| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
CANCER STEM CELLS |
aMaxine Dunitz Neurosurgical Institute, Los Angeles, California, USA;
bDepartment of Surgery, Cedars-Sinai Medical Center, Los Angeles, California, USA
Key Words. Spheres • Monolayer • Mitogens • Cancer stem-like cells • Glioma
Correspondence: John S. Yu, M.D., 8631 West Third Street, Suite 800E, Los Angeles, California 90048, USA. Telephone: 310-423-0845; Fax: 310-423-0810; e-mail: Yuj{at}cshs.org
Received October 2, 2006;
accepted for publication March 29, 2007.
First published online in STEM CELLS EXPRESS April 5, 2007.
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
Disclosure of potential conflicts of interest is found at the end of this article.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
At present, there is no animal brain tumor model that exactly simulates a glioma. However, of the rat models that are available for use, such as 9L, C6, T9, and F98, in vivo and in vitro properties of brain tumors and their responses to therapy have been previously demonstrated. Both the 9L and C6 cell lines have been extensively used in a variety of tumor models [3]. In comparison, the C6 glioma has no syngeneic rat host strain available for experimentally induced tumors, making it immunogenic in allogeneic hosts [3]. The 9L gliosarcoma cell line, nonetheless, does possess its own limitations. When used in gene therapy and boron neutron capture therapy models, the 9L gliosarcoma is immunogenic, generating a tumor-associated immune response that may help in eliminating residual tumor left over after treatment [3, 4]. Nonetheless, the 9L glioma is still useful for a variety of other studies, including experiments involving glioma therapy protocols, and remains the most widely used glioma cell line for rat brain tumor models [3].
Stem cells have been defined as multipotent self-renewing progenitors with the potential to differentiate into multiple cell types [5–10]. Reynolds and Weiss [11, 12] were one of the first groups to introduce a system to identify neural stem cells in a defined medium, whereby striatal embryonic progenitors could be isolated, harvested, and grown in culture as undifferentiated neurospheres (clonally derived aggregates of cells derived from a single stem cell) under the influence of the mitogens epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF). Many of these cells expressed Nestin (an intermediate filament found in neuroepithelial stem cells) but not markers for the more differentiated principal neural cell types of the central nervous system (CNS)—neuronal and glial cells. However, when these neurospheres were allowed to differentiate by removing mitogens from the medium, many of the cells within the neurospheres differentiated into neurons and astrocytes. The isolated striatal cells fulfilled the critical features defined for neural stem cells: an unlimited capacity for self-renewal and the capacity to differentiate into principal mature neural cells [5–10]. Under similar environmental conditions, a subpopulation of tumor-derived cells has been suggested to have an ability to behave like endogenous stem cells and hence is referred to as "cancer stem cells."
The cancer stem cell hypothesis proposes that a subpopulation of tumor cells proliferate, self-renew, and eventually differentiate into a phenotypically diverse and heterogeneous tumor cell population [13]. In recent years, several groups have isolated brain tumor stem cells (BTSCs) from human brain tumors and demonstrated their ability to recapitulate tumors under environmental conditions that are prohibitive for non-BTSCs to repropagate tumors [14–18]. BTSCs have been characterized by their ability to form clonally derived neurospheres, self-renew, and recapitulate the tumor from which they were isolated. Furthermore, human glioma-derived multipotent BTSCs that express the neural stem cell surface marker CD133 have the capacity to self-renew, and when injected into adult mice brains at relatively low densities, they histologically recapitulated the original tumor [14, 16, 19, 20].
