(2010) 29, 4362–4368 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 www.nature.com/onc SHORT COMMUNICATION Hypoxia-regulated microRNA-210 modulates mitochondrial function and decreases ISCU and COX10 expression

Z Chen1,YLi1, H Zhang2, P Huang2 and R Luthra1

1Department of Hematopathology, University of Texas MD Anderson Center, Houston, TX, USA and 2Department of Molecular Pathology, University of Texas MD Anderson Cancer Center, Houston, TX, USA

The mechanisms of compromised mitochondrial function formation and downregulation of mitochondrial activity under various pathological conditions, including hypoxia, through pyruvate dehydrogenase kinase-1 (Papandreou remain largely unknown. Recent studies have shown that et al., 2006; Bristow and Hill, 2008). microRNA-210 (miR-210) is induced by hypoxia under the MicroRNAs (miRNAs) are small noncoding regulation of hypoxia-inducible factor-1a and has an that silence the expression of target through either important role in cell survival under hypoxic microenviron- mRNA degradation or suppression of ment. Hence, we hypothesized that miR-210 has a role in procedure. miRNAs have roles in almost all aspects of regulating mitochondrial metabolism and investigated miR- cell biology, and their deregulation has been reported in 210 effects on mitochondrial function in cancer cell lines various diseases, especially cancer (Lee and Dutta, 2009). under normal and hypoxic conditions. Our results demon- Recently, a specific group of hypoxia-regulated miRNAs strate that miR-210 decreases mitochondrial function and that are under the regulation of HIF-1a was identified. upregulates the glycolysis, thus make cancer cells more miR-210 is one of the most hypoxia-sensitive miRNAs, sensitive to glycolysis inhibitor. miR-210 can also activate and its overexpression protected breast cancer cells from the generation of reactive oxygen species (ROS). ISCU hypoxia-induced (Kulshreshtha et al., 2007). (iron-sulfur cluster scaffold homolog) and COX10 (cyto- miR-210 has also been shown to have a role in chrome c oxidase assembly ), two important factors regulation (Giannakakis et al., 2008) and DNA damage of the mitochondria and the and repair (Crosby et al.,2009).However,whether tricarboxylic acid cycle have been identified as potential miR-210 has role in compromised mitochondrial function targets of miR-210. The unique means by which miR-210 under hypoxic conditions remain unknown. We thus regulates mitochondrial function reveals an miRNA- hypothesized that miR-210 regulates mitochondrial func- mediated link between microenvironmental stress, oxidative tion by targeting key electron transport/tricarboxylic acid phosphorylation, ROS and iron homeostasis. cycle genes and investigated its effects on mitochondrial Oncogene (2010) 29, 4362–4368; doi:10.1038/onc.2010.193; function in cancer cells. published online 24 May 2010

Keywords: miR-210; mitochondria; hypoxia; COX10; Results and Discussion ISCU; ROS Hypoxia and chemical mimetic regulate miR-210 We first confirmed that miR-210 expression is upregulated by hypoxia in different cancer cell lines using quantitative Introduction real-time PCR analysis. Cell lines tested included the colon cancer cell lines HCT116, HCT116 p53À/À Hypoxia is known to have an important role in various (HCTÀ/À), HT29 and RKO; breast cancer cell line physiological processes and pathological conditions, MCF-7; esophageal adenocarcinoma cell line BE3. The including cancer, Parkinson’s disease, myocardial ischemia/ HCTÀ/À cells were derived from HTC116 cells with reperfusion and stroke. A hypoxic microenvironment wild-type through somatic disruption of p53 (Bunz can cause mitochondrial dysfunction and activate et al., 1998). In agreement with previous publications various signaling pathways through hypoxia-inducible (Kulshreshtha et al., 2007), substantial induction of factor-1a (HIF-1a). The activation of HIF-1a can miR-210 expression was observed in every cell line tested trigger various biological events, including activation (Figure 1a). The iron chelator deferoxamine mesylate of the glycolytic pathway, increased blood vessel (DFO), widely used as a hypoxia mimetic that induces accumulation of HIF-1a, also induced miR-210 expres- sion in the same set of cell lines (Figure 1b). Correspondence: Professor R Luthra, Department of Hematopathol- ogy, University of Texas MD Anderson Cancer Center, 8515 Fannin miR-210 downregulates mitochondrial respiration and St., NAO1.061a, Houston, TX 77054, USA. E-mail: [email protected] mediates metabolic switch to glycolysis Received 21 August 2009; revised 13 April 2010; accepted 15 April 2010; To test whether miR-210 is able to regulate mitochon- published online 24 May 2010 drial respiratory activity, HCT116 cells were transfected miR-210 regulates mitochondria and targets ISCU COX10 ZChenet al 4363

