegur,J. and Mechta-Grigoriou,F. (2006) Redox regulation of the hypoxia inducible

factor. Biol. Chem., 387, 1337-1346.

[41] Schroedl,C., McClintock,D.S., Budinger,G.R. and Chandel,N.S. (2002) Hypoxic but not

anoxic stabilization of HIF1- alpha requires mitochondrial reactive oxygen species, Am. J.

Physiol. Lung Cell Mol. Physiol., 283, L922–31.

[42] Okamoto,K., Toyokuni,S., Uchida,K., Ogawa,O., Takenewa,J., Kakehi,Y., Kinoshita,H.,

Hattori-Nakakuki,Y., Hiai,H. and Yoshida, O. (1994) Formation of 8-hydroxy-2'-

deoxyguanosine and 4-hydroxy-2-nonenal-modified proteins in human renal-cell carcinoma.

Int. J. Cancer., 58, 825-9.

[43] Boveris,A., Docampo,R., Turrens,J.F. and Stoppani,A.O. (1978) Effect of

beta-lapachone on superoxide anion and hydrogen peroxide production in

Trypanosoma cruzi. Biochem J., 175, 431-9.

[44] Sarto,C., Deon,C., Doro,G., Hochstrasser,D.F., Mocarelli,P. and Sanchez,J.C. (2001)

Contribution of proteomics to the molecular analysis of renal cell carcinoma with an emphasis

on manganese superoxide dismutase. Proteomics, 2, 1288-94.

[45] Unwin,R.D., Craven,R.A., Harnden,P., Hanrahan,S., Totty,N., Knowles,M., Eardley,I.,

Selby,P.J. and Banks,R.E. (2003) Proteomic changes in renal cancer and co-ordinate

demonstration of both the glycolytic and mitochondrial aspects of the Warburg effect.

Proteomics, 3, 1620-32.

[46] Genova,M.L., Pich,M.M., Biondi,A., Bernacchia,A., Falasca,A., Bovina,C.,

Formiggini,G., Parenti Castelli,G. and Lenaz,G. (2003) Mitochondrial production of oxygen



ownloaded from


radical species and the role of Coenzyme Q as an antioxidant. Exp. Biol. Med. (Maywood),

228, 506-13.

[47] Burdon,R.H. (1995) Superoxide and hydrogen peroxide in relation to mammalian cell

proliferation. Free Radic Biol Med., 18, 775-94.

[48] Chandel,N.S., McClintock,D.S., Feliciano,C.E., Wood,T.M., Melendez,J.A.,

Rodriguez,A.M. and Schumacker, P.T., (2000) Reactive oxygen species generated at

mitochondrial complex III stabilize hypoxia-inducible factor-1alpha during hypoxia, a

mechanism of O2 sensing, J. Biol. Chem., 275, 25130–8.

[49] Bell,E.L., Emerling,B.M. and Chandel,N.S. (2005) Mitochondrial regulation of oxygen

sensing. Mitochondrion, 5, 322-32.

[50] Drapier,J.C. and Hibbs,J.B. (1996) Aconitases: a class of metalloproteins highly sensitive

to nitric oxide synthesis. Methods Enzymol., 269, 26-36.



ownloaded from


Figure legends

Figure 1: Increase of respiratory chain subunit contents in clear cell renal carcinoma cells

after inhibition of HIF2-α mRNA expression by treatment with siRNA targeted to HIF2-α: A.

