Mitochondrial pyrimidine nucleotide carrier (PNC1) regulates mitochondrial biogenesis and the invasive phenotype of cancer cells

Article metrics

Abstract

The insulin-like growth factor (IGF-I) signalling pathway is essential for metabolism, cell growth and survival. It induces expression of the mitochondrial pyrimidine nucleotide carrier 1 (PNC1) in transformed cells, but the consequences of this for cell phenotype are unknown. Here we show that PNC1 is necessary to maintain mitochondrial function by controlling mitochondrial DNA replication and the ratio of transcription of mitochondrial genes relative to nuclear genes. PNC1 suppression causes reduced oxidative phosphorylation and leakage of reactive oxygen species (ROS), which activates the AMPK-PGC1α signalling pathway and promotes mitochondrial biogenesis. Overexpression of PNC1 suppresses mitochondrial biogenesis. Suppression of PNC1 causes a profound ROS-dependent epithelial–mesenchymal transition (EMT), whereas overexpression of PNC1 suppresses both basal EMT and induction of EMT by TGF-β. Overall, our findings indicate that PNC1 is essential for mitochondria maintenance and suggest that its induction by IGF-I facilitates cell growth whereas protecting cells from an ROS-promoted differentiation programme that arises from mitochondrial dysfunction.

Introduction

The importance of mitochondria in cancer cell proliferation has been studied since Warburg (1956) proposed that tumour cells are less dependant on mitochondria for ATP production than normal cells. Increased glycolysis and rapid production of ATP could allow tumour cells to survive in low oxygen conditions found inside solid tumours in vivo (Pedersen, 1978; Gatenby and Gillies, 2004). It has long been suggested that this adaptation is due to an energy deficiency caused by non-functional mitochondria such as may occur with mitochondrial DNA (mtDNA) mutations (Penta et al., 2001). However, the Warburg hypothesis has been revisited notably because observed increases in glycolysis in cancer cells can occur along with normal mitochondrial function or respiration. There is now considerable evidence to indicate that enhanced glycolysis in cancer cells cooperates with oxidative metabolism, and that the anabolic functions of mitochondria in glutamine and fatty acid metabolism are essential for tumour cell proliferation (reviewed by Deberardinis et al., 2008; Jones and Thompson, 2009).

The insulin-like growth factor (IGF-I) signalling pathway through PI3-kinase, Akt and mTOR is important in promoting cell growth and proliferation. This signalling pathway can directly enhance both glycolysis and oxidative phosphorylation (OxPhos). It controls the expression and trafficking of glucose transporters (Quon et al., 1995; Hajduch et al., 1998; Huang et al., 2005), enhances glycolysis (Elstrom et al., 2004; Pankratz et al., 2009) and may also directly enhance OxPhos (Unterluggauer et al., 2008). IGF-I-mediated activation of the mTOR pathway is regulated by AMP-activated kinase (AMPK), a serine or threonine kinase that is considered to be a master regulator of cellular and systemic metabolism by acting to restore energy balance (Marshall, 2006). However, although AMPK has an important role in mitochondrial biogenesis, the links between the IGF-I signalling pathway and mitochondrial function and biogenesis are not clear.

The Warburg effect suggests that increased glycolysis in cancer cells harbouring mitochondria with mtDNA mutations confers a growth advantage in hypoxic conditions. However, damaged mitochondria may also directly participate in increasing metastatic potential by activating a reactive oxygen species (ROS)-mediated retrograde signalling pathway (Ferraro et al., 2006; Owusu-Ansah et al., 2008). Earlier studies have correlated elevated levels of ROS in cancer cells with carcinogenesis (Klaunig et al., 1998; Mori et al., 2004). Numerous studies have implicated ROS in DNA mutation and apoptosis, low levels of ROS produced in response to growth factor signalling may also function as a secondary messenger to activate signalling pathways (Finkel, 2000).

The intense interest in targeting the IGF-I pathway for cancer therapy (reviewed in Pollak, 2008) highlight the need for better insights into the mechanistic links between IGF-I or insulin signalling, mitochondrial function, ROS signalling and metabolic regulation. For these reasons we investigated the role of IGF-I signalling in mitochondrial activity in tumour cells by analyzing the function of the mitochondrial pyrimidine nucleotide carrier 1 (PNC1), which we previously identified as an IGF-I and insulin-responsive protein in transformed cells (Floyd et al., 2007). Here we show that PNC1 is essential for maintenance of mtDNA and respiration, and that suppression of PNC1 leads to activation of an ROS-mediated signalling pathway that controls nuclear gene expression. The findings show that IGF-I signalling is integrated with mitochondrial function to determine the phenotype of cancer cells.

Results

PNC1 suppression is associated with increased mitochondrial ROS production

We have previously shown that PNC1 expression levels are inversely correlated with cellular ROS levels (Floyd et al., 2007). To investigate the mechanism of this, we generated HeLa cell lines in which PNC1 expression was stably suppressed using shRNA. Clones showing an approximately 40% (ShRNA PNC1 clone 1) and 60% (ShRNA PNC1 clone 2) reduction, respectively in PNC1 transcription were isolated (Figure 1a). Increased basal expression of cellular ROS was observed in these clones compared with controls (Figure 1b). The subcellular origin of the observed ROS was investigated using immunofluorescence with the ROS-specific H2DCF-DA dye. As shown in Figure 1c, when compared with controls, cells with suppressed PNC1 expression show increased accumulation of the H2DCF-DA dye in regions in which it is colocalized with the mitochondrial marker MitoTracker. We then used inhibitors that target different complexes of the mitochondrial respiratory chain to determine which complex may contribute to this ROS production. Rotenone (blocks transfer of electron from complex I to complex II) induced ROS production in both control and shRNA clones whereas antimycin A (complex III inhibitor) and CCCP (Electron Transport Chain, uncoupler; unpublished data) did not increase ROS production in cells with suppressed PNC1 (Figure 1d). The observation that antimycin A does not affect ROS production in PNC1-deficient cells suggests that complex III in these cells is already leaking ROS. Altogether, these data indicate that the increased cellular ROS observed in cells with suppressed PNC1 is derived from mitochondria and may be derived from a leak in the respiratory chain at complex III.

