BMS-777607 decreases glioma cell viability and induces glioma cell apoptosis in vitro

The exposure of U118MG and SF126 glioma cells with 12.5 μM BMS-777607 resulted in significantly reduced cell numbers in MTT assay after 24 hours of treatment (Figure ​(Figure2D,2D, left image). Single application of 12.5 μM BMS-777607 led to a significant increase of CPP32 activity in U118MG and SF126 cell lines after 24 hours (Figure ​(Figure2D,2D, middle image). We concluded that reduced cell viability by BMS-777607 resulted from the induction of apoptosis and reduction of proliferation.

BMS-777607 blocks glioma cell migration and invasive growth pattern

Next, we addressed the effects of BMS-777607 on glioma cell migration and invasion. Blockage of RTK-AXL signaling with initial dose of 12.5 μM BMS-777607 resulted in a significant decrease of migration rate in U118MG and SF126 cells measured by Boyden Chamber Migration assay after 3 hours (Figure ​(Figure2D,2D, left image). Invasive growth pattern of glioma cells was studied in the orthotropic brain slice culture model, a three-dimensional invasion assay closely related to an in vivo situation as the structure and organization of the brain tissue are preserved [19]. Under control conditions, only U118MG cells, but not SF126 cells, demonstrated invasive growth in this assay. Inhibition of RTK-AXL signaling with daily administration of 12.5 μM BMS-777607 abrogated tumor cell invasion into the brain tissue (Figure ​(Figure3A,3A, right image). Quantitative analysis using confocal microscopy confirmed this significant anti-invasive effect of BMS-777607 in this ex vivo model on day 8 (Figure ​(Figure3A,3A, left image).

Open in a separate window

Figure 3

(A) Orthotopic brain slice invasion assay with implanted, DiI stained tumor cell spheroids of cell line U118MG. Images show data at day 8 under daily administration of BMS-777607 with corresponding statistical analysis (n = 8, **p = 0.0040). Results were obtained with confocal microscopy. Scale bar indicates 10 μm. (B) Expression of AXl and P-AXL in HUVECs (− = negative control). (C) IC₅₀ value of BMS-777607 treatment in HUVECs. (D) left images: Tube formation assay with HUVECs after 4 and 20 hours and single administration of 12.5 μM BMS-777607. (D) right images: Statistical analysis at 4 and 20 hours of tube length (n = 5, 4 h: ***p < 0.0001, 20 h: ***p < 0.0001) and branching points (n = 5, 4 h: *p = 0.035, 20 h: ***p < 0.0001).

RTK-AXL is phosphorylated in HUVECs and inhibition displays antiangiogenic effect

RTK-AXL signaling has been also shown to be involved in endothelial cell proliferation and angiogenesis (5). Thus, we additionally studied the effects of BMS-777607 on endothelial cell biology using in vitro assays with HUVECs. Figure ​Figure3B3B shows RTK-AXL expression and RTK-AXL phosphorylation in HUVECs. The IC₅₀ value of BMS-777607 was assessed as shown in Figure ​Figure3C.3C. In accordance to this result, tube formation assay was carried out with 12.5 μM BMS-777607.

Consistent with in vivo findings we proved direct effect of BMS-777607 on endothelial cells HUVEC in tube formation assay. We observed a significant decrease of tube formation and branching points after 4 hours, which was still persistent after single treatment at the 20 hours time point (Figure ​(Figure3D3D).

I.p. administration of BMS-777607 selectively blocks RTK-AXL signaling in vivo and inhibits intracranial tumor growth in mouse xenografts

For in vivo experiments, U118MG and SF126 cells were implanted stereotactically into the brains of CD1NuNu mice. While U118MG xenografts showed a more invasive growth pattern, SF126 xenografts developed vascular proliferates, central necrosis and had a stronger proliferation rate (Supplementary Figure S2).

In both tumor models, treatment was initiated after proven tumor manifestation using MRI. In the more aggressive tumor SF126 model, treatment was started at day 3 after implantation while treatment in the U118MG model was started on day 7 after implantation (Supplementary Figure S3). In both models, BMS-777607 treatment (i.p. 2x/day) resulted in a significant decrease of intracranial tumor growth on day 14 after implantation. The most significant anti-tumor effect was seen in U118MG tumor model (Figure ​(Figure4A).4A). Here, 30 mg/kg BW BMS-777607 resulted in more than 90% tumor reduction after 6 days of treatment, in two cases we even observed complete tumor regression (n = 5, Figure ​Figure4C,4C, left image). In the SF126 tumor model we detected a 56% tumor volume reduction. Increase of dosage up to 100 mg/kg BW lead to further regression. We showed that BMS-777607 displayed a dose dependent effect when applying it at 30 mg/kg BW and 100 mg/kg BW (Figure ​(Figure4C,4C, right image).

