Background
The standard treatment of glioblastomas includes surgical resection, fractionated radiation and concomitant as well as adjuvant chemotherapy with temozolomide. This treatment has improved survival but despite these improvements the median survival is only 14.6 months [
1]. Two biological aspects believed to be highly responsible for tumor recurrence are the resistant tumor stem-like cells [
2] and the invasive properties of glioblastomas [
3]. Treatments targeting the tumor stem cells and the invasive cells are therefore of great interest.
The lysosomal cell death pathway involves lysosomal membrane permeabilization, thus being a cell death pathway still functional in the tumor cells. By lysosomal membrane permeabilization the lysosomal content translocates to the cytosol and may cause programmed cell death [
4,
5]. Among the proteases responsible for this cell death are the cathepsins, which are still active at neutral pH [
4]. Most excitingly the cathepsins are capable of inducing a caspase-independent and mitochondrial-independent cell death promoting cell death in tumor cells with multiple defects in the classical apoptosis pathway [
6].
A compound shown to accumulate in the lysosomes causing lysosomal membrane permeabilization and release of the cathepsins to the cytosol is the σ 2 receptor agonist, siramesine (Lu-28-179; 1 V-[4-[1-(4-fluorphenyl)-1H-indol-3-yl] butan-1-yl]spiro[isobenzofuran-1(3H),4 V-piperidine]). Siramesine was originally designed to treat anxiety and depression and it was shown to successfully enter the brain in mice ex-vivo binding studies [
7]. The drug was well tolerated and non-toxic in humans but the effect was not satisfactory [
8]. Because of the lack of side effects and the suggested role of σ 2 receptors triggering cell death, siramesine was investigated as an anti-cancer drug [
9]. Indeed, siramesine induced cell death in immortalized and tumorigenic cells [
9] by lysosomal leakage of cathepsins and oxidative stress [
10]. Siramesine was found to directly destabilize the lysosomal membrane followed by lysosomal dysfunction leading to permeabilization of the membrane and release of cathepsins to the cytosol resulting in cathepsin mediated cell death [
9,
10]. Importantly, the cell death was independent of caspases and P53 tumor suppressor protein and insensitive to the anti-apoptotic effect of Bcl-2 [
9‐
11].
The aim of the current study was to investigate the effect of siramesine on glioblastoma cells using approaches comprising both immature tumor stem-like cells, differentiated tumor cells and migrating tumor cells. Accordingly, we used both standard glioma cell lines and patient-derived spheroids cultures with tumor-initiating stem-like cells [
12,
13]. To perform a thorough testing, spheroid cultures were implanted in three dimensionel organotypic brain slice cultures [
13,
14] and used for generation of patient-like tumors in a glioblastoma xenograft mice model [
12,
15,
16]. Using these approaches the in vitro results with tumor cells and spheroids suggested a potential of lysosomal destabilizing drugs in killing glioblastoma cells, but in the organotypic spheroid-brain slice culture model and the in vivo xenograft model siramesine was without effect.
Methods
Cells and treatments
In the present study we used the commercial human glioma cell lines, U87, A172, T98G (all from European Collection of Cell Cultures (ECACC), catalogue numbers: 89081402, 88062428 and 92090213, respectively) and U251 (from CLS, cell lines service, Germany, catalogue number: 300385) cultured in serum containing medium as described earlier [
12].
The glioblastoma stem cell-like containing spheroid (GSS) cultures T78, T86 and T87 were established in our laboratory and cultured under stem cell promoting conditions as neurospheres (spheroids). The spheroids were cultured in serum-free medium as described earlier [
12]. The GSS cultures have the ability to form new spheroids at clonal density, a karyotype typical of glioblastomas, the ability to differentiate into cells expressing neuronal, astrocytic and oligodendrocyte markers upon culturering in serum-containing medium and the ability to form highly invasive tumors upon orthotopic xenografting [
15].
