Oncogenic signaling by NOTCH is elevated in brain tumor-initiating cells (BTIC) in malignant glioma, but the mechanism of its activation is unknown. Here we provide evidence that tenascin-C (TNC), an extracellular matrix protein prominent in malignant glioma, increases NOTCH activity in BTIC to promote their growth. We demonstrate the proximal localization of TNC and BTIC in human glioblastoma specimens and in orthotopic murine xenografts of human BTIC implanted intracranially. In tissue culture, TNC was superior amongst several extracellular matrix proteins in enhancing the sphere-forming capacity of glioma patient-derived BTIC. Exogenously applied or autocrine TNC increased BTIC growth through an α2β1 integrin-mediated mechanism that elevated NOTCH ligand Jagged1 (JAG1). Microarray analyses and confirmatory PCR and Western analyses in BTIC determined that NOTCH signaling components including JAG1, ADAMTS15, and NICD1/2 were elevated in BITC after TNC exposure. Inhibition of γ-secretase and metalloproteinase proteolysis in the NOTCH pathway, or silencing of α2β1 integrin or JAG1, reduced the proliferative effect of TNC on BTIC. Collectively, our findings identified TNC as a pivotal initiator of elevated NOTCH signaling in BTIC and define the establishment of a TN-α2β1-JAG1-NOTCH signaling axis as a candidate therapeutic target in glioma patients. Cancer Res; 77(12); 3231–43. ©2017 AACR.

Malignant gliomas are thought to be maintained by a rare population of transformed stem-like cells referred to as glioma stem cells or brain tumor-initiating cells (BTIC; refs. 1, 2). BTICs exhibit increased resistance to radiation (3, 4) and chemotherapy (5, 6) as compared with their more differentiated transformed progeny (henceforth referred to as differentiated glioma cells), and they account for glioma recurrence following efficient chemotherapy in mice (7). Thus, the identification of mechanisms that maintain the growth of BTICs will be important for therapeutic purposes.

NOTCH proteins (NOTCH 1-4 in mammals) are transmembrane proteins activated by ligands such as delta 1-4 and serrate/jagged 1 (JAG1) and JAG2. The interaction of NOTCH with its ligand results in the proteolytic cleavage of the NOTCH receptor and release of the intracellular domain (NOTCH-intracellular domain, NICD) that translocates to the nucleus to affect NOTCH-dependent transcription of genes (8, 9). Proteases implicated in the generation of NICD include metzincin metalloproteinases [e.g., matrix metalloproteinases (MMP), a disintegrin and metalloproteinases (ADAM), and ADAMs with thrombospondin motifs (ADAMTS)] at the proximal extracellular loop of NOTCH, and the γ-secretase complex at the intracellular juxtamembrane (10).

The NOTCH signaling pathway maintains normal stem cell turnover in the brain. Excessive or aberrant NOTCH signaling, however, is a feature of tumorigenesis (8). The BTICs present in malignant glioma exhibit elevated NOTCH activity (11–13) that has recently been implicated in BTIC growth and self-renewal, and their resistance to radiation and chemotherapy (11, 14, 15). However, the cause of the elevated NOTCH activation in BTICs is unknown.

The microenvironment of tumors includes the surrounding extracellular matrix (ECM). For most tumor types, the availability of ECM proteins is a critical factor for tumorigenesis (16–18) as ECM components activate integrins on the cell surface to trigger survival and proliferative signaling, and because the ECM sequesters a rich source of growth factors. The ECM in malignant glioma is unique and includes vitronectin, collagen-I and -IV, osteopontin, and tenascin-C (TNC; refs. 19, 20). Of these, arguably the most important component is TNC as its amount in situ is correlated proportionally with increased glioma grade (21–25) and proliferative capacity (21, 23, 26–29). Furthermore, antibodies that block TNC interactions reduce the motility of differentiated glioma cells in culture and suppress the growth of gliomas in mice (30). These preclinical results have evolved to clinical trials of glioma patients using locally introduced small interfering RNA (siRNA) to TNC (31), or 131I-labeled anti-TNC antibody (32), where the survival results have been encouraging.

TNC also plays an important role in promoting tumor stemness. For instance, Wnt/β-catenin signaling has an important role in regulating the balance between differentiation and stemness in a number of adult stem cell niches (33), and TNC modulates Wnt/β-catenin signaling in glioma cells (27). In addition, TNC has been shown to elevate components of stem cell signaling, such as musashi homolog 1, a positive regulator of NOTCH signaling in breast cancer cells, thus linking TNC to NOTCH signaling pathway (34). Whether or not TNC can promote epithelial-to-mesenchymal transition in glioma stem-like cells is likely an important future area of research.

We have reported that TNC facilitates the invasiveness of differentiated glioma cells in culture, through regulating MMP-12 (35) and protein kinase C (36); moreover, we found that TNC promotes BTIC invasion through ADAM-9 and the c-Jun NH2-terminal kinase pathway (37). However, the role of TNC on BTIC biology has not been investigated in detail. A recent study has identified TNC as a novel marker for BTICs (38). Herein, we sought to test the hypothesis that the elevated NOTCH signaling that regulates BTIC growth is mediated by TNC. Our results implicate TNC in the maintenance of glioma stem cells in the cancer stem cell niche, by modulating NOTCH activity.

BTICs generated from human glioma patients

BTICs were generated and stable lines were formed from resected specimens of patients with malignant glioma as described before (39, 40). We used two BTIC lines for most of the tissue culture experiments in the present study designated BT025 and BT048 with divergent genetic background; these lines have been referred to previously as 25EF and 48EF, respectively (39). Other patient-derived BTIC lines used in culture or xenograft experiments included BT012, BT53M, BT067, BT069, BT075, BT134, BT143, BT147, BT157, and BT161 (Supplementary Figs. S1–S4; ref. 41). All of these lines were cultured chronologically, maintained, and authenticated to the present time within the University of Calgary BTIC Core directed by Dr. Sam Weiss and Dr. Greg Cairncross. Their stemness features have been previously described in our publications (39, 41) and our citations therein. Lines were initiated in culture from resected specimens in year 2005 (BT012 and BT025), 2007 (BT048, BT53M, BT067, BT069, and BT075), 2009 (BT134, BT143, and BT147), and 2010 (BT157 and BT161).

Neurosphere assay

To study the growth of BTICs, dissociated cells from BTIC spheres were plated at a density of 10,000 cells/100 μL in serum-free BTIC medium (39). BTICs were treated with or without TNC or other ECM proteins (10–50 μg/mL; Millipore), inhibitors or antibodies. We used soluble TNC in the majority of our study as a matter of convenience as our previous studies found that this ECM molecule exerts its effects on differentiated glioma cells either in the soluble form or when embedded in a 3-dimensional matrix of collagen gel (35, 37). Although TNC purified from the U251 human glioma line (Millipore, catalog # CC065, molecular weight 280–300, and glycosylated) was used in the majority of experiments unless otherwise stated, we also tested TNC (recombinant human, R&D Systems, catalog #3358-TC) from a different commercial source.

Cultures were maintained at 37°C in 5% CO2 incubator. After 72 hours, four random fields per well were photographed under a 10× objective in a phase contrast microscope, and the number of spheres over 60 μm in diameter was tabulated. In some experiments, cells were treated with metalloproteinase inhibitors [BB94 (500 nmol/L, British Biotech) or GM6001 (10 μmol/L, Calbiochem)], or a NOTCH/γ-secretase inhibitor (DAPT, with a final concentration of 2 or 5 μmol/L). TNC was added 1 hour after the addition of inhibitors. In selected experiments, cells were treated with neutralizing antibodies to α2β1, α9β1, and αVβ6 integrins or isotype antibodies (Abcam) followed by TNC after 1 hour.

Western blot analysis and immunoprecipitation

GBM tumor samples were used for Western blot analyses. The samples were probed with rabbit human TNC antibody as described before (37).

For immunoprecipitation, 500 μg of cell lysates was incubated for 3 hours at 4°C with 4 μg of the α2β1 antibody (Abcam) and 30 μL of protein A/G-agarose beads (Santa Cruz Biotechnology). To control for nonspecific binding, an isotype antibody replaced the primary antibody. The immunoprecipitates were subjected to Western blot analysis as described before (42).