Figure 1
figure1

PNC1 suppression increases mitochondrial ROS production. (a) Pnc1 mRNA expression in the different HeLa stable clones was estimated by real-time PCR using pnc1- and gapdh-specific primers. Levels of pnc1 mRNA relative to gapdh are represented. (b) The cellular ROS level was measured in HeLa stable clones by flow cytometry using the H2DCFDA fluorescent. The green open histogram represents the level of cellular ROS in cells expressing control shRNA1. The purple closed histograms represent ROS levels for each clone as indicated. Results shown are representative of three independent experiments that gave similar results. (c) Visualization of cellular ROS in HeLa clones by fluorescence microscopy in digitonin-permeabilized cells using the H2DCFDA (green) or MitoTracker (red)-specific probes. (d) Cellular ROS level in HeLa clones treated with complete media (CM), 10 μM rotenone (Rot) or 10 μM antimycin A (AntA). The cellular ROS level was measured by FACS using the H2DCFDA probe. A Student's t-test comparing each condition to complete media for each clone was calculated. *P<0.025.

PNC1 suppression inhibits mitochondrial ATP production and activates glycolysis

We next tested the effects of PNC1 suppression on respiratory chain activity. To do this, we measured the levels of intracellular ATP in cells cultured in glucose-containing media (which produce ATP by both respiration and glycolysis) and in galactose-containing media (which inhibit glycolytic production of ATP resulting in reliance on respiration (OxPhos) (Marroquin et al., 2007)). No significant difference was observed between PNC1-deficient cells and control cells cultured in glucose (Figure 2a). No difference was also observed in control cells after 2 h culture in galactose. In contrast, PNC1-deficient clones showed a reduction of 15–20% in cellular ATP level in the presence of galactose. This reduction in mitochondrial ATP production suggests that cells with PNC1 suppressed are deficient in OxPhos and cannot compensate for inhibition of glycolysis by galactose as efficiently as controls.

Figure 2
figure2

Suppression of PNC1 inhibits mitochondrial ATP production and oxygen consumption, and increases glycolysis. (a) Total cellular ATP production was measured in HeLa clones stably expressing shRNA control or shRNA PNC1. Cells were grown in either glucose or galactose-containing media as indicated. (b) Oxygen consumption was measured using the phosphorescent oxygen-sensing probe MitoXpress in HeLa cells expressing shRNA control or a PNC1 shRNA as indicated. The mitochondrial inhibitor Antimycin A was used as a negative control (AA). Data are presented as oxygen consumption over time. (c) Glycolysis was measured with a phosphorescent pH-sensitive probe and presented as extracellular cellular acidification rate. FCCP (0.5 μM) was applied to the control cells to increase glycolysis and mimic a deficiency in OxPhos. In (a), (b) and (c), the data are presented as the mean and standard deviation of values obtained for each time point in three independent experiments. Statistical significance (Student's t-test *P<0.01, #P<0.025).

To further assess the effect of PNC1 suppression on respiration, we investigated the effect of PNC1 suppression on oxygen consumption using the phosphorescent probe MitoXpress (Luxcel Biosciences, Cork, Ireland), whose fluorescence emission is quenched by oxygen. When cell cultures are isolated from air, the cells consume dissolved oxygen and the fluorescence emission of the probe increases (Will et al., 2006). As can be seen in Figure 2b, HeLa cells with suppressed PNC1 grown in glucose show decreased oxygen consumption compared with controls. This effect on oxygen consumption is consistent with the decrease in mitochondrial ATP production in PNC1-deficient cells grown in galactose (Figure 2a).

Cells with suppressed PNC1 could compensate for the loss of mitochondrial ATP by increasing activation of the glycolytic pathway for ATP production. We therefore investigated glycolytic activity in PNC1-deficient cells by measuring the rate of extracellular acidification, which is linked to lactic acid production during glycolysis. We found that cells with suppressed PNC1 have increased extracellular acidification (increased emission of the pH-sensitive probe). This indicates that glycolysis is increased in these cells compared with controls (Figure 2c).

Altogether our study data indicate that in conditions of normal oxygen supply suppression of PNC1 causes a defect in the respiratory chain and increases the rate of glycolysis to compensate for overall cellular ATP production.

PNC1 expression regulates mitochondrial mass and mtDNA replication

We next turned our attention to the mechanisms by which PNC1 regulates mitochondrial function. Having observed decreased OxPhos in cells with suppressed PNC1, we first investigated whether mitochondrial membrane potential (MMP) was altered using the tetramethylrhodamine (TMRE) probe, which accumulates inside the mitochondria in a potential-dependent manner. Surprisingly, neither PNC1 suppression nor PNC1 overexpression had a detectable affect on MMP (Figure 3a). This suggests that although the mitochondria of these cells contain a defective respiratory chain, they maintain a normal MMP.

Figure 3
figure3

PNC1 suppression does not affect mitochondrial membrane potential but controls mitochondrial mass and mtDNA transcription per mitochondrion. (a) The mitochondrial membrane potential of MCF-7 cells with PNC1 overexpressed or suppressed was measured by FACS using the TMRE probe. Expression levels of pnc1 in cells transfected with siRNA measured by RT–PCR are indicated in left panel below (abbreviations: (c) siRNA control, 1 PNC1 siRNA1, 2 PNC1 siRNA2)). PNC1 protein expression and Ha-PNC1 overexpression were determined by western blot with anti-Ha antibody and are shown in right panel below (abbreviations: N, pcDNA3-Ha; 1, PNC1 Ha clone 1; 2, PNC1 Ha clone 2). (b) RT–PCR showing levels of three mitochondria-encoded genes (Cox1, mtND4 and CytoB) and one mitochondria-encoded ribosomal RNA (rRNA 16S) in MCF-7 or HeLa cells transfected for 72 h with an siRNA control or two siRNA specific for PNC1. As control, the level of transcription of PNC1 and two housekeeping genes (gapdh and actin) are also indicated. (c) The mitochondrial mass of MCF-7 cells was measured by FACS using the MitoTracker Green probe. The data for cells with PNC1 suppressed are shown in the left panel and are presented as the percentage of fluorescence in controls (cells transfected with control siRNA), The right panel shows the mitochondrial mass of MCF-7 cells overexpressing PNC1 with data presented as percentage of fluorescence in controls (cells expressing empty vector). For both panels the data represent the mean and standard deviation of data from three independent experiments (Student's t-test #P<0.01). (d) Analysis of mitochondrial transcription per mitochondria. The levels of transcription of the mitochondrial genes (Cyclooxygenase-1 (COX1) and cytochrome B (CytoB) relative to gapdh were measured by real-time PCR. In the same cultures mitochondrial mass was measured using MitoTracker Green staining with FACS. To determine the rate of transcription per mitochondria the data for gene expression were expressed as a fraction of the data for mitochondrial mass. For each condition the data are presented as a percentage of control (either control siRNA or empty vector), and as the mean and standard deviation from three separate experiments (Student's t-test *P<0.05, #P<0.01). (e) Mitochondrial DNA content per cell population was determined by real-time PCR using primers specific for mtDNA or genomic DNA. The data are presented as the ratio of mtDNA to genomic DNA. For each condition the data are presented as a percentage of control (either control siRNA or empty vector), and as the mean and standard deviation from three separate experiments (Student's t-test #P<0.01).

