Inhibition of both focal adhesion kinase and insulin-like growth factor-I receptor kinase suppresses glioma proliferation in vitro and in vivo

Gliomas are the third most frequent cause of cancer-related deaths in adults and the second most in children (1). Glioblastomas, the most malignant form of primary gliomas, are characterized by high infiltration of surrounding normal brain without metastasis to distant sites. Despite numerous advances in chemotherapeutic, radiation, and surgical procedures, the survival rate for patients with gliomas/glioblastomas has essentially remained unchanged for the past 30 years. Even when different treatment regiments are combined, resistant tumor cells infiltrate normal brain tissue, so that tumor recurrence is inevitable. Because tumor invasion into the local environment is an important factor in the morbidity and mortality of cancer patients with brain tumors, development of novel therapeutic strategies or agents that disrupt tumor invasion will be beneficial to patients with glioma.

Neoplastic transformation in the normal human brain occurs as a result of the accumulation of a series of genetic alterations. These genetic alterations include the loss, gain, or amplification of different chromosomes, which lead to altered expression of proteins that play important roles in the regulation of cell proliferation. For example, focal adhesion kinase (FAK), a non–receptor-type tyrosine kinase, is expressed in most tumor and normal cells and is required for cell proliferation, migration, and survival (2). A handful of studies have shown the association between FAK and grade of malignancy (3), vascularity (4), and proliferation and migration of gliomas (5). FAK can be activated upon ligand binding and clustering of integrin receptors as well as activation of other cell surface receptors such as epidermal growth factor receptor (EGFR; ref. 6). Integrin family binding of transmembrane receptors with the extracellular matrix induces recruitment of FAK from the cytoplasm, which stimulates autophosphorylation of FAK at its Tyr³⁹⁷ residue. This mechanism makes SH2-binding sites available for the recruitment of Src family tyrosine kinase(s). Further phosphorylation of FAK creates binding sites for growth factor receptor binding protein 2 and Cas adaptor and docking proteins. Similar to the cytoplasmic domain of a receptor tyrosine kinase, FAK functions as both a kinase and docking protein in these complexes. Therefore, FAK participates in multiple cell functions required for cell proliferation, survival, motility, and invasion (7, 8). Experiments designed to disrupt FAK signaling, such as overexpressing a kinase-dead FAK mutant (FRNK; ref. 9) or a small interfering RNA–induced silencing of the expression of FAK (10), resulted in antiproliferative activity. These observations suggest that FAK inactivation could be a potential targeted therapy for treatment of gliomas.

The insulin-like growth factor-I receptor (IGF-IR) has been shown to play a role in growth and proliferation of many tumor types. It is a transmembrane receptor tyrosine kinase, which is primarily activated by IGF-I and IGF-II. The activated IGF-IR signaling pathway either through high levels of IGF-IR and/or its ligands IGF-I and IGF-II has been associated with various types of human cancers including gliomas and medulloblastomas (11, 12). This high level of expression has also been correlated with increased tumor cell invasiveness and metastatic potential (13–15). IGF-IR signaling cascades activate both Akt and mitogen-activated protein kinase (MAPK) and has been shown to protect cancer cells from apoptosis induced by various anticancer drugs and radiation (16, 17). In the central nervous system, IGF-IR has been implicated in fetal and postnatal growth and development of the brain (18) as well as in brain tumorigenesis. The presence of an active IGF system has been shown in astrocytomas (12) and glioblastomas (19). These findings strongly indicate that the IGF-IR signaling pathway might contribute to the highly infiltrative characteristic of gliomas. Inhibiting the IGF-IR signaling pathway by blocking its expression via antisense (19, 20), by preventing ligand binding with an antibody (21), or by blocking its signaling capacity with a dominant-negative form of the receptor (22) has been shown to reduce tumor progression and/or metastasis formation in animal models.

