Abstract
Glioblastoma (GBM) is a devastating brain tumor with a poor survival outcome. It is generated and propagated by a small subpopulation of rare and hierarchically organized cells that share stem-like features with normal stem cells but, however, appear dysregulated in terms of self-renewal and proliferation and aberrantly differentiate into cells forming the bulk of the disorganized cancer tissues. The complexity and heterogeneity of human GBMs underlie the lack of standardized and effective treatments. This study is based on the assumption that available markers defining cancer stem cells (CSCs) in all GBMs are not conclusive and further work is required to identify the CSC. We implemented a method to isolate CSCs independently from cell surface markers: four patient-derived GBM neurospheres containing stem, progenitors, and differentiated cells were labeled with PKH-26 fluorescent dye that reliably selects for cells that divide at low rate. Through in vitro and in vivo assays, we investigated the growth and self-renewal properties of the two different compartments of high- and slow-dividing cells. Our data demonstrate that only slow-dividing cells retain the ability of a long-lasting self-renewal capacity after serial in vitro passaging, while high-dividing cells eventually exhaust. Moreover, orthotopic transplantation assay revealed that the incidence of tumors generated by the slow-dividing compartment is significantly higher in the four patient-derived GBM neurospheres analyzed. Importantly, slow-dividing cells feature a population made up of homogeneous stem cells that sustain tumor growth and therefore represent a viable target for GBM therapy development.
Introduction
Glioblastoma (GBM; World Health Organization grade IV astrocytoma) is the most aggressive and common primary brain tumor. With a dismal prognosis, GBM is the most challenging tumor to effectively treat; patients affected by GBM have a life expectancy of less than 1 year [1]. The most common approach for tackling tumors abides in the isolation of that population of rare cells that are thought to reseed cancers after chemotherapy and radiotherapy in hematopoietic as well as in solid tumors. Since the leukemia tumor-initiating cell was first described and characterized [2], the origins of these cells are still controversial, and their biology is not yet fully elucidated.
Stem cells are defined by their ability to divide asymmetrically, resulting in the formation of two daughter cells, one of which is another stem cell and the other one is a committed progenitor capable of further differentiation and, even if limited, proliferation but lacks the ability to self-renew. A cancer stem cell (CSC) functions likewise to sustain the growth and spread of tumors while repopulating the distinct cell types represented within the tumor.
Tumor heterogeneity represents a paramount feature supporting tumor robustness and inconclusive therapies. The bulk of malignant cells in GBM is generated by a rare fraction of self-renewing, multi-potent tumor-initiating cells (CSCs) able to maintain and propagate the tumor through their capacity of continued growth and resistance to chemotherapy and radiotherapy [3–6].
Moreover, CSCs are able to reinitiate the tumors following transplantation with the key features of the GBMs from which they derived, e.g., infiltrative phenotype, hypercellularity, pseudopalisading necrosis, and angiogenesis.
Processing and isolating the tumor-initiating cells in some tumors, such as breast, prostate, pancreas, skin, colon, and blood cancer, demonstrated the presence of a slowly cycling and highly tumorigenic cell fraction [2,4,7–13], while the same approach applied to brain tumors, and in particular to glioblastomas, manifests a remarkable challenge. Only little experimental evidence exists so far in glioblastoma. The demonstration of a small pool of cells slowly dividing and that retains long-term self-renewal ability was only recently obtained in vivo in a mouse model of glioblastoma through genetically engineered mice [14]: The chemotherapeutic drug temozolomide killed the highly proliferating cells within the tumor but did not eliminate the mostly dormant cells accountable for new bursts of tumor growth when therapy was stopped. Recently, it has also been reported that a population of dye-retaining brain tumor cells is able to generate GBMs in immunocompromised mice [15].
Although several markers are enlightening and/or significant for brain tumor stem cell identification, the segregation of universal specific markers suitable to pinpoint this tumor population representing clinically relevant target is still an unachievable goal. Most current treatments target rapidly dividing cells that generally constitute the non-stem cell component of tumors, leaving the quiescent, rather slow-dividing, stem cells to reinitiate the tumor. The slow-dividing population holds the ability to self-renew and to revert to the quiescent state to maintain the pool, while the fast-dividing majority population undergoes a limited but rapid replication and will eventually exhaust.
The isolation of a pure population of slow-dividing CSCs will serve as a valuable tool to allow the development of new therapeutic strategies to target the tumor-initiating cell compartment in patients with brain cancer.
In this study, we used PKH-26 fluorescent dye to identify and isolate a slow-cycling subpopulation responsible for tumor growth and progression in human GBMs. Following the binding to the cell membrane, the fluorescent dye is equally distributed between daughter cells upon division, leading to the halving of the fluorescence intensity with every cell division. The intensity of PKH-26 staining correlates with the cell division and then discriminates the highly cycling from the slowly dividing cells (putative stem cells).
Our findings suggest the application of a marker-independent method on GBM-derived neurospheres isolated from patients to isolate a small homogeneous slow-dividing population capable of long-term self-renewal and tumor maintenance and progression out of a tumoral bulk with proliferative heterogeneity.
Materials and Methods
Preparation of Cell Suspensions from Patient GBM
This study was approved by the Ethical Committee for human experimentation of the European Institute of Oncology, and all patients signed an approved consent document before surgery. Surgical specimens of tumors were collected at the Neurosurgery Department, Istituto Neurologico Carlo Besta (Milan, Italy), and tissue fragments without necrotic areas were processed. The tissue was mechanically dissociated into single-cell suspension with papain (2 mg/ml; Worthington Biochemical, Lakewood, NJ; http://worthington-biochem.com) at 37°C for 2 hours.
