The resistance of glioma cells to a number of antitumor agents and the highly invasive nature of glioma cells that escape the primary tumor mass are key impediments to the eradication of tumors in glioma patients. In this study, we evaluated the therapeutic efficacy of a novel PI3-kinase/mTOR inhibitor, PI-103, in established glioma lines and primary CD133+ glioma-initiating cells and explored the potential of combining PI-103 with stem cell–delivered secretable tumor necrosis factor apoptosis-inducing ligand (S-TRAIL) both in vitro and in orthotopic mouse models of gliomas. We show that PI-103 inhibits proliferation and invasion, causes G0–G1 arrest in cell cycle, and results in significant attenuation of orthotopic tumor growth in vivo. Establishing cocultures of neural stem cells (NSC) and glioma cells, we show that PI-103 augments the response of glioma cells to stem cell–delivered S-TRAIL. Using bimodal optical imaging, we show that when different regimens of systemic PI-103 delivery are combined with NSC-derived S-TRAIL, a significant reduction in tumor volumes is observed compared with PI-103 treatment alone. To our knowledge, this is the first study that reveals the antitumor effect of PI-103 in intracranial gliomas. Our findings offer a preclinical rationale for application of mechanism-based systemically delivered antiproliferative agents and novel stem cell–based proapoptotic therapies to improve treatment of malignant gliomas. Cancer Res; 71(1); 154–63. ©2010 AACR.
Glioblastoma multiforme (GBM) is the most common and aggressive form of malignant brain tumors. Despite the advances in the understanding of glioma biology, GBMs remain very difficult to eradicate and the median survival after diagnosis is still less than 12months (1). The inherent or acquired resistance of tumor cells to a number of antitumor agents and the highly invasive nature of tumor cells that escape the primary tumor mass and subsequently cause recurrence (2, 3) are among the major obstacles in finding an optimal cure for GBMs. Furthermore, the lack of efficient delivery of promising anti-GBM agents contribute to therapeutic failure (4).
A number of genetic alterations are implicated in GBM progression, contributing to high proliferation, resistance to apoptosis, and extreme invasion capabilities of GBM tumors (5). One pathway that is commonly deregulated and a promising target for therapies is the PI3 kinase (PI3K)-Akt pathway (6). Although PI3K exerts most of its effects through Akt, its downstream signaling also activates mammalian target of rapamycin (mTOR). Because of the high frequency of PI3K pathway alterations in several cancers, including GBMs, there has been a major interest in discovering novel modulators of this pathway's components that are eventually compatible with preclinical models and clinical studies. Recently, a novel kinase inhibitor, PI-103, which inhibits both PI3K and mTOR signaling, has been utilized in several preclinical tumor models including subcutaneous models of glioma (7–14). PI-103 has been shown to inhibit cell proliferation and tumor growth through its direct effects on the inhibition of PI3K and mTOR (13, 15, 16). Although PI-103 has shown promising results in mouse models of different tumor types, its efficacy in mouse models of intracranial GBMs has not been characterized.
Current treatment regimens for GBMs involve targeting one or the other hallmark of GBM tumors through surgery, radiation or chemotherapy with systemically delivered drugs (5). However, combination therapies that aim to simultaneously tackle different components of glioma progression, such as proliferation, invasion, and apoptosis resistance, would ultimately reveal more efficient eradication of these tumors. There has been a major interest in the development of tumor-specific cytotoxic therapies for cancers including GBM. As such, TRAIL (tumor necrosis factor–related apoptosis-inducing ligand), is a promising protein due to its tumor-specific induction of apoptosis in a death receptor–dependent manner. Although systemically delivered TRAIL has short half-life and requires repeated administration (17), an on-site TRAIL delivery approach would result in a significantly better outcome. To this end, we previously established that a secreted form of TRAIL (S-TRAIL) reduces glioma growth when delivered by various stem cell types, such as neural stem cells (NSC) or mesenchymal stem cells (MSC), in orthotopic glioma models (18–21). We also showed that stem cell–mediated delivery of TRAIL is a very effective method of delivering the tumor-specific agent on-site of the gliomas through utilizing the tumor-tracking properties of stem cells. However, certain genetic alterations in gliomas might counteract the response to TRAIL (22–24). Indeed, over activation of PI3K/Akt pathway has been reported to antagonize TRAIL-mediated apoptosis (25) by modulating the TRAIL downstream effectors. Therefore, targeting cell proliferation through the inhibition of PI3K signaling might not only slow down the tumor growth and cause a cytostatic response but also enhance the tumor cytotoxicity in response to TRAIL.