PNC1 is a UTP carrier, so it is likely that reduced availability of UTP in PNC1-deficient cells (Floyd et al., 2007) could impair mtDNA transcription. This could result in reduced expression of mitochondrial-encoded components of the respiratory chain leading to decreased ATP output and increased ROS production. To test this hypothesis, we analyzed the transcription levels of three genes (cyclooxygenase-1, NADH dehydrogenase subunit IV (mtND4) and cytochrome B) and one ribosomal RNA (rRNA 16s) encoded by mtDNA, and compared them with two cellular housekeeping genes (gapdh and actin). Surprisingly, despite the effect of PNC1 on mitochondrial ATP production, PNC1 suppression had no discernable effect on the overall transcription of mtDNA-encoded genes in MCF-7 and HeLa cells (Figure 3b).

To test whether the observed lack of difference in the level of mtDNA transcription between control cells and cells with suppressed PNC1 was related to different numbers of mitochondria in these cells, we assessed mitochondrial mass. To do this we used the MitoTracker Green (MTG) probe that accumulates in the mitochondria in an MMP-independent manner. As shown in Figure 3c, cells with suppressed PNC1 have increased mitochondrial mass whereas cells with overexpressed PNC1 have reduced mitochondrial mass. The increased mitochondrial mass observed in PNC1-deficient cells would mask a reduction in mtDNA transcription in whole cell lysates derived from cells deficient in PNC1, and could account for the lack of observable differences (Figure 3b). To gain insight into the relative levels of mitochondrial transcription per mitochondrion, we analyzed mtDNA transcription (Cox1 and cytochrome B) and mitochondrial mass in the same pool of cells, and then normalized the amount of mtDNA transcription to mitochondria mass. As can be seen in Figure 3d, relative mtDNA transcription per mitochondrion was decreased in cells with suppressed PNC1, whereas it was increased in cells with overexpressed PNC1.

UTP is a cofactor for the DNA helicase twinkle, which is necessary for mtDNA replication (Korhonen et al., 2003). Thus, PNC1 expression would also be expected to alter mtDNA replication. We assessed the levels of mtDNA by real-time PCR using primers specific for mtDNA (spanning Cox1 and 2 genes) normalized to the level of nuclear genomic DNA (Figure 3e). Cells with suppressed PNC1 had reduced mtDNA, whereas cells with overexpressed PNC1 had increased mtDNA compared with controls.

Overall, the data indicate that PNC1 is essential for the transcription and replication of mtDNA, and suggest that control of mitochondrial mass may be a compensatory mechanism to maintain respiratory chain efficiency.

PNC1 controls mitochondrial biogenesis by ROS-dependent activation of the AMPK signalling pathway

Increased mitochondrial mass or mitochondrial biogenesis requires induction of nuclear genes encoding mitochondrial proteins. To investigate whether mitochondrial biogenesis was affected in cells with suppressed or overexpressed PNC1, we compared the expression of mitochondria-encoded and nuclear-encoded mRNA in total RNA extracts. Cells with suppressed PNC1 showed increased transcription of the nuclear-encoded mitochondrial genes ADP/ATP translocator and Aralar (Figure 4a), whereas cells overexpressing PNC1 showed decreased transcription of these genes (Figure 4b). This suggests that PNC1 controls mitochondrial mass by regulating transcription of nuclear genes that encode mitochondrial proteins.

Figure 4
figure4

PNC1 expression controls the transcription of nuclear-encoded mitochondrial genes (a) and (b) total mitochondrial transcription compared with nuclear transcription. The mRNA encoded by mtDNA (COX1 and CytoB) or nuclear DNA (ATP/ADP translocator and Aralar) were measured by real-time PCR in cells with (a) suppressed or (b) overexpressed PNC1 relative to gapdh. Data are presented as a percentage of control (either control siRNA or empty vector), and as the mean and standard deviation from three separate experiments (Student's t-test *P<0.025, #P<0.01). (c, d) Activity of AMPK in MCF-7 cells with suppressed PNC1 (c) or overexpressing PNC1 (d). Cells were serum-starved for 4 h and then stimulated for the indicated time with IGF-I. Lysates were prepared and analysed by western blot with anti-phospho-ACC, anti-phospho-AMPK, AMPK and actin antibodies.

To determine whether signalling pathways associated with mitochondrial biogenesis were activated in cells with suppressed or overexpressed PNC1, we first investigated AMPK, a known mitochondrial biogenesis promoter (Lage et al., 2008). MCF-7 cells with suppressed PNC1 showed increased basal AMPK phosphorylation on Thr172 and increased phosphorylation of its substrate Acetyl-CoA carboxylase (ACC; Figure 4c).

Interestingly, IGF-I-induced phosphorylation of both AMPK and ACC was not observed in cells overexpressing PNC1 (Figure 4d).