Cell cycle analysis

BTIC cells were treated with or without TNC for 48 hours, and cell-cycle analysis was performed with propidium iodide using a standard flow cytometry protocol (39).

Microarray and data analysis

To determine the effect of TNC on BTIC gene expression, BT025 cells were treated with TNC for 6 hours and subjected to microarray analysis as described elsewhere (39). All microarray data have been submitted to the US National Institute of Health GEO database under accession number GSE94640.

Quantitative real-time PCR for JAG1

BTICs were lysed in 1 mL of TRIzol reagent (Invitrogen) by leaving plates at room temperature for 5 minutes before the content of the well was harvested and stored at −80°C prior to use. Following extraction, RNA was reverse transcribed and the resulting cDNA was used as a template for the Bio-Rad iCycler MyiQ detection system and 2× SYBR green mastermix (Qiagen). Primers (10× QuantiTect Primer Assay) were purchased from Qiagen. Expression of gene transcripts was normalized against at least two housekeeping genes, that is, GAPDH and β-actin. Relative expression levels for genes of interest were determined using the formula 2−ΔCT, where ΔCT = CT (gene of interest) − CT (housekeeping gene).

NOTCH activity

BTICs were plated in 12-well plates (100,000 cells/well) and treated with or without TNC either in the presence or absence of inhibitors or integrin antibodies. After 6 hours, cell lysate was prepared and analyzed for the detection of cleaved NOTCH product, NICD1 (Rabbit anti-cleaved NOTCH1 (1:1,000, Cell Signaling Technology) or NICD2 (rabbit anti-NOTCH2, cleaved; Millipore), by Western blot analysis following protein separation by gradient gel or 10% SDS-PAGE.

Small interfering RNA to TNC, JAG1, and α2β1 integrin

A predesigned small interfering RNA (siRNA, Santa Cruz Biotechnology), designated JAG1a siRNA, was used to target human JAG1. A second siRNA designated JAG1b siRNA was used to confirm the first siRNA results. Similarly, two predesigned siRNAs were used to target human TNC. For α2β1 integrin knockdown, two sets of predesigned silencer select siRNAs were used (Ambion, Life Technologies); each set comprised two siRNAs that targeted the respective α2 and β1 genes. A matrix-assisted laser desorption/ionization-time of flight mass spectrometer was used to identify the correct mass of the single-stranded RNA oligonucleotides. The annealed siRNAs were analyzed by nondenaturing PAGE. A negative control siRNA, composed of a 19-bp scrambled sequence with three deoxythymidine overhangs, was used; the sequences have no significant homology to any known gene sequences from mouse, rat, or human. For transfection with siRNA, BTICs were plated in 12-well plates and were incubated with 30 nmol/L siRNA and Lipofectamine (Invitrogen). After 24 hours, cells were harvested for neurosphere assays, Western blot, and cell-cycle analysis.

Lentiviral-mediated TNC knockdown

Lentivirus-mediated transfection was performed to stably knockdown TNC. Stable knockdown of TNC was carried out using shRNA constructs in BT025 and BT048 using Sigma Mission TNC shRNAs or Scrambled controls in pLKO.1-puro vector as described elsewhere. Briefly, the shRNA vector was cotransfected using Lipofectamine 2000 (Invitrogen) into HEK-293 cells with pMD2.G (VSV.G env) and pCMV-ΔR8.91. After a 12-hour transfection in DMEM, viral containing media were collected over 2 days in BTIC media. Collected media were then syringe-filtered with a 0.22-μm filter, and then ultracentrifuged at 26,000 RPM for 90 minutes at 4°C. Viral pellets were resuspended in BTIC media and added to cultures overnight. One μg/mL of puromycin (Invitrogen) was added to cultures 3 days after infection.

Detection of TNC by immunofluorescence or immunohistochemistry in human glioblastoma specimens

Paraffin sections were deparaffinized and subjected to antibody staining and confocal microscopy as described elsewhere (39). We evaluated paraffin sections from 3 autopsied glioblastoma patients (obtained from the University of Calgary Neuropathology archive) for the expression of TNC (using rabbit anti-TNC antibody, 1:500, Novus Biologicals) in proximity to presumed BTICs labeled for nestin (mouse anti-Nestin, 1:500, Chemicon) or CD133 (mouse anti-CD133, 1:500, Milteny Biotech). A fluorescence-conjugated secondary antibody was then applied to reveal the antigen. We also used a biotinylated secondary antibody, ABC reagent (Vectastain ABC kit, Vector Laboratories) and diaminobenzidine to detect antigens in glioblastoma specimens for corroboration.

Mice and BTIC implantation

BTIC spheres were dissociated into single cells using Accumax solution, and 10,000 viable cells in 2 μL of saline were stereotactically implanted into the right striatum of each 6–8 week-old female SCID mice (Charles River) as described elsewhere (37, 39). Animals were returned to their cages and allowed free access to food and water. Mice were weighed every other day and observed for symptoms of neurologic deficits; they were sacrificed 7 weeks after implantation while still asymptomatic. The whole brain was removed, cut into blocks, fixed in 10% buffered formalin, and embedded in paraffin. Sections of 6 μm were taken every 120 μm apart, through the entire brain. The sections were deparaffinized, rehydrated, and stained with hematoxylin and eosin (H&E) for general histology. Sections were also labeled for TNC (rabbit anti-TNC antibody, 1:500, Novus Biologicals), nestin (1:500, Chemicon), human implanted cells (mouse anti-human nucleolin, 1:500, R&D; ref. 38), or for Jagged1 (mouse anti-Jagged1, 1:100, Santa Cruz Biotechnology). Following a biotinylated secondary antibody, ABC reagent (Vectastain ABC kit, Vector Laboratories), and diaminobenzidine reaction, the slides were lightly counterstained with hematoxylin or nuclear fast red, dehydrated and mounted. All protocols were approved by the Animal Care Committee at the University of Calgary in accordance with research guidelines from the Canadian Council for Animal Care.

Statistical analyses

For analyses of differences in sphere formation in culture, the one-way ANOVA with post hoc Tukey comparisons was used for multiple groups, while the t test was used for comparisons of two groups.

BTICs are found in TNC-enriched areas in human glioblastoma specimens

TNC has been documented to be markedly elevated in malignant glioma compared with surrounding normal brain tissue (22), and we have demonstrated by Western blot analyses the substantial increase in TNC content in glioblastoma compared with nontransformed human brain specimens (37). However, the localization of TNC in relation to BTICs in human glioblastoma specimens has not been characterized. Figure 1A and B shows that by immunohistochemistry, the high expression of TNC within the glioblastoma of patient 1, where presumed BTICs, detected through CD133+ and nestin+ staining, was present. We corroborated by immunofluorescence microscopy the high expression of TNC in glioblastoma specimen from 2 other patients, in proximity to nestin-positive cells (Fig. 1C).

Figure 1.

Staining of human glioblastoma and xenografts shows BTICs to be in the vicinity of TNC. A, TNC expression in a resected human glioblastoma specimen showing elevated level of TNC protein (brown) compared with normal human brain tissue (original magnification,×100). B, At higher ×200 magnification, TNC in GBM specimen (patient 1) is in the same area as BTICs; the latter were identified by nestin and CD133 immunoreactivity. These sections were counterstained with hematoxylin (for nestin) or nuclear fast red (for TNC and CD133). C, Sections from two other autopsied GBM specimens show TNC (red) in proximity to nestin-positive cells (green); nuclei were labeled with DAPI in blue. D, TNC immunoreactivity (brown) was observed in the brain of mice implanted with human BTICs (BT048, BT53M, and BT147). Implanted BTICs were identified in adjacent sections by nestin staining and by the presence of human nucleolin (see Supplementary Fig. S1).

Figure 1.

