Tenascin-C is a member of the matricellular protein family, and its expression level is correlated to poor prognosis in cancer, including glioblastoma, whereas its substantial role in tumor formation and malignant progression remains controversial. We reported previously that peptide TNIIIA2 derived from the cancer-associated alternative splicing domain of tenascin-C molecule has an ability to activate β1-integrin strongly and to maintain it for a long time. Here, we demonstrate that β1-integrin activation by TNIIIA2 causes acquisition of aggressive behavior, dysregulated proliferation, and migration, characteristic of glioblastoma cells. TNIIIA2 hyperstimulated the platelet-derived growth factor–dependent cell survival and proliferation in an anchorage-independent as well as -dependent manner in glioblastoma cells. TNIIIA2 also strongly promoted glioblastoma multiforme cell migration, which was accompanied by an epithelial–mesenchymal transition–like morphologic change on the fibronectin substrate. Notably, acquisition of these aggressive properties by TNIIIA2 in glioblastoma cells was abrogated by peptide FNIII14 that is capable of inducing inactivation in β1-integrin activation. Moreover, FNIII14 significantly inhibited tumor growth in a mouse xenograft glioblastoma model. More importantly, FNIII14 sensitized glioblastoma cells to temozolomide via downregulation of O6-methylguanine-DNA methyltransferase expression. Consequently, FNIII14 augmented the antitumor activity of temozolomide in a mouse xenograft glioblastoma model. Taken altogether, the present study provides not only an interpretation for the critical role of tenascin-C/TNIIIA2 in aggressive behavior of glioblastoma cells, but also an important strategy for glioblastoma chemotherapy. Inhibition of the tenascin-C/β1-integrin axis may be a therapeutic target for glioblastoma, and peptide FNIII14 may represent a new approach for glioblastoma chemotherapy.

Significance:

These findings provide a proposal of new strategy for glioblastoma chemotherapy based on integrin inactivation.

Glioblastoma multiforme (GBM) is the most common and aggressive glial tumor in adults. Despite multimodal therapies, including surgical resection, radiotherapy, and chemotherapy, the prognosis remains extremely poor (1). GBM is characterized by a high degree of proliferation and disseminative migration throughout the brain parenchyma, thus preventing complete surgical resection. In order to develop an effective chemotherapy for GBM, it is crucial to identify the molecular mechanism that leads gliomas to acquire aggressive properties and advance to GBM.

Tenascin-C is an adhesion modulatory protein present in extracellular matrix (ECM), so-called “matricellular proteins” (2), and is characterized by its regulated expression. Expression of tenascin-C in normal adult tissues is generally low, whereas it increases transiently in pathologic states including inflammation and malignant tumors (3). Various isoforms of tenascin-C are generated by alternative splicing within the fibronectin type III (FNIII) repeats (4). Among them, tenascin-C variants containing the FNIII A2 repeat are highly expressed in neoplasms (5, 6). Therefore, these variants have been implicated in tumor formation and/or progression. Notably, the expression of tenascin-C is especially high in gliomas/GBM (7, 8). In fact, high tenascin-C levels are correlated with poor prognosis in GBM patients (3), and its expression is therefore considered a negative prognostic factor. However, the role of tenascin-C in the acquisition of the aggressive phenotype of GBM has not been clarified. ECM remodeling often occurs in malignant tumors where ECM protein fragments with biological functions are released through cleavage by inflammatory proteinases (9, 10). To clarify the substantial role of tenascin-C in the pathogenesis of GBM, both the full-length tenascin-C molecule and its peptide fragments must be considered.

We reported previously that the cancer-associated FNIII A2 repeat of tenascin-C molecule has a cryptic functional site (11) and that a 22-mer peptide containing this functional site, TNIIIA2, is capable of inducing potent activation in integrin α5β1 that was sustained for a long period (12). Based on this effect, TNIIIA2 was shown to influence various cellular processes. Notably, TNIIIA2 not only made nontransformed fibroblasts (NIH3T3 cells) resistant to anoikis (12), but also induced hyperstimulation of platelet-derived growth factor (PDGF)–dependent proliferation in NIH3T3 cells through activation of the PDGF receptor/Ras/MAPK pathway, consequently leading to the formation of dense multilayered cell aggregates, so-called transformed foci (12), suggesting disruption of the normal cell phenotype (“contact inhibition of cell proliferation”). These observations raise the possibility that the FNIII A2 repeat–containing variants of tenascin-C molecule highly expressed in tumor tissues may be involved in oncogenic transformation and malignant progression. In particular, TNIIIA2, which is released from the cancer-associated tenascin-C variants containing the FNIII A2 repeat by inflammatory proteinases (13), may contribute to tumorigenesis and/or malignant progression. Because the PDGF signal is accelerated in some subclasses of GBM, especially proneural GBM (14, 15), stimulation of PDGF signaling based on TNIIIA2-induced activation of integrin α5β1 may be associated with the acquisition of a highly aggressive phenotype in GBM.

In the present study, we demonstrated that the activation of β1-integrins by TNIIIA2 is involved in the acquisition of malignant properties, such as excessive survival/proliferation and disseminative migration in GBM cells, which are major causes of poor prognosis in GBM. We also showed that the TNIIIA2-stimulated aggressive behavior of GBM cells can be abrogated through inactivation of β1-integrins by another peptide factor, FNIII14 (16). Furthermore, FNIII14 has the ability to sensitize GBM cells to temozolomide (TMZ) via downregulation of O6-methylguanine-DNA methyltransferase (MGMT), which is known to cause resistance to TMZ treatment. Our results thus suggest that peptide FNIII14 has potential for new approaches to treat this refractory disease.

Reagents

Human plasma fibronectin was purified as described previously (17). Peptide TNIIIA2, peptide FNIII14, and its analogous inactive scrambled control peptide FNIII14scr have been described previously (11, 18). Peptide (GRGDSP)n ((GRGDSP)8K4K2KY), which was used as an antagonist for integrin α5β1, was kindly provided by Dr. Motoyoshi Nomizu (Department of Pharmacy, Tokyo University of Pharmacy and Life Sciences). PDGF-BB, poly(2-hydroxyethyl methacrylate) (poly-HEMA) and TMZ were purchased from WAKO Pure Chemicals, Sigma-Aldrich Japan, and Tokyo Chemical Industry, respectively.

