Biochemical Journal

Research article

Androgen stimulates glycolysis for de novo lipid synthesis by increasing the activities of hexokinase 2 and 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase 2 in prostate cancer cells

Jong-Seok Moon , Won-Ji Jin , Jin-Hye Kwak , Hyo-Jeong Kim , Mi-Jin Yun , Jae-Woo KIM , Sahng Wook Park , Kyung-Sup Kim

Abstract

Up-regulation of lipogenesis by androgen is one of the most characteristic metabolic features of LNCaP prostate cancer cells. The present study revealed that androgen increases glucose utilization for de novo lipogenesis in LNCaP cells through the activation of HK2 (hexokinase 2) and activation of the cardiac isoform of PFKFB2 (6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase). Activation of PKA (cAMP-dependent protein kinase) by androgen increased phosphorylation of CREB [CRE (cAMP-response element)-binding protein], which in turn bound to CRE on the promoter of the HK2 gene resulting in transcriptional activation of the HK2 gene. Up-regulation of PFKFB2 expression was mediated by the direct binding of ligand-activated androgen receptor to the PFKFB2 promoter. The activated PI3K (phosphoinositide 3-kinase)/Akt signalling pathway in LNCaP cells contributes to the phosphorylation of PFKFB2 at Ser466 and Ser483, resulting in the constitutive activation of PFK-2 (6-phosphofructo-2-kinase) activity. Glucose uptake and lipogenesis were severely blocked by knocking-down of PFKFB2 using siRNA (small interfering RNA) or by inhibition of PFK-2 activity with LY294002 treatment. Taken together, our results suggest that the induction of de novo lipid synthesis by androgen requires the transcriptional up-regulation of HK2 and PFKFB2, and phosphorylation of PFKFB2 generated by the PI3K/Akt signalling pathway to supply the source for lipogenesis from glucose in prostate cancer cells.

  • androgen
  • glycolysis
  • hexokinase 2
  • 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase 2 (PFKFB2)
  • prostate cancer

INTRODUCTION

Prostate cancer is the most common cancer in males, where androgen is a critical factor in regulating cell proliferation and growth. The most characteristic metabolic change by androgen in prostate cancer cells is the induction of de novo lipid synthesis through the up-regulation of lipogenic enzyme expression [15]. Since the metabolites from glycolysis are the main carbon sources for lipid synthesis, glycolysis should be controlled along with lipogenesis. Although the function of androgen in lipogenesis has been extensively investigated, little is known about the role of androgen in glycolysis for de novo lipid synthesis. In the present study, we carried out a set of experiments to elucidate the mechanism by which androgen controls glycolysis for support of vigorous lipogenesis in LNCaP prostate cancer cells.

A critical step for a high glycolytic state is the phosphorylation of glucose, which is catalysed by HK (hexokinase) [6,7]. Among the four mammalian HK types (HK 1–4), HK2 is frequently the predominant overexpressed form in tumours [810]. Furthermore, mounting evidence indicates that HK2 plays a pivotal role in highly malignant cancer cells by promoting cell growth, survival, enhancing biosynthesis and helping immortalization of the cells under hypoxic conditions. Another key step in controlling the glycolytic flux is the conversion of Fru-6-P (fructose 6-phosphate) into Fru-1,6-P2 (fructose 1,6-bisphosphate) by PFK-1 (phosphofructokinase). The activity of PFK-1 is allosterically controlled by Fru-2,6-P2 (fructose 2,6-bisphosphate), which is produced by the bifunctional PFKFB (6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase), which catalyses both the biosynthesis and degradation of Fru-2,6-P2 [11]. In mammals, the isoforms of PFKFB are encoded by four separate genes (PFKFB1, PFKFB2, PFKFB3 and PFKFB4), which are characterized by tissue expression pattern, the ratio of kinase to phosphatase activity, and the response to protein kinases [12]. In cancer cells, the level of Fru-2,6-P2 is generally elevated [11,13,14] due to overexpression and the activation of PFKFB3 and PFKFB4 [15,16]. PFKFB2, the cardiac isoform of PFKFB, is an essential enzyme in the regulation of glycolysis in the heart. Adrenaline, insulin and a work load increase the level of Fru-2,6-P2 by elevated PFK-2 (6-phosphofructo-2-kinase) activity of PFKFB2 [1719]. The C-terminal regulatory domain of PFKFB2, which is absent in the liver isoform PFKFB1, can be phosphorylated on both Ser466 and Ser483 by PDK-1 (3-phosphoinositide-dependent kinase-1), PKA (cAMP-dependent protein kinase), p70S6K (p70 ribosomal S6 kinase) and MAPK-1 (mitogen-activated protein kinase-1) [19,20]. These findings suggest that increased Fru-2,6-P2 production by elevated PFK-2 activity of PFKFB2 may play an important role in transformation of non-malignant cells or proliferation of cancer cells.