Staining of human glioblastoma and xenografts shows BTICs to be in the vicinity of TNC. A, TNC expression in a resected human glioblastoma specimen showing elevated level of TNC protein (brown) compared with normal human brain tissue (original magnification,×100). B, At higher ×200 magnification, TNC in GBM specimen (patient 1) is in the same area as BTICs; the latter were identified by nestin and CD133 immunoreactivity. These sections were counterstained with hematoxylin (for nestin) or nuclear fast red (for TNC and CD133). C, Sections from two other autopsied GBM specimens show TNC (red) in proximity to nestin-positive cells (green); nuclei were labeled with DAPI in blue. D, TNC immunoreactivity (brown) was observed in the brain of mice implanted with human BTICs (BT048, BT53M, and BT147). Implanted BTICs were identified in adjacent sections by nestin staining and by the presence of human nucleolin (see Supplementary Fig. S1).

Close modal

To further emphasize that TNC and BTICs are in proximity for potential interactions to occur, we determined the expression of TNC in SCID mice implanted with human BTIC cells (BT048, BT53M, BT134, BT143, BT147, BT157, and BT161) in the right striatum. We sacrificed asymptomatic mice 7 weeks after implantation and found tumor mass in mice. TNC expression was detected in brain areas where nestin+ and human nucleolin+ cells were found (Fig. 1D and Supplementary Fig. S1). Elevated level of TNC expression was observed in the tumor mass and at the invasive front of the tumor (Supplementary Figs. S2 and S3).

We addressed the localization of TNC in relation to cells or the ECM. As TNC is an ECM component, it would be reasonable to find some staining outside of cells (e.g., the diffuse TNC staining in Fig. 1A). However, during its production and subsequent export out of cells, some signal for TNC immunoreactivity should be inside cells. This appears to be the case in Fig. 1C (patient 2) where there are yellow profiles that are overlap of nestin-positive BTICs (green) and TNC (red). The latter result also suggests that BTICs in glioblastomas have the potential to produce TNC.

TNC promotes growth of glioma patient-derived BTICs

Although others have reported on the capacity of TNC to promote the invasiveness and proliferation of differentiated glioma cells, we have not found reports of the potential of TNC to promote BTIC growth. Several BTIC lines isolated from individual glioma patients with diverse genetic mutation profiles were used to address whether their responses were uniform to TNC. These lines were assessed to be BTICs rather than differentiated glioma cells based on their expression of nestin, Sox2 and Musashi-1; they exhibited self-renewal property and multilineage capacity by differentiation into GFAP+ and βIII tubulin+ cells; and they formed invasive tumors when implanted orthotopically as xenografts into mice (39, 40). The BTICs grew as progressively larger spheres in suspension cultures and we mechanically dissociated the enlarging spheres into single cells periodically, so that these could regrow as spheres to allow the lines to be maintained indefinitely.

To determine the role of ECM proteins on BTIC growth, the patient-derived spheres were plated as dissociated cells (10,000 cells/100 μL) and exposed to 10 μg/mL TNC, type III collagen (CL III), type IV collagen (CL IV), fibronectin (FN), vitronectin (VN), or chondroitin sulfate proteoglycans (CSPG). Using a sphere size cut off of 60 μm for ease of quantitation (39), we determined that both TNC and CL IV stimulated sphere formation of the BT025 and BT048 lines (Fig. 2A–F), with TNC having a greater effect. Indeed, a sphere after 3 days of exposure to TNC (Fig. 2C) regularly contains ∼1,500 cells. TNC from a different commercial source also promoted sphere formation of BTICs (Supplementary Fig. S4, D). CSPG promoted sphere formation in one of Two lines, whereas cells grew as an adherent monolayer on VN. Moreover, TNC increased the number of viable cells (Fig. 2E), and the total number of cells in the S and G2–M phases of the cell cycle (Fig. 2F and G). Importantly, TNC-enhanced sphere formation was found across all human-derived BTIC lines tested (Supplementary Fig. S4). Although TNC increased the number of cells in the S-phase of the cell cycle supporting a proliferative effect, an additional effect of TNC on cell survival is not ruled out.

Figure 2.

Of several ECM proteins, TNC most efficiently increases the sphere-forming capacity of BTICs. A and B, TNC stimulates sphere-forming capacity of the BT025 and BT048 lines in comparison to other ECM proteins or control (no ECM); 3 days treatment, *, P < 0.05; **, P < 0.01; ***, P < 0.001 compared with control (ANOVA with Tukey's post hoc comparisons). The TNC enhancement of sphere growth is pictorially depicted in C for the BT025 cell line. D, TNC also stimulated growth of BTIC when anchored in a 3-dimensional (3D) matrix of type I collagen (CL). The effect of soluble TNC on BTIC growth was supported by Alamar blue assays (E), cell counts (F), and PI flow cytometry (G). **, P < 0.01; ***, P < 0.001 compared with controls (t test, D–G). Error bars, SEM of 3–4 analyses.

Figure 2.

Of several ECM proteins, TNC most efficiently increases the sphere-forming capacity of BTICs. A and B, TNC stimulates sphere-forming capacity of the BT025 and BT048 lines in comparison to other ECM proteins or control (no ECM); 3 days treatment, *, P < 0.05; **, P < 0.01; ***, P < 0.001 compared with control (ANOVA with Tukey's post hoc comparisons). The TNC enhancement of sphere growth is pictorially depicted in C for the BT025 cell line. D, TNC also stimulated growth of BTIC when anchored in a 3-dimensional (3D) matrix of type I collagen (CL). The effect of soluble TNC on BTIC growth was supported by Alamar blue assays (E), cell counts (F), and PI flow cytometry (G). **, P < 0.01; ***, P < 0.001 compared with controls (t test, D–G). Error bars, SEM of 3–4 analyses.

Close modal

The above experiments involved the provision of soluble TNC to cells to simulate the release of TNC from its cellular source to affect autocrine or paracrine interactions. To represent TNC anchored to the ECM, we encased TNC along with BTICs within a 3-dimensional matrix of collagen gel as described previously (35). In this environment, the increase in cell numbers in response to TNC was still evident (Fig. 2D).

Integrin-mediated TNC-induced autocrine growth of BTICs

We examined the receptors responsible for TNC-mediated BTIC growth. TNC binds to integrin receptors on cells, including α2β1, α9β1, and αVβ6 (43). BTICs were treated with neutralizing antibodies to these integrins in the presence of exogenously added TNC, and the number of spheres was quantified. Cells treated with α2β1-neutralizing antibody in the presence of exogenously TNC showed significant reduction of BTIC growth compared with control (Fig. 3A and B), whereas α9β1- and αVβ6-neutralizing antibodies had no effect. Thus, the data suggest that exogenously added TNC acted through α2β1 integrin on BTICs for growth. Unexpectedly, the α2β1 integrin antibody added to BTICs in the absence of exogenously applied TNC also diminished sphere formation (Fig. 3A and B), suggesting that TNC was produced by BTICs to regulate autocrine growth. Consistent with this observation, we previously reported that TNC was detected in the conditioned media of BTICs (37). As noted above, there is overlap of nestin and TNC immunoreactivity in a glioblastoma specimen (Fig. 1C).

Figure 3.

TNC is produced by BTICs for autocrine growth. A and B, Basal (no TNC added) or TNC-promoted BTIC growth (sphere-forming capacity or cell counts) of both the BT025 and BT048 lines is abrogated by neutralizing α2β1, but not α9β1 or αVβ6 integrin. ***, P < 0.001 compared with control. C and D, Knockdown by siRNA of α2β1 integrin in BT025 and BT048 cells, as determined by FACS analysis (red, background staining; blue, positive population). E and F, When α2β1 integrin knockdown cells were subjected to neurosphere assay either in presence or absence of TNC, both sphere growth and total cell numbers were reduced. ***, P < 0.001 compared with control siRNA. G and H, TNC knockdown with siRNA as evaluated by Western blot and cell-cycle progression of the transfected cells. I, Control siRNA– or TNC siRNA–treated cells were subjected to neurosphere assay and evaluated after 72 hours. TNC siRNA–treated cells showed significant reduction of sphere formation compared with control, and this could be overcome by exogenous TNC. Control + TNC was used as a positive control for TNC stimulated BTIC growth. ***, P < 0.001 compared with their respective siRNA control. J, Immunofluorescence staining showing focal adhesion kinase (FAK) expression 6 hours after TNC stimulation of freshly dissociated cells compared with control (a representative image from four cultures each). All error bars represent SEM of 3–4 analyses, with statistical evaluation through ANOVA with Tukey multiple comparisons.