Siramesine was kindly provided by H. Lundbeck A/S, Valby, Denmark. Dimethylsulfoxid (DMSO) was used as a solvent for siramesine in all in vitro studies. The cathepsin and caspase inhibitors used were z-FA-FMK (cathepsin B, L and S), Ca-074-Me (irreversible capthepsin B inhibitor), z-DEVD-FMK (caspase 3 and 7), and z-DEVD-CMK (caspase 3), all from Bachem. The inhibitors were added 1 h before the addition of siramesine. The glioma cell lines were incubated with inhibitors and siramesine for 48 h before measuring cell proliferation and viability.
Cell proliferation and viability in adherent glioma cell lines
The cell proliferation was analysed using the cell proliferation reagent WST-1 (2-(4-Iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt, from Roche) according to the manufactures instructions. After 24 or 48 h of incubation of cells with siramesine, the WST-1 reagent was added and the absorbance at 450 nm was found using an absorbance microplate reader (BioTek ELx808, Holm and Halby, Denmark).
A lactate dehydrogenase kit (LDH Cytotoxicity Detection Kit, Roche) was used to detect cell death according to the manufactures instructions, detecting LDH released from cells with a permeabilized membrane. Medium from each well in a 96-well plate was transferred to a new plate where after the LDH reaction mix were added and the absorbance measured at 450 nm in the absorbance microplate reader.
Change in lysosomal acidity
The glioma cell lines were incubated in 12 well glass bottom plates until approximately 80% confluence and exposed to siramesine (5–30 μM). After 1 h, Acridine orange (5 μg/ml, Invitrogen) was added to the cells and incubated for 15 min at 37 °C. Images were recorded using confocal microscopy (Nikon, Inverted Microscope, ECLIPSE TE2000-E). Analysis was made using the image analysis tool Visiomorph™ (Visiopharm, Hørsholm, Denmark). The results represent the proportion of red/green + yellow area, accounting for the loss of red staining and gain of green and yellow staining in the cells.
Cell death in GSS spheroid cultures
Cell death in the GSS spheroid cultures was analyzed using propidium iodide (PI) (2 μM, Invitrogen) and Hoechst 33342 (10 mM, Invitrogen). The spheroids were treated with siramesine for 24 h before addition of PI for 1 h followed by Hoechst for 15 min. The spheroids were then analyzed using a confocal microscope recording z-stacks which afterwards were superimposed into one image. The images were analyzed using the image analysis tool Visiomorph establishing a classifier identifying the PI uptake as red staining area and the total nuclear area as the sum of red staining and blue Hoechst staining. The area ratio between red staining and red plus blue staining was calculated.
Spheroids treated with siramesine for 24 h were dissociated, the cells were counted and 1000 cell/ml were added to each well in a six-well plate. The cells were allowed to form secondary spheroids and after 2–3 weeks the number of spheroids in each well was counted.
