Extracellular vesicle-mediated transfer of CLIC1 protein is a novel mechanism for the regulation of glioblastoma growth.

Little progresses have been made in the treatment of glioblastoma (GBM), the most aggressive and lethal among brain tumors. Recently we have demonstrated that Chloride Intracellular Channel-1 (CLIC1) is overexpressed in GBM compared to normal tissues, with highest expression in patients with poor prognosis. Moreover, CLIC1-silencing in cancer stem cells (CSCs) isolated from human GBM patients negatively influences proliferative capacity and self-renewal properties in vitro and impairs the in vivo tumorigenic potential. Here we show that CLIC1 exists also as a circulating protein, secreted via extracellular vesicles (EVs) released by either cell lines or GBM-derived CSCs. Extracellular vesicles (EVs), comprising exosomes and microvesicles based on their composition and biophysical properties, have been shown to sustain tumor growth in a variety of model systems, including GBM. Interestingly, treatment of GBM cells with CLIC1-containing EVs stimulates cell growth both in vitro and in vivo in a CLIC1-dose dependent manner. EVs derived from CLIC1-overexpressing GBM cells are strong inducers of proliferation in vitro and tumor engraftment in vivo. These stimulations are significantly attenuated by treatment of GBM cells with EVs derived from CLIC1-silenced cells. However, CLIC1 modulation appears to have no direct role in EV structure, biogenesis and secretion. These findings reveal that, apart from the function of CLIC1 cellular reservoir, CLIC1 contained in EVs is a novel regulator of GBM growth.


Cell culture
This study was approved by the Ethical Committee for human experimentation of IEO (European Institute of Oncology) and all patients signed an approved consent document prior to surgery. Surgical specimens of tumors were collected at the Neurosurgery Dpt. at IRCCS Istituto Clinico Humanitas and examined by a neuropathologist to verify that each case met criteria for GBM and to select a tissue fragment with high content of viable tumor tissue. Each tissue specimen was dissociated into single cell suspension in warmed EBSS (Earle's Balanced Salt Solution) containing papain (2 mg/ml) (Worthington Biochemical), EDTA (0.8 mg/ml) and L-Cystein (0.8 mg/ml) at 37°C for 1-2 hours. The dissociated tumor was filtered through a 70 μm filter and washed a minimum of three times prior culturing. GBM-derived CSCs were maintained as neurosphere in DMEM-F12 1:1 medium (Dulbecco's Modified Eagle Medium -Ham's F12 Nutrient Mixture) medium (Invitrogen) supplemented with B27 Supplement (Invitrogen), EGF (20 ng/ml), b-FGF (10 ng/ml) (PeproTech) and 2 μg/ml Heparin (Sigma), at 37°C in a 5% CO 2 humidified incubator. GBM CSC cultures were passaged by mechanically dissociation when spheres reached approximately 300-500 microns in diameter, and cell counts were performed at the time of passage. Human GBM cell lines U87MG, A172, LN405, U118MG, T98G, DBTRG-05MG and U373 MG were purchased from American Type Culture Collection (ATCC) and maintained in DMEM (Lonza) supplemented with 2 mM glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin, and 10% FBS. All cell cultures were maintained at 37°C in a humidified 5% CO 2 incubator.

Lentivirus production and transduction
Recombinant lentiviruses were produced by transient transfection of 293T cells according to standard procedures. In brief, 293T cells were transfected with the vector, packaging plasmids, and a plasmid coding for the G protein of the vesicular stomatitis virus by the calcium-phosphate method. Virus was harvested 48 h later and concentrated by ultracentrifugation (68,000 × g). The multiplicity of infection (moi) was determined by infecting 293T cells, followed by flow cytometric quantification of eGFP-positive cells.
Lentiviral particles (multiplicity of infection = 1) targeting CLIC1 mRNA or bearing CLIC1-FLAG cDNA were added to the culture medium, along with 8 μg/mL polybrene. 48 hours after infection, GBM cells were incubated with 1 μg/mL puromycin for 3 days.