p=0.0135 0.4 p=0.0401 p=0.0068 Negative control 21 p=0.0021 DFO 12 hours Normoxia p=0.0067 18 Hypoxia 24 hours p=0.0365 0.3 15

12 0.2 9 p=0.0096 p=0.0058 p=0.0632 p=0.0007 p=0.0049 6 p=0.0584 0.1 miR-210 Fold induction 3

0 0.0 HCT116 HCT -/- HT29 RKO MCF-7 BE3 miR-210 Relative expression 2^(-dCT) HCT116 HCT -/- HT29 RKO MCF-7 BE3 Figure 1 miR-210 induction caused by hypoxia or DFO treatment. (a) reverse transcription (RT)–PCR results of miR-210 induction in colon cancer, breast cancer and esophageal cancer cells under hypoxia for 24 h. (b) RT–PCR results of miR-210 induction in cells treated with DFO for 12 h. Data shown are the means and s.d.s of triplicate quantitative real-time PCR reactions. Corresponding P-values are shown.

Negative control Negative control 240 miR-210 125 p=0.1688 miR-210 transfection p=0.0495 220 100

75 200 50

Oxygen (nmol/ml) 180 Relative Oxygen 25 Consumption Rate % 160 0 1 234 5678 0 Time (min) 24 Hrs 48 Hrs

Negative control Negative control 120 miR-210 0.8 miR-210 p=0.0020 100 p=0.0066 cells 5 0.6 80

60 0.4 40 0.2 20 Survival cells, % controlal 0.0 0 Lactate mmol/L/1X 10 24 Hrs 48 Hrs 0 102030405070 3-BrOP (μM) Figure 2 miR-210 can decrease mitochondrial respiration and activate glycolysis activity. (a) Oxygen consumption assays were performed on HCT116 cells treated with miR-210 mimic and negative control for 24 and 48 h. Left panel is representative of the original data; right panel shows the means and s.d.s of duplicate experiments. (b) Cellular lactate production of HCT116 cells treated by miR-210 mimic and mimic-negative control were analyzed at 24 and 48 h. Data shown are the means and s.d.s of triplicate experiments. Corresponding P-values are shown. (c) MTT assay of miR-210 mimic-transfected HCT116 cells treated with 3-BrOP. with miRNA mimic of miR-210 and oxygen consump- resulted in increased lactate levels compared with the tion activity was measured. Exposure to miR-210 mimic mimic-negative control. for 48 h resulted in an approximately 24% reduction in Cytotoxicity assay indicated that this metabolic the oxygen consumption rate (Figure 2a). Similar result switch was accompanied by an elevated sensitivity to was observed in HCTÀ/À cells (data not shown). 3-brommo-2-oxopropionate-1-propyl ester (3-BrOP), Lactate acid, the end product of the glycolysis which is a potent inhibitor of the glycolysis pathway pathway, increases under circumstances of cellular targeting the hexokinase (Figure 2c). 3-BrOP is hypoxia or mitochondrial dysfunction because cells known to preferentially kill cancer cells with elevated have to upregulate glycolysis activity to meet their glycolytic activity (Chen et al., 2007). These data demand for ATP and metabolic intermediates. If miR- suggested that miR-210 can trigger the metabolic switch 210 does inhibit mitochondrial respiratory activity, one from aerobic respiration to glycolysis. This metabolic would expect to observe elevated lactate levels. Hence, switch was unlikely due to the initiation of cell death, as we measured lactate levels in culture medium after the these cells did not undergo apoptosis, although cell treatment of HCT116 cells with miR-210 mimic. As proliferation did appear to slow down, as described illustrated in Figure 2b, treatment with miR-210 mimic below (Figure 3).