Cell culture

Human GBM cell line T98G, which was obtained from the American Type Culture Collection, was maintained in RPMI 1640 medium supplemented with 10% FBS (SAFC Biosciences). Rat GBM cell line 9L, which was kindly provided by Dr. Yoshida Fumiyo (Department of Neurosurgery, University of Tsukuba, Japan), was maintained in DMEM with 10% FBS. These cell lines were passaged soon after receipts, divided, and stocked in liquid nitrogen. Each experiment was carried out using thawed cells without further authentication. These cell lines were also authenticated by routine monitoring of cell morphology and proliferation, kept in a humidified incubator at 37°C with 5% CO2, and cultured up to 15 passages.

Cell survival and proliferation

T98G cells (3.5 × 103 cells/well) were seeded on 96-well plates coated with fibronectin (0.25 μg/mL) or poly-HEMA in serum-free medium. The number of viable cells was evaluated by the WST-8 assay, as described previously (12).

Colony formation assay

Solution of a 1:1 mixture of 1.4% agar (BD Bioscience) and 2 × RPMI 1640 growth medium was poured into 12-well plates. After solidification, T98G (1.0 × 104 cells/well) suspended in growth medium containing 0.7% agar in the presence or absence of TNIIIA2 and/or PDGF were overlaid on top of a base layer. After solidification of the top agar layer, RPMI 1640 growth medium in the presence or absence of TNIIIA2 and/or PDGF was added. Media were changed every 4 to 5 days. After 10 to 14 days, colonies were stained with crystal violet, and the number of them was counted for five randomly selected fields under the microscope at 20× magnification.

Western blotting

Cells (2.5 × 105 cells/well) were allowed to adhere in 6-well plates coated with fibronectin in the FBS-containing medium and then starved overnight in serum-free medium. Subsequent steps were conducted using antibodies shown in Supplementary Table S1, as described previously (12).

Immunoprecipitation

Immunoprecipitation studies were performed using antibodies shown in Supplementary Table S1, as described previously (12).

Wound-healing assay

Cells (2.1 × 105 cells/well) suspended in the FBS-containing medium were seeded onto 12-well plates coated with fibronectin. A single scratch wound was created using a p200 pipette tip into confluent cells. After the incubation in the assay medium for 6 hours, cells were fixed with 4% paraformaldehyde and stained with crystal violet. Images were captured by phase-contrast microscopy at 0 and 6 hours, and wound width was measured by Motic Image Plus 2.2S (Shimadzu Rika). The distances of cell migration were determined by the wound width at different time points. Cell migration in each group was expressed as a percentage of the control.

Scattering assay

Cells (1.0 × 103 cells/well) suspended with growth medium were seeded onto 6- or 12-well plates, cultured up to formation of a cobblestone-like cell sheet, and then cultured in the assay medium for 6 hours. Cell images were captured by phase-contrast microscopy and analyzed by Motic Image Plus 2.2S.

Confocal microscopy

Confocal microscopy was performed using the antibodies presented in Supplementary Table S1 by confocal microscopy Fluoview FV1000 (Olympus) or AF6000 (Leica Microsystems) as described previously (19).

Animal study

All the animal procedures were approved by the Institutional Animal Care and Use Committee of Tokyo University of Science. 9L cells (5 × 105) suspended in PBS (−) (100 μL) were s.c. injected into the left flank of 6-week-old female Balb/c nude mice (Sankyo Laboratory Service). Mice were randomized into treatment groups after tumors were established (mean volume = 50 mm3). Tumor volume (mm3) determined by measuring with calipers was calculated: Volume = (length) × (width)2 × 0.5.

Combination index

The combination index is based on the following multiple drug effect equation (20): Combination index = 1, >1, and <1 are considered to be additive, antagonistic, and synergistic, respectively.

Semiquantitative PCR

The cell RNA was extracted using the RNeasy Mini Kit (QIAGEN) according to the manufacturer's instructions. cDNA was gathered by the reverse transcription reaction using QuantiTect Reverse Transcription (QIAGEN) and amplified in TaKaRa PCR Thermal Cycler Dicer (Takara Bio). The PCR products were electrophoresed using 2% agarose gel TBE including 0.5 μg/mL ethidium bromide and developed with a trans-illuminator. Primers are shown in Supplementary Table S2.

Statistical analysis

Data are expressed as the mean ± S.D. Two-tailed Student t test or one-way ANOVA analysis was used to determine statistical differences. Values of *, P < 0.05 were considered significant.

Aggressive progression in GBM cells through activation of integrin α5β1 by peptide TNIIIA2

Expressions of tenascin-C and integrin subunits α5 and β1 were clearly detected in human (T98G, U251, and U87), rat (C6 and 9L), and mouse (GL261) GBM cell lines (Supplementary Table S1A). The tenascin-C molecule expressed in T98G cells contained the TNIIIA2-related matricryptic site (Supplementary Table S1B). Among these GBM cell lines, we used two cell lines, human T98G and rat 9L, in this study. Because fibronectin is known to be overexpressed in GBM tissue (21), we investigated the effects of TNIIIA2 and PDGF on proliferation of GBM cells adhering to the fibronectin substrate. Flow cytometry using mAb (AG89) recognizing an active conformation-specific epitope of the β1-integrin subunit showed that TNIIIA2 activated β1-integrins of GBM cells (Supplementary Table S2A). Based on this effect, TNIIIA2 promoted the adhesion of GBM cells to the fibronectin substrate (Supplementary Table S2B). We then examined the effects of β1-integrin activation by TNIIIA2 on GBM cell proliferation. The effects of TNIIIA2 on the PDGF-dependent proliferation of GBM cells were similar to those on NIH3T3 cells (12). PDGF stimulated GBM cell proliferation in a dose-dependent manner but reached a plateau at around 10 ng/mL for T98G cells (Fig. 1A) and 20 ng/mL for 9L cells (Supplementary Table S3A). The PDGF-dependent proliferation of T98G cells at submaximal PDGF concentration on fibronectin substrate was further stimulated by addition of TNIIIA2 (Fig. 1B). Similar observations were made for 9L cells (Supplementary Table S3B). To verify the integrin isoform involved in hyperstimulation of PDGF-dependent cell proliferation by TNIIIA2, we examined the effects of a function-blocking antibody for integrin subunits on cell proliferation. The results indicated that TNIIIA2-enhanced proliferation was specifically abrogated by antibodies against integrin α5 and β1 subunits, but not those against α4, αv, or β3 subunits (Fig. 1C). Similar to the effect of TNIIIA2, the β1-integrin–activating antibody, HUTS-4, also enhanced PDGF-dependent GBM cell proliferation (Supplementary Table S3C). These results suggested that GBM cell survival and proliferation, which are secured by PDGF in GBM, are strongly stimulated by activation of β1-integrin induced by TNIIIA2 derived from tenascin-C that is highly expressed in GBM.