The GBM samples collected in our laboratory were classified in the three molecular subtypes defined on the basis of gene expression profile that has been previously described [16,17]. The GBM-derived neurospheres used for this study were derived from different GBM classes (i.e., hGBM#7 and hGBM#9 classified as mesenchymal subtype and hGBM#8 and hGBM#18 as proneural subtype).
Neurosphere Culture
Neurosphere culture was maintained in Dulbecco's modified Eagle's medium/Ham's F12 nutrient mixture (DMEM/F12) supplemented with B27 supplement (Life Technologies, Paisley, United Kingdom; www.invitrogen.com), 20 ng/ml epidermal growth factor, 10 ng/ml basic fibroblast growth factor (PeproTech, Rocky Hill, NJ; http://www.peprotech.com), and 0.0002% heparin (Sigma-Aldrich, St Louis, MO; http://www.sigmaaldrich.com) at 37°C in a 5% CO2 humidified incubator, as previously described. All cultures were passaged by mechanical dissociation of the spheres, and the cells were seeded at a density of 104 cells/cm2. Cell growth was measured by seeding the cells in 96-well plates (2000 cells/well) in DMEM/F12 complete medium and using the CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (Promega, Madison, WI; http://www.promega.com).
PKH-26 Assay
Human primary GBM cells were stained for 5 minutes with 1:2000 PKH-26 dye (Sigma), blocked with 1% BSA, washed twice, and plated to obtain primary neurospheres that are then passaged for 2 to 3 weeks to allow adequate time for cell division and PKH-26 dilution. Fluorescence-activated cell sorting (FACS/sorting) was performed on single-cell suspensions 2 weeks after labeling using a FACS Vantage SE Flow Cytometer (BD Biosciences, Franklin Lakes, NJ) to obtain PKH-positive (PKH-Pos) and PKH-negative (PKH-Neg) cells. The cells were then isolated according to their label intensity: The brightest cells (1–3% of the entire population) represent the stem cells characterized by a mean fluorescence compatible with very few cycles of cell divisions. Purity of sorted populations was verified by flow cytometry.
FACS/Sorting Analysis
Cells labeled with PKH-26 were subjected to FACS/sorting analysis with a FACS Vantage SE Flow Cytometer (BD Biosciences; http://bdbiosciences.com) to yield PKH-Pos (≤3% of the entire population) and PKH-Neg (≤70% of the cells analyzed) cells. After gating to eliminate debris and dead cells, additional gates were drawn in the PKH-26 channel and in the forward scatter channel to select human PKH-Pos cells. Cells showing signal for PKH-26 above the gate established by the unlabeled control samples were deemed to be PKH-Pos cells.
3-(4, 5-Dimethylthiazol-2-yl)-2, 5-Diphenyltetrazolium Bromide Assay
GBM neurospheres ready for passaging were mechanically dissociated and seeded in 96-well plates at the density of 3000 cells per well. 3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT, 50 mg/ml) was added, and after a 4-hour incubation, crystals were dissolved in DMSO and added to the wells. Cell viability was evaluated by CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (Promega; http://www.promega.com). Three independent replicates were considered for each experiment.
Clonogenic Assay
The clonogenic efficiency was analyzed by seeding human GBM neurospheres at a low density in methylcellulose whose viscous composition does not allow cell aggregation. To evaluate the clonogenic capacity, immediately after sorting analysis, PKH-Pos, PKH-Neg, and unsorted cells were resuspended in DMEM/F12 with methylcellulose (StemCell Technologies, Vancouver, British Columbia; http://www.stemcell.com) and seeded in a 35-mm culture plate (100 cells/cm2). Reported numbers represent a minimum of three plates per condition. Ten to fifteen days after plating, neurospheres containing more than 20 cells were scored: The ratio between neurospheres formed and number of single cells plated corresponds to the percentage of neural stem cells in the plate.
Replating Procedure
Soon after clone dissociation and seeding at a low density in methylcellulose, both PKH-Pos and PKH-Neg cells are able to efficiently clone. The investigation of a possible progressive decline of this capability due to subsequent subculturing was carried out. Sorted single cells were seeded in methylcellulose plates and allowed to form colonies for 2 weeks. Representative clones from each population were picked with a Gilson tip, dissociated, and then reseeded in the stem cell medium and methylcellulose. We serially passaged GBM-derived neurospheres from five different patients up to the tertiary neurospheres to estimate the overall clonogenic efficiency of PKH-Pos cells compared to PKH-Neg cells. After several serial passaging on the viscous medium, PKH-Neg cells exhibited limited proliferative potential exhausting, while PKH-Pos generated a large number of progeny even if this population showed a slight reduction in self-renewal potential (Figure 2D;*P < .05).
Figure 2.
Distinct clonogenic efficiency of the PKH-Pos and PKH-Neg cell compartments. (A) Bars represent the clonogenic potential of PKH-Pos cells (ranging from 7.5% for hGBM#7 to 25.8% for hGBM#18) collected immediately after sorting (post sorting) and compared to unsorted cells and the negative counterpart. All cells were grown in methylcellulose for 15 days, and then each plate was examined under a light microscope, and the total number of formed neurospheres was determined. *P < .05; **P < .01 with one-way ANOVA with Bonferroni correction for multiple comparisons. (B) The graph depicts the characteristic discrepancy in spheres' size between neurospheres derived from PKH-Pos cells and PKH-Neg cells and assessed after 2 weeks of growth at low density in methylcellulose substrate. The data are the means ± SD from four patients (n = 60 spheres for each fraction for each patient; **P < .01). The scale bar represents 100 µm. (C) The panel shows the fluctuation of spheres' size among all the studied neurospheres. Black represents spheres with diameter from 50 µm to up to 100 µm distributed among unsorted, PKH-Pos, and PKH-Neg cells, and gray illustrates spheres with a diameter ≥200 µm, while the bars with solid line show the contribution of spheres with a not determined diameter (N.D.). (D) Clonogenic efficiency of GBM cells after serial replating. Random single clones from both populations were picked up and, after dissociation, consecutively plated for at least three times in methylcellulose to form new secondary spheres. The majority of the cells die during passaging except for the responsive sphere-forming cells: The PKH-Neg pool goes down to 0% efficiency, while the PKH-Pos putative stem cells can be indefinitely replated. Average values derived from six independent experiments/patients were represented. Error bars represent 95% confidence intervals (*P < .05). GLM plus Tukey post-hoc test was used.