Figure 3.

TNC is produced by BTICs for autocrine growth. A and B, Basal (no TNC added) or TNC-promoted BTIC growth (sphere-forming capacity or cell counts) of both the BT025 and BT048 lines is abrogated by neutralizing α2β1, but not α9β1 or αVβ6 integrin. ***, P < 0.001 compared with control. C and D, Knockdown by siRNA of α2β1 integrin in BT025 and BT048 cells, as determined by FACS analysis (red, background staining; blue, positive population). E and F, When α2β1 integrin knockdown cells were subjected to neurosphere assay either in presence or absence of TNC, both sphere growth and total cell numbers were reduced. ***, P < 0.001 compared with control siRNA. G and H, TNC knockdown with siRNA as evaluated by Western blot and cell-cycle progression of the transfected cells. I, Control siRNA– or TNC siRNA–treated cells were subjected to neurosphere assay and evaluated after 72 hours. TNC siRNA–treated cells showed significant reduction of sphere formation compared with control, and this could be overcome by exogenous TNC. Control + TNC was used as a positive control for TNC stimulated BTIC growth. ***, P < 0.001 compared with their respective siRNA control. J, Immunofluorescence staining showing focal adhesion kinase (FAK) expression 6 hours after TNC stimulation of freshly dissociated cells compared with control (a representative image from four cultures each). All error bars represent SEM of 3–4 analyses, with statistical evaluation through ANOVA with Tukey multiple comparisons.

Close modal

We further corroborated the role of α2β1 integrin in TNC-stimulated BTIC growth using RNAi approaches. Indeed, the knockdown of α2β1 integrin attenuated the stimulatory effect of TNC on BT025 and BT048 growth, both in presence and absence of TNC (Fig. 3C–F).

We evaluated the TNC–α2β1 integrin interaction further using lenti-viral-mediated knockdown of TNC. In TNC-silenced shTNC BTICs, the addition of the function blocking antibody to α2β1 integrin in both BT025 and BT048 lines no longer reduced BTIC growth below the level of control, supporting the above contention that TNC was produced by BTICs to regulate autocrine growth (Supplementary Fig. S5). Furthermore, exogenous TNC that promoted BTIC growth in TNC-silenced cells was blocked of this activity in the presence of the α2β1 integrin antibody (Supplementary Fig. S5). These results support a critical role for α2β1 integrin in mediating autocrine or exogenously applied TNC in stimulating BTIC growth.

BTICs with knockdown of α2β1 integrin or those exposed to the α2β1 integrin blocking antibody, with or without TNC exposure, had lower number of cells in the S-phase and corresponding more in G1 of the cell cycle, and some death was evident after 3 days, indicating that cell cycle block and eventual death occurred. Notably, α2β1 integrin seems to be predominantly expressed in BTICs (BT025: 99.4% of cells; BT048: 65.5%) compared with other TNC binding integrins (Fig. 3C and D and Supplementary Fig. S5). TNC is also known to bind the αVβ3 integrin; very low expression of αVβ3 was detected in BTIC lines (Supplementary Fig. S5), and a neutralizing antibody to αVβ3 integrin did not alter BTIC growth upon TNC stimulation.

We evaluated TNC-mediated autocrine signaling further by reducing TNC expression in BTICs using siRNAs (Fig. 3G). Consequently, cells were reduced in cell-cycle progression (Fig. 3H) and they formed spheres less readily; the exogenous application of TNC to TNC siRNA-treated BTICs resumed their capacity for sphere formation (Fig. 3I).

Finally, to begin to provide data (see also results below) that soluble TNC interacting with integrin promotes signaling in BTICs, we evaluated BTICs for focal adhesion kinase, a known downstream target of this interaction. We noted by immunofluorescence the pronounced focal adhesion kinase expression in TNC-treated BTICs (Fig. 3J).

TNC upregulates NOTCH ligand expression in BTIC

We subjected BTICs to global gene expression analysis to provide insights into the mechanisms by which TNC stimulates BTIC growth. After 6 hours of TNC treatment, 917 genes (P < 0.05) were differentially expressed compared with controls (Fig. 4A; GEO accession number GSE94640), with 178 genes altered by a fold change of at least 1.5 (Supplementary Tables S1 and S2); these included previously known genes that regulate glioma growth, such as SMAD7, ID2, TGFβ, ADAMTS15, and the NOTCH ligand, JAG1. Pathway analysis using Panther database demonstrated several signaling pathways were upregulated (P < 0.05) with TNC treatment, including the Notch signaling pathway (Fig. 4B). When analyzing the data using a volcano plot of fold change (log2) of genes against their statistical significance, two of the most highly upregulated genes, JAG1 and ADAMTS15, are key components of NOTCH signaling pathways, and other upregulated genes are SKIL and ELK3 implicated for TGFβ and interleukin signaling pathways, respectively (Fig. 4C and Supplementary Table S1) reported to play a critical role in glioma stem cell proliferation and self-renewal (12, 44). The increase in expression of JAG1 was confirmed by PCR transcript and protein analysis (Fig. 4D and E).

Figure 4.

A, Microarray analysis of BTICs 6 hours after exposure to TNC. Hierarchical clustering upon 6 hours of TNC treatment (GEO accession number GSE94640). B, Upregulated genes (fold change ≥1.5) were classified according to their involvement in different pathways via Panther database. Numbers show the percentage of genes involved in the respective pathway. C, Volcano plot of all present genes. The most significantly upregulated genes are shown in blue. Among them, four genes are involved in different signaling pathways. The results were obtained from three experiments performed at different times, where each experiment involved a control and a TNC-treated culture in order to capture consistent changes. D, PCR validation of JAG1 gene in BT025 cells (GADPH normalized); **, P < 0.001 compared with control (t test); error bars, SEM. E, The JAG1 expression was increased as well with TNC treatment at the protein level (BT025 and BT048 lines). F, Metacore analysis of microarray data, showing interactome pathway in TNC-treated BT025 cells. G, Brain samples from 6 GBM patients; all had evidence of JAG1, NICD1, and NICD2 expression.

Figure 4.

A, Microarray analysis of BTICs 6 hours after exposure to TNC. Hierarchical clustering upon 6 hours of TNC treatment (GEO accession number GSE94640). B, Upregulated genes (fold change ≥1.5) were classified according to their involvement in different pathways via Panther database. Numbers show the percentage of genes involved in the respective pathway. C, Volcano plot of all present genes. The most significantly upregulated genes are shown in blue. Among them, four genes are involved in different signaling pathways. The results were obtained from three experiments performed at different times, where each experiment involved a control and a TNC-treated culture in order to capture consistent changes. D, PCR validation of JAG1 gene in BT025 cells (GADPH normalized); **, P < 0.001 compared with control (t test); error bars, SEM. E, The JAG1 expression was increased as well with TNC treatment at the protein level (BT025 and BT048 lines). F, Metacore analysis of microarray data, showing interactome pathway in TNC-treated BT025 cells. G, Brain samples from 6 GBM patients; all had evidence of JAG1, NICD1, and NICD2 expression.

Close modal

Metacore analysis of the microarray data was used to probe interactive pathways or interactome (Fig. 4F). Mining of the microarray data publicly available from the Oncomine (Compendia Biosciences, http://www.oncomine.org/) database revealed that glioma patients with higher TNC expression have a significantly poorer survival advantage. The mining data also showed high JAG1 and NOTCH1 expression in glioma patients (Supplementary Fig. S6), and improved survival for glioma patients with low expression of JAG1 or NOTCH1 based on other available data sets from Rembrandt (Currently available from G-Doc plus database) or The Cancer Genome Atlas (TCGA; using Agilentg Platform; Supplementary Fig. S7). Our microarray data also showed increased expression of α2 integrin (ITGA2) with TNC treatment (Supplementary Table S1). Notably, TCGA data analysis showed significant correlation between TNC and NOTCH1, JAG1, or α2β1 integrin in GBM samples that strongly support our findings of these interacting components (Supplementary Fig. S8). To validate these findings, we analyzed tumor samples from 6 GBM patients and readily observed JAG1 expression. Moreover, 4 of these patients showed elevated levels of active or cleaved forms of NOTCH (NICD1 and NICD2 expression; Fig. 4G).