Figure 1.

Hyperstimulation of GBM cell proliferation through β1-integrin activation by TNIIIA2. AC, Effect of TNIIIA2 on the PDGF-dependent proliferation of T98G cells. T98G cells adhered on the fibronectin substrate were stimulated with PDGF in the presence or absence of TNIIIA2 for 1 day, as described in Materials and Methods. In C, integrin isotypes associated with the TNIIIA2-dependent stimulation were examined by addition of anti-integrin function-blocking antibodies (10 μg/mL). D, Colony formation assay was performed using T98G cells. Cells suspended in the absence or presence of TNIIIA2, PDGF, or their combination were cultured as described in Materials and Methods. E, Effect of PDGF and TNIIIA2 on T98G cell survival under the nonadhered conditions. Each point represents the mean ± SD of triplicate determinations. **, P <0.01.

Figure 1.

Hyperstimulation of GBM cell proliferation through β1-integrin activation by TNIIIA2. AC, Effect of TNIIIA2 on the PDGF-dependent proliferation of T98G cells. T98G cells adhered on the fibronectin substrate were stimulated with PDGF in the presence or absence of TNIIIA2 for 1 day, as described in Materials and Methods. In C, integrin isotypes associated with the TNIIIA2-dependent stimulation were examined by addition of anti-integrin function-blocking antibodies (10 μg/mL). D, Colony formation assay was performed using T98G cells. Cells suspended in the absence or presence of TNIIIA2, PDGF, or their combination were cultured as described in Materials and Methods. E, Effect of PDGF and TNIIIA2 on T98G cell survival under the nonadhered conditions. Each point represents the mean ± SD of triplicate determinations. **, P <0.01.

Close modal

Hyperstimulation of PDGF-dependent GBM cell proliferation by TNIIIA2 was further substantiated in anchorage-independent cell proliferation, as determined by soft-agar colony formation assay to evaluate aggressive behavior of GBM cells by β1-integrin activation with TNIIIA2. T98G cells were able to form colonies in soft agar, and this was enhanced by PDGF, albeit weakly, while combined treatment with TNIIIA2 further enhanced colony formation (Fig. 1D). On the other hand, under detached conditions on poly-HEMA–coated plates, the number of viable cells was reduced after 3 days, whereas addition of either PDGF or TNIIIA2 protected T98G cells from this decrease (Fig. 1E). Addition of PDGF in combination with TNIIIA2 not only protected T98G cells from cell death, but also strongly stimulated the proliferation of T98G cells (Fig. 1E). Taken together with the data shown in Fig. 1D, it was suggested that the activation of β1-integrin by TNIIIA2 was responsible for the resistance to cell death due to loss of adhesion, which led to hyperproliferation of GBM cells in the presence of PDGF. These results suggested that TNIIIA2 derived from tenascin-C induced hyperstimulation of GBM cell survival and proliferation, both of which are characteristic of GBM, based on strong activation of β1-integrins.

Ras activation assay (Supplementary Table S3D) and Western blotting analysis (Supplementary Table S3E, left) showed that treatment of T98G cells with TNIIIA2 synergistically promoted PDGF-dependent phosphorylation/activation of PDGF receptors, which in turn enhanced activation of the classical MAP kinase signaling pathway. Similar results were obtained using 9L cells (Supplementary Table S3E, right). Immunoprecipitation experiments and immunostaining analysis showed that stimulation with PDGF in the presence of TNIIIA2 induced a physical association between integrin α5β1 and the phosphorylated/activated form of the PDGF receptor (Supplementary Table S3F), concomitant with colocalization of β1-integrin and PDGF receptor (Supplementary Table S3G). Furthermore, T98G cells treated with PDGF and TNIIIA2 markedly increased both activation of the prosurvival protein Akt and expression of the protooncoprotein c-myc (Supplementary Table S3H). These results suggested that activation of integrin α5β1 via TNIIIA2-induced PDGF receptor activation in a PDGF-dependent manner, which activated two downstream signaling pathways for cell survival and proliferation (Akt and MAPK signaling, respectively), resulted in hyperstimulation of GBM cell proliferation.

TNIIIA2 highly activates migration of GBM cells on the fibronectin substrate

GBM is characterized by dissemination throughout the brain parenchyma, as well as active proliferation. Wound-healing assay showed that TNIIIA2 stimulated the ability of cell migration of GBM cell lines (Fig. 2A, data using 9L cells also shown in Supplementary Table S4A). This TNIIIA2-enhanced ability of GBM cell migration was abrogated by the RGD peptide, an antagonist of integrin α5β1, but not by CS-1 peptide, an antagonist of integrin α4β1 (Fig. 2B; Supplementary Table S4B), suggesting that activation of integrin α5β1 is responsible for the enhanced ability of cell migration induced by TNIIIA2. In contrast, during the experiments, we observed that T98G cells cultured on the 2D fibronectin substrate adhered to each other to form cobblestone-like cell sheets (Fig. 2C). Interestingly, the cell-to-cell adhesive interactions of T98G cells were halted by the addition of TNIIIA2, resulting in a mesenchymal morphology (Fig. 2C; Supplementary Movies S1 and S2). This epithelial–mesenchymal transition (EMT)–like change was also induced by TNIIIA2 in 9L cells (Supplementary Table S4C), suggesting that the EMT-like change induced by TNIIIA2 may be one of the driving forces behind the enhanced ability of GBM cells to migrate. The EMT-like change via TNIIIA2 was overridden by the addition of BV7, a function-blocking antibody to β1-integrin (Fig. 2D), indicating the implication of β1-integrin activation by TNIIIA2. In addition, β-catenin was shown to be localized around the cell-to-cell contact area in GBM cells without stimulation, whereas TNIIIA2 treatment markedly reduced this localization, concomitant with cell scattering from the cobblestone-like sheet (Fig. 2E). No obvious changes in expression of EMT-related markers, including β-catenin, N-cadherin, and Twist, were detected upon exposure of GBM cells to TNIIIA2 (Supplementary Fig. S4D). These results thus suggested that β1-integrin activation by TNIIIIA2 may be also implicated in active migration of GBM cells, which is one of the important characteristics of aggressive behavior in GBM.

