Mutant IDH1 inhibits PI3K/Akt signaling in human glioma


  • Peter Birner MD,

    1. Department of Neuropathology, Institute of Pathology, Ruprecht-Karls-University Heidelberg, Heidelberg, Germany
    2. Clinical Institute of Pathology, Medical University of Vienna, Vienna, Austria
    Search for more papers by this author
  • Stefan Pusch PhD,

    1. Department of Neuropathology, Institute of Pathology, Ruprecht-Karls-University Heidelberg, Heidelberg, Germany
    2. Clinical Cooperation Unit Neuropathology, German Cancer Research Center (DKFZ), Heidelberg, Germany
    Search for more papers by this author
  • Christo Christov MD,

    1. Department of Neurosurgery, St. Ivan Rilski Hospital, Sofia, Bulgaria
    Search for more papers by this author
  • Stiliana Mihaylova MD,

    1. Department of Neurosurgery, St. Ivan Rilski Hospital, Sofia, Bulgaria
    Search for more papers by this author
  • Kalina Toumangelova-Uzeir MS,

    1. Department of Neuropathology, St. Ivan Rilski Hospital, Sofia, Bulgaria
    Search for more papers by this author
  • Sevdalin Natchev MD,

    1. Department of Neuropathology, St. Ivan Rilski Hospital, Sofia, Bulgaria
    Search for more papers by this author
  • Sebastian F. Schoppmann MD,

    1. Department of Surgery, Medical University of Vienna, Vienna, Austria
    Search for more papers by this author
  • Andrey Tchorbanov PhD,

    1. Department of Immunology, Institute of Microbiology, Bulgarian Academy of Sciences, Sofia, Bulgaria
    Search for more papers by this author
  • Berthold Streubel MD,

    Corresponding author
    1. Department of Obstetrics and Gynecology, Medical University of Vienna, Vienna, Austria
    • Corresponding authors : Marin Guentchev, Department of Neurosurgery, Klinikum Idar-Oberstein, Dr. Ottmar-Kohler Str. 2, 55743 Idar-Oberstein, Germany; Fax: (011) +49/6781/661562;; Berthold Streubel, Department of Obstetrics and Gynecology, Medical University of Vienna, Währinger Gürtel 18-20, 1090 Vienna, Austria; Fax: (011) 43 1 404 00-2820;

    Search for more papers by this author
  • Jochen Tuettenberg MD,

    1. Department of Neurosurgery, Klinikum Idar-Oberstein, Idar-Oberstein, Germany
    Search for more papers by this author
    • Drs. Guentchev and Tuettenberg share senior authorship.

  • Marin Guentchev MD

    Corresponding author
    1. Department of Neurosurgery, Klinikum Idar-Oberstein, Idar-Oberstein, Germany
    • Corresponding authors : Marin Guentchev, Department of Neurosurgery, Klinikum Idar-Oberstein, Dr. Ottmar-Kohler Str. 2, 55743 Idar-Oberstein, Germany; Fax: (011) +49/6781/661562;; Berthold Streubel, Department of Obstetrics and Gynecology, Medical University of Vienna, Währinger Gürtel 18-20, 1090 Vienna, Austria; Fax: (011) 43 1 404 00-2820;

    Search for more papers by this author
    • Drs. Guentchev and Tuettenberg share senior authorship.

  • We are grateful to Andreas von Deimling, Pieter Wesseling, and Herbert Budka for critical discussions and support. We thank Gerda Ricken and Milena Moneva for excellent technical assistance. The kind support of David Reuss is highly appreciated.



Recently, isocitrate dehydrogenase 1 (IDH1) was identified as a major participant in glioma pathogenesis. At present, the enzymatic activity of the protein has been the main topic in investigating its physiological function, but its signaling pathway allocation was unsuccessful. Interestingly, proteins regulated by phosphoinositide 3-kinase (PI3K)/Akt signaling, are among the top downregulated genes in gliomas associated with high percentage of IDH1 and IDH2 mutations. The aim of this study was to investigate a hypothetical relation between IDH1 and PI3K signaling.


The presence of mutant IDH1 and markers for active PI3K/Akt signaling, present as phosphorylated Akt and podoplanin (PDPN), were investigated in a discovery cohort of 354 patients with glioma. In vitro experiments were used to confirm functional links.