One hypothesis for glioma oncogenesis embraces the notion that mutational transformations in normal neural stem cells may be the origin of cancer stem cells, which leads to unchecked self-renewal and the evolution of CNS tumors. Several critical signaling pathways have been implicated to play a role in this process, including the Shh, Wnt, and EGF signaling pathways [17]. Another hypothesis suggests that brain tumors are a result of terminally differentiated cells dedifferentiating through a reprogramming process as a consequence of oncogenic mutations [14, 19]. Irrespective of the cause, the relatively small proportion of cancer stem cells that may exist within a tumor may explain the lack of therapeutic efficacy in the treatment of malignant brain tumors. Hence, even after gross resection, chemotherapy, or radiation therapy, the residual tumor cells that remain presumably contain cancer stem cells with the potential to form a lethal recurrent tumor. To better understand the potential of these cancer stem cells, we have used the 9L gliosarcoma rat tumor cell line for both in vitro and in vivo analyses to determine whether cancer stem cells can be derived from this cell line and whether they behave as stem cells under specific environmental conditions. Here, we characterize the experimentally induced 9L gliosarcoma rat tumor model and demonstrate that neurospheres generated from 9L gliosarcoma cells initially grown as a monolayer in vitro and then subsequently grown under serum-free conditions supplemented with EGF and bFGF mitogens possess a subpopulation of cells with phenotypic and functional characteristics of cancer stem-like cells (CSLCs). In addition, when implanted in vivo, these cells are more aggressive at generating tumors than their monolayer counterparts maintained in serum-containing medium devoid of mitogens. In addition, our data demonstrate that 9L neurospheres containing CSLCs are more chemoresistant than the 9L monolayer cells, suggesting that a more effective therapeutic strategy must be devised to combat the more aggressive chemoresistant CSLCs.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
Infection of 9L Gliosarcoma and Tumor Spheres with Luciferase Gene
9L gliosarcoma cells grown as a monolayer expressing the Luciferase gene were obtained from Dr. Rehemtulla from the University of Michigan. Luciferase-expressing tumor stem-like cells of 9L were derived from stable transfection of tumor stem-like cells isolated from 9L with a pcDNA3.1-luciferase vector that encodes the firefly luciferase gene by the calcium phosphate precipitation method as described previously [23]. The stable transfectants were selected with gentamicin-418. Transgene expression from the transduced tumor stem-like cells was confirmed by a luciferase assay with the IVIS Imaging System 200 Series (Xenogen Corporation, Hopkinton, MA, http://www.xenogen.com). We found no differences in phenotypic protein expression or proliferation rates for 9L luciferase cells compared with the parent 9L cell line (data not shown). For all experiments, we used the 9L luciferase cells except for the in vitro immunocytochemical data for Figure 1, in which we used the parental 9L cell line.
|
Immunocytochemistry Staining
To examine the expression of NSC markers and lineage markers, immuncyto- and immunohistochemical staining was performed. For staining of differentiated tumor spheres, tumor spheres, and 9L monolayers, cells growing in chamber slides were fixed with 4% paraformaldehyde for 15 minutes at 4°C, treated with 5% normal horse serum (NHS)/0.1% Triton X-100, and then immunostained with the following antibodies: rabbit anti-Nestin (1:200; Chemicon, Temecula, CA, http://www.chemicon.com), rabbit anti-Sox2 (1:1,000; Chemicon), rabbit anti-mitogen-activated protein 2 (anti-MAP2) (1:1,000; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), mouse anti-ß-tubulin III (1:200; Chemicon), rabbit anti-glial fibrillary acidic protein (anti-GFAP) (1:1,000; Chemicon), and mouse anti-myelin/oligodendrocyte marker (1:1,000; Chemicon). The primary antibodies were detected with Cy3 or fluorescein isothiocyanate (FITC)-conjugated anti-mouse or anti-rabbit IgG antibodies (1:200; Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com). The cells were counterstained with 4',6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) to identify all nuclei. The stained sections were examined and photographed using a QED cell scanner program and a Nikon Eclipse TE2000-E microscope (Nikon, Tokyo, http://www.nikon.com) and analyzed using ImageJ (NIH). For immunostaining of tumor spheres, the spheres where allowed to adhere to chamber slides (Nunc, Rochester, NY, http://www.nuncbrand.com) precoated with laminin (16 µg/ml) for 3 hours before fixation in 4% paraformaldehyde, whereas monolayers were allowed to adhere overnight on nonfixed chamber slides and subsequently fixed in 4% paraformaldehyde.