In the present study, we demonstrate that androgen up-regulates the expression of HK2 and PFKFB2, resulting in elevation of the utilization of glucose for de novo lipid synthesis in LNCaP prostate cancer cells. We show that androgen-induced HK2 expression is dependent on PKA signalling, and the induction of PFKFB2 expression is achieved by a direct binding of the androgen receptor to the PFKFB2 promoter. Constitutive activation of the PI3K (phosphoinositide 3-kinase)/Akt signalling pathway in LNCaP cells plays a critical role in the activation of PFK-2 activity of PFKFB2. In conclusion, these results suggest that androgen stimulates the utilization of glucose to undergo a metabolic conversion for both production of ATP and lipogenesis in androgen-dependent prostate cancer cells.

MATERIALS AND METHODS

Reagents

R1881 (methyltrienolone), a synthetic androgen, was purchased from Dupont-New England Nuclear and dissolved in DMSO. Forskolin, H-89 and LY294002 were purchased from Sigma–Aldrich and dissolved in DMSO.

Cell culture and routine procedures

Details of cell culture condition and routine procedures, including transfection, reporter assays, construction of plasmids, quantitative RT-PCR (real-time PCR), Western blotting and ChIP (chromatin immunoprecipitation) assays, are described in the Supplementary Materials and methods section (at http://www.BiochemJ.org/bj/433/bj4330225add.htm).

RNAi (RNA interference)

siRNA (small interfering RNA) oligonucleotides targeting PFKFB2 and control siRNA were purchased from Shanghai Genepharma. The siRNA oligonucleotide sequences were: PFKFB2 #1, 5′-AGAGCAAGAUAGUCUACUATT-3′; PFKFB2 #2, 5′-AGGAAAUAACAGACCUCAATT-3′; PFKFB2 #3, 5′-GUGGAAACAAUUAAACUUATT-3′; and control siRNA, 5′-UUCUCCGAACGUGUCACGUTT-3′. Transfections were performed twice over 2 days using Lipofectamine™ RNAiMAX (Invitrogen) according to the manufacturer's instructions. Details of RNAi are shown in the Supplementary Materials and methods section.

PFK-2 activity assay

PFK-2 activity was measured according to the method described by Manzano et al. [23]. Briefly, cleared cell lysates were obtained by homogenization in 500 μl of buffer A [50 mM Tris/HCl (pH 7.4), 0.1 M KCl, 0.5 mM EDTA, 5 mM MgCl2, 50 mM potassium fluoride, 1 mM DTT (dithiothreitol), 10 mg/ml leupeptin, 10 mg/ml aprotinin and 0.5 mM PMSF], and centrifuged at 13000 g at 4 °C for 10 min. Supernatants were mixed with 1 vol. of buffer B [100 mM Tris/HCl (pH 7.1), 200 mM KCl, 10 mM ATP, 4 mM MgCl2, 2 mM KH2PO4, 10 mM Fru-6-P, 35 mM glucose 6-phosphate and 2 mM DTT] and incubated at 30 °C for 30 min. Reactions (200 μl volume) were stopped by the addition of 1 vol. of 0.2 M NaOH followed by heating at 80 °C for 10 min, then neutralized with 40 μl of ice-cold 1 M acetic acid in 20 mM Hepes. After centrifugation (13000 g for 10 min at 4 °C), 350 μl aliquots of supernatant was mixed with 500 μl of reaction mixture [100 mM Tris/HCl (pH 8.0), 4 mM magnesium acetate, 2 mM Fru-6-P and 0.3 mM NADH], and then the mixtures were incubated with 100 μl of enzyme stock solution [1 unit/ml PPi-PFK (Fru-6-P 1-phosphotransferase), 500 units/ml triosephosphate isomerase, 170 units/ml α-glycerophosphate dehydrogenase and 45 units/ml aldolase] at 25 °C for 5 min. All enzymes were purchased from Sigma–Aldrich. The assay was started by the addition of 50 μl of 10 mM PPi solution in the mixtures. Absorbance at 340 nm was measured at room temperature (25 °C) every 1 min for a period of 10 min with an Ultrospec 1100 pro-spectrometer as described previously [22,23]. Assays were performed in triplicate.