Open in a separate window

Figure 4

(A) MRI results of U118MG after 7 days of treatment with 30 mg/kg BW BMS-777607 (BMS30, n = 7) and vehicle at day 14 (n = 7). (B) MRI results of SF126 after 11 days of treatment with 30 mg/kg BW (BMS30, n = 5), 100 mg/kg BW BMS-777607 (BMS100, n = 7), and vehicle (n = 7) at day 14. MRI images are contrast enhanced. (C) Both images display tumor volumetric analysis after 14 days experiment. U118MG vehicle vs. BMS30 **p = 0.0012. SF126 vehicle vs. BMS30: *p = 0.015, vehicle vs. BMS100: **p = 0.0013, BMS30 vs. BMS100: ns. (D) left image: Western blot analysis with regulation of RTK-AXL and MET kinase phosphorylation on intracranial tumor tissue after i.p. administration of BMS-777607 at day 14 (− = negative control). (D) right image: Statistical analysis of three independent experiments of Western blot with tumor lysates (n = 3, ***p < 0.0001).

Protein analysis of the intracranial tumors revealed reduction of RTK-AXL phosphorylation on Western blot analysis. Furthermore, we did not observe compensatory MET kinase activation after 14 days of treatment (Figure ​(Figure4D4D).

BMS-777607 exerts multiple anti-tumor effects in vivo

The treatment with BMS-777607 revealed multiple anti-tumor effects in vivo. In accordance with our in vitro data, BMS-777607 treatment resulted in a reduction of tumor cell proliferation and increase of apoptotic events in SF126 tumor xenografts compared to the control (p = 0.012, Figure ​Figure5A5A and Figure ​Figure5C,5C, left image). Qualitative analysis of Ki67 expression within the tumor tissue showed a decreased proliferative activity under treatment with BMS-777607 (Figure ​(Figure5B5B).

Open in a separate window

Figure 5

(A) Intratumoral apoptotic events in SF126 xenografts under treatment with BMS-777607 and vehicle. (B) CD31 staining (red) shows differences in vessel size and density under treatment with BMS-777607 in vivo. Proliferative activity is displayed qualitatively with Ki67 staining (green). (C) Images show corresponding statistical analysis of apoptotic events within tumor tissue (n = 3, *p = 0.0121) and statistical analysis of intratumoral vessel density (n = 3, ***p < 0.0001). Scale bar indicates 50 μm.

Having shown that inhibition of RTK-AXL phosphorylation in endothelial cells also resulted in an impaired tube formation, we also investigated the effect of BMS-777607 treatment on the glioma blood vessel surface and vascular architecture using immunohistochemistry. BMS-777607 reduced glioma vessel surface and vascular size, revealing its anti-angiogenic efficacy (Figure ​(Figure5B5B and Figure ​Figure5C,5C, right image). These findings were consistent in the border zone and center of the tumor where different angiogenic activities are usually observed (data not shown).

Cell culture

Human high-grade glioma cells U118MG were obtained from American Type Culture Collection (ATCC). Human high-grade glioma cells SF126 were obtained from the JCRB Cell Bank. Both were primary tumor cell cultures derived from surgical specimens of human GBM, WHO Grade IV. Cell authentication was carried out with LGC Standards Cell line Authentication service in June 2014 with 16 loci service of short tandem repeat profile. Tumor cells were maintained as monolayer cultured in tumor growth medium at 37°C, 5% CO₂, 95% humidity in a tissue culture incubator. Growth medium was comprised of Dulbecco's modified Eagle's medium (DMEM, Invitrogen) supplemented with 10% fetal calf serum (FCS) and 1% antibiotics (penicillin/steptamycin). Prior starting experiments, each cell population was grown in equal (80%) confluence in culture dishes under normal conditions or starving conditions (culture medium containing 0% FCS). Cell count and cell vitality was assessed with CASY^® Cell Counter TT (OLS).