Interference of NOTCH activation reduces TNC-enhanced BTIC growth

Binding of ligand to the NOTCH receptor results in the stimulation of its cleavage by metalloproteinases and γ-secretase. This results in the generation of the NICD that can translocate to the nucleus to modulate gene expression. We assayed NOTCH activation by measuring NICD1 and NICD2 in BTICs exposed to TNC for 6 hours and found elevated levels of this cleavage fragment (Fig. 5A). Moreover, an inhibitor of the γ-secretase complex, DAPT (N-[N-(3,5-Diflurophenaacetyl-L-alanyl)]-S-phenylglycine t-Butyl Ester), decreased BTIC sphere formation (Fig. 5B and Supplementary Fig. S9) with a concordant reduction in the total number of cells (Fig. 5C) in the presence of TNC. Addition of the metzincin metalloproteinase inhibitors BB94 and GM6001 also reduced TNC-stimulated BTIC growth in BT025 and BT048 lines (Fig. 5C and Supplementary Fig. S9). Collectively, these data suggest that NOTCH signaling is engaged following TNC treatment, and involved in BTIC growth.

Figure 5.

The TNC-stimulated NOTCH activation or growth of BTICs is reduced by blocking γ-secretase or metalloproteinases, or in α2β1 integrin knockdown cells. A, TNC stimulates NOTCH activity in BTICs as evidenced by elevation of cleaved NOTCH receptors, NICD1 and NICD2, after 6 hours of TNC treatment and evaluated through Western blot analysis. B and C, Addition of the NOTCH inhibitor, DAPT (2 or 5 μmol/L), or the metalloproteinase inhibitors, BB94 (500 nmol/L) or GM6001 (10 μmol/L), reduced total cell numbers under basal and TNC-stimulated condition. ***, P < 0.001 compared with TNC and compared with control (in absence of TNC experiment) for all the bars enveloped by the drawn line (ANOVA with Tukey's multiple comparisons). Error bars represent SEM of four analyses, and the results were reproduced in two other experiments. D, TNC-stimulated NOTCH activity was abrogated by a blocking antibody to α2β1 integrin after 6 hours of exposure to TNC in both the BT025 and BT048 lines (Supplementary Fig. S9). In contrast, isotype antibody treatment had no effect. Actin was used as loading control. Results were reproduced in another set of experiment. E, The TNC-stimulated JAG1 expression was abrogated by blocking γ-secretase (DAPT), metalloproteinase (GM6001), or α2β1 integrin. F, TNC-stimulated NOTCH activity was abrogated by siRNA knockdown of α2β1 integrin in BT025 cells. The result was reproduced two other times. G, When α2β1 integrin knockdown cells were treated with TNC, BTIC growth was significantly reduced; however, CLIV or LN had no effect in α2β1 knockdown cells for BTIC growth. H, Immunoprecipitation demonstrates the association of α2β1 integrin with TNC in both the BT025 and BT048 lines.

Figure 5.

The TNC-stimulated NOTCH activation or growth of BTICs is reduced by blocking γ-secretase or metalloproteinases, or in α2β1 integrin knockdown cells. A, TNC stimulates NOTCH activity in BTICs as evidenced by elevation of cleaved NOTCH receptors, NICD1 and NICD2, after 6 hours of TNC treatment and evaluated through Western blot analysis. B and C, Addition of the NOTCH inhibitor, DAPT (2 or 5 μmol/L), or the metalloproteinase inhibitors, BB94 (500 nmol/L) or GM6001 (10 μmol/L), reduced total cell numbers under basal and TNC-stimulated condition. ***, P < 0.001 compared with TNC and compared with control (in absence of TNC experiment) for all the bars enveloped by the drawn line (ANOVA with Tukey's multiple comparisons). Error bars represent SEM of four analyses, and the results were reproduced in two other experiments. D, TNC-stimulated NOTCH activity was abrogated by a blocking antibody to α2β1 integrin after 6 hours of exposure to TNC in both the BT025 and BT048 lines (Supplementary Fig. S9). In contrast, isotype antibody treatment had no effect. Actin was used as loading control. Results were reproduced in another set of experiment. E, The TNC-stimulated JAG1 expression was abrogated by blocking γ-secretase (DAPT), metalloproteinase (GM6001), or α2β1 integrin. F, TNC-stimulated NOTCH activity was abrogated by siRNA knockdown of α2β1 integrin in BT025 cells. The result was reproduced two other times. G, When α2β1 integrin knockdown cells were treated with TNC, BTIC growth was significantly reduced; however, CLIV or LN had no effect in α2β1 knockdown cells for BTIC growth. H, Immunoprecipitation demonstrates the association of α2β1 integrin with TNC in both the BT025 and BT048 lines.

Close modal

α2β1 integrin and NOTCH activity

From the observations that a function blocking antibody to α2β1 integrin, or the siRNA knockdown of α2β1 integrin, blocked the effect of TNC on BTIC growth, and that NOTCH activity was enhanced by TNC treatment, we evaluated whether blocking α2β1 integrin would alter NOTCH activity. Figure 5D shows elevated level of NOTCH activity (NICD1) following 6 hours of TNC treatment and this was abolished by the neutralizing antibody to α2β1 integrin. In contrast, a neutralizing antibody to α9β1 integrin did not alter TNC-stimulated NOTCH activity (Supplementary Fig. S9). The neutralizing antibody to α2β1 integrin attenuated TNC-stimulated JAG1 expression (Fig. 5E and Supplementary Fig. S9), which was also decreased in the presence of γ-secretase (DAPT) and metalloproteinase (GM6001) inhibitors. We could not obtain a consistent basal (in absence of TNC) expression of JAG1 to address whether the low level of JAG1 in unstimulated cells would also be reduced by these inhibitors.

We determined that BTICs deficient of α2β1 integrin through siRNA knockdown had reduced NOTCH activity (NICD1) compared with control siRNA-treated cells in response to TNC stimulation (Fig. 5F), or in absence of TNC (Supplementary Fig. S9, E). Unexpectedly, unlike the function blocking antibody to α2β1 integrin that ameliorated TNC-stimulated JAG1 expression (Fig. 5E and Supplementary Fig. S9), we did not find reduction of JAG1 expression in TNC-stimulated α2β1 knockdown cells; the turnover of JAG1 in siRNA transfected or antibody treated cells may be markedly different to account for this discrepancy.

α2β1 integrin also binds to CLIV and laminin (LN), but these ECM proteins unlike TNC did not alter BTIC growth in α2β1 integrin knockdown cells (Fig. 5G). It is possible that CLIV and LN promote BTIC growth through a α2β1-independent mechanism that compensates for lack of interaction between endogenous TNC and α2β1. Finally, the presence of TNC bound to the α2β1 integrin was demonstrated by coimmunoprecipitation experiments (Fig. 5H) although it is still conceivable that there is an intermediary molecule that anchors TNC to the α2β1 integrin. Overall, these results support the result that α2β1 integrin mediates the TNC-induced NOTCH activation in BTICs.

TNC as a regulator of NOTCH signaling for BTIC growth

Because TNC stimulates NOTCH activity and elevates the expression of the NOTCH1 ligand JAG1 in BTICs, we sought to determine the functional implication of TNC–JAG1–NOTCH1 link. We sought to stably knockdown TNC in BT025 and BT048 cells using lentiviral transduction. We used 4 different TNC shRNAs to reduce TNC (Supplementary Fig. S10) in both lines, and shTNC (1) and shTNC (3) were effective (Fig. 6A). Remarkably, JAG1 expression was reduced in the TNC knocked down clones but not in BTIC clones where TNC was not successfully lowered. This observation further links the functional association between TNC and NOTCH1 ligand JAG1 expression.

Figure 6.