In the present study, we investigated the antitumor activity of a novel compound TAE226, a potent ATP-competitive inhibitor of several tyrosine protein kinases, in particular FAK and IGF-IR kinases. In a cell-based kinase assays, FAK, IGF-IR kinase, and IR kinase were inhibited with an IC₅₀ range of 100 to 300 nmol/L compared with the other kinases tested, which were >10-fold less sensitive. We show that TAE226 antagonized both FAK and IGF-IR signaling pathways, resulting in growth inhibition, cell cycle arrest, attenuation of tumor invasion, and induction of apoptosis. More importantly, TAE226 significantly prolonged the survival of animals bearing intracranial glioma xenografts, showing the potential therapeutic efficacy of using TAE226 to treat patients with glioblastomas.

To assess the levels of IGF-IR and FAK in human gliomas as well as the effect of TAE226, we used the following cell lines: U87, U87/EGFR (U87 cells stably expressed wild-type EGFR [wt-EGFR]; ref. 23), U87/vIII (U87 cells stably expressed EGFR variant III [EGFRvIII] mutant; ref. 23), U251, U251/EGFR (U251 stably expressed wt-EGFR; ref. 24), U373, LN229, LN382T, LN18, LN308, A172, D54, SNB19, U343 (25). Cell lines expressing wild-type p53 (wt-p53) used in the detailed study were D54, U343, U87, and U87 isogenic cell lines. Cell lines bearing mutant-p53 used in the detailed study were U251, LN18, and LN229. All cell lines were maintained as monolayer cultures in DMEM/F12 supplemented with 10% fetal bovine serum and penicillin/streptomycin (all from Life Technologies, Inc., Grand Island, NY). Cell cultures were maintained and incubated as described previously (26).

NVP-TAE226 was synthesized and generously provided by Novartis Pharma AG (Basel, Switzerland) through a material transfer agreement with the University of Texas M.D. Anderson Cancer Center. For use in vitro, TAE226 was dissolved in DMSO (Sigma-Aldrich Corp., St. Louis, MO) to a concentration of 10 mmol/L, stored at −20°C, and further diluted to an appropriate final concentration in DMEM. DMSO in the final solution did not exceed 0.1% (v/v).

The antiproliferative activity of TAE226 on cells growing in culture was determined using a tetrazolium-based colorimetric [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay as described previously (24).

Cell cultures were harvested with 0.05% trypsin and seeded in triplicate at 2 × 10⁴ in 24-well culture plates for 24 h before drug treatment. Culture medium was used for mock treatment. Cells were harvested at the indicated day after treatment, and viable cells were counted using the Vi-cell viability analyzer.

Invasion of glioma cells in vitro was measured by the invasion of cells through Matrigel-coated 12-well Boyden inserts as described previously (27). Quantification of invasion was done by dissolving the inserts in 2% deoxycholate solution and measuring the absorbance at 595 nm.

Glioma cells culture treated with TAE226 for 48 h were replaced with fresh serum-free medium for an additional 24 h. Conditioned medium was collected and concentrated with equal volumes of ethanol, resuspended in radioimmunoprecipitation assay buffer, and subjected to analysis as described previously (27).

Glioma cells were plated on a six-well culture plate at 2 × 10⁴ per well for 24 h and subsequently treated with TAE226 for 48 h. Thereafter, the cells were fixed with 4% paraformaldehyde in PBS for 20 min and then permeabilized with 1.0% Triton X-100 in PBS for 5 min. Tubulin was visualized by incubating first with mouse anti-α-tubulin monoclonal antibody and then with Texas red–conjugated goat antibody to mouse IgG. Filamentous-actin filaments were stained with phalloidin conjugated with Alexa Fluor 488 (Invitrogen, Carlsbad, CA). Nuclei were stained with 1 μg/mL Hoescht 33258 for 10 min at room temperature. Pictures were taken under a Zeiss Axioskop 40 microscope equipped with appropriate filters with aid of Axio Vision 4.4 software.

Glioma cells were plated at a density of 3 × 10⁵ per well in 60-mm plates (Costar, Cambridge, MA) and maintained in 10% fetal bovine serum medium overnight. The next day, the cells were washed twice and maintained in 10% fetal bovine serum medium before treatment with TAE226 and subjected to flow cytometric analysis as described previously (26).