In Vivo Limiting Dilution Transplantation Assay
FACS/sorting was applied on GBM single-cell suspensions to sort the cells into PKH-Pos and PKH-Neg cells 2 weeks after labeling. Different cell concentrations (from 105 to 10 cells) of the subpopulations obtained post-sorting and the unsorted cells were resuspended in 2 µl of phosphate-buffered saline (PBS) and stereotaxically injected into the nucleus caudatus (coordinates from bregma: 1 mm posterior, 3 mm left lateral, and 3.5 mm in depth) of 5-week-old female nu/nu CD1 mice (Charles River, Wilmington, MA; http://www.criver.com). In vivo experiments in CD-1 nude mice were performed in accordance with the Italian laws (D.L.vo 116/92 and following additions), which enforce EU 86/609 Directive (Council Directive 86/609/EEC of 24 November 1986 on the approximation of laws, regulations and administrative provisions of the Member States regarding the protection of animals used for experimental and other scientific purposes). The mice were maintained until development of neurologic signs, and the brains of killed mice were collected.
Immunohistochemistry
Xenografted mouse brains were formalin-fixed and paraffin-embedded according to established procedures. Each brain was sliced into several 5-µm coronal sections and placed onto polysine slides (Thermo Scientific, Waltham, MA). Sections were cut, deparaffinized with xylene, rehydrated in a graded alcohol series, and stained in Mayer's hematoxylin and eosin (H&E)-phloxine B solution.
For immunohistochemistry (IHC), tissue endogenous peroxidases were blocked with 3% hydrogen peroxide and antigen retrieval was carried out in sodium citrate (0.01 M, pH 6.0) in a water bath (at 95°C for 55 minutes). Paraffin-embedded sections were blocked with a PBS solution containing 2% goat serum and 1% BSA for 120 minutes and incubated overnight in humidified chambers with primary antibodies: human nuclei (monoclonal mouse, 1:1000; Millipore, Billerica, MA), Olig2 (goat polyclonal, 1:500; Santa Cruz Biotechnology, Dallas, TX), Sox2 (rabbit polyclonal, 1:50; StemCell Technologies), human Nestin (mouse monoclonal, 1:200; Chemicon Millipore), human CD31 (mouse monoclonal, 1:40; Dako, Carpinteria, CA), CD15 (mouse monoclonal, 1:100; BD Pharmigen, San Jose, CA), CD133 (mouse monoclonal, 1:100; Miltenyi, Bergisch Gladbach, Germany), and Ki-67 (mouse monoclonal, 1:50; Millipore). Tissue sections were washed in PBS and incubated with a secondary biotinylated antibody (1:200; Vector Labs, Burlingame, CA) for 1 hour. Antibody binding was detected using a VECTASTAIN Elite Avidin-Biotin Complex Peroxidase Kit following the manufacturer's instructions. All sections were counterstained with Mayer' shematoxylin and visualized using a bright-field microscope.
Statistical Analysis
Statistical analysis was performed using Statistical Package for Social Science software. Significance of differences for all parametric variables has been tested by means of analysis of variance (ANOVA) or generalized linear models (GLMs). Data are graphed as mean + 95% confidence intervals. Differences were considered “statistically significant” when *P < .05, **P < .001, and ***P < .0001.
Results
Isolation of a Slow-Dividing Cell Population Present in GBM-Derived Neurospheres
We have established neurosphere cultures from GBM patient specimens by culturing freshly dissociated cells in conditions commonly used to culture neural stem cells [16,18,19]. To be defined as cancer stem/progenitor cell populations, these cells are required to be clonogenic in vitro, to express stem cell markers (like Olig2 or Sox2) [20,21], to be capable of neuronal and/or glial differentiation, and to be tumorigenic in vivo, generating xenograft tumors that recapitulate the biologic features of the parental GBM.
We performed proliferative assays based on the property of the PKH-26 fluorescent dye to bind cell membranes and to segregate in daughter cells after each cell division, such that intensity of staining correlates with the number of prior cell divisions at the single-cell level [22].
We examined a series of four GBM patient-derived neurospheres cultured for few passages to identify and isolate slow-dividing putative stem cells, since the neurospheres are composed of a mixed population of stem, progenitors, and differentiated cells. FACS/sorting analyses of the two different compartments of slow- and high-dividing cells (making up most of the neurospheres) were carried out 2 weeks after dye labeling. In Figure 1A, examples of typical FACS/sorting PKH-stained cell profiles derived from two different patients are represented. PKH-Pos (slow-dividing) cells are defined as the ∼1% to 3% most fluorescent cells, characterized by a mean fluorescence compatible with very few cycles of cell divisions, while PKH-Neg cells represent up to 70% of highly dividing gated cells. The two fractions (PKH-Pos and PKH-Neg cells) were immediately expanded in liquid culture after sorting, and the growth was assessed by MTT assay. We found reduced cellular growth kinetics of PKH-Pos cells in all patient-derived GBM neurospheres analyzed (Figure 1B; **P < .001 with ANOVA among cell fractions). These results suggest that the PKH-Pos cells are not quiescent but are rather capable to divide at low rate.