AMPK can be activated by cellular AMP and phosphorylation on Thr172 by LKB1. Interestingly, the basal increase in AMPK activation observed in cells with PNC1 suppressed was also observed in LKB1-deficient HeLa cells, suggesting that the pathway for activation of AMPK is LKB1 independent (Figure 5a). AMPK can also be activated through signalling pathways that require intracellular calcium and cellular ROS (Witczak et al., 2008). Since ROS is increased in cells with suppressed PNC1, this was a likely candidate. To test this we used N-acetyl cysteine (NAC) to scavenge ROS. As shown in Figure 5a and Supplementary Figure S1, NAC treatment totally blocked AMPK activation in HeLa and MCF-7 cells with PNC1 suppressed. This indicates that increased AMPK activation is mediated by elevated cellular ROS in PNC1-deficient cells.

Figure 5
figure5

PNC1 expression levels control activation of an ROS–AMPK–PGC1α signalling pathway. (a) AMPK activity was measured in HeLa cells with stably suppressed PNC1 and that were cultured in complete medium (CM) incubated for 30 min with or without 5 mM NAC by western blotting with anti-phospho-ACC, anti-phospho-AMPK, AMPK or anti-actin antibodies. (b) Pnc1, pgc1α and nrf2α mRNA expression was measured by real-time PCR in MCF-7 cells with (c) suppressed or (d) overexpressed PNC1. Data are presented as a percentage of control (either control siRNA or empty vector), and as the mean and standard deviation from three separate experiments (Student's t-test *P<0.05, #P<0.01).

We next measured the activity of a downstream target of AMPK, the transcription and activation of proliferator-activated receptor-γ coactivator-1 (PGC-1α) (Terada et al., 2002; Jager et al., 2007; Irrcher et al., 2008), which acts as a transcriptional co-factor for transcription factors including nuclear respiratory factors (NRF-1 and NRF-2) that control transcription of nuclear-encoded mitochondrial genes. In cells with PNC1 suppressed, both PGC1α and NRF-2a expression levels were increased (Figure 5b), whereas in cells with PNC1 overexpressed, PGC1α and NRF-2a expression levels were reduced (Figure 5c). The effects on PGC1α and NRF-2a correlate closely with PNC1 expression and mitochondrial mass. NAC and the AMPK inhibitor compound C were both able to abrogate the increase in mitochondrial mass and transcription of nuclear-encoded genes in PNC1-deficient cells (Supplementary Figure S2).

Taken together the data indicate that PNC1 regulates mitochondrial biogenesis through an ROS-dependent signalling pathway that controls AMPK activity and transcription of PGC1α and NRF-2a.

PNC1 regulates ROS-dependent epithelial–mesenchymal transition in MCF-7 and HeLa cells

MCF-7 and HeLa cells with PNC1 suppressed both showed altered morphology (Supplementary Figure S3) and size. We therefore investigated whether mitochondria dysfunction in these cells was associated with phenotypic changes in migration, adhesion or clonogenic growth. In wound-healing assays, cells with PNC1 suppressed showed higher motility than controls (Figure 6a). This was not due to increased proliferation as both controls and PNC1 shRNA clones showed similar proliferation rates (Figure 6b). Strikingly, cells with suppressed PNC1 had greatly decreased adhesion to fibronectin (Figure 6c) and a twofold increased capacity to form colonies in soft agarose (Figure 6d).

Figure 6
figure6

PNC1 suppression initiates an epithelial–mesenchymal transition in MCF-7 and HeLa cells. (a) HeLa cells were assayed for cell migration in a wound-healing assay for 18 h at which cells were stained with crystal violet and photographed at × 20 magnification. (b) Cell proliferation was assessed in cultures that were seeded at 104 cells per well in 24-well plate and complete media. At the indicated time cells were trypsinized and counted. The graph represents the average number of cells per well of three independent experiments. (c) HeLa cells with suppressed PNC1 were assessed for adhesion to fibronectin-coated plates for 30 min. The graph represents crystal violet staining of attached cells measured by spectrophotometry as a mean and standard deviation of three separate experiments. (d) Anchorage-independent growth of HeLa cell clones assayed in soft agar assay. Cells were allowed to grow for 3 weeks before staining with crystal violet and counting. The graph represents the mean and standard deviation of number of colonies per well from three separate experiments. (one way ANOVA #P<0.01,). (e) E-cadherin and β-catenin were detected by immunofluorescence with specific antibodies in formaldehyde-fixed cells and a Cy2-labelled secondary antibody (green). Nuclei were stained with Hoechst (blue). Images were obtained using a Nikon epifluorescence microscope and represent similar fields in more than three experiments.

Because the cellular changes observed in MCF-7 and HeLa cells with suppressed PNC1 resembled those of cells undergoing epithelial–mesenchymal transition (EMT), we investigated expression of markers of EMT by immunofluorescence. As shown in Figure 6e, MCF-7 cells with suppressed PNC1 exhibited a strong reduction in E-cadherin at the cell–cell contacts compared with controls (Figure 6e). By contrast, the expression of the mesenchymal marker β-catenin was increased (Figure 6e). Vimentin, fibronectin and N-cadherin were also increased (Supplementary Figure S4; unpublished data).

Mitochondrial ROS signalling has been strongly implicated in motility, EMT and metastatic process (Wu, 2006; Ishikawa et al., 2008). To determine whether PNC1 regulates EMT through mitochondrial ROS production, we used the ROS scavenger NAC. Exposure of cells to NAC completely blocked EMT in cells with PNC1 suppressed as indicated by the expression levels and location of E-cadherin (Figure 7a).

Figure 7
figure7

Morphological changes observed on PNC1 suppression are due to an ROS-dependent EMT programme in MCF-7 cells. (a) E-cadherin expression was assessed in MCF-7 cells with suppressed PNC1 that were incubated in complete media (control) or incubated overnight with 5 mM NAC. (b) The expression level of β-catenin was measured by FACS using an intracellular labelling in HeLa clones stably transfected with an shRNA control or an shRNA specific for PNC1 cultured for 8 h in complete media (CM) in presence or absence of 5 mM NAC or 10 μM compound C (CC). Statistical significance (one-way ANOVA #P<0.01). (c) E-cadherin expression was assessed in cells overexpressing PNC1 after overnight treatment with 20 ng/ml of TGF-β either in the presence or in the absence of 5 mM NAC. Images were obtained using a Nikon epifluorescence microscope and represent similar fields in more than three experiments.

