RNAi screening in glioma stem-like cells identifies PFKFB4 as a key molecule important for cancer cell survival


The concept of cancer stem-like cells (CSCs) has gained considerable attention in various solid tumors including glioblastoma, the most common primary brain tumor. This sub-population of tumor cells has been intensively investigated and their role in therapy resistance as well as tumor recurrence has been demonstrated. In that respect, development of therapeutic strategies that target CSCs (and possibly also the tumor bulk) appears a promising approach in patients suffering from primary brain tumors. In the present study, we utilized RNA interference (RNAi) to screen the complete human kinome and phosphatome (682 and 180 targets, respectively) in order to identify genes and pathways relevant for the survival of brain CSCs and thereby potential therapeutical targets for glioblastoma. We report of 46 putative candidates including known survival-related kinases and phosphatases. Interestingly, a number of genes identified are involved in metabolism, especially glycolysis, such as PDK1 and PKM2 and, most prominently PFKFB4. In vitro studies confirmed an essential role of PFKFB4 in the maintenance of brain CSCs. Furthermore, high PFKFB4 expression was associated with shorter survival of primary glioblastoma patients. Our findings support the importance of the glycolytic pathway in the maintenance of malignant glioma cells and brain CSCs and imply tumor metabolism as a promising therapeutic target in glioblastoma.


Glioblastoma is the most common and most malignant primary brain tumor in adults. Despite multimodal aggressive treatment, comprising surgical resection, local radiation therapy and systemic chemotherapy, the average patient survival time is still in the range of 1 year after diagnosis (Stupp et al., 2007). Several recent studies suggested a hierarchical cellular organization in glioblastomas, which consist of heterogeneous cell populations that differ in their tumorigenic potential. Indeed, it is believed that gliomas are initiated and maintained by a small sub-population of cells that is capable of extensive self-renewal and recapitulation of the original tumor in nonobese diabetic/severe combined immunodeficient mice (Singh et al., 2003; Lee et al., 2006; Vescovi et al., 2006). These so-called brain cancer stem-like cells (CSCs) have been implicated not only in tumor initiation, but also in therapy resistance and tumor recurrence. It was shown that brain CSCs are enriched in recurrent malignant gliomas and are resistant to chemo- and radiotherapy (Bao et al., 2006; Eramo et al., 2006; Liu et al., 2006; Al-Hajj, 2007).

Given their central role, it has appeared that therapeutically targeting this particular population of cells is crucial for successful glioblastoma treatment (Reya et al., 2001; Pardal et al., 2003; Park et al., 2009). One possible direction is to uncover the mechanisms, by which brain CSCs escape apoptosis, and ultimately provide knowledge for the development of drug therapeutics targeting the involved genes and pathways.

Kinases and phosphatases control the reversible process of phosphorylation, which is involved in each intracellular pathway, including cell survival signalling. Therefore, identification of the kinases and phosphatases whose inhibition induces the death of brain CSCs may pave the way toward novel therapeutic targets. In this study, we performed a loss-of-function screen using a lentiviral short-hairpin RNA (shRNA) library representing the entire human kinome and phosphatome, and identified 46 candidates that are essential for the survival of brain CSCs in vitro. These candidates included key regulators of the glycolytic pathway, underlining the central role of metabolism for the survival of brain CSCs. One of the glycolysis-associated candidate genes, PFKFB4, was functionally characterized in brain CSCs in vitro, and the importance of lactate and adenosine triphosphate (ATP) production for the survival of brain CSCs was assessed.


Kinome and phosphatome RNA interference (RNAi) screen to induce brain CSC death

Lentiviral short-hairpin libraries were used recently to conduct loss-of-function screens in mammalian cell lines and turned out to be particularly useful in slowly dividing cells (Wurdak et al., 2010). We performed a large-scale silencing screen in vitro to identify kinases and phosphatases that are essential for the survival of brain CSCs. We transduced NCH421k, whose stem-like cell properties have been previously characterized (Campos et al., 2010), with a lentiviral vector-based shRNA library that targets each of the 682 and 180 known and putative kinases and phosphatases, respectively (average of 4.5 shRNAs per target, The RNAi Consortium, Broad Institute from MIT and Harvard, MA). Viral stocks were produced in a 96-well plate format and used at a sufficient titer to transduce brain CSCs.

In order to assess the efficiency of the viral transduction, cells transduced with lentiviral particles produced from a control plasmid encoding green fluorescent protein (GFP) were included in each plate. Viral transduction was efficient as indicated by 70–90% GFP-expressing cells. As positive control for the cell death phenotype, brain CSCs were transduced with a specific shRNA targeting BCL2. After 6 days, cell death was measured by propidium iodide staining and fluorescence-activated cell sorting (FACS). In the screen, we considered only those genes as positive hits whose median over all shRNAs targeting this gene showed an increased level of cell death of 1.5-fold (median of all shRNAs plus 2.5-fold median absolute deviation) relative to that of the control (scrambled) shRNA. Overall, 39 kinases and 7 phosphatases fulfilled this criterion (Figure 1 and Tables 1 and 2) and were defined as survival genes. Moreover, each of these genes was targeted by at least two shRNAs that showed a fold increase of cell death rate 1.5-fold.

