SEARCH

SEARCH BY CITATION

Keywords:

  • Glioma;
  • Akt;
  • Targeted therapy;
  • Small molecule inhibitor;
  • Cancer stem cell

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Malignant brain tumors are among the most lethal cancers, and conventional therapies are largely limited to palliation. Novel therapies targeted against specific molecular pathways may offer superior efficacy and less toxicity than conventional therapies, but initial clinical trials of molecular targeted agents in brain cancer therapy have been frequently disappointing. In brain tumors and other cancers, subpopulations of tumor cells have recently been characterized by their ability to self-renew and initiate tumors. Although these cancer stem cells, or tumor initiating cells, are often only present in small numbers in human tumors, mounting evidence suggests that cancer stem cells contribute to tumor maintenance and therapeutic resistance. Thus, the development of therapies that target cancer stem cell signal transduction and biology may improve brain tumor patient survival. We now demonstrate that populations enriched for cancer stem cells are preferentially sensitive to an inhibitor of Akt, a prominent cell survival and invasion signaling node. Treatment with an Akt inhibitor more potently reduced the numbers of viable brain cancer stem cells relative to matched nonstem cancer cells associated with a preferential induction of apoptosis and a suppression of neurosphere formation. Akt inhibition also reduced the motility and invasiveness of all tumor cells but had a greater impact on cancer stem cell behaviors. Furthermore, inhibition of Akt activity in cancer stem cells increased the survival of immunocompromised mice bearing human glioma xenografts in vivo. Together, these results suggest that Akt inhibitors may function as effective anticancer stem cell therapies.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Author contributions: C.E.E.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; W.-C.F.: conception and design, collection and/or assembly of data, data analysis and interpretation; C.C.E. and W.-C.F. contributed equally to the manuscript. K.M.L.: collection and/or assembly of data; R.E.M.: provision of study material or patients, data analysis and interpretation; A.B.H.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; J.N.R.: conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.

Glioblastomas are the most common and deadly primary tumors of the central nervous system in adults [1]. Current therapy consists of maximal surgical resection followed by cytotoxic therapies that nonspecifically damage DNA or inhibit mitosis [2]. Despite the significant toxicities of these therapies, a large fraction of brain cancer patients suffer tumor recurrence due to the resistance of tumors to therapy [2]. The rapidity of tumor regrowth even after evidence of tumor response suggests that some tumor cells are resistant to therapy at treatment initiation. We and other groups have demonstrated that a subset of tumor cells (called cancer stem cells or tumor-initiating cells) are functionally defined by their abilities to undergo sustained self-renewal and form tumors recapitulating the phenotypes of the parental tumors [3, 4, 5, 6, 78] and are relatively resistant to radiotherapy and chemotherapy [3, 9, 10, 11, 1213]. Although the degree to which cancer stem cells contribute to therapeutic resistance remains to be defined, the identification of a tumor cell subpopulation with an enriched capacity to promote tumor growth [3, 4, 5, 6, 78], angiogenesis [4], and metastasis [12] may better inform cancer biology. Clarke and colleagues have hypothesized that current therapies might target the bulk of the tumor (the majority of which is the nonstem cancer cell compartment), but unless the cancer stem cells themselves are targeted, they may remain viable and repopulate the tumor following treatment [14, 15]. Thus, it is paramount to devise therapies that target cancer stem cells if the tumor is to be eliminated entirely.

Glioblastomas commonly display hyperactivation of the phosphatidylinositol-3-kinase (PI3K)-Akt (also known as protein kinase B) pathway, a protumorigenic signaling cascade that contributes to the pathogenesis of several human cancers [16, 17, 18, 19, 2021]. The PI3K-Akt pathway can be activated in tumors through a number of mechanisms, including activation of upstream growth factor receptors, mutations of the PI3K catalytic subunit, overexpression or amplification of Akt family members, or inactivation of the inhibitory effects of the phosphatase and tensin homolog (PTEN) tumor suppressor [16, 17, 18, 1920]. Hyperactive Akt signaling promotes tumorigenic cell behaviors by increasing cell survival, proliferation, invasion, and angiogenesis, and it has been directly associated with in vitro conversion of grade III anaplastic astrocytoma to grade IV glioblastoma [16, 17, 18, 19, 2021].

