|Home | About | Journals | Submit | Contact Us | Français|
The cAMP-responsive element binding protein CREB is frequently overexpressed and activated in tumors of distinct histology, leading to enhanced proliferation, migration, invasion and angiogenesis as well as reduced apoptosis. The de-regulated expression of CREB might be linked with transcriptional as well as post-transcriptional regulation mechanisms. We show here that altered CREB expression levels and function are associated with changes in the cellular metabolism. Using comparative proteome-based analysis an altered expression pattern of proteins involved in the cellular metabolism in particular in glycolysis was found upon CREB down-regulation in HER-2/neu-transfected cell lines. This was associated with diminished expression levels of the glucose transporter 1, reduced glucose uptake and reduced glycolytic activity in HER-2/neu-transfected cells with down-regulated CREB when compared to HER-2/neu+ cells. Furthermore, hypoxia-induced CREB activity resulted in changes of the metabolism in HER-2/neu transfected cells. Low pH values in the supernatant of HER-2/neu transformants were restored by CREB down-regulation, but further decreased by hypoxia. The altered intracellular pH values were associated with a distinct expression of lactate dehydrogenase, and its substrate lactate. Moreover, enhanced phosphorylation of CREB on residue Ser133 was accompanied by a down-regulation of pERK and an up-regulation of pAKT. CREB promotes the detoxification of ROS by catalase, therefore protecting the mitochondrial activity under oxidative stress. These data suggest that there might exists a link between CREB function and the altered metabolism in HER-2/neu-transformed cells. Thus, targeting these altered metabolic pathways might represent an attractive therapeutic approach at least for the treatment of patients with HER-2/neu overexpressing tumors.
The 43 kDa cyclic AMP (cAMP)-responsive element binding protein (CREB) is a member of the large family of basic leucine zipper (bZIP)-containing transcription factors (TF) and plays a key role in many physiologic as well as pathophysiologic processes including cell growth, differentiation, apoptosis, cell cycle progression and immune response amongst others . CREB exerts distinct activities in context and time-dependent manners [2–4], which depend on its phosphorylation status mediated by various kinases. Phosphorylated CREB interacts with the CREB-binding protein CBP/p300 via its kinase inducible domain (KID) and the KID-interacting domain (KIX) in CBP/p300 . The CREB/CBP complex induces a transcriptional machinery leading to activation of cAMP-responsive element (CRE) containing gene promoters. CREB mediates transcriptional responses to different stimuli including hormones, growth factors and neurotransmitters thereby acting as mediator between signaling pathways and their downstream activation target [3, 6].
However, based on its frequent overexpression and persistent activation CREB is also involved in the malignant transformation process . This is mediated by aberrant activation of cAMP signal transduction-dependent pathways such as G-coupled receptors, receptor tyrosine kinase (RTK) like human epidermal growth factor receptor 2 (HER-2/neu) and the cytokine/JAK/STAT signaling cascades, as well as signaling events downstream of CREB.
HER-2/neu is a member of the transmembrane epidermal growth factor receptor (EGF-R) family and is physiological expressed in many epithelial cells. However, it is overexpressed and/or amplified in human tumor of distinct origin including mammary carcinoma  and colorectal cancer . HER-2/neu transformation is associated with an increased cell proliferation, migration, enhanced cell cycle progression as well as a survival of patients . Recently, we could demonstrate in the murine model system used also in this study, that HER-2/neu can increase the phosphorylation rate of CREB and that HER-2/neu overexpressing cells had more detectable CREB protein than the parental NIH3T3 cells .
As recently reviewed CREB overexpression and its constitutive activation was found in many solid tumors, including non-small cell lung carcinoma, glioblastoma, melanoma, mammary carcinoma, renal cell carcinoma and mesothelioma as well as in different hematopoietic malignancies [4, 12–18], leading to enhanced cell proliferation, reduced apoptosis, increased angiogenesis and migration rates. Downregulation of CREB expression inhibits proliferation, migration and/or invasion, which was accompanied by suppression of matrix metallopeptidases and proteins of the epithelial mesenchymal transition . Furthermore, CREB overexpression can be correlated with clinico-pathological parameters, in particular with tumor grade, stage, metastasis formation, poor prognosis and reduced patients’ survival [14, 19–22]. In addition, CREB overexpression is often caused by an up-regulation of downstream target genes involved in neoplastic transformation as demonstrated by different high throughput analysis .