In this study, we first investigated the potency of PI-103 in intracranial models of established and invasive gliomas. Furthermore, we addressed the effect of PI-103 in combination with primary NSC-delivered S-TRAIL in NSC–tumor cell cocultures in vitro and in mouse models of intracranial tumors in vivo.
Materials and Methods
Established human glioma lines (Gli36, U87MG, U251, Gli79, LN229, A172), Gli36 expressing a constitutively active variant of EGFR (EGFRvIII), herein referred to as Gli36-EvIII, and Gli36-EvIII engineered to express Fluc-mCherry (Gli36-EvIII-FmC) were grown in DMEM supplemented with 10% FBS and penicillin/streptomycin. CD133+ GBM8 glioma cells were cultured in neurobasal medium (Invitrogen/GIBCO) supplemented with 3mmol/L of l-Glutamine (Mediatech), B27 (Invitrogen/GIBCO), 2 μg/mL of heparin (Sigma), 20 ng/mL of human EGF (R&D Systems), and 20 ng/mL of human FGF-2 (fibroblast growth factor; PeproTech) as described (26). GBM6 and GBM12 primary GBM cells were kind gifts of Dr. Paul Dent (Virginia Commonwealth University) and cultured in DMEM supplemented with 5% FBS. mNSCs, which are kindly provided by Dr. Angelo Vescovi (Raffaele Scientific Institute), were grown in NeuroCult mouse neural stem cell basal medium (Stem Cell Technologies) supplemented with NeuroCult proliferation supplements, 2 μg/mL of heparin (Sigma), 20 ng/mL of human EGF (R&D Systems), and 20 ng/mL of human FGF-2 (PeproTech) and penicillin/streptomycin.
Viral vectors, lentiviral packaging, and transduction of cells
Three lentiviral vectors were used: (a) Pico2-Fluc.mCherry, a kind gift from Dr. Andrew Kung (Dana Farber Cancer Institute); (b) LV-S-TRAIL that bears S-TRAIL driven by the CMV (cytomegalovirus) promoter and also contains an IRES-GFP cassette (27); and (c) control GFP, that is driven by a CMV promoter (27). Lentiviral packaging was performed by transfection of 293T cells as previously described (28). Glioma cells (Gli36-EvIII) were transduced with LV-Pico2-Fluc.mCherry at a multiplicity of infection (MOI) of 2 in medium containing protamine sulfate (4 μg/mL) and selected with puromycin. mNSCs were dissociated and cultured as monolayers on laminin-coated (5 μg/mL) 6-well plates for 24 hours and transduced with LV-GFP or LV-S-TRAIL at an MOI of 5. All cells were visualized by fluorescence microscopy for mCherry or GFP expression 36 hours posttransduction.
Viability and caspase assays
To determine the effects of PI-103 (Cayman Biochemicals) and/or S-TRAIL on glioma viability, glioma cells were seeded on 96-well plates (0.5 × 104 per well) and treated with different doses of PI-103 (0–5 μmol/L) and/or S-TRAIL (0–1,000 ng/mL) 24 hours after plating. The effect of PI-103 on viability was assessed at 24, 48, and 72 hours post–PI-103 treatment. The effect of S-TRAIL was assessed at 24 hours posttreatment. For combination treatments, glioma cells were treated with PI-103 for 24 hours followed by S-TRAIL treatment. Cell viability was measured using an ATP-dependent luminescent reagent (CellTiterGlo; Promega); and caspase activity was determined using DEVD-aminoluciferin (CaspaseGlo 3/7; Promega) according to manufacturer's instructions. All experiments were performed in triplicates.
