Glioblastomas display cellular hierarchies with self-renewing tumor-initiating cells (TIC), also known as cancer stem cells, at the apex. Although the TIC hypothesis remains controversial and the functional assays to define the TIC phenotype are evolving, we and others have shown that TICs may contribute to therapeutic resistance, tumor spread, and angiogenesis. The identification of TICs has been informed by the use of markers characterized in normal stem cells, but this approach has an inherent limitation to selectively identify TICs. To develop reagents that enrich TICs but not matched non-TICs or tissue-specific stem cells, we adopted Cell-Systematic Evolution of Ligands by Exponential Enrichment (Cell-SELEX) to identify glioblastoma TIC-specific nucleic acid probes—aptamers—that specifically bind TICs. In this study, using Cell-SELEX with positive selection for TICs and negative selection for non-TICs and human neural progenitor cells, we identified TIC aptamers that specifically bind to TICs with excellent dissociation constants (Kd). These aptamers select and internalize into glioblastoma cells that self-renew, proliferate, and initiate tumors. As aptamers can be modified to deliver payloads, aptamers may represent novel agents that could selectively target or facilitate imaging of TICs. Cancer Res; 73(15); 4923–36. ©2013 AACR.

Cancers invoke molecular programs expressed during development and wound responses to promote the initiation and maintenance of complex neoplastic tissue systems that include not only transformed cells but also supportive vasculature, immune components, stroma, and extracellular matrix (1). Standard human cancer models based on established cell lines are subjected to passage under conditions that select for rapid proliferation and survival through mechanisms often distinct from the original tumor. Although proliferation and resistance to apoptosis are hallmarks of cancer, other aspects of cancer—invasion of normal tissues, metastasis, resistance to cytotoxic insults, and vascular recruitment–critically contribute to the lethality of cancer (1). An increasing number of cancers have been shown to display cellular hierarchies with a subset of the neoplastic compartment activating molecular mechanisms and cellular phenotypes similar—but not identical—to embryonic or tissue-specific stem cells (2, 3). These tumor-initiating cells (TIC), also called cancer stem cells or tumor-propagating cells, are functionally defined through assays of self-renewal and tumor propagation (4). We and others have shown that TICs are relatively resistant to conventional cancer therapies (radio- and chemotherapy) and promote tumor growth through angiogenesis (5–8). The TIC hypothesis has been questioned because of potential plasticity of the cellular hierarchy and difficulties with TIC identification but these challenges are products of our attempts to simplify complex systems with limited technical resources. On the basis of this background, creating agents that prospectively identify TICs may not only permit the interrogation of the cellular hierarchy in cancers but also serve as a platform for the development of novel targeted therapies and imaging reagents.

Glioblastoma (GBM) is the most prevalent and lethal primary brain tumor and ranks among the most lethal of all cancers (8). Through the work of many groups the presence of a cellular hierarchy has been supported in not only glioblastoma but also other central nervous system cancers (9, 10). The characterization of TICs is based on paradigms developed from embryonic and tissue-specific stem cells, but TICs are distinct from these normal cells so the immunophenotypes may not fully overlap. Several surface TIC markers have shown promise in glioblastoma, including CD133 (Prominin-1; refs. 5, 11, 12), CD15/Lewis X-antigen/stage-specific embryonic antigen-1 (SSEA-1; ref. 13), CD44 (14), L1CAM (15), integrin α6 (16), EGF receptor (EGFR; refs. 17, 18), platelet-derived growth factor receptor β (PDGFRβ; ref. 19), and the EphA2 receptor (20). Functional assays, including ALDEFLUOR and side population, have been less reliable in glioblastoma but useful in other cancer types (21). Neurosphere formation has been used to enrich TICs but this method prevents the prospective separation of tumorigenic and nontumorigenic cells that define a cellular hierarchy. Although these markers have been useful in some studies to prospectively enrich or deplete TICs, many of these have been limited by a shared antigen with normal neural progenitors. Several groups have taken other approaches to identify TIC targets. RNA interference screens have identified key transcription factors, kinases, phosphatases, or ubiquitin-modifying enzymes (22, 23). One group reported a TIC screen in prostate cancer stem cells (24) and another identified CD133-binding aptamers (25). However, we are not aware of a large screen to identify novel TIC-enrichment reagents in glioblastoma. On the basis of this background, we hypothesized that the application of a large, unbiased screening method could develop highly specific reagents to identify functional glioblastoma TICs.

Aptamers are short nucleic acids capable of specific and tight binding to target structures that can include proteins, lipids, other nucleic acids, or any three-dimensional structures. Aptamers are named from the Latin aptus (fitting) and Greek meros (part). Aptamers can be selected by the in vitro iterative process SELEX (Systematic Evolution of Ligands by EXponential enrichment) that enriches RNA or DNA aptamers capable of discriminating between molecular targets with even subtle differences (26–29). Cell-SELEX is a modified procedure that uses cells as an evolutionary selection source and does not require foreknowledge of a molecular target (28, 30). In general, Cell-SELEX is a combination of in vitro evolution and combinatorial chemistry involving a series of steps including incubation, partitioning, and amplification.

In Cell-SELEX, the initial pool of library is chemically created by solid-phase technology. This library must contain at least one, but preferably several, sequences(s) having the unique folded conformations required to facilitate selective binding with the target. Cell-SELEX typically involves positive selection to isolate target cell-interacting sequences and counter selection to eliminate nonspecific sequences recognizing nontarget cells (28). Counter selection results in the depletion of aptamers that bind common cell surface receptors in the resulting pools. Thus, the probability of recognizing unique molecules exclusively expressed in the target population is greatly enhanced. Aptamers bind to their targets with high affinity with typical binding dissociation constants (Kd) in the pmol/L to nmol/L range. Furthermore, aptamers are structurally stable under a wide range of temperature and storage conditions while instantly reconstituting their active tertiary structures in some conditions. Aptamers can be synthesized quickly and cost-effectively with minimal interbatch variability, in contrast to monoclonal antibodies. Aptamers can also be easily modified to increase stability and affinity for therapeutic and diagnostic purposes (26). Together, these data suggest that there may be strong benefits of identifying TIC-binding aptamers.