To determine whether TAE226 acted through its purported mechanism in vitro, we evaluated the effect of this agent on inhibition of IGF-I–stimulated tyrosine phosphorylation of IGF-IR, MAPK, and AKT in human U87 and U251 glioma cells. Under basal conditions in serum-free media, both cells showed a low level of autophosphorylation, which was enhanced after exposure to IGF-I for 30 min. These cell lines were plated onto 10-cm plates at a concentration of 5 × 10⁵ per plate and then incubated for 24 h. On the next day, cells were incubated with serum-free media for 24 h with various concentrations of TAE226, ranging from 0.25 to 1 μmol/L, and controls. The cells were then activated with recombinant IGF-I (100 ng/mL) for 30 min. The membrane was probed with primary antibodies. Primary antibodies used in this study were IGF-IRα, IGF-IRβ, phosphorylated IGF-IR (p-IGF-IR; Tyr^1135/1136), p-Akt (Ser⁴⁷³), p-MAPK (Thr²⁰²/Tyr²⁰⁴), total FAK, total MAPK, cleaved caspase-3 (Asp¹⁷⁵), caspase-7 (Asp¹⁹⁸), poly(ADP-ribose) polymerase (PARP), Bax, cdc2, p-cdc2, and cyclin B1 (Cell Signaling, Boston, MA). P-FAK (pY397) was purchased from Biosource International (Camarillo, CA). Anti-β-actin antibody was purchased from Sigma-Aldrich.

Apoptosis was quantified using the Annexin V-FLUOS Apoptosis kit (Roche, Indianapolis, IN) according to the manufacturer's instructions. U87, D54, U343, U251, and LN18 glioma cells were continuously incubated in drug for 72 h before subjected to the flow cytometric analysis.

Male nude mice were purchased from the National Cancer Institute-Frederick Cancer Research and Development Center (Frederick, MD). Mice were maintained in a pathogen-free environment and used in accordance with the Animal Care and Use Guidelines of the University of Texas M.D. Anderson Cancer Center. Mice used for this study were 6 to 8 weeks old. In DMEM/F12 serum-free media (5 μL), 5 × 10⁵ of U87 cells and 1 × 10⁶ of LN229 cells per mouse were implanted intracranially through a guide-screw system as described previously (28). Four days after injection of the tumor cells, mice were randomized into three groups for each cell line (n = 6). Mice in group 1 were treated with 50 mg/kg TAE226 in 200 μL of 0.5% methylcellulose, via an oral gavage. The mice in group 2 received 75 mg/kg TAE226 in 200 μL of 0.5% methylcellulose. The mice in group 3 the same vehicle used for administration of TAE226 (control). Treatment frequency was once a day for 5 days and off for 2 days, for a duration of 4 weeks. Mice were monitored daily. Mice were euthanized when they were moribund, and the whole brain was extracted for rapid freezing in liquid nitrogen and storage at −70°C.

For the in vitro experiments, statistical analyses were done using a two-tailed Student's t test. Data are mean ± SD. The in vivo therapeutic efficacy of TAE226 was assessed by plotting survival curves according to the Kaplan-Meier method, and groups were compared using the log-rank test.

The potency and selectivity of NVP-TAE226 was assessed in in vitro kinase assays with several recombinant kinase domains and peptide substrates (Table 1). NVP-TAE226 inhibited the recombinant FAK kinase domain with an IC₅₀ of 0.0055 μmol/L and was equipotent against the recombinant proline-rich tyrosine kinase 2 domain. NVP-TAE226 was ∼10- to 100-fold less potent against IR, IGF-IR, ALK, and c-Met kinases and >100-fold less potent against the remaining kinases tested.