Figure 1

Identification of survival kinases and phosphatases. NCH421k cells were seeded in 96-well plates and transduced with shRNAs directed against all known and putative kinases and phosphatases in the human genome. Cells were incubated for 6 days to allow target knockdown and cell death was measured by PI staining and analyzed by FACS. The graph depicts the median fold change of cell death for all kinases and phosphatases normalized to control shRNA. The genes whose loss of function increased cell death by >1.5-fold were considered as positive hits. The red line represents the median of all shRNAs plus 2.5-fold median absolute deviation (MAD).

Table 1 Putative survival kinases identified in the RNA interference (RNAi) screen
Table 2 Putative survival phosphatases identified in the RNA interference (RNAi) screen

Validation of survival genes

Next, we generated a focused library comprising the top 10 candidates from the first screen, each represented by 2 to 3 independent shRNAs that gave the highest rate of cell death. To ensure that the differences observed were not caused by ‘off-target’ shRNA effects, we verified the level of silencing of each shRNA by quantitative real-time reverse transcription–PCR (qRT–PCR). Of these top 10 genes, 9 were represented by at least two independent shRNAs that decreased their target mRNA transcript levels by 2- to 10-fold, suggesting that the observed phenotype was because of the silencing of the intended targets (Supplementary Figure S1A). Only silencing of NTRK1 was not confirmed, as its basal level of expression in NCH421k was too low to be measured by qRT–PCR. Thus, its detection in our screen was most probably because of an ‘off-target’ effect.

We performed a validation screen on two additional brain CSC cultures (NCH441 and NCH644), as well as one normal neural stem cell (NSC) line (BLV4), monitoring the activation of caspases 3 and 7, as they are key mediators of apoptosis. Measurements were done 4, 5, 6 and 7 days after transduction in three independent experiments (Figure 2 and Supplementary Figures S1B and S2). The results confirmed that knockdown of each candidate gene resulted in a higher fold change of apoptosis when compared with scrambled shRNA. Moreover, as shown in Figure 2, the silencing increased significantly the fold change of apoptosis of brain CSCs compared with normal NSCs (Student's t-test for ranked transformed data; P<0.05). However, shRNA-1 targeting PAK2 (Figure 2) and shRNA-2 specific for PRKCA (Supplementary Figure S2) showed an increased fold change of apoptosis in all cell lines.

Figure 2

Candidate genes identified induce apoptosis upon silencing. Verification of candidate gene knockdown-induced apoptosis as determined by caspase assays in three different brain CSC lines (NCH421k, NCH441 and NCH644) and in normal NSCs (BLV4). Apoptosis was measured 4, 5, 6 and 7 days after transduction. The time points that showed the highest caspase activity are shown. The fold change of apoptosis was normalized to negative control scrambled shRNA and is depicted as mean of independent biological triplicates for one shRNA (shRNA-1) per targeted gene. BCL2 shRNA was used as positive control. Error bars represent the s.d. of three independent biological replicates. Gene knockdown induced a significantly increased rate of apoptosis in brain CSCs when compared with normal NSCs (P<0.05), with the exception of PAK2 shRNA (P=0.59). Statistical significance was calculated using Student's t-test for ranked transformed data.

Gene expression of PFKFB4 is associated with glioblastoma prognosis

To assess the clinical relevance of the validated candidates, we compared their mRNA expression among a series of astrocytic gliomas of different grades (WHO (World Health Organization) grades II, III and IV, including primary and secondary glioblastomas) using previously published microarray data (GSE15698510) (Toedt et al., 2011). Remarkably, the mRNA expression of PFKFB4, whose gene product plays a key role in energy metabolism, significantly differentiated IDH1 wild-type primary glioblastomas from the secondary glioblastomas as well as diffuse and anaplastic astrocytomas, which mostly carried IDH1 gene mutations (Figure 3a). Moreover, primary glioblastoma patients whose tumors demonstrated PFKFB4 mRNA expression levels above average had a significantly (P<0.0001) shorter overall survival time than glioblastoma patients whose tumors displayed lower than average PFKFB4 expression (Figure 3b). We confirmed the results using the National Cancer Institute's REMBRANDT (Repository of Molecular Brain Neoplasia Data) database (http://rembrandt.nci.nih.gov). Kaplan–Meier survival analysis revealed that increased PFKFB4 expression >3-fold compared with normal brain was associated with significantly reduced overall survival of glioblastoma patients (P=0.0201; Supplementary Figure S3A). To test whether the prognostic value of PFKFB4 is independent of IDH1 mutation status in grade IV gliomas, we calculated Cox hazard models for the overall survival time considering death of disease only and censoring all other events. A model with PFKFB4 alone with n=39 showed a significant effect of PFKFB4 expression on survival (P=0.022). IDH1 alone showed no significant association with survival (n=68, P=0.95). In a model (n=39) stratified for IDH1, PFKFB4 was associated with survival (P=0.017). Proportional hazard assumption was fulfilled for all models.

Figure 3

Expression of PFKFB4 and PFKFB3 in different glioma entities. (a) PFKFB4 and PFKFB3 mRNA expression levels in the glioma expression data set GSE15698510. Box plot shows expression levels of PFKFB4 and PFKFB3 in diffuse astrocytomas: World Health Organization (WHO) grade II (AII, n=8), anaplastic astrocytomas WHO grade III (AAIII, n=13), secondary glioblastoma WHO grade IV (sGBIV, n=10) and primary glioblastoma WHO grade IV (pGBIV, n=38) relative to the normal brain control tissue (n=4). Note that the median PFKFB3 and PFKFB4 mRNA expression levels are significantly higher in primary glioblastoma compared with the normal brain tissue samples and all other tumor groups (*P<0.02 and **P<0.002). (b) Kaplan–Meier analysis showing the association of PFKFB4 expression with overall survival of patients with IDH1 wild-type primary glioblastoma. Low- and high-expression groups were separated according to the mean expression value of all samples studied. The P-value was calculated applying the Mantel–Haenszel test (P<0.0001). Using a bootstrapping approach with 1000 draws, the P-value for differences in survival based on the expression of PFKFB4 was confirmed by reselecting the mean expression value.