Isolation and culture of cells including TICs

We designed a two-stage process for Cell-SELEX with an initial discovery phase to enrich TIC-binding aptamers using functionally validated TICs followed by a validation phase–interrogating aptamer specificity with minimally cultured bulk cancer cells derived directly from in vivo environments. To obtain the large numbers of cells necessary for the initial phase, functionally validated TICs were derived from pooled tumors of the same parental glioblastoma xenograft. Tumors were dissociated using the Worthington Biochemical Papain Dissociation System according to the manufacturer's instructions. To facilitate recovery from enzymatic digestion, cells were then cultured in Neurobasal Media supplemented with B27 without vitamin A (Invitrogen), basic fibroblast growth factor (bFGF; 10 ng/mL), and EGF (10 ng/mL) for at least 6 hours before magnetic sorting. In selecting an enrichment method for TICs, we considered several cell surface markers that we and others have used. Although CD133 has been controversial in some reports as some glioblastomas may not express CD133 and others have tumorigenic CD133-negative cells, CD133 can identify TIC phenotypes in specific glioblastoma models in our studies (5, 31, 32). After minimal culture, we separated CD133-positive and -negative cells using microbead-conjugated CD133 antibodies (Miltenyi Biotech). To prevent the effects of culture, we conducted Cell-SELEX in parallel with confirmatory TIC assays. CD133-enriched neoplastic cells were functionally validated in assays of self-renewal and tumorigenesis (data not shown) then designated as TICs and used as targets to enrich TIC-specific aptamers and were maintained in the above stem cell media. CD133-negative cells depleted for tumorigeneic or self-renewing capacity (data not shown) were used as non-stem glioma cells (non-TICs) and cultured in the presence of 10% FBS. Human glioblastoma xenografts maintained exclusively in vivo (GBM10, GBM12, and GBM14) were kindly provided by Jann Sarkaria (Mayo Clinic, Rochester, MN). These xenografts were maintained as previously described (33). Human neural progenitors were commercially obtained (Lonza).

DNA primers and library

The DNA library used for TIC-specific aptamer Cell-SELEX was a pool of DNA sequences consisting of a combination of common and unique nucleotides. Nineteen common nucleotides were present on the 5′-end, 50 randomized base sequences in the middle, and an additional 18 common nucleotides on the 3′-end. The 5′-end was labeled by fluorescein isothiocyanate (FITC; ref. 5′-FITC-sequence-N45-sequence-3′) to monitor enrichment of selection using fluorescence-activated cell sorting (FACS). The forward primer was labeled at the 5′-end with FITC (5′-FITC-sequence-3′) and the reverse primer with biotin at the 5′-end (5′-biotin-sequence-3′). To amplify each eluted pool, PCR was used, and ssDNA was isolated by capturing the biotinylated complementary strand by streptavidin–biotin interactions and denaturing double-stranded DNA with 200 mmol/L NaOH. PCR mixtures were prepared and PCR reactions were carried out according to the manufacturer's instructions.

Cell-SELEX

Minimally cultured TICs isolated from pooled 08–387 xenografts were used as target (positive cell) and matching non-TICs and neural progenitors were used for counter selection (negative cell). Of note, 5 nmol of DNA library was dissolved in 1,000 μL of binding buffer containing 4.5 g/L glucose, 5 mmol/L MgCl2, 0.1 mg/mL tRNA, and 1 mg/mL bovine serum albumin in Dulbecco's PBS. The DNA library or enriched pool was denatured at 95°C for 5 minutes, cooled on ice for 10 minutes, and incubated with TICs on ice in an orbital shaker for 1 hour. After TICs were washed three times to remove unbound DNA sequences, the bound DNA sequences were eluted using 500 μL binding buffer at 95°C for 10 minutes with centrifugation. To carry out a counter selection, each aptamer pool was incubated with non-TICs for 1 hour, and then the supernatant was collected to carry out the positive selection. Pool enrichment was monitored using FACS and subjected to cloning into Escherichia coli using TOPO TA Cloning Kit for sequencing (Invitrogen) to identify the aptamer candidates.

Flow cytometry to monitor aptamer binding

Each FITC-labeled aptamer candidate was incubated with TICs or non-TICs in binding buffer on ice for 30 minutes. The cells were washed three times with binding buffer, and the pellets with the bound sequences were resuspended in 200 μL binding buffer. The fluorescence intensity was determined with a LSR II (BD Immunocytometry Systems) by counting 10,000 events. The FITC-labeled unselected ssDNA library was used as a negative control. To determine the specificity of the selected aptamers, aptamer binding to additional glioblastoma cells (08–322, 4121, U87MG) or neural progenitors (RenCell c-myc immortalized human neural progenitors, Millipore; human fetal neural progenitors, Lonza) was determined. For determination of aptamer binding to differentiated TICs, TICs were plated on Geltrex-coated plates in the presence of 10% serum and harvested after 5 days for FACS analysis.

Determination of aptamer affinity for TICs

To determine the binding affinity of the aptamers, 1 × 106 target cells were incubated with varying concentrations of FITC-labeled aptamer or library and analyzed by flow cytometry. The mean fluorescence intensity of the unselected library was subtracted from that of the aptamer with the target cells to determine the specific binding of the labeled aptamer. The apparent equilibrium Kd of the aptamer–cell interaction was obtained by fitting the dependence of intensity of specific binding on the concentration of the aptamers to the equation Y = BmaxX/(Kd + X).

Isolation and in vitro characterization of aptamerhigh and aptamerlow glioblastoma cells

To validate the specificity of the selected aptamers, we interrogated key TIC functional phenotypes in minimally cultured and unselected bulk tumors from either xenografts or patient surgical resected tumors. After overnight recovery of glioblastoma cells from tissue dissociation, each Cy5-labeled aptamer was incubated with each specimen for 30 minutes. After washing three times, labeled glioblastoma cells were subjected to cell sorting. For the sphere formation assay, 10 cells of either aptamerhigh and aptamerlow glioblastoma cells sorted by FACS were directly plated into each well of a 96-well plate containing 200 μL of TIC-culturing medium similar to our prior descriptions. After 10 to 14 days, the number of tumorspheres in each well was counted and plotted. For the cell proliferation assay, aptamerhigh and aptamerlow glioblastoma cells were sorted into each 15 mL tube, and equal numbers of cells were plated into 96 wells. Differences in cell growth were measured using CellTiter-Glo (Promega) according to the manufacturer's instructions and our prior reports (16, 19, 31, 34). GraphPad Prism software was used to complete t test comparisons of aptamerhigh and aptamerlow values for individual aptamers.

In vivo tumorigenic potential

To determine tumorigenic potential, 300 or 3,000 viable aptamerhigh or aptamerlow glioblastoma cells were intracranially implanted into 4- to 6-week-old athymic/nude immunocompromised mice obtained from Charles River Laboratories, as we have previously described. Animals were maintained until the development of neurologic signs including lethargy, seizure, ataxia, or paralysis or for a maximum of 60 days, when they were sacrificed. All animal procedures were carried out in accordance with a Cleveland Clinic (Cleveland, OH) Institutional Animal Care and Use Committee–approved protocol. Kaplan–Meier curves and log-rank statistical analysis were completed with GraphPad Prism software.

SELEX screening against TICs enriched for a pool of aptamers

SELEX is a method of repeated rounds of in vitro selection used to identify DNA or RNA aptamers, which bind specific targets, including cells (29, 30). As the identification and targeting of TICs has become a focus of some therapeutic-targeting strategies (35), we hypothesized that Cell-SELEX could selectively distinguish TICs from normal neural progenitors and differentiated tumor cells (non-TICs). We developed a two-stage approach with an initial discovery phase that used minimally cultured functionally validated TICs as a positive selector and functionally validated non-TICs and normal neural progenitors in negative selection. A validation phase described later involved testing the use of aptamers on minimally cultured and unselected bulk tumors.

