The evidence of glioblastoma heterogeneity

Abstract

Cancers are composed of heterogeneous combinations of cells that exhibit distinct phenotypic characteristics and proliferative potentials. Because most cancers have a clonal origin, cancer stem cells (CSCs) must generate phenotypically diverse progenies including mature CSCs that can self-renew indefinitely and differentiated cancer cells that possess limited proliferative potential. However, no convincing evidence exists to suggest that only single CSCs are representative of patients' tumors. To investigate the CSCs' diversity, we established 4 subclones from a glioblastoma patient. These subclones were subsequently propagated and analyzed. The morphology, the self-renewal and proliferative capacities of the subclones differed. Fluorescence-activated cell sorting and cDNA-microarray analyses revealed that each subclone was composed of distinct populations of cells. Moreover, the sensitivities of the subclones to an inhibitor of epidermal growth factor receptor were dissimilar. In a mouse model featuring xenografts of the subclones, the progression and invasion of tumors and animal survival were also different. Here, we present clear evidence that a brain tumor contains heterogeneous subclones that exhibit dissimilar morphologies and self-renewal and proliferative capacities. Our results suggest that single cell-derived subclones from a patient can produce phenotypically heterogeneous self-renewing progenies in both in vitro and in vivo settings.

Figure 1. Various clones grown in vitro.

(a) Distinct cellular morphologies observed after plating dissociated tumor cells in culture dishes. (b) Procedures used in establishing the 4 clones. Immediately after tumor spheres were formed from the patient's tissue, mechanically dissociated single cells were plated in small culture dishes. (c) The clones #2 and #4 formed sphere-like aggregates, whereas #3 and #5 attached to the uncoated culture dishes. (d) The 4 clones expressed the stem-cell marker nestin (red, nestin: blue, DAPI). (e) Cell-doubling time. Scale bars = 100 μm.

Figure 2. Dissimilar tumorigenesis in the in vivo animal model.

(a) Kaplan-Meier survival plots; 3 mice were used for cells of each clone. (b) Representative brain tumors of NOD-SCID mice harboring xenografts of the clones #2 (upper left), #3 (lower left), #4 (upper right) and #5 (lower right); H.E. staining, arrows: tumor, scale bar = 5 mm. (c) A representative xenograft of the clone #4; nestin-positive cells infiltrated the contralateral hemisphere (arrows). Scale bars = 5 mm (left), 1 mm (middle), 50 μm (right).

Figure 3. Analysis of cell-surface markers and cDNA arrays.

Flow-cytometry data were collected in at least triplicate and at distinct culture periods in order to avoid ongoing culture selection. ORI: original cells not sorted into single cells; DIF: differentiated cells obtained using 10% serum. (a) CD133 expression (no statistically significant difference). (b) CD44 expression. (c) CD24 expression (*P < 0.01). (d) CD56 expression (*P < 0.01). (e) Analysis of cDNA arrays. Fold-changes in gene expression were calculated as the ratio of the signal values of the original cells to the value of each of the clonal cells after duplication.

Figure 4. Inhibition of EGF-EGFR/PI3K/Akt and/or MAPK-Erk1/2 pathways by an EGFR inhibitor.

(a) Cell numbers determined after treatment with the EGFR inhibitor gefitinib (1 μM); cells were counted after one week. *P < 0.01. Cont: control; EGFRi: EGFR inhibitor. (b) Phosphorylation of Akt, Stat3, Erk1/2, and Smad1/5 in cells incubated with 1 μM gefitinib for 1 h. C: control, Ei: EGFR inhibitor. Full-length blots are shown in Supplementary Fig. 4.