Oncogene miR-210 regulates mitochondria and targets ISCU COX10 Z Chen et al 4364 Negative control Negative control miR-210 transfection 150 miR-210 24 hours P=0.0091 miR-210 48 hours 2 P=0.0233 100 P=0.4845 1 50 Fluorescence % Absorbance (570 nm) Relative Rhodamine 123 0 0 24 Hrs 48 Hrs 72 Hrs HCT116 HCT -/-

Negative control 2.0 Negative Control 3 miR-210 transfection 24 hours miR-210 transfection miR-210 transfection 48 hours P=0.0116 Negative control P=0.1412 1.5 p=0.0328 2 P=0.1112 p=0.0151 miR-210 mimic P=0.0111 1.0

1 peroxide level 0.5 Relative Hydrogen

0 Relative Superoxide Level 0.0 HCT116 HCT -/- HCT116 HCT -/- Figure 3 Effect of miR-210 on cell proliferation, cellular ROS and mitochondrial membrane potential. (a) Cell proliferation assay. HCT116 cells were transfected with miR-210 mimic and mimic-negative control. At three different time points post-transfection, cell proliferation was analyzed using MTT assay. (b) The change in mitochondrial membrane potential of HCT116 and HCTÀ/À cells treated with miR-210 mimic was analyzed by flow cytometry using fluorescent dye rhodamine-123. (c)H2O2 levels of HCT116 and HCTÀ/À cells treated with miR-210 mimic for 24 and 48 h were measured by flow cytometry using fluorescent dye DCF-DA. Left panel is representative of the original data; right panel shows the means and s.d.s of duplicate experiments. (d) Superoxide levels of HCT116 and HCTÀ/À cells treated with miR-210 mimic for 24 h were measured by flow cytometry using fluorescent dye dihydroethidium. Data shown are the means and s.d.s of duplicate experiments. Corresponding P-values are shown. H2O2, hydrogen peroxide. DCF-DA, CM-H2DCFDA.

Effects of miR-210 on cell proliferation, membrane 48 h after transfection. The amount of H2O2 induced potential and ROS was more significant than that of superoxide, and miR-210 mimic started having a moderate inhibitory similar observations were made in HCT116 and effect on the proliferation rate of HCT116 cells 48 h HCTÀ/À cells. One possible explanation is that post-transfection (Figure 3a). This observation is accumulated superoxide is readily converted into H2O2 supported by a recent publication showing that miR- by called superoxide dismutases. These results 210 has a role in cell cycle regulation (Giannakakis indicate that miR-210 negatively regulates mitochon- et al., 2008). Meanwhile, miR-210 had no noticeable drial respiration activity and activates ROS generation, effect on mitochondrial membrane potential in HCT116 which may contribute to cell survival under hypoxic or HCTÀ/À cells within 48 h after transfection microenvironment. (Figure 3b). Mitochondrial membrane potential is an important parameter of mitochondrial function and has ISCU and COX10 are target genes of miR-210 related to been widely used as an indicator of cell health. Our data mitochondrial metabolism indicate that although cell proliferation moderately We first attempted to identify miR-210’s target genes slowed down, there was no apparent cell death caused associated with mitochondrial metabolism by in silico by increased miR-210 levels. prediction algorithms using Sanger miRBase (http:// Elevated oxidative stress has been observed in many www.microrna.sanger.ac.uk/) and TargetScan (http:// cancer cell types, which correlates with the aggressive- www.targetscan.org/). We identified several promising ness of tumors and poor prognosis. Reactive oxygen putative miR-210 target genes related to OXYPHOS species (ROS) can activate various survival pathways, (oxidative phosphorylation), including ISCU (iron- promote genetic instability and render cells resistant sulfur cluster scaffold homolog), COX10 (cytochrome to drug treatment (Trachootham et al., 2009). The c oxidase assembly protein), CS (citrate synthase), production of ROS can be triggered by hypoxia and NDUFA4 (NADH dehydrogenase (ubiquinone) 1a mitochondria initiated ROS can stabilize HIF-1a by subcomplex), SDHALP2 (succinate dehydrogenase partially inhibiting prolyl hydroxylase activity (Guzy complex, subunit A, flavoprotein 2) and et al., 2005). We, therefore, questioned whether miR-210 NOX4 (NADPH oxidase 4). We observed that the also contributes to the induction of ROS under hypoxic mRNA levels of two of these targets, ISCU and COX10, conditions. As illustrated in Figures 3c and d, treatment were consistently downregulated in HCT116, HCTÀ/À, with miR-210 mimic activated the generation of HT29 and RKO cells under hypoxia or DFO treatment, hydrogen peroxide (H2O2) and superoxide within 24 or which lead to robust induction of miR-210. This