Following treatment of glioma cells with PI-103 (1 μmol/L) for 24 hours, control and treated cells were pulsed for 30 minutes with 10 μmol/L of bromodeoxyuridine (BrdU; Amersham). The cells were trypsinized and fixed with ice-cold EtOH and denatured with 2N HCL/0.5% TritonX-100. BrdU incorporation was detected by incubation of cells with anti-BrdU antibody (Becton Dickinson) followed by FITC (fluorescein isothiocyanate)-conjugated anti-mouse antibody (Vector Labs) for 30 minutes at room temperature (RT). Cells were washed and incubated with propidium iodide (10 μg/mL; Sigma) and RNase A (250 μg/mL) and analyzed using flow cytometry (Becton Dickinson) and CellQuest software.
The invasive capacity of primary human GBM cells was tested using in vitro invasion assays (Becton Dickinson BioCoat Matrigel Invasion Chamber). GBM6, GBM8, and GBM12 lines were collected with Versene and 1 × 105 cells were seeded on each well of a 24-well, matrigel-coated invasion plate in serum-free and phenol red–free DMEM. The cells were induced to invade toward a chemoattractant placed in the lower chambers (5% FBS in phenol red–free DMEM). After incubation for 24 hours, the noninvading cells were removed from the upper surface of the invasion membrane and the cells on the lower surface were stained with Diff-Quick staining kit (IMEB Inc.). The average number of cells per field was determined by counting the cells in 6 random fields per well in 10× images of each well captured using (Olympus IX51). The effect of PI-103 on invasion was tested by adding PI-103 at the time of seeding cells in invasion chambers. All experiments were performed in triplicates.
Following sequential treatment with PI-103 and S-TRAIL (each for 24 hours) glioma cells were lysed with NP-40 buffer supplemented with protease (Roche) and phosphatase inhibitors (Sigma). Thirty micrograms of harvested proteins from each lysate was resolved on 10% SDS-PAGE, immunoblotted with antibodies against p-Akt (Ser473), total Akt, p-S6 (Ser235/236), total S6, poly(ADP-ribose) polymerase (PARP; Cell Signaling) or α-tubulin (Sigma), and detected by chemiluminescence after incubation with HRP (horse radish peroxidase)-conjugated secondary antibodies.
To establish cocultures of glioma cells and mNSCs, 0.5 × 105 Gli36-EvIII-FmC cells were seeded on 24-well plates and grown to 80% confluence in DMEM supplemented with 5% FBS for 24 hours. mNSC neurospheres expressing S-TRAIL or GFP (consisting of 0.5 × 105 cells per neurospheres) were placed on top of untreated or PI-103–treated glioma monolayers and cultured in mNSC medium. Cells were visualized 24 hours post–mNSC addition using fluorescence microscope (Olympus IX51). Representative images were processed with DP2-BSW Software (Olympus). Glioma cell viability was measured by measuring the Fluc activity of cells with d-luciferin as described previously (20).
In vivo experiments
SCID mice (3 weeks of age; Charles River Laboratories) were implanted with Gli36-EvIII-FmC cells (5 × 106 per mouse; n = 20) subcutaneously. To test the effect of PI-103 in vivo, each mouse was administered 25 mg/kg of PI-103 in a mixture of saline (prepared from stocks 50 μmol/L in DMSO) or control solution intraperitoneally (i.p.) every day. Mice were imaged for Fluc activity (tumor volumes) by bioluminescence imaging (BLI) on days0 (the first day of drug administration), 7, 11, 14, and 17 as described previously (28). To establish intracranial gliomas, 0.5 × 105 Gli36-EvIII-FmC cells were implanted stereotactically [from bregma, AP: −2 mm, ML: 1.5mm V (from dura): 2 mm; n = 24]. Three days postimplantation, PI-103 was administered to mice daily for a period of 17days (for PI-103 alone) or shorter (3 or 7days for combination therapies). For combination therapy, mice were implanted with mNSC-S-TRAIL (n = 6) or control mNSC-GFP (n = 6) on the last day of PI-103 administration. The effects on intracranial tumor growth were monitored by bioluminescence imaging as described above. On day 15, mice were perfused with 4% PFA (paraformaldehyde) and tissue was processed for histopathologic analysis. All in vivo procedures were approved by the Subcommittee on Research Animal Care at Massachusetts General Hospital.