Figure 1.
Flow cytometric cell sorting of PKH-26-positive and PKH-26-negative cells and their growth kinetics. (A) Schematic representation of PKH-26 cell labeling and culture expansion. After gating to eliminate debris and dead cells, additional gates were drawn to select PKH-Pos cells. Representative flow cytometry plots depict the gating strategy for collections of PKH-Pos (1–3%) and PKH-Neg cells (up to 70%; right panel) and not stained cells (left panel). The scale bar in boxed areas represents 20 µm. (B) The growth properties of PKH-Pos and PKH-Neg cells and unsorted cells from hGBM#7, hGBM#8, hGBM#9, and hGBM#18 were investigated through MTT assay. Cell growth was measured every 24 hours for 7 days. SD bars are shown for each measurement. **P < .01 with one-way ANOVA with Bonferroni correction for multiple comparisons.
Quantitative Clonal Analysis
To further characterize the slowly dividing cell fraction retaining the dye (PKH-Pos) inside the membrane, we analyzed its clonogenic potential compared to the negative counterpart and the total population (pre-sorting; Figure 2A) by methylcellulose assay. Following fractioning of established neurosphere cultures through FACS, clonogenic capacity analysis was performed by counting single clones obtained after 15 days of culture, and the clonogenic cells were calculated as the percentage of the total number of seeded cells (up to 3000 cells) on 35-mm petri dishes. We observed that both the cell fractions retaining the dye (PKH-Pos) and the negative counterpart (PKH-Neg) were able to form expandable neurospheres, but the clonogenic efficiency was considerably variable: The cells retaining the PKH-26 dye showed a significant increased capacity to form spheres (ranging from 7.5% for hGBM#7 to 25.8% for hGBM#18) compared to the fast proliferating sibling cells and the total population. The greater clonogenic capacity of PKH-Pos cells might be justified by the urge to self-maintain and to generate a more frequently dividing population, while the PKH-Neg fraction, however, exhibited a lower clonogenic ability since it has already made an extensive number of cycles toward the differentiation (see also Figure W1).
We found a divergence in size of the PKH-Neg cell-derived clones and PKH-Pos cell-derived clones: The diameter was 30% bigger in spheres composed by PKH-Pos cells (Figure 2B; **P < .0001 using unpaired t test with t = 9.585 and df = 77). The number of the cells composing the neurospheres was about three times more in neurospheres derived from PKH-Pos cells (data not shown). The neurosphereforming process ability was assessed in terms of fluctuation in size and number of clones 2 weeks after plating in methylcellulose at low density for the PKH-Pos, PKH-Neg, and the entire population (unsorted). The proportion of larger neurospheres (with a diameter bigger than 200 µm) fell in the long-term self-renewable PKH-Pos population. Clones with a diameter smaller than 50 µm were represented in greater proportion in the PKH-Neg population and, in a lesser extent, in the unsorted fraction (Figure 2C).
We then performed experiments to evaluate the clonogenic capacity of PKH-Pos, PKH-Neg, and unsorted cells: Dissociated cells were serially passaged up to the tertiary neurospheres to estimate the extensive clonogenic efficiency of PKH-Pos cells compared to PKH-Neg cells in all the patients analyzed. After several serial passaging on the viscous medium, PKH-Neg cells exhibited limited proliferative potential exhausting, while PKH-Pos cells still generated a large number of progeny (Figure 2D; *P < .05 with Tukey post-hoc test). This is consistent with the definition of stem cell (uncommitted cell that divides repeatedly and indefinitely while maintaining the potency to generate differentiated cell types) and progenitor, that is, a cell with a limited division capacity before a change in potency or overt differentiation.
Carrying out these experiments of neurosphere-forming proficiency, we were able to split off slowly and highly dividing cells and demonstrate that the former were able to produce large colonies with a greater self-renewal potential allowing several consecutive “pick and replate,” while the negative fraction gave rise predominantly to small clones derived from cells that do not exhibit stem cell criteria. These results suggest that PKH-Neg cells contain mostly progenitor cells that actively divide for few cycles and then exhaust, whereas the slow-cycling cell pool (PKH-Pos) remains relatively quiescent to expand in number.
PKH-Pos Cells Exhibit Greater In Vivo Tumor-Initiating Capacity and Stem Cell Frequency
We carried out an orthotopic transplantation assay stereotaxically injecting dissociated neurospheres derived from four different patients (hGBM#7, hGBM#8, hGBM#9, and hGBM#18) in the brain of nude mice (n = 3 minimum per group) to test their ability to generate xenograft tumors recapitulating the biologic and genomic features of the parental GBM.
To assess the number of cells required to generate tumors in vivo, we intracerebrally injected the PKH-Pos, PKH-Neg, and unsorted populations at different concentrations (from 100,000 cells to 10 cells).
The 100% of mice injected with 105 or 104 cells derived from the three subpopulations developed tumors with no differences in survival. PKH-Pos cells engrafted with lower cell doses (100 or 10 cells) formed tumors in all the injected animals, while PKH-Neg cells generated tumors only at low frequency or not at all (as shown by H&E and IHC analysis; Figures W2, W3, and W4), with a statistically significant difference (*P < .0001 for both doses; Table 1).
Table 1.