Glucose uptake and lactate assay

LNCaP cells (6×105 cells/well) or PC3 cells (2×105 cells/well) were plated on six-well plates and treated as indicated. For the glucose-uptake assay, cells were incubated in glucose-free RPMI 1640 medium (Invitrogen) for 6 h, then supplied with approx. 0.3 MBq (10 μCi) [18F]FDG ([18F]fluoro-2-deoxyglucose) per well. After incubation at 37 °C for 20 min, cells were washed twice with PBS. The cells were harvested in PBS, and the radioactivity was measured using a Wallac 148 Wizard 3 γ-counter (PerkinElmer Life and Analytical Science). All experiments were performed in triplicate. The lactate levels in the culture medium were measured using a lactate assay kit according to the manufacturer's instructions (Biovision).

Lipid synthesis assay

Lipid synthesis from glucose was assayed by measuring the incorporation of 14C from D-[1-14C]glucose (PerkinElmer Life and Analytical Science) into lipid. Cells in six-well plates were incubated for 6 h in glucose-free RPMI 1640 medium and then were treated with 72 μM of D-[1-14C]glucose (4 μCi/ml) for 2 h. Cells were harvested and washed three times with PBS, and disrupted by adding 400 μl of 0.5% Triton X-100. The lipids were extracted with 500 μl of a hexane/isopropanol solution [3:2 (v:v)]. Aliquots of solvent layer were collected and dried under nitrogen gas. The dried samples were resuspended in 50 μl of chloroform, and the radioactivity of 14C was counted on a LS 6500 scintillation counter (Beckman Coulter).

Statistical analysis

SPSS for Windows version 17.0 was used for all statistical analyses.

RESULTS

Androgen activates glycolysis and lipid synthesis in LNCaP cells

To observe the changes in glycolysis by androgen in androgen-dependent LNCaP prostate cancer cells, glucose uptake and lactate production were measured after treatment with R1881, an androgen agonist. The uptake of [18F]FDG, which accumulates as 2-deoxyglucose 6-phosphate, over 20 min was markedly elevated up to 3-fold in cells treated with R1881 (Figure 1A), whereas lactate production was remained unchanged (Figure 1B). To determine whether the increase in glucose uptake serves as a substrate for the synthesis of lipids in LNCaP cells, we measured the amount of hexane/isopropanol-extractable lipids synthesized with D-[1-14C]glucose. R1881 drastically increased the incorporation of 14C from glucose into lipids up to 7-fold (Figure 1C), and this result was supported by marked induction of lipogenic enzymes, such as FASN (fatty acid synthase), ACLY (ATP citrate lyase) and ACACA (acetyl-CoA carboxylase α) (Figure 1D). These results suggest that the androgen-dependent activation of glucose uptake provides an important carbon source for androgen-dependent lipogenesis.

Figure 1 Androgen increases glucose uptake and lipid synthesis in LNCaP cells

Glucose uptake (A), lactate production (B) and lipid synthesis from D-[1-14C]glucose (C) were measured in LNCaP cells in the absence or the presence of 10 nM R1881 for 72 h. Results shown are the means±S.D. of triplicate samples from a representative experiment. (D) In each group, the amount of FASN, ACLY and ACACA were measured by Western blot analysis, using tubulin as an internal control. (E) Quantitative RT-PCR analysis was performed to measure the level of mRNAs for HK2 and PFKFB isoforms in LNCaP cells in the absence (open columns) or the presence (solid columns) of 10 nM R1881 for 72 h. Relative amounts of mRNAs were represented using GAPDH (glyceraldehyde-3-phosphate dehydrogenase) mRNA as an invariant control. *P< 0.05. EtOH, ethanol.