HUVEC cell line was obtained from PromoCell. Cells were cultured accordingly to manufactures instructions.

Spheroids

Cells were relabeled prior forming spheroids with DiI (invitrogen) according to manufactures instructions. Cells were seeded in an uncoated, non-adhesive 96-well plate (Sarstedt). Culture medium contained 20% Methocell medium consisting of 6 gr. carboxymethylcellulose (Sigma Aldrich) and 250 ml of the preheated ECBM (Endothial cell growth medium, Lonza). 5.000 cells were plated per well in 100 μl Methocell medium. Spheroids grew within 48 hours. Round-shaped spheroids were selected for experiments with a diameter size of 300–500 μm.

Cell viability assay-MTT assay

5.000 cells were seeded per well of a 96-well plate in triplicates and incubated with 0.1 ml DMEM supplemented with 5% FCS. Fractions were treated with DMSO or treated with compound at different concentrations (1.56 μM–50 μM). Culture medium was replaced or exchanged after 24, 48 and 72 hours. 100 μl of MTT reagent (Thiazolyl Blue Tetrazolium Bromide, Sigma) was added and incubated for 3 hour. Crystals were solved with 100 μl DMSO-isopropanol (Sigma Aldrich) (1:1) under constant shacking for 10 minutes. Absorption of supernatant was measured at 570 nm with plate reader (infinite200).

Apoptosis assay

Apoptosis was measured with caspase3/CPP32 Colorimetric Assay Kit (Biovision). Cells were cultured for 24 hours with 50 ng/ml TNF-alpha (Dianova) or with compound at a concentration of 12.5 μM. The non-induced control was treated with solvent DMSO at a final concentration of 0.5% (Sigma Aldrich). 200 μg protein were incubated with 2 × Reaction buffer containing 10 mM DTT and DEVD-pNA substrate. Absorption was determined at 405 nm. Fold increase in CPP32 activity was determined by comparing results of treated group with the level of the non-induced control. Assay was performed in triplicates for the calculation of mean values and standard deviation (n = 3).

Boyden chamber migration assay

Experiments were performed with 50.000 cells per chamber. Prior starting the experiment the inserts were coated with fibronectin (Sigma Aldrich) at a final concentration of 10 μg/ml in cold PBS. Inserts were incubated with fibronectin for 2 hours in the cell incubator. Cell suspension and compound were added to the upper chamber (compound concentration: 12.5 μM). The lower chamber was filled with 110 μl of fibroblast-conditioned medium (FCM) to stimulate cell migration. A filter with 8 μm pore size separated the upper and lower chamber. After a migration time of 3 hours, cells adhering to the bottom side of the filters were fixed with methanol for 5 minutes. The inserts were dried for 5 minutes and then stained with DAPI (Sigma Aldrich, diluted in PBS 1:100). Cells per field were counted at a magnification of 10 fold, 4 fields per filter were averaged (n = 5).

Orthotopic brain slice invasion assay

For preparation of brain slice cultures, 3 to 6-days-old mouse pups (C57BL/6NCrl, Charles River) were used. Brain hemispheres were sectioned in 300 μm thick coronary slices (McIllwain tissue chopper). Slices were separated in ice cold dissection medium containing 99 ml MEM and 1 ml glutamine (Gibco) and cultured onto semipermeable membrane of inserts (Transparent PET membrane, 1.6 × 10⁶-pores/cm², BD, Falcon^®) with 1 ml culture medium containing 23 ml MEM (Gibco21575), 12.5 ml Horse serum, 12.5 ml BME (Gibco 41010), 1.5 ml of 20% glucose (Roth) and 0.5 ml Glutamine (Life Technologies, L-Glutamine 200 mM). Medium was replaced every two days. 24 hours after cultivation pre-labeled spheroids were placed in the area between striatum and corpus callosum. Invasion of DiI-labeled cells were monitored using confocal microscopy. After 8 days slices were fixed in 4% PFA for 2 hours at 37°C. Nuclear staining of slices was performed with To-PRO-3 according to manufacturer's instructions (Life Technologies). Slices were transferred onto glass slides (Langenbrinck) and mounted in mounting medium (Immuno mount, Thermo Scientific) for confocal analysis. Confocal analysis was performed with Zeiss Confocal microscope. For quantitative analysis distance of invaded cells from border of spheroid was measured (n = 8).