Pharmacokinetic/pharmacodynamic assessment of PI-103
Mice bearing intracranial or subcutaneous gliomas were dosed i.p. with 25 mg/kg of PI-103 in 20% hydroxypropyl β-cyclodextrin daily for 5 days. Two or 5 hours after the last dose, blood was collected from vehicle- or PI-103–treated mice, centrifuged, and plasma was kept at −80°C. Subcutaneous and intracranial tumors were isolated, snap frozen in liquid nitrogen, and kept at −80°C until analysis. Quantitative analysis of PI-103 was achieved by liquid chromatography tandem mass spectrometry (LC-MS/MS). Specifically, tumor tissues were homogenized with Tissuelyzer (Qiagen) and all samples were precipitated in acetonitrile containing 250 ng/mL of cerbutamide (internal standard) and diluted in mouse serum. Analysis was performed using multiple reaction monitoring on API 4000 triple quadrupole system that is equipped with electrospray ionization source (Applied Biosystems). Chromatographic separation was done on LC-10A DvP pump (Shimadzu) and Polar-RP column (50 × 2.0mm, 4 μm; Phenomenex) in 0.1% formic acid at 0.5 mL/min for 3 minutes. Quantitation was done against the mouse serum standard curves. For pharmacodynamic assessment of PI-103, tissue proteins were extracted and analyzed for p-Akt by Western blotting as described above.
Mice were perfused and tumors were processed for paraffin embedding. For p-Akt staining, paraffin sections were subject to de-paraffinization and antigen retrieval in citrate buffer (pH=6.0), blocked with 5% normal goat serum in PBS-TritonX and incubated with primary antibody at 4°C overnight. Sections were washed 3 times with PBS, incubated with biotinylated secondary antibody for 30 minutes at RT and with ABC reagent (Vectorlabs), followed by development with diaminobenzidine. After hematoxylin counterstaining, sections were visualized with light microscopy (Olympus IX51)
Data were analyzed by Student's t test when comparing 2 groups. Data were expressed as mean ± SEM and differences were considered significant at P < 0.05.
PI-103 inhibits glioma cell proliferation and attenuates tumor growth in vivo
To test the effect of PI-103 on glioma cell proliferation, we chose 7 different established glioma lines, Gli36, Gli36-EvIII (as previously characterized; ref. 29; fast-growing Gli36 line that overexpresses a constitutively active form of EGFR), Gli79, U87MG, U251, LN229, and A172. Among these cell lines, U87MG, U251, and A172 harbor PTEN mutations with a comparably overactive PI3K-Akt pathway. PI-103 caused inhibition of Akt phosphorylation as well as S6 phosphorylation as assessed by immunoblotting (Fig. 1A). PI-103 also resulted in a dose-dependent (Fig. 1B) and time-dependent (Supplementary Fig. 1) inhibition of cell proliferation in all tested glioma lines. We further analyzed the effects of PI-103 on a selected glioma line, Gli36-EvIII, and showed that PI-103 resulted in a proliferative arrest in the cell cycle with a proportional increase in G0–G1, a decrease in the S phase and no change in the G2 phases of cell cycle (Fig. 1C).
To assess the effect of PI-103 in both subcutaneous and orthotopic models of glioma, we used the highly malignant Gli36-EvIII glioma line, engineered to express Fluc-mCherry, Gli36-EvIII-FmC (Supplementary Fig. 2). A significant attenuation of tumor growth was observed in mice bearing subcutaneous tumors treated with daily PI-103 i.p. injections as compared with controls (Fig. 1E), revealing that PI-103 has the potential to block the growth of this fast-growing glioma in mice. To test the effect of PI-103 orthotopically, we established Gli36-EvIII-FmC intracranial gliomas in mice. PI-103 treatment slowed down the growth of intracranial tumors as compared with the controls (Fig. 1F). The presence of PI-103 in the intracranial tumors was confirmed by the assessment of PI-103 levels in the brains of tumor-bearing mice (Table 1). To assess the concentration of PI-103 in the subcutaneous and intracranial tumor compartments relative to plasma levels, we performed LC-MS/MS on tumors and plasma collected from mice treated with (i) a single dose i.p. injection of PI-103 and (ii) repeated daily i.p. injection of PI-103. Accordingly, the daily injections of PI-103 resulted in higher levels of the drug in both compartments possibly due to the establishment of steady-state levels (Table 1). Importantly, we observed that the intracranial tumor compartment had comparable PI-103 levels to the subcutaneous tumor compartment relative to plasma levels in both treatment regimens (Table 1), suggesting that PI-103 has the potential to reveal efficacy in orthotopic gliomas. The growth-limiting effects of PI-103 in the in vivo models were partly due to a downregulation of Akt activity as evident by the decreased p-Akt staining in the paraffin sections extracted from the tumor tissue (Fig. 1G and H) and decreased p-Akt levels in the tumor tissue extracts (Supplementary Fig. 3). These results demonstrate that PI-103 is a potent inhibitor of growth in mouse models of malignant gliomas.