Because of the association of Akt activity with a wide range of tumorigenic properties, we hypothesized that brain cancer stem cells may exhibit a dependence on the Akt pathway. Indeed, chemoresistance in hepatocarcinoma stem cells may be conferred by activation of Akt [11], and Akt regulates the survival of tumor cells in the perivascular niche bearing stem cell markers in mouse medulloblastoma models [21]. To further investigate the dependence of brain cancer stem cells on Akt signaling, we pharmacologically treated matched populations of glioblastoma cancer stem cells and nonstem cells with a small molecule inhibitor of Akt. We sought to determine if preferential targeting of brain tumor stem cells could be achieved through inhibition of Akt by decreasing the capacity of these cells to survive, proliferate, and invade, thereby decreasing their malignant potential.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Isolation of CD133+ and CD133 Tumor Cells

T3359 cultures were isolated from primary glioblastoma samples transiently amplified in immunocompromised mice. Tumor specimens were obtained from surgical biopsies of consenting patients under a protocol approved by the Duke University Medical Center Institutional Review Board. D456MG xenografts were originally derived from a pediatric glioblastoma biopsy specimen and have been maintained in immunocompromised mice under a Duke Institutional Animal Care and Use approved protocol. Of note, T3359 and D456MG express wild-type PTEN. Tumors were dissociated into single cells using an enzyme dissociation kit (Worthington Biochemical, Lakewood, NJ, http://www.worthington-biochem.com). For fluorescence-activated cell sorting (FACS) into CD133+ and CD133 enriched populations, cells were labeled with an allophycocyanin-conjugated CD133 antibody (Miltenyi Biotec, Auburn, CA, http://www.miltenyibiotec.com) before sorting by FACS. For magnetic bead sorting into CD133+ and CD133 enriched cell populations, cells were incubated with CD133 antibodies conjugated with biotin and magnetic beads that bind biotin prior to separation by a magnetic column (Miltenyi Biotec). CD133+ cells were maintained in their undifferentiated state using neurobasal medium supplemented with epidermal growth factor (EGF) and fibroblastic growth factor (FGF) (each at 10 μg/500 ml medium), sodium pyruvate, glutamine, B27, nonessential amino acids, and penicillin/streptomycin (Gibco, Grand Island, NY, http://www.invitrogen.com). CD133 cells were maintained in their differentiated state with Dulbecco's Modified Eagle's Medium (Gibco) supplemented with 10% fetal bovine serum (FBS) (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) and penicillin/streptomycin.

Small Molecule Inhibitor

The small molecule inhibitors of Akt (AktIII/SH-6, AktII/Sh-5), PI3K (LY294002), and mammalian target of rapamycin (mTOR) (rapamycin) were purchased from Calbiochem (San Diego, http://www.emdbiosciences.com). For all assays, stock solutions created by dissolving the drug in dimethyl sulfoxide (DMSO) (Sigma-Aldrich) were stored at −80°C. Immediately prior to the experiment, stock solutions were diluted in DMSO to 1,000× the final concentrations indicated. For each experiment, 1 μl/ml of DMSO as a control or inhibitor 1,000× stock solutions in DMSO were added to the medium of cells to make the indicated final concentrations of inhibitor.

Antibodies and Western Blotting

CD133+ and CD133 cells were plated in appropriate media in six-well plates at 5 × 105 cells per well and allowed to recover overnight. CD133 medium was changed to CD133+ growth medium before each experiment for the indicated times. All cells were harvested together and then lysed in buffer (62.5 mM Tris-HCl, 2% w/v SDS, 10% glycerol, 40 mM dithiothreitol, and protease inhibitors). Protein concentrations were quantified (Bio-Rad Protein Assay Reagent, Bio-Rad, Hercules, CA, http://www.bio-rad.com), and equal amounts of protein were run on polyacrylamide gels (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), followed by transfer to polyvinyldifluoride membranes (Millipore, Billerica, MA, http://www.millipore.com), which were then probed with antibodies. Total Akt and phospho-Akt (Ser473) antibodies were purchased from Cell Signaling Technology (Beverly, MA, http://www.cellsignal.com), whereas α-tubulin antibodies were purchased from Sigma-Aldrich. Olig2 antibodies were purchased from R&D Systems, Inc. (Minneapolis, http://www.rndsystems.com). Proteins were detected with an enhanced chemiluminescence system (Pierce Biotechnology, Rockford, IL, http://www.piercenet.com).

Akt Kinase Assay

CD133+ and CD133 cells were plated in appropriate media in 10-cm dishes at 106 cells per plate and allowed to recover overnight. CD133 medium was changed to CD133+ medium before treatment with DMSO or increasing concentrations of the Akt inhibitor. After 24 hours of treatment, cells were changed to neurobasal medium without growth factors for 12 hours. Cells were then stimulated for 20 minutes with EGF and FGF at a final concentration of 50 ng/ml before harvest and lysis. Equal amounts of total protein (50 μg in 200 μl) were used for the in vitro kinase assay, which was performed as per the manufacturer's instructions (Cell Signaling).

Proliferation and Survival Assay

CD133+ and CD133 cells were plated in appropriate media in six-well plates in triplicate at 1 × 105 cells per well and left to recover overnight. CD133 medium was changed to CD133+ medium before treatment with DMSO or increasing concentrations of the Akt inhibitor. After 48 hours of treatment, cells were harvested, and live and dead cells were quantified via a hemocytometer with trypan blue stain (Gibco).