Oncogene miR-210 regulates mitochondria and targets ISCU COX10 ZChenet al 4365 ISCU COX10 Negative Control 0.6 Negative Control p=0.0018 microRNA transfection 6 microRNA transfection p=0.0138 p=0.0190 p=0.0432

0.4 4 2^(-dCT) 2^(-dCT) Control miR-210 2 0.2 ISCU1/2 ISCU Relative expression COX10 Relative expression 0 0.0 β-actin miR-210 mimic miR-210 inhibitor miR-210 mimic miR-210 inhibitor Normoxia Hypoxia Normoxia Hypoxia Control miR-210 150 Negative control Negative control miR-210 p=0.0109 miR-210 COX10 100 p=0.0036 p=0.0216 p=0.2571 β 100 -actin

50 50 % Luciferase activity % Luciferase activity

0 0 PGL3C-ISCU PGL3C-ISCU Reverse PGL3C-COX10 PGL3C-COX10 Revers Figure 4 ISCU and COX10 are targets of miR-210. (a) HCT116 cells were treated with miR-210 mimic under normoxic condition or mir-210 hairpin inhibitor under hypoxic condition. miRNA mimic-negative control or hairpin inhibitor-negative control was applied as control. After 24 h, mRNA levels of ISCU and COX10 were analyzed by quantitative real-time PCR. (b) miR-210 mimic directly targets the 30-UTR of ISCU/COX10 mRNA. Luciferase reporter assay was performed in the HCT116 cells transfected with PGL3C luciferase reporter containing the 30-UTR regions of ISCU/COX10 mRNA. Data shown are the means and s.d.s of triplicate reactions. Corresponding P-values are shown. (c) Western blot analysis using cell extracts of HCT116 cells transfected with miR-210 for 48 h was performed to monitor ISCU and COX10 protein levels. suggests that ISCU and COX10 may be potential targets effect on the pGLS3 vector containing the reverse of miR-210 (Supplementary Figure S1). sequences of the ISCU and COX10 30-UTR regions To test if ISCU and COX10 are targets of miR-210, (Figure 4b). Downregulation of ISCU and COX10 by we first evaluated their mRNA levels in HCT116 cells miR-210 was further confirmed at the protein level by after transfection with miR-210 mimic under normoxic western blot analysis (Figure 4c). condition. mRNA levels of ISCU and COX10 under To confirm the specificity of miR-210’s regulation of normoxic culture conditions were significantly decreased ISCU and COX10, we tested the effect of altering miR- after transfection with miR-210 mimic compared with 210 levels on additional genes encoding subunits of the the miRNA mimic-negative control (Figure 4a). We electron transport chain under similar experimental then determined if treatment with miR-210 hairpin conditions as described in Figure 4a (Supplementary inhibitor under hypoxic conditions could abrogate the Figure S3). Genes we tested are NDUFA4 (complex I reduction of ISCU or COX10 mRNA level caused by subunit), SDHALP2 (complex II subunit) and COX15 hypoxia. As shown in Figure 4a, the reduction of (complex IV subunit). Among them, NDUFA4 has been mRNA level under hypoxic conditions was abrogated picked up by in silico prediction algorithms as potential by transfection with the miR-210 hairpin inhibitor target of miR-210. The results confirm that miR-210 compared with the miRNA inhibitor negative control, mimic or inhibitor treatment does not significantly alter which indicate that miR-210 is the major cause of mRNA level of these three additional genes we tested. downregulation of ISCU and COX10 genes under To determine that repression of ISCU and COX10 hypoxic condition. A similar trend was observed in contributes to miR-210-mediated effects on mitochon- three additional cancer cell lines tested (Supplementary dria function, we performed knockdown experiments Figure S2). using small interfering RNA targeting ISCU and We next used the luciferase reporter gene assay to COX10. As shown in Supplementary Figure S4 A and confirm that miR-210 directly targets mRNA of ISCU B, knockdown of ISCU or COX10 caused reduction of and COX10 by cloning the 30-UTR regions of ISCU and oxygen consumption activity by approximately 30 and COX10 mRNA separately into the luciferase reporter 48%, respectively. However, ROS level was not pGLS3 control vector. Co-transfection of significantly altered. As there are various sources of HCT116 cells with pGLS3 vector and miR-210 mimic cellular ROS in addition to mitochondrial OXPHOS, resulted in 40 and 70% reduction of luciferase activity of for example, NADPH oxidase, nitric oxide synthase, ISCU and COX10, respectively. miR-210 mimic had no lipoxygenases, cyclooxygenases, xanthine oxidase and