Figure 2.

TNIIIA2 induced disseminative migration of T98G cells. A and B, Effects of TNIIIA2 on T98G cell migration on the fibronectin substrate were evaluated in the presence or absence of (GRGDSP)n, a potent antagonist for integrin α5β1, by the wound-healing assay, as described in Materials and Methods. Each point represents the mean ± SD of triplicate determinations. **, P <0.01. Scale bar, 200 μm. CE, Scattering assay was performed as described in Materials and Methods. Cobblestone-like cell sheets were formed by culturing T98G cells on the fibronectin substrate. In C and E, cells were then treated in the absence or presence of TNIIIA2 (25 μg/mL) for 4 hours. Staining with crystal violet (C) and immunofluoro-staining with anti–β-catenin antibody (E) were presented. Scale bar, 100 μm (C) or 100 μm (E). In D, the cobblestone-like cell sheets were treated with or without BV7 (10 μg/mL), β1-integrin–blocking antibody, for 1 hour, and then cultured in the absence or presence of TNIIIA2 (25 μg/mL) for 4 hours. Scale bar, 100 μm.

Figure 2.

TNIIIA2 induced disseminative migration of T98G cells. A and B, Effects of TNIIIA2 on T98G cell migration on the fibronectin substrate were evaluated in the presence or absence of (GRGDSP)n, a potent antagonist for integrin α5β1, by the wound-healing assay, as described in Materials and Methods. Each point represents the mean ± SD of triplicate determinations. **, P <0.01. Scale bar, 200 μm. CE, Scattering assay was performed as described in Materials and Methods. Cobblestone-like cell sheets were formed by culturing T98G cells on the fibronectin substrate. In C and E, cells were then treated in the absence or presence of TNIIIA2 (25 μg/mL) for 4 hours. Staining with crystal violet (C) and immunofluoro-staining with anti–β-catenin antibody (E) were presented. Scale bar, 100 μm (C) or 100 μm (E). In D, the cobblestone-like cell sheets were treated with or without BV7 (10 μg/mL), β1-integrin–blocking antibody, for 1 hour, and then cultured in the absence or presence of TNIIIA2 (25 μg/mL) for 4 hours. Scale bar, 100 μm.

Close modal

Peptide FNIII14 abrogates the acquisition of TNIIIA2-enhanced GBM-aggressive properties by blocking β1-integrin activation

We reported previously that peptide FNIII14 derived from fibronectin is capable of inducing the conformational changes in β1-integrins necessary for functional inactivation. As shown in Supplementary Tables S2A and S5, peptide FNIII14 abrogated the proadhesive effects of TNIIIA2 by inducing inactivation of β1-integrins. Based on this effect, peptide FNIII14 was shown to abrogate GBM cell proliferation to the control level, overcoming the stimulation with PDGF and TNIIIA2 (Fig. 3A; Supplementary Table S6A). Soft agarose colony formation assay showed that peptide FNIII14 inhibited the anchorage-independent cell growth supported by PDGF and TNIIIA2 (Fig. 3B). Scattering of GBM cells from the cobblestone-like cell sheet and their migration stimulated by TNIIIA2 were also inhibited by peptide FNIII14 (Fig. 3C and D; Supplementary Movie S3). Moreover, peptide FNIII14 impeded TNIIIA2-enhanced T98G cell survival either in the presence (Fig. 3E, left) or in the absence (Fig. 3E, right) of PDGF under the nonadhesive conditions. Similar effects of TNIIIA2 and peptide FNIII14 on cell survival were also observed in rat GBM cell line 9L (Supplementary Table S6B). RGD peptide could not impede the TNIIIA2-enhanced T98G cell survival (Fig. 3E, right). Peptide FNIII14 inhibited the TNIIIA2-induced PDGF-dependent phosphorylation/activation of PDGF receptors, Erk, and Akt (Fig. 3F; Supplementary Table S6C). These results indicated that any of the malignant properties stimulated through β1-integrin activation by TNIIIA2, such as hyperstimulation of anchorage-dependent and anchorage-independent cell proliferation and disseminative migration, could be successfully abrogated by peptide FNIII14.

Figure 3.

β1-Integrin inactivation by peptide FNIII14 attenuated the acquisition of TNIIIA2-induced aggressive phenotype of GBM cells. A, T98G cells suspended in the serum-free medium with or without TNIIIA2 were seeded with or without peptide FNIII14 at the indicated concentrations for 2 days. Cell proliferation was assayed by WST-8 assay. B, Colony formation assay was performed in T98G cells. Cells suspended in the absence or presence of TNIIIA2, PDGF, and peptide FNIII14 were cultured for 10 days. Effect of peptide FNIII14 on the TNIIIA2-enhanced migration (C) and TNIIIA2-induced cell scatter (D). Scale bar, 200 μm. E, T98G cells were suspended in poly-HEMA–coated plates. In left plot, cells were treated with or without peptide FNIII14 in the presence or absence of PDGF and TNIIIA2. In right plot, cells were treated with or without TNIIIA2 in the presence or absence of peptide FNIII14 or peptide (GRGDSP)n. Each point represents the mean ± SD of triplicate determinations. **, P < 0.01. F, Western blotting analysis of the effect of peptide FNIII14 on TNIIIA2-enhanced PDGF receptor phosphorylation and Erk phosphorylation in T98G cells.

Figure 3.

