A Prominin-1-Rich Pediatric Glioblastoma: Biologic Behavior Is Determined by Oxygen Tension-Modulated CD133 Expression but Not Accompanied by Underlying Molecular Profiles

Abstract

Few studies on the biologic and molecular properties of pediatric glioblastoma have been performed. Until now, differential genomic analysis of CD133(+)ve and CD133(-)ve fractions has not been described in pediatric glioma. We hypothesize not only that the presence of CD133 could be the source of tumor resistance but also that maintenance of this molecule by hypoxia dictates cellular and molecular behavior. From a series of human glioblastoma biopsies investigated, only one, IN699 (from a pediatric glioblastoma), generated greater than 4% of the total cell volume as CD133(+)ve cells. Using this pediatric glioblastoma, containing unprecedented high levels of the putative brain tumor stem cell marker CD133, as a study model, we report biologic and molecular characteristics of the parent culture and of CD133(+)ve and CD133(-)ve populations derived therefrom under atmospheric and hypoxic culture conditions. Immunocytochemistry and flow cytometry were performed with antigenic markers known to characterize neural stem cells and associated glioma behavior. Behavioral analysis was carried out using proliferation, adhesion, migration, and invasion assays. Cell cycle analysis and array comparative genomic hybridization were used to assess copy number profiles for parental cells and CD133(+)ve and CD133(-)ve fractions, respectively. With regard to invasion and proliferation, CD133(+)ve and CD133(-)ve fractions were inversely proportional, with a significant increase in invasive propensity within the CD133(-)ve cells (P < .005) and a significant increase in proliferation within CD133(+)ve cells (P < .005). Our observations indicate identical genomic imbalances between CD133(+)ve and CD133(-)ve fractions. Furthermore, our research documents a direct link between decreasing oxygen tension and CD133 expression.

Figure 1

(A) Growth of the parental cell culture established under NSCs conditions and 3% O₂ plus 5% CO₂. When cultured under defined stem cell conditions, the parental culture forms true spheres, which are derived from a single cell, as well as some spherically aggregated clusters (B) representative of the classic neurosphere.

Figure 2

The percentage mean (percentage-gated expression) of cells displaying positivity for each antigen for the CD133⁺ve (blue) and CD133^-ve (red) cell fractions, cultured under hypoxia (3% O₂ and 5% CO₂), analyzed by flow cytometry. Student's t test was used for statistical evaluation (*P < .005). All data presented are representative of three independent experiments performed in triplicate (n = 3). All values are shown as arithmetic means ± standard errors of the mean.

Figure 3

Fluorescence immunocytochemistry of isolated CD133⁺ve and CD133^-ve cells characterized with stem cell markers and markers of multilineage differentiation. The cellular location of Nestin seems to be within the cytosol and cellular processes of both isolated cell cultures (A and B). GFAP was highly expressed in both cell types, localized within the cytoplasm and cellular processes (C and D). Fluorescence staining for Olig-4 (E and F) was also observed within both cell cultures, primarily located around the periphery of the nucleus. Staining for stemness and multilineage differentiation are shown in green (AlexaFluor 488) and counterstained with Hoescht blue.

Figure 4

(A) Percentage mean (percentage-gated expression) of the parental IN699 cells displaying positivity for the CD133 antigen under atmospheric and hypoxic oxygen tensions as analyzed by flow cytometry. (B) Mean amount of CD133 antigen expression as shown by an increase in fluorescence fold intensity (i.e., overall fluorescence level of the total cell population) in the parental pediatric GB cell culture (IN699) under atmospheric and hypoxic culture conditions as analyzed by flow cytometry. Student's t test was used for statistical evaluation (P < .005). All data presented are representative of three independent experiments performed in triplicate (n = 3). All values are shown as arithmetic means ± standard errors of the mean.

Figure 5

Characterization of the parental pediatric GB cell culture (IN699) with CD133 (green) and counterstained with Hoescht blue under atmospheric (A) and hypoxic (B) oxygen tensions with fluorescent immunocytochemistry. The cellular location of CD133 was typically identified as spanning the plasma membrane under both oxygen tensions with a significant increase in CD133 expression with hypoxia. Scale bar, 32 µm.

Figure 6

An overview of the DNA copy number alterations in IN699 for (A) CD133⁺ve and (B) CD133^-ve cell fractions. Green blocks indicate regions of chromosome gain and red blocks regions of loss.

Figure 7

(A) Percentage mean (percentage-gated expression) of the parental pediatric GB cell culture (IN699) expressing the antigens CD133, HIF-1α, and HIF-2α at the following time points: 0, 24, 48, 72, and 96 hours with relation to the incubation period with hypoxia. Although the levels of HIF-2α remain constant with hypoxia, the levels of HIF-1α seem to be inversely proportional to decreasing oxygen tension. A positive correlation between decreasing oxygen tension and CD133 is also apparent. Student's t test was used for statistical evaluation (P < .005). All data presented are representative of three independent experiments performed in triplicate (n = 3). All values are shown as arithmetic means ± standard errors of the mean. (B) Mean amount of CD133, HIF-1α, and HIF-2α antigen expression as shown by an increase in fluorescence fold intensity (i.e., overall fluorescence level of the total cell population) in the parental pediatric GB cell culture (IN699) at the following time points: 0, 24, 48, 72, and 96 hours with relation to the incubation period with hypoxia. A positive correlation between HIF-2α and CD133 was observed with decreasing oxygen tension. However, in contrast, an inversely proportional relationship between CD133 and HIF-1α with decreasing oxygen tension can be seen. Student's t test was used for statistical evaluation (P < .005). All data presented are representative of three independent experiments performed in triplicate (n = 3). All values are shown as arithmetic means ± standard errors of the mean.

Figure 8

(A) PI analysis of the parental pediatric GB cell culture (IN699) with BrdU under atmospheric and hypoxic growth conditions. (B) PI analysis of CD133⁺ve and CD133^-ve cells with BrdU with hypoxia. (C) Cell attachment propensity analysis of the parental pediatric GB cell culture (IN699) under atmospheric and hypoxic oxygen tensions, cultured in the presence of vitronectin, fibronectin, Geltrex, and BSA (control sample). (D) Cell attachment propensity analysis of CD133⁺ve and CD133^-ve cell fractions, cultured in the presence of vitronectin, fibronectin, Geltrex, and BSA (control sample) under hypoxic oxygen tension. (E) Invasion propensity analysis of the parental pediatric GB cell culture (IN699) under atmospheric and hypoxic oxygen tensions, cultured in the presence of vitronectin, fibronectin, and Geltrex. (F) Invasion propensity analysis of isolated CD133⁺ve and CD133^-ve cell fractions, cultured in the presence of vitronectin, fibronectin, and Geltrex with decreasing oxygen tensions. (G) PI analysis of the parental pediatric GB cell culture (IN699) under atmospheric and hypoxic oxygen tensions, cultured in the presence of vitronectin, fibronectin, Geltrex, and BSA (control sample). (H) PI analysis of isolated CD133⁺ve and CD133^-ve cell fractions, cultured in the presence of vitronectin, fibronectin, Geltrex, and BSA (control sample) with decreasing oxygen tension. All data presented in Figure 7 are representative of three independent experiments performed in triplicate (n = 3). All values are shown as arithmetic means ± standard errors of the mean. Student's t test was used for statistical evaluation (P < .005).