VHL-deficient 786-0 cells transfected with a void vector (786-0 PRC3: VHL-) or with a

vector expressing HIF2-α (786-0 WT8: VHL+) were treated for 24 h with siRNA targeted to

HIF2-α and the HIF2-α mRNA expression were analyzed by semi-quantitative RT-PCR. The

amplicon amounts observed after N or N+6 PCR cycles were analyzed by agarose gel

electrophoresis and the HIF2-α mRNA contents was compared to control (L32 ribosomal

protein). B. Decrease in HIF2-α protein content 48 h after siRNA treatment: western blot

analysis using an anti-HIF2-α antibody was performed with proteins extracted from 786-0

WT8 (VHL+) or 786-0 PRC3 (VHL-) cells treated or not with siRNA directed to HIF2-α for

24 h or 48 h. Treatment with 250 µM CoCl2 was performed for 4 h. The 24 h- treatment with

the anti HIF2-α siRNA was sufficient to prevent the CoCl2-induced HIF2-α expression.

HIF2-α expression was also prevented 48 h (or 72 h, not shown) after anti HIF2-α siRNA

transfection in VHL- (786-0 PRC3) cells. Loading control: western blot stained with Ponceau

Red indicating similar protein contents in the different samples. C. D. E. Restoration of

respiratory chain subunit contents by the anti HIF2-α siRNA treatment: representative

western blot analysis of OXPHOS subunit contents using COX 4, Core 2 and ATPase β-

subunit (5G11) antibodies (C) performed with 786-0 PRC3 (VHL-) and 786-0 WT8 (VHL+)

treated or not for 72 h with siRNA directed to HIF2-α. Anti-ATPase β subunit antibody was

used as a loading control. Data (D, E) are given as the mean ± SD of five different

experiments.

Figure 2: Influence of VHL presence or HIF-2α stabilization on COX4-1 level and ROS

production. A, B: Effects of CoCl2 or DFO stabilizing HIF2-α on COX4-1 subunit content.



ownloaded from


The COX4-1 subunit content was estimated by western blot, using VHL+ (786-0 WT8) cells

untreated or treated with 250 µM CoCl2 or DFO to stabilize HIF2-α. 20 µg of proteins were

used in each lane. A representative blot is shown in A. The graph B shows the mean ±SD of

three experiments. C: Increased expression of Mn-SOD and in 786-0 cells devoid of pVHL:

the content of superoxide dismutases (Mn-SOD and CuZn-SOD) was quantified by Western

blot. For analysis, 15 µg of protein from cell extracts or 5 µg of mitochondrial protein were

loaded. D. Effects of HIF expression on ROS production, as measured by DCF fluorescence.

The DCF fluorescence intensity was measured as described in the experimental section. Data

are given as means ± SD of 3 to 5 experiments, each one being made in triplicates. CoCl2 was

incubated for 8 hours at a concentration of 250 µM. E, F. Effects of HIF expression on

aconitase activity and contents. Aconitase activity was measured by the method of Drapier

and Hibbs [50] in 786-0 PRC3 cells (VHL-) or 786-0 WT8 cells (VHL+) untreated or treated

for 8 h with 250 µM CoCl2. To lyse the cells, 2 to 5 x 106 cells homogenized in 300 µl of 50

mM Tris-HCl, pH 7.4 containing 0.625 mM MgCl2 were sonicated 10 times for 1 sec (Vibra

Cell 72405 sonicator) and centrifuged at 2,000 g for 6 min. The aconitase amounts were

determined by western blot analysis quantified by densitometry. The different cells and

treatments were compared to the untreated 786-0 WT8 cells taken as 100%. Data are given as

means ± SD of 3 to 9 experiments.

Figure 3: A, B: Effect of H2O2 on COX 4 and Core 2 contents: VHL+ cells were treated for 8

h with increasing H2O2 concentrations. Total proteins (10 or 20 µg) were analyzed by

Western blot using anti COX 4 or anti Core 2 antibodies. B: Estimation of protein amounts

remaining after 8 hrs of treatment with the indicated H2O2 concentrations. The intensity of the

assays tested with the COX4 or Core 2 antibodies in the absence of inhibitor was arbitrary

taken as 100% and the intensities observed in the presence of H2O2 was compared to that of

this reference. Data are given as mean ± SD of 4 experiments. Stars (*P<0.05; **P<0.01)



ownloaded from


indicate significant differences observed between assays made in the presence of the indicated

H2O2 concentrations and the assays made in the absence of H2O2.