The increase in β-catenin expression and cell motility in PNC1-deficient cells were abrogated by NAC (Figure 7b; Supplementary Figure S5). Interestingly, the AMPK inhibitor compound C had no effect on β-catenin expression suggesting that although ROS controls both EMT and mitochondrial biogenesis, EMT does not require AMPK.

To test whether overexpression of PNC1 would suppress EMT, we used TGF-β to stimulate the onset of EMT in MCF-7 cells. TGF-β can induce an ROS-dependent EMT in epithelial cells (Rhyu et al., 2005). As shown in Figure 7c, control MCF-7 cells lost E-cadherin expression when stimulated with TGF-β, whereas cells overexpressing PNC1 retained E-cadherin levels. This indicates that cells expressing PNC1 are protected from induction of EMT. Exposure of cells to NAC also reversed the effects of TGF-β in control cells and had no effect on E-cadherin expression in cells with PNC1 overexpressed.

Altogether, these data show that PNC1 controls cellular phenotype in transformed cells by regulating mitochondrial production of cellular ROS.

Discussion

Regulation of mitochondrial function by growth factors and their effects on growth, proliferation and the phenotype of cancer cells is an important but still poorly understood area of cancer research. Here we have shown that the IGF-I- and insulin-responsive mitochondrial UTP carrier PNC1 controls mtDNA replication and transcription, mitochondrial ROS production, bioenergetics/metabolism and the invasive potential of transformed cells. As summarized in Figure 8, we found that PNC1 is essential for maintenance of mtDNA synthesis and transcription, the ratio of mitochondria- to nuclear-encoded components of the electron transport chain, and mitochondrial function. Suppression of PNC1 leads to release of ROS, which results in increased mitochondrial biogenesis and an altered cellular differentiation programme (EMT). In contrast, overexpression of PNC1 suppresses ROS and EMT. Thus PNC1 maintains mitochondrial function and regulates an ROS-induced cellular differentiation programme.

Figure 8
figure8

Model illustrating function of PNC1 as an IGF-I-responsive protein in maintaining mitochondrial function. IGF-I induces expression of IGF-1. When PNC1 is expressed mitochondrial DNA synthesis and transcription produce components of the electron transport chain that together with nuclear-encoded components are essential for maintaining Oxphos and prevent production of ROS. When PNC1 is suppressed DNA synthesis and transcription are impaired; there is an imbalance in mitochondrial-encoded components of the electron transport chain compared with nuclear-encoded components, a defect in Oxphos occurs; and ROS are produced. This leads to activation of an AMPK-dependent signalling pathway that activates mitochondrial biogenesis and expression of genes that change cellular phenotype (epithelial to mesenchymal transition). IGF-I also activates AMPK indicated by dotted line.

Our study results strongly indicate that the defect in oxygen consumption and mitochondrial ATP production is caused by the effects of PNC1 on expression of both nuclear- and mitochondria-encoded components of the respiratory chain. This is most likely due to decreased mitochondrial UTP (Floyd et al., 2007) and the consequences of this for mtDNA replication and transcription. As components of the respiratory chain are encoded by both nuclear and mtDNA, an imbalance in expression could lead to defective respiratory chain function and ROS production. Our study results are consistent with previous studies showing that mitochondrial ROS activates a retrograde pathway leading to altered expression of nuclear-encoded genes (Butow and Avadhani, 2004; Owusu-Ansah et al., 2008).

We previously showed that PNC1 expression is strongly induced in response to IGF-I and insulin stimulation, which indicates direct regulation of mitochondrial function. Although IGF-I has long been implicated in cellular metabolism through its activity on mTOR, glycolysis and glucose transporters, there is also recent evidence for a direct effect of IGF-I on OxPhos in tumour cells (Unterluggauer et al., 2008) and in an experimental cirrhosis model in rat in which a low dose of IGF-I can increase the mitochondrial activity (Perez et al., 2008). Our study results suggest that by regulating the level of PNC1 expression, IGF-I may enhance mitochondrial function (OxPhos) when cells require this energy to support the transformed phenotype. This would also increase overall mitochondrial bioenergetics including fatty acid synthesis and glutamine metabolism, which has recently been shown to be essential for Myc oncogene-driven tumour cell growth (Jones et al., 2005; Deberardinis et al., 2008). Our observation that IGF-I directly activates AMPK also indicates a direct role for IGF-I in enhancing expression of nuclear-encoded genes necessary for mitochondrial biogenesis.

Our findings support an essential function for PNC1 in maintenance of mtDNA. A linear correlation was observed between mtDNA content and the level of PNC1 expression in transformed cells. When PNC1 was reduced by 80% (as routinely obtained with siRNA number 2) this resulted in a reduction in mtDNA content of approximately 80%, whereas cells with PNC1 suppressed by 40–60% of PNC1 (siRNA1 and shRNA) showed a 40–60% reduction in mtDNA content (Figure 3e; unpublished data). The necessity for PNC1 in mtDNA maintenance is apparently conserved throughout evolution because deletion of the yeast homologue Rim2p causes loss of mtDNA (Van Dyck et al., 1995). Interestingly, Rim2p overexpression can also rescue mtDNA deletion observed in yeast lacking the DNA helicase PIF1. The mtDNA maintenance in mammalian cells is controlled by several nuclear-encoded gene products. Among these are the DNA polymerase-γ (DNA polG), the DNA helicase twinkle and mtDNA-interacting proteins including mtSSB and TFAM (Falkenberg et al., 2007). Mutations in these proteins result in a loss of mtDNA and have been associated with various mitochondrial diseases (Copeland, 2008). The minimal replication system consists of the DNA polG and twinkle. Interestingly, although twinkle can use different 5′-triphosphates as cofactors for its helicase activity, UTP is proposed to be its most potent cofactor (Korhonen et al., 2003). Because PNC1 imports UTP into mitochondria, it is rational to propose that decreased mitochondrial UTP in PNC1-deficient cells (Floyd et al., 2007) would reduce the efficiency of twinkle activity and reduce mtDNA levels.