Overexpression of PFKFB4 isozyme is cancer specific

PFKFB4 is one of the four genes that encode 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase (PFK2/FBPase-2), and originally was found to be expressed in the testes (Sakata et al., 1991). It is a bifunctional enzyme having both kinase and phosphatase functions, and plays a key role in energy metabolism by controlling glucose utilization.

To determine the cancer specificity of PFKFB4 expression, we compared its mRNA expression levels in the different brain CSC lines with that in a pool of normal brain samples. As shown in Figure 4a, qRT–PCR analysis demonstrated that brain CSCs have a PFKFB4 mRNA expression that is 10-fold higher when compared with that of adult normal brain tissue. This result was confirmed at the protein level (Figure 4b).

Figure 4

Differential expression of PFKFB3 and PFKFB4 in brain CSCs. (a) Analysis of PFKFB3 and PFKFB4 mRNA expression as determined by qRT–PCR. Total mRNA was extracted from three different brain CSC lines (NCH421k, NCH441 and NCH644). As reference, a pool of normal brain mRNA was included. Results were normalized to mRNA levels of four housekeeping genes. (b) PFKFB4 protein expression in three brain CSC cultures. As reference, a pool of normal brain samples was used.

We next asked whether the overexpression of PFKFB4 was related to the expression of PFKFB3, another isozyme of PFK2 that is expressed in normal brain and overexpressed in numerous tumors (Manzano et al., 1998; Kessler and Eschrich, 2001; Atsumi et al., 2002; Gomez-Brouchet et al., 2007). As depicted in Figure 3a, PFKFB3 showed a 1.3-fold higher mean mRNA expression level in primary glioblastomas relative to normal brain tissue and was not overexpressed in the other glioma groups investigated. Moreover, the level of PFKFB3 expression was not significantly related to overall survival of primary glioblastoma patients (P=0.518, data not shown). In contrast to PFKFB4, PFKFB3 mRNA expression was decreased in brain CSCs when compared with a pool of normal brain mRNAs (Figure 4a). It should be noted that, based on microarray data (Toedt et al., 2011), no expression was observed for the PFKFB1 isoform and that PFKFB2 was equally expressed in the glioma and normal brain samples (Supplementary Figure S3B).

Silencing of PFKFB4 suppresses viability of brain CSCs and inhibits lactate and ATP production

To determine the effect of PFKFB4 knockdown on overall brain CSC population in vitro, transduction of NCH421k with PFKFB4 shRNA was performed in six-well plates. The knockdown level was monitored by western blot analysis (Figure 5a) and cell death was assessed by propidium iodide (PI) staining and FACS. shRNA-2 induced >95% cell death, whereas shRNA-1 showed 50% PI-positive cells (Figures 5b and c).

Figure 5

PFKFB4 knockdown induces an increase of PI-positive brain CSCs. (a) Western blot analysis of PFKFB4 knockdown in NCH421k cells. Levels of silencing were assessed 3, 4 and 5 days after lentiviral-mediated transduction with scrambled shRNA (Ctl), PFKFB4-shRNA1 and -shRNA2. α-Tubulin antibody was used as loading control. (b) Induction of cell death 6 days after lentiviral-mediated transduction with scrambled shRNA, PFKFB4-shRNA1 and -shRNA2. Left panels show representative phase contrast microphotographs of transduced NCH421k. Scale bar, 500 μm. Histogram plots in the right panels depict the PI fluorescence intensity, acquired with a FACSCanto flow cytometer. Transduction efficiency, as determined by GFP-positive cells, was >95% (data not shown). (c) Quantitative analysis of PI fluorescence intensity for 60 000 cells, 6 days after lentiviral transduction.

Malignant glioma cells are known to be resilient to anaerobic conditions created by a hypoxic microenvironment, thanks to their potential to increase glycolysis and thereby production of lactic acid (Floridi et al., 1989). This metabolic adaptation is mediated by changes in gene expression patterns triggered by the activation of hypoxia-inducible factor 1α (HIF-1α) (Lu et al., 2002). Previous studies have highlighted the immediate targeting of PFKFB4 by HIF-1α (Minchenko et al., 2004). Analysis of HIF-1α protein and mRNA expression on microarrays in brain CSCs and normal NSCs (Ernst et al., 2009) revealed a higher expression level in brain CSCs (Supplementary Figure S4A). This observation was made regardless of the normoxic culture conditions, suggesting that glycolysis is increased despite the presence of oxygen, a phenomenon described as the Warburg effect (Lu et al., 2002). To verify that silencing of PFKFB4 impairs metabolism and thereby induces apoptosis of brain CSCs, we measured the level of lactate and of ATP at 3 and 4 days after PFKFB4 knockdown with two independent shRNAs. Lactate secretion and ATP level was significantly reduced in NCH421k, confirming that PFKFB4 silencing induces metabolic changes (Figures 6a and b). As shown previously, inhibition of glycolysis induces the activation/phosphorylation of the adenosine monophosphate (AMP)-activated protein kinase (AMPK), which inhibits the mTOR (mammalian target of rapamycin) pathway and therefore sensitizes tumor cells to death receptor-induced apoptosis (Pradelli et al., 2010). This activation is mainly due to a loss of balance between AMP and ATP content because of a drop of the ATP level. As shown by western blot analysis in Figure 6c, silencing of PFKFB4 with shRNA-1 induces the phosphorylation of AMP kinase. This phenomenon could not be demonstrated when using shRNA-2, most likely because of the high apoptosis rate induced by shRNA-2. To determine whether the mTOR pathway has a relevant role in brain CSC survival, we treated NCH421k cells with rapamycin at different concentrations. Cell viability was assessed by MTS assay 3 days after treatment. The 20–40% decrease of cell viability upon rapamycin treatment suggests a potential role of mTOR in the survival of brain CSCs (Supplementary Figure S4B).