PI-103 augments the response of glioma cells to S-TRAIL
After confirming the in vivo potential of PI-103 in gliomas, we tested whether PI-103 acts in concert with locally delivered TRAIL through simultaneous inhibition of cell proliferation and activation of cell death pathways. A combination treatment of PI-103 and S-TRAIL resulted in a significant inhibition of cell viability compared with control conditions or S-TRAIL alone in the majority of glioma lines with varying sensitivities to TRAIL-mediated apoptosis (Fig. 2A). Next, we assessed the effect of PI-103 on TRAIL response of the highly proliferating and malignant Gli36-EvIII line. PI-103 treatment prior to S-TRAIL treatment elevated the TRAIL-induced cell killing as shown with varying doses of PI-103 and S-TRAIL (Fig. 2B). The augmented TRAIL response with PI-103 treatment was also evident at the caspase activation level. Specifically, whereas TRAIL treatment induced caspase 3/7 activity up to 2.5-fold of basal levels in control cells, cells previously exposed to PI-103 exerted up to 4-fold increase in caspase activity with TRAIL (Fig. 2C). One of the effector proteins downstream of caspases, PARP, was not affected by PI-103 alone but its cleavage upon TRAIL treatment was enhanced by PI-103 (Fig. 2D). The augmented TRAIL response was also observed in the Gli36 line that does not express EGFRvIII, suggesting that the effect of PI-103 on TRAIL-induced cell killing is independent of EGFR variant status in gliomas (Supplementary Fig. 4). These results reveal that PI-103 works in concert with TRAIL-induced apoptosis in glioma cells.
Effect of PI-103 and stem cell–derived TRAIL in cocultures of mNSCs and glioma cells and in vivo
To test the combined effect of PI-103 and stem cell–delivered TRAIL on glioma viability in vitro, we first engineered primary mouse NSCs (mNSC) to express a secretable and a highly potent variant of TRAIL, S-TRAIL (from here on referred as mNSC-TRAIL; Supplementary Fig. 5). To observe the effect of mNSC-TRAIL on glioma cells in vitro, we established cocultures of glioma cells (Gli36-EvIII-FmC) and mNSCs (mNSC-TRAIL; and control mNSC-GFP). mNSCs initially spread and infiltrated into the Gli36-EvIII-FmC monolayer (Fig. 3A) and mNSC-TRAIL resulted in glioma killing as confirmed by significant decrease in glioma cell viability as compared with the controls (Fig. 3A). To test the effect of PI-103 in this system, the monolayer of Gli36-EvIII-FmC cells were treated with PI-103 for 24 hours prior to the seeding of mNSCs. A significantly increased glioma cell killing was observed in PI-103 and mNSC-TRAIL treatment over 24 and 48 hours as compared with the controls (Fig. 3B). These results reveal that PI-103 acts in concert with mNSC-TRAIL and results in increased cell killing in vitro.
To investigate the combined antiglioma potential of PI-103 and mNSC-TRAIL in vivo, we first demonstrated the functionality of mNSC-TRAIL by admixing glioma cells and mNSCs in subcutaneous model. Tumors growing in the presence of mNSC-TRAIL were significantly smaller than controls (Fig. 3C). To ultimately examine the effect of combined PI-103 and mNSC-TRAIL therapy in intracranial gliomas, mice bearing established Gli36-EvIII-FmC gliomas were treated with PI-103 over a brief period of 3 or 7days and then implanted with mNSC-TRAIL or control mNSC-GFP in the close vicinity of the tumors (Fig. 3D). A significant decrease in tumor volumes was observed in mice implanted with mNSC-TRAIL after 3day PI-103 pretreatment (Fig. 3E). To test the effect of a longer PI-103 administration in this setting, we administered PI-103 for 7days prior to mNSC implantation. A significantly reduced tumor size compared with control mice was seen after mNSC-TRAIL implantation in PI-103–treated mice (Fig. 3F). Taken together, these results reveal that systemic delivery of PI-103 combined with mNSC-TRAIL results in marked attenuation of intracranial tumor growth.