Endogenous levels of FAK and IGF-IR were assessed in a battery of glioma cell lines. It has been shown that the FAK signaling pathway is activated by signaling of growth factor receptors such as EGFR (6). Isogenic cell lines of U87 and U251 expressing either wt-EGFR or EGFRvIII were therefore included in the analysis for detection of FAK levels. Results from Western blotting analyses showed that glioma cell lines express different levels of p-FAK (Tyr³⁹⁷), an index of FAK activity (Fig. 1A). Protein extract from 293 cells transiently transfected with a FAK expression vector was used as a positive control for size determination of FAK. Of interest, among U87 isogenic cell lines, it seemed that EGFRvIII expression resulted in the highest p-FAK levels followed by wt-EGFR and the parental cells. It is not surprising because EGFRvIII-expressing cells possessed constitutive tyrosine kinase activity in the absence of EGF ligand, whereas the tyrosine kinase activity in the wt-EGFR–expressing cells required ligand stimulation (23). The expression of both IGF-IRα (130 kDa) and IGF-IRβ subunits (90 kDa) also varied among the different cell lines (Fig. 1B).

The signaling events that are evoked immediately upon integrin ligation and attachment of cells to extracellular matrix have been described (29). Engaging integrins in this manner activates FAK by augmenting phosphorylation at its Tyr³⁹⁷ residue. Accordingly, we wanted to determine if TAE226 treatment abrogates integrin-induced FAK autophosphorylation. As shown in Fig. 2A, the level of p-FAK (Tyr³⁹⁷) was induced by 1.5-fold in U87 and 1.8-fold in U251 cell line on collagen-coated plates compared with uncoated control (Fig. 2B). A concentration of 0.25 μmol/L TAE226, which is smaller than 1 μmol/L [the IC₅₀ concentration established by a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay; data not shown], effectively inhibited collagen-stimulated p-FAK (Tyr³⁹⁷) expression. Interestingly, an obvious dose-dependent inhibition of TAE226 on FAK activity was noted in U87 but not in U251 cell line (Fig. 2B) possibly due to the level of integrin in U251 that might require higher concentration of TAE226 to antagonize the FAK activity.

To determine whether TAE226 could inhibit IGF-I-stimulated signaling pathways, U87 and U251 glioma cell lines were serum depleted and incubated in the presence or absence of various concentrations of TAE for 24 h. Cells were then stimulated with or without 100 ng/mL IGF-I for 30 min. Cellular extracts were then assessed by Western blotting analysis to determine the expression of p-IGF-IR and total IGF-IRβ, Akt, and MAPK protein. Upon binding of its ligands, phosphorylation of the triple tyrosine cluster (Tyr¹¹³¹, Tyr¹¹³⁵, and Tyr¹¹³⁶) within the kinase domain is the earliest major site of autophosphorylation (30). P-IGF-IRβ (Tyr^1135/1136) was significantly induced by IGF-I more so in U251 cells than in U87 cells (Fig. 2C) likely due to the difference in IGF-IR expression between the two cell lines (Fig. 1B). Such an induction was reduced to and lower than the baseline of the untreated and unstimulated control levels by 0.25 μmol/L TAE226 in U251 and U87, respectively. Activity of two notable downstream target genes in the IGF-IR signaling pathway (Akt and MAPK) were also inhibited by TAE226 as assessed by the decrease in expression of p-Akt (Ser⁴⁷³) and p-MAPK (Thr²⁰²/Tyr²⁰⁴) levels. Total levels of IGF-IRβ, MAPK, and Akt were also determined (Fig. 2C). Quantitative analysis of the effect of TAE226 on both Akt and MAPK activities shown in Fig. 2D showed that TAE226 was able to effectively antagonize the kinase activity of FAK and IGF-IR, subsequently inactivating the downstream target genes Akt and MAPK. Cells treated with media containing 1% of vehicle (DMSO) did not affect signaling or viability (data not shown); therefore, vehicle control was substituted with media control throughout the investigation.