Flow Cytometric and Annexin V Analysis

CD133+ and CD133 cells were plated in appropriate media in six-well plates in triplicate at 1 × 105 cells per well and were allowed to recover overnight. CD133 medium was changed to CD133+ growth medium before treatment with DMSO control or increasing concentrations of Akt inhibitor for 24 hours. Cells were then harvested with their conditioned medium to ensure collection of floating cells along with adherent cells. Using an annexin V kit, apoptotic cells were labeled with fluorescein isothiocyanate whereas necrotic cells were labeled with propidium iodide as per the manufacturer's instructions (EMD Chemicals, San Diego, http://www.emdbiosciences.com). Proportions of apoptotic cells were then quantified by FACScan gated to exclude cellular debris.

Neurosphere Formation Assay

CD133+ cells were plated in 24-well plates at 20 cells per well with DMSO or Akt inhibitor at increasing concentrations. Wells were inspected every 3 days, numbers of neurospheres were quantified at each time point, and individual neurospheres were imaged with an Olympus CK40 digital camera (Olympus, Tokyo, http://www.olympus-global.com) mounted to a light microscope.

Migration and Invasion Assay

Migration/invasion plates were purchased and used according to the manufacturer's instructions (BD Biosciences, San Diego, http://www.bdbiosciences.com). CD133+ and CD133 cells were pretreated for 1 hour with increasing concentrations of Akt inhibitor, then plated at 5 × 104 cells per well in upper transwell chambers of inserts uncoated (migration) or coated with Matrigel (invasion) in serum- and growth factor-free media. The bottoms of the chambers were filled with 500 μl of medium containing 2% FBS. After 24 hours, migration inserts were fixed and stained with Diff-Quick Fixative Solutions (Dade Behring, Newark, DE, http://diagnostics.siemens.com). After 48 hours, Matrigel-coated inserts were fixed and stained. Attached cells were imaged with an Olympus CK40 digital camera mounted to a light microscope and quantified using ImageJ software (http://rsb.info.nih.gov/ij/).

Statistical Analysis

Statistical significance was calculated with GraphPad Prism Software (GraphPad Software Inc., La Jolla, CA, http://www.graphpad.com). Data are presented as the mean ± standard error. Statistical significance in the human xenograft mouse models was calculated using Kaplan-Meier survival curves and statistical analysis was conducted with MedCalc Software (Mariakerke, Belgium, http://www.medcalc.be).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Brain Tumor Stem Cells Exhibit Greater Sensitivity to Akt Inhibition

Cancer stem cells are defined through functional assays to determine the capacity for sustained self-renewal and the ability to recapitulate the full diversity of the parental tumor upon xenotransplanation [3, 4, 5, 6, 78]. Several groups, including our own laboratory, have demonstrated that cancer stem cells can be enriched prospectively through the use of the CD133 (Prominin-1) cell surface marker [3, 4, 5, 6, 78]. The CD133 marker does not absolutely segregate for tumorigenesis because some tumors may contain CD133 cells that form tumors, although this frequently requires transplantation of very high numbers of cells [22, 23]. However, we have found that CD133+ tumor cells from patient biopsy specimens display both potent neurosphere formation potential in cell culture and effective tumor generation in immunocompromised animal models, whereas CD133 cells do not form neurospheres and rarely, if ever, form tumors upon xenotransplanation [3]. We, therefore, used models that we have previously characterized in functional assays (sustained neurosphere generation, multilineage differentiation, tumor initiation) to define “cancer stemness” in our studies. In addition, the CD133+ cells used for these novel studies highly expressed stem cell markers such as Nestin, Oct4, Olig2, and Sox2, whereas CD133 cells did not (supplemental online Figs. 1, 2). Taken together these results suggest that CD133+ cells are enriched for cancer stem cells.

To determine whether CD133+ tumor cells exhibit differential activation of the PI3K-Akt pathway compared with their nonstem counterparts, the activation state of pathway components was examined through immunoblotting. In short-term cultures (<5 passages) of T3359 (Fig. 1A, 1B and supplemental online Fig. 3A, 3B) or D456MG (Suppl. Fig. 3C, 3D) glioblastoma cells, total levels of Akt protein were similar between CD133+ and CD133 cells. In contrast, basal levels of phosphorylated (activated) Akt were higher in the CD133 populations (Fig. 1A, 1B and supplemental online Fig. 3). Consistent with these data, we found decreased Akt kinase activity in CD133+ cells in comparison with matched CD133 cells (Fig. 1C, 1D). Although we expected cancer stem cells to have a higher basal activation of Akt, the relatively lower activation of Akt in CD133+ cells may be due to differences in cell attachment. Cancer stem cells grow in three-dimensional neurospheres whereas nonstem cancer cells are adherent to tissue culture plates. Because there may be differences in cell–cell and cell–matrix interactions between these growth conditions, we evaluated Akt phosphorylation in short-term cultures of adherent and nonadherent CD133 cells to ensure survival of non-adherent CD133 cells (which is compromised by the lack of serum and cell adhesion). We found adherent CD133 cells had greater Akt phosphorylation than either nonadherent CD133 or CD133+ cells, which had very similar levels of basal Akt phosphorylation (supplemental online Fig. 4).