PI-103 inhibits the proliferation and invasion of primary glioma–initiating cells
Although our established glioma lines serve as potential tools to examine the effect for PI-103 and TRAIL, they still lack the characteristics that mimic clinical glioma settings. As such, established glioma lines mainly form solid tumors when implanted into mice as opposed to the highly infiltrative nature of human gliomas. To this end, we utilized 3 primary human GBM lines that are enriched in CD133+ glioma-initiating cells and form highly invasive tumors in mice (26, 30). The treatment of GBM6, GBM8, and GBM12 lines with PI-103 slowed down their rate of proliferation in a dose- and time-dependent manner (Fig. 4A), with GBM8 being the most responsive cell line. Moreover, PI-103 treatment reduced the invasive ability of GBM cells in a matrigel-coated invasion assay to a far more extent than the inhibition of proliferation (Fig. 4B and C). Accordingly, GBM6 and GBM8 invasion was most affected by PI-103 after 24 hours (∼90% inhibition of invasion in these cells compared with only ∼40% inhibition in viability). PI-103 also inhibited PI3K and mTOR pathways as indicated by reduced Akt and S6 phosphorylation in a dose-dependent manner (Fig. 4D). To test whether PI-103 and TRAIL can work together to inhibit the proliferation of these clinically relevant GBM cells, we selected GBM8 line, which we previously characterized for its TRAIL response (18). Treatment of GBM8 cells with PI-103 prior to S-TRAIL augmented the TRAIL response in these cells (Fig. 4E). These results reveal that PI-103 inhibits invasive properties of GBMs and that PI-103 augments the TRAIL response, thus attenuating the aggressive characteristic of primary GBMs.
In this study, we demonstrate the effect of a dual PI3K/mTOR inhibitor, PI-103, in a panel of established and primary invasive glioma cell lines and provide evidence of the antitumor effects of systemically delivered PI-103 in intracranial glioma models. Furthermore, we show that PI-103 co-operates with stem cell–delivered S-TRAIL in mouse models of gliomas.
PI3K pathway is one of the most commonly deregulated signaling networks in gliomas (6); therefore, the components of this pathway are candidates for targeted therapies. A novel pyridinylfuranopyrimidine inhibitor, PI-103, has been shown to dually inhibit PI3K and mTOR and block cell proliferation in several cancer cell lines including gliomas (7–16, 31–34). In this study, we show that PI-103 inhibits cell proliferation in a dose- and time-dependent manner. Although these established glioma lines are the most commonly used models in vitro, they fail to recapitulate the clinical properties of tumors. Therefore, recent studies have focused on primary glioma lines and indicated a role for tumor-initiating cells in these lines. In an effort to test the effect of PI-103 in such models, we employed 3 invasive primary glioma lines with high CD133+ expression (26, 30) and showed that PI-103 alone is effective in attenuating growth as well as invasiveness of these cells. This suggests that, once delivered efficiently, PI-103 might block not only the proliferation but also the invasion of these tumors. Our results agree with previous reports that PI-103 can serve as a widely effective antiglioma agent (13). Furthermore, we show that PI-103 is effective in reducing the growth of orthotopic gliomas in vivo by using bioluminescence imaging. Our assessment of PI-103 levels in intracranial tumors in comparison with subcutaneous tumors by LC-MS/MS demonstrates the availability of PI-103 in both tumor compartments, suggesting that PI-103 might serve as a good tool compound for efficacy studies in orthotopic gliomas as well as subcutaneous models. Also demonstrated by our analysis of p-Akt levels in the tumor tissues, which can be regarded as a pharmacodynamic marker, we can now be certain about the effect of PI-103 in our glioma models. To our knowledge, this is the first study demonstrating the effect of this dual PI3K/mTOR inhibitor in an orthotopic glioma and is in line with the findings of a previous report on the pharmacologic characterization of PI-103 (15).