Frequency of CSCs in hGBM#7, hGBM#8, hGBM#9, and hGBM#18 Neurospheres.
| Number of Transplanted Cells | CSC Frequency (Estimate) | |||||||
| 105 | 104 | 103 | 102 | 10 | ||||
| hGBM#7 | Unsorted | Incidence | 14/14 | 5/5 | 6/7 | 4/9 | 0/9 | 1/353 |
| Median survival | 47.4 ± 7.3 | 51.8 ± 7.9 | 70.7 ± 11.4 | 60.7 ± 8.8 | 0 ± 0 | |||
| PKH-Neg | Incidence | 8/8 | 3/3 | 0/3 | 0/3 | 1/433 | ||
| Median survival | 65.6 ± 15.1 | 65.3 ± 9.6 | 0 ± 0 | 0 ± 0 | ||||
| PKH-Pos | Incidence | 8/8 | 3/3 | 4/4 | 3/4 | 1/8* | ||
| Median survival | 63.6 ± 4.5 | 77 ± 5.2 | 84.7 ± 14.4 | 117.3 ± 43.5 | ||||
| hGBM#8 | Unsorted | Incidence | 6/6 | 8/8 | 7/7 | 8/7 | 2/4 | 1/100 |
| Median survival | 77.7 ± 10.5 | 84.5 ± 14.3 | 111.8 ± 5.3 | 133 ± 9.8 | 156 ± 4.2 | |||
| PKH-Neg | Incidence | 3/3 | 3/3 | 3/6 | 0/3 | 1/154 | ||
| Median survival | 120 ± 3.5 | 108.3 ± 6.6 | 198 ± 37 | 0 ± 0 | ||||
| PKH-Pos | Incidence | 3/3 | 3/3 | 5/5 | 3/3 | 1/1* | ||
| Median survival | 118.7 ± 19.6 | 119.7 ± 4.2 | 130.8 ± 17.2 | 258.8 ± 79.1 | ||||
| hGBM#9 | Unsorted | Incidence | 5/5 | 4/4 | 3/4 | 2/4 | 2/4 | 1/304 |
| Median survival | 92.3 ± 10.6 | 137.6 ± 6.3 | 183.7 ± 38 | 166.2 ± 59 | 243 ± 15.5 | |||
| PKH-Neg | Incidence | 3/4 | 1/4 | 1/4 | 0/4 | 1/4802 | ||
| Median survival | 160 ± 0 | 233 ± 0 | 123 ± 0 | 0 ± 0 | ||||
| PKH-Pos | Incidence | 4/4 | 4/4 | 4/4 | 4/4 | 1/1* | ||
| Median survival | 175.3 ± 35 | 123.7 ± 20 | 202.7 ± 9 | 234.7 ± 19 | ||||
| hGBM#18 | Unsorted | Incidence | 4/4 | 7/7 | 6/7 | 4/6 | 3/7 | 1/240 |
| Median survival | 66.7 ± 10.3 | 80.5 ± 10.6 | 92.3 ± 5.2 | 148.6 ± 23.5 | 131.7 ± 10.6 | |||
| PKH-Neg | Incidence | 6/6 | 4/6 | 2/6 | 0/6 | 1/681 | ||
| Median survival | 88.9 ± 8.7 | 114 ± 8.2 | 197.5 ± 27.4 | 0 ± 0 | ||||
| PKH-Pos | Incidence | 6/6 | 6/6 | 6/6 | 5/6 | 1/6* | ||
| Median survival | 121.3 ± 21.5 | 101.3 ± 4 | 142.1 ± 8.4 | 143.1 ± 3.2 | ||||
Limiting dilution analysis of CSC frequency calculated with ELDA algorithm in unsorted, PKH-Pos, and PKH-Neg cells of four different patients immediately after sorting analysis. Data show the CSC frequency per transplanted cell (estimate) from the indicated cell groups. In the table, the incidence of tumor formation and the survival time (in days; mean ± SD) of tumor-bearing mice after injection of hGBM#7, hGBM#8, hGBM#9, and hGBM#18 cells are also indicated. Differences in CSC frequency between the PKH-Pos population and the other two experimental groups are calculated and resulted statistically significant for *P < .0001.
Using the ELDA algorithm (http://bioinf.wehi.edu.au/software/elda/), we calculated the CSC frequency in the tumors derived from the three cell populations transplanted: The analysis showed a significantly higher CSC frequency (1/1 to 1/8) in the positive fraction compared to the negative counterpart (1/154 to 1/4802). The unsorted population of all the patients analyzed showed an intermediate estimation of CSC frequency. Histologic analysis of tumors that arose in all mice transplanted with the three cell populations showed that all the tumors were GBMs according to the World Health Organization classification and they all were similar in cytoarchitecture, vascularization, and invasiveness (H&E of brain coronal sections analyzed with ImageJ software, rsbweb.nih.gov/ij/; Figure W2). After testing the human origin of injected cells through the antigen human nuclei, the expression pattern of putative stem cell markers (i.e., CD133, CD15, CD31 Sox2, Olig2, and Nestin) was examined and quantified, but no consistent differences were found in the patients studied nor among the tumors derived from the three different cell fractions. Ki-67 expression was also evaluated in the same sections analyzed by IHC and no differences were found (Figure W3).
Expression of stem cell-associated genes was evaluated in vitro in PKH-Pos, PKH-Neg, and the total population of all the studied GBM-derived neurospheres. The putative stem/progenitor cell markers Olig2 and Nestin were enriched in PKH-Pos fraction, while Sox2 showed statistically significant difference (Figure 3; *P < .05 with paired samples t test); other markers as CD133 and CD15 were expressed in variable degrees among the samples.
Figure 3.
Gene expression levels in patient-derived neurospheres. Putative stem-like cell gene expression was quantified by quantitative reverse transcription-polymerase chain reaction in four different GBM-derived neurospheres (of hGBM#7, hGBM#8, hGBM#9, and hGBM#18) with n = 5. Experiments were performed in triplicate. Error bars represent 95% confidence intervals (*P < .05). GLM plus paired samples t test was used.
These results show that both PKH-Pos and PKH-Neg populations are able to form tumors in vivo, but PKH-Pos cells exhibit higher in vivo tumor-initiating capacity.