To explore the antitumor activity of TAE226 on glioma cell lines and determine if FAK and IGF-IR expression correlates with sensitivity to TAE226 treatment, we chose four cell lines expressing different level of FAK and IGF-IR, as shown in Fig. 1A. U87, U87/EGFR, and U87/EGFRvIII cell lines were parental cells, cells stably expressing EGFR, and cells stably expressing the EGFRvIII mutant (23), respectively. The U251 cell line expressed the lowest level of FAK but highest level of IGF-IR. Cell lines were treated with different concentrations of TAE226 (1 and 10 μmol/L), and viable cells were counted at days 1, 3, and 5 after treatment using a trypan blue exclusion method. The effect of TAE226 on tumor cell growth is shown in Fig. 3A. TAE226 treatment significantly retarded the growth of all cell lines tested in a concentration-dependent manner, albeit having no direct correlation between FAK and IGF-IR expression and sensitivity to TAE226 treatment. However, comparisons among the different U87 cell lines suggested that U87/EGFR and U87/EGFRvIII cell lines, which express a higher level of p-FAK (Tyr³⁹⁷) than parental U87 (Fig. 1A), were more sensitive to TAE226. Day 5 after TAE226 treatment, there were 22%, 14%, and 17% of viable U87, U87/EGFR, and U87/EGFRvIII cells over their untreated control, respectively. Interestingly, in addition to their flat, enlarged morphology, resembling cellular senescence (31, 32), TAE226-treated cells became multinucleated, which suggested that TAE226 may disrupt cell cycle progression (Fig. 3B). Moreover, in comparison with U87 cells, prolonged incubation of U251 cells with TAE226 generated floating cells with the appearance of membrane blebs, echoing the apoptotic process (ref. 33; data not shown). These observations raised an intriguing possibility that TAE226 treatment might induce apoptosis correlating with p53 status in tumor cells because the U87 cell line contains the wt-p53 gene, whereas U251 cell line possesses the mutant p53 (mut-p53) gene.

We examined the effect of TAE226 on glioma cell cycle distribution, in particular the sub-G₁ population indicative of apoptosis using flow cytometric analysis. Three glioma cell lines containing mut-p53 (U251, LN18, and LN229) and one with wt-p53 (U87) were used for the investigation. Glioma cells were treated either with or without 1 μmol/L TAE for 48 h before harvesting for cell cycle analysis. When the U87 glioma cell line was exposed to TAE226, the number of cells in G₂-M increased from 7.0% to 35.3%, whereas cells in G₀-G₁ were markedly decreased from 77.4% to 32.8% compared with the controls. The percentage of cells in sub-G₁ was minimally affected by TAE226 treatment (1.2-fold increase over control). In sharp contrast, sub-G₁ population of U251, LN18, and LN229 cell lines following TAE226 treatment showed a 4.8-fold, 12.8-fold, and 13.0-fold increase, respectively, as shown in Fig. 3B. The population of cells in the G₂-M phase was similar between the treated and untreated in these three mut-p53 cell lines.

Interestingly, the M5 cell population with an 8N DNA content increased significantly in all treated cell lines, suggesting that TAE226 induced the formation of tetraploid cells (Fig. 3B). These observations were further confirmed by immunofluorescence staining of cell nuclei with Hoescht 33258. TAE226-treated cells became multinucleated compared with untreated cells with single nuclear staining (Fig. 3C). The size of the nuclei and the number of giant cells increased as much as 3-fold in the presence of the drug compared with controls. The microtubule network was examined after staining with α-tubulin, showing that the intracellular microtubule network extends throughout the cytoplasm in both TAE226-treated and untreated cells. When TAE226-treated cells were stained, multiple nuclei were found in single tubulin-surrounded cells. The cytoskeletal network was visualized by staining stress actin with Alexa Fluor 448–conjugated phalloidin.

Because these results suggest that TAE226 disrupts G₂-M progression, we examined the expression of proteins that control the G₂-M cell cycle checkpoint. For cells to enter mitosis, the mitosis initiator MPF (cdc2/cyclin B1) has to be properly regulated, accompanied by the dephosphorylation of p34cdc2 kinase on Tyr¹⁵ (34). It has been shown that depletion of cyclin B1 with small interfering RNA resulted in G₂-M cell cycle arrest and, subsequently, induction of apoptosis (35). In our experiments, TAE226 treatment of glioma cell lines reduced the expression of p-cdc2 (Tyr¹⁵) and cyclin B1 in a dosage-dependent manner (Fig. 3D). These results indicate that TAE226-induced G₂-M phase cell cycle arrest in glioma cells is associated with a marked decline in cyclin B1 and p-cdc2 protein expression.