The expression level and activity of CREB is dynamically regulated by different mechanisms  and critical for maintaining a homeostatic cellular environment . CREB transcriptionally controls a large number of putative target genes, which are involved in numerous cellular processes, such as signal transduction, cell structure, differentiation, cell survival as well as proliferation . So far, amplifications and/or deletions in CREB have only been rarely detected suggesting that rather deregulation mechanisms of CREB appear to be responsible for its overexpression in tumors [26, 27]. Recently, a post-transcriptional control mechanism for CREB has been described, which seems to be mediated by microRNAs (miRs) known to for their binding to the 3' untranslated region (UTR) of CREB thereby contributing to tumorigenesis in different tumor models as demonstrated both in vitro and in vivo [28–31]. In addition, there is increasing evidence that different extra-cellular signals have an impact on the tumor microenvironment (TME), like hypoxia, pH variation and oxidative stress . Furthermore, post-translational modifications (PTM) of CREB, which can be quite diverse including phosphorylation, ubiquitination, methylation, glycosylation and SUMOylation, might have an impact on CREB function(s) [3, 17, 33]. So far, a link between CREB expression levels/function(s) and tumor metabolism has not been identified. Therefore, this study analyzed the effects of CREB on the metabolism using a murine model of HER-2/neu transformation with distinct CREB expression and activation levels, which has been previously well characterized and was able to induce tumors in immunocompetend DBA mice [11, 17, 34].
Since the level of CREB and HER-2/neu expression has been correlated with growth characteristics and altered signaling cascades , the protein expression pattern of HER-2/neu+ versus CREB-diminished HER-2/neu+ (shCREB) cells (Supplementary Figure 1A), with a knock down of up to 80% on the protein level (Supplementary Figure 1B, 1C) were determined by using two-dimensional gel electrophoresis (2-DE)-based proteome analysis and differentially expressed protein spots, defined by a 2-fold regulation, were identified by mass spectrometry. Overall 23 differentially expressed protein spots have been identified from three biological replicates (merged gels of all three experiments can be found in Supplementary Figure 2A), from which 13 proteins were down-regulated including four different forms of alpha-tubulin and 10 up-regulated upon CREB down-regulation. The differentially expressed proteins were mainly involved in metabolic processes (Table (Table1,1, Figure Figure1A,1A, Supplementary Figure 2B), in particular in glycolysis (Figure (Figure1B).1B). Based on their distinct expression pattern the following candidate CREB-regulated proteins were selected and their expression validated by qPCR and/or Western blot analyses: The panel of potential targets includes the phosphoglycerate kinase (PKG)1, prolyl endopeptidase, peroxiredoxin (PRX)4, enolase (ENO), triose phosphate isomerase (TPI), pyruvate kinase M (PKM) and citrate synthase. In line with the proteomic profiling data reduced transcription levels of PKM, citrate synthase and TPI were found in CREB down-regulated HER-2/neu+ cells (Table (Table2),2), while the mRNA expression level of PGK1 remained unchanged and that of the prolyl endopeptidase (PEP) induced. In addition, a CREB-mediated transcriptional control was detected for cofilin and α-crystalline (Table (Table2).2). The decreased mRNA levels were associated with decreased protein expression levels of ENO, PRX4, PGK1, PGAM1, PKM and TPI in HER-2/neu+ shCREB versus HER-2/neu+ cells (Figure (Figure1C),1C), which was further confirmed by a down-regulated PKM, TPI, and PGK1 enzyme activity (Table (Table3).3). Other differentially expressed proteins were enzymes important for detoxification mechanisms (catalase, PRX4, superoxide dismutase [Cu-Zn]) or linked to the protein degradation process (26S proteasome non-ATPase regulatory subunit 13, PEP, leukocyte elastase inhibitor A) (Table (Table11).
In silico analysis of gene promoters from differentially expressed proteins upon CREB down-regulation revealed that most of the identified proteins were controlled by half CRE sites (TGACG or CGTCA), whereas full CRE sites (TGACGTCA) were merely found in promoters of up-regulated proteins (Tables (Tables44 and and5).5). Since the promoter of the oncogene HER-2/neu contains no CRE elements, its expression was not affected by CREB down-regulation .
The comparative proteomics profiling of HER-2/neu+ cells and its shCREB derivatives suggests a link between the CREB expression levels/activation status and metabolic changes. In order to determine the functional relevance of this altered expression pattern, intracellular pH values were determined in the model system. Reduced intracellular pH values were detected in HER-2/neu+ cells when compared to NIH3T3 cells, while the intracellular pH values were restored to that of NIH3T3 cells by CREB down-regulation in HER-2/neu+ cells (Figure (Figure2A).2A). As expected treatment of the cells with the proton pump inhibitor Esomeprazole (ESOM), which served as control, strongly decreased the intracellular pH value in CREB down-regulated HER-2/neu+ cells (Figure (Figure2A2A).
Although a correlation between hypoxic conditions and metabolic reprogramming as well as the survival of tumor cells has been demonstrated [35–37], the role of CREB in this process has not yet been analyzed in detail. Therefore, the effect(s) of hypoxia on the energy metabolism of HER-2/neu+ and HER-2/neu+ shCREB as well as HER-2/neu- (NIH3T3) cells was compared by determining the intracellular and extracellular pH values, lactate and lactate dehydrogenase (LDH) concentrations. Under hypoxic conditions the intracellular pH was even further reduced in HER-2/neu+ cells, but also reduced in CREB down-regulated HER-2/neu+ cells (Figure (Figure2A).2A). The hypoxia-mediated decline of the intracellular pH values in HER-2/neu+ cells could be restored over time by culturing these cells under normoxia suggesting that this is a rather reversible process (Figure (Figure2B2B).