β1-Integrin inactivation by peptide FNIII14 attenuated the acquisition of TNIIIA2-induced aggressive phenotype of GBM cells. A, T98G cells suspended in the serum-free medium with or without TNIIIA2 were seeded with or without peptide FNIII14 at the indicated concentrations for 2 days. Cell proliferation was assayed by WST-8 assay. B, Colony formation assay was performed in T98G cells. Cells suspended in the absence or presence of TNIIIA2, PDGF, and peptide FNIII14 were cultured for 10 days. Effect of peptide FNIII14 on the TNIIIA2-enhanced migration (C) and TNIIIA2-induced cell scatter (D). Scale bar, 200 μm. E, T98G cells were suspended in poly-HEMA–coated plates. In left plot, cells were treated with or without peptide FNIII14 in the presence or absence of PDGF and TNIIIA2. In right plot, cells were treated with or without TNIIIA2 in the presence or absence of peptide FNIII14 or peptide (GRGDSP)n. Each point represents the mean ± SD of triplicate determinations. **, P < 0.01. F, Western blotting analysis of the effect of peptide FNIII14 on TNIIIA2-enhanced PDGF receptor phosphorylation and Erk phosphorylation in T98G cells.

Close modal

Based on these results, we investigated the therapeutic effects of peptide FNIII14 in a mouse xenograft GBM model. Preliminary experiments indicated that T98G cells hardly developed tumors in athymic nude mice. Therefore, the therapeutic experiments were conducted using 9L rat GBM cells, which expressed both tenascin-C and integrin α5β1 (Supplementary Table S1) and responded to PDGF stimulation (Supplementary Fig. S3A). Suspensions of 9L cells were injected s.c. into athymic nude mice. After the development of tumors (mean volume = 50 mm3), the tumor-bearing mice were randomized and treated with peptide FNIII14 or control peptide for 21 consecutive days. As shown in Fig. 4A–C, tumor growth in FNIII14-treated xenografts was markedly delayed compared with control mice, suggesting that peptide FNIII14 may have therapeutic potential for GBM. There were no significant differences in body weight between the groups (Fig. 4D).

Figure 4.

Peptide FNIII14 delayed GBM growth in a mouse xenograft model. A, Tumor growth curves for 9L xenografts (n = 4, each group). Mice were treated every day by i.v. injection of peptide FNIII14 (250 μg/mouse) or FNIII14scr peptide (250 μg/mouse). B, Dot plots show the tumor sizes in individual mice in each group on day 29, with the bars representing the mean ± SD. *, P < 0.05. C, Photographs were taken on day 29. D, Body weight profiles for the animals in A.

Figure 4.

Peptide FNIII14 delayed GBM growth in a mouse xenograft model. A, Tumor growth curves for 9L xenografts (n = 4, each group). Mice were treated every day by i.v. injection of peptide FNIII14 (250 μg/mouse) or FNIII14scr peptide (250 μg/mouse). B, Dot plots show the tumor sizes in individual mice in each group on day 29, with the bars representing the mean ± SD. *, P < 0.05. C, Photographs were taken on day 29. D, Body weight profiles for the animals in A.

Close modal

Peptide FNIII14 increases the cytotoxicity of TMZ

It was noteworthy that peptide FNIII14 made GBM cells expressing MGMT susceptible to TMZ, an oral alkylating agent, which is the standard of care for newly diagnosed GBM patients. As shown in Fig. 5A, when T98G cells were treated with increasing concentrations of TMZ in combination with or without peptide FNIII14, the antitumor effect of TMZ was clearly increased in the presence of peptide FNIII14. The IC50 of TMZ in peptide FNIII14–treated T98G cells was 0.24 μmol/L, whereas that in control cells was 1.0 μmol/L. The combination index value (see Materials and Methods) of the peptide FNIII14 effect was <1 (Fig. 5B). Moreover, combined use of peptide FNIII14 increased the TMZ cytotoxicity in a dose- and time-dependent (Fig. 5C and D) manner. Similar results were obtained using 9L cells (Supplementary Table S7). Interestingly, cells treated with peptide FNIII14 showed remarkable reduction of MGMT expression at both mRNA and protein levels (Fig. 5E). MGMT expression was slightly suppressed by TMZ alone (Fig. 5F). A combination treatment of MGMT and FNIII14 additively reduced the expression level of MGMT (Fig. 5F). It was thus shown that peptide FNIII14 is able to strongly enhance the susceptibility of GBM cells to TMZ, probably by reducing MGMT expression in conjunction with TMZ. To ascertain the potentiation of TMZ cytotoxicity by FNIII14, we examined whether the FNIII14-mediated reduction in MGMT level was enabled through alteration of gene methylation status. We therefore conducted an experiment using a methylation-specific PCR. As a result, no significant change in methylation status was observed upon peptide FNIII14 treatment (Supplementary Table S8), suggesting that peptide FNIII14 suppressed MGMT expression in a MGMT promoter methylation-independent manner. Based on these results, we then examined the effects of peptide FNIII14 on cytotoxic effect of TMZ in a mouse xenograft model. Notably, coadministration of peptide FNIII14 and TMZ clearly augmented the antitumor effect of TMZ in comparison with TMZ alone (Fig. 5G).

Figure 5.

Peptide FNIII14 potentiated the cytotoxic effects of TMZ in GBM cells and in a xenograft model. A, WST-8 assay of glioma cells treated with peptide FNIII14 and TMZ. T98G cells were cultured in the presence or absence of peptide FNIII14 (12.5 μg/mL) and TMZ for 3 days. B, Combination index of TMZ and peptide FNIII14. C, T98G cells were treated with the indicated concentration of peptide FNIII14 and/or TMZ for 3 days. D, T98G cells were treated for the indicated times with peptide FNIII14 (12.5 μg/mL) and/or TMZ (250 μmol/L). Cell viability was determined by WST-8 assay. Each point represents the mean ± SD of triplicate determinations. **, P < 0.01. E and F, T98G cells were treated with or without peptide FNIII14 in the presence or absence of TMZ. After 3 days, MGMT mRNA levels were evaluated by semiquantitative PCR. After 5 days, MGMT protein levels were evaluated by Western blotting. G, Tumor growth curves for 9L xenografts (n = 4–6). Mice were treated every other day by i.v. injection of peptide FNIII14 (250 μg/mouse) or every 3 days by intraperitoneal injection of TMZ (12 mg/kg). Tumor volumes are shown as the mean ± SD. *, P < 0.05.

Figure 5.