Figure 4: Combined effects of H2O2 and DFO on cell growth and COX4-1 contents. A: The

cells were plated in 96-well plates at a density of 104 cells per well and incubated on the

following day for 24 hrs in the presence of H2O2 and/or DFO at the indicated concentrations.

Cell number was estimated by crystal violet staining. Each point represents an average of six

determinations. (Top left): Effects of increasing H2O2 concentrations on the viability of 786-0

cells transfected with a void vector (VHL-) or with a vector expressing functional pVHL

(VHL+). (Top right): Effects of increasing DFO concentrations on the viability of 786-0 cells

transfected with a void vector (VHL-) or with a vector expressing functional pVHL (VHL+).

Influence of DFO at concentrations of 50 or 100 µM on H2O2-induced cell toxicity in VHL-

(Bottom left) or VHL+ (bottom right) cells. B,C,D: Combined effects of H2O2 and DFO on

COX 4-1 subunit contents in VHL- or VHL+ cells. The cells (106 cells per 25 cm2 flask) were

treated for 24 hrs in the presence or absence of 100 µM H2O2 and/or 100 µM DFO. For

analysis, 20 µg of proteins were loaded in each lane. The COX4-1 content was quantified by

Western blot. B: Loading control stained with Ponceau Red showing similar protein contents

in the different samples. C: Western blot obtained using anti COX 4 antibody. In each

experiment, the sample corresponding to VHL+ cells in the absence of H2O2 and DFO

(sample 2) was arbitrarily fixed at 100. The spot intensities were then calculated by

comparison with this arbitrary value. The mean ± SD of three experiments is reported on

graph D. Bars with grey and white background represent VHL- and VHL+ cells, respectively.

Stars (*P<0.05; **P<0.01) indicate significant differences observed between treatments when

comparing assays linked by brackets.

Figure 5: Scheme summarizing the main effects of HIF and of various ROS components on

cell bioenergetics. Thick plain arrows indicate genes targeted by HIF; thick dotted lines with



ownloaded from


arrows show targets for ROS; T means direct inhibition. See introduction for details on

HIF/VHL mechanisms and discussion for ROS involvement.



ownloaded from


Table 1: Modification of gene expression in 786-0 WT8 cells (VHL+) compared to 786-0

parental cells (VHL-) (p < 10-4)