Altered extracellular pH conditions also affected both CREB activity and CREB-relevant signaling pathways as determined by Western blot analysis of HER-2/neu+ cells cultured at pH 6.3 and pH 7.3, respectively (Figure (Figure2C).2C). A shift from slightly acidic towards neutral pH conditions resulted in a decreased pCREBSer133 phosphorylation rate in HER-2/neu+ cells, which was accompanied by a down-regulation of pERK and an up-regulation of pAkt as well as of total CREB.
The modulation of the intracellular pH values was linked to changes of the LDH expression levels and of lactate concentrations as determined in the supernatants of normoxic and hypoxic HER-2/neu+ and CREB down-regulated HER-2/neu+ cells: The lactate concentrations were increased in HER-2/neu+ vs. HER-2/neu+ shCREB and NIH3T3 cells, while hypoxia enhanced lactate concentrations in all cell supernatants tested with the highest increase in supernatants of NIH3T3 cells (Figure (Figure2D).2D). Total LDH concentrations were comparable across all cells under normoxic conditions. Furthermore, culturing the cells under hypoxic conditions caused an up-regulation of total LDH activity in all cell lines to a similar extent (Figure (Figure2E),2E), while CREB knock down had no significant influence.
Since hypoxia (1% O2) can increase the phosphorylation of CREB at serine residue 133 , it was analyzed whether hypoxia-related TFs, like HIF-1α, can influence the CREB activity. Therefore, HER-2/neu+ cells were incubated with the HIF-1α inhibitor 2-methoxyestradiol (2-ME) for 24 h resulting in a decreased HIF-1α protein expression (Figure (Figure3A),3A), which was linked with a down-regulation of CREB phosphorylation at serine 133 in HER-2/neu+ cells. This effect was already visible upon treatment of HER-2/neu+ cells with 1 μM 2-ME and was not further enhanced in the presence of higher 2-ME concentrations (Figure (Figure3B).3B). In contrast, NIH3T3 cells exhibited the highest CREB phosphorylation at 1 μM 2-ME, which was strongly reduced with higher concentrations of this substance. Furthermore, electromobility shift assays (EMSA) revealed that 2-ME inhibited the formation of CREB-CRE complexes by decreasing the phosphorylation of CREB (Figure (Figure3C3C).
In addition, blocking HIF-1α changed the morphology of NIH3T3 from a fibroblast-like to a foci-like grown pattern, while HER-2/neu+ cells lost their ability to form bigger foci (Figure (Figure3D).3D). It is noteworthy that the treatment of both cell lines with 10 μM 2-ME did not altered the cell vitality significantly (data not show), but decreased concentration-dependent the proliferation (Figure (Figure3E)3E) and the metabolic / glycolytic activity (Figure (Figure3F3F).
In order to determine whether HER-2/neu+ cells are more resistant to apoptosis compared to shCREB cells, cells were cultured in medium supplemented with different FCS and glucose concentrations. CREB down-regulated HER-2/neu+ cells, but not the parental HER-2/neu+ cells cultured in 0.1 % FCS showed a reduced cell vitality as determined by flow cytometry upon annexin V / PI stainings (Figure (Figure4A).4A). Cultivation of HER-2/neu+ cells with increasing FCS concentrations correlated with higher CREB and pCREB levels, which were further correlated with the higher frequency rates of vital cells (Figure (Figure4B4B).
In addition, glucose deprivation resulted in reduced cell vitalities in particular in CREB down-regulated HER-2/neu+ cells. The increased apoptosis rates were linked to reduced pCREB levels (Figure (Figure4C)4C) suggesting that growth properties, in particular the neoplastic features of HER-2/neu+ cells, are tightly linked with cell vitality. Moreover, supplementation of high glucose concentrations caused a significant induction of the corresponding pCREB levels (Figure (Figure4D4D).
Based on the results of the proteomics profiling and glucose deprivation experiments alterations within the glycolytic pathway in HER-2/neu+ cells vs. CREB down-regulated HER-2/neu+ cells are expected. Therefore, modulations of glycolysis pathways were determined by the extracellular acidification rates (ECAR). CREB down-regulation in HER-2/neu+ cells led to a marked decrease of the ECAR under baseline conditions as well as induction of glycolysis by adding glucose (Figure (Figure5A).5A). The glycolytic capacity of the cells, meaning the highest glycolytic activity under inhibition of the mitochondrial ATPase with oligomycin was lower in CREB knock down cells. This was accompanied with a diminished glycolytic activity observed in these cells (Figure (Figure5B).5B). In line with this observation, a reduced glucose uptake rate was recorded in CREB down-regulated HER-2/neu+ cells (Figure (Figure5C),5C), which was further associated with significantly decreased expression levels of the glucose transporter GLUT1 (Figure (Figure5D)5D) as well as with diminished pyruvate and citrate concentrations (Figure 5E, 5F). Interestingly, the concentrations of both glycolytic metabolites were reduced in the oncogenic HER-2/neu+ cell line after CREB knock down, but not in the parental NIH3T3 after diminishing of CREB by shRNA (Figure 5E, 5F). Under hypoxia the loss of pyruvate was even more pronounced in HER-2/neu+ shCREB cells, while it was unchanged in NIH3T3 shCREB cells.