Implantation of Tumor into Syngeneic Fisher Rats
Fisher rats 6–8 weeks old (Harlan Sprague-Dawley, Indianapolis, http://www.harlan.com) were anesthetized with i.p. ketamine (80 mg/kg) and xylazine (10 mg/kg) and stereotactically implanted with isolated 9L tumor sphere cells containing the luciferase gene (5,000 cells) or non-sphere-forming monolayer cells containing the luciferase gene (5,000 cells) in a 4-µl volume of 1.2% methylcellulose/phosphate-buffered saline (PBS) into the right striatum. Rats were portioned to either the tumor volume group (n = 10, with five animals in both monolayer and tumor sphere groups) or the survival group (n = 18, with nine animals in both monolayer and tumor sphere groups), with control rats (n = 6) receiving 4 µl of 1.2% methylcellulose/PBS only. Because each group consisted of both monolayer-implanted and tumor sphere-implanted animals, tumor aggression could be determined. Animals in the tumor volume group were sacrificed 18 days after tumor implantation. Tumor volume was assessed by using the formula for an ellipsoid, (length x width x height)/2 [24], with the height and the width of the tumor being approximately equal because of the well-defined circumference of the tumors generated by the 9L gliosarcoma [25]. Average and SEM were determined for each group. Animals in the survival group were followed for survival and euthanized via CO2 asphyxiation when terminal neurological signs developed—inability to access food or water, seizure activity, weakness, and paralysis—or if animals exceeded a survival period of 40 days. The animals in the survival group underwent a luciferase scan 14 days after implantation, which was conducted 15 minutes after i.p. injection of 150 mg/kg of the substrate D-luciferin (Biosynth International, Inc., Naperville, IL, http://www.biosynth.com). Upon euthanization, brains were harvested and frozen in 2-methylbutane (Sigma-Aldrich) between –10°C and –20°C and stored at –80°C until sectioning as described below. All the animal experiments were performed in strict accordance with the Institutional Animal Care and Use Committee guidelines of Cedars-Sinai Medical Center.
Bioluminescent Imaging
Bioluminescent imaging was performed using a charge-coupled device cooled camera and Living Image software from Xenogen. In vitro imaging was performed by adding 20,000 bioluminescent cells of either monolayers or neurospheres to a 96-well plate in 0.2 ml of DMEM/Ham's F-12 medium with 10% FBS. Five minutes before imaging, 0.2 ml of a 150 µg/ml stock of D-luciferin (Biosynth) was added to each well to compare the intensity of the two cell populations. For in vivo imaging, D-luciferin was given i.p. at 150 mg/kg; animals were anesthetized with 3%–5% isoflurane for approximately 5 minutes and then placed inside the camera box (Xenogen) that maintained 1%–2% isoflurane for the duration of imaging. Images were taken 15 minutes after i.p. luciferin delivery.
Immunohistochemical Staining of Brain Sections
The tumor cell-implanted rat brains were cryostat-cut at 20 µm thickness in coronal sections, fixed in 4% paraformaldehyde for 30 minutes, washed with PBS, and air-dried. To characterize the brain tissue by immunohistochemistry, free-floating sections 20 µm thick were blocked with 5% NHS for 30 minutes at room temperature and then immunostained with the following antibodies: rabbit anti-Nestin (1:200; Chemicon), rabbit anti-Sox2 (1:1,000; Chemicon), rabbit anti-MAP2 (1:1,000; Sigma-Aldrich), mouse anti-ß-tubulin III (1:200; Chemicon), rabbit anti-GFAP (1:1,000; Chemicon), and mouse anti-myelin/oligodendrocyte (1:1,000; Chemicon). The primary antibodies were detected with secondary antibodies as described above. The cells were counterstained with DAPI (Vector Laboratories) to identify all nuclei. The stained sections were examined and photographed using the Zeiss Axiovision 3.1 program in conjunction with Zeiss Axioskop 2 microscope (Carl Zeiss, Jena, Germany, http://www.zeiss.com) and analyzed using ImageJ (NIH).