This study shows an inverse correlation between mutant IDH1 and markers for active PI3K/Akt signaling. In support of a functional link between these molecules, in vitro expression of mutant IDH1 inhibited Akt phosphorylation in a 2-hydroxyglutarate–dependent manner.


This study provides patient tumor and in vitro evidence suggesting that mutant IDH1 inhibits PI3K/Akt signaling. Cancer 2014;120:2440–2447. © 2014 American Cancer Society.


Recent reports have identified a subgroup of gliomas with better prognosis characterized by mutations in the genes encoding isocitrate dehydrogenase 1 (IDH1) and IDH2.[1] Despite intense research, it is still unclear how IDH1 and IDH2 influence prognosis in patients with cancer. Recent reports show that the mutated form decreases proliferation of glioma cell lines[2] and blocks differentiation of nontransformed cells.[3]

PI3K/Akt signaling is aberrantly activated in many cancer types, including human glioma.[4] In most cases, loss of the PI3K inhibitor PTEN (phosphatase and tensin homolog) or a gain-of-function mutations of PI3KCA[5] are the underlying cause. Activation of PI3K/Akt signaling is associated with a more aggressive behavior in human glial tumors[6] and causes faster clinical and histological tumor progression in a platelet-derived growth factor (PDGF)-driven mouse glioma model.[7] Intriguingly, 2 proteins linked to PI3K signaling, podoplanin (PDPN) and retinol binding protein 1 (RBP1), are some of the top downregulated genes in a subset of gliomas associated with high frequency of IDH1 mutations.[8] Interestingly AKT1 was downregulated on the RNA level, and immunocytochemistry showed a reduced staining intensity for phospho-Akt in IDH1R132H-expressing glioma cells in vitro.[2]

There is a growing body of evidence suggesting that PDPN, a downstream effector of PI3K signaling,[9] plays a major role in glioma pathogenesis. PDPN contributes to cancer progression by inducing collective cell migration and tumor invasion,[10] and its expression correlates with tumor grade in human glioma[11] and serves as a marker of reduced overall survival in glioblastoma patients.[12-14] PDPN knockdown impairs[15] and upregulation induces[16] migration of glioma cells, suggesting a direct involvement of PDPN in disease progression.[15]

Here, we show that in human glioma mutant IDH1 inversely correlates with PI3K/Akt pathway activation, and its in vitro expression blocks Akt phosphorylation suggesting that mutant IDH1 inhibits the transition to a more aggressive subtype of glioma by blocking PI3K/Akt signaling.



Our discovery cohort included 354 patients with diffuse gliomas who underwent surgery at St. Ivan Rilski University Hospital (Sofia, Bulgaria) between 2004 and 2011 (Table 1). Our discovery cohort was composed by 72 cases diagnosed as WHO grade II (33 astrocytomas, 26 oligodendrogliomas, and 13 oligoastrocytomas), 32 cases diagnosed as WHO grade III (12 anaplastic astrocytomas, 8 anaplastic oligodendrogliomas, and 12 anaplastic oligoastrocytomas), and 250 cases diagnosed as WHO grade IV (249 glioblastomas and 1 gliosarcoma). Postoperative radiotherapy was the standard adjuvant therapy in Bulgaria at that given period for grade III and IV gliomas. After 2006 temozolamide treatment was added to grade IV glioma. Only cases with frontal, parietal, temporal or occipital localizations of the tumor were used. No biopsy cases were studied, but only full neurosurgical specimens.

Table 1. Patients Cohort
 No. of CasesMean Age (y)Mean Follow-Up Time (d)EventsSurgery for Recurrence
  1. Abbreviations: SD, standard deviation; SE, standard error.

All cases35452 ± 15 (SD)928 ± 35 (SE)23039
WHO grading
II7240 ± 14799 ± 444144
III3245 ± 16657 ± 558193
IV25056 ± 13338 ± 30219732

Tissue Microarray Construction and Immunohistochemistry

Each case has been reviewed by an experienced neuropathologist (in most cases, S.N.). Based on hematoxylin and eosin staining, representative areas have been selected and were used for tissue core sampling. Tissue cores of 4-mm diameter from each case were placed into a recipient block. Immunostaining for PTEN, pAkt, PDPN and mutated IDH1 (R132H) was investigated in 4-μm-thick histological slides from tissue microarray (TMA) blocks using a Benchmark Ultra immunostainer (Ventana, Tucson, Ariz). The antibodies used are listed in Table 2.