To validate if the loss of CREB binding activity has an effect on the cellular metabolism, two CREB inhibitors were used: 666-15, a potent inhibitor of CREB-regulated gene transcription , which disturbs the interaction of CREB and the CREB binding protein (CBP) resulting also in in vivo anticancer effects, and surfen, affecting the CREB-CRE complex [17, 39]. As determined by EMSA surfen prevented the formation of the CREB-CRE complex, while 666-15 had no effect on this interaction (Supplementary Figure 3A). Furthermore, surfen cannot displace ethidium bromide from the CRE oligonucleotide in a cell-free assay system (Supplementary Figure 3B) indicating that surfen did not bind to the CRE DNA element, but could interact with the basic leucine zipper of CREB. Both inhibitors diminished the proliferation by slowing down the cell cycle transition from the G1 to the S phase (Supplementary Figure 3C). This effect was more pronounced upon treatment with 666-15 than with surfen. Both 666-15 and surfen lowered the metabolic activity of HER-2/neu+ cells, while 666-15 but not surfen diminished the metabolic activity of NIH3T3 cells (Figure (Figure6A6A).
Using specific fluorescence dyes we could show that 666-15 treatment resulted in an accumulation of dysfunctional mitochondria in HER-2/neu+ and NIH3T3 cells, while for surfen treatment this effect was only detectable in the HER-2/neu-transformed cell line (Figure (Figure6B).6B). This accumulation of dysfunctional mitochondria can be explained by the CREB-regulated “master switch for mitochondrial biogenesis” peroxisome proliferator-activated receptor gamma coactivator 1α  or by the loss of autophagy-related proteins ATG5 and ATG7 (Supplementary Figure 4A), a mechanism that can prevent mitochondrial dysfunctions in several cell lines [40, 41]. Furthermore, under normoxia and hypoxia an increased intracellular lactate level was determined in HER-2/neu+ shCREB cells (Supplementary Figure 4B). This can be caused by dysfunctional mitochondria and leading to an acidosis. Since pyruvate and citrate levels were decreased upon CREB knock down, we analyzed the levels of the metabolite acetyl-CoA. In both tested cell lines the concentration of acetyl-CoA was was increased when compared to the parental and to the vector control (Supplementary Figure 4C).
Since a loss of mitochondrial membrane potential was detected under treatment with both inhibitors (Figure (Figure6C),6C), the release of cytochrome c and ROS into the cytoplasm is suggested.
One of the CREB-regulated proteins identified to be differentially expressed upon CREB down-regulation was catalase, an enzyme involved in the detoxification of hydrogen peroxide. Since oxidative stress can increase CREB phosphorylation and activity, the effect of different concentrations of H2O2 on HER-2/neu+ cells was determined. While low concentrations (10 μM) increased the CREB phosphorylation rate at Ser133, but not at Ser121 during 4 h, high concentrations of 50 μM H2O2 significantly increased CREB phosphorylation after 30 min. to 4 h (Figure (Figure7A).7A). The phosphorylation rate of AKT was strongly increased in response to exposure to 50 μM H2O2, while the phosphorylation level of ERK reached the maximal signal after a 4 h incubation period.
The most efficient detoxification of H2O2 was determined in the presence of 30 μM H2O2. Under these conditions almost 90 % of H2O2 was degraded after 2 h, while at higher or lower concentrations more than 50 % of the starting H2O2 levels remained still detectable in the given culture supernatants (Figure (Figure7B).7B). Exposure to 50 μM H2O2 significantly induced necrosis after 24 h (Figure (Figure7C),7C), which was absent at lower H2O2 concentrations (data not shown). Furthermore, incubation with formaldehyde over the same time had no effect on the vitality of the cells. However, long-term intervals (24 h) exposure resulted in increased phosphorylation rates of CREB (Figure (Figure7D)7D) regardless of the type of toxic compound (H2O2, formaldehyde).