UP-REGULATED GENES

Gene Name Accession

Adjusted

p Value

Fold

Change

1 von Hippel-Lindau tumor suppressor isoform 1 NM_000551 4.58E-11 9.29

2 DEP domain containing 6 NM_022783 4.34E-10 13.07

3 S100 calcium binding protein A1 NM_006271 2.28E-09 5.12

4 alpha-methylacyl-CoA racemase isoform 1 NM_014324 3.20E-08 3.26

5 CD74 antigen isoform b NM_004355 6.79E-08 6.25

6 8D6 antigen NM_016579 9.99E-08 4.74

7 glutathione transferase NM_000852 1.96E-07 8.22

8 alpha 1 type IV collagen preproprotein NM_001845 2.42E-07 4.34

9 placenta-specific 8 NM_016619 1.05E-06 5.64

10 protein kinase C. epsilon NM_005400 1.68E-06 3.50

11 crystallin. alpha B NM_001885 1.98E-06 3.09

12 NADP-dependent retinol dehydrogenase/reductase NM_005771 3.04E-06 4.27

13 interferon. alpha-inducible protein 27 isoform NM_032036 3.36E-06 5.76

14 mitochondrial ribosomal protein S35 NM_021821 4.25E-06 2.33

15 hypothetical protein LOC55147 NM_018107 1.04E-05 2.94

16 secreted phosphoprotein 1 NM_000582 1.04E-05 3.77

17 solute carrier family 20 (phosphate NM_005415 1.20E-05 2.79

18 FERM. RhoGEF. and pleckstrin domain protein 1 NM_005766 2.14E-05 2.49

19 protein tyrosine phosphatase. non-receptor type NM_002830 2.75E-05 2.26

20 S-adenosylhomocysteine hydrolase NM_000687 3.27E-05 2.37

21 RAP guanine-nucleotide-exchange factor 3 NM_006105 4.07E-05 2.40

22 ubiquitin carboxyl-terminal esterase L1 NM_004181 4.10E-05 2.44



ownloaded from


23 hypothetical protein LOC54502 NM_019027 4.30E-05 2.59

24 carboxypeptidase E precursor NM_001873 4.88E-05 2.70

25 phosphatidate cytidylyltransferase 1 U60808 6.28E-05 2.22

26 claudin 16 NM_006580 7.49E-05 2.38

27 atrophin-1 NM_001940 7.68E-05 2.04

DOWN-REGULATED GENES

Gene Name Accession

Adjusted

p Value

Fold

Change

1 dedicator of cytokinesis 10 NM_017718 7.30728E-13 7.2

2 solute carrier family 43. member 3 NM_014096 2.72146E-12 14.4

3 chemokine (C-X-C motif) ligand precursor NM_002993 5.41434E-11 35.1

4 lysyl oxidase preproprotein NM_002317 5.41434E-11 12.5

5 retinoic acid receptor responder (tazarotene NM_002889 5.41434E-11 10.3

6 keratin 19 NM_002276 2.41972E-10 10.4

7 interleukin 8 precursor NM_000584 1.19334E-09 41.7

8 tumor necrosis factor (ligand) superfamily. NM_003810 1.30794E-09 13.1

9 - AJ011772 1.81134E-08 6.6

10 EGF-like repeats and discoidin I-like NM_005711 2.26346E-08 6.9

11 zinc finger. DHHC domain containing 4 NM_018106 2.63261E-08 3.7

12 ADP-ribosylation factor-like 7 NM_005737 5.09985E-08 6.7

13 - NM_024724 1.14234E-07 4.7

14 CD44 antigen precursor NM_000610 1.20183E-07 4.3

15 gap junction protein. beta 2. 26kDa (connexin NM_004004 1.21195E-07 4.1

16 transforming growth factor. beta-induced. 68kDa NM_000358 1.74752E-07 5.0

17 cadherin 2. type 1 preproprotein X54315 1.78777E-07 4.7

18 vascular cell adhesion molecule NM_001078 1.78777E-07 8.5

19 insulin receptor substrate 1 NM_005544 1.9619E-07 3.4

20 5' nucleotidase. ecto NM_002526 2.69075E-07 5.2

21 dihydropyrimidinase-like 3 NM_001387 2.71377E-07 5.3



ownloaded from


22 sushi-repeat-containing protein. X-linked NM_006307 3.19297E-07 3.9

23 cadherin 2. type 1 preproprotein NM_001792 3.34252E-07 3.2

24 cyclin-dependent kinase 6 NM_00125 9 4.30299E-07 4.5

25 UDP glycosyltransferase 1 family. polypeptide AF297093 6.12535E-07 5.4

26 interleukin 6 (interferon. beta 2) NM_000600 6.82333E-07 8.1