Table 2. Antibodies Used and Evaluation Procedure
AntigencloneDilutionVendorPositive ControlSemiquantitative Analysis
PTEND4.31:100Cell Signaling TechnologyNormal gastric tissueLoss of PTEN was rated if = 80% of glioma cells showed no expression of PTEN
phospho-Akt (Ser473)D9E1:100Cell Signaling TechnologyBreast cancerPositive is = 10% of glioma cells showed distinct cytoplasmic staining
PDPND2-40Ready to useCell MarqueCervical cancerPositive if = 50% of glioma cells showed distinct cytoplasmic or membranous staining
IDH1 (R132H)H091:3Hybridoma supernatantGlioma with known mutIDH1Positive if = 80% of glioma cells showed distinct staining

Rembrandt Database mRNA Expression Analysis

As validation cohort, we performed a analysis of gene expression microarray data of human glioma samples from the National Cancer Institute Rembrandt public data repository.[17] On March 3, 2014, we downloaded clinical data from 342 glioma samples: 106 astrocytomas (50 grade II, 45 grade III, and 11 grade unknown), 49 oligodendrogliomas (18 grade II, 21 grade III, and 10 grade unknown), 7 oligoastrozytomas (2 grade II, 3 grade III, and 2 grade unknown), and 180 glioblastomas. Single-gene survival analysis of human glioma was performed on cases separated based on high or low mRNA copy numbers (highest geometric mean intensity value). Clinical data were derived from contributing center institutional diagnoses at Henry Ford Hospital, UCSF, Lee Moffitt Cancer Center, Dana-Farber Cancer Center, University of Wisconsin, and National Cancer Institute. PDPN high-expressing samples were considered the ones with more than 5-fold upregulation of PDPN mRNA (n=188). These were subsequently compared to all other samples (n=154). PTEN low-expressing samples were the ones with more than 2-fold PTEN mRNA downregulation (n=45), subsequently compared to all other samples (n=297). All data were transferred to Excel (Microsoft, Redmond, Wash) and subsequently to SPSS version 20.0 (IBM, Armonk, NY).

Next-Generation DNA Sequencing

DNA was extracted from FFPE using QIAamp DNA FFPE Tissue Kit (Qiagen, Hilen, Germany), DNA quality was evaluated using the Illumina FFPE QC Kit (Illumina, San Diego, CA). Mutation detection within PI3K/Akt pathway-related genes was performed using the TruSeq Amplicon - Cancer Panel on the MiSeq System (Illumina, San Diego, Calif).

Cell Culture

To generate an inducible cell line, we used the pT-REx-DEST system (Invitrogen, Grand Island, NY). As first step we transfected the cell line LN319 (provided by Dr. Capper, Tübingen, Germany) with pcDNA6/TR. From this transfection, we generated single-cell clones and tested their reliability, by introducing enhanced green fluorescent protein (EGFP) in pT-REx-DEST30. We have chosen the clone with no EGFP expression in tetracycline-free media and highest expression after induction with 1 μM doxycycline (Sigma Aldrich, St. Louis, MO, #D3072). This clone 09 (K09) was then used for all further experiments and has not been authenticated in the past 6 months.

For the experimental setup, we used LN319 K09 with wtIDH1 and R132H mutIDH1 in pT-REx-DEST30. All cell lines were seeded with 100,000 cells per 10-cm dish into 2 dishes each. One of these dishes was induced with 1 μM doxycycline, and the other one was treated with the comparable amount of solvent (dimethylsulfoxide; Sigma Aldrich, St. Louis, MO). MutIDH1 expression was controlled via western blot (Supporting Fig. 1; see online supporting information).

We further used wtIDH1-expressing cell line LN229 and 3 clones of LN229-transfected with mutIDH1 producing increasing amounts of 2-hydroxyglutarate (D2HG): H3 (∼8μM in the supernatant), H114 (∼20 μM), and H77 (∼60 μM).[18]


RNA from LN319 K09 cells with wtIDH1 and mutIDH1 was isolated with TRIzol, and cDNA was synthesized using SuperScript II (Invitrogen, Grand Island, NY). For PDPN cDNA detection, the ABI TaqMan Gene Expression Assay Hs00366766_m1 (PDPN) was used. For normalization, we used Hs99999905_m1 (GAPDH) and HS99999903_m1 (ACTB) respectively. Relative expression was calculated using the comparative cycle threshold (ΔΔCt) method.