Using the ROS sensitive dye 2,7-DCFH an accumulation of ROS in CREB down-regulated cells and in HER-2/neu+ transformants was detected, which was higher than that of parental NIH3T3 cells (Figure (Figure8A).8A). In the presence of the HIF-1α inhibitor 2-ME an accumulation of ROS was also found in HER-2/neu+ cells (Figure (Figure8B).8B). CREB knock down decreased the cell viability in the presence of H2O2 but not of formaldehyde-treated cells up to 60 % when compared to untreated cells (Figure (Figure8C).8C). We therefore analyzed the mRNA expression of three “detoxifying” enzymes important in cellular homeostasis: catalase, esterase D, an enzyme an enzyme involved in the detoxification of formaldehyde and glutathione synthetase, which is an important antioxidant. Only catalase mRNA expression levels were decreased in response to CREB down-regulation, while the mRNA expression levels of esterase D was not altered (Figure (Figure8D).8D). The decreased mRNA expression levels of catalase was accompanied by diminished catalase protein expression (Figure (Figure8E)8E) and consequently by a lower catalase activity (Figure (Figure8F).8F). In addition, glutathione synthetase was 2.1-fold up-regulated in CREB down-regulated HER-2/neu+ cells, but not in the parental NIH3T3 cells. However, free glutathione remained unaltered in all cell lines analyzed (Figure (Figure8G8G).
Next, we tested the effect of H2O2 as a ROS on the glycolysis and the mitochondrial activity. In the culture supernatant of H2O2-treated cells lower concentrations of lactate was detectable than in the DMSO control (Figure (Figure9A).9A). Furthermore, H2O2 reduced the uptake of glucose as determined using the glucose analogue 2-NBDG (Figure (Figure9B).9B). Both mechanisms were concentration dependent. This led to lower metabolic activity (Figure (Figure9C)9C) and reduced ATP production (Figure (Figure9D),9D), which was not only due to a decreased cell proliferation (Supplementary Figure 5). However, higher concentrations of H2O2 decreased mitochondrial membrane potential (Figure (Figure9E),9E), as well as induced the cleavage of caspase-3 (Figure (Figure9F9F).
Major advances have been made in our understanding of the molecular mechanism of CREB function in malignant transformation and its association with hypoxia , but so far there exists only little information about the role of CREB in modulation of the cellular metabolism. Therefore, the aim of this study was to determine the effects of CREB and/or hypoxia-induced pathways on the cellular metabolism. Using 2-DE-based proteomic profiling of HER-2/neu+ versus HER-2/neu+ shCREB cells differentially expressed proteins were identified mainly involved in cell metabolism and localized in the cytoplasm. In particular, the expression of proteins associated with the glycolysis was altered. Validation of the CREB-regulated genes revealed in some cases a coordinated down-regulation of the respective mRNA and protein expression levels, while e.g. PGK 1 transcription remained unchanged upon CREB modulation, but altered at the protein level suggesting a posttranscriptional control. Thus, metabolic changes mediated by CREB seem to occur at both the transcriptional as well as at the posttranscriptional level. In this context it is also noteworthy that hypoxia could lead to posttranslational modifications of CREB . The increased glycolysis of HER-2/neu+ cells was in accordance with the neoplastic properties of these cells. Different phosphorylation of CREB serine residues can affect metabolic enzymes, like PEPCK .
The altered metabolism due to HER-2/neu-controlled CREB overexpression modulates the TME. High lactate concentrations and low pH values were found in the supernatant of HER-2/neu+ cells, which might affect the anti-tumoral immune response. Indeed, acidic pH values could negatively interfere with the activity and function of effector T cells, dendritic cells as well as NK cells . Furthermore, the direct influence of CREB on the cellular metabolism was strengthened by altered intra-cellular LDH levels. LDH representing the key enzyme of the glycolytic cascade and converting pyruvate to lactate is a CREB target . Up-regulation of LDH levels led to an increased aerobic glycolysis, while CREB knock down not only reduced LDH levels in our HER-2/neu model system, but also in breast cancer . The activation of the MEK/ERK pathway by HER-2/neu increased the CREB binding to the CREB-binding site (CRE) in the LDH promoter leading to its transcriptional up-regulation . Furthermore, loss or reduced LDH expression abrogated cell proliferation in vitro and tumor formation in vivo in particular in triple negative mammary carcinoma , while high levels of LDH expression could be associated with poor patients’ outcome and inhibition of immune surveillance .
Comparative analysis of HER-2/neu+ and HER-2/neu+ shCREB cells demonstrated that CREB suppression correlated with an increased respiration and respiratory reserve. Thus, CREB-controlled altered metabolism was also associated with a reduced sensitivity to undergo apoptosis since HER-2/neu+ cells were able to grow under limited serum and glucose concentrations. This could be even further altered under hypoxic conditions leading to pronounced changes of the intracellular lactate dehydrogenase levels, extracellular lactate concentrations as well as extracellular pH values. In silico analysis of the transcriptional profile of chronic myeloid leukemia cells, in which CREB was knocked out demonstrated an upregulation of genes involved in tumor initiation and progression as well as in hypoxic signaling . Since tumors with a higher glycolytic rate were associated with a worse prognosis accompanied by a poor clinical outcome of patients [48, 50], targeting of metabolic pathways might become an attractive therapeutic approach to treat cancer patients.