thumbnail image

Figure Figure 1.. The Akt pathway is differentially targeted by an Akt inhibitor in CD133+ and CD133 brain tumor cells. CD133+ and CD133 cells were isolated from a T3359 glioblastoma patient specimen passaged short term in immunocompromised mice and treated with the indicated concentrations of Akt III inhibitor (SH-6, Calbiochem, San Diego, http://www.emdbiosciences.com) for 2 hours. Lysates were analyzed by Western blotting (A) and the intensities were quantified using ImageJ software and normalized to the tubulin loading control (B). The levels of phospho-Akt, normalized to tubulin, decreased in a concentration-dependent manner in the CD133+ population but not the CD133 population. Akt was also immunoprecipitated from lysates and used for an in vitro Akt kinase assay with GSK3α/β fusion protein as the substrate (C, D). Western blotting analysis (C) and intensities quantified using ImageJ software and normalized to total GSK3α/β indicated decreased basal and Akt-mediated phosphorylation of GSK3α/β in CD133+ cells in comparison with matched CD133 cells. AktIII inhibitor decreased the Akt-mediated phosphorylation of GSK3α/β in CD133+ cells, but not matched CD133 cells.

Download figure to PowerPoint

Several pharmacologic agents have been designed to inhibit Akt function. Phosphatidylinositol ether lipid analogs may target the pleckstrin homology domain of Akt to selectively inhibit cell survival in cancers with high Akt activity [24]. One of these drugs, AktIII inhibitor (SH-6), reduced Akt activation in CD133+ glioma cells in a concentration-dependent manner but had no beneficial effect on CD133 cells (Fig. 1A, 1B). The temporal course of AktIII inhibitor was also assessed in matched tumor cell populations and demonstrated transient effects in both populations with relatively greater effects on the CD133+ cells when normalized to the basal level of activation (supplemental online Fig. 3). This decrease in activated Akt is not due to a decrease in total Akt levels, nor to improper loading of protein samples, as assessed by tubulin controls (Fig. 1 and supplemental online Fig. 3). Therefore, these data demonstrate that CD133+ and CD133 cells exhibit differential activation levels of the Akt pathway, and the CD133+ population has a greater sensitivity to the Akt inhibitor.

Akt Activity Is Necessary for Cancer Stem Cell Proliferation and Survival

To determine if targeting Akt activity would preferentially decrease the protumorigenic behaviors of CD133+ cells, the proliferation and survival of matched CD133+ and CD133 short-term brain tumor cell cultures were interrogated. Despite having identical numbers of cells plated at the initiation of studies, at the end of the experiments CD133+ cultures yielded higher numbers of viable cells at baseline than the matched CD133 cells, likely due to the long-term proliferative potential of these cells. However, the pharmacologic AktIII inhibitor demonstrated a significant concentration-dependent effect in reducing the number of viable CD133+ cells with more modest effects on CD133 cells (Fig. 2A, 2B) as determined by trypan blue staining. The preferential decrease in CD133+ cell numbers is not likely to be due to potential differences in basal Akt phosphorylation in this assay with adherent CD133 cells. A similar preferential decrease in CD133+ cell growth was observed when both CD133 and CD133+ cells were cultured in stem cell medium under nonadherent conditions with two different Akt inhibitors (Fig. 3 and supplemental online Fig. 5). Use of the PI3K inhibitor LY290042 to inhibit an upstream component of the Akt pathway also demonstrated greater effects on CD133+ cell growth with increasing concentrations of drug (Fig. 3 and supplemental online Fig. 5). When mTOR, a downstream target of Akt, was targeted with rapamycin there was a more modest preferential targeting of CD133+ cell growth (Fig. 3 and supplemental online Fig. 5). Together these results suggest that inhibition of PI3K or Akt more potently regulates the growth of CD133+ cells than CD133 cells.

thumbnail image

Figure Figure 2.. A small molecule inhibitor of Akt targets CD133+ cell growth and survival. CD133+ and CD133 cells were isolated from an established D456MG pediatric glioblastoma xenograft (A, C) or a T3359 glioblastoma patient specimen passaged short term in immunocompromised mice (B, D). Cells were treated with the indicated concentrations of AktIII inhibitor for 48 hours and the numbers of live cells (A, B) and the percentage of dead cells (C, D) as a fraction of the control was determined through trypan blue staining. *, p < .01 with an ANOVA comparison of AktIII-treated CD133+ cells with DMSO control-treated CD133+ cells; ≈, p < .01 with an ANOVA comparison of AktIII-treated CD133 cells with DMSO control-treated CD133 cells;. #, p < .01 with ANOVA comparison of CD133+ cells with identically treated CD133 cells. Abbreviations: ANOVA, analysis of variance; DMSO, dimethyl sulfoxide.