Flow Cytometry
To describe the differentiation patterns of tumor spheres, tumor spheres were differentiated for 2 weeks in Dulbecco's modified Eagle's medium/Ham's F-12 medium containing 10% FBS in 75 cm2 flasks. Cells were then removed using 0.1% trypsin and EDTA (Gibco-BRL). Differentiated tumor sphere cells, tumor sphere cells, and monolayer cells were fixed with 4% paraformaldehyde for 15 minutes at 4°C, treated with 5% NHS/0.1% Triton X-100, and then stained with the following antibodies: rabbit anti-Nestin (1:1,000; Chemicon), rabbit anti-Sox2 (1:1,000; Chemicon), rabbit anti-MAP2 (1:1,000; Sigma-Aldrich), mouse anti-ß-tubulin III (1:500; Chemicon), rabbit anti-GFAP (1:1,000; Chemicon), and mouse anti-myelin/oligodendrocyte (1:1,000; Chemicon). The primary antibodies were detected with FITC-conjugated anti-mouse or anti-rabbit IgG antibody (1:200; Jackson Immunoresearch) using a FACSVantage fluorescence-activated cell sorter (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com).
H&E and Reticulin Staining of Brain Sections
For H&E staining, 20-µm coronal brain sections described above were mounted on SuperFrost plus slides (Fisher Scientific International, Hampton, NH, http://www.fisherscientific.com), stained with Harris' hematoxylin for 2 minutes, and then counterstained with alcoholic eosin. Reticulin stains were performed on 12-µm coronal brain sections.
Proliferation and Chemoresistance Assay
To compare the differential proliferation rates between cells grown as a monolayer versus cells grown as tumor spheres and determine their resistance to chemotherapeutic agents, 2,000 healthy 9L tumor sphere and monolayer cells were exposed either to Dulbecco's modified Eagle's medium/Ham's F-12 medium containing 10% FBS or to 100 µM stock solution of temozolomide or carboplatin dissolved in PBS at concentrations of 1,000, 500, 250, and 125 µM for 2 days. The viability of the cells was scored by measurement of the absorption of formazan dye—the amount of formazan dye formed directly correlates to the number of metabolically active cells—using the cellular proliferation assay WST-1 (Roche Molecular Biochemicals, Mannheim, Germany, http://www.roche.com). Formazan was measured with the use of the plate reader (Tecan, San Jose, CA, http://www.tecan.com) and spectrophotometer set at a wavelength of 440 nm and a reference wavelength of 890 nm. Cellular viability was determined by exposing cells to WST-1 for 4 hours and calculating the percentage of viable cells. Proliferation was also assessed by using manual cell counting after 7 days in culture, with an initial cellular concentration of 100,000 cells per milliliter in a 25-mm2 flask.
RNA Isolation and cDNA Synthesis
Total RNA was extracted from 9L monolayer cells and 9L tumor sphere cells using an RNA4PCR kit (Ambion, Austin, TX, http://www.ambion.com) according to the manufacturer's protocol. For cDNA synthesis, 
1 µg of total RNA was reverse-transcribed into cDNA using oligo(dT) primer and iScript cDNA synthesis kit reverse transcriptase (Life Science Research, Hercules, CA, http://www.bio-rad.com). cDNA was stored at –20°C for polymerase chain reaction (PCR).
Real-time Quantitative Reverse Transcription-PCR
Gene expression was quantified by real-time quantitative reverse transcription-PCR using QuantiTect SYBR Green dye (Qiagen, Valencia, CA, http://www1.qiagen.com). DNA amplification was carried out using iCycler (Bio-Rad, Hercules, CA, http://www.bio-rad.com), and the detection was performed by measuring the binding of the fluorescence dye SYBR Green I to double-stranded DNA. The relative quantities of target gene mRNA against an internal control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), was possible by following a 
CT method. An amplification plot of fluorescence signal versus cycle number was drawn. The difference (CT) between the mean values in the duplicated samples of target gene and those of GAPDH were calculated using Microsoft Excel (Redmond, WA, http://www.microsoft.com), and the relative quantified value was expressed as 2–CT. The relative expression of each gene presented in this report was compared between 9L monolayer and 9L tumor sphere cells. All primers were designed for rat. Primers were as follows: GAPDH forward, 5'-ATG TAT CCG TTG TGG ATC TGA C-3'; GAPDH reverse, 5'-CCT GCT TCA CCA CCT TCT TG-3'; cIAP1 forward, 5'ACA TTT CCC CAG CTG CCC ATT C-3'; cIAP1 reverse, 5'-CTC CTG CTC CGT C TG CTC CTC T-3'; Survivin forward, 5' CAA CCT GGA CCT GAG TGA CAT-3'; Survivin reverse, 5'-CCA CCC ATA GAT CCT GTC AGA-3'; Bcl-2 forward, 5'-GGG ATG ACT TCT CTC GTC GCT AC; Bcl-2 reverse, 5' GTT GTC CAC CAG GGG TGA CAT; Mrp2 forward, 5'-GAC GAC GAT GAT GGG CTG AT-3'; Mrp2 reverse, 5'-CTT CTC ATG GCC AAG GAA GCT-3'; Mrp3 forward, 5'-TCC CAC TTC TCG GAG ACA GTA ACT-3'; Mrp3 reverse, 5'-CTT AGC ATC ACT GAG GAC CTT GAA-3'; Mrp6 forward, 5'-CTC TCC CAT TGG CTT CTT TGA G-3'; Mrp6 reverse, 5'-GTC CAC ATC CAC TAT GTC CGT CT-3'; Bcrp forward, 5'-CAG GTA GGC AAT TGT GAG GAA GA-3'; Bcrp reverse, 5'-AAT CAG GGC ATC GAT CTG TCA-3'.