To identify TIC-specific aptamers, a library of randomized DNA sequences was exposed to glioblastoma cells that fulfilled functional criteria of TICs or non-TICs isolated from pools of xenografts originally generated from a human glioblastoma specimen (Fig. 1A). Aptamers were first exposed to non-TICs or non-neoplastic human neural progenitor cells (NPC) as a counter selection to eliminate sequences not specific for TICs. DNA sequences that did not bind NPCs or non-TICs were incubated with TICs to select for binding aptamers as a positive selection (Fig. 1A). To select highly specific aptamers to TICs with high affinity, the number of NPCs and non-TICs was increased, whereas the number of TICs was decreased in the subsequent rounds of enrichment (Fig. 1B). In addition, concentration of competitors was increased, whereas the reaction time decreased to further enrich high-affinity TIC-specific aptamers (Fig. 1B). Eight rounds of selection enriched a pool of DNA sequences preferentially bound to TICs in comparison with non-TICs and NPCs as determined via flow cytometry with fluorescently labeled aptamers (Fig. 1C). Having confirmed the ability of the enriched DNA pool to recognize TICs, we then cloned and sequenced the enriched DNA pool to identify individual TIC-specific aptamer candidates. We obtained nine DNA sequences that were repeatedly detected in multiple colonies of bacteria out of 96 colonies analyzed for sequencing. The most frequent sequence domains repeated 8 to 27 times (Fig. 1D; Supplementary Table S1).

Figure 1.

TIC-specific aptamer candidates were identified from Cell-SELEX. A, to develop TIC-specific aptamers, glioblastoma cells were obtained from subcutaneous xenografts initiated from a glioblastoma patient specimen. To obtain enough cells for experiments, cells dissociated from multiple xenografts of GBM1 (08–387) were pooled. TICs and non-TICs prospectively sorted from these xenografts using CD133 were used for positive or counter selection, respectively. To enhance the cancer-specific activity of selected aptamers, human NPCs were also used for counter selection. B, as the number of rounds of enrichment increased, more stringent selection conditions were applied to increase the specificity and affinity of the TIC-specific aptamers identified. These conditions included increasing the number of non-TICs, decreasing the number of TICs, increasing the concentration of competitor, and decreasing the incubation period. C, FACS analysis confirmed that a pool of aptamers present after 8 rounds of selection was enriched with DNA sequences, which bind to TICs but have minimal binding to non-TICs or normal neural progenitors. D, DNA sequencing of the DNA sequences present in the pool of aptamers enriched after selection revealed nine different sequences with more than two copy numbers from 96 clones.

Figure 1.

TIC-specific aptamer candidates were identified from Cell-SELEX. A, to develop TIC-specific aptamers, glioblastoma cells were obtained from subcutaneous xenografts initiated from a glioblastoma patient specimen. To obtain enough cells for experiments, cells dissociated from multiple xenografts of GBM1 (08–387) were pooled. TICs and non-TICs prospectively sorted from these xenografts using CD133 were used for positive or counter selection, respectively. To enhance the cancer-specific activity of selected aptamers, human NPCs were also used for counter selection. B, as the number of rounds of enrichment increased, more stringent selection conditions were applied to increase the specificity and affinity of the TIC-specific aptamers identified. These conditions included increasing the number of non-TICs, decreasing the number of TICs, increasing the concentration of competitor, and decreasing the incubation period. C, FACS analysis confirmed that a pool of aptamers present after 8 rounds of selection was enriched with DNA sequences, which bind to TICs but have minimal binding to non-TICs or normal neural progenitors. D, DNA sequencing of the DNA sequences present in the pool of aptamers enriched after selection revealed nine different sequences with more than two copy numbers from 96 clones.

Close modal

Individual aptamers bind TICs

To confirm the ability of individual aptamer candidates to specifically bind TICs, flow cytometry was conducted with FITC-labeled individual aptamer candidates (Fig. 1D; sequences A1–A6) and compared with signal obtained from FITC-labeled library containing randomized DNA sequences as a negative control. A positive shift along the FITC axis was observed with each TIC-enriched aptamer compared with the library when incubated with TICs isolated from GBM1 (Supplementary Fig. S1A). In contrast, negligible shifts in the histogram occurred with non-TICs isolated from the same tumor (Supplementary Fig. S1B) or TICs differentiated with FBS (Supplementary Fig. S1C). Similar preferences for aptamer binding to TICs in comparison with matched non-TICs was observed for cells derived from two additional glioblastoma xenografts for the majority of the aptamers tested (Supplementary Fig. S2A–S2D). TIC-enriched aptamers also failed to bind to the well-characterized U87MG glioblastoma cell line (Supplementary Fig. S2E), which was passaged under prodifferentiating conditions in serum. To further characterize the ability of the aptamers to bind non-neoplastic cells in the brain, flow cytometry was also conducted with ReNCell CX immortalized human neural progenitors (Supplementary Fig. S1D) or primary human fetal neural progenitors (Supplementary Fig. S1E). The majority of the TIC-enriched aptamers showed minimal binding to non-neoplastic neural progenitors, although A4 bound both sets of non-neoplastic brain cells, indicating some of the aptamers could more broadly recognize neural progenitors (Supplementary Fig. S2D and S2E).

To further evaluate the specificity of TIC-enriched aptamers, the fold change in fluorescence was calculated. Fluorescence shift was calculated using the equation (FaptamerFcell)/(FlibraryFcell), where Faptamer, Flibrary, and Fcell refer to the fluorescence value of the aptamer, library, and cell itself. Using this calculation, specificity of aptamers A2–A6 for TICs in comparison with non-TICs was shown for all three glioblastomas tested (Fig. 2A–C). A1 also showed greater binding to TICs in GBM1 and GBM2, but not GBM3 (Fig. 2A–C). The affinity of A1–A6 binding to TICs was also very high with Kd in the nmol/L range (Fig. 2D). Together, these data strongly suggest that we have identified several aptamers that specifically bind to TICs with high affinity. From antibody and minibody studies, it is apparent that optimal targeting of tumor cells with cell killing payloads (toxin or radioisotope) is achieved when specificity of binding to the target cells is also associated with internalization. Using matched TICs and non-TICs, we found that aptamers bind with specificity to the cell surface and also internalized specifically in TICs (Fig. 3A and Supplementary Fig. S3), suggesting their potential to target TICs.

Figure 2.

Aptamer candidates bind with high affinity to TICs. FACS analysis was conducted using fluorescent aptamers with TICs, non-TICs, and differentiated TICs isolated from GBM1 (08–387) [A], GBM2 (4121) [B], or GBM3 (08–322) [C] xenografts. Fluorescence shift was calculated using the equation (FaptamerFcell)/(FlibraryFcell). Faptamer, Flibrary, and Fcell refer to the fluorescence value with the aptamer, library, and cell itself. D, the equilibrium Kd of the aptamer–cell interaction was obtained by fitting the dependence of intensity of specific binding on the concentration of the aptamers to the equation: Y = BmaxX/(Kd + X).

Figure 2.