Download figure to PowerPoint

thumbnail image

Figure Figure 3.. Targeting Akt preferentially decreases CD133+ brain tumor cell growth over time. CD133+ and CD133 cells isolated from a T3359 glioblastoma patient specimen passaged short term in immunocompromised mice were plated in neurobasal medium with EGF and FGF, allowed to recover overnight, and then treated with 10 μM AktIII (A), 10 μM AktII (B), 25 μM AktIII (C), 25 μM AktII (D), 50 μM LY290042 (E), and 100 nM rapamycin (F) inhibitors. Cell growth was measured on the indicated days after inhibitor treatment began using the Cell Titer Glo assay (Promega, Madison, WI, http://www.promega.com) according to the manufacturer's instructions. The data for each time point were standardized to the DMSO-treated controls for the same cell type on each day. *, p < .05 with t-test comparison of inhibitor-treated CD133+ cells with DMSO-treated control CD133+ cells on the same day; ≈, p < .01 with t-test comparison of inhibitor-treated CD133 cells with DMSO-treated control CD133 cells on the same day; #, p < .05 with t-test comparison of CD133+ cells with similarly treated CD133 cells on the same day. Abbreviations: DMSO, dimethyl sulfoxide; EGF, epidermal growth factor; FGF, fibroblast growth factor.

Download figure to PowerPoint

To evaluate the impact of pharmacologic Akt inhibition on CD133+ cell survival, the killing efficiency was measured over a range of inhibitor concentrations. Although the percentage of dead cells increased in a concentration-dependent manner with Akt inhibition in both the CD133+ and CD133 cells (Fig. 2C, 2D), the CD133+ populations were more significantly affected than the CD133 cells. After determining that AktIII preferentially targets CD133+ tumor cell survival, we investigated the mechanism of cell death using annexin V staining to assess apoptosis in CD133+ and CD133 cells. In parallel with our earlier results, each CD133+ tumor cell culture demonstrated a concentration-dependent increase in apoptosis upon treatment with the Akt inhibitor (Fig. 4 and supplemental online Fig. 6). In contrast, CD133 cells displayed little to no increase in the apoptotic cell fraction in response to AktIII inhibitor treatment (Fig. 4 and supplemental online Fig. 6). These results indicate that the preferential cell death in CD133+ cancer stem cells is at least partially due to increased apoptosis, consistent with the known effects the Akt pathway has on cellular apoptosis and survival.

thumbnail image

Figure Figure 4.. Targeting the Akt pathway results in preferential induction of CD133+ cell apoptosis. CD133+ and CD133 cells isolated from an established D456MG pediatric glioblastoma xenograft (A, B) or a T3359 glioblastoma patient specimen passaged short term in immunocompromised mice (C, D) were treated with the indicated concentration of AktIII inhibitor for 24 hours, trypsinized, labeled with an annexin V kit according to manufacturer's instructions and analyzed by FACS. Apoptosis was induced in CD133+ cells at significantly higher levels than in CD133 cells with increasing concentrations of Akt inhibitor. *, p < .001 with an ANOVA comparison of AktIII-treated CD133+ cells with DMSO control-treated CD133+ cells; #, p < .01 with an ANOVA comparison of CD133+ cells with identically treated CD133 cells. Representative FACS gates of CD133+ and CD133 cells from D456MG (B) and T3359 (D) are shown. Abbreviations: ANOVA, analysis of variance; DMSO, dimethyl sulfoxide; FACS, fluorescence-activated cell sorting.

Download figure to PowerPoint

Cancer Stem Cell Neurosphere Formation Requires Akt Activity

Both brain tumor stem cells and neural stem cells display the ability to form complex three-dimensional spherical structures (neurospheres) when cultured in serum-free medium. In our studies, neurosphere formation potential is restricted to the prospectively enriched CD133 + tumor population, although other labs have generated neurospheres that did not express CD133 [20]. To further investigate the effects of Akt inhibition on the cancer stem cell behaviors of CD133+ brain tumor cells, we examined the effects of the inhibitor on neurosphere formation in the presence of increasing concentrations of Akt inhibitor. CD133+ tumor cells displayed a striking concentration-dependent decrease in the ability to generate neurospheres across each time point examined (Fig. 5). When neurospheres that did form with Akt inhibitor treatment were further analyzed, there was a clear qualitative decrease in size and the cells were unable to form secondary neurospheres (Fig. 5 and data not shown). These results indicate that CD133+ brain tumor cells require Akt activity for neurosphere formation.

thumbnail image

Figure Figure 5.. Targeting the Akt pathway decreases the neurosphere formation efficiency of CD133+ cells. CD133+ cells isolated from an established D456MG pediatric glioblastoma xenograft (A) or a T3359 glioblastoma patient specimen passaged short term in immunocompromised mice (B) were plated at an approximate density of 20 cells per well and treated with the indicated concentration of AktIII inhibitor. The number of neurospheres per well was quantified over three time points. The number of neurospheres per well significantly decreased in the presence of the Akt inhibitor. Representative images of neurospheres photographed at day 9 are shown and demonstrate a clear qualitative difference in size between those grown in the control conditions and those treated with the Akt inhibitor. *, p < .001 with an ANOVA comparison of AktIII inhibitor treatment with the DMSO-treated control on the same day. Abbreviations: ANOVA, analysis of variance; DMSO, dimethyl sulfoxide.