Although targeting the highly proliferative state of gliomas is a favorable therapeutic approach, it might not be sufficient in the eradication of tumors that are left behind after surgical intervention. Therefore, combinatorial approaches that target tumor cell proliferation and induce tumor-specific cytotoxicity would be very effective in eliminating recurrence and subsequent therapeutic failure. To this end, apoptosis-inducing reagents, such as TRAIL, have recently gained attention for the preclinical studies. However, the short half-life and the off-target toxicity of systemically delivered TRAIL pose a challenge in translating it into the clinics (17). On the basis of these limitations, we have previously established that stem cell–delivered TRAIL is very effective in glioma eradication (18–20, 27, 35, 36) due to its on-site delivery and long-term expression compared with systemically administered purified TRAIL. In this study, we report on the combination of PI-103 and mNSC-TRAIL in our glioma models and demonstrate the in vitro effects of PI-103 and S-TRAIL focusing on the viability and apoptosis of a panel of established glioma cell lines. Prior to our study, a few other studies reported the combination effect of PI-103 with EGFR inhibition (32), radiation (31), and chemotherapy-induced apoptosis (37). To our knowledge, this is the first study that examines the combined effect of PI-103 with TRAIL in glioma cells both in vitro and in vivo. Our findings suggest that PI-103 augments the response of glioma cells to TRAIL. However, it should be noted that whereas we did not observe a switch in the response of TRAIL-resistant gliomas to TRAIL by PI-103 treatment, we noticed an increase in the TRAIL response of glioma cells that were already fairlysensitive to TRAIL. Therefore, we do not argue that PI-103 is a TRAIL-sensitizing agent, but it has the potential tosignificantly augment the TRAIL-induced apoptosis in a subset of gliomas. Addressing the molecular interactions of PI-103–mediated PI3k/mTOR inhibition with TRAIL will further our understanding of their combination as a treatment strategy.
Our in vivo studies reveal that systemic delivery of PI-103 combined with mNSC-TRAIL results in marked attenuation of intracranial tumor growth, suggesting that combining systemically delivered PI3K/mTOR inhibitors with stem cell–delivered agents might lead to marked tumor eradication. It should be noted that we performed our combination therapy experiments with caution. We only briefly administered PI-103 prior to mNSC implantation to avoid any possible collateral damage to mNSCs from PI-103. When we briefly treated mice for only 3days with PI-103 prior to mNSC implantation, we did not observe a difference in tumor growth due to drug alone. However, upon administration of the drug for 7 days, we noticed a difference in tumor growth due to PI-103, which was further enhanced with mNSC-TRAIL. Therefore, optimal conditions for alternating the systemically delivered drugs with implanted stem cells should be defined for future translation into clinics.
Our results with intracranial glioma models suggest that the combination of systemically delivered PI3K/mTOR inhibitors and stem cell–delivered TRAIL offers promise in gliomas. Testing this approach with other PI3K/mTOR inhibitors, such as GDC-0941, which is undergoing phase I clinical trials (38), would extend our findings and help with the translation of our combination strategy to the clinics. This could be performed in brain tumor patients by administering the clinically approved PI3K/mTOR inhibitor systemically and implanting stem cell–secreting S-TRAIL in the tumor resection cavity after surgical resection of the tumor. This combination therapy would enhance the eradication of the residual tumor cells and prevent tumor recurrence. Taken together, our study might serve as an excellent foundation for future therapies combining systemically delivered cytostatic drugs, such as PI3K/mTOR inhibitors, with stem cell–delivered cytotoxic agents, such as TRAIL in gliomas.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
We thank Dr. Philip Lambert (VivoPath) for help with LC-MS/MS analysis; Dr. Andrew Kung (Dana Farber Cancer Institute) for providing pico2-Fluc-mCherry lentiviral vector; and Massachusetts General Hospital Pathology Core for help with histologic analysis.
This work was supported by American Cancer Society (K. Shah) and Alliance for Cancer Gene Therapy (K. Shah).
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.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received May 10, 2010.
- Revision received October 13, 2010.
- Accepted November 9, 2010.
- ©2010 American Association for Cancer Research.