The knockdown of JAG1 reduces TNC-stimulated BTIC growth. A, The reduction of TNC expression in BTIC lines (BT025 and BT048) by lentiviral transfection of small hairpin loop RNA (shRNA) to TNC also exhibited reduced JAG1 expression, as determined using cell lysates by Western blots. The clones generated using shRNA construct is designated as shRNA(1) to shRNA(4). shRNA(1) and shRNA(3) were found to express the least amount of TNC and corresponding JAG1 compared with control shRNA-generated cells. This evaluation was reproduced with another set of analysis. B, In a separate experiment, JAG1 gene was transiently knocked down with two different siRNAs for JAG1 (designated JAG1a and JAG1b) in BT025 and BT048 cell lines. BTICs treated with TNC for 6 hours after siRNA-mediated knockdown of JAG1 have reduced S-phase cell-cycle kinetics (C) and sphere forming capacity (D). ***, P < 0.001 compared with TNC + control siRNA. E, JAG1 and corresponding TNC immunoreactivity (brown) in the tumor-containing area of the brain of 5 mice at 7 weeks following implantation with BT025 cells.

Figure 6.

The knockdown of JAG1 reduces TNC-stimulated BTIC growth. A, The reduction of TNC expression in BTIC lines (BT025 and BT048) by lentiviral transfection of small hairpin loop RNA (shRNA) to TNC also exhibited reduced JAG1 expression, as determined using cell lysates by Western blots. The clones generated using shRNA construct is designated as shRNA(1) to shRNA(4). shRNA(1) and shRNA(3) were found to express the least amount of TNC and corresponding JAG1 compared with control shRNA-generated cells. This evaluation was reproduced with another set of analysis. B, In a separate experiment, JAG1 gene was transiently knocked down with two different siRNAs for JAG1 (designated JAG1a and JAG1b) in BT025 and BT048 cell lines. BTICs treated with TNC for 6 hours after siRNA-mediated knockdown of JAG1 have reduced S-phase cell-cycle kinetics (C) and sphere forming capacity (D). ***, P < 0.001 compared with TNC + control siRNA. E, JAG1 and corresponding TNC immunoreactivity (brown) in the tumor-containing area of the brain of 5 mice at 7 weeks following implantation with BT025 cells.

Close modal

Next, we knocked down JAG1 using two siRNAs in BT025 and BT048 lines (Fig. 6B). The resultant cells were no longer responsive to TNC for BTIC growth, suggesting that JAG1 is integral to TNC's effect (Fig. 6C and D). This is supported by JAG1 immunoreactivity in TNC-expressing areas in brain sections of asymptomatic mice at 7 weeks after implantation of BT25 BTICs (Fig. 6E).

Overall, these results strongly suggest that NOTCH signaling is critically involved in TNC stimulated BTIC growth. Figure 7 shows an overview of TNC–NOTCH signaling pathway, which is presumed to be operative in BTICs in glioblastoma.

Figure 7.

Postulated mechanism of TNC-stimulated BTIC growth. The diagram depicts that TNC elevates level of JAG1 expression in BTICs upon binding to α2β1 integrin on the same or proximal cell. The interaction of JAG1 and its receptor NOTCH1 then results in the proteolytic cleavage of the NOTCH receptor and release of the intracellular domain (NICD) that translocates to the nucleus to affect NOTCH-dependent transcription of genes. The generation of NICD is promoted by metzincin metalloproteinases (denoted as S1) at the proximal extracellular loop of NOTCH and the γ-secretase complex (denoted as S2) at the intracellular juxtamembrane. TNC is either from the extracellular matrix or produced by BTICs for autocrine growth.

Figure 7.

Postulated mechanism of TNC-stimulated BTIC growth. The diagram depicts that TNC elevates level of JAG1 expression in BTICs upon binding to α2β1 integrin on the same or proximal cell. The interaction of JAG1 and its receptor NOTCH1 then results in the proteolytic cleavage of the NOTCH receptor and release of the intracellular domain (NICD) that translocates to the nucleus to affect NOTCH-dependent transcription of genes. The generation of NICD is promoted by metzincin metalloproteinases (denoted as S1) at the proximal extracellular loop of NOTCH and the γ-secretase complex (denoted as S2) at the intracellular juxtamembrane. TNC is either from the extracellular matrix or produced by BTICs for autocrine growth.

Close modal

Within the CNS, glioma cells show preference to infiltrate along the periphery of blood vessel walls, the subpial glial surface (glial limitans externa) or white matter tracts (19). Besides the occurrence of glial cells or their processes, these structures are enriched in ECM proteins. For most tumor types, including gliomas, the availability of ECM proteins is considered a critical step in cell invasion and tumor growth as they activate integrins on the cell surface to trigger signaling and because the ECM is a rich source of growth factors. Among many glioma ECM proteins, TNC is arguably the most prominent component (45), and plays important roles in glioma angiogenesis, growth, and invasion. However, there is dearth of such information for the influence of TNC on BTICs in contrast to differentiated glioma cell lines.

Here, we demonstrate the potential of interactions between BTICs and TNC by displaying their proximity in human glioblastoma specimens and in mouse brains implanted with human BTICs (Fig. 1). We emphasize the capacity of exogenous TNC to promote BTIC growth in culture (Fig. 2) and we have uncovered an autocrine pathway of TNC-mediated cellular proliferation that engages the α2β1 integrin (Fig. 3); an association between TNC and the α2β1 integrin was also found by coimmunprecipitation and is corroborated by the TCGA mining data (Fig. 5 and Supplementary Fig. S8). Moreover, failure of exogenous TNC to promote BTIC growth in α2β1 integrin blocked–TNC-silenced BTIC cells further emphasize the critical requirement of functional α2β1 integrin heterodimers for TNC-promoted BTIC growth (Supplementary Fig. S5). We note that while several TNC-interacting integrins other than α2β1 were not prominently expressed on BTICs, and that their function blocking antibodies did not abrogate TNC-promoted sphere formation by BTICs, we have not exhausted the examination of integrins with the capacity to bind TNC, such as α5β1 (46).

Importantly, our results link TNC/α2β1 integrin to the NOTCH elevation that is noted in gliomas. This was supported by microarray analyses and confirmatory PCR and Western blots that TNC-elevated components of the NOTCH signaling pathway in BTICs, including JAG1, ADAMTS15, and NICD1/2 (Figs. 4 and 5); and by perturbation experiments in culture involving knockdown of JAG1 or pharmacologic inhibitors to γ-secretase and metalloproteinases that abrogate TNC-promoted BTIC growth (Figs. 5 and 6).

Although we have posited TNC as an important factor in activating NOTCH signaling in BTICs, it is likely that other factors in the tumor microenvironment can also activate NOTCH. For instance, TGFβ signaling is elevated in TNC-stimulated BTICs as indicated by the microarray data (Fig. 4), and whether this is relevant to the enhanced BTIC growth though activating NOTCH remains to be explored in future experiments. In addition, other cell types may also activate NOTCH in BTICs; one example is endothelial cells that express NOTCH ligands (47) and that may regulate brain tumor stem cell niches via NOTCH signaling (48). Another area for investigation in future studies is whether the TNC–JAG1–NOTCH signaling pathway plays a vital role in the proliferation of transformed cells differentiated from BTICs, or whether this cascade is a property of stemness within glioblastomas.

In our study, we suggest the presence of autocrine TNC signaling for BTICs as TNC and nestin immunoreactivity in BTICs in glioblastoma specimens may overlap (Fig. 1C), and because blocking the α2β1 integrin in unstimulated BTICs (Fig. 3) result in reduced basal sphere growth. Exogenously applied TNC further stimulates growth, raising the question of interactions of TNC from autocrine and paracrine sources. Indeed, there appears to be some interaction between endogenous and exogenous TNC because in the TNC siRNA-treated cells without endogenous TNC for autocrine signaling, the addition of exogenous TNC does not return sphere growth to the extent seen with control + TNC BTICs (Fig. 3I).

In reviewing the literature, we note that NOTCH signaling in gliomas has been found to upregulate TNC levels (49); however, the converse that TNC accounts for the elevated NOTCH signaling in gliomas, particularly in BTICs, has not been previously reported. Moreover, we are unaware of publications that addressed whether TNC is a growth-promoting factor for BTICs. One publication reported that TNC-expressing BTICs were better at sphere formation in limiting dilution assays than BTICs that were not expressing TNC (38), but the direct role of TNC in promoting proliferation was not addressed. Nonetheless, that manuscript supports our present analysis that BTICs produce TNC, and that the TNC in turn can regulate BTIC growth.