Because FAK is instrumental in cell motility, and because IGF-IR plays an important role in tumor invasion, we wanted to determine if TAE226 affects glioma cell invasion. Glioma cell lines U87, U87/EGFRvIII, U251, and LN18 were treated with or without 1 μmol/L TAE226 for 24 h before subjected to an in vitro Matrigel invasion assay. The results shown in Fig. 4A (top) showed that treatment with TAE226 significantly decreased the invasive propensity of glioma cells. Quantification (Fig. 4B) of these results showed that TAE226 treatment inhibits glioma cell invasion by at least 50% of untreated control.

Matrix metalloproteases (MMP), in particular MMP-2 and MMP-9, degrading extracellular matrix components have been implicated in glioma invasion. We examined if inhibiting tumor invasion in vitro was associated with reduced MMP-2 activity. Serum-free conditioned media collected from 1 μmol/L TAE226–treated or untreated cells were subjected to a gelatin zymographic analysis. Results in Fig. 4C showed that MMP-2 enzymatic activity in TAE226-treated cells was similar to untreated cells, suggesting that TAE226 inhibition of glioma cell invasion is likely due to affect motility.

Cell cycle analysis showed that TAE226 was responsible for an elevated sub-G₁ population in U251, LN18, and LN229 but not in U87, suggesting that TAE226 induced apoptosis in a subset of glioma cell lines containing the mut-p53 gene. We confirmed apoptosis by examining the expression of Annexin V, PARP cleavage, and caspase-7 activation in TAE226-treated glioma cells. We also included two well-characterized wt-p53–containing cell lines (D54 and U343) in the study (25). Results from a quantitative analysis of apoptosis using an Annexin V-FLUOS Apoptosis kit (Roche) are summarized in Fig. 5A. Incubation with 1 μmol/L TAE226 resulted in increased apoptosis in U251 (29%), LN18 (47%), and LN229 (38%) cell lines. However, TAE226 induced only minimal Annexin V–positive staining in the wt-p53–containing cell lines U87, D54, and U343 (<5%). This finding confirmed the cell cycle results, which showed that the U251, LN18, and LN229 cell lines, but not U87 cell line, had a higher sub-G₁ cell population (Fig. 3B). These results strongly correlated p53 status in tumor cells with their sensitivity to TAE226-induced apoptosis. Cleavage of PARP by caspase-3/7 is a hallmark of apoptosis. We examined the expression of PARP cleavage during TAE226-induced apoptosis using Western blot analysis. After 48 h of exposure to TAE226, an 89-kDa band representing cleaved PARP was readily detectable both in U251 and LN18 cell lines, but not in U87 (Fig. 5B). We examined caspase-3/7 activation in tumor cells after TAE226 treatment. Western blotting analysis was carried out to determine if there was a cleaved caspase-3 and caspase-7 product using an antibody specific for detection of both cleaved caspase-3 (17 and 19 kDa) and caspase-7 (20 kDa), which would serve as an index of its activation. Our results clearly showed that after treatment with TAE226, only mut-p53–containing cell lines, such as U251, LN18, and LN229, showed cleaved caspase-3 and caspase-7 bands but not in TAE-treated wt-p53–containing tumor cell lines (Fig. 5C). The proapoptotic gene bax has been shown to be a direct target of p53 (36). As shown in Fig. 5D, TAE226 treatment did not result in induction of bax expression in apoptotic sensitive cell lines such as U251 and LN18. These results established the correlation of p53 status to the sensitivity of TAE226-induced apoptosis and showed that TAE226-induced apoptosis is mediated by caspases, which might be activated by proapoptotic genes other than bax in glioma cell lines.