Oncogene miR-210 regulates mitochondria and targets ISCU COX10 Z Chen et al 4366 cytochrome P450 enzymes, the upregulation of ROS effect) and promoting tumor cell survival under hypoxic production caused by miR-210 is likely due to yet conditions (Papandreou et al., 2006; Hsu and Sabatini, unidentified targets of miR-210, which appear to be 2008), our study indicates that miR-210 has an many according to in silico prediction algorithms. important role in cancer cells’ adaptation to hypoxic We further tested whether overexpression of the microenvironment. This is supported by observations coding region of ISCU (without 30-UTR region, thus, that expression of miR-210 is elevated in pancreatic missing miR-210 target sequence) can reverse the cancer (Greither et al., 2010), diffuse large B-cell effects of miR-210 on mitochondrial respiration. Con- lymphoma (Lawrie et al., 2008) and breast cancer structs of the coding regions of ISCU1/2 were generated (Camps et al., 2008). using pcDNA3.1 vector, and stable HCT116 cell lines The discovery that hypoxia-regulated miRNA de- expressing those vectors were established. MiR-210 creases mitochondrial respiration activity and activates mimic treatment caused significantly less repression ROS generation provides a better understanding of at the protein levels of ISCU1/2 in stable HCT116 cancer cell metabolism and may serve as a biochemical cells expressing ISCU1/2. Furthermore, the downregu- basis for developing new chemotherapy strategies. lation of mitochondrial respiration activity caused by Future studies are warranted to see if miR-210 has a miR-210 mimic was partially reversed in the presence role in other hypoxia- and mitochondrial dysfunction- of constitutively expressed ISCU1/2 (Supplementary related diseases such as Parkinson’s disease, myocardial Figure S4 C and D). Unfortunately, we were unable to ischemia and stroke. establish a stable cell line overexpressing COX10 due to technical issues for validating causative role of COX10. ISCU is essential for the assembly of Fe-S clusters, Materials and methods which are major components in the mitochondrial electron transport chain/tricarboxylic acid cycle and Reagents define the electron transport pathways in redox enzymes DFO was purchased from Sigma-Aldrich (St Louis, MO, (Johnson et al., 2005). Loss of function of ISCU can USA). TRIzol reagent was from Invitrogen (Carlsbad, CA, cause decline of mitochondrial function and disrupt iron USA). 3-BrOP (cell-permeable ester of 3-bromopyruvate) and homeostasis (Tong and Rouault, 2006). COX10 encodes human colon cancer cell lines HCT116 and HCT116 p53À/À a protoheme: heme O farnesyl transferase that partici- (HCTÀ/À) were gifts from Dr Peng Huang. miRNA miRIDIAN pates in the of heme-a, which is an essential mimic, hairpin inhibitor and relative negative controls, component of COX (, complex ON-TARGETplus SMARTpool small interfering RNA for IV). The loss of function of COX10 can cause ISCU and COX10, DharmaFECT 1 and DharmaFECT Duo Transfection reagent were