Androgen induces the expression of HK2 and PFPKB2 in LNCaP cells

On the basis of the assumption that the increase in glycolysis supports androgen-dependent lipogenesis, microarray analyses were carried out to reveal glycolytic genes whose expression is under the control of androgen. The expression of HK2 and PFKFB2 genes were increased 2.6- and 3.0-fold by R1881 treatment. Quantitative RT-PCR confirmed these significant increases in HK2 and PFKFB2, whereas other PFKFB isoforms remained unchanged (Figure 1E). These results indicate that HK2 and PFKFB2 could have an important role in the androgen-dependent increase in glycolysis in LNCaP cells.

PKA signalling activated by androgen stimulates the expression of HK2

CREB [CRE (cAMP-response element)-binding protein] plays a critical role in the strong transcriptional activation of the HK2 promoter, which contains a well-characterized CRE [2426]. CREB is the transcriptional activator associated with the cAMP/PKA signalling pathway, which is stimulated in LNCaP cells after androgen treatment [26a]. On the basis of these previous observations, we examined whether the induction of HK2 by androgen in LNCaP cells is achieved by transcriptional activation of the HK2 promoter (–2054/+21) via the cAMP/PKA signalling pathway. The activity of the HK2 promoter was increased 8-fold by androgen and this activation was significantly attenuated by treatment with H-89, an inhibitor of PKA (Figure 2A). Consistent with the HK2 promoter activity, the HK2 protein level was markedly increased by androgen. However, whereas H-89 only partially inhibited luciferase activity (Figure 2A), it fully blocked the induction of protein accumulation (Figure 2B). The difference between the HK2 protein mass and luciferase activity might be due to different stabilities of these proteins in cells treated with H-89. To determine whether androgen-mediated activation of the HK2 promoter is dependent on CRE, the CRE was mutated as shown in Figure 2(C). Mutation of CRE completely suppressed the responsiveness to androgen, indicating that CRE plays a critical role in the regulation of the HK2 promoter by androgen (Figure 2C). We have shown previously that androgen induces PKA activity by increasing the expression of PKA-Cβ2 (PKA catalytic subunit β transcript variant 2) in LNCaP cells [26a]. In the present study, we confirmed that PKA activity, the PKA-Cβ2 protein level and phosphorylation of CREB were significantly increased by R1881 treatment in LNCaP cells (Figures 2D and 2E). Exogenous overexpression of PKA-Cβ2 markedly activated the HK2 promoter in PC3 cells (Figure 2F). These results suggest that the induction of HK2 expression by androgen in LNCaP cells resulted from activation of PKA signalling.

Figure 2 Androgen activates the transcription of the HK2 gene via cAMP/PKA signalling in LNCaP cells

(A) The luciferase activity of the HK2 promoter was measured in LNCaP cells transfected with the pHK2-Luc construct, after incubation for 72 h in the absence or the presence of 10 nM R1881. At 24 h before harvest, when used, H-89 was added at a final concentration of 10 μM. Luciferase activities were normalized to the activities of SV40 (simian virus 40) promoter-driven Renilla luciferase. Values are the means±S.D. of triplicate samples from a representative experiment. *P< 0.05. (B) The levels of HK2 were analysed by Western blot analysis with the protein extracts of LNCaP cells treated as described in (A). Tubulin was used as an invariant control. (C) The conserved CRE in the HK2 promoter located at –504 from the ATG codon was mutated as shown. LNCaP cells transfected with promoter–Luc constructs harbouring wild-type or mutant CRE were incubated in the absence (open column) or the presence (solid column) of 10 nM R1881 for 72 h. Values are the means±S.D. of triplicate samples from a representative experiment. (D) PKA activities were measured as described in Supplementary Materials and methods section (at http://www.BiochemJ.org/bj/433/bj4330225add.htm) in LNCaP cells treated as described in (A). Data shown are relative to that of the group not treated with R1881. (E) In each group, the levels of PKA-Cβ2, phospho-CREB and total CREB in nuclear fractions were measured by Western blot analyses. (F) The activities of the HK2 promoter were assayed in PC3 cells exogenously overexpressing PKA-Cβ2. Cells were transfected with the pHK2-Luc construct in combination with the empty vector or pSG5-PKA-Cβ2. The relative luciferase activities were analysed 48 h after transfection. Values are the means±S.D. of triplicate samples from a representative experiment. EtOH, ethanol.