Statistical Analysis
Student's t test was used to analyze statistical significance for all analyzed data except the survival data. Survival data were analyzed using SPSS software (SPSS, Inc., Chicago, http://www.spss.com) and a Kaplan-Meier survival curve with a log-rank test.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
9L Tumor Spheres Express NSC, Neuronal Cell, and Glial Cell Markers In Vitro
To determine the phenotype of cells within the tumor spheres, we determined the protein expression of neural stem cell markers Nestin and Sox2 (Fig. 1). Cells immunopositive for Nestin and Sox2 could be found throughout the tumor spheres. Cells throughout the tumor spheres were also found to be positive for the astrocyte lineage marker GFAP, whereas relatively few or no cells expressed neuronal or oligodendrocyte lineage markers, including ß-tubulin III, MAP2, and myelin/oligodendrocyte marker. Furthermore, we examined protein expression by fluorescence-activated cell sorting (FACS) analysis and showed that 98.7% ± 0.8% of these cells express Nestin and 29.9% ± 3.8% express Sox2, demonstrating that a large population of cells within these tumor spheres express some neural stem cell markers.
To test whether these tumor spheres produce progenies of different lineages, spheres were differentiated for 14 days in the absence of bFGF and EGF and immunostained for various neural lineage markers (Fig. 1). We found that tumor sphere cells maintained under these conditions were immunopositive for GFAP, MAP2, and myelin/oligodendrocyte markers, but ß-tubulin III was not detected. FACS analysis exhibited results similar to those derived via immunocytochemistry; 87.7% ± 4.7% were Nestin-positive, 24.2% ± 3.5% were Sox2-positive, 96.3% ± 2.8% were GFAP-positive, 29.1% ± 3.9% were MAP2-positive, and 5.7% ± 1.2% were myelin/oligodendrocyte-positive. These results suggest that these 9L tumor spheres, unlike endogenous stem/progenitor cells, differentiate into aberrant cells that coexpress stem/progenitor markers, including Nestin, along with multiple phenotypic neural lineage markers, notably GFAP and MAP2 (Fig. 1). The expression pattern of the differentiated progeny was similar in profile to that of primary cultured tumor cells, 9L cells grown as monolayer, from which the tumor spheres had originally been isolated and predominantly differentiated into GFAP- and MAP2-positive cells. To better understand how the environmental conditions altered the 9L cell phenotype, we examined the protein expression of 9L cells grown as a monolayer. The immunostaining pattern for the monolayer population was similar to that of tumor spheres differentiated for 14 days (Fig. 1), except for the expression of Sox2, which was not expressed in the parental 9L tumor cell line (data not shown). FACS analysis demonstrated that monolayer cells labeled for Nestin (92.2% ± 3.2%), GFAP (93.4% ± 4.2%), MAP2 (20.4% ± 4%), and myelin/oligodendrocyte (5.1% ± 1.2%), but neither ß-tubulin III nor Sox2 was detected. These results suggest that 9L tumor spheres posses cells that express multineural lineage markers similar to those expressed in 9L cells maintained as a monolayer; however, they also express Sox-2, a marker for neural stem/progenitor cells.