Tube formation assay

10 μl BD matrigel basement membrane matrix (BD 354230) was pipetted to a 15-well plate (μ-Slide Angiogenesis, ibiTreat), matigel solution formed gel within 30 minutes incubation time at 37°C. Harvested HUVEC cells were suspended in a concentration of 200 000 cells/ml with 12.5 μM concentration of compound or vehicle (0.5% DMSO). 50 μl of cell suspension pro well was added onto solidified gel and incubated for 4 to 20 hours. Tube formation was analyzed with Zeiss Axiovision fluorescence microscope. Tube length and branching points were quantified using online service Wimasis Image Analysis (n = 5).

Experimental tumor models

Athymic CD-1 Nu/Nu nude mice (female, obtained from Charles River) were maintained within a pathogen germ-free environment and were used at 6–10 weeks of age. The weight ranges was in a range of 20 to 28 gram. Experiments were performed in accordance with the approved institutional protocol and the guidelines of the Institutional Animal Care and Use Committee. General anesthesia was induced with 7% Ketaminhydrochlorid (Ketavet, Pfizer), 8% Xylaxine (Rompun 2%, Bayer) dissolved in aqua via i.p. injection. Phenoxymethylpenicillin (InfectoCilin, 5 Mega) was administered via intramuscular injection. For stereotactic tumor cell implantation, mice were positioned in the stereotactic platform. A longitudinal skin incision was performed in the mid scalp extending from the ears caudally. The burr hole was drilled with a syringe of 23′G 2 mm posterior and 1.5 mm laterally to bregma. Hamilton syringe was loaded with glioma cells dissolved in DMEM without FCS. Hamilton syringe was inserted gently to a depth of 4 mm. After injection wound was closed with suture (Prolene 2.0, Ethicon). Finally, mice were placed on a heating plate set to 37°C until regaining consciousness. Drinking water was enriched with tramadol calculating 15 mg/kg BW (Grünenthal). Contrast enhanced MRI (gadopentetate dimeglumine; Magnesvist, Bayer) scans were performed under anesthesia with 7T Bruker MRI (PharmaScan 70/16 US, Bruker Software Paravision 5.1). Tumor volumetric analysis was carried out with Analyze 10.0 and ImageJ software. Mice were sacrificed at day 14 and the body was perfused with PBS under anesthesia. The whole brain or tumor was dissected and immediately frozen in liquid nitrogen and stored at −80°C (n = 7).

Immunoprecipitation (IP)

IP was performed using Protein A-Sepherose 4A beads (Life Technologies). Low salt IP-lysis buffer was enriched with Halt^™ Phosphatase Inhibitor Single-Use Cocktail (Thermo Scientific) and Halt Protease Inhibitor Single-Use Cocktail (Thermo Scientific). Immuncomplexes were prepared with 200 μg protein and 1:50 dilution of anti-AXL (C89E7) rabbit antibody (Cell Signaling). Assay was performed according to previous protocol (14). Detection of RTK-AXL phosphorylation was determined using antibody 4G10 mouse anti-tyrosine antibody (Max Planck Institute of Biochemistry, Dilution 1:2.000).

Western blotting (WB)

Protein samples were boiled for 5 min with Laemmli sample buffer (Bio-Rad Laboratories) and loaded into pre-poured Tris-HCl-glycine SDS-PAGE gels (stacking gel 4%, resolving gel 6% or 8%). Gels run at 150 V for 1.5 hours following transfer to a polyvinylidene difluoride membrane (PVDF, Bio-Rad Laboratories) at 40 mA constant current for 2 h. Blots were blocked with 5% BSA in 1xTBST, primary and secondary antibodies were dissolved in TBST. Following primary antibodies were used. R & D Systems: human phopho-AXL (Y779) mAb (Clone 713610), Cell Signaling: anti-AXL C89E7 rabbit mAb, phospho-MET (Tyr1234/1235) rabbit mAb (D26) XP^®. Santa Cruz: Gas6 antibody (C-20). Sigma Aldrich: mouse monoclonal β-actin antibody (clone 1A4). Immunocomplexes were visualized using a second HRP-conjugated anti-rabbit or anti-mouse antibody (Pierce Biotechnology). For development we used ECL Kit (Sigma). ImageJ Software was used for densitometry analysis. Reported values were first normalized to the loading control and then multiplied by a constant to reach the lowest whole integral. Each western blot was carried out with a negative control consistent of sample diluent or beads incubated with antibody and sample diluent only.