Figure 6

PFKFB4 knockdown induces cellular metabolism changes. (a, b) Lactate and ATP contents were measured using a lactate assay and an ATP assay, respectively, 3 and 4 days after transduction with scrambled shRNA, shRNA specific for BCL2 and shRNA-1 and -2 targeting PFKFB4. Results are depicted as mean of biological triplicates. Data were normalized to 20 000 living cells and are shown as percentage to the lactate content after transduction with scrambled shRNA. shRNA targeting BCL2 was used as an additional negative control, ensuring that the effect observed was not due to apoptosis initiation. Error bars represent the minimum and the maximum values of the biological triplicates (*P=0.06, **P=0.01). (c) NCH421k cells were transduced with scrambled shRNA (Ctl) and shRNA-1 and -2 targeting PFKFB4. Total AMPK and phosphorylation of AMPK on Thr172 were assessed by western blot analysis 5 days after transduction. α-Tubulin antibody was used as loading control.


Several reports have provided evidence that glioblastomas are cellularly heterogeneous tumors containing the so-called brain CSCs that show stem-like features in vitro and give rise to tumor development when xenotransplanted in immunocompromised mice. Furthermore, brain CSCs have been implicated in the resistance of glioblastomas to chemo- and radiotherapy, which emphasizes the importance to target these cells to achieve a more effective therapy (Bao et al., 2006; Eramo et al., 2006; Liu et al., 2006; Al-Hajj, 2007). In this respect, an appealing treatment approach would be to specifically target proteins encoded by the genes that are essential for the survival of brain CSCs.

In this study, we performed a kinome- and phosphatome-wide loss-of-function screen in order to identify genes whose silencing induces apoptosis in brain CSCs. Our results indicate that 5% of the kinome and phosphatome is required for the survival of brain CSCs. However, false-positive and false-negative results due to technical pitfalls inherent in large-scale screening cannot be excluded. As expected, known survival kinases and phosphatases, such as SPHK1, FLT1 and SGK1 (Kasahara et al., 2000; Brunet et al., 2001; Gomez-Brouchet et al., 2007), were identified, as well as DUSP13, CHEK2, ADRBK1, MAP3K1, PPM1B and PRKCA, reported as survival genes in previous screens (MacKeigan et al., 2005; Giroux et al., 2006). This reflects the robustness of our screening settings.

Interestingly, each gene knockdown that was validated in the secondary screen demonstrated a similar phenotype in the three independent brain CSC lines evaluated, but at variable levels, which might be explained by the heterogeneity existing between patient samples. Nonetheless, the phenotype obtained after silencing, which was observed in each of the brain CSC lines tested, and was specific to the tumor cells, as indicated by the comparison with normal NSCs, is of great interest for further characterization and ultimately therapeutics development.

Among the putative kinases and phosphatases that we identified as relevant for brain CSC survival, several are known to be activated by HIF proteins and thereby regulated by the hypoxic microenvironment of the tumor. Indeed, FLT1 (Fms-related tyrosine kinase 1), the vascular endothelial growth factor receptor 1, is regulated by HIF proteins and plays a key role in vascular development (Okuyama et al., 2006). Increased FLT1 expression has been shown in various types of carcinomas, glioblastoma, multiple myeloma and nephroblastoma (Fischer et al., 2008). In addition, our approach highlighted three other kinases, namely PFKFB4, the pyruvate dehydrogenase kinase 1 (PDK1) and the M2 splice isoform of pyruvate kinase (PKM2), which are direct targets of HIF proteins (Kress et al., 1998; Minchenko et al., 2004; Wigfield et al., 2008). This response to low level of oxygen is particularly important in solid tumors, such as malignant gliomas, that are characterized by aberrant neovascularization and poor oxygen diffusion. Most importantly, recent studies have shown that hypoxia also influences the survival and the proliferation of brain CSCs within the tumor and promotes reprogramming toward a stem-like phenotype (Keith and Simon, 2007; Heddleston et al., 2009; Soeda et al., 2009). Another important common feature of these three kinases is their involvement in the glycolytic pathway. Indeed, PKM2 has been reported to be overexpressed in tumor tissue and to be advantageous for tumor cell growth. It is required for aerobic glycolysis and catalyzes the last step by dephosphorylating phosphoenolpyruvate to pyruvate, which leads to a net production of ATP (Christofk et al., 2008).