Increased ROS accumulation could be caused by mitochondrial dysfunction through the endoplasmic reticulum . Furthermore, decreased mitochondrial activity implemented by mitochondrial dysfunction could activate CREB , which could in turn promote mitochondrial biogenesis . Therefore, we tested whether oxidative stress like hydrogen peroxide could induce CREB activity in vitro. By upregulation of the AKT pathway the phosphorylation of CREB at Ser 133 was increased. Similar results were reported in a PKA-independent pathway by Pregni and co-authors  in neuronal cells, while treatment of melanoma cells with H2O2 increased CREB expression by the cAMP/PKA pathway . Furthermore, H2O2 could stimulate phosphorylation of CREB by AKT thereby promoting mitochondrial respiration and biosynthesis , while the expression levels of esterase D (formyl glutathione hydrolase), an enzyme important for formaldehyde detoxification and upregulated by K-Ras mutations in murine fibroblasts , were not altered under these conditions.
CREB supports metabolism in several ways: First it can up-regulate the expression of Glut-1, therefore increasing uptake potential of glucose and intracellular glucose levels. Second, it can positive regulates some glycolytic enzymes (PKM2, TPI, alpha-enolase). Third it can enable mitochondrial respiration and biogenesis under stress conditions like hypoxia and ROS, therefore allowing a better survival of tumor and tumor cell (Figure (Figure1010).
However, further research is required for better understand the role of TFs, like CREB, in the mitochondria and their effect on the regulation of the cellular metabolism and oxidative stress under physiologic, but also pathophysiologic conditions .
The generation and culture conditions of HER-2/neu- NIH3T3 cells, their HER-2/neu+ derivative and their respective CREB variants have been recently described [11, 34]. HER-2/neu-overexpressing cells with high levels of pCREB were termed HER-2/neu+ cells, HER-2/neu+ cells, in which CREB was down-regulated by shCREB were termed HER-2/neu+ shCREB cells.
For determination of the impact of FCS and glucose concentrations on the cell growth, cells were cultivated in 1 – 10 % FCS and 0.5 – 4.5 g/l glucose, respectively.
For hypoxia induction, cells were incubated in a 1 % (v/v) O2, 5% (v/v) CO2 atmosphere at 37°C for the indicated time spans . Normoxic controls were cultured using standard conditions.
Cells were treated for 24 – 48 h with different inhibitors (HIF-1α inhibitor 2-methoxyestradiol, CREB-CBP inhibitor 666-15 (Tocris), CREB-CRE inhibitor surfen (Sigma), proton pump inhibitor Esomeprazole) and toxic components (H2O2, formaldehyde). Cytotoxicity of the substances was analyzed by treatment of cells with increasing concentrations of these inhibitors to determine the IC50 value by using the cytoto assay kit (Promega). Glucose uptake was determined as recently described in detail .
Total cellular RNA from 5 x 105 cells/sample was extracted and subjected to qPCR analysis as recently described . The specific primer sequences and PCR conditions used were provided in Supplementary Table 1. Relative expression levels were calculated according to the ΔCt method and normalized against β-actin and/or GAPDH as internal controls. Three independent experiments were performed.
Total protein was isolated from 5 x 106 cells/sample and the resulting lysates subjected to Western blot analysis as recently described  using the panel of antibodies listed in Supplementary Table 2. Blot membranes were stripped with 2.5 M glycine (pH 2) for 15 min. and were then re-probed with a new antibody after re-blocking with skim milk. For data analysis, three independent experiments were used.
The loss of CREB phosphorylation and protein was further quantified with a CREB specific ELISA system (Cayman). 10 μg total protein were incubated with the antibody coated plates for 1 h at 25°C before CREB phosphorylation and CREB protein expression were determined by adding the primary antibody. After three washing steps the samples were incubated with the HRP-linked secondary antibody for 30 min., washed three times and reaction was started by adding TNB substrate. After 30 min. the reaction was stopped with 1 mM acetic acid and absorption was measured af 490 nm. Three biological samples with three reactions were measured with each cell line.
Cells were harvested after cultivation for 24 h under normoxia or hypoxia, respectively, and then stained for 20 min. with an anti-HER-2/neu-PE-conjugated (Becton Dickinson) antibody or with the IgG1-PE conjugated isotype control (BD), respectively, as recently described . Fluorescence intensity was determined by flow cytometry on a FACSCalibur (BD). The results were expressed as mean specific fluorescence intensity (MFI) of three independent experiments.
For measuring the mitochondrial mass and the mitochondrial activity per single cell, the fluorescent dyes MitoTracker green (Cell signaling) staining the mitochondrial membrane, and MitoTracker red CFXRos (Invitrogen), which changed to a fluorescent dye under mitochondrial activity, were used. Cell pellets were resuspended in PBS supplemented with 100 nM of both dyes and then stained for 30 min. at 37°C in the dark. Cells were washed with PBS and directly analyzed by flow cytometry. Three biological replicates were used for every condition.