Western Blot

Cells were harvested after 7 days of incubation. They were washed with phosphate-buffered saline (PBS) twice and scratched off the plate with a cell scratcher (Cell Lifter; Corning Inc., Corning, NY) in 1 ml PBS. The cells were then centrifuged at 5000 rpm, and supernatant was discarded. The pellet was resuspended in cell lysis buffer (NP-40 lysis buffer; 150 mM NaCl, 0,1% NP-40, 50 mM Tris-HCl pH 8.0, Complete Protease Inhibitor). The cells were then frozen in liquid nitrogen and thawed 3 times. Subsequently, the probe was centrifuged at 14,800 rpm for 15 minutes, and the supernatant was transferred into a new tube for further processing.

Aliquots (about 20 μg of protein) were boiled for 10 minutes in SDS-PAGE sample buffer with β-mercaptoethanol, run on 10% SDS-PAGE gels and transferred to a nitrocellulose membrane (0.45μm, Sartorius, Germany). The membranes were blocked overnight at 4°C in TBS pH 7.4 with 0.05% Tween 20 and 5% BSA followed by incubation with following antibodies: anti-phospho-Akt (Ser 473) (clone D9E; Cell Signaling Technology, Danvers, Mass), anti-PDPN (clone NZ 1.2, Sigma Aldrich, St. Louis, MO), anti-mutIDH1[19] and anti–β-actin (Cell Signaling Technology). The blots were probed according to manufacturer recommendations, and the signal for the immunoreactive proteins was developed with peroxidase-conjugated secondary antibodies (Cell Signaling Technology) and visualized by sodium nitroprusside and o-dianisidin-dihydrochloride-3,3′-dimethoxybenzidine (Sigma Aldrich, St. Louis, MO). The mutant IDH1 western blot and the one on lysates from H3, H114, and H77 cells was developed by chemiluminescence at a different laboratory (by S.P.). Blots were photographed and arranged in MS PowerPoint (Microsoft, Redmond, Wash). Band quantifications were done using the ImageJ 1.47 software (NIH Image).


Mann-Whitney U test and Kruskal-Wallis test were used as appropriate. Overall survival was defined as the period from primary surgery until death, regardless of cause. Survival until the end of the observation period was considered a censored observation. Multivariate analysis of survival was performed with the Cox proportional hazards model. A 2-tailed P value of ≤ .05 was considered as significant. Univariate analysis of overall survival was performed using log-rank analysis or univariate Cox regression in case of patients' age. Due to the exploratory design of the study, no correction for multiple testing was performed.[20] SPSS version 20.0 (IBM, Armonk, NY) was used for all calculations.


PI3K Signaling Correlates With PDPN Expression in Human Glioma

To find evidence in human glioma supporting previous in vitro data[9] that PDPN is a downstream target of PI3K signaling, we performed immunostaining for PDPN, PTEN, and pAkt in 354 glioma cases (Fig. 1).

Figure 1.

Immunostaining is shown for PTEN, pAkt, PDPN, and mutIDH1 in representative glioma cases: (A) case with mutIDH, (B) PDPN-positive, (C) PTEN-positive, (D) PTEN-negative, and (E) pAkt-positive. Scale bars represent ∼100 μm.

In the univariate analysis, PDPN (P < .001, log-rank test, 1-year survival rate 33.1% ± 4.2% SE versus 65.5% ± 3.2% SE), PTEN loss (P < .001, log-rank test, 1-year survival rate 36.9% ± 5.7% SE versus 60.1% ± 3.2% SE), and pAkt (P < .001, log-rank test, 1-year survival rate 46.3% ± 4.6% SE versus 60.5% ± 3.8% SE) were associated with shorter survival in all gliomas (Fig. 2, Supporting Table 1). A multivariate analysis (Cox regression) of survival of all cases including WHO grade, age, and PDPN, PTEN and pAkt expression showed that only WHO grade (P < .001, hazard ratio = 2.223, 95% CI = 1.642-3.01), age (P = .036, hazard ratio 1.015, 95% CI = 1.001-1.029) and PDPN expression (P = .001, hazard ratio = 1.736, 95% CI = 1.269-2.375) remained independent prognostic factors (Supporting Table 1). The proportion of PTEN-negative, pAkt- and PDPN-positive cases increased with tumor grade (Table 3). When analyzing the Rembrandt dataset, high levels of PDPN mRNA correlated with low levels of PTEN mRNA (P < .001, chi-square test). High levels of PDPN mRNA (P < .001, log-rank test) and low levels of PTEN (P < .001, log rank test) were associated with worse prognosis (Supporting Fig. 2).