PDK1 is another important gene whose gene product has an effect on pyruvate metabolism and is responsible for phosphorylation and concomitant inactivation of pyruvate dehydrogenase (PDH). Its expression has been shown to be associated with poor prognosis in head and neck squamous cell carcinoma (Wigfield et al., 2008). Suppression of PDH by PDK1 inhibits the conversion of pyruvate to acetyl-CoA, thereby attenuating mitochondrial function and increasing the production of lactate. Recently, dichloroacetate (DCA), an inhibitor of PDK1, has been shown to induce apoptosis in vitro, thus confirming the phenotype we observed after knockdown of PDK1 in brain CSCs (Wigfield et al., 2008; Michelakis et al., 2010).

The third kinase that has elicited a strong cell death phenotype after silencing in brain CSCs was PFKFB4, which encodes the bifunctional enzyme PFK2/FBPase-2 (6-phosphofructo-2 kinase/fructose-2,6-biphosphatase (PFKFB1-4)). This enzyme modulates the intracellular concentration of the allosteric glycolytic activator, fructose-2,6-biphosphate (F2,6BP), which is a key regulator of glycolysis (Yalcin et al., 2009). It has been originally identified in the testes and shown to be highly expressed in both colon and breast cancers (Minchenko et al., 2004; Gomez et al., 2005). The homologous PFKFB3 encodes an inducible form of PFK2/FBPase-2 that is ubiquitously expressed in many normal tissues, such as brain and liver, and overexpressed in different cancers (Manzano et al., 1998; Kessler and Eschrich, 2001; Atsumi et al., 2002). Both PFKFB3 and PFKFB4 are induced by hypoxia in various tumors. Interestingly, primary glioblastomas showed a significant higher expression of PFKFB3 and PFKFB4 when compared with secondary glioblastomas as well as with the lower-grade astocytomas. Furthermore, high expression levels of PFKFB4 correlated with poor survival in primary glioblastoma patients, whereas PFKFB3 expression was not related to the clinical outcome of these patients. In addition, mRNA expression of both PFKFBs showed an opposite pattern in brain CSCs, in which PFKFB4 was overexpressed whereas PFKFB3 was downregulated when compared with normal brain tissue. These findings suggest that PFKFB4 is the main PFK2/FBPase-2 isozyme that regulates glycolytic flux in malignant glioma cells. This can be partly explained by the common loss of chromosome 10p in glioblastoma that contains the locus of PFKFB3. However, we cannot exclude that PFKFB3 plays a role in brain CSCs survival as suggested by its mRNA expression level in glioblastoma, but it is also required for normal brain tissue homeostasis, as shown by its expression in normal brain. Our results thus suggest PFKFB4 as a key molecule for the survival of brain CSCs. Moreover, its inhibition might affect not only brain CSCs, but also the tumor bulk, as implied by the expression level in glioblastoma. To test this hypothesis, brain CSCs should be treated with small-molecule inhibitors, such as 3PO, that bind the substrate binding site of PFKFBs (Clem et al., 2008). Such studies were initiated but are hampered by the limited specificity of 3PO, as all four PFKFBs share a high homology in their substrate-binding domain. Therefore, further efforts are necessary to develop inhibitors specifically targeting PFKB4.

Silencing of PDK1 and PKM2 resulted in increased apoptosis of brain CSCs. Knockdown of PFKFB4 showed a downregulation of lactate and ATP production in brain CSCs and ultimately induced apoptosis possibly by the phosphorylation/activation of AMPK because of the higher ratio of AMP/ATP. The identification of glycolytic kinases in our screen reflects the importance of the Warburg effect by liberating the cells from oxygen dependence for ATP production. This adaptation to the tumor microenvironment has been shown to confer a survival advantage to cancer cells (Hsu and Sabatini, 2008). Our findings highlight the importance of the lactate and ATP production in the maintenance of brain CSCs in the tumor microenvironment and support a promise for therapies that target proteins involved in the regulation of tumor metabolism.

Materials and methods

Glioma patients

See Supplementary Materials and methods.

Cell culture

The investigated three brain CSC lines, NCH421k, NCH441 and NCH644, were established from primary glioblastoma patients undergoing surgical resection according to the research proposals approved by the institutional review board at the Medical Faculty, University of Heidelberg. The brain CSC lines were characterized genotypically and phenotypically in a previous study (Campos et al., 2010). An adult human neural stem cell line (BLV4) was originally cultivated from intraoperative ventricular tissue after informed consent was given for the scientific use of anonymized samples and after approval by the ethics committee of the Neurosurgery Department/Charité, University Hospital Berlin, Germany (registration number 125/2001). The normal genotype was verified at passage 10 (data not shown).

All cell lines were cultivated at 37 °C in a humidified incubator with 5% CO2 as floating aggregates, the so-called neurospheres on uncoated tissue culture dishes. Brain CSC medium consisted of Dulbecco's modified Eagle's medium/F-12 medium containing 20% BIT serum-free supplement, basic fibroblast growth factor and epidermal growth factor at a concentration of 20 ng/ml each (all from Provitro, Berlin, Germany). Normal NSCs were maintained in Dulbecco's modified Eagle's medium/F-12 medium supplemented with serum-free B27 (Life Technologies, Darmstadt, Germany), 1 M HEPES (Gibco, Darmstadt, Germany), basic fibroblast growth factor and epidermal growth factor (20 ng/ml each).