Aptamer candidates bind with high affinity to TICs. FACS analysis was conducted using fluorescent aptamers with TICs, non-TICs, and differentiated TICs isolated from GBM1 (08–387) [A], GBM2 (4121) [B], or GBM3 (08–322) [C] xenografts. Fluorescence shift was calculated using the equation (FaptamerFcell)/(FlibraryFcell). Faptamer, Flibrary, and Fcell refer to the fluorescence value with the aptamer, library, and cell itself. D, the equilibrium Kd of the aptamer–cell interaction was obtained by fitting the dependence of intensity of specific binding on the concentration of the aptamers to the equation: Y = BmaxX/(Kd + X).

Close modal
Figure 3.

TIC-specific aptamers are internalized into TICs and identify stem-like tumor cells. A, matched TICs and non-TICs were exposed to fluorescently labeled aptamer A3 (Cy3) and imaged. TICs from GBM1 specifically internalized the aptamer, whereas non-TICs did not. Quantification is provided in Supplementary Fig. S3. B, bulk, minimally cultured glioblastoma cells were derived from a transiently passaged human glioblastoma specimen (CW1553 at passage 0) and exposed to FITC-labeled aptamer A4. FACS-sorted aptamerhigh and aptamerlow cells were stained for markers of stem cells (SOX2) or differentiated cells (GFAP).

Figure 3.

TIC-specific aptamers are internalized into TICs and identify stem-like tumor cells. A, matched TICs and non-TICs were exposed to fluorescently labeled aptamer A3 (Cy3) and imaged. TICs from GBM1 specifically internalized the aptamer, whereas non-TICs did not. Quantification is provided in Supplementary Fig. S3. B, bulk, minimally cultured glioblastoma cells were derived from a transiently passaged human glioblastoma specimen (CW1553 at passage 0) and exposed to FITC-labeled aptamer A4. FACS-sorted aptamerhigh and aptamerlow cells were stained for markers of stem cells (SOX2) or differentiated cells (GFAP).

Close modal

TIC-specific aptamers enrich glioblastoma cells with increased growth and tumorsphere formation capacity

Having identified aptamers that bound functional TICs preferentially, we conducted validation studies to determine the ability of the aptamers to select TICs from unselected bulk xenografts and patient specimens. We found that cells that bound the aptamer with high affinity (aptamerhigh) expressed high levels of stem cell markers such as Sox2, whereas cells with low aptamer (aptamerlow) binding expressed differentiation markers including glial fibrillary acidic protein (GFAP; Fig. 3B). As TICs are functionally defined, we examined the ability of the TIC-enriched aptamers to identify cells that could proliferate and self-renew. In cell titer assays, each of the five aptamerhigh fractions (A1–A5) showed significantly enhanced cell growth over time in the three glioblastomas tested (Fig. 4).

Figure 4.

Aptamerhigh cells are enriched for cell growth in vitro. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387) [A], GBM2 (4121) [B], and GBM4 (3691) [C] xenografts. Cells were plated at equivalent density and growth measured over time using the CellTiter-Glo assay. *, P < 0.05. Error bars represent SD.

Figure 4.

Aptamerhigh cells are enriched for cell growth in vitro. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387) [A], GBM2 (4121) [B], and GBM4 (3691) [C] xenografts. Cells were plated at equivalent density and growth measured over time using the CellTiter-Glo assay. *, P < 0.05. Error bars represent SD.

Close modal

As neurosphere formation assay has been used as a surrogate measure of self-renewal, we used this assay to determine if TIC-binding aptamers could enrich glioblastoma cells with self-renewing capacity. We found that A2, A3, and A5 significantly segregated cells with neurosphere formation capacity in all glioblastomas tested (Fig. 5). A1- and A4-enriched cells also showed greater neurosphere formation capacity in the majority of glioblastomas tested (Fig. 5). We confirmed these studies in three xenografts propagated without cell culture. When we dissociated these xenografts and segregated cells based on aptamer binding without culture, aptamerhigh cells were enriched for neurosphere formation capacity (Fig. 5E–G). These data show that TIC-enriched aptamers can be used to segregate fractions of glioblastoma cells with TIC properties as measured in vitro.

Figure 5.

Aptamerhigh cells are enriched for neurosphere formation capacity. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387) [A], GBM2 (4121) [B], GBM3 (08–322) [C], and GBM4 (3691) [D] xenografts, and cells plated for neurosphere formation assays and the number of spheres per well is shown. In separate studies, xenografts maintained exclusively in vivo [GBM10 (E), GBM12 (F), and GBM14 (G)] were dissociated without culture and separated into aptamer A4high and aptamer A4low cells to confirm results from other tumors. The percentage of wells with neurospheres is shown (E and F) or an example LDA (G).

Figure 5.

Aptamerhigh cells are enriched for neurosphere formation capacity. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387) [A], GBM2 (4121) [B], GBM3 (08–322) [C], and GBM4 (3691) [D] xenografts, and cells plated for neurosphere formation assays and the number of spheres per well is shown. In separate studies, xenografts maintained exclusively in vivo [GBM10 (E), GBM12 (F), and GBM14 (G)] were dissociated without culture and separated into aptamer A4high and aptamer A4low cells to confirm results from other tumors. The percentage of wells with neurospheres is shown (E and F) or an example LDA (G).

Close modal

Aptamerhigh cells are enriched for tumorigenic potential

As TICs by definition must be able to propagate the parental tumor in vivo, we determined whether the TIC-specific aptamers could segregate for differences in tumorigenic potential. Using Cy5-labeled A3 (Fig. 6A and B) and A4 (Fig. 6A and B) as probes, aptamerhigh and aptamerlow fractions were sorted from bulk cells isolated from GBM1 xenografts. Three thousand (Fig. 6A and C) or 300 (Fig. 6B and D) live aptamerhigh and aptamerlow cells were implanted into the brains of immunocompromised mice, which were then monitored daily for neurologic signs. We sacrificed a small cohort of the group as well as animals implanted with cells from an additional xenograft (GBM43) upon the initial development of neurologic signs and compared tumor size. The aptamerhigh cells displayed substantially greater tumor volumes than the aptamerlow cells (Fig. 6E). We further noted that aptamerhigh cells displayed greater propensity to invade into normal brain (Supplementary Fig. S4). These differences translated into survival differences as the survival of mice implanted with either A3 or A4 aptamerhigh cells was significantly reduced with decreased mean survival (Fig. 6F). All mice implanted with the lower number of aptamerhigh cells (300) developed tumors, whereas 3 of 4 tumors developed with A3 aptamerlow cells and only 2 of 4 tumors developed with A4 aptamerlow cells (Fig. 6F). These data suggest that TIC-enriched aptamers can segregate for glioblastoma cells with tumorigenic potential.

Figure 6.

Aptamerhigh cells are enriched for tumorigenic potential in comparison with aptamerlow cells. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387) using A3 (A and B) or A4 (C and D). Three thousand or 300 of matched aptamerhigh or aptamerlow cells were isolated and transplanted into mouse brains intracranially and their neurologic signs were monitored and Kaplan–Meier survival curves are shown (A–D). E, a cohort of tumor-bearing mice was sacrificed at the onset of tumor signs and compared in tumor volume with hematoxylin and eosin staining. F, mean survival as well as tumor incidence are displayed. *, P < 0.05 with log-rank analysis of Kaplan-Meier survival curves.