Androgen induces the recruitment of androgen receptors to the PFKFB2 promoter resulting in induction of PFKFB2

To determine whether PFKFB2 is a direct target gene of the androgen receptor, we compared the regulation of PFKFB2 with the KLK3 (kallikrein-related peptidase 3) gene, which is a known target of the androgen receptor in LNCaP cells. As shown in Figure 3(A), both KLK3 and PFKFB2 mRNA accumulated in a similar manner by treatment with R1881. PFKFB2 mRNA was significantly induced after treatment with R1881 by 1.8- and 2.5-fold at 6 and 12 h respectively. The amount of KLK3 mRNA was increased to a slightly higher level by 2.7- and 7-fold at 6 and 12 h respectively (Figure 3A). In accordance with the mRNA data, PFKFB2 protein was markedly increased 24 h after R1881 treatment and continued to increase until 72 h (Figure 3B). To elucidate whether the androgen receptor directly binds to the PFKFB2 promoter, a ChIP assay was carried out in LNCaP cells 6 h after R1881 treatment using an antibody against the androgen receptor. Five regions located at approx. 1 kb intervals on the PFKFB2 promoter were amplified by PCR using chromatin immunoprecipitates as templates (Figure 3C). Treatment with R1881 increased the recruitment of the androgen receptor only on the –1361 to –1077 region from the transcription start site (GenBank® accession number NM_006212.2), suggesting that a functional ARE (androgen-response element) is located at this region and is involved in the androgen-mediated regulation of PFKFB2 expression.

Figure 3 Androgen increases the expression of PFKFB2 in LNCaP cells by recruiting liganded androgen receptor to the PFKFB2 promoter

(A) The relative amounts of mRNA for PFKFB2 and KLK3 were determined by quantitative RT-PCR in a time course after treatment with 10 nM R1881. Each amount of mRNA was normalized using GAPDH (glyceraldehyde-3-phosphate dehydrogenase) mRNA as an invariant control. (B) The levels of protein for PFKFB2 were measured by Western blot analysis in the extracts of LNCaP cells treated with R1881 for the indicated times. Tubulin was used as an invariant control. (C) ChIP analyses were performed on LNCaP cells treated with 10 nM R1881 for 6 h using an androgen-receptor(AR)-specific antibody or non-specific IgG. The indicated chromosomal regions, denoted as nucleotide numbers from the transcription start of the PFKFB2 gene, were amplified by PCR and analysed in 1.2% agarose-gel electrophoresis. EtOH, ethanol; IP, immunoprecipitation.

Next, we examined the effects of androgen on PFKFB2 promoter activity, using a pPFKFB2–Luc construct which contains the PFKFB2 promoter fragment (–1949/+92) complexed with the luciferase gene. Androgen increased the luciferase activity by 3-fold (Figure 4A), which correlates with the androgen-induced increase in endogenous PFKFB2 gene expression. In order to find the key sequence element involved in androgen-mediated regulation, the putative ARE in the PFKFB2 promoter was searched using the MatInspector program of Genomatix software (http://www.genomatix.de). Only one putative ARE was found at –1204 and its mutation resulted in complete unresponsiveness to R1881 (Figure 4B), suggesting that the ARE at –1204 mediates the androgen-dependent activation of the PFKFB2 promoter. Taken together, these results indicate that the PFKFB2 gene is a direct binding target of the androgen receptor and is activated by androgen in the LNCaP cells.

Figure 4 Androgen activates the PFKFB2 promoter in LNCaP cells

(A) LNCaP cells were transfected with the pPFKFB2–Luc construct, then incubated for 72 h in the absence or the presence of 10 nM R1881. The relative luciferase activities were analysed as described in Figure 1(A). Data shown are the means±S.D. of triplicate samples from a representative experiment. (B) A putative consensus ARE is underlined in the PFKFB2 promoter, and the mutant construct was generated by the destruction of this consensus as indicated by lowercase characters. The responsiveness to R1881 was measured in LNCaP cells transfected with the wild-type pPFKFB2–Luc construct or the mutant construct as described in (A). Data shown are the means±S.D. of triplicate samples from a representative experiment. EtOH, ethanol.