Peptide FNIII14 potentiated the cytotoxic effects of TMZ in GBM cells and in a xenograft model. A, WST-8 assay of glioma cells treated with peptide FNIII14 and TMZ. T98G cells were cultured in the presence or absence of peptide FNIII14 (12.5 μg/mL) and TMZ for 3 days. B, Combination index of TMZ and peptide FNIII14. C, T98G cells were treated with the indicated concentration of peptide FNIII14 and/or TMZ for 3 days. D, T98G cells were treated for the indicated times with peptide FNIII14 (12.5 μg/mL) and/or TMZ (250 μmol/L). Cell viability was determined by WST-8 assay. Each point represents the mean ± SD of triplicate determinations. **, P < 0.01. E and F, T98G cells were treated with or without peptide FNIII14 in the presence or absence of TMZ. After 3 days, MGMT mRNA levels were evaluated by semiquantitative PCR. After 5 days, MGMT protein levels were evaluated by Western blotting. G, Tumor growth curves for 9L xenografts (n = 4–6). Mice were treated every other day by i.v. injection of peptide FNIII14 (250 μg/mouse) or every 3 days by intraperitoneal injection of TMZ (12 mg/kg). Tumor volumes are shown as the mean ± SD. *, P < 0.05.

Close modal

Although a high expression level of tenascin-C is known to correlate with poor prognosis in malignant tumors, including GBM (3, 22, 23), its roles in gliomagenesis and/or GBM aggression remain controversial. Recent studies on the correlation of tumor stiffness with cancer cell malignancy would provide important insight into the roles of tenascin-C in cancer development. Miroshnikova and colleagues reported that glioma aggression and patient prognosis correlate with HIF1α levels and the stiffness of tenascin-C–enriched ECM (24). Furthermore, Barnes and colleagues recently demonstrated that the glycocalyx/ECM-integrin loop promotes GBM aggression in a tissue tension-dependent manner; human recurrent GBMs, which often exhibit a mesenchymal, stem-like phenotype, have an increased bulky glycocalyx, tenascin-C–enriched stiffened ECM, and elevated integrin mechanosignaling (25). They also showed that tumor xenografts derived from GBM cells expressing an autoclustering β1-integrin mutant (V737N) exhibit augmented integrin mechanosignaling, increased tenascin-C–enriched ECM stiffness, and increased tumor burden. Of note, these findings were validated by experiments using mouse models as well as clinical samples from GBM patients. Our results suggest that TNIIIA2 derived from the tenascin-C variants strongly expressed in GBM may contribute to the acquisition of aggressive properties in GBM, such as hyperproliferation and disseminative migration, by activating β1-integrins. We previously showed that TNIIIA2-induced β1-integrin activation is especially potent and persistent because the effect of TNIIIA2 is based on the physical support of the active conformation of β1-integrin through the induction of lateral binding with syndecan-4, a transmembrane proteoglycan (11). This potent and persistent activation of β1-integrin is responsible for the induction of malignant transformation in NIH3T3 cells (12). In addition to the conformational and resultant functional activation of β1-integrin, TNIIIA2 seems to be capable of inducing β1-integrin clustering on the membrane, because the appearance of focal adhesions and reorganization of actin stress fibers became evident after stimulation with TNIIIA2 (11). Therefore, TNIIIA2 may be able to induce phenotypic alterations in cells similar to the forced expression of the integrin mutant V737N. The present study not only supports the proposal by Barnes and colleagues, but may also provide a molecular mechanism for the role of tenascin-C in GBM aggressiveness.

Many studies also indicate that ECM stiffening enhances downstream signaling of integrin, resulting in malignant progression of tumor cells. For example, ECM stiffening enhances the β1-integrin/PI3K/Akt pathway and VEGF production in breast malignancy and hepatocellular carcinoma, respectively (26, 27). Moreover, increasing matrix stiffness augments integrin aggregation to enhance growth factor–dependent ERK and ROCK activation, leading to the promotion of malignant transformation in mammary epithelial cells (28). In the present study, activation of integrin α5β1 by TNIIIA2 caused dysregulated survival/proliferation in GBM cells through hyperactivation of PDGF receptors and subsequent MAPK and Akt signaling pathway. The effects of TNIIIA2 on β1-integrin are characterized by potent and sustained activation, in contrast to moderate and transient activation through “inside-out signaling” by well-known integrin activators, such as cytokines and chemokines (12). Our recent study showed that potentiated and sustained activation of integrin α5β1 by TNIIIA2 markedly enhanced PDGF-dependent proliferation in NIH3T3 cells through augmentation of PDGF receptor-β kinase activity by inducing PDGF receptor-β dimerization (12). TNIIIA2 seems to have the ability to maximize growth factor signaling, including PDGF, leading to hyperproliferation of GBM cells. Highly expressed tenascin-C in malignant tumors may contribute to not only increased ECM stiffness, as mentioned above, but also highly enhanced activation of β1-integrin by TNIIIA2-related fragments released from tenascin-C. It is unlikely that at least the antiadhesive effect of tenascin-C, which has been considered to be a major biochemical function of this protein, is responsible for ECM stiffening and consequent enhanced integrin signaling. Further investigations are required to determine whether β1-integrin activation by TNIIIA2 could actually result in increased ECM/tissue stiffness.

GBM is characterized not only by active proliferation, but also by dissemination throughout the brain parenchyma, both of which are the major cause for poor prognosis in GBM patients. It has been reported that integrin-mediated adhesive interaction plays a critical role in tumor cell growth, migration, invasion, apoptosis resistance, and therapy resistance (29–31). A number of previous studies showed that fibronectin, a ligand for α5β1, is overexpressed in GBM (21, 32). In addition, Janouskova and colleagues showed that α5-integrin expression is increased with increasing glioma grade and is correlated with poor prognosis in high-grade glioma (33). In the present study, TNIIIA2 promoted disseminative migration, as well as PDGF-dependent hyperproliferation, both of which were induced by activation of integrin α5β1 by TNIIIA2. Blandin and colleagues recently showed that although α5-integrin induces cell-to-cell cohesion and limits cell scattering from this cohesion in a fibronectin-poor microenvironment, α5-integrin promoted cell dissemination of GBM cells in a fibronectin-rich microenvironment (34). In a fibronectin-rich microenvironment, ligation and activation of α5-integrin induced cell dissemination, similar to the present observation that activation of integrin α5β1 by TNIIIA2 stimulated migration on the fibronectin substrate. The above observations and our findings support the suggestion that active proliferation and disseminative migration in GBM would be regulated by integrin α5β1-mediated adhesion to fibronectin substrate and subsequent activation of integrin signaling. They also suggested that α5-integrin antagonists may be anti-invasive agents for GBM (34). In support of this suggestion, the present study indicated that peptide FNIII14 has potent ability to abrogate the disseminative migration of GBM cells by inhibiting cell scattering. Peptide FNIII14 may be a promising tool to prevent acquisition of the aggressive GBM phenotype.