Download figure to PowerPoint

Targeting Akt Decreases Cancer Stem Cell Migration and Invasion

The ability of malignant gliomas to invade into normal neural structures leads to the inability for these tumors to be completely surgically resected [2]. Although the mechanisms underlying brain tumor invasion remain incompletely understood, the PI3K–PTEN–Akt axis has been recognized as a contributor to invasion. Therefore, we expected that Akt inhibition may negatively regulate tumor cell invasion [16, 20]. Using Boyden chamber assays, the migration and invasion capacity of matched CD133+ and CD133 brain tumor cells was evaluated. Interestingly, there was a striking basal difference in the capacity of CD133+ and CD133 cells to migrate through either an uncoated membrane or a membrane coated with an artificial extracellular matrix (Matrigel). In one model (D456MG), the CD133 cells displayed much greater migratory and invasive potential (Fig. 6), perhaps due to the long-term passage of these cells in a xenograft. In contrast, the CD133+ cells in the T3359 model were more invasive than the CD133 cells (supplementary online Fig. 7), consistent with the notion that CD133+ cells contribute to invasion and migration. Increasing concentrations of Akt inhibitor significantly attenuated the capacity of both CD133+ and CD133 cells to both migrate and invade (Fig. 6), but in both models the CD133+ cells displayed greater sensitivity to the inhibitory effects of Akt inhibition than the cancer stem cell-depleted CD133 population (Fig. 6 and supplemental online Fig. 7). Thus, the migratory and invasive potential of cancer stem cells relative to the nonstem cell population may depend on the tumor model, but cancer stem cells depend on Akt activity for these proinvasive behaviors.

thumbnail image

Figure Figure 6.. Targeting the Akt pathway decreases CD133+ cell migration and invasion. CD133+ cells isolated from an established D456MG pediatric glioblastoma xenograft were plated in serum-free medium in the upper chambers of uncoated inserts (A-C) or Matrigel-coated inserts (D-F) and were allowed to migrate toward 2% FBS for 48 hours. The migrating or invading cells were then stained and quantified with ImageJ software, demonstrating that both migration (A) and invasion (D) decrease in CD133+ cells with increasing concentrations of Akt inhibitor. When the percent change from baseline migration (B) or invasion (E) was calculated, CD133+ cells exhibited a greater sensitivity to the effects of the Akt inhibitor. Representative images of migrating (C) or invading (F) cells are shown. *, p < .05 with an ANOVA comparison of AktIII inhibitor-treated CD133+ cells with DMSO-treated control CD133+ cells; ≈, p < .05 with an ANOVA comparison of AktIII inhibitor-treated CD133 cells with DMSO-treated control CD133 cells; #, p < .001 with an ANOVA comparison of CD133+ cells with similarly treated CD133 cells. Abbreviations: ANOVA, analysis of variance; DMSO, dimethyl sulfoxide; FBS, fetal bovine serum.

Download figure to PowerPoint

Inhibition of Akt Activity in Cancer Stem Cells Increases Survival of Immunocompromised Mice Bearing Human Glioma Xenografts

To further determine the potential therapeutic benefit of targeting Akt activity in cancer stem cells, we determined the tumorigenic potential of cells treated with Akt III inhibitor. When 1 × 104 or 1 × 103 CD133+ cells were injected into the forebrains of immunocompromised mice, we observed neurologic signs due to the development of brain tumors regardless of Akt inhibitor treatment. These data demonstrate that inhibition of Akt activity alone for the treatment period was not sufficient to significantly reduce the tumor formation potential of cancer stem cells (Table 1). However, the time to the development of neurologic signs was significantly increased with Akt inhibition. The median survival time until the development of neurologic signs of animals bearing 1 × 104 CD133+ tumor cells treated with DMSO control was 24 days, whereas animals bearing identical CD133+ cells treated with the Akt inhibitor survived for a median of 42 days (p < .03). Similarly, the median survival time until the development of neurologic signs of animals bearing 1 × 103 CD133+ cells was 35 days for the DMSO control and 66 days for Akt inhibitor treatment (p < .03). These data demonstrate that reducing Akt activity in CD133+ cells can increase the survival of mice bearing intracranial xenografts. Prior work in our laboratory has implicated cancer stem cells in promoting tumor angiogenesis via elevated vascular endothelial growth factor (VEGF) secretion [4]. However, the increased survival of mice injected with Akt inhibitor-treated CD133+ cells is unlikely to be due to changes in VEGF levels, because the Akt inhibitor did not reduce VEGF expression (supplemental online Fig. 8), and is more likely to be attributed to the induction of apoptosis (Figs. 2, 4 and supplemental online Fig. 6).