Immunohistochemistry and immunofluorescence cells

Adherent cells were plated onto cover slips and fixed with 4% PFA for 10 minutes at room temperature. Cells were blocked with 1% casein for 45 minutes at room temperature. Antibodies were diluted in 0.5% casein. The primary antibody was incubated for one to two hours, the secondary antibody for one hour. Wash steps were carried out with TBST. Negative control was carried out without primary antibody incubation and is shown in Supplementary Figure S1 for anti-AXL antibody and anti-phospho-AXL antibody.

For immunofluorescence human phopho-AXL (Y779) mAb (Clone 713610, R & D Systems) and anti-AXL C89E7 rabbit mAb (Cell Signaling) were used. The following secondary, fluorescently labeled antibodies were used: FITC-conjugated donkey anti-rabbit IgG (Dianova, 711-095-152), Cy3-conjugated donkey anti-mouse IgG (Dianova, 115-165-146).

Mouse tissue

Mouse brain was mounted in paraffin block in a.p. orientation. 8–20 μm coronary sections were prepared with cryostat (Microm Cryo-Star HM 560 Cryostat, GMI). Sections were fixed with PFA 4%. Staining of intratumoral vessels and proliferation was carried out with purified rat anti-mouse CD31 antibody (clone MEC13.3, BD Pharmingen^™) and rabbit anti-Ki-67 antigen monoclonal antibody (clone SP6, Diganostic BioSystems). The following secondary, fluorescently labeled antibodies were used: FITC-conjugated donkey anti-rabbit IgG (Dianova, 711-095-152), Cy3-conjugated donkey anti-rat IgG (Dianova, 712-165-153). Apoptotic activity within the tumor was assessed using Apoptaq kit (Millipore). Staining procedure was done according to manufacturer's instructions.

Tissue staining for AXL and phospho-AXL was carried out on PFA fixed slices. Cells and brain slices were finally counterstained with DAPI (Thermo Scientific). Slices were analyzed using a fluorescence microscope (Zeiss).

For statistical analysis we used ImageJ software. For qualitative analysis of fluorescent signal, all parameters were maintained as set for the initial image.

Human tissue

Immunohistochemical staining of formalin-fixed, paraffin-embedded (FFPE) tissue sections (4 μm-thick) was performed on a VENTANA Benchmark XT automated staining instrument according to the manufacturer's instructions. Slides were de-paraffinized using EZ prep solution (Ventana Medical Systems, Tucson, AZ) for 30 minutes at 75°C. Antigen retrieval was accomplished on the automated stainer using CC1 solution (Ventana Medical Systems, Tucson, AZ) for 60 minutes at 95°C. Briefly, the anti-phospho-AXL antibody (Human Phospho-AXL (Y779) monoclonal mouse IgG Clone 713610, R & D Systems, dilution 1:50) was applied and developed using the iVIEW DAB Detection Kit (Ventana Medical Systems). All slides were then counterstained with hematoxylin for 4 minutes.

Patient data

Clinical data was assessed under an institutional review board-approved protocol and de-identified for patient confidentiality. We included 16 GBM patients, which have been treated in our institution in the year 2015. GBM diagnosis was assured of two independent neuropathologists. RTK-AXL and phospho-AXL expression was quantified with IP and WB analysis in 8 patients with primary GBM and in 8 patients with recurrent disease, which received re-resection. FFPE tissue sections of these individual 16 patients were stained with anti-phospho-AXL antibody.

Microscopy

Images were recorded by fluorescent microscope from Zeiss (Obeserver Z1). Following objectives were used: 5× EC PlnN, 5×/0.16 DIC0 (resolution: 2.0 μm), 10× Pln Apo, 10x/0.45 DIC II (resolution: 0.74 μm), 20× Pln Apo, 20×/0.8 DIC II (resolution: 0.42 μm). We use a HAL 100 and detectors for DAPI, GFP and DSRed. Pictures were processed and recorded with Image software Axio Vision Rel. 4.8. Statistical analysis was carried out with ImageJ Software 1.46r.

The authors thank Petra Matylewski for establishing the staining of human tissue and excellent technical assistance. We further thank Susanne Müller for technical support with animal MRI.

We thank ShangHai Biochempartner Co. for purchasing BMS777607.