Figure 6.

Aptamerhigh cells are enriched for tumorigenic potential in comparison with aptamerlow cells. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387) using A3 (A and B) or A4 (C and D). Three thousand or 300 of matched aptamerhigh or aptamerlow cells were isolated and transplanted into mouse brains intracranially and their neurologic signs were monitored and Kaplan–Meier survival curves are shown (A–D). E, a cohort of tumor-bearing mice was sacrificed at the onset of tumor signs and compared in tumor volume with hematoxylin and eosin staining. F, mean survival as well as tumor incidence are displayed. *, P < 0.05 with log-rank analysis of Kaplan-Meier survival curves.

Close modal

Principal component analysis shows aptamerhigh and aptamerlow cells are distinct

Having shown that the TIC-enriched aptamers could segregate for functional measures of TIC biology, we sought to determine if there were similarities in the molecular profile of aptamerhigh or aptamerlow cells as determined via gene expression and principal component analysis (PCA). PCA is a multivariate analysis resulting in a three-dimensional visual representation of the data useful for identifying similarities within large datasets. We used FACS analysis with fluorescently labeled aptamers (A1–A5) in GBM1, GBM2, and GBM4 xenografts to segregate aptamerhigh and aptamerlow cells. Microarray data from the aptamerhigh and aptamerlow populations for each of the individual aptamers in each of the cell types (total of 15 datasets with matched aptamerhigh and aptamerlow cells) was then processed via PCA. The resulting analysis showed two distinct groups (shown as ellipsoids) corresponding aptamerhigh and aptamerlow datasets (Fig. 7A). The PCA analysis therefore indicates the data from the aptamerhigh and aptamerlow cells is distinct showing different molecular profiles between the groups.

Figure 7.

Microarray analysis reveals clustering of aptamerhigh and aptamerlow cells. A, PCA revealed clustering of aptamerhigh and aptamerlow cells. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387), GBM2 (4121), and GBM4 (3691) xenografts using fluorescently labeled aptamers A1, A2, A3, A4, or A5. The data for the aptamerhigh fractions from all xenografts were pooled and compared with data for the corresponding group of aptamerlow cells. PCA comparing these aptamerhigh and aptamerlow groups recognized two distinct groups of data as indicated by the ellipsoids. B and C, aptamerA3high cells have a differential expression profile associated with downregulation of genes involved in cellular development. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387) [A], GBM2 (4121) [B], and GBM4 (3691) [C] xenografts using fluorescently labeled A3. B, harvested RNA was used for microarray analysis on an Affymetrix platform and Ingenuity software was used to identify genes commonly differentially expressed between A3high and A3low cells, which are shown. The genes differentially expressed in the same direction across all glioblastomas tested were then used to create a database, which was analyzed in Ingenuity to identify the top network containing these targets. C, the cellular development, hematologic system development and function, and hematopoiesis network, which was the top hit, are shown.

Figure 7.

Microarray analysis reveals clustering of aptamerhigh and aptamerlow cells. A, PCA revealed clustering of aptamerhigh and aptamerlow cells. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387), GBM2 (4121), and GBM4 (3691) xenografts using fluorescently labeled aptamers A1, A2, A3, A4, or A5. The data for the aptamerhigh fractions from all xenografts were pooled and compared with data for the corresponding group of aptamerlow cells. PCA comparing these aptamerhigh and aptamerlow groups recognized two distinct groups of data as indicated by the ellipsoids. B and C, aptamerA3high cells have a differential expression profile associated with downregulation of genes involved in cellular development. FACS analysis was used to segregate aptamerhigh and aptamerlow cells from GBM1 (08–387) [A], GBM2 (4121) [B], and GBM4 (3691) [C] xenografts using fluorescently labeled A3. B, harvested RNA was used for microarray analysis on an Affymetrix platform and Ingenuity software was used to identify genes commonly differentially expressed between A3high and A3low cells, which are shown. The genes differentially expressed in the same direction across all glioblastomas tested were then used to create a database, which was analyzed in Ingenuity to identify the top network containing these targets. C, the cellular development, hematologic system development and function, and hematopoiesis network, which was the top hit, are shown.

Close modal

Aptamerhigh cells are enriched for molecular signatures associated with the regulation of cellular development

To identify specific molecular differences between aptamerhigh and aptamerlow cells, microarray data from GBM1, GBM2, and GBM4 cells segregated with A3 (Fig. 7B and C) or A4 (Supplementary Fig. S5) fluorescently labeled aptamers was further analyzed. Comparison of genes differentially elevated or repressed in A3high cells in comparison with A3low cells, revealed a set of 27 molecules that were commonly altered across all three lines (Fig. 7B). Only one gene transcript, chemokine (C–C motif) ligand 2 (CCL2) was consistently elevated, whereas the expression of the remaining 26 targets were repressed in A3high cells (Fig. 7B). A similar analysis of A4high cells in comparison with A4low cells, identified a set of 25 targets whose expression levels were commonly altered across all three glioblastoma cultures tested (Supplementary Fig. S5A). DEAD box polypeptide 60 (DDX60) was commonly elevated, whereas expression in the remaining targets was repressed in A4high cells (Supplementary Fig. S5A). When the A3 or A4 gene expression profiles were analyzed using Ingenuity software, the top biologic networks represented by these profiles were both found to regulate cellular development (Fig. 7C, Supplementary Fig. S5B). In fact, several genes were commonly downregulated in the A3high- and A4high-enriched cells (Fig. 7B and Supplementary Fig. S5A), suggesting that independent TIC-specific aptamers can select a subfraction of glioblastoma cells with common phenotypes and some similarities in their expression profiles.

Glioblastomas rank among the most lethal of human cancers with median survival of only 12 to 15 months despite surgical resection followed by radio- and chemotherapy with the DNA-alkylating agent temozolomide. The highly infiltrative and invasive feature of glioblastomas hinders complete resection of most tumors, and recurrence is nearly universal (8). TICs are thought to be the portion of glioblastoma cells that are responsible for both cellular invasion and therapeutic resistance (9–12), indicating the need for therapies that kill these cells and for imaging reagents that can identify them in vivo. To identify TIC-specific aptamers, we used a two-stage approach with initial selection using functionally validated TICs followed by independent validation in bulk tumor models with minimal culture to avoid drift. We believe that the TIC-specific aptamers identified in this study may help to fill these needs by providing methods to better target TICs for detection and treatment.