To assess the potential therapeutic efficacy of TAE226, we did a study to examine the effect of TAE226 treatment on animal survival in an intracranial glioma xenograft model (28). Two glioma cell lines, U87 (wt-p53) and LN229 (mut-p53), were chosen for the study. On day 4 after tumor cell implantation, animals were treated with either vehicle or TAE226 (50 and 75 mg/kg) as described in Materials and Methods. The median survival was 28 and 29 days in vehicle control group for LN229 and U87, respectively (Fig. 6A). Treatment with TAE226 at 50 or 75 mg/kg extended the median survival of U87 xenograft animals by 6 and 7 days, respectively (P = 0.084 and P = 0.042, respectively, compared with vehicle-treated animals). However, TAE226 treatment of LN229-engrafted animals significantly prolonged their median survival by 19 days (Fig. 6B; P < 0.004 for both dosages, compared with vehicle-treated animals). These results not only showed the potential therapeutic efficacy of TAE226 for the treatment of human gliomas but also correlated well with the findings from the in vitro studies showing that LN229 cell line was more sensitive to the treatment of TAE226 than U87 cell line.

Recent advances in the development of targeted molecular therapies have the advantage of blocking the cell signaling events that govern tumor cell growth, migration, and survival. TAE226 is a potent ATP-competitive inhibitor of several tyrosine protein kinases, including FAK and IGF-IR kinase. We examined the antitumor activity of TAE226 on glioma cells in vitro. We found that TAE226 treatment antagonized extracellular matrix–induced autophosphorylation of FAK and IGF-I–induced phosphorylation of IGF-IR, clearly indicating that TAE226 can effectively inhibit its intended targets. It profoundly affected proliferation and invasion of glioma cells. Its ability to inhibit cell cycle progression, particularly at the G₂-M checkpoint, and its effect on inducing apoptosis showed that TAE226 could potentially be introduced into the clinical setting for the treatment of patients with gliomas.

The expression of FAK and IGF-1R is high, albeit varied among the glioma cell lines tested, suggesting that these kinases are suitable candidates for targeted intervention. The signaling of these two kinases and activity of their downstream target genes was inhibited by TAE226 at concentrations lower than the IC₅₀. It is conceivable that higher concentrations of TAE226 are required to elicit its growth-suppressive effect. Consistent with such a notion, in an effort to assess the effect of TAE226 on cell growth, we found that 1 μmol/L TAE226 can exert >80% reduction of cell viability on day 5 of treatment. Concomitant with the loss of viable cells was the detection of multinucleated and membrane-blebbing cells, suggesting that TAE226 perturbed cell cycle progression and induced programmed cell death. Indeed, cell cycle analysis of TAE226 treatment revealed a high percentage of G₂-M and sub-G₁ populations. The pronounced effect of TAE226 on G₂-M cell cycle arrest was accompanied by a concomitant reduction of the expression of p-cdc2 (Tyr¹⁵) and cyclin B1. It has been shown that progression of cells from the G₂ phase of the cell cycle to mitosis (M) is a tightly regulated cellular process that requires cdc2 kinase activation, which precedes the onset of mitosis (37, 38). The activity of cdc2 is regulated in part by the phosphorylation status of the Tyr¹⁵ residue, which is rapidly dephosphorylated by cdc25 tyrosine phosphatase, triggering entry into mitosis. Decreased expression of p-p34cdc2 (Tyr¹⁵), but not total p34cdc2 level by TAE226, may in part force cells to stall at the mitotic phase. The enlarged and multinucleated cells seen in triple immunofluorescence staining assays (Fig. 3C) suggested that several rounds of cell cycle activity occur without mitosis in TAE226-treated cells. Such an observation has been documented previously in that a staurosporine analogue K-252a inhibited p34cdc2 activity, induced G₂-M cell cycle arrest, and produced giant cells containing polyploid nuclei with DNA contents greater than 8N (39). These results also suggested that TAE226 might affect cytokinesis. In contrast to our results, Zhao et al. showed that inhibition of FAK signaling by overexpression of FRNK, a dominant-negative form of FAK, in NIH3T3 cells induced G₁ cell cycle arrest (40). This disparate finding from ours could reflect different cell types and the diverse mechanisms by which TAE226 inactivates FAK activity as well as IGF-IR signaling, whereas FRNK competes with the endogenous FAK in focal contacts for binding signaling molecules such as Src (40).