PI3K/Akt signalling is involved in the phosphorylation of PFKFB2 in LNCaP cells

The unique C-terminal region of PFKFB2, which is not conserved in the other isoforms of PFKFB, contains the critical residues (Ser466 and Ser483) that control the enzyme activity through their phosphorylation status [17,19,20,27]. R1881 increased the phosphorylation at Ser466 and Ser483 of PFKFB2 as well as increasing the total enzyme level (Figure 5A). To determine the signalling pathway involved in the phosphorylation of these residues, LNCaP cells were treated with forskolin (a PKA activator), H-89 (a PKA inhibitor) or LY294002 (a PI3K inhibitor) for 40 min after treatment with R1881 for 72 h. The level of phosphorylation at Ser466 and Ser483 was not affected by treatment with forskolin or H-89 (Figure 5B), whereas LY294002 significantly decreased the phosphorylation of both Ser466 and Ser483 (Figure 5C). Consistent with this decreased level of phosphorylation, LY294002 markedly inhibited the PFK-2 activity (Figure 5D) and glucose uptake (Figure 5E) was increased by R1881. In LNCaP cells, it has been reported that the PI3K/Akt signalling pathway is constitutively activated by mutations in the PTEN (phosphatase and tensin homologue) gene and is likely to participate in activation of PFK-2 activity. Inhibition of glucose uptake by LY294002 resulted in the block of the flux of carbon from [14C]glucose into lipid synthesis (Figure 5F) without significant changes in amounts of lipogenic enzymes and the androgen receptor (Figure 5G). These results suggest that the PFK-2 activity of the bifunctional enzyme PFKFB2 should be activated by PI3K/Akt signalling for the high rate of glycolysis to support androgen-induced lipogenesis.

Figure 5 The PI3K signalling pathway is involved in the phosphorylation of Ser466 and Ser483 of PFKFB2

(A) The amount of total or phosphorylated forms of PFKFB2 (Ser466 and Ser483) were measured by Western blot analyses in LNCaP cells treated with 10 nM R1881 for the times indicated. (B and C) LNCaP cells grown in medium containing 10 nM R1881 for 72 h were treated with 10 μM forskolin or H-89 (B) or with 25 μM LY294002 for 40 min before harvest (C). The amounts of indicated proteins were determined by Western blot analyses using tubulin as an invariant control. (DG) LNCaP cells were grown in the absence or the presence of 10 nM R1881 for 72 h, then treated with 25 μM LY294002 for 40 min before harvest. PFK-2 activity (D), glucose uptake (E) and lipid synthesis from D-[1-14C]glucose (F) were assayed in each group. PFK-2 activity was denoted as the relative value to that of the group without any treatment. Data shown are the means±S.D. of triplicate samples from a representative experiment. The amount of FASN, ACLY, ACACA and androgen receptor (AR) were measured by Western blot analyses in each group, using tubulin as an invariant control (G). EtOH, ethanol.

PFKFB2 has a critical role in glucose uptake and glucose-dependent lipid synthesis

Inhibition of the PI3K/Akt signalling pathway drastically decreased PFK-2 activities and glucose uptake. Because the PI3K/Akt signalling pathway affects diverse metabolic pathways, the direct contribution of the PFK-2 activity of PFKFB2 to glucose uptake should be confirmed. Overexpression of PFKFB2 by transient transfection in PC3 cells (Figure 6C) increased the PFK-2 activity 3-fold (Figure 6A), which was accompanied by an increase in glucose uptake (Figure 6B). In contrast with the overexpression experiment, knocking-down of the PFKFB2 expression using siRNAs (Figures 6D and 6E) blocked the androgen-induced increase in glucose uptake (Figure 6F). Accordingly, glucose-dependent lipid synthesis was inhibited by knocking-down PFKFB2 (Figure 6G); however, the increased levels of lipogenic enzymes induced by androgen were not affected (Figure 6H). These results indicate that an increase in glucose uptake mediated by up-regulation of PFKFB2 is an important requirement for androgen-dependent induction of lipid synthesis in LNCaP cells.

Figure 6 PFKFB2 has a critical role in glucose uptake and lipid production in prostate cancer cells

PFK-2 activity (A) and glucose uptake (B) were measured in PC3 cells transfected with an empty vector or the pSG5-PFKFB2 expression vector. PFK-2 activity was denoted as the relative value to that of mock-transfected cells. Data shown are the means±S.D. of triplicate samples from a representative experiment. The expression of exogenous PFKFB2 was confirmed by Western blot analysis using an antibody against the HA-tag. Tubulin was used as an invariant control (C). (D and E) The levels of PFKFB2 mRNA (D) and protein (E) were measured in LNCaP cells transfected with the control or three independent PFKPB2 siRNAs (#1–#3) and then grown in the presence of R1881 for 72 h. Glucose uptake (F) and lipid synthesis from D-[1-14C]glucose (G) were determined in LNCaP cells which were transfected with a control or the PFKPB2 siRNA #3 and then grown in the absence or presence of R1881 for 72 h as indicated. Values shown are the means±S.D. of triplicate samples from a representative experiment. The amount of FASN, ACLY and ACACA were measured by Western blot analyses in each group using tubulin as an invariant control (H). EtOH, ethanol.