The function of PNC1 in mtDNA maintenance may also explain why suppression of PNC1 leads to impaired cell proliferation only in certain conditions. We have previously reported that PNC1 suppression affects cell proliferation and halts the cell cycle in G1 (Floyd et al., 2007). However, this reduction was only observed when PNC1 was suppressed by more than 70%. HeLa clones stably expressing shRNA targeting PNC1 exhibited a 40% and 60% reduction in protein did not show reduced proliferation (Figure 6b). Previous studies of mtDNA suppression after ethidium bromide treatment or suppression of critical components of the mtDNA maintenance machinery have shown impaired proliferation only when a critical level of mtDNA (less than 25% of initial content) is reached, which suggests that a minimal level of mtDNA is sufficient for cell proliferation (Jazayeri et al., 2003; Jeng et al., 2008). Interestingly, the effects of PNC1 on cell size and ROS production were not as tightly correlated with the degree of PNC1 suppression (Floyd et al., 2007 and this study). In addition, cells overexpressing PNC1 showed increased cell size and decreased ROS without altered proliferation (Floyd et al., 2007). Altogether, these observations suggest that PNC1 primarily regulates ROS production and cell size and that impaired proliferation occurs as a consequence of drastic loss of mtDNA in cells with greatly suppressed PNC1.

PNC1 expression is enhanced in transformed cell lines and tumours compared with non-transformed cells, which suggests a role in facilitating transformation or tumour growth. In this report we found that suppression of PNC1 in transformed cells leads to acquisition of a more aggressive phenotype because of an ROS-dependent EMT programme. Mitochondria, and specifically, mutations in mtDNA, have previously been implicated in cancer progression (reviewed in Brandon et al., 2006; Chatterjee et al., 2006). Transfer of mutated mtDNA from highly metastatic cell lines is sufficient to initiate a metastatic phenotype in non-metastatic cell lines by promoting mitochondrial ROS production (Ishikawa et al., 2008). Moreover, mtDNA depletion in MCF-7 cells induces an EMT programme (Naito et al., 2008). Our results show that PNC1, by regulating the mitochondrial ROS production, has the potential to regulate initiation of an ROS-dependent EMT programme in tumour cells. Furthermore, PNC1 overexpression in MCF-7 cells was sufficient to prevent TGF-β-induced EMT. Thus, increased PNC1 expression could enable a tumour to optimally produce ATP and grow whilst it would also protect cells from metabolic stress and prevent progression toward an invasive aggressive phenotype. Our findings also suggest that PNC1 may have a more general function in regulating cellular differentiation programmes that are controlled by mitochondria retrograde signalling.

In summary, we have shown that PNC1 controls mitochondrial integrity and regulates signalling pathways that control the growth and differentiation of cancer cells. Our findings also highlight one of the limitations associated with the concept of targeting mitochondria in cancer. Partial inhibition of mitochondrial function may exacerbate the aggressiveness of cancer due to ROS production and subsequent effects on EMT signalling.

Materials and methods

Cell lines and plasmids

MCF-7 and HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum, 10 mM L-glutamine and antibiotic (all from Biowhittaker, Verviers, Belgium). The pSuper-PNC1-ShRNA and pSuper-Scramble vectors were generated by insertion of the PNC1 shRNA (sequence: IndexTermgatccccggctgtatactttgcatgttattgatatcctaacatgcaaagtatacagcctttttc) or scramble (sequence: IndexTermgatccccctggaagtcttcattaaggtgttgatatcccaccttaatgaagacttccagtttttc) between the BglII and XhoI sites of pSUPER.neo (OligoEngine, Seattle, WA, USA).

PNC1 siRNAs and transfection procedure were previously described (Floyd et al., 2007).

RT–PCR and real-time PCR

Whole RNA was isolated using the Trizol method (Invitrogen, Carlsbad, CA, USA). And cDNA synthesis was carried out by reverse transcription with equal amounts of RNA (2 μg) using a cDNA synthesis kit (Invitrogen). Equal amounts of cDNA were amplified using HotStar Taq DNA polymerase (Qiagen, Hilden, Germany) for regular reverse transcriptase (RT)–PCR or QuantiTect SYBR Green PCR kit (Qiagen) for real-time PCR. The various levels of amplification were normalized to the levels of amplification of gapdh or actin.

Measurement of ATP levels, O2 consumption and glycolysis

For ATP analysis, we incubated cells for 6 h in 96-well plates (Sarstedt, Wexford, Ireland) at 2 × 105 cells per well in serum-free Dulbecco's modified Eagle's medium containing P/S, 1 mM pyruvate and either 10 mM of glucose or 10 mM of galactose. Cellular total ATP was quantified using CellTiterGlo Assay (Promega, Madison, WI, USA) and Victor2 plate reader (PerkinElmer Life Science, Waltham, MA, USA) on white 96-well plates (Greiner Bio One, Frickenhausen, Germany) following the manufacturer's instructions.

For O2 consumption assay, we used the phosphorescent oxygen-sensing probe MitoXpress (Luxcel Biosciences) as previously described (Zhdanov et al., 2008). The extracellular acidification assay was performed as previously described (Hynes et al., 2009).

Determination of mtDNA levels

Total DNA was isolated as described before (Miller et al., 1988). Briefly, cells were lysed in lysis buffer (100 mM Tris-HCL (pH 8.5), 5 mM EDTA, 0.2% SDS and 10 mg/ml proteinase K) at 55 °C for 2 h. Total DNA was extracted using ethanol precipitation. Owing to the absence of introns in the mitochondrial genome, mtDNA was assessed using primer pairs that are complementary to sequences spanning two genes: Cox1 for mtDNA F and Cox2 for mtDNA R (see Table 1 in Supplementary Figures). The levels of nuclear DNA were measured using two pairs of primers located on either one exon and one intron of pnc1 gene or two introns of the actin gene (pnc1 exon4 and intron4, and α-actin intron3 and intron4 for DNA F, DNA R, actDNA F and actDNA R, respectively).

Western blot analysis and antibodies

Cellular protein extracts and western blot were prepared as previously described (Floyd et al., 2007). Anti-phospho ACC, anti-phospho-p70SK1, anti-AMPK and anti-phospho-AMPK antibodies were all purchased from Cell Signaling Technology (Beverly, MA, USA). The anti-actin monoclonal antibody was purchased from Sigma-Aldrich (Dublin, Ireland).