One method through which the TIC-enriched aptamers identified here may be developed for clinical use is through conjugation to increase the affinity and specificity of therapeutic nanoparticles for subgroups of cancer cells that drive tumor maintenance. Nanoparticle technology has already been approved by the U.S. Food and Drug Administration to treat patients with breast cancer in the form of Nab-paclitaxel, and paclitaxel-loaded nanoparticles showed efficacy against glioblastoma cell lines (36). Paclitaxel-loaded nanoparticles cross the blood–brain barrier (37, 38) and reduce glioblastoma growth in animal models (36, 38), suggesting the potential of enhancing the efficacy of these reagents through conjugation to TIC-specific aptamers. Aptamer–drug conjugates for delivery of doxorubicin (39) have recently been shown to have efficacy with liposomal-mediated delivery in glioblastoma animal models (40). Postoperative/chemoradiotherapy with doxorubicin liposomes was well tolerated in phase II clinical trials with liposomal delivery, but did not have significant improvement in outcome over historical averages (41). Directed delivery to TICs through conjugating to TIC-specific aptamers may generate more promising results. These data suggest that TIC-targeting aptamers conjugated to nanoparticles may provide additional benefit for glioblastoma therapies.

In addition to therapeutic use, labeled TIC-specific aptamers may be useful as imaging reagents. In glioblastoma, tumor recurrence most often occurs within 2 cm of the primary resection site, indicating the difficulty of complete resection of the tumor cells due to their infiltrating behavior. The ability to visualize invasive cells during surgery could therefore provide significant benefit. Fluorescent dyes encapsulated in nanoparticles have shown some success for crossing the blood–brain barrier to serve as imaging reagents in glioma, and quantum dot-labeled nanoparticles with aptamers targeting tenascin-C on human glioma cells have shown the ability to image glioma cells in vitro (42). Recently Hwang do and colleagues showed the ability to use MRI, fluorescent imaging, and radionucleotide imaging together to visualize tumors after treatment of mice bearing C6 glioma xenografts with radiolabeled nanoparticles containing rhodamine and a cancer-targeting aptamer (against nucleolin; ref. 43). Although intracranial tumors were not imaged, the data suggest the use of aptamer-conjugated nanoparticles as promising for imaging reagents.

In addition to the ability to either identify or target TICs, aptamers could have many potential applications for brain tumor biology, which have not yet been fully explored. For example, aptamers could be identified that specifically recognize proneural, mesenchymal, or classical glioblastoma subtypes, allowing for rapid profiling of patient biopsies. Aptamers that recognize glioblastoma cells in specific regions of the tumor associated with therapeutic resistance, such as the hypoxic niche, could also be identified and be useful for directing therapies to these difficult to target areas. Even aptamers showing specificity for non-neoplastic neural stem cells can be useful as these cells are being considered for delivery of antiglioma therapies.

TIC-specific aptamers also serve as a powerful tool to delineate critical molecular mechanisms in glioblastoma. We have used microarray analysis to determine a set of genes that are differentially regulated in aptamerhigh cells. These studies identified common targets that were involved in pathways associated with the regulation of cellular development and have recognized roles in cancer. For example, CCL2 that was elevated in the aptamerhigh cells is implicated in an interleukin-6–dependent paracrine loop to promote cellular invasion in vivo (44) and has been linked to the promotion of a cancer stem cell phenotype through IFN-regulatory factor 7 (45). Neutralizing antibodies against CCL2 have also shown efficacy in xenografts models as an antiglioma therapy (46). Hexamethylene bisacetamide-inducible protein 1 (HEXIM1) promotes the stability of p53 (47) and inhibits metastasis (48), suggesting that the reduced expression levels present in aptamerhigh cells would promote tumorigenesis. Growth arrest and DNA damage-inducible 45b (GADD45b) is a cellular stress sensor, which is suggested to be a tumor suppressor (49), confirming it would be beneficial for cancer cells to have decreased GADD45B expression as shown in our study. Furthermore, GADD45b has been shown to be overexpressed in non-side population cells depleted for cancer stem cell characteristics in embryonic carcinoma cells (50) consistent with the notion it is enriched in TICs. Thus, our molecular analysis suggests TICs identified with specific aptamers have gene expression profiles, which could promote tumorigenesis.

These studies represent an initial development of tools that may inform analytic techniques for molecular pathways involved in the tumor hierarchy and growth as well as therapeutic targeting. We are currently investigating the expression levels and possible functional contributions of the genetic targets identified in the gene expression studies through comparison of TICs and non-TICs in different glioblastoma subgroups with targeting of expression in these models. It is notable that five different aptamers with completely different sequences identified tumor cell populations with very tightly clustered gene expression programs, suggesting the identification of a convergent tumor population (Fig. 7A). This is an important biologic finding as the widely diverse set of putative TIC markers in brain cancer, such as other cancers, has not clearly defined a common population. In addition, preliminary studies suggest that our TICs can be bound to nanoparticles to act as delivery vehicles into the brain (J. Heddleston, Q. Wu, A. Hjelmeland and J. Rich; unpublished data). Collectively, these results suggest that aptamer technology reinforces the presence of a phenotypically similar stem-like tumor population that may be selectively identified and–in future studies–targeted. Aptamer-based therapies may be designed for combinations with either conventional therapeutics or targeted therapies to disrupt tumor subpopulations that contribute to therapeutic resistance, invasion, and angiogenesis.

No potential conflicts of interest were disclosed.

Conception and design: Y. Kim, A.B. Hjelmeland, J.N. Rich

Development of methodology: Y. Kim, M. Hitomi, A.B. Hjelmeland, J.N. Rich

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Kim, Q. Wu, P. Hamerlik, A.E. Sloan, G.H. Barnett, R.J. Weil, J.N. Rich

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Kim, P. Hamerlik, M. Hitomi, P. Leahy, A.B. Hjelmeland, J.N. Rich

Writing, review, and/or revision of the manuscript: Y. Kim, P. Hamerlik, M. Hitomi, A.E. Sloan, G.H. Barnett, R.J. Weil, A.B. Hjelmeland, J.N. Rich

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Kim, A.B. Hjelmeland, J.N. Rich

Study supervision: A.B. Hjelmeland, J.N. Rich

The authors thank Cleveland Clinic Foundation Tissue Procurement Service and S. Staugatis and M. McGraw for providing tissue. The authors also thank the assistance of the Flow Cytometry Core, including Cathy Shemo, Sage O'Bryant, and Moneen Morgan.