Annexin V - propidium iodide staining was used for the detection of apoptotic and necrotic cells as previously described . For fluorescence analysis, 10 μM acridine orange and 10 μM ethidium bromide were added to the given cell culture. After a 10 min incubation period the samples were analyzed by fluorescence microscopy using both green / red filter settings. Three biological replicates with three biological replicated each were photographed.
To determine the caspase-3 activation through cleavage, the FITC Active Caspase-3 Apoptosis Kit (BD) was used following the manufacturer’s instruction. Briefly, 2.5 x 106 cells were fixed in 500 μl cytofix/cytoperm at 4°C for 30 min., and then washed once with 1 ml cytowash, before they were stained in 100 μl cytowash and 10 μl anti-active-caspase-3-FITC for 30 min. After one additional washing step, 1x105 cells/sample were analyzed on a FacsCalibur (BD). The experiment was independently repeated twice.
The concentration of H2O2 in the culture media was determined by the xylenol orange-Fe2+ method described by Gay and co-authors . The enzymatic activity of PKM in cell lysates was measured by the method of Tietz and Ochoa . 100 μl cell lysate were incubated with reaction puffer (50 mM Tris HCl pH 7.5, 4 mM MgCl2, 75 mM KCl, 0.2 mM ADP, 1.5 mM PEP, 0.2 mM NADH, 12 U/ml LDH) for 5 min. at room temperature. The reaction was started by addition of pyruvate kinase (end concentration 0.3 U/ml) and the absorbance change at 340 nm was recorded for 5 minutes.
Activity of prolyl endopeptidase was determined fluorometrically by the cleavage of the substrate Z-glycyl-prolyl-4-methylcoumarinyl-7-amide as described by Goossens and co-workers  with the following modifications: 100 μl of cell lysate were diluted with 100 μl incubation buffer (100 mM KH2PO4 pH 7.4, 100 μM NaN3, 1 mM DTT, 1 mM EDTA). Next the samples were incubated for 15 min. at 37°C and 10 μl of substrate was added (5 mM in DMSO). After incubation for 2 h in the dark at 37°C 500 μl 1.5 mM acetic acid were added and mixed. Fluorescence was analyzed with an ELISA reader (Tecan) with λex 370 nm and λem 440 nm.
PGAM1 activity was measured following the protocol published by Hallows and co-authors .
TPI enzyme activity in cell lysate was analyzed by adding 2 ml reaction buffer (220 mM triethanolamine HCl, pH 7.6, 3 μM DL-glyceraldehyde-3-phosphate, 0.26 μM NADH, 50 mU glycerol-3-phosphate dehydrogenase, 0.2 U/ml triose phosphate isomerase) to 100 μl of cell lysate. The decrease in absorbance at 240 nm was recorded by a spectrometer.
PGK activity was determined by ATP formation: 100 μl cell lysate were mixed with 50 mM potassium phosphate, 5 mM 1,3-bisphosphoglycerate, 1 mM ADP, 5 mM CaCl2. The formation of Ca3(PO4)2 was measured by 340 nm after incubation by 25°C for 15 min.
For determination of the PRX 4 activity the protocol from Nelson and Parsonage  was used.
The enzymatic activity was normalizied to the total protein concentration in the cell lysate. All experiments were performed with three cell preparations.
Bioenergetics of CREB down-regulated and CREB expressing HER-2/neu+ and HER-2/neu- control cells were determined using the XF96e Extracellular Flux analyzer (Seahorse Bioscience) as recently described . Briefly, 2x105 cells/well were reseeded in specialized tissue culture plates (96FX micro well plate) and subsequently immobilized using CELL-TAK (BD Biosciences). One hour prior measurement, cells were incubated at 37°C in a CO2-free atmosphere. First, the basal extracellular acidification rate (ECAR), as an indicator for lactic acid production or glycolysis, was recorded. In the next step ECAR in response to the application of 1 μM oligomycin (XF Cell Mito Glyco Test Kit, Seahorse Bioscience) as well as 10 mM glucose was evaluated. All experiments were performed in four independent experiments with at least hexaplicates.
The amount of lactate in 100 μl cell culture supernatants was analyzed with a lactate detection kit (Sigma) according to the manufacturer’s instructions in a 96 well plate format. The deproteinated samples were measured at 570 nm in an ELISA reader (Tecan) after enzymatic reaction. Three independent experiments were performed using three technical replicates in each run and the results were expressed as mean ± SD using bar charts.
For the detection of pyruvate and citrate in cell lysates commercial kits (Biovision) were used following the manufacturer’s protocol. Samples and standards were measured with an ELISA reader as described above.
Acetyl-CoA was measured fluorimetrically with an acetyl-CoA assay kit following the manufacturer’s instructions (Sigma).
The amount of glucose in the culture medium was controlled with the DNSA method as described by Miller .
Proliferation of treated cells was determined by measuring the DNA synthesis with the cell proliferation ELISA, BrdU (colorimetric) kit (Roche). 5 x 104 cells were seeded per 96/well and were treated with or without H2O2. The inclusion of BrdU was analyzed following the manufacturer’s instructions. Three biological replicates with three biological sample each were analyzed at an ELISA reader (Tecan).