High-throughput lentivirus production and transduction

The Mission RNAi library was purchased from Sigma-Aldrich (St Louis, MO, USA). High-quality DNA preparations were obtained using a large-scale plasmid purification kit (Qiagen, Hilden, Germany and Roche, Mannheim, Germany). The lentivirus particles were produced in a 96-well format according to the TRC (The RNAi Consortium) protocol. Briefly, 293T packaging cells were co-transfected with the pLKO.1 vector encoding the shRNA, as well as the necessary helper plasmids for virus production (psPAX2 and pMGD2), using Trans-IT (Mirus, Madison, WI, USA). The titer was measured using lentiviral particles that contained the pLKO.1 vector expressing GFP and ranged from 3 × 106 to 5 × 106 IU/ml. Before transduction, cultured spheres were dissociated with trypsin or accutase treatment. For cell death analysis, NCH421k cells were seeded in 96-well plates at 20 000 cells per well in a final volume of 100 μl and transduced at a multiplicity of infection (MOI) of 5 in the presence of 8 μg/ml polybrene. At 24 h after transduction, medium was replaced with fresh brain CSC medium. Cells were analyzed 6 days after viral transduction using PI staining and a flow-cytometer equipped with a high-throughput sampler. Our experimental settings ensured a maximal efficiency, but resulted in some transduction- and readout-associated cytotoxicity and an average of cell survival of 70.15% for all 3838 shRNAs (s.d.±10.25%) and of 26.24% in wells containing scrambled shRNAs (n=768; s.d.±11.38%).

To minimize potential artifacts due to position effects, the rows 1 and 12 were not used. In addition, a plate that contained a scrambled shRNA in each well was included and normalization was performed by matching the cell death rate to the corresponding well from that control plate. The production of single shRNAs was performed in 6-cm petri dishes. After 72 h, produced lentiviruses were concentrated by ultracentrifugation of the HEK293T supernatant. Transductions were performed at the MOI of 5 with 8 μg/ml polybrene for brain CSCs, and at the MOI of 10 without polybrene for normal NSCs.

RNA extraction and qRT–PCR

See Supplementary Materials and methods.

Gene expression analysis

See Supplementary Materials and methods.

FACS and cell death analysis

Flow cytometry analysis was performed for NCH421k cells 6 days after shRNA transduction in 96-well cell culture plates. Cells were transferred onto a 96-well filter plate (Millipore, Bellerica, MA, USA), centrifuged, trypsinized and resuspended in 100 μl staining solution (phosphate-buffered saline, 1.5 μg/ml PI). Samples were analyzed using a flow cytometer equipped with a high-throughput sampler (FACSArray, BD Biosciences, San Jose, CA, USA). Validation of the cell death phenotype upon silencing of PFKFB4 with shRNA-1 (5′-CCTGTGGCATATGGTTGTAAA-3′) and shRNA-2 (5′-GACGTGGTCAAGACCTACAAA-3′) was performed in a six-well plate. Samples were run on a FACSCanto flow cytometer and analyzed with FACS Diva software (Becton Dickinson, San Jose, CA, USA).

Apoptosis and cell viability assays

To study apoptosis we used a caspase assay. Cells were seeded at 20 000 cells per 100 μl per well in 96-well clear plates and transduced with lentiviral shRNA. Caspase-3/7 activity was measured 4, 5, 6 and 7 days after transduction, using a homogeneous luminescent method (Caspase-Glo 3/7 Assay, Promega, Madison, WI, USA). Transduced cells were transferred to white/solid bottom assay plates and 100 μl of Caspase-Glo 3/7 reagent was added. Plates were incubated at room temperature for 1 h, and luminescence intensity determined using the luminometer MITHRAS plate reader (PerkinElmer, Shelton, CT, USA).

To assess brain CSC viability, the MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium)-based CellTiter 96 AQueous Assay (Promega) was used following the manufacturer's recommendations. All samples were assayed in three biological replicas. The 3PO (3-(3-pyridinyl)-1-(4-pyridinyl)-2-propen-1-one) was kindly provided by Dr J Chesney (Louisville, KY, USA). Rapamycin InSolution was provided in dimethyl sulfoxide (Merck, Darmstadt, Germany).

Western blot analysis

Cell lysates were separated in 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. The monoclonal mouse antibody against α-tubulin (Sigma-Aldrich) was used in a dilution of 1:2000, the polyclonal rabbit antibody against PFKFB4 (Sigma-Aldrich) in a dilution of 1:1000 and the monoclonal rabbit and mouse antibodies against phospho-AMPK (Thr172) and total AMPK, respectively (Cell Signaling Technology, Danvers, MA, USA), in a dilution of 1:1000. Horseradish peroxidase-conjugated secondary antibodies (Abcam, Cambridge, UK) were used in a 1:5000 dilution before the chemiluminescent detection of bound antibodies (ECL plus kit; GE Healthcare, Little Chalfont, UK).

Lactate and ATP assays

Brain CSCs were seeded at 2 × 104 cells per 100 μl in 96-well plate and immediately transduced with lentiviruses at an MOI of 5. Lactate levels were measured 3 and 4 days later using a lactate colorimetric assay read at 570 nm according to the manufacturer's instructions (MBL International, Woburn, MA, USA). ATP levels were measured 3 and 4 days after silencing using the ATP luminometric determination kit according to the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA). Concentration was calculated based on a standard curve and normalized to the total number of living cells. The cell number was estimated by the MTS-based CellTiter 96 AQueous Assay (Promega) and calculated based on a standard curve.