This study was financially supported by the NIH: CA129958, CA116659, and CA154130 (J.N. Rich); CA151522 (A.B. Hjelmeland); CA137443, NS063971, CA128269, CA101954, and CA116257 (A.E. Sloan); James S. McDonnell Foundation (J.N. Rich) and the Ohio Department of Development Tech 09–071 (A.E. Sloan); Melvin Burkhardt Chair in Neurosurgical Oncology and the Karen Colina Wilson Research Endowment within the Burkhardt Brain Tumor and Neuro-oncology Center at the Cleveland Clinic Foundation for additional support and funding (R.J. Weil). Y. Kim is supported by a post-doctoral fellowship from the American Brain Tumor Association. P. Hamerlik is supported by the Danish Council for Independent Research/Medical Sciences 1D4765/11-105457.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Hanahan
D
,
Weinberg
RA
. 
Hallmarks of cancer: the next generation
.
Cell
2011
;
144
:
646
74
.
2.
Reya
T
,
Morrison
SJ
,
Clarke
MF
,
Weissman
IL
. 
Stem cells, cancer, and cancer stem cells
.
Nature
2001
;
414
:
105
11
.
3.
Hjelmeland
AB
,
Lathia
JD
,
Sathornsumetee
S
,
Rich
JN
. 
Twisted tango: brain tumor neurovascular interactions
.
Nat Neurosci
2011
;
14
:
1375
81
.
4.
Galli
R
,
Binda
E
,
Orfanelli
U
,
Cipelletti
B
,
Gritti
A
,
De Vitis
S
, et al
Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma
.
Cancer Res
2004
;
64
:
7011
21
.
5.
Bao
S
,
Wu
Q
,
McLendon
RE
,
Hao
Y
,
Shi
Q
,
Hjelmeland
AB
, et al
Glioma stem cells promote radioresistance by preferential activation of the DNA damage response
.
Nature
2006
;
444
:
756
60
.
6.
Bao
S
,
Wu
Q
,
Sathornsumetee
S
,
Hao
Y
,
Li
Z
,
Hjelmeland
AB
, et al
Stem cell-like glioma cells promote tumor angiogenesis through vascular endothelial growth factor
.
Cancer Res
2006
;
66
:
7843
8
.
7.
Cheng
L
,
Wu
Q
,
Guryanova
OA
,
Huang
Z
,
Huang
Q
,
Rich
JN
, et al
Elevated invasive potential of glioblastoma stem cells
.
Biochem Biophys Res Commun
2011
;
406
:
643
8
.
8.
Dunn
GP
,
Rinne
ML
,
Wykosky
J
,
Genovese
G
,
Quayle
SN
,
Dunn
IF
, et al
Emerging insights into the molecular and cellular basis of glioblastoma
.
Genes Dev
2012
;
26
:
756
84
.
9.
Singh
SK
,
Hawkins
C
,
Clarke
ID
,
Squire
JA
,
Bayani
J
,
Hide
T
, et al
Identification of human brain tumour initiating cells
.
Nature
2004
;
432
:
396
401
.
10.
Hemmati
HD
,
Nakano
I
,
Lazareff
JA
,
Masterman-Smith
M
,
Geschwind
DH
,
Bronner-Fraser
M
, et al
Cancerous stem cells can arise from pediatric brain tumors
.
Proc Natl Acad Sci U S A
2003
;
100
:
15178
83
.
11.
Salmaggi
A
,
Boiardi
A
,
Gelati
M
,
Russo
A
,
Calatozzolo
C
,
Ciusani
E
, et al
Glioblastoma-derived tumorospheres identify a population of tumor stem-like cells with angiogenic potential and enhanced multidrug resistance phenotype
.
Glia
2006
;
54
:
850
60
.
12.
Singh
SK
,
Clarke
ID
,
Terasaki
M
,
Bonn
VE
,
Hawkins
C
,
Squire
J
, et al
Identification of a cancer stem cell in human brain tumors
.
Cancer Res
2003
;
63
:
5821
8
.
13.
Son
MJ
,
Woolard
K
,
Nam
DH
,
Lee
J
,
Fine
HA
. 
SSEA-1 is an enrichment marker for tumor-initiating cells in human glioblastoma
.
Cell Stem Cell
2009
;
4
:
440
52
.
14.
Anido
J
,
Saez-Borderias
A
,
Gonzalez-Junca
A
,
Rodon
L
,
Folch
G
,
Carmona
MA
, et al
TGF-beta receptor inhibitors target the CD44(high)/Id1(high) glioma-initiating cell population in human glioblastoma
.
Cancer Cell
2010
;
18
:
655
68
.
15.
Bao
S
,
Wu
Q
,
Li
Z
,
Sathornsumetee
S
,
Wang
H
,
McLendon
RE
, et al
Targeting cancer stem cells through L1CAM suppresses glioma growth
.
Cancer Res
2008
;
68
:
6043
8
.
16.
Lathia
JD
,
Gallagher
J
,
Heddleston
JM
,
Wang
J
,
Eyler
CE
,
Macswords
J
, et al
Integrin alpha 6 regulates glioblastoma stem cells
.
Cell Stem Cell
2010
;
6
:
421
32
.
17.
Inda
MM
,
Bonavia
R
,
Mukasa
A
,
Narita
Y
,
Sah
DW
,
Vandenberg
S
, et al
Tumor heterogeneity is an active process maintained by a mutant EGFR-induced cytokine circuit in glioblastoma
.
Genes Dev
2010
;
24
:
1731
45
.
18.
Jin
X
,
Yin
J
,
Kim
SH
,
Sohn
YW
,
Beck
S
,
Lim
YC
, et al
EGFR–AKT–Smad signaling promotes formation of glioma stem-like cells and tumor angiogenesis by ID3-driven cytokine induction
.
Cancer Res
2011
;
71
:
7125
34
.
19.
Kim
Y
,
Kim
E
,
Wu
Q
,
Guryanova
O
,
Hitomi
M
,
Lathia
JD
, et al
Platelet-derived growth factor receptors differentially inform intertumoral and intratumoral heterogeneity
.
Genes Dev
2012
;
26
:
1247
62
.
20.
Binda
E
,
Visioli
A
,
Giani
F
,
Lamorte
G
,
Copetti
M
,
Pitter
KL
, et al
The EphA2 receptor drives self-renewal and tumorigenicity in stem-like tumor-propagating cells from human glioblastomas
.
Cancer Cell
2012
;
22
:
765
80
.
21.
Broadley
KW
,
Hunn
MK
,
Farrand
KJ
,
Price
KM
,
Grasso
C
,
Miller
RJ
, et al
Side population is not necessary or sufficient for a cancer stem cell phenotype in glioblastoma multiforme
.
Stem Cells
2011
;
29
:
452
61
.
22.
Goidts
V
,
Bageritz
J
,
Puccio
L
,
Nakata
S
,
Zapatka
M
,
Barbus
S
, et al
RNAi screening in glioma stem-like cells identifies PFKFB4 as a key molecule important for cancer cell survival
.
Oncogene
2012
;
31
:
3235
43
.
23.
Hubert
CG
,
Bradley
RK
,
Ding
Y
,
Toledo
CM
,
Herman
J
,
Skutt-Kakaria
K
, et al
Genome-wide RNAi screens in human brain tumor isolates reveal a novel viability requirement for PHF5A
.
Genes Dev
2013
;
27
:
1032
45
.
24.
Sefah
K
,
Bae
KM
,