Table 3. Expression of Proteins in Relation to WHO Grading
 mutIDH1 (n = 348)PDPN (n = 354)PTEN (n = 321)pAkt (n = 295)
All cases83 (23.9%)129 (36.4%)243 (75.7%)123 (41.7%)
WHO grade    
II51 (70.8%)3 (4.2%)61 (88.4%)11 (17.5%)
III13 (41.9%)6 (18.8%)24 (82.8%)6 (30.0%)
IV19 (7.8%)120 (48%)158 (70.9%)106 (50.0%)
Figure 2.

Kaplan-Meier curves representing the survival impact of pAkt, PTEN, and PDPN are shown in all cases and in high-grade gliomas only. The x-axis represents time since surgery (days), and y-axis represents cumulative overall survival.

Statistically PDPN expression was associated with loss of PTEN and pAkt (P < .001, respectively, chi-square test). In 26 patients, however, PDPN was expressed in cases where no PTEN loss or pAkt were detected. To test whether in these cases PDPN is upregulated by an alternative signaling pathway or there are mutations causing aberrant PI3K/Akt signaling, we subjected samples of 9 PDPN+/PTEN+ cases and 1 PDPN+/PTEN− control case to Next Generation Sequencing using a 50-gene cancer panel. In 7 of 9 PDPN+/PTEN+ cases, mutations of the PI3KCA and/or PTEN genes were found. In the control case, no mutations of PI3KCA, Akt1, or PTEN were detected.

These human data support previous in vitro results[9] that PDPN is a downstream effector of PI3K signaling in human glioma.

Mutant IDH1 Inversely Correlates With Active PI3K Signaling in Human Glioma

IDH1 has been identified as a major determinant of prognosis in human glioma.[1] Because PDPN, a protein regulated by PI3K/Akt signaling,[9] is one of the top downregulated genes in cases associated with high percentage of IDH1 and 2 mutations[8] we hypothesized that IDH1 is involved in PI3K/Akt and PDPN regulation.

As a first step, we studied how the most common IDH1 mutation overlaps with markers for active PI3K signaling as Akt phosphorylation and PDPN expression. We immunohistochemically determined the presence of R132H mutated IDH1 (mutIDH1), representing more than 90% of all IDH1 mutations[21] in our series and correlated it with pAkt and PDPN. Consistent with previous reports,[1] mutIDH1 was associated with a better prognosis (P < .001, log-rank test) in all cases (Supporting Fig. 3). MutIDH1 was inversely associated with the presence of pAkt (P < .001, chi-square test) and expression of PDPN (P < .001, chi-square test). Interestingly, only 2 of 83 mutIDH1 cases (2.4%) were positive for PDPN and 12 of 83 (17.6%) were positive for pAkt.

Because activation of PI3K signaling is strongly associated with high-grade gliomas, ie, WHO grade II gliomas usually lack marked activation of PI3K/Akt signaling. To exclude the possibility that malignancy grade is a major confounder of the reported inverse association between immunopositivity for pAkt and mutIDH1, we investigated the proportion of pAkt+ cases among the mutIDH1-positive and mutIDH1-negative glioblastoma population. In the mutIDH1+ glioblastoma cases, only 21.4% (3 of 14) were pAkt positive. In the mutIDH1-negative population, the proportion of pAkt-positive cases was 52.1% (101 of 194) (P = .027, chi-square test).

The fact that Akt phosphorylation or PDPN expression rarely takes place in the presence of mutIDH1 is an in vivo result in humans that is compatible with the hypothesis that mutant IDH1 might inhibit PI3K/Akt signaling.