DISCUSSION

The androgen receptor has been implicated in proliferation and progression of prostate tumours [5,28,29]. Androgen-induced lipogenesis is an important metabolic change necessary for survival and proliferation of these tumours and the mechanisms responsible for transcriptional activation of lipogenic enzyme genes by the androgen receptor have been extensively studied [1,3]. Glucose utilization has also been reported to be controlled in an androgen-dependent manner in these tumours. In a clinical study, androgen-ablation therapy suppressed glucose utilization in prostate cancers [30] and most metastatic prostate cancers that express high levels of androgen receptors demonstrate high glucose uptake [31]. In general, glycolysis is known to be the main carbon source for vigorous lipogenesis, but the studies regarding how glycolysis is controlled by androgen in prostate cancer cells are strictly limited. In the present study, we focused on how androgen functions to induce glycolysis in LNCaP cells and how this induced glycolysis is linked to changes in lipogenesis required for cancer cell proliferation.

To address which glycolytic enzymes were transcriptionally regulated by androgen in prostate cancer cells, we performed a microarray analysis using total RNA isolated from LNCaP cells treated with or without R1881. The transcripts for HK2 and PFKFB2 among glycolytic enzymes were selectively elevated by androgen treatment. Cancer cells display a high rate of glucose uptake owing to elevated expression of glucose transporters and increased HK activity [32]. Among the four mammalian HK isoforms, it is proposed that the elevated HK activity in cancer cells is mainly due to high expression of HK2 [25]. Even if prostate cancer is known to have a relatively low glucose catabolic rate, the high rate of de novo lipid synthesis induced by androgen requires a carbon source from glycolysis. For this reason, the increase in glycolysis by androgen should be accompanied with a high rate of lipid synthesis. The present study revealed that HK2 and PFKFB2 are the major target genes for inducing glycolysis by androgen in LNCaP cells. HK2 is the principal isoform of HKs in skeletal muscle, heart and adipose tissue [33], and its activity is stimulated by exercise and β-adrenergic agonists, which involve the cAMP/PKA signalling pathway [34,35]. The 5′ flanking region of the HK2 promoter is rich in putative response elements, including six GC boxes, an E2F element, a CRE and a CCAAT box [25,26]. Our results showed that the CRE in this region plays a critical role in androgen-dependent activation of HK2 in LNCaP (Figures 2B and 2C). In a previous report, we have shown that R1881 enhanced the activity of PKA, which plays an important role in proliferation [26a,36], and this increase in PKA activity was correlated with the up-regulation of PKA-Cβ2 by androgen treatment in LNCaP cells [26a]. We further elucidated that the androgen-dependent induction of PKA-Cβ2 in the nucleus coincides with an increase in phosphorylation of CREB, which binds to CRE and activates transcription of the HK2 gene (Figure 2E). Taken together, it is strongly suggested that the HK2 gene is regulated in an androgen-dependent manner via up-regulation of the PKA signalling pathway in LNCaP cells.

In addition to an increase in HK activity, persistent consumption of glucose 6-phosphate into the glycolytic pathway is necessary to achieve a high rate of glucose uptake. The fate of glucose 6-phosphate to catabolic glycolysis is determined by the amount of Fru-2,6-P2, whose synthesis and degradation are catalysed by the bifunctional enzyme PFKFB. Mammals possess the distinct isoforms of PFKFB encoded by four different genes (PFKFB1, PFKFB2, PFKFB3 and PFKFB4) [12]. Rapidly proliferating cancer cells produce markedly elevated levels of Fru-2,6-P2, which is generally thought to be mediated by high expression of PFKFB3 and PFKFB4 isoforms. However, in the present study, we found that LNCaP cells uniquely express PFKFB2 which shows a high ratio of kinase/phosphatase activity, as PFKFB3 does [15]. Expression of PFKFB2 was induced by the direct recruitment of the ligand-activated androgen receptor to the PFKFB2 promoter in LNCaP cells. The PFKFB2 promoter was activated by R1881 in androgen-independent PC3 cells when the androgen receptor was exogenously overexpressed (results not shown). Androgen-dependent induction of PFKFB2 is assumed to induce glucose uptake and glucose-dependent lipid synthesis in LNCaP cells, because knocking-down PFKFB2 by siRNA severely impaired both processes (Figures 6F and 6G). These results demonstrate that the expression of PFKFB2 is directly regulated by the ligand-activated androgen receptor and plays a key role in androgen-induced glucose uptake and glucose-dependent lipid synthesis in LNCaP cells.