Phillips
JA
,
Siemann
DW
,
Su
Z
,
McClellan
S
, et al
Cell-based selection provides novel molecular probes for cancer stem cells
.
Int J Cancer
2013
;
132
:
2578
88
.
25.
Shigdar
S
,
Qiao
L
,
Zhou
S
,
Xiang
D
,
Wang
T
,
Li
Y
, et al
RNA aptamers targeting cancer stem cell marker CD133
.
Cancer Lett
2013
;
330
:
84
95
.
26.
de Franciscis
V
,
Esposito
CL
,
Catuogno
S
,
Cellai
L
,
Cerchia
L
. 
Aptamers as innovative diagnostic and therapeutic agents in the central nervous system
.
CNS Neurol Disord Drug Targets
2009
;
8
:
393
401
.
27.
Nimjee
SM
,
Rusconi
CP
,
Sullenger
BA
. 
Aptamers: an emerging class of therapeutics
.
Annu Rev Med
2005
;
56
:
555
83
.
28.
Ye
M
,
Hu
J
,
Peng
M
,
Liu
J
,
Liu
J
,
Liu
H
, et al
Generating aptamers by cell-SELEX for applications in molecular medicine
.
Int J Mol Sci
2012
;
13
:
3341
53
.
29.
Sefah
K
,
Shangguan
D
,
Xiong
X
,
O'Donoghue
MB
,
Tan
W
. 
Development of DNA aptamers using Cell-SELEX
.
Nat Protoc
2010
;
5
:
1169
85
.
30.
Kim
Y
,
Liu
C
,
Tan
W
. 
Aptamers generated by Cell SELEX for biomarker discovery
.
Biomarkers Med
2009
;
3
:
193
202
.
31.
Li
Z
,
Bao
S
,
Wu
Q
,
Wang
H
,
Eyler
C
,
Sathornsumetee
S
, et al
Hypoxia-inducible factors regulate tumorigenic capacity of glioma stem cells
.
Cancer Cell
2009
;
15
:
501
13
.
32.
Eyler
CE
,
Wu
Q
,
Yan
K
,
MacSwords
JM
,
Chandler-Militello
D
,
Misuraca
KL
, et al
Glioma stem cell proliferation and tumor growth are promoted by nitric oxide synthase-2
.
Cell
2011
;
146
:
53
66
.
33.
Giannini
C
,
Sarkaria
JN
,
Saito
A
,
Uhm
JH
,
Galanis
E
,
Carlson
BL
, et al
Patient tumor EGFR and PDGFRA gene amplifications retained in an invasive intracranial xenograft model of glioblastoma multiforme
.
Neurooncology
2005
;
7
:
164
76
.
34.
Hjelmeland
AB
,
Wu
Q
,
Wickman
S
,
Eyler
C
,
Heddleston
J
,
Shi
Q
, et al
Targeting A20 decreases glioma stem cell survival and tumor growth
.
PLoS Biol
2010
;
8
:
e1000319
.
35.
Zhou
BB
,
Zhang
H
,
Damelin
M
,
Geles
KG
,
Grindley
JC
,
Dirks
PB
. 
Tumour-initiating cells: challenges and opportunities for anticancer drug discovery
.
Nat Rev Drug Discov
2009
;
8
:
806
23
.
36.
Gao
H
,
Qian
J
,
Cao
S
,
Yang
Z
,
Pang
Z
,
Pan
S
, et al
Precise glioma targeting of and penetration by aptamer and peptide dual-functioned nanoparticles
.
Biomaterials
2012
;
33
:
5115
23
.
37.
Dilnawaz
F
,
Singh
A
,
Mewar
S
,
Sharma
U
,
Jagannathan
NR
,
Sahoo
SK
. 
The transport of non-surfactant based paclitaxel loaded magnetic nanoparticles across the blood brain barrier in a rat model
.
Biomaterials
2012
;
33
:
2936
51
.
38.
Guo
J
,
Gao
X
,
Su
L
,
Xia
H
,
Gu
G
,
Pang
Z
, et al
Aptamer-functionalized PEG-PLGA nanoparticles for enhanced anti-glioma drug delivery
.
Biomaterials
2011
;
32
:
8010
20
.
39.
Huang
YF
,
Shangguan
D
,
Liu
H
,
Phillips
JA
,
Zhang
X
,
Chen
Y
, et al
Molecular assembly of an aptamer-drug conjugate for targeted drug delivery to tumor cells
.
Chembiochem
2009
;
10
:
862
8
.
40.
Yang
T
,
Wang
Y
,
Li
Z
,
Dai
W
,
Yin
J
,
Liang
L
, et al
Targeted delivery of a combination therapy consisting of combretastatin A4 and low-dose doxorubicin against tumor neovasculature
.
Nanomedicine
2012
;
8
:
81
92
.
41.
Beier
CP
,
Schmid
C
,
Gorlia
T
,
Kleinletzenberger
C
,
Beier
D
,
Grauer
O
, et al
RNOP-09: pegylated liposomal doxorubicine and prolonged temozolomide in addition to radiotherapy in newly diagnosed glioblastoma—a phase II study
.
BMC Cancer
2009
;
9
:
308
.
42.
Chen
XC
,
Deng
YL
,
Lin
Y
,
Pang
DW
,
Qing
H
,
Qu
F
, et al
Quantum dot-labeled aptamer nanoprobes specifically targeting glioma cells
.
Nanotechnology
2008
;
19
:
235105
.
43.
Hwang do
W
,
Ko
HY
,
Lee
JH
,
Kang
H
,
Ryu
SH
,
Song
IC
, et al
A nucleolin-targeted multimodal nanoparticle imaging probe for tracking cancer cells using an aptamer
.
J Nuclear Med
2010
;
51
:
98
105
.
44.
Zhang
J
,
Sarkar
S
,
Cua
R
,
Zhou
Y
,
Hader
W
,
Yong
VW
. 
A dialog between glioma and microglia that promotes tumor invasiveness through the CCL2/CCR2/interleukin-6 axis
.
Carcinogenesis
2012
;
33
:
312
9
.
45.
Jin
X
,
Kim
SH
,
Jeon
HM
,
Beck
S
,
Sohn
YW
,
Yin
J
, et al
Interferon regulatory factor 7 regulates glioma stem cells via interleukin-6 and Notch signalling
.
Brain
2012
;
135
:
1055
69
.
46.
Zhu
X
,
Fujita
M
,
Snyder
LA
,
Okada
H
. 
Systemic delivery of neutralizing antibody targeting CCL2 for glioma therapy
.
J Neurooncol
2011
;
104
:
83
92
.
47.
Lew
QJ
,
Chia
YL
,
Chu
KL
,
Lam
YT
,
Gurumurthy
M
,
Xu
S
, et al
Identification of HEXIM1 as a positive regulator of p53
.
J Biol Chem
2012
;
287
:
36443
54
.
48.
Ketchart
W
,
Smith
KM
,
Krupka
T
,
Wittmann
BM
,
Hu
Y
,
Rayman
PA
, et al
Inhibition of metastasis by HEXIM1 through effects on cell invasion and angiogenesis
.
Oncogene.
2012 Sep 10
. [Epub ahead of print].
49.
Ying
J
,
Srivastava
G
,
Hsieh
WS
,
Gao
Z
,
Murray
P
,
Liao
SK
, et al
The stress-responsive gene GADD45G is a functional tumor suppressor, with its response to environmental stresses frequently disrupted epigenetically in multiple tumors
.
Clin Cancer Res
2005
;
11
:
6442
9
.
50.
Inowa
T
,
Hishikawa
K
,
Matsuzaki
Y
,
Isagawa
T
,
Takeuchi
T
,
Aburatani
H
, et al
GADD45beta determines chemoresistance and invasive growth of side population cells of human embryonic carcinoma
.
Stem Cells Int
2010
;
2010
:
782967
.