Statistical analysis

R 2.12 (http://www.r-project.org) was used for statistical analyses. The significance was calculated using Student's t-test assuming unequal variances. The fold changes in apoptotic rate were compared after rank transformation (Zimmerman and Zumbo, 1993).

For the comparison of the Kaplan–Meier curves, the Mantel–Haenszel test was applied. Using a bootstrapping approach with 1000 draws, the P-value for differences in survival based on the expression of PFKFB4 was confirmed by reselecting the mean expression value and calculating the Mantel–Haenszel P-value. Multivariate analyses were performed calculating Cox hazard models.


  1. Al-Hajj M . (2007). Cancer stem cells and oncology therapeutics. Curr Opin Oncol 19: 61–64.

  2. Atsumi T, Chesney J, Metz C, Leng L, Donnelly S, Makita Z et al. (2002). High expression of inducible 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (iPFK-2; PFKFB3) in human cancers. Cancer Res 62: 5881–5887.

  3. Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB et al. (2006). Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444: 756–760.

  4. Brunet A, Park J, Tran H, Hu LS, Hemmings BA, Greenberg ME . (2001). Protein kinase SGK mediates survival signals by phosphorylating the forkhead transcription factor FKHRL1 (FOXO3a). Mol Cell Biol 21: 952–965.

  5. Campos B, Wan F, Farhadi M, Ernst A, Zeppernick F, Tagscherer KE et al. (2010). Differentiation therapy exerts antitumor effects on stem-like glioma cells. Clin Cancer Res 16: 2715–2728.

  6. Christofk HR, Vander Heiden MG, Harris MH, Ramanathan A, Gerszten RE, Wei R et al. (2008). The M2 splice isoform of pyruvate kinase is important for cancer metabolism and tumour growth. Nature 452: 230–233.

  7. Clem B, Telang S, Clem A, Yalcin A, Meier J, Simmons A et al. (2008). Small-molecule inhibition of 6-phosphofructo-2-kinase activity suppresses glycolytic flux and tumor growth. Mol Cancer Ther 7: 110–120.

  8. Eramo A, Ricci-Vitiani L, Zeuner A, Pallini R, Lotti F, Sette G et al. (2006). Chemotherapy resistance of glioblastoma stem cells. Cell Death Differ 13: 1238–1241.

  9. Ernst A, Hofmann S, Ahmadi R, Becker N, Korshunov A, Engel F et al. (2009). Genomic and expression profiling of glioblastoma stem cell-like spheroid cultures identifies novel tumor-relevant genes associated with survival. Clin Cancer Res 15: 6541–6550.

  10. Fischer C, Mazzone M, Jonckx B, Carmeliet P . (2008). FLT1 and its ligands VEGFB and PlGF: drug targets for anti-angiogenic therapy? Nat Rev Cancer 8: 942–956.

  11. Floridi A, Paggi MG, Fanciulli M . (1989). Modulation of glycolysis in neuroepithelial tumors. J Neurosurg Sci 33: 55–64.

  12. Giroux V, Iovanna J, Dagorn JC . (2006). Probing the human kinome for kinases involved in pancreatic cancer cell survival and gemcitabine resistance. FASEB J 20: 1982–1991.

  13. Gomez-Brouchet A, Pchejetski D, Brizuela L, Garcia V, Altie MF, Maddelein ML et al. (2007). Critical role for sphingosine kinase-1 in regulating survival of neuroblastoma cells exposed to amyloid-beta peptide. Mol Pharmacol 72: 341–349.

  14. Gomez M, Manzano A, Navarro-Sabate A, Duran J, Obach M, Perales JC et al. (2005). Specific expression of pfkfb4 gene in spermatogonia germ cells and analysis of its 5′-flanking region. FEBS Lett 579: 357–362.

  15. Heddleston JM, Li Z, McLendon RE, Hjelmeland AB, Rich JN . (2009). The hypoxic microenvironment maintains glioblastoma stem cells and promotes reprogramming towards a cancer stem cell phenotype. Cell Cycle 8: 3274–3284.

  16. Hsu PP, Sabatini DM . (2008). Cancer cell metabolism: Warburg and beyond. Cell 134: 703–707.

  17. Kasahara Y, Tuder RM, Taraseviciene-Stewart L, Le Cras TD, Abman S, Hirth PK et al. (2000). Inhibition of VEGF receptors causes lung cell apoptosis and emphysema. J Clin Invest 106: 1311–1319.

  18. Keith B, Simon MC . (2007). Hypoxia-inducible factors, stem cells, and cancer. Cell 129: 465–472.

  19. Kessler R, Eschrich K . (2001). Splice isoforms of ubiquitous 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase in human brain. Brain Res Mol Brain Res 87: 190–195.

  20. Kress S, Stein A, Maurer P, Weber B, Reichert J, Buchmann A et al. (1998). Expression of hypoxia-inducible genes in tumor cells. J Cancer Res Clin Oncol 124: 315–320.

  21. Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM et al. (2006). Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serum-cultured cell lines. Cancer Cell 9: 391–403.

  22. Liu G, Yuan X, Zeng Z, Tunici P, Ng H, Abdulkadir IR et al. (2006). Analysis of gene expression and chemoresistance of CD133+ cancer stem cells in glioblastoma. Mol Cancer 5: 67.