Mutant IDH1 Inhibits PI3K Signaling In Vitro

To test our hypothesis in vitro, we expressed wild-type and mutant IDH1 R132H in LN-319 glioma cells via a Tet-inducible system. Western blot analysis showed that the expression of mutant IDH1 inhibited PI3K signaling, but was unable to downregulate already expressed PDPN in glioma cells (Fig. 3A, Supporting Fig. 4). A quantitative RT-PCR (qRT-PCR) for PDPN did not show any significant change of mRNA levels within these groups as well (Supporting Fig. 5).

Figure 3.

(A) Mutant IDH expression blocks PI3K/Akt signaling. In LN-319 glioblastoma cells, wild-type or mutant IDH1 was expressed using a TeT-inducible system. PDPN and pAkt levels were determined by western blot. Lane: 1) wild-type IDH1 with doxycycline, 2) wild-type IDH1 with no doxycycline, 3) mutant IDH1 with doxycycline, 4) mutant IDH1 with no doxycycline. (B) Mutant IDH expression blocks Akt phosphorylation in a D2HG-dependent manner. Western blot analysis of Akt phosphorylation in wild-type LN229 and mutIDH1-transfected LN229 lines H3, H114, and H77 producing increasing amounts of D2HG (as published previously). Lane: 1) wild-type LN229, 2) H3, 3) H114, and 4) H77.

The fact that Akt phosphorylation is inhibited by mutIDH1 expression suggests a functional link between IDH1 and PI3K/Akt signaling.

D2HG-Dependent Akt Inhibition

To further test our hypothesis in vitro, we analyzed the IDH wild-type glioma cell line LN229 and a set of mutIDH1-transfected LN229 cells with stable, but different expression levels of mutant IDH1 R132H. Wild-type LN229 without IDH mutation exhibited basal levels of D2HG. The mutIDH1 transfected LN229 lines H3, H77, H114 exhibited elevated levels of D2HG.[18] Western blot analysis showed that the expression of mutant IDH1 inhibited Akt phosphorylation in a D2HG level-dependent manner (Fig. 3B, Supporting Fig. 6).

The fact that the level of Akt inhibition correlated well with amount of D2HG produced suggests that D2HG is the PI3K/Akt-inhibiting agent.


Despite its obvious importance in glioma pathogenesis[1, 22] and the extensive research conducted, the disease relevant function of IDH1 is still enigmatic. Here, we provide evidence that mutated IDH1 blocks PI3K/Akt signaling, a pathway associated with the development of a more aggressive glioma phenotype.[7, 11, 23]

Point mutations in IDH1 and IDH2 occur in a significant portion of gliomas,[22] acute myeloid leukemia,[24] chondrosarcoma,[25] and cholangiocarcinoma.[26] The first proposed function for IDH mutants is a dominant-negative effect on the wild-type enzyme by inactivating the enzymatic activity of IDH1 and IDH2 to convert isocitrate to α-ketoglutarate.[27] Mutated IDH1 can build a heterodimer with the wild-type protein resulting in a decreased isocitrate dehydrogenase activity compared to wild-type homodimer in vitro.[28]

On the cell level, blocking mutant IDH1 was shown to delay growth and promote differentiation of glioma cells.[29] Our results add a further clue about the role of IDH1, namely an inhibiting effect on PI3K signaling that fits well with the concept that mutant IDH1 regulates key gene expression programs, especially those characterizing G-CIMP–positive gliomas.[30] Since Akt activation is sufficient to promote glioma progression,[31] we hypothesize that mutant IDH1, via blocking PI3K/Akt signaling, slows glioma malignization.

Recent reports[9] and our findings show that PDPN is associated with and regulated by PI3K signaling. Despite the link demonstrated here between mutIDH1 and Akt phosphorylation, we were not able to show that mutIDH1 inhibits PDPN expression. Several scenarios might be able to explain this phenomenon. We believe that the most likely one is that PI3K signaling is involved in upregulation, but not in perpetuation of PDPN's expression.

Our study shows that the proportion of IDH1 mutations in grade III tumors is lower than previously described.[1, 21] A main factor accounting for that discrepancy is most likely the lack of normal distribution due to the lower number of grade III cases used in our study. Furthermore, it is possible that in our series some primary glioblastomas were misdiagnosed as grade III tumors.

In summary, we show direct and indirect evidence that mutant IDH1, a molecule with an important role in glioma pathogenesis, inhibits PI3K/Akt signaling, thus slowing glioma malignization.


No specific funding was disclosed.


Dr. Tuettenberg receives royalties from Roche. All other authors made no disclosure.