Immunohistochemical staining of GSS cultures exposed to siramesine
After having analyzed the spheroids exposed to siramesine with PI uptake, the spheroids were fixed in 4% formalin for 24 h followed by paraffin embedding. Three-micrometer sections of the paraffin-embedded spheroids were cut on a microtome. Thereafter one section was hematoxylin eosin stained and adjacent sections immunohistochemically stained on a Dako autostainer, Universal Staining System. For immunohistochemical staining, paraffin sections were deparaffinized and heat-induced epitope retrieval was performed by incubation in a TEG buffer solution of 10 mmol/L Trisbase and 0.5 mmol/L EGTA (CD133, nestin, Bmi-1, Sox 2, Ki67, Lamp-2, cathepsin B) or EDTA buffer (Cathepsin B). After blocking of endogenous peroxidase activity by incubation in 1.5% hydrogen peroxide (H2O2), the sections incubated for 60 min with primary antibodies against CD133 (1 + 40, CD133/1 W6B3C1, Miltenyi Biotec), Nestin (1 + 3000, 196908, R&D Systems), Bmi-1 (1 + 400, F6, Upstate Biotechnology), Sox 2 (1 + 400, 245610, R&D Systems), Ki67 (1 + 200, MIB-1, Dako), Lamp-2 (1 + 2000, H4B4, Developmental Studies Hybridoma Bank), Cathepsin B (1 + 200, polyclonal, Abgent) and Cathepsin D (1 + 750, EPR3057Y, Epitomics). Detection of immunohistochemical staining CD133 was detected by CSAII (Catalysed Signal Amplification II kit, Dako), nestin, Bmi-1 and sox 2 with Power Vision (Dako) and Ki-67, Lamp-2, cathepsin B and cathepsin D was performed by use of the detection system EnVision (Dako). The visualization was performed using diaminobenzidine as chromogen. Finally, the sections were counterstained with Haematoxylin and Eosin (H&E) and cover slips were mounted with Aquatex. Paraffin sections of tissue microarrays with 28 normal tissues and 12 cancers were used as positive control. Primary antibody omission was used as negative control.
Spheroid migration assay
Briefly, a reduced growth factor basement membrane matrix (Geltrex™, Life Technologies™, Denmark) solution was mixed with neurobasal medium, and added to each well in a 12-well plate. Subsequently, the supernatant was aspirated and the spheroids were placed individually (1 spheroid/well). Afterwards neurobasal medium was added to each well and cells were allowed to settle and migrate for 24 h, while being exposed to siramesine in increasing concentrations. The cells were monitored by light microscopy and imaging. The outer diameter of the migrating cells was measured using ImageJ software, relatively to spheroid diameter measured at day 0. Cell death in spheroids was visualized using PI (2 μM).
Preparation of organotypic brain slice cultures
Newborn Wistar rat pups (Taconic Europe, Denmark) used in the present study were treated according to the procedures at the Biomedical Laboratory, University of Southern Denmark.
Organotypic corticostriatal slice cultures were prepared as previous described [
13,
14]. Seven days after the start of culturing the brain slice cultures were exposed to siramesine and cell death was examined by PI uptake as described earlier by Nørregaard et al. [
14].
Preparation of co-cultures
GSS culture spheroids were implanted in the area between cortex and striatum close to corpus callosum. The medium was changed to serum-free medium before implanting the spheroids. The spheroids (200–300 μm) were incubated in DiO solution (1 mM, Molecular Probes, Invitrogen) for 24 h, before implanting them in the brain slice cultures. PI (2 μM, Molecular Probes, Invitrogen) was added to the medium to monitor cell death in spheroids and brain slices. Before start of exposure (Day 0) confocal z-stacks with 20 μm steps were recorded after 1 h of incubation. Thereafter the z-stacks were superimposed into one image representing the entire spheroid and surrounding brain tissue. This procedure was repeated at day 3 and day 6, whereafter the co-cultures were fixed in 4% formalin and paraffin embedded.
As a control assay to ensure cell death in the spheroids, DiO stained spheroids derived from T78 and T86 were placed on the same type of membranes used for the co-cultures. Cell death was investigated in the spheroids exposed to 20 μM siramesine with PI uptake recording confocal z-stacks as for the brain slice cultures. Cell death in the tumor cells was quantified using Visiomorph software. A classifier identifying PI uptake as red staining, co-expression of PI and DiO as yellow staining and DiO as green staining was created. The data were illustrated as area of cell death (red + yellow staining) divided by total spheroid area (red + yellow + green staining).