27 sparc/osteonectin. cwcv and kazal-like domains NM_004598 8.03785E-07 3.6

28 - AB028974 8.22674E-07 5.3

29 mitogen-inducible gene 6 protein NM_018948 1.05461E-06 4.3

30 microtubule associated monoxygenase. calponin NM_014632 1.31417E-06 3.8

31 tetraspan NET-6 NM_014399 1.32853E-06 3.8

32 a disintegrin and metalloprotease with NM_007038 2.14011E-06 3.3

33 plasminogen activator inhibitor-1 NM_000602 2.14011E-06 5.0

34 tissue factor pathway inhibitor 2 NM_006528 2.14011E-06 6.2

35 zygin 1 isoform 1 NM_022549 2.14011E-06 3.3

36 manganese superoxide dismutase isoform A NM_000636 2.46854E-06 7.2

37 leupaxin NM_004811 3.36264E-06 4.4

38 adipocyte-specific adhesion molecule AK026068 4.25407E-06 2.7

39 dual specificity phosphatase 10 isoform a NM_007207 5.76809E-06 3.2

40 putative NFkB activating protein 373 isoform 1 NM_024911 6.26746E-06 3.2

41 semaphorin 3C NM_006379 7.79244E-06 3.1

42 matrix metalloproteinase 2 preproprotein NM_004530 7.82768E-06 4.2

43 toll-like receptor 2 NM_003264 1.00966E-05 2.3

44 chemokine (C-X-C motif) ligand 1 BC015753 1.03528E-05 6.2

45 cyclin D1 NM_001758 1.09045E-05 4.2

46 fibronectin 1 isoform 1 preproprotein U42593 1.33447E-05 5.9

47 protein tyrosine phosphatase. receptor type. B NM_002837 1.35149E-05 3.1

48 transglutaminase 2 isoform a NM_004613 1.54398E-05 3.4

49 tumor necrosis factor. alpha-induced protein 8 NM_014350 2.09543E-05 3.0

50 - NM_002089 2.14424E-05 5.0

51 moesin NM_002444 2.25658E-05 3.8



ownloaded from


52 follistatin-like 1 precursor NM_007085 2.32619E-05 3.1

53 insulin receptor substrate 2 NM_003749 2.32619E-05 3.2

54 melanophilin NM_024101 2.56728E-05 3.0

55 ubiquitin-conjugating enzyme E2L 6 isoform 1 NM_004223 2.74716E-05 2.7

56 v-ets erythroblastosis virus E26 oncogene NM_005238 2.74716E-05 2.4

57 myristoylated alanine-rich protein kinase C XM_047795 3.46722E-05 2.3

58 membrane alanine aminopeptidase precursor NM_001150 3.78546E-05 2.7

59 transmembrane prostate androgen-induced protein NM_020182 4.09053E-05 2.5

60 adipose differentiation-related protein NM_001122 4.12914E-05 2.9

61 solute carrier family 1. member 1 NM_004170 4.12914E-05 2.2

62 endothelial differentiation. lysophosphatidic NM_001401 4.64841E-05 2.1

63 nidogen 2 NM_007361 4.78417E-05 4.8

64 colony stimulating factor 2 precursor NM_000758 4.82949E-05 5.9

65 colony stimulating factor 1 precursor NM_000757 5.15274E-05 3.3

66 pleckstrin homology-like domain. family A. NM_007350 5.18454E-05 3.3

67 prostaglandin H2 D-isomerase NM_000954 0.000052799 8.1

68 forkhead box P1 isoform 1 NM_016477 6.16318E-05 2.0

69 myoferlin isoform b NM_013451 6.82454E-05 3.0

70 cytochrome P450. family 1. subfamily B. NM_000104 7.48794E-05 2.6

71 dickkopf homolog 3 precursor NM_015881 7.48794E-05 2.1

72 EGF-containing fibulin-like extracellular matrix NM_016938 7.51841E-05 2.4

73 CUB domain-containing protein 1 isoform 1 NM_022842 7.76466E-05 2.5

74 apolipoprotein L1 isoform a precursor NM_003661 8.02538E-05 2.6

75 ring finger protein 12 NM_016120 8.03288E-05 2.1

76 large conductance calcium-activated potassium NM_002247 8.82167E-05 3.0

77 prostaglandin E receptor 2 (subtype EP2). 53kDa NM_000956 0.000097913 2.3



ownloaded from




ownloaded from


HIF and reactive oxygen species regulate oxidative phosphorylation in cancer

Documents

Transcript of HIF and reactive oxygen species regulate oxidative phosphorylation in cancer