The Aggressiveness of 9L Cells In Vivo Is Reliant on the CSLCs Encompassing the Tumor Sphere
To determine whether 9L cells grown as monolayer and those grown as tumor spheres differ in their ability to grow as a tumor in a syngeneic rat experimentally induced tumor model, we implanted 5,000 cells from both populations of cells and subsequently assessed tumor volumes and survival times. To assess tumor volumes, we sacrificed the rats 18 days after implantation of 9L cells. We subsequently removed the brains and generated 20-µm sections throughout the tumor. Tumor length and greatest cross-sectional diameter were determined, and tumor volume was assessed using the formula for an ellipse [24], which showed a significantly (p < .02) greater tumor burden in rats implanted with 9L tumor sphere compared with rats implanted with monolayer cells (Fig. 2A).
|
|
Furthermore, we sought to examine the phenotype of implanted 9L tumor spheres by immunohistochemistry, H&E, and reticulin staining. We observed by immunohistochemistry that tumors generated from the 9L monolayer population were immunopositive for Nestin, GFAP, and myelin/oligodendrocyte markers (Fig. 4). H&E staining revealed no tumor in the saline control animals (Fig. 5A) and a high-grade glioma with necrosis consistent with a glioblastoma, displaying the dual nature of the gliosarcoma (Fig. 5B). In addition, in control rats, we found no reticulin staining (Fig. 5C). However, in tumor sphere-implanted rats, most of the tumor was positive for reticulin (Fig. 5D). These staining patterns suggest that 9L tumor spheres have the potential to recapitulate a gliosarcoma tumor by differentiating into both neural and glial lineages in vivo.
|
|
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
In this study, we aimed at characterizing the 9L gliosarcoma cell line, the most commonly used cell line for the generation of a syngeneic rat model for brain tumors, thereby enabling efficient testing of both immunotherapeutic and chemotherapeutic approaches in an experimentally induced brain tumor animal model. We also sought to determine whether different environmental conditions could alter the functionality of the 9L gliosarcoma line and generate a subpopulation of cells that behave more similarly to endogenous stem/progenitors or to primary tumor-derived tumor stem cells. To this end, we used a defined medium consisting only of EGF and bFGF mitogens along with B-27 supplement and heparin to maintain the 9L cell line. These conditions have been used to maintain endogenous neural stem/progenitors and tumor stem cells from human tumors and tumor cell lines. For example, Kondo et al. characterized the C6 rat tumor cell line in vitro and in vivo and demonstrated a small population of stem-like cells derived from the well-characterized Side population [28]. Setoguchi et al. demonstrated cancer stem-like cell potential from other cancer cell lines, including the MCF7 breast cancer line, B104 neuroblastoma, and HeLa adenocarcinoma [29].
We show that the 9L gliosarcoma cell line grown in a defined serum-free medium containing EGF and bFGF mitogens forms tumor spheres that contain cells that behave as cancer stem-like cells. 9L tumor spheres maintain their ability to proliferate and differentiate into the heterogeneous cellular population observed in the original 9L tumor cell line maintained under standard cell line serum-based medium conditions: monolayer cells. In addition, 9L tumor spheres contain cells that also express Sox-2, a neural stem cell marker, whereas the 9L monolayer cells do not. It has been suggested that Sox-2 does not affect endogenous neural progenitor proliferation but plays a role in the maintenance of a progenitor state [30]. We show that unlike endogenous neural progenitors, 9L tumor spheres maintain their ability to self-renew and generate spheres after 10 or more passages when grown at clonal density. We also show that the efficiency of generating tumor spheres is not changed after additional passages. These data suggest that the proliferative potential of 9L tumor spheres is maintained after extended passages and 9L spheres maintain expression of phenotypes generated from the parent cell line. Surprisingly, we show that the proliferative potential of 9L tumor spheres in vitro is less than that of 9L monolayer cultures. However, in vivo, 9L tumor spheres generate experimentally induced tumors at a faster rate than the monolayer-derived cells, suggesting, furthermore, that the environment plays a critical role in the behavior of 9L gliosarcoma cells.