Electron microscopy
For routine electron microscopy (EM), purified EVs were fixed with 1% glutaraldehyde for 1 hour, washed, post-fixed with 1% reduced osmium tetroxide for 1 hour, washed, post-stained with 0,3% thiocarbohydrazide; refixed in the OsO4 and embedded into Epon. Ultrathin sections were placed on formvar-coated grids or slotgrids. Immune-EM analysis was performed as previously described (54, 55). Briefly, purified EVs were fixed with 1% glutaraldehyde and centrifuged. The pellet was embedded into gelatine and cryo-sections were prepared according to the standard procedure and cryo-sections were placed on slot-grid and labelled with antibodies against CLIC1 (rabbit polyclonal, 1:500, sc-134859 Santa Cruz, CA, USA) and CD63 (rabbit polyclonal, 1:1000, sc-15363, Santa Cruz, CA, USA) with subsequent labelling with protein A conjugated with 10 and 15 nm gold particles (UMC Utrecht, 1:60). Grids were observed at 200 kV with a Tecnai 20 electron microscope (FEI).

PKH26 labeling
Purified EVs were labeled with PKH26 (Sigma) according to manufacturer's instructions. Briefly, EVs (20 μg) were resuspended in 1 ml of diluent C, then mixed rapidly with a freshly prepared PKH26 solution in diluent C (final concentration during labeling step: 5 × 10 -6 M) and incubated for 3 minutes at room temperature. The labeling step was stopped by addition of an equal volume of 1% BSA for 1 minute. After three washes in PBS by ultracentrifugation, EVs were resuspended in PBS (100 ng/μl).

EVs uptake assay
2 × 10 5 U87 MG cells were harvested in a 6-well plate with 500 μl of complete medium. Labeled EVs (1 μg) were diluted in the growth medium for the indicated times. The uptake was stopped by washing cells in cold PBS, followed by fixation in 4% paraformaldehyde (PFA). Cells were oserved under a confocal microscopy, and the  Immuno-precipitation was carried out with 20 μg of anti-GFP antibody (sc-9996, Santa Cruz, CA, USA) on 2 mg of whole cell extract overnight at 4°C. Membranes were blotted anti-CLIC1.

Apoptosis detection
For apoptosis analysis, cells were first fixed in 1% formaldehyde for 20 minutes on ice, washed once in PBS and fixed again in ethanol 75% for 30 minutes on ice. Fixed cells were incubated in Propidium Iodide (2.5 μg/ml) and RNAse (250 μg/ml) for 12 hours at +4°C and analyzed by flow cytometry.

MS-based proteomics analysis
EVs sample pellets from NT, CLIC1 FLAG and siCLIC1 cells were directly resuspended in Laemmli buffer and loaded on SDS-PAGE on a gradient gel (4-12% Tris-HCl Precast Gel, Invitrogen) for protein separation. Gels were stained with Colloidal Coomassie. Enzymatic in-gel digestion was performed essentially as previously described; briefly, samples were subjected to reduction in 10 mM DTT for 1 hat 56°C, followed by alkylation with 55 mM iodoacetamide for 45 min at RT, in the dark. Digestion was carried out saturating the gel with 12.5 ng/ml sequencing grade-modified trypsin (Promega) in 50 mM ammonium bicarbonate. After one overnight, peptide mixtures were acidified with tri-fluoro acetic acid (TFA, final concentration 3%), extracted from gel slices with two rounds of washes (in 30% acetonitrile (ACN)/3% TFA and then in 100% ACN, respectively). Extracted peptides were subsequently loaded onto homemade C18-Stage Tips, for concentration and desalting prior to LC-MS/MS analysis.