From the literature, it is evident that TNC is present in neural stem cell niches in the normal central nervous system (50). Whether TNC maintains the stemness of neural stem cells through α2β1 integrin–NOTCH signaling remains to be determined. Whether TNC maintains the tumor stemness phenotype through α2β1 integrin–NOTCH signaling in other cancer types should also be of interest. In breast cancer, tumor cells that metastasize to the lungs support their metastatic capacity by expressing TNC that enhances the expression of a stem cell signaling component, musashi homolog 1, a positive regulator of NOTCH signaling (34).

In summary, and to the best of our knowledge, this study is the first to identify TNC for the elevated NOTCH signaling in BTICs; moreover, we demonstrate that the TNC-activated NOTCH signaling within BTICs regulates their proliferation and sphere-forming capacity. The TNC/NOTCH/JAG1 axis has pathologic relevance as mining of the Oncomine, Rembrandt or TCGA databases shows each component to have survival disadvantage for patients with gliomas; moreover, TCGA data analysis showed significant correlation between TNC and NOTCH1, JAG1 and α2 integrin in GBM specimens. As the relative resistance of BTICs to chemo- or radiotherapy helps account for the intractability of human gliomas, our findings of TNC-triggered activation of NOTCH signaling could provide new insights and therapeutics to target the components of this axis, such as by inhibiting the proteases or translocation of NICD involved in the activation of NOTCH, in order to control the growth of BTICs and improve the prognosis of those with glioblastomas.

No potential conflicts of interest were disclosed.

Conception and design: S. Sarkar, R. Mirzaei, V.W. Yong

Development of methodology: S. Sarkar, R. Mirzaei, V.W. Yong

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Sarkar, R. Mirzaei, F.J. Zemp, D.L. Senger, S.M. Robbins, V.W. Yong

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Sarkar, R. Mirzaei, W. Wu, V.W. Yong

Writing, review, and/or revision of the manuscript: S. Sarkar, R. Mirzaei, F.J. Zemp, D.L. Senger, S.M. Robbins, V.W. Yong

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D.L. Senger, V.W. Yong

Study supervision: V.W. Yong

We acknowledge the technical help of Claudia Silva, Yan Fan, Xiuling Wang, and Fiona Yong. We thank the University of Calgary BTIC Core headed by Drs. Sam Weiss and Greg Cairncross for isolating BTIC lines from patient-resected specimens.

We acknowledge grant support from Alberta Cancer Foundation/Alberta Innovates – Health Solutions to V.W. Yong, and the Canadian Institutes of Health Research to S.M. Robbins and V.W. Yong. R. Mirzaei is supported by a University of Calgary Eyes High postdoctoral fellowship.