For cell cycle analysis the DNA content per cell was measured as described previously . Briefly, after synchronization of the cells by serum deprivation (0.5% FCS) for 48 h, cells were incubated with media containing inhibitors and 10% FCS for the indicated time span. The cells were fixed after harvesting in 70% ethanol for at least 24 h by 4°C, following the isolation of nucleus by washing once with PBS, 0.5% tween 20 and then PBS, 0,5% tween 20 and 100 mM citrate. RNA was digested with RNase A treatment and DNA was stained with 0.1 mg/ml PI. DNA content was measured with a FacsAria III (BD) with three biological replicates for each condition.
Binding of CREB to the CRE DNA element was analyzed by an EMSA. Nuclear extracts of HER-2/neu+ cells were incubated with a biotinylated oligonucleotide (TGACGTCA) and the testing substances as described by Bukur and co-authors .
Furthermore, the ethidium bromide displacement assay was used to determine whether surfen interacts with the CRE element and therefore prevents binding of CREB or whether surfen binds to the CREB protein. The assay was conducted as published in Rishi and co-workers  for three times independently. The following hairpin oligonucleotids were used: TGACGTCAAAAAATGACGTCA (CREB), TGACTCAAAAAATGACTCA (AP1), TTAATTAAAAAA ATTAATTAA (AT rich), GGCCGGCCAAAAAGGCC GGCC (GC rich).
Cells were incubated in 10 μM 2',7'-dichlorodihydrofluorescein diacetate (2,7-DCFH) for 15 min. at room temperature and then were washed twice with PBS. Fluorescence intensity was analysed by flow cytometry (BD) as recently described .
The glutathione levels were determined using the GSH/oxidized GSH ratio kit (Calbiochem) as previously described . Briefly, 1 x 106 cells were incubated for 48 h, washed twice with PBS, scratched of the plates and resuspended in 100 μl PBS. After cell lysis obtained by 6 cycles of freezing and thawing 900 μl 5 % (w/v) meta phosphoric acid/sample was added. 50 μl of the cell supernatant was analyzed according to the manufacturer’s instructions at 412 nm in a spectrometer (Merck). The amount of glutathione was correlated to the protein concentration determined with the bicinchoninic acid kit (Thermo).
Catalase staining was performed according to Weydert and Cullen . 150 μg non-denaturated total protein lysate was separated with a native 8 % PAGE gel in a cold room (6°C) at ~40 mA for 6 to 8 h. After electrophoresis the gel was washed with aqua dest. and incubated for 10 min by RT in 0.03 % H2O2. After washing twice with aqua dest. the gel was stained in 2 % (w/v) FeCl3 and 2 % (w/v) K3[Fe(CN)6]) until clear bands appeared. Finally, the gels were extensively washed with water and documented with a gel scanner. Four single protein lysates were used for every condition.
1 x 106 cells were stained with 0.1 μM BCECF-AM for 30 min. at 4°C and washed twice with culture medium. 1 x 104 cells/well were seeded into a black 96 well plate with a clear bottom and incubated at 37 °C for 24 h under normoxic or hypoxic conditions. The fluorescence intensity/well was determined in a fluorescence reader (Tecan, excitation wavelength 480 nm, emission wavelength 535 nm). For the analysis of pH stability and the regeneration time under normoxia, the fluorescence intensity was monitored for one h. For the standard curve the cells were incubated in buffer with known pH whereas nigericin-KCl (25 μM and 125 mM) was used for permeabilization.
For determination of the differential protein expression pattern upon CREB down-regulation, 2-DE-based proteome analysis followed by mass spectrometry was performed as recently described . Three biological replicates with four to six technical replicates (gels) were used for proteome analysis. Coomassie stained gels were documented with a gel scanner and gel pictures were compared with the Delta2D software (Decodon). The cut-off for up and down-regulation of protein spot was < 0.5 and > 2.5. Putative CREB-regulated proteins were cut-out with an automatic spot picker (Spothunter, Herolab) and prepared for MALDI TOF/TOF-MS with a α-cyano-hydroxycinnamic acid matrix. MS data were identified with the Mascot database. Protein spots with a score value higher than 55 and found in all three biological replicates were used for further validation processes.
Differences between groups were analyzed by Mann-Whitney test, Student’s t test or ANOVA as appropriate. Data are expressed as mean ± SD and the corresponding p values reported are 2-sided. Significance was accepted if p values were ≤ 0.05 (*). The statistical analysis of the graphs and tables compares the CREB diminished cells with their parental cell lines (NIH3T3 or HER-2/neu+) if nothing else is stated in the figure legend. For the analysis of putative CRE regions in the listed gene promoter sequence the CREB target gene database (natural.salk.edu/CREB/) was used .
We would like to thank Sylvi Magdeburg for excellent secretarial work.
This work was supported by the Mildred Scheel Foundation (BS).
CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.