Table Table 1.. Effect of AktIII/SH-6 pretreatment of xenograft-derived cells on glioma initiation
Thumbnail image of

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

The essential pathways regulating cancer stem cell biology remain poorly defined, but molecular targets with defined roles in normal stem cell biology and aberrant activity or expression in cancers (including Notch, Hedgehog, Wnt/β-catenin, bone morphogenic protein, Myc, epidermal growth factor receptor, fibroblast growth factor receptor, and PTEN) are likely to be important. Activation of these pathways by autocrine signals from the cancer stem cells themselves or paracrine signals from the cancer stem cell niche could be essential for stem cell maintenance [25, 26]. Recognizing that these stem cell maintenance cues regulate cell survival and differentiation, these signals are logical targets for anticancer stem cell-directed therapies.

The potential for inhibition of the PI3K-Akt pathway to target cancer stem cells is supported by the known involvement of Akt signaling in tumorigenesis and normal stem cell biology as well as the beneficial effects of Akt inhibition on glioma cell growth [27, 28, 29, 30, 31, 32, 3334]. Multiple protumorigenic behaviors (such as the promotion of cell proliferation, survival, and invasion) now suggested to be driven by cancer stem cells [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 1415] are known to be regulated by Akt signaling [20]. Activation of the PI3K-Akt pathway is common in malignant gliomas and associated with increased tumor grade and decreased glioma patient survival [16, 17, 18, 1920]. Genetically engineered glioma models also demonstrate that that constitutively activated Akt contributes to tumor initiation [27, 28].

In addition to having roles in tumor formation, the PI3K-Akt pathway is implicated in stem cell biology. Loss of the PI3K inhibitor PTEN increases neural and hematopoeitic stem cell proliferation and survival [29, 30]. Recent studies in the hematopoietic system have shown that activation of the PI3K-Akt pathway in normal hematopoietic stem cells can produce leukemia within weeks [31]. However, leukemic stem cells are also particularly sensitive to the effects of mTOR antagonism [31], suggesting a potential therapeutic window for targeting the PI3K-Akt pathway in cancer stem cells with minimal effects on normal somatic stem cells. Until recently, this strategy has not been fruitfully applied to the cancer stem cells of solid tumors. Ma et al. [11] demonstrated that the CD133+ cells from two established hepatocellular carcinoma cell lines are less sensitive to chemotherapy and express higher levels of survival proteins involved in the Akt and Bcl-2 pathway than the CD133 cells, and respond to Akt inhibition by reducing key survival proteins. Hambardzumyan and coworkers used a murine model of medulloblastoma (a primary brain tumor that is distinct from the gliomas that commonly occur in children and displays greater radiosensitivity than gliomas) to show that a nestin-expressing perivascular tumor cell population survives radiation, activates downstream effectors of Akt, undergoes p53-dependent cell-cycle arrest, and re-enters the cell cycle at 72 hours [13]. In addition, inhibition of Akt signaling sensitizes cells in the perivascular region to radiation-induced apoptosis [13]. We have extended these findings into a cancer type in which the Akt–PTEN axis plays an essential role using models derived from human specimens or maintained in vivo then only briefly cultured to maintain a cancer stem cell phenotype. We sought to further determine if inhibition of Akt signaling in glioma stem cells may be a beneficial mechanism for reducing cancer stem cell growth in vitro and increasing survival in vivo.

To evaluate the effects of Akt inhibition in cancer stem cells and nonstem cancer cells, we have built on prior investigations, including our own, that demonstrated that cancer stem cells may be enriched through the use of the CD133 (Prominin-1) cell surface marker [3, 4, 5, 6, 78]. The use of CD133 must be viewed with caution as CD133 has not been linked to a contributory role in “cancer stemness” and some tumors may have CD133 cells with tumor-initiation abilities [22, 23]. However, experiments in our laboratory find a striking enrichment of cancer stem cells in the CD133+ tumor cell population even though CD133 should not be considered a surrogate for a cancer stem cell phenotype.

When we compared the effects of Akt inhibition on populations of brain tumor cells, we found a preferential targeting of Akt activity in CD133+ cells in comparison with matched CD133 cells. The AktIII inhibitor that we employed for the majority of our experiments is a phosphatidylinositol ether lipid analog that may also activate the stress kinase p38α. This small molecule inhibitor of Akt effectively reduced the growth, survival, migration, and invasion of glioma cells and did so with greater potency in the CD133+ subpopulation than in matched CD133 cells. Inhibiting the Akt pathway in brain cancer stem cells also increased survival in animal models, suggesting a potential therapeutic benefit. Together these data suggest that many of the malignant characteristics of brain tumor stem cells are dependent on Akt signals.