from Dharmacon (Chicago, IL, suppression of mitochondrial function at the level of USA). Lipofectamine 2000 was from Invitrogen (Carlsbad, both complex I and IV (Diaz et al., 2006). CA, USA). TaqMan miRNA assays for hsa-miR-210, hsa- The possibility that ISCU and COX10 are under miR-16, and RNU48 and TaqMan assays for direct regulation of miR-210 indicates that miR-210 ISCU and COX10 were from Applied Biosystems (Foster City, may have a role in the regulation of the mitochondria CA, USA). pGL3 luciferase reporter control vector and the electron transport chain and tricarboxylic acid cycle. It Dual Luciferase Reporter Assay System kit were from is interesting to note that DFO treatment increased Promega (Madison, WI, USA). ISCU and COX10 antibodies miR-210 expression and decreased the mRNA levels were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). of ISCU and COX10. Given the fact that DFO and 0 other hypoxia mimetics (2, 2 -dipyridyl and COCl2) Cell culture and treatment are all iron chelators, there might be an intrinsic Colon cancer cells (HCT116, HCTÀ/À, HT29 and RKO) were correlation between hypoxia and cellular iron home- cultured in McCoy’s 5A medium supplemented with 10% fetal ostasis, and miR-210 may have an important role in iron bovine serum. MCF-7 and BE3 cells were maintained in metabolism. Dulbecco’s modified Eagle’s medium F12 medium supplemen- In conclusion, our data demonstrate that miR-210, ted with 10% fetal bovine serum. The hypoxic culture condition (2% O2) was reached by culturing cells in a modular which is a highly upregulated miRNA induced by incubator chamber (Billups-Rothenberg, Del Mar, CA, USA) hypoxia, can negatively regulate mitochondrial respira- flushed with 2% O2,5%CO2 and 95% N2. The working tion activity and increase ROS generation. ISCU and concentration of DFO was 50 mM. COX10 genes, which are key factors of the mitochon- drial electron transport chain and the tricarboxylic acid Mitochondrial membrane potential and cellular ROS cycle, have been identified as potential targets of miR- To measure mitochondrial membrane potential, the fluores- 210. Our results confirm causative role of ISCU1/2 in cent dye rhodamine 123 (1 mM) was applied. To measure miR-210 modulated mitochondrial respiration; further cellular superoxide or H2O2 levels, 100 ng/ml dihydroethidium experimentation is required to validate similar role for or 3 mM DCF-DA (CM-H2DCFDA) were applied. Samples COX10. As loss of function of both ISCU and COX10 were analyzed by flow cytometry (Pelicano et al., 2003). seems to have no effect on ROS generation, the mechanism of miR-210’s effect on cellular ROS needs Oxygen consumption and lactate generation to be further investigated. Given that hypoxia/HIF-1a After trypsinization, 3 million HCT116 cells were sealed in the pathway is capable of triggering the metabolic switch respiration chamber containing a Clark-type oxygen electrode from respiration to the glycolytic pathway (the Warburg disc (Oxytherm; Hansatech Instrument, Cambridge, UK). The