We also examined how inactivation of FAK and IGF-IR signaling affects glioma invasion in an in vitro Matrigel invasion assay. The propensity of glioma cells to invade was reduced to at least 50% of the control, supporting TAE226 anti-invasion activity. The possibility of TAE226 affecting MMP-2 enzymatic activity, thereby inhibiting invasion, was assessed in a gelatin zymographic assay. The result showed that despite inhibiting glioma cell invasion, TAE226 treatment did not significantly affect the MMP-2 enzymatic activity. Invasive tumors possess not only high levels of MMPs but also a high rate of cell motility. Our results suggest that TAE226 inhibition of glioma invasion is likely due to its effect on cell motility. However, this assay yielded opposite results to reported findings that overexpression of kinase-dead FAK mutants inhibited tumor metastasis of v-src–transformed cells by suppressing MMP-2 promoter activity, resulting in decreased MMP-2 expression (9). Different experimental system or the abovementioned possibility might contribute to such contradictory findings. It is also possible that TAE226 antagonizes multiple pathways, resulting in different biological outcomes.

Another intriguing finding in this study was that in addition to inducing cell cycle arrest, TAE226 induced apoptosis in a subset of glioma cell lines. This might have been accomplished by the perturbation of IGF-IR signaling resulting from TAE226 treatment. Studies have shown that antagonizing the IGF-IR signaling pathway, either with a kinase-defective IGF-IR (41) or a soluble form of IGF-IR (42), led to apoptosis of tumor cells. Interestingly, the sensitivity of glioma cells to TAE22-induced apoptosis was linked to their p53 status. Those with wt-p53 (e.g., U87, D54, and U343 cell lines) were less sensitive to TAE226-induced apoptosis than cell lines bearing mut-p53 (e.g., U251, LN18, and LN229). Coincidentally, two other studies have also reported a finding of apoptosis induction by different stimuli in a p53-independent manner. First, rapamycin, a mammalian target of rapamycin inhibitor, induced apoptosis in tumor cells lacking functional p53 but not in those with wt-p53 (43). Second, U87 glioma cells were resistant to Adp53-induced apoptosis, in contrast to U251 cells (44). These results strongly suggest that functional p53 protects tumor cells from apoptosis induced by various stimuli. Apoptosis induced by TAE226 clearly progressed through a caspase-dependent pathway, consistent with the detection of cleaved PARP, a substrate of caspase-3 and caspase-7 (45). Several unanswered questions are being scrutinized in continuing experiments regarding the mode of TAE-induced apoptosis. For example, does TAE226 induce apoptosis by activating the mitochondrial (intrinsic) pathway, the death receptor–mediated (extrinsic) pathway, or both pathways? What is the mechanism by which p53 protects tumor cells from TAE226-induced apoptosis? Although we did not detect any change in the level of Bax after TAE226 treatment, we can not exclude the involvement of other proapoptotic and/or antiapoptotic proteins that ultimately alters mitochondria potential leading to caspase activation.

Consistent with the in vitro findings of TAE22 exerting a cytotoxic effect on LN229 but cytostatic effect on U87, TAE226 treatment of nude mice bearing LN229 tumor xenografts significantly prolonged animal survival compared with the vehicle-treated animals. On the other hand, TAE226 treatment of U87 xenografts yielded a shorter but statistically significant median survival rate. Although LN229 cell line expressed less FAK and p-FAK than U87, it expressed higher level of IGF-IR than U87. Thus, it is possible that IGF-IR but not FAK signaling weighs more responsibility to the effectiveness of TAE226 treatment both in vitro and in vivo. Furthermore, our results did not rule out the possibility that inhibition of other pathways may also contribute to the overall cellular sensitivity to TAE226. Taken together, our results show that TAE226 potently suppresses glioma cell growth by disruption of cell cycle progression and/or induction of apoptosis. Simultaneous blockade of the FAK and IGF-IR signaling pathways can potentially be efficacious for treating patients with gliomas upon further development.

We thank Bill Sphon for assistance in fluorescence microscopy and Jennifer Edge and Verlene Henry for their excellent effort in performing animal study.