Taking Apart Clonogenic and Tumorigenic Capacity
We analyzed the frequency of stem cells as cells able to form neurospheres in vitro (by methylcellulose assay) and to initiate tumors in vivo (by limiting dilution assay) in all the patients studied (hGBM#7, hGBM#8, hGBM#9, and hGBM#18). Interestingly, in the unsorted and PKH-Neg cell populations, there was a relevant difference in the stem cell frequency calculated by neurosphere-forming assay and by in vivo limiting dilution (Figure 4; *P < .05 with Tukey post-hoc test); CSCs in vitro resulted significantly higher than in vivo. This finding means that not all the cells able to form neurospheres in vitro are able to form tumors in vivo. Indeed, we demonstrated by methyl-cellulose assay that PKH-Neg cells form in vitro small clones that exhaust after picking and replating, being composed of progenitor cells. Moreover, the stem cell frequency of PKH-Pos cells was similar in in vitro and in vivo studies, demonstrating that the PKH-Pos fraction is the unique population composed of homogeneous stem cells.
Figure 4.
Stem cell frequency calculated in vitro by methylcellulose assay and in vivo by limiting dilution assay. Comparison of stem cell frequency calculated by in vivo and in vitro assays in unsorted, PKH-Pos, and PKH-Neg cell fractions of a pool of four patients. Results suggest that the PKH-Pos fraction is a unique population consisting of homogeneous cells, maintaining the same stem cell frequency when calculated using the neurosphere-forming assay and the in vivo limiting dilution assay. Error bars represent 95% confidence intervals (*P < .05). GLM plus Tukey post-hoc test was used.
Discussion
Marker-independent Method to Isolate CSCs
Fluorescent labeling has been already used to track and to define the biology of long-term repopulating hematopoietic stem cells because of their quiescent nature [22] and, more recently, to purify stem cells from cultured mammospheres isolated from normal mammary gland [23]. The skeletal muscle tissue comprises an undifferentiated slow-dividing cell population with long-term self-renewal ability that is restricted to the PKH-26 label-retained cell fraction [24].
While a slow-cycling cell compartment highly enriched in tumorigenic stem cells has been already described in several solid tumors such as breast [25], pancreas [26], and melanoma [27], in the human brain tumors a small subpopulation of slow cycling stem-like cells driving glioblastoma formation still needs to be identified. A CSC would function similarly to a normal neural stem cell to sustain the growth and spread of tumors, repopulating in the meanwhile the distinct cell types represented within the tumor. However, a CSC would not be subjected to the same intrinsic and extrinsic controls as normal stem cells.
Several studies report discrepancies on the segregation of universal markers for CSCs in gliomas mainly due to a lack of uniform layout for cell sorting and “stemness” determination. Furthermore, the isolation of CSCs implies the identification of possible markers: Here, we suggest an alternative method based on the functional properties of the membrane-labeling vital dye PKH-26 to isolate a “pure” population of CSCs.
Slow-dividing Cells Exhibit Indefinite Self-Renewal Capacity In Vitro
As model system, we chose to use the neurospheres because they represent a good surrogate for in vitro study of glioma stem and progenitor cells [19,28]. A recent study indicates that in vitro expansion of GBM stem/progenitor cells as neurospheres does not alter the differentiation ability and the tumorigenic potential of these cells, neither their karyotype and gene expression pattern [29].
The label retaining cells make divisions at a slow pace but show a more clonogenic efficiency compared to the other fractions in all the patients studied that might be the result of an asymmetric division in which PKH-26-positive stem cells are slow dividing because they have to maintain the pool.
To note, the free-floating culture system (MTT assay shown in Figure 1) allows an indefinite growth for all subpopulations because it contains a mixed population of cells at various stage of commitment while solid and viscous medium (e.g., methylcellulose) sustains the growth of a less heterogeneous population, in which only cells endowed with “stemness” characteristics are able to indefinitely expand. Surprisingly, the unsorted cell fraction grows faster than the PKH-26-negative fraction. We can hypothesize that the growth of PKH-26-negative cells can be influenced by the PKH-26-positive cells within the bulk population.
It is important to note that clone sizes varied among the colonies on methylcellulose: While PKH-26-negative cells formed clones extremely variable in size (diameter smaller than 50 µm and up to 100 µm), PKH-26-positive cells gave rise to clones displaying a significantly bigger diameter—and in a greater proportion—in all patients considered, indicating that slow-dividing cells composing neurospheres identify a homogeneous population. Due to fewer cell divisions completed, PKH-Pos fractions exhibit an increased number of cells endowed with “stemness” properties (then more prone to cancer), while the PKH-Neg fraction is mainly composed of progenitor cells set up to a more advanced stage of differentiation. This cell fraction is made up of cells at different levels of differentiation: There are cells in immature state, cells moving toward a more differentiated state, and cells that have already completed the major part of their life span.
The high heterogeneity in cell population composition present in each GBM is reflected in a dearth of markers (or pattern of markers) sufficiently robust to convincingly identify GBM stem cells. In this study, the expression of the putative stem cell markers Olig2, Nestin, Sox2, CD133, and CD15 has been analyzed by quantitative reverse transcription-polymerase chain reaction, and the PKH-Pos cells of the four human GBM-derived neurospheres investigated showed an up-regulation of three of five genes that are enriched in CSCs, though only Sox2 exhibited a difference that is statistically significant. These results emphasize the concept that PKH-26-positive population is enriched of CSCs heterogeneously expressing more than one putative stem cell marker.