Oncogene miR-210 regulates mitochondria and targets ISCU COX10 ZChenet al 4367 oxygen consumption rate was calculated and plotted as Cytotoxicity and proliferation assay described previously (Pelicano et al., 2003). The cellular lactate After transfection with miR-210 mimic for 24 h, HCT116 cells production was analyzed using an Accutrend lactate analyzer were trypsinized and re-plated into 96-well plates. After 24 h, (Roche, Mannheim, Germany) according to the manufac- different concentrations of 3-BrOP were added and incubated turer’s instructions and normalized by cell amount. for an additional 72 h. The cytotoxicity of 3-BrOP was analyzed as described previously (Xu et al., 2005). Cell proliferation assay was performed using MTT assay (Luthra Western blot analysis et al., 2008). Western blot analysis of miR-210 mimic-treated cells was performed as described previously (Pelicano et al., 2003). Small interfering RNA and miRNA mimic/inhibitor transfection The transfection of small interfering RNA of ISCU and COX10, miRNA mimic of miR-210, miR-210 hairpin inhibitor Quantitative real-time PCR analysis for mRNA and miRNA were performed following the manufacturer’s protocol (Dhar- Total RNA was isolated using TRIzol reagent, following the macon). The working concentration of RNA was 100 nM. manufacturer’s protocol (Invitrogen). For mRNA expression profile of ISCU and COX10, real-time PCR analysis was performed using TaqMan gene expression assays as described Establishment of stable cell lines expressing coding regions of previously (Luthra et al., 2008). GUSB (glucuronidase b) was ISCU1/2 and COX10 0 used as the normalization control. miR-210 levels were The coding regions (without 3 -UTR) of ISCU1, ISCU2 and determined by TaqMan MicroRNA assay using stem-loop COX10 were amplified separately from normal female real-time quantitative PCR analysis (Applied Biosystems). genomic cDNA by PCR using the following specific primers: 0 0 miR-16 or RNU48 were applied as normalization controls. As ISCU1 forward primer, 5 -cttaagcttatggttctcattgacatg-3 ; 0 0 the levels of miRNAs, including miR-16 and RNU48, reverse primer, 5 -cggaattctcatttcttctctgcctc-3 . ISCU2 forward 0 0 fluctuated significantly after hypoxia treatment, the miR-210 primer, 5 -cttaagcttatggcggcggctgggg-3 ; reverse primer, 0 0 0 expression levels after hypoxia were normalized to the 5 -cggaattctcatttcttctctgcctctcc-3 . COX10 forward primer, 5 - 0 0 normoxic samples. accaagcttatggccgcatctccgcacac-3 ; reverse primer, 5 -cggaattct cagctgggagggggcc-30. Constructs were generated using pcDNA3.1( þ ) vector, and stable HCT116 cell lines expressing Luciferase reporter assay ISCU1/2 and COX10 were established using Lipofectamine The 30-UTR regions of ISCU and COX10 containing the 2000, following the manufacturer’s protocol (Invitrogen). predicted binding sites of miR-210 were amplified separately from normal female genomic DNA by PCR using the 0 following specific primers: ISCU forward primer, 5 -tagtctag Conflict of interest agccctccctcggcgaag-30; reverse primer, 50-gcgtctagaactagagaac 0 0 accatcatc-3 . COX10 forward primer, 5 -tagtctagagagcactggga The authors declare no conflict of interest. cgcccac-30; reverse primer, 50-tcgtctagaaggccctggctcagtcac-30. The duplexes were cloned into the PGL3 control vector at the XbaI site in the correct and reverse (negative control) directions. Co-transfection and measure- Acknowledgements ment of luciferase activity were performed as described previously (Luthra et al., 2008). We thank Dawn Chalaire for critical editing of the article.