TMZ, an oral alkylating agent, is the first-line chemotherapeutic agent for newly diagnosed GBM. However, approximately 60% of all GBM cases are positive for DNA repair protein MGMT, and resistance to TMZ via DNA repair by MGMT represents a significant barrier to successful treatment of GBM patients (35). Therefore, new strategies to overcome TMZ resistance in GBM are urgently needed. Recent studies showed that U87MG GBM cells overexpressing α5-integrin lead to TMZ resistance and that depletion of α5-integrin sensitizes GBM cells to TMZ (33). In our study, peptide FNIII14, which inactivated β1-integrin, sensitized GBM cells to TMZ via downregulation of MGMT at the mRNA and protein levels. This decrease in MGMT expression was in a MGMT promoter methylation-independent manner, e.g., Hedgehog/Gli1 signaling pathway (36). Furthermore, combination with peptide FNIII14 augmented the antitumor efficacy of TMZ in a mouse xenograft model. Further investigations regarding the molecular basis of FNIII14-induced MGMT downregulation are needed.

Taken altogether, the present study suggested that tenascin-C/TNIIIA2 may play a role in the acquisition of aggressive behavior in GBM through potentiated and sustained activation of β1-integrins. Therefore, inactivation of β1-integrin would provide a useful strategy to abrogate the acquisition of aggressive properties in GBM. The administration of peptide FNIII14 in combination with TMZ represents a promising therapeutic strategy for GBM.

No potential conflicts of interest were disclosed.

Conception and design: F. Fukai

Development of methodology: M. Fujita, T. Yamamoto, F. Fukai

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Fujita, T. Fujisawa, M. Sasada, R. Nagai, C. Kudo, S. Kamiya

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Fujita, T. Iyoda, F. Fukai

Writing, review, and/or revision of the manuscript: M. Fujita, T. Iyoda, F. Fukai

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T. Yamamoto, T. Iyoda, H. Kodama, F. Fukai

Study supervision: F. Fukai

Other (financial support (partial)): T. Iyoda

This work was supported by a Grant-in-Aid for Scientific Research (grant no. 23590090) from the Japan Science and Technology Agency.