Experiments of consecutive neural colony-forming assay showed that after plating PKH-26-positive and PKH-26-negative fractions, only the former was able to produce large colonies with a greater proliferative potential (extensive self-renewal and proliferative potential), while the negative fraction was exhausted starting from the second replating. These results reveal that PKH-Neg population is mainly composed of progenitor cells that have done several cell cycles and are committed toward a more differentiated state. However, PKH-Pos cells showed an extended clonogenic efficiency through which it is possible to maintain the pool of undifferentiated cells and at the same time to generate early progenitors.
Our results are in accordance with previous experimental data of clonal studies on semisolid substrate of glioma cells demonstrating that large colonies contain numerous self-renewed progeny, suggesting that the cells generating large colonies possess stem cell characteristics [30,31].
PKH-Pos Cells Exhibit Greater In Vivo Tumor-Initiating Capacity and Stem Cell Frequency
Through in vivo limiting dilution analysis, we demonstrated that the stem cell frequency (calculated by the ELDA algorithm) was significantly lower in the negative counterpart (1/154 to 1/4802) in comparison with the label retaining slow-dividing cell proportion (1:1 to 1:8) in all patients analyzed. Only for the slow-dividing cells the frequency of stem cells measured in vitro by methylcellulose assay was similar to that measured in vivo by limiting dilution assay, emphasizing the data that PKH-Pos fraction contains a homogeneous population of cells. On the contrary, the frequency of stem cells assessed by methylcellulose assay was higher than the frequency estimated by in vivo limiting dilution assay for the PKH-Neg cells that indeed formed in vitro small, exhausting clones. This disparity between clonogenicity in vitro and tumorigenicity in vivo points out the fact that not all tumorigenic cells in the PKH-Neg cell fraction are stem cells. In fact, while PKH-26-positive cells display self-renewal ability in vitro, the PKH-Neg cells denote self-extinguishing behavior.
Our data are supported by the recent identification in vivo in a mouse model of glioma of quiescent glioma stem-like cells that propagate tumor growth after chemotherapy [14].
Conclusion
A major conclusion of this investigation is the finding of relatively quiescent stem-like cells able to expand indefinitely to sustain tumor growth, and this capacity counteracts proliferating progenitors, a heterogeneous pool of cells near the differentiation stage.
Our study demonstrates the feasibility to take advantage of fluorescent dye as PKH-26 to enrich for slow-dividing cells retaining high capability to self-renew to identify a distinct category of cells that constitute the tumor.
Therefore, a new approach to GBM therapy might focus on specific targeting of these populations.
Supplementary Material
Acknowledgments
We thank D. Osti and M. Setti for technical assistance and S. Ronzoni for FACS/sorting assistance.
Abbreviations
- GBM
glioblastoma
- CSCs
cancer stem cells
- FACS
fluorescence-activated cell sorting
- MTT
3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide
Footnotes
G.P. was supported by Ministry of Health grants. Disclosure of potential conflicts of interest: No potential conflicts of interest were disclosed.
This article refers to supplementary materials, which are designated by Figures W1 to W4 and are available online at www.neoplasia.com.
References
- 1.Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, Scheithauer BW, Kleihues P. The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol. 2007;114:97–109. doi: 10.1007/s00401-007-0243-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Baum CM, Weissman IL, Tsukamoto AS, Buckle AM, Peault B. Isolation of a candidate human hematopoietic stem-cell population. Proc Natl Acad Sci USA. 1992;89:2804–2808. doi: 10.1073/pnas.89.7.2804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, Dewhirst MW, Bigner DD, Rich JN. Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature. 2006;444:756–760. doi: 10.1038/nature05236. [DOI] [PubMed] [Google Scholar]
- 4.Clarke MF, Dick JE, Dirks PB, Eaves CJ, Jamieson CH, Jones DL, Visvader J, Weissman IL, Wahl GM. Cancer stem cells—perspectives on current status and future directions: AACR Workshop on cancer stem cells. Cancer Res. 2006;66:9339–9344. doi: 10.1158/0008-5472.CAN-06-3126. [DOI] [PubMed] [Google Scholar]
- 5.Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S, Fiocco R, Foroni C, Dimeco F, Vescovi A. Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res. 2004;64:7011–7021. doi: 10.1158/0008-5472.CAN-04-1364. [DOI] [PubMed] [Google Scholar]
- 6.Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB. Identification of human brain tumour initiating cells. Nature. 2004;432:396–401. doi: 10.1038/nature03128. [DOI] [PubMed] [Google Scholar]
- 7.Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci USA. 2003;100:3983–3988. doi: 10.1073/pnas.0530291100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dick JE. Stem cell concepts renew cancer research. Blood. 2008;112:4793–4807. doi: 10.1182/blood-2008-08-077941. [DOI] [PubMed] [Google Scholar]
- 9.Kelly K, Yin JJ. Prostate cancer and metastasis initiating stem cells. Cell Res. 2008;18:528–537. doi: 10.1038/cr.2008.50. [DOI] [PubMed] [Google Scholar]
- 10.Lee CJ, Dosch J, Simeone DM. Pancreatic cancer stem cells. J Clin Oncol. 2008;26:2806–2812. doi: 10.1200/JCO.2008.16.6702. [DOI] [PubMed] [Google Scholar]