References

Bristow RG, Hill RP. (2008). Hypoxia and metabolism. Hypoxia, Greither T, Grochola L, Udelnow A, Lautenschlager C, Wurl P, DNA repair and genetic instability. Nat Rev Cancer 8: 180–192. Taubert H. (2010). Elevated expression of microRNAs 155, 203, 210 Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Brown JP and 222 in pancreatic tumours associates with poorer survival. Int J et al. (1998). Requirement for p53 and to sustain G2 arrest after Cancer 126: 73–80. DNA damage. 282: 1497–1501. Guzy RD, Hoyos B, Robin E, Chen H, Liu L, Mansfield KD et al. Camps C, Buffa FM, Colella S, Moore J, Sotiriou C, Sheldon H (2005). Mitochondrial complex III is required for hypoxia-induced et al. (2008). hsa-miR-210 Is induced by hypoxia and is an ROS production and cellular oxygen sensing. Cell Metab 1: independent prognostic factor in breast cancer. Clin Cancer Res 401–408. 14: 1340–1348. Hsu PP, Sabatini DM. (2008). Cancer cell metabolism: Warburg and Chen Z, Lu W, Garcia-Prieto C, Huang P. (2007). The Warburg effect beyond. Cell 134: 703–707. and its cancer therapeutic implications. J Bioenerg Biomembr 39: Johnson DC, Dean DR, Smith AD, Johnson MK. (2005). Structure, 267–274. function, and formation of biological iron-sulfur clusters. Annu Rev Crosby ME, Kulshreshtha R, Ivan M, Glazer PM. (2009). MicroRNA Biochem 74: 247–281. regulation of DNA repair gene expression in hypoxic stress. Cancer Kulshreshtha R, Ferracin M, Wojcik SE, Garzon R, Alder H, Agosto- Res 69: 1221–1229. Perez FJ et al. (2007). A microRNA signature of hypoxia. Mol Cell Diaz F, Fukui H, Garcia S, Moraes CT. (2006). Cytochrome c oxidase Biol 27: 1859–1867. is required for the assembly/stability of respiratory complex I in Lawrie CH, Gal S, Dunlop HM, Pushkaran B, Liggins AP, Pulford K mouse fibroblasts. Mol Cell Biol 26: 4872–4881. et al. (2008). Detection of elevated levels of tumour-associated Giannakakis A, Sandaltzopoulos R, Greshock J, Liang S, Huang J, microRNAs in serum of patients with diffuse large B-cell Hasegawa K et al. (2008). miR-210 links hypoxia with cell cycle lymphoma. Br J Haematol 141: 672–675. regulation and is deleted in human epithelial ovarian cancer. Cancer Lee YS, Dutta A. (2009). MicroRNAs in cancer. Annu Rev Pathol 4: Biol Ther 7: 255–264. 199–227.

Oncogene miR-210 regulates mitochondria and targets ISCU COX10 Z Chen et al 4368 Luthra R, Singh RR, Luthra MG, Li YX, Hannah C, Romans AM Tong WH, Rouault TA. (2006). Functions of mitochondrial ISCU and et al. (2008). MicroRNA-196a targets annexin A1: a microRNA- cytosolic ISCU in mammalian iron-sulfur cluster biogenesis and mediated mechanism of annexin A1 downregulation in . iron homeostasis. Cell Metab 3: 199–210. Oncogene 27: 6667–6678. Trachootham D, Alexandre J, Huang P. (2009). Targeting cancer cells Papandreou I, Cairns RA, Fontana L, Lim AL, Denko NC. (2006). by ROS-mediated mechanisms: a radical therapeutic approach? Nat HIF-1 mediates adaptation to hypoxia by actively downregulating Rev Drug Discov 8: 579–591. mitochondrial oxygen consumption. Cell Metab 3: 187–197. Xu RH, Pelicano H, Zhou Y, Carew JS, Feng L, Bhalla KN Pelicano H, Feng L, Zhou Y, Carew JS, Hileman EO, Plunkett W et al. et al. (2005). Inhibition of glycolysis in cancer cells: a novel (2003). Inhibition of mitochondrial respiration: a novel strategy to strategy to overcome drug resistance associated with enhance drug-induced apoptosis in human leukemia cells by a reactive mitochondrial respiratory defect and hypoxia. Cancer Res 65: oxygen species-mediated mechanism. JBiolChem278: 37832–37839. 613–621.

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

Oncogene