Adhesion, migration, soft agar and proliferation assays

These assays were conducted as previously described (Ayllon and O'Connor, 2007) with the following modifications. For adhesion assays, 1 × 104 cells per well were added to previously coated wells (with fibronectin) in triplicate and allowed to attach for 30 min. The cells were then stained with crystal violet and the number of attached cells was estimated by spectrophotometer. For soft agar assays, 4 × 103 HeLa cells per well were resuspended in 0.33% low-melting point agarose in Dulbecco's modified Eagle's medium–10% fetal bovine serum and plated in triplicate onto 35-mm dishes containing a 2 ml base agarose layer (0.6%).

Immunofluorescence and flow cytometry

Immunofluorescence assays were conducted as previously described (Floyd et al., 2007). For ROS detection, the H2DCF-DA probe (Molecular Probes, Eugene, OR, USA) was added at 50 μM to the media in the absence of fetal calf serum and incubated for 15 min at 37 °C. After one wash in PHEM, the cells were fixed in PHEM 3.7% formaldehyde for 15 min and treated or not with 10 μM of digitonin (Calbiochem, San Diego, CA, USA) for 30 min in Hank's balanced salt solution (Gibco BRL, Paisley, Scotland, UK) at room temperature. The discs were then mounted on slides and photographed using a Nikon E600 and × 100 objective.

For fluorescence-activated cell sorting (FACS) analysis the following probes were used: 40 nM MitoTracker Green dye (Molecular Probes) for mitochondria mass, 10 μM H2DCF-DA (Molecular Probes) for ROS detection, 20 nM TMRE (Sigma) for MMP. For β-catenin detection, 150 000 cells were cultured for 8 h in six-well plates in normal media in presence or absence of 5 mM NAC or 10 μM compound C. Cells were then collected using phosphate-buffered saline 0.5 mM EDTA, washed in phosphate-buffered saline 0.5%, bovine serum albumin 0.1% saponin, incubated for 20 min with β-catenin antibody (BD Transduction Laboratories, Erembodegem, Belgium) in the presence of saponin. Cells were then washed and incubated for 20 min with anti-mouse Cy3-conjugated antibody (Jackson ImmunoResearch Laboratories, Newmarket, UK). Cells were then washed and analysed by FACS.

References

  1. Ayllon V, O'Connor R . (2007). PBK/TOPK promotes tumour cell proliferation through p38 MAPK activity and regulation of the DNA damage response. Oncogene 26: 3451–3461.

  2. Brandon M, Baldi P, Wallace DC . (2006). Mitochondrial mutations in cancer. Oncogene 25: 4647–4662.

  3. Butow RA, Avadhani NG . (2004). Mitochondrial signaling: the retrograde response. Mol Cell 14: 1–15.

  4. Chatterjee A, Mambo E, Sidransky D . (2006). Mitochondrial DNA mutations in human cancer. Oncogene 25: 4663–4674.

  5. Copeland WC . (2008). Inherited mitochondrial diseases of DNA replication. Annu Rev Med 59: 131–146.

  6. Deberardinis RJ, Sayed N, Ditsworth D, Thompson CB . (2008). Brick by brick: metabolism and tumor cell growth. Curr Opin Genet Dev 18: 54–61.

  7. Elstrom RL, Bauer DE, Buzzai M, Karnauskas R, Harris MH, Plas DR et al. (2004). Akt stimulates aerobic glycolysis in cancer cells. Cancer Res 64: 3892–3899.

  8. Falkenberg M, Larsson NG, Gustafsson CM . (2007). DNA replication and transcription in mammalian mitochondria. Annu Rev Biochem 76: 679–699.

  9. Ferraro D, Corso S, Fasano E, Panieri E, Santangelo R, Borrello S et al. (2006). Pro-metastatic signaling by c-Met through RAC-1 and reactive oxygen species (ROS). Oncogene 25: 3689–3698.

  10. Finkel T . (2000). Redox-dependent signal transduction. FEBS Lett 476: 52–54.

  11. Floyd S, Favre C, Lasorsa FM, Leahy M, Trigiante G, Stroebel P et al. (2007). The insulin-like growth factor-I mTOR signaling pathway induces the mitochondrial pyrimidine nucleotide carrier to promote cell growth. Mol Biol Cell 18: 3545–3555.

  12. Gatenby RA, Gillies RJ . (2004). Why do cancers have high aerobic glycolysis? Nat Rev Cancer 4: 891–899.

  13. Hajduch E, Alessi DR, Hemmings BA, Hundal HS . (1998). Constitutive activation of protein kinase B alpha by membrane targeting promotes glucose and system A amino acid transport, protein synthesis, and inactivation of glycogen synthase kinase 3 in L6 muscle cells. Diabetes 47: 1006–1013.

  14. Huang C, Thirone AC, Huang X, Klip A . (2005). Differential contribution of insulin receptor substrates 1 versus 2 to insulin signaling and glucose uptake in l6 myotubes. J Biol Chem 280: 19426–19435.

  15. Hynes J, O′Riordan TC, Zhdanov AV, Uray G, Will Y, Papkovsky DB . (2009). In vitro analysis of cell metabolism using a long-decay pH-sensitive lanthanide probe and extracellular acidification assay. Anal Biochem 390: 21–28.

  16. Irrcher I, Ljubicic V, Kirwan AF, Hood DA . (2008). AMP-activated protein kinase-regulated activation of the PGC-1alpha promoter in skeletal muscle cells. PLoS ONE 3: e3614.

  17. Ishikawa K, Takenaga K, Akimoto M, Koshikawa N, Yamaguchi A, Imanishi H et al. (2008). ROS-generating mitochondrial DNA mutations can regulate tumor cell metastasis. Science 320: 661–664.

  18. Jager S, Handschin C, St-Pierre J, Spiegelman BM . (2007). AMP-activated protein kinase (AMPK) action in skeletal muscle via direct phosphorylation of PGC-1alpha. Proc Natl Acad Sci USA 104: 12017–12022.

  19. Jazayeri M, Andreyev A, Will Y, Ward M, Anderson CM, Clevenger W . (2003). Inducible expression of a dominant negative DNA polymerase-gamma depletes mitochondrial DNA and produces a rho0 phenotype. J Biol Chem 278: 9823–9830.