In addition to the changes in the amount of PFKFB2, the changes in kinase activity of PFKFB2 were addressed in the present study, because the concentration of Fru-2,6-P2 can be elevated only when its PFK-2 activity is activated [11]. The regulation of PFK-2 activity is achieved by phosphorylation or dephosphorylation of PFKFB2 at Ser466 and Ser483, mediated by several kinases, such as calcium/calmodulin-dependent protein kinase [27], PKA [37] or PI3K and Akt [19,20]. We have revealed in the present study that the PFK-2 activity of PFKFB2 was regulated by the PI3K/Akt signalling pathway, which is constitutively activated in LNCaP cells as a result of mutations of PTEN, an important regulator of the PI3K/Akt signalling pathway [38,39]. However, despite PKA activation by androgen, PKA did not affect the phosphorylation status of PFKFB2. PFKFB4, which is one of the isoforms expressed in LNCaP cells and known to be overexpressed in various malignancies [16,40], was not induced by androgen (results not shown). However, PFKFB4 might contribute little to glucose flux in prostate cancer cells, considering that the inhibition of phosphorylation of PFKFB2 by LY294002 inhibited androgen-induced PFK-2 activity to below the basal level (Figure 5E).

In conclusion, the present study demonstrates that the stimulation of HK2 and PFKFB2 by androgen induces high glucose uptake and the activation of glucose utilization, resulting in increased lipogenesis in prostate cancer.

AUTHOR CONTRIBUTION

Jong-Seok Moon, Won-Ji Jin, Jin-Hye Kwak, Hyo-Jeong Kim, Sahng Wook Park and Kyung-Sup Kim performed the experiments. Jong-Seok Moon, Won-Ji Jin, Jin-Hye Kwak, Hyo-Jeong Kim, Sahng Wook Park and Kyung-Sup Kim conceived and designed the experiments. Mi-Jin Yun and Jae-Woo Kim performed the experiments of glucose uptake using [18F]FDG. Jong-Seok Moon, Sahng Wook Park and Kyung-Sup Kim analysed the data and wrote the paper. All of the contributing authors critically reviewed the manuscript.

FUNDING

This work was supported by the National Research Foundation funded by the Korea government (MEST) [grant number 2010-0028364]; and by the 2007 Korean National Cancer Control Program, Ministry of Health and Welfare, Republic of Korea [grant number 0720400].

Abbreviations: ACACA, acetyl-CoA carboxylase α; ACLY, ATP citrate lyase; ARE, androgen-response element; ChIP, chromatin immunoprecipitation; CRE, cAMP-response element; CREB, CRE-binding protein; DTT, dithiothreitol; FASN, fatty acid synthase; [18F]FDG, [18F]fluoro-2-deoxyglucose; Fru-2,6-P2, fructose 2,6-bisphosphate; Fru-6-P, fructose 6-phosphate; HK, hexokinase; KLK3, kallikrein-related peptidase 3; PFK-1, phosphofructokinase; PFK-2, 6-phosphofructo-2-kinase; PFKFB, 6-phosphofructo-2-kinase/fructose 2,6-bisphosphatase; PI3K, phosphoinositide 3-kinase; PKA, cAMP-dependent protein kinase; PKA-Cβ2, PKA catalytic subunit β transcript variant 2; PTEN, phosphatase and tensin homologue; RNAi, RNA interference; siRNA, small interfering RNA; RT-PCR, real-time PCR

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 26a.
  28. 27.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
  33. 32.
  34. 33.
  35. 34.
  36. 35.
  37. 36.
  38. 37.
  39. 38.
  40. 39.
  41. 40.
View Abstract