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.
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
.
2.
Rennert
RC
,
Achrol
AS
,
Januszyk
M
,
Kahn
SA
,
Liu
TT
,
Liu
Y
, et al
Multiple subsets of brain tumor initiating cells co-exist in glioblastoma
.
Stem Cells
2016
;
34
:
1702
7
.
3.
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
.
4.
Blazek
ER
,
Foutch
JL
,
Maki
G
. 
Daoy medulloblastoma cells that express CD133 are radioresistant relative to CD133- cells, and the CD133+ sector is enlarged by hypoxia
.
Int J Radiat Oncol Biol Phys
2007
;
67
:
1
5
.
5.
Eramo
A
,
Ricci-Vitiani
L
,
Zeuner
A
,
Pallini
R
,
Lotti
F
,
Sette
G
, et al
Chemotherapy resistance of glioblastoma stem cells
.
Cell Death Differ
2006
;
13
:
1238
41
.
6.
Piccirillo
SG
,
Spiteri
I
,
Sottoriva
A
,
Touloumis
A
,
Ber
S
,
Price
SJ
, et al
Contributions to drug resistance in glioblastoma derived from malignant cells in the sub-ependymal zone
.
Cancer Res
2015
;
75
:
194
202
.
7.
Chen
J
,
Li
Y
,
Yu
TS
,
McKay
RM
,
Burns
DK
,
Kernie
SG
, et al
A restricted cell population propagates glioblastoma growth after chemotherapy
.
Nature
2012
;
488
:
522
6
.
8.
Ranganathan
P
,
Weaver
KL
,
Capobianco
AJ
. 
Notch signalling in solid tumours: a little bit of everything but not all the time
.
Nat Rev Cancer
2011
;
11
:
338
51
.
9.
Espinoza
I
,
Miele
L
. 
Deadly crosstalk: Notch signaling at the intersection of EMT and cancer stem cells
.
Cancer Lett
2013
;
341
:
41
5
.
10.
Brou
C
,
Logeat
F
,
Gupta
N
,
Bessia
C
,
LeBail
O
,
Doedens
JR
, et al
A novel proteolytic cleavage involved in Notch signaling: the role of the disintegrin-metalloprotease TACE
.
Mol Cell
2000
;
5
:
207
16
.
11.
Fan
X
,
Khaki
L
,
Zhu
TS
,
Soules
ME
,
Talsma
CE
,
Gul
N
, et al
NOTCH pathway blockade depletes CD133-positive glioblastoma cells and inhibits growth of tumor neurospheres and xenografts
.
Stem Cells
2010
;
28
:
5
16
.
12.
Saito
N
,
Fu
J
,
Zheng
S
,
Yao
J
,
Wang
S
,
Liu
DD
, et al
A high Notch pathway activation predicts response to gamma secretase inhibitors in proneural subtype of glioma tumor-initiating cells
.
Stem Cells
2014
;
32
:
301
12
.
13.
Xu
R
,
Shimizu
F
,
Hovinga
K
,
Beal
K
,
Karimi
S
,
Droms
LA
, et al
Molecular and clinical effects of Notch inhibition in glioma patients: a phase 0/I trial
.
Clin Cancer Res
2016
;
22
:
4786
96
.
14.
Gursel
DB
,
Berry
N
,
Boockvar
JA
. 
The contribution of Notch signaling to glioblastoma via activation of cancer stem cell self-renewal: the role of the endothelial network
.
Neurosurgery
2012
;
70
:
N19
21
.
15.
Wang
H
,
Sun
T
,
Hu
J
,
Zhang
R
,
Rao
Y
,
Wang
S
, et al
miR-33a promotes glioma-initiating cell self-renewal via PKA and NOTCH pathways
.
J Clin Invest
2014
;
124
:
4489
502
.
16.
Kim
Y
,
Stolarska
MA
,
Othmer
HG
. 
The role of the microenvironment in tumor growth and invasion
.
Prog Biophys Mol Biol
2011
;
106
:
353
79
.
17.
Lathia
JD
,
Li
M
,
Hall
PE
,
Gallagher
J
,
Hale
JS
,
Wu
Q
, et al
Laminin alpha 2 enables glioblastoma stem cell growth
.
Ann Neurol
2012
;
72
:
766
78
.
18.
Grossman
M
,
Ben-Chetrit
N
,
Zhuravlev
A
,
Afik
R
,
Bassat
E
,
Solomonov
I
, et al
Tumor cell invasion can be blocked by modulators of collagen fibril alignment that control assembly of the extracellular matrix
.
Cancer Res
2016
;
76
:
4249
58
.
19.
Bellail
AC
,
Hunter
SB
,
Brat
DJ
,
Tan
C
,
Van Meir
EG
. 
Microregional extracellular matrix heterogeneity in brain modulates glioma cell invasion
.
Int J Biochem Cell Biol
2004
;
36
:
1046
69
.
20.
Niibori-Nambu
A
,
Midorikawa
U
,
Mizuguchi
S
,
Hide
T
,
Nagai
M
,
Komohara
Y
, et al
Glioma initiating cells form a differentiation niche via the induction of extracellular matrices and integrin alphaV
.
PLoS One
2013
;
8
:
e59558
.
21.
Herold-Mende
C
,
Mueller
MM
,
Bonsanto
MM
,
Schmitt
HP
,
Kunze
S
,
Steiner
HH
. 
Clinical impact and functional aspects of tenascin-C expression during glioma progression
.
Int J Cancer
2002
;
98
:
362
9
.
22.
Leins
A
,
Riva
P
,
Lindstedt
R
,
Davidoff
MS
,
Mehraein
P
,
Weis
S
. 
Expression of tenascin-C in various human brain tumors and its relevance for survival in patients with astrocytoma
.
Cancer
2003
;
98
:
2430
9
.
23.
Behrem
S
,
Zarkovic
K
,
Eskinja
N
,
Jonjic
N
. 
Distribution pattern of tenascin-C in glioblastoma: correlation with angiogenesis and tumor cell proliferation
.
Pathol Oncol Res
2005
;
11
:
229
35
.
24.
Brosicke
N
,
Faissner
A
. 
Role of tenascins in the ECM of gliomas
.
Cell Adh Migr
2015
;
9
:
131
40
.
25.
Xia
S
,
Lal
B
,
Tung
B
,
Wang
S
,
Goodwin
CR
,
Laterra
J
. 
Tumor microenvironment tenascin-C promotes glioblastoma invasion and negatively regulates tumor proliferation
.
Neuro Oncol
2016
;
18
:
507
17
.
26.
Kim
CH
,
Bak
KH
,
Kim
YS
,
Kim
JM
,
Ko
Y
,
Oh
SJ
, et al
Expression of tenascin-C in astrocytic tumors: its relevance to proliferation and angiogenesis
.
Surg Neurol
2000
;
54
:
235
40
.
27.
Ruiz
C
,
Huang
W
,
Hegi
ME
,
Lange
K
,
Hamou
MF
,
Fluri
E
, et al
Growth promoting signaling by tenascin-C
.
Cancer Res
2004
;
64
:
7377
85
.
28.
Orend
G
. 
Potential oncogenic action of tenascin-C in tumorigenesis
.
Int J Biochem Cell Biol
2005
;
37
:
1066
83
.
29.
Giblin
SP
,
Midwood
KS
. 
Tenascin-C: form versus function
.
Cell Adh Migr
2015
;
9
:
48
82
.
30.
Deryugina
EI
,
Bourdon
MA
. 
Tenascin mediates human glioma cell migration and modulates cell migration on fibronectin
.
J Cell Sci
1996
;
109
(
Pt 3
):
643
52
.
31.
Zukiel
R
,
Nowak
S
,
Wyszko
E
,
Rolle
K
,
Gawronska
I
,
Barciszewska
MZ
, et al
Suppression of human brain tumor with interference RNA specific for tenascin-C
.
Cancer Biol Ther
2006
;
5
:
1002
7
.
32.
Reardon
DA
,
Zalutsky
MR
,
Akabani
G
,
Coleman
RE
,
Friedman
AH
,
Herndon
JE
 2nd
, et al
A pilot study: 131I-antitenascin monoclonal antibody 81c6 to deliver a 44-Gy resection cavity boost
.
Neuro Oncol
2008
;
10
:
182
9
.
33.
Garikapati
KR
,
Patel
N
,
Makani
VK
,
Cilamkoti
P
,
Bhadra
U
,
Bhadra
MP
. 
Down-regulation of BORIS/CTCFL efficiently regulates cancer stemness and metastasis in MYCN amplified neuroblastoma cell line by modulating Wnt/beta-catenin signaling pathway
.
Biochem Biophys Res Commun
2017
;
484
:
93
9
.
34.
Oskarsson
T
,
Acharyya
S
,
Zhang
XH
,
Vanharanta
S
,
Tavazoie
SF
,
Morris
PG
, et al
Breast cancer cells produce tenascin C as a metastatic niche component to colonize the lungs
.
Nat Med
2011
;
17
:
867
74
.
35.
Sarkar
S
,
Nuttall
RK
,
Liu
S
,
Edwards
DR
,
Yong
VW
. 
Tenascin-C stimulates glioma cell invasion through matrix metalloproteinase-12
.
Cancer Res
2006
;
66
:
11771
80
.
36.
Sarkar
S
,
Yong
VW
. 
Reduction of protein kinase C delta attenuates tenascin-C stimulated glioma invasion in three-dimensional matrix
.
Carcinogenesis
2010
;
31
:
311
7
.
37.
Sarkar
S
,
Zemp
FJ
,
Senger
D
,
Robbins
SM
,
Yong
VW
. 
ADAM-9 is a novel mediator of tenascin-C-stimulated invasiveness of brain tumor-initiating cells
.
Neuro Oncol
2015
;
17
:
1095
105
.
38.
Nie
S
,
Gurrea
M
,
Zhu
J
,
Thakolwiboon
S
,
Heth
JA
,
Muraszko
KM
, et al
Tenascin-C: a novel candidate marker for cancer stem cells in glioblastoma identified by tissue microarrays
.
J Proteome Res
2015
;
14
:
814
22
.
39.
Sarkar
S
,
Doring
A
,
Zemp
FJ
,
Silva
C
,
Lun
X
,
Wang
X
, et al
Therapeutic activation of macrophages and microglia to suppress brain tumor-initiating cells
.
Nat Neurosci
2014
;
17
:
46
55
.
40.
Yang
R
,
Sarkar
S
,
Korchinski
DJ
,
Wu
Y
,
Yong
VW
,
Dunn
JF
. 
MRI monitoring of monocytes to detect immune stimulating treatment response in brain tumor
.
Neuro Oncol
2016
[Epub ahead of print]
.
41.
Lun
X
,
Wells
JC
,
Grinshtein
N
,
King
JC
,
Hao
X
,
Dang
NH
, et al
Disulfiram when combined with copper enhances the therapeutic effects of temozolomide for the treatment of glioblastoma
.
Clin Cancer Res
2016
;
22
:
3860
75
.
42.
Besson
A
,
Wilson
TL
,
Yong
VW
. 
The anchoring protein RACK1 links protein kinase Cepsilon to integrin beta chains. Requirements for adhesion and motility
.
J Biol Chem
2002
;
277
:
22073
84
.
43.
Orend
G
,
Chiquet-Ehrismann
R
. 
Tenascin-C induced signaling in cancer
.
Cancer Lett
2006
;
244
:
143
63
.
44.
Turchi
L
,
Debruyne
DN
,
Almairac
F
,
Virolle
V
,
Fareh
M
,
Neirijnck
Y
, et al
Tumorigenic potential of miR-18A* in glioma initiating cells requires NOTCH-1 signaling
.
Stem Cells
2013
;
31
:
1252
65
.
45.
Kong
X
,
Ma
W
,
Li
Y
,
Wang
Y
,
Guan
J
,
Gao
J
, et al
Does tenascin have clinical implications in pathological grade of glioma patients?: a systematic meta-analysis
.
Medicine (Baltimore)
2015
;
94
:
e1330
.
46.
Jachetti
E
,
Caputo
S
,
Mazzoleni
S
,
Brambillasca
CS
,
Parigi
SM
,
Grioni
M
, et al
Tenascin-C protects cancer stem-like cells from immune surveillance by arresting T-cell activation
.
Cancer Res
2015
;
75
:
2095
108
.
47.
Zhu
TS
,
Costello
MA
,
Talsma
CE
,
Flack
CG
,
Crowley
JG
,
Hamm
LL
, et al
Endothelial cells create a stem cell niche in glioblastoma by providing NOTCH ligands that nurture self-renewal of cancer stem-like cells
.
Cancer Res
2011
;
71
:
6061
72
.
48.
Nyfeler
Y
,
Kirch
RD
,
Mantei
N
,
Leone
DP
,
Radtke
F
,
Suter
U
, et al
Jagged1 signals in the postnatal subventricular zone are required for neural stem cell self-renewal
.
EMBO J
2005
;
24
:
3504
15
.
49.
Sivasankaran
B
,
Degen
M
,
Ghaffari
A
,
Hegi
ME
,
Hamou
MF
,
Ionescu
MC
, et al
Tenascin-C is a novel RBPJkappa-induced target gene for Notch signaling in gliomas
.
Cancer Res
2009
;
69
:
458
65
.
50.
Faissner
A
,
Roll
L
,
Theocharidis
U
. 
Tenascin-C in the matrisome of neural stem and progenitor cells
.
Mol Cell Neurosci
2016
[Epub ahead of print] doi: 10.1016/j.mcn.2016.11.003
.