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.
Stupp
R
,
Mason
WP
,
van den Bent
MJ
,
Weller
M
,
Fisher
B
,
Taphoorn
MJB
, et al
Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma
.
N Engl J Med
2005
;
352
:
987
96
.
2.
Bornstein
P
,
Sage
EH
. 
Matricellular proteins: extracellular modulators of cell function
.
Curr Opin Cell Biol
2002
;
14
:
608
16
.
3.
Midwood
KS
,
Hussenet
T
,
Langlois
B
,
Orend
G
. 
Advances in tenascin-C biology
.
Cell Mol Life Sci
2011
;
68
:
3175
99
.
4.
Giblin
SP
,
Midwood
KS
. 
Tenascin-C: form versus function
.
Cell Adh Migr
2015
;
9
:
48
82
.
5.
Brösicke
N
,
van Landeghem
FKH
,
Scheffler
B
,
Faissner
A
. 
Tenascin-C is expressed by human glioma in vivo and shows a strong association with tumor blood vessels
.
Cell Tissue Res
2013
;
354
:
409
30
.
6.
Dueck
M
,
Riedl
S
,
Hinz
U
,
Tandara
A
,
Möller
P
,
Herfarth
C
, et al
Detection of tenascin-C isoforms in colorectal mucosa, ulcerative colitis, carcinomas and liver metastases
.
Int J Cancer
1999
;
82
:
477
83
.
7.
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
.
8.
Brösicke
N
,
Faissner
A
. 
Role of tenascins in the ECM of gliomas
.
Cell Adh Migr
2015
;
9
:
131
40
.
9.
Stamenkovic
I
. 
Extracellular matrix remodelling: the role of matrix metalloproteinases
.
J Pathol
2003
;
200
:
448
64
.
10.
Davis
GE
,
Bayless
KJ
,
Davis
MJ
,
Meininger
GA
. 
Regulation of tissue injury responses by the exposure of matricryptic sites within extracellular matrix molecules
.
Am J Pathol
2000
;
156
:
1489
98
.
11.
Saito
Y
,
Imazeki
H
,
Miura
S
,
Yoshimura
T
,
Okutsu
H
,
Harada
Y
, et al
A peptide derived from tenascin-C induces beta1 integrin activation through syndecan-4
.
J Biol Chem
2007
;
282
:
34929
37
.
12.
Tanaka
R
,
Seki
Y
,
Saito
Y
,
Kamiya
S
,
Fujita
M
,
Okutsu
H
, et al
Tenascin-C-derived peptide TNIIIA2 highly enhances cell survival and platelet-derived growth factor (PDGF)-dependent cell proliferation through potentiated and sustained activation of integrin α5β1
.
J Biol Chem
2014
;
289
:
17699
708
.
13.
Saito
Y
,
Owaki
T
,
Fukai
F
. 
Cell regulation through membrane rafts/caveolae
. In:
Ohshima
H
,
editor
. Electrical phenomena at interfaces and biointerfaces.
Hoboken, NJ
:
John Wiley & Sons, Inc.
; 
2012
.
p.
767
81
.
14.
Brennan
C
,
Momota
H
,
Hambardzumyan
D
,
Ozawa
T
,
Tandon
A
,
Pedraza
A
, et al
Glioblastoma subclasses can be defined by activity among signal transduction pathways and associated genomic alterations
.
PLoS One
2009
;
4
:
e7752
.
15.
Ozawa
T
,
Riester
M
,
Cheng
YK
,
Huse
JT
,
Squatrito
M
,
Helmy
K
, et al
Most human non-GCIMP glioblastoma subtypes evolve from a common proneural-like precursor glioma
.
Cancer Cell
2014
;
26
:
288
300
.
16.
Kato
R
,
Ishikawa
T
,
Kamiya
S
,
Oguma
F
,
Ueki
M
,
Goto
S
, et al
A new type of antimetastatic peptide derived from fibronectin
.
Clin Cancer Res
2002
;
8
:
2455
62
.
17.
Miekka
SI
,
Ingham
KC
,
Menache
D
. 
Rapid methods for isolation of human plasma fibronectin
.
Thromb Res
1982
;
27
:
1
14
.
18.
Fukai
F
,
Hasebe
S
,
Ueki
M
,
Mutoh
M
,
Ohgi
C
,
Takahashi
H
, et al
Identification of the anti-adhesive site buried within the heparin binding domain of fibronectin
.
J Biochem
1997
;
121
:
189
92
.
19.
Iyoda
T
,
Nagamine
Y
,
Nakane
Y
,
Tokita
Y
,
Akari
S
,
Otsuka
K
, et al
Coadministration of the FNIII14 peptide synergistically augments the anti-cancer activity of chemotherapeutic drugs by activating pro-apoptotic Bim
.
PLoS One
2016
;
11
:
e0162525
.
20.
Drewinko
B
,
Green
C
,
Loo
TL
. 
Combination chemotherapy in vitro with cis-dichlorodiammineplatinum(II)
.
Cancer Treat Rep
1976
;
60
:
1619
25
.
21.
Kumar Chintalaa
S
,
Sawayaa
R
,
Levent Gokaslana
Z
,
Fullerb
G
,
Sambasiva Raoh
J
. 
Immunohistochemical localization of extracellular matrix proteins in human glioma, both in vivo and in vitro
.
Cancer Lett
1996
;
101
:
107
14
.
22.
Gocheva
V
,
Naba
A
,
Bhutkar
A
,
Guardia
T
,
Miller
KM
,
Li
CM
, et al
Quantitative proteomics identify Tenascin-C as a promoter of lung cancer progression and contributor to a signature prognostic of patient survival
.
Proc Natl Acad Sci U S A
2017
;
114
:
E5625
34
.
23.
Ishihara
A
,
Yoshida
T
,
Tamaki
H
,
Sakakura
T
. 
Tenascin expression in cancer cells and stroma of human breast cancer and its prognostic significance
.
Clin Cancer Res
1995
;
1
:
1035
41
.
24.
Miroshnikova
YA
,
Mouw
JK
,
Barnes
JM
,
Pickup
MW
,
Lakins
JN
,
Kim
Y
, et al
Tissue mechanics promote IDH1-dependent HIF1α-tenascin C feedback to regulate glioblastoma aggression
.
Nat Cell Biol
2016
;
18
:
1336
45
.
25.
Barnes
JM
,
Kaushik
S
,
Bainer
RO
,
Sa
JK
,
Woods
EC
,
Kai
F
, et al
A tension-mediated glycocalyx–integrin feedback loop promotes mesenchymal-like glioblastoma
.
Nat Cell Biol
2018
;
20
:
1203
14
.
26.
Levental
KR
,
Yu
H
,
Kass
L
,
Lakins
JN
,
Egeblad
M
,
Erler
JT
, et al
Matrix crosslinking forces tumor progression by enhancing integrin signaling
.
Cell
2009
;
139
:
891
906
.
27.
Dong
Y
,
Xie
X
,
Wang
Z
,
Hu
C
,
Zheng
Q
,
Wang
Y
, et al
Increasing matrix stiffness upregulates vascular endothelial growth factor expression in hepatocellular carcinoma cells mediated by integrin β1
.
Biochem Biophys Res Commun
2014
;
444
:
427
32
.
28.
Paszek
MJ
,
Zahir
N
,
Johnson
KR
,
Lakins
JN
,
Rozenberg
GI
,
Gefen
A
, et al
Tensional homeostasis and the malignant phenotype
.
Cancer Cell
2005
;
8
:
241
54
.
29.
Sarkar
S
,
Mirzaei
R
,
Zemp
FJ
,
Wei
W
,
Senger
DL
,
Robbins
SM
, et al
Activation of NOTCH signaling by tenascin-C promotes growth of human brain tumor-initiating cells
.
Cancer Res
2017
;
77
:
3231
43
.
30.
Blandin
AF
,
Renner
G
,
Lehmann
M
,
Lelong-Rebel
I
,
Martin
S
,
Dontenwill
M
. 
β1 integrins as therapeutic targets to disrupt hallmarks of cancer
.
Front Pharmacol
2015
;
6
:
279
.
31.
Jahangiri
A
,
Aghi
MK
,
Carbonell
WS
. 
β1 Integrin: critical path to antiangiogenic therapy resistance and beyond
.
Cancer Res
2014
;
74
:
3
7
.
32.
Sallinen
SL
,
Sallinen
PK
,
Haapasalo
HK
,
Helin
HJ
,
Helén
PT
,
Schraml
P
, et al
Identification of differentially expressed genes in human gliomas by DNA microarray and tissue chip techniques
.
Cancer Res
2000
;
60
:
6617
22
.
33.
Janouskova
H
,
Maglott
A
,
Leger
DY
,
Bossert
C
,
Noulet
F
,
Guerin
E
, et al
Integrin α5β1 plays a critical role in resistance to temozolomide by interfering with the p53 pathway in high-grade glioma
.
Cancer Res
2012
;
72
:
3463
70
.
34.
Blandin
AF
,
Noulet
F
,
Renner
G
,
Mercier
M-C
,
Choulier
L
,
Vauchelles
R
, et al
Glioma cell dispersion is driven by α5 integrin-mediated cell-matrix and cell-cell interactions
.
Cancer Lett
2016
;
376
:
328
38
.
35.
Hegi
ME
,
Liu
L
,
Herman
JG
,
Stupp
R
,
Wick
W
,
Weller
M
, et al
Correlation of O6-methylguanine methyltransferase (MGMT) promoter methylation with clinical outcomes in glioblastoma and clinical strategies to modulate MGMT activity
.
J Clin Oncol
2008
;
26
:
4189
99
.
36.
Wang
K
,
Chen
D
,
Qian
Z
,
Cui
D
,
Gao
L
,
Lou
M
. 
Hedgehog/Gli1 signaling pathway regulates MGMT expression and chemoresistance to temozolomide in human glioblastoma
.
Cancer Cell Int
2017
;
17
:
117
.