Liquid chromatography and tandem mass spectrometry (LC-MS/MS)
Peptide mixtures were separated by nanoliquid chromatography using an EASY-nLC system (ProxeonBiosystems, Odense, Denmark) connected to the hybrid dual-pressure linear ion trap/orbitrap mass spectrometer (LTQ OrbitrapVelos, Thermo Scientific). The nano-liter flow LC was operated in one column set-up with a 15 cm analytical column (75 μm inner diameter, 350 μm outer diameter) packed with C18 resin (ReproSil, Pur C18AQ 3 μm, Dr.Maisch, Germany). Solvent A was 0.1% FA and 5% ACN in ddH2O and solvent B was 95% ACN with 0.1% FA. Separation was performed with a gradient of 0-40% solvent B over 90 minutes, followed by a gradient of 40-60% for 10 minutes and 60-80% over 5 minutes at a flow rate of 250 nl/min. The LTQ OrbitrapVelos MS was used in the data-dependent mode. CID-fragmentation method when acquiring MS/MS spectra consisted of an orbitrap full MS scan followed by up to 10 LTQ MS/MS experiments (TOP10) on the most abundant ions detected in the full MS scan. Essential MS settings were as follows: full MS (AGC 1,000,000; resolution 30,000; m/z range 300-1500; maximum ion time 500 ms); MS/MS (AGC 30,000; maximum ion time 100 ms; minimum signal threshold 500; isolation width 2 Da; dynamic exclusion time setting 30 s (±10 ppm relative to the precursor ion m/z); singly charged ions and ions for which no charge state could be determined were excluded from selection. Normalized collision energy was set to 35%, and activation time to 10 ms; spray voltage, 2.2 kV; no sheath and auxiliary gas flow; heated capillary temperature, 275ºC; predictive automatic gain control (pAGC) enabled, and an S-lens RF level of 65%. For all full-scan measurements with the Orbitrap detector, a lock mass ion from ambient air (m/z 445.120024) was used as an internal calibrant.

Protein identification by MaxQuant software and data analysis
The mass spectrometric raw data were analyzed with the MaxQuant software (version 1.3.0.5) (http://www. maxquant.org/downloads.htm). A false discovery rate (FDR) of 0.01 for proteins and peptides, and a minimum peptide length of 6 aminoacids, were required. In order to improve mass accuracy of the precursor ions, the timedependent recalibration algorithm of MaxQuant was used. The MS/MS spectra were searched by Andromeda engine against a human Uniprot sequence database (containing 86 725 protein sequences) combined with 262 common contaminants and concatenated with the reversed versions of all sequences. Enzyme specificity was set to Trypsin and maximum of two missed cleavages were allowed. Peptide identification was based on a search with an initial mass deviation of the precursor ion of up to 7 ppm. The fragment mass tolerance was set to 20 ppm on the m/z scale. Cysteine carbamidomethylation (Cys +57.021464 Da) was searched as fixed modification, whereas N-acetylation of protein (N-term, +42.010565 Da), oxidized Methionine (+15.994915 Da) were searched as variable modifications. Peptide and protein identification was performed automatically with MaxQuant using default settings for parameters. Briefly, peptide matches are assembled into protein groups. Posterior error probability (PEP) is calculated using Bayesian statistics as a probability of false hit using the peptide identification score and length of peptide. The protein group is assigned a PEP score by multiplying their peptide PEPs. Only peptides with distinct sequences and only highest scoring identified spectra were used. In the final list, the proteins identified were accepted only if they contain at least to two peptides, of which at least one unique (peptide > 1, unique > 0).

Gene ontology (GO) analysis
GO analysis was carried out by DAVID algorithm (http://david.abcc.ncifcrf.gov/) using the UNIPROT IDs obtained from the high confidence proteome identified for NT, CLIC1 FLAG and siCLIC1 EVs population. Particularly, the list of 2642 (NT), 2328 (CLIC1 FLAG) and 2310 (siCLIC1 EVs) were independently used as target list against a Human Uniprot sequence database (containing 86 725 protein sequences). GO terms were accepted as significantly enriched only when having a