We were somewhat surprised to find that CD133 tumor cells expressed higher basal levels of activated Akt than CD133+ cells, considering the recognized role of Akt in survival and prior reports indicating higher levels of phosphorylated Akt in human hepatocellular carcinoma and mouse medulloblastoma stem cells [11, 13]. Because our studies were conducted in different tumor types and culture conditions, it is difficult to make direct comparisons. However, we found that differences in attachment between CD133 and CD133+ cells can alter basal Akt phosphorylation, suggesting that different culture conditions may contribute to the discrepancy in basal Akt activation. All of our primary experiments were performed on short-term cultures in which CD133+ cells formed neurospheres (were nonadherent) and CD133 cells were adherent in the presence of identical media in an effort to make the best possible comparisons. We also used AktII and AktIII inhibitors rather than the previously used AktI inhibitor [11] or perifosine [13]. Using these conditions and inhibitors, we found that CD133+ cells exhibited greater sensitivity to Akt inhibition. The precise molecular differences between CD133 and CD133+ cells that could contribute to the observed higher potency of Akt inhibition in CD133+ cells remain to be fully elucidated.

The concept of cancer stem cells is still evolving, but data implicating these cells in tumor maintenance and therapeutic resistance indicate the potential benefit of targeting cancer stem cells in combination with conventional therapies. Whereas targeting Akt activation in CD133+ cells increased the survival of mice bearing intracranial human glioblastoma xenografts, Akt inhibition alone is unlikely to directly translate to therapeutic benefit for human patients. Monotherapies against any signaling pathway have been largely ineffective in the clinic. However, data from our laboratory and others suggest the utility of targeting Akt signaling components alone or in combination with other pathways in gliomas [32, 3334]. As these prior studies focus on the effects of Akt inhibitors in glioma cell lines that are passaged long term in the presence of serum (a condition that promotes the differentiation of cancer stem cells), future studies determining the effect of combining Akt inhibition with chemo- and radiotherapies on cancer stem cell biology may prove enticing. Ensuring that both glioblastoma stem cells and the more prevalent CD133 cells are targeted may offer the opportunity to eliminate the last vestiges of the primary tumor after surgical resection, an absolute requirement for preventing recurrence.

In conclusion, CD133+ glioblastoma stem cells were shown to have increased sensitivity to the effects of a small molecule Akt inhibitor, despite exhibiting decreased baseline activation of the Akt pathway compared with CD133 cells. Akt inhibition produced a preferential reduction in CD133+ cell growth, survival, migration, and invasion in comparison with CD133 cells. The AktIII small molecule inhibitor also targeted characteristics unique to cancer stem cells, such as the ability to form neurospheres. Targeting Akt activity in CD133+ cells also increased the survival of immunocompromised mice bearing glioma xenografts, indicating a potential therapeutic benefit. Although many protumorigenic behaviors of CD133+ cells were reduced to a greater extent than in the CD133 cells, the secretion of VEGF was not one of them, indicating that not all the unique features of CD133+ cells are dependent on Akt signaling.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

We thank Kathryn Lattimore, Pei Miao, Sarah Wickman, and Qiulian Wu for technical assistance and Dr. Mike Cook and Dr. Beth Harvat for assistance with flow cytometry.

Financial support was provided by the Childhood Brain Tumor Foundation, the Pediatric Brain Tumor Foundation of the United States, Accelerate Brain Cancer Cure, Alexander and Margaret Stewart Trust, Brain Tumor Society, Goldhirsh Foundation, Duke Comprehensive Cancer Center Stem Cell Initiative Grant (J.R.), and NIH grants NS047409, NS054276, CA129958, and CA116659 (J.R.). W.F. is a Howard Hughes Research Training Fellow. J.R. is a Damon Runyon-Lilly Clinical Investigator supported by the Damon Runyon Cancer Research Foundation and a Sidney Kimmel Foundation for Cancer Research Scholar. C.E. receives Medical Scientist Training Program support from the National Institute of General Medical Sciences grant 2T32GM007171.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
SC-07-1073_Suppl_Fig_7.pdf183KSupplemental Figure 7
SC-07-1073_Suppl_Fig_Legends.pdf29KSupplemental Figure Legends
SC-07-1073_Suppl_Fig_1.tif1692KSupplemental Figure 1
SC-07-1073_Suppl_Fig_2.tif907KSupplemental Figure 2
SC-07-1073_Suppl_Fig_3.tif1199KSupplemental Figure 3
SC-07-1073_Suppl_Fig_4.tif948KSupplemental Figure 4
SC-07-1073_Suppl_Fig_5.tif845KSupplemental Figure 5
SC-07-1073_Suppl_Fig_6.tif1421KSupplemental Figure 6
SC-07-1073_Suppl_Fig_8.tif183KSupplemental Figure 8

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.