More importantly, 9L tumor spheres generate experimentally induced brain tumors with larger volumes and lead to shorter survival times compared with tumors generated from the monolayer-derived cells when implanted into syngeneic Fisher rats. In addition, 9L tumor spheres in vitro are more resistant to Temodar and carboplatin and proliferate at a slower rate in vivo compared with the 9L cells grown as a monolayer. These data suggest that 9L tumor spheres consist of enriched numbers of cancer stem-like cells, which potentially makes them less susceptible to the actions of both temozolomide and carboplatin, which act as alkylating agents that disrupt DNA replication in cells that replicate at faster rates. Our data also demonstrate an upregulation of antiapoptosis-related genes and drug resistance-related genes in the tumor sphere population, which is not evident in the monolayer group. As a result, 9L gliosarcoma cells derived from tumor spheres are more resistant to the effects of temozolomide and carboplatin than the monolayer-derived cells.
Previous studies have shown that glioblastoma multiforme cells grown as monolayers, regardless of the cellular density, are unable to form experimentally induced tumors in vivo after the presumed cancer stem cells have been removed [20]. Similar results have been achieved by other laboratories [16, 28]. However, in this study, we show that tumor formation does occur from 9L monolayer-derived cells. Nonetheless, our monolayer-derived cells presumably contained a relatively low percentage of cancer stem-like cells, as no method for isolation or removal of these cells was used. The difference stems from the fact that we did not isolate and remove the small percentage of cancer stem cells that reside within the 9L gliosarcoma grown as a monolayer. It may be argued that it is this small population of cancer stem cells that is responsible for tumor formation. Regardless, more cancer stem cells are evident in the tumor sphere population, which may explain its more aggressive behavior. The purpose of this decision was to compare the 9L gliosarcoma tumor in its entirety—the same population of cells that is used in brain tumor models—to CSLCs isolated and propagated as tumor spheres.
Another apparent difference was the high level of GFAP expression by both differentiated tumor spheres and cells grown as monolayers. Previous studies involving glioblastomas [20] and the C6 cell line [28] have demonstrated a relatively greater proportion of cells positive for neuronal markers as opposed to astrocytic markers. In our study, we demonstrated a greater proportion of cells positive for astrocytic markers. In addition, we used vimentin (data not shown), which is believed to be an astrocytic marker, and witnessed a pattern similar to that of GFAP.
In vitro luciferase scans demonstrated a large difference in the luciferase signal between the two cell populations (monolayer vs. tumor spheres) cotransfected with the luciferase gene. This was demonstrated in a scan comparing 20,000 cells from each population, with nearly a twofold larger signal in the monolayer population, measured in photons/second. A question remains as to the cause of such a large difference in the luciferase signal. One argument is that tumor spheres containing CSLCs focus gene transcription, favoring genes regulating cellular proliferation, as opposed to less warranted genes, including the luciferase gene. A more reasonable explanation is that more luciferase-negative cells were selected for in culture in the tumor sphere group compared with the monolayer group because of the lower concentration of gentamicin-418 present in the neurosphere medium. With a lower concentration of gentamicin-418, more gentamicin-sensitive cells lacking the luciferase gene were allowed to proliferate, resulting in the difference observed in luciferase expression between the two cell populations.
Because CD-133/prominin could not be used as a NSC marker because of the lack of a rat homologue, we used Sox2 in addition to Nestin. Kondo et al. showed that even after 10 days of differentiation of CSLCs, there was a disproportionately high level of Nestin expression [28]. This led us to believe that Nestin alone may not be a specific NSC marker, persuading us to use another marker in addition to Nestin, Sox2. Sox2 expression was elevated in both tumor spheres and tumor spheres differentiated for 14 days, but not in the monolayer population. These data corroborate the assertion that tumor spheres generated from the 9L gliosarcoma are enriched with CSLCs.
In summary, our study elucidates properties of the 9L gliosarcoma cell line. Not only can these cells be used for simulation of glioma tumors, but other strategies can be harnessed using the understanding of cancer stem cells and chemoresistance to pursue more elaborate therapeutic strategies in a laboratory setting that may prove beneficial in a clinical setting.
|
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| STEM CELLS | THE ONCOLOGIST | CME | ALPHAMED PRESS JOURNALS |