- 11.Quintana E, Shackleton M, Sabel MS, Fullen DR, Johnson TM, Morrison SJ. Efficient tumour formation by single human melanoma cells. Nature. 2008;456:593–598. doi: 10.1038/nature07567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Reya T, Morrison SJ, Clarke MF, Weissman IL. Stem cells, cancer, and cancer stem cells. Nature. 2001;414:105–111. doi: 10.1038/35102167. [DOI] [PubMed] [Google Scholar]
- 13.Ricci-Vitiani L, Lombardi DG, Pilozzi E, Biffoni M, Todaro M, Peschle C, De Maria R. Identification and expansion of human colon-cancer-initiating cells. Nature. 2007;445:111–115. doi: 10.1038/nature05384. [DOI] [PubMed] [Google Scholar]
- 14.Chen J, Li Y, Yu TS, McKay RM, Burns DK, Kernie SG, Parada LF. A restricted cell population propagates glioblastoma growth after chemotherapy. Nature. 2012;488:522–526. doi: 10.1038/nature11287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Deleyrolle LP, Harding A, Cato K, Siebzehnrubl FA, Rahman M, Azari H, Olson S, Gabrielli B, Osborne G, Vescovi A, et al. Evidence for label-retaining tumour-initiating cells in human glioblastoma. Brain. 2011;134:1331–1343. doi: 10.1093/brain/awr081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ortensi B, Osti D, Pellegatta S, Pisati F, Brescia P, Fornasari L, Levi D, Gaetani P, Colombo P, Ferri A, et al. Rai is a new regulator of neural progenitor migration and glioblastoma invasion. Stem Cells. 2012;30:817–832. doi: 10.1002/stem.1056. [DOI] [PubMed] [Google Scholar]
- 17.Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, Misra A, Nigro JM, Colman H, Soroceanu L, et al. Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell. 2006;9:157–173. doi: 10.1016/j.ccr.2006.02.019. [DOI] [PubMed] [Google Scholar]
- 18.Brescia P, Ortensi B, Fornasari L, Levi D, Broggi G, Pelicci G. CD133 is essential for glioblastoma stem cell maintenance. Stem Cells. 2013;31:857–869. doi: 10.1002/stem.1317. [DOI] [PubMed] [Google Scholar]
- 19.Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM, Pastorino S, Purow BW, Christopher N, Zhang W, et al. Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serum-cultured cell lines. Cancer Cell. 2006;9:391–403. doi: 10.1016/j.ccr.2006.03.030. [DOI] [PubMed] [Google Scholar]
- 20.Ligon KL, Huillard E, Mehta S, Kesari S, Liu H, Alberta JA, Bachoo RM, Kane M, Louis DN, Depinho RA, et al. Olig2-regulated lineage-restricted pathway controls replication competence in neural stem cells and malignant glioma. Neuron. 2007;53:503–517. doi: 10.1016/j.neuron.2007.01.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Yuan H, Corbi N, Basilico C, Dailey L. Developmental-specific activity of the FGF-4 enhancer requires the synergistic action of Sox2 and Oct-3. Genes Dev. 1995;9:2635–2645. doi: 10.1101/gad.9.21.2635. [DOI] [PubMed] [Google Scholar]
- 22.Lanzkron SM, Collector MI, Sharkis SJ. Hematopoietic stem cell tracking in vivo: a comparison of short-term and long-term repopulating cells. Blood. 1999;93:1916–1921. [PubMed] [Google Scholar]
- 23.Pece S, Tosoni D, Confalonieri S, Mazzarol G, Vecchi M, Ronzoni S, Bernard L, Viale G, Pelicci PG, Di Fiore PP. Biological and molecular heterogeneity of breast cancers correlates with their cancer stem cell content. Cell. 2010;140:62–73. doi: 10.1016/j.cell.2009.12.007. [DOI] [PubMed] [Google Scholar]
- 24.Ono Y, Masuda S, Nam HS, Benezra R, Miyagoe-Suzuki Y, Takeda S. Slow-dividing satellite cells retain long-term self-renewal ability in adult muscle. J Cell Sci. 2012;125:1309–1317. doi: 10.1242/jcs.096198. [DOI] [PubMed] [Google Scholar]
- 25.Cicalese A, Bonizzi G, Pasi CE, Faretta M, Ronzoni S, Giulini B, Brisken C, Minucci S, Di Fiore PP, Pelicci PG. The tumor suppressor p53 regulates polarity of self-renewing divisions in mammary stem cells. Cell. 2009;138:1083–1095. doi: 10.1016/j.cell.2009.06.048. [DOI] [PubMed] [Google Scholar]
- 26.Dembinski JL, Krauss S. Characterization and functional analysis of a slow cycling stem cell-like subpopulation in pancreas adenocarcinoma. Clin Exp Metastasis. 2009;26:611–623. doi: 10.1007/s10585-009-9260-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Roesch A, Fukunaga-Kalabis M, Schmidt EC, Zabierowski SE, Brafford PA, Vultur A, Basu D, Gimotty P, Vogt T, Herlyn M. A temporarily distinct subpopulation of slow-cycling melanoma cells is required for continuous tumor growth. Cell. 2010;141:583–594. doi: 10.1016/j.cell.2010.04.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Rietze RL, Reynolds BA. Neural stem cell isolation and characterization. Methods Enzymol. 2006;419:3–23. doi: 10.1016/S0076-6879(06)19001-1. [DOI] [PubMed] [Google Scholar]
- 29.Vik-Mo EO, Sandberg C, Olstorn H, Varghese M, Brandal P, Ramm-Pettersen J, Murrell W, Langmoen IA. Brain tumor stem cells maintain overall phenotype and tumorigenicity after in vitro culturing in serum-free conditions. Neuro Oncol. 2010;12:1220–1230. doi: 10.1093/neuonc/noq102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Driessens G, Beck B, Caauwe A, Simons BD, Blanpain C. Defining the mode of tumour growth by clonal analysis. Nature. 2012;488:527–530. doi: 10.1038/nature11344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Louis SA, Rietze RL, Deleyrolle L, Wagey RE, Thomas TE, Eaves AC, Reynolds BA. Enumeration of neural stem and progenitor cells in the neural colony-forming cell assay. Stem Cells. 2008;26:988–996. doi: 10.1634/stemcells.2007-0867. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.