  23. Lu H, Forbes RA, Verma A . (2002). Hypoxia-inducible factor 1 activation by aerobic glycolysis implicates the Warburg effect in carcinogenesis. J Biol Chem 277: 23111–23115.

  24. MacKeigan JP, Murphy LO, Blenis J . (2005). Sensitized RNAi screen of human kinases and phosphatases identifies new regulators of apoptosis and chemoresistance. Nat Cell Biol 7: 591–600.

  25. Manzano A, Rosa JL, Ventura F, Perez JX, Nadal M, Estivill X et al. (1998). Molecular cloning, expression, and chromosomal localization of a ubiquitously expressed human 6-phosphofructo-2-kinase/ fructose-2, 6-bisphosphatase gene (PFKFB3). Cytogenet Cell Genet 83: 214–217.

  26. Michelakis ED, Sutendra G, Dromparis P, Webster L, Haromy A, Niven E et al. (2010). Metabolic modulation of glioblastoma with dichloroacetate. Sci Transl Med 2: 31ra34.

  27. Minchenko O, Opentanova I, Minchenko D, Ogura T, Esumi H . (2004). Hypoxia induces transcription of 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase-4 gene via hypoxia-inducible factor-1alpha activation. FEBS Lett 576: 14–20.

  28. Okuyama H, Krishnamachary B, Zhou YF, Nagasawa H, Bosch-Marce M, Semenza GL . (2006). Expression of vascular endothelial growth factor receptor 1 in bone marrow-derived mesenchymal cells is dependent on hypoxia-inducible factor 1. J Biol Chem 281: 15554–15563.

  29. Pardal R, Clarke MF, Morrison SJ . (2003). Applying the principles of stem-cell biology to cancer. Nat Rev Cancer 3: 895–902.

  30. Park CY, Tseng D, Weissman IL . (2009). Cancer stem cell-directed therapies: recent data from the laboratory and clinic. Mol Ther 17: 219–230.

  31. Pradelli LA, Beneteau M, Chauvin C, Jacquin MA, Marchetti S, Munoz-Pinedo C et al. (2010). Glycolysis inhibition sensitizes tumor cells to death receptors-induced apoptosis by AMP kinase activation leading to Mcl-1 block in translation. Oncogene 29: 1641–1652.

  32. Reya T, Morrison SJ, Clarke MF, Weissman IL . (2001). Stem cells, cancer, and cancer stem cells. Nature 414: 105–111.

  33. Sakata J, Abe Y, Uyeda K . (1991). Molecular cloning of the DNA and expression and characterization of rat testes fructose-6-phosphate,2-kinase:fructose-2,6-bisphosphatase. J Biol Chem 266: 15764–15770.

  34. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J et al. (2003). Identification of a cancer stem cell in human brain tumors. Cancer Res 63: 5821–5828.

  35. Soeda A, Park M, Lee D, Mintz A, Androutsellis-Theotokis A, McKay RD et al. (2009). Hypoxia promotes expansion of the CD133-positive glioma stem cells through activation of HIF-1alpha. Oncogene 28: 3949–3959.

  36. Stupp R, Hegi ME, Gilbert MR, Chakravarti A . (2007). Chemoradiotherapy in malignant glioma: standard of care and future directions. J Clin Oncol 25: 4127–4136.

  37. Toedt G, Barbus S, Wolter M, Felsberg J, Tews B, Blond F et al. (2011). Molecular signatures classify astrocytic gliomas by IDH1 mutation status. Int J Cancer 128: 1095–1103.

  38. Vescovi AL, Galli R, Reynolds BA . (2006). Brain tumour stem cells. Nat Rev Cancer 6: 425–436.

  39. Wigfield SM, Winter SC, Giatromanolaki A, Taylor J, Koukourakis ML, Harris AL . (2008). PDK-1 regulates lactate production in hypoxia and is associated with poor prognosis in head and neck squamous cancer. Br J Cancer 98: 1975–1984.

  40. Wurdak H, Zhu S, Romero A, Lorger M, Watson J, Chiang CY et al. (2010). An RNAi screen identifies TRRAP as a regulator of brain tumor-initiating cell differentiation. Cell Stem Cell 6: 37–47.

  41. Yalcin A, Telang S, Clem B, Chesney J . (2009). Regulation of glucose metabolism by 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatases in cancer. Exp Mol Pathol 86: 174–179.

  42. Zimmerman DW, Zumbo BD . (1993). Rank transformations and power of the Student t test and Welch t test for nonnormal populations with unequal variances. Can J Exp Psychol 47: 523–539.

Download references


We thank Dr S Pfister and D Haag for their helpful comments. We thank Mrs V Lang and G Iren for their technical help. We thank Dr J Chesney for kindly providing 3PO. This work was supported by the German Federal Ministry of Education and Research (BMBF) within the National Genome Research Network (NGFNplus; 01GS0883, 01GS0884), a stipend from the Alexander von Humboldt foundation (to SN) and by the grant ‘Young Investigator Fellowship’ of the Medical Faculty of Heidelberg (to VG).

Author information

Correspondence to V Goidts.

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

Cite this article

Goidts, V., Bageritz, J., Puccio, L. et al. RNAi screening in glioma stem-like cells identifies PFKFB4 as a key molecule important for cancer cell survival. Oncogene 31, 3235–3243 (2012). https://doi.org/10.1038/onc.2011.490

Download citation


  • loss-of-function screen
  • apoptosis
  • glioblastoma
  • cancer stem-like cells
  • glycolysis

Further reading