  20. Jeng JY, Yeh TS, Lee JW, Lin SH, Fong TH, Hsieh RH . (2008). Maintenance of mitochondrial DNA copy number and expression are essential for preservation of mitochondrial function and cell growth. J Cell Biochem 103: 347–357.

  21. Jones RG, Plas DR, Kubek S, Buzzai M, Mu J, Xu Y et al. (2005). AMP-activated protein kinase induces a p53-dependent metabolic checkpoint. Mol Cell 18: 283–293.

  22. Jones RG, Thompson CB . (2009). Tumor suppressors and cell metabolism: a recipe for cancer growth. Genes Dev 23: 537–548.

  23. Klaunig JE, Xu Y, Isenberg JS, Bachowski S, Kolaja KL, Jiang J et al. (1998). The role of oxidative stress in chemical carcinogenesis. Environ Health Perspect 106 (Suppl 1): 289–295.

  24. Korhonen JA, Gaspari M, Falkenberg M . (2003). TWINKLE Has 5′-&gt;3′ DNA helicase activity and is specifically stimulated by mitochondrial single-stranded DNA-binding protein. J Biol Chem 278: 48627–48632.

  25. Lage R, Dieguez C, Vidal-Puig A, Lopez M . (2008). AMPK: a metabolic gauge regulating whole-body energy homeostasis. Trends Mol Med 14: 539–549.

  26. Marroquin LD, Hynes J, Dykens JA, Jamieson JD, Will Y . (2007). Circumventing the Crabtree effect: replacing media glucose with galactose increases susceptibility of HepG2 cells to mitochondrial toxicants. Toxicol Sci 97: 539–547.

  27. Marshall S . (2006). Role of insulin, adipocyte hormones, and nutrient-sensing pathways in regulating fuel metabolism and energy homeostasis: a nutritional perspective of diabetes, obesity, and cancer. Sci STKE 2006: re7.

  28. Miller SA, Dykes DD, Polesky HF . (1988). A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 16: 1215.

  29. Mori K, Shibanuma M, Nose K . (2004). Invasive potential induced under long-term oxidative stress in mammary epithelial cells. Cancer Res 64: 7464–7472.

  30. Naito A, Cook CC, Mizumachi T, Wang M, Xie CH, Evans TT et al. (2008). Progressive tumor features accompany epithelial–mesenchymal transition induced in mitochondrial DNA-depleted cells. Cancer Sci 99: 1584–1588.

  31. Owusu-Ansah E, Yavari A, Mandal S, Banerjee U . (2008). Distinct mitochondrial retrograde signals control the G1–S cell cycle checkpoint. Nat Genet 40: 356–361.

  32. Pankratz SL, Tan EY, Fine Y, Mercurio AM, Shaw LM . (2009). Insulin receptor substrate-2 regulates aerobic glycolysis in mouse mammary tumor cells via glucose transporter 1. J Biol Chem 284: 2031–2037.

  33. Pedersen PL . (1978). Tumor mitochondria and the bioenergetics of cancer cells. Prog Exp Tumor Res 22: 190–274.

  34. Penta JS, Johnson FM, Wachsman JT, Copeland WC . (2001). Mitochondrial DNA in human malignancy. Mutat Res 488: 119–133.

  35. Perez R, Garcia-Fernandez M, Diaz-Sanchez M, Puche JE, Delgado G, Conchillo M et al. (2008). Mitochondrial protection by low doses of insulin-like growth factor-I in experimental cirrhosis. World J Gastroenterol 14: 2731–2739.

  36. Pollak M . (2008). Insulin and insulin-like growth factor signalling in neoplasia. Nat Rev Cancer 8: 915–928.

  37. Quon MJ, Chen H, Ing BL, Liu ML, Zarnowski MJ, Yonezawa K et al. (1995). Roles of 1-phosphatidylinositol 3-kinase and ras in regulating translocation of GLUT4 in transfected rat adipose cells. Mol Cell Biol 15: 5403–5411.

  38. Rhyu DY, Yang Y, Ha H, Lee GT, Song JS, Uh ST et al. (2005). Role of reactive oxygen species in TGF-beta1-induced mitogen-activated protein kinase activation and epithelial–mesenchymal transition in renal tubular epithelial cells. J Am Soc Nephrol 16: 667–675.

  39. Terada S, Goto M, Kato M, Kawanaka K, Shimokawa T, Tabata I . (2002). Effects of low-intensity prolonged exercise on PGC-1 mRNA expression in rat epitrochlearis muscle. Biochem Biophys Res Commun 296: 350–354.

  40. Unterluggauer H, Hutter E, Viertler HP, Jansen-Durr P . (2008). Insulin-like growth factor-induced signals activate mitochondrial respiration. Biotechnol J 3: 813–816.

  41. Van Dyck E, Jank B, Ragnini A, Schweyen RJ, Duyckaerts C, Sluse F et al. (1995). Overexpression of a novel member of the mitochondrial carrier family rescues defects in both DNA and RNA metabolism in yeast mitochondria. Mol Gen Genet 246: 426–436.

  42. Warburg O . (1956). On the origin of cancer cells. Science 123: 309–314.

  43. Will Y, Hynes J, Ogurtsov VI, Papkovsky DB . (2006). Analysis of mitochondrial function using phosphorescent oxygen-sensitive probes. Nat Protoc 1: 2563–2572.

  44. Witczak CA, Sharoff CG, Goodyear LJ . (2008). AMP-activated protein kinase in skeletal muscle: from structure and localization to its role as a master regulator of cellular metabolism. Cell Mol Life Sci 65: 3737–3755.

  45. Wu WS . (2006). The signaling mechanism of ROS in tumor progression. Cancer Metastasis Rev 25: 695–705.

  46. Zhdanov AV, Ward MW, Prehn JH, Papkovsky DB . (2008). Dynamics of intracellular oxygen in PC12 cells upon stimulation of neurotransmission. J Biol Chem 283: 5650–5661.

Download references

Acknowledgements

We thank Kurt Tidmore for assistance with illustrations and to our colleagues in the Cell Biology Laboratory for helpful discussions. This work was funded by Science Foundation Ireland and the Health Research Board.

Author information

Correspondence to R O'Connor.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • PNC1
  • IGF-I
  • mitochondria
  • ROS
  • EMT in cancer
  • metabolism

Further reading