Immunohistochemical studies of co-cultures
The fixed and paraffin embedded co-cultures were cut in three μm sections and immunhistochemical stained. Immunohistochemistry was performed as described earlier, using the antibodies CD56 (1 + 100, CO4-NCAM, Neomarkers) and Vimentin (1 + 200, EP20, Epitomics) to identify human tumor cells in the rat tissue. A panel of the stem cell markers CD133, Nestin and Podoplanin (1 + 100, D2-40, Dako) as well as the proliferation marker Ki-67 was furthermore used as described above. The staining of the stem cell markers were assessed in the implanted spheroids by semi-quantitative scoring (0, 1+, 2+ and 3+). A Ki-67 labeling index was measured using the software program Tissuemorph (Visiopharm, Hørsholm, Denmark).
Glioblastoma tumor xenografts
The experimental procedure was performed as previously described [
12]. Female Balb/c nude (BALBNU-F, Taconic) mice were anesthetized subcutaneously and placed in a stereotactic instrument. Through a burr hole a 2-μL suspension of 300,000 single cells was injected into the striatum. Mice (
n = 42) were implanted with the standard cell line U87 (
n = 22), and the patient-derived cell line T78 (
n = 20).
Siramesine was dissolved in 0.5% methylcellulose 15 (M7140, Sigma-Aldrich, Denmark) in 0.9% NaCl. Siramesine treatment (100 mg/kg) was administered orally using a stomach tube. Control animals received 0.5% methylcellulose 15 in 0.9% NaCl. The treatment included biweekly treatment, initiated 1 week after tumor implantation for U87 implanted mice and 2 weeks after implantation for T78 implanted mice. U87 mice were euthanized after 1 week of treatment, whereas T78 mice were euthanized after 6 weeks of treatment.
The mice were euthanized at the same time point. When symptoms were observed as described below in the first mice, all mice were sacrificed to be able to compare the volumes among groups. The brains were removed immediately after death and fixed in 4% formaldehyde for 48 h. Before paraffin embedding the brains were divided by coronal sections (1 mm). Subsequently, brain sections were cut on a microtome (3 μM), and stained with H&E as well as Vimentin immunohistochemical staining as described previously. The tumor volume was determined using the digital software NanoZoomer Digital Pathologi, NDP viewer (Hamamatsu).
Ethics
The official Danish ethical review board named the Regional Scientific Ethical Committee of the Region of Southern Demark approved the use of human glioma tissue (permission J. No. S-VF-20040102) in the current study. Written informant consent was obtained from all participants.
The use of animals for organotypic brain slice cultures was approved by The Animal Experiments Inspectorate in Denmark (permission J. No. 2008/561-1572). The rats (newborn wistar rats, Taconic Denmark, n = 60; 4–6 slice cultures were obtained per rat) were decapitated and the brains were removed.
The use of animals for glioblastoma mice xenografts were approved by The Animal Experiments Inspectorate in Denmark (permission J. Nr. 2013-15-2934-00973). Mice (Female Balb c nu/nu mice 7–8 weeks, Taconic Denmark, n = 42) were anesthetized by a subcutaneous injection with a mixture of hypnorm and dormicum (0.12 ml/10 g). The mice were euthanized in a carbon dioxide chamber upon symptoms such as weight loss (20% loss of body weight) and general poor state including lethargy, hunched posture and failure to groom. The animals were housed according to national guidelines (National declaration for animal experiments 2013), and had free access to food and water.
Statistics
Data following a Gaussian distribution was analyzed using one-way ANOVA with Dunnett’s post test to compare treated cultures with control cultures. Non-parametric data was analyzed using Kruskal-Wallis with Dunn’s post test to compare the difference in the sum of ranks between two columns. Statistical significance was defined as *P < 0.05, **P < 0.01, ***P < 0.001. EC50 values were estimated by nonlinear regression. The Pearson correlation was calculated to quantify the association between the two variables, WST-1 and LDH. Tumor volume was compared using unpaired t-test. All statistics were carried out using Graph Pad Prism 5.0 (Graphpad Software, San Diego California USA).
Acknowledgements
We would like to thank Helle Wohlleben and Tanja Dreehsen Højgaard for assistance with the immunohistochemical stainings.