A combination of r-ketorolac, bumetanide and niflumic

Abstract

The present invention relates to a pharmaceutical composition for use in the treatment of both low-grade and high-grade glioblastoma in a human subject. The composition comprises a synergistic combination of the following active agents: R-Ketorolac, Bumetanide, and Niflumic acid. This specific combination is designed to exert a synergistic therapeutic effect against glioblastoma by targeting multiple molecular pathways involved in tumor progression, inflammation, and ion transport. In a further aspect of the invention, there is provided a method for selecting a subpopulation of glioblastoma subjects who are responsive to the aforementioned treatment. This method comprises identifying genetic profiles predictive of a favorable therapeutic response to the combined administration of R-Ketorolac, Bumetanide, and Niflumic acid.

Description

[0001] P1732PC00

[0002] "A GLIOBLASTOMA TREATMENT AND A GENOMIC METHOD TO PREDICT THE EFFICACY OF THE SAME"

[0003] BACKGROUND

[0004] The replication of high-grade glioma, commonly referred to as glioblastoma (GBM), is a highly complex and fatal process. Both replication and cell cycle progression are regulated by molecular mechanisms and interactions within the nucleus, involving chromosomal duplication and modification. Decrease of GBM replication and prolongation of the life of patients with GBM have been recently achieved for subtypes of diffuse glioma, i.e., for those with a mutated isocitrate dehydrogenase 1 (IDH1). For this kind of glioma, a specific peptide vaccine has been developed with clear benefit for the patients (Flatten M, et al. A vaccine targeting mutant IDH1 in newly diagnosed glioma. Nature. 2021; 592: 463-468), demonstrating the efficacy of immuno-oncology approaches.

[0005] The malignancy of GBM is driven by its unusually high infiltration and proliferation rate. Cellular infiltration is closely associated with cell motility, and is mediated by volume changes and dynamic rearrangements of the cytoskeleton. As in most cell types, GBM replication proceeds through a series of cell cycle phases, culminating in cytokinesis, during which the cellular volume must approximately double.

[0006] When neurons and GBM are co-cultured in the same 2D or 3D environment, GBM cells exhibit significantly higher motility compared to healthy neurons or glial cells. Indeed, red fluorescent U87 glioblastoma cells, when co-cultured in a 3D environment made by carbon nanotube (CNT), have a markedly increased motility than that of neurons and glial cells, even in a short time frame (Xiao M, et al. A Fully 3D Interconnected Graphene-Carbon Nanotube Web Allows the Study of Glioma Infiltration in Bioengineered 3D Cortex-Like Networks. Adv Mater. 2018; 30:el806132).

[0007] GBM cells not only exhibit motility, but can also shrink and dynamically alter their shape and volume, enabling them to permeate through narrow pores with radii significantly smaller than their soma or even their nuclear radius. To quantify this infiltrative ability, the ability of U87 GBM to permeate through narrow pores has been evaluated by the trans-well assay and their motility by the scratch assay (Xu J, et al. Rael Promotes Cell Motility by Controlling Cell Mechanics in Human Glioblastoma. Cancers. 2020; 12:1667).

[0008] (R)-Ketorolac is a 5-benzoyl-2,3-dihydro-lH-pyrrolizine-l -carboxylic acid that has R configuration. Unlike the S-enantiomer, it does not exhibit COXI and COX2 inhibition but does exhibit analgesic activity. (R)-Ketorolac is used as a potent analgesic for the short-term management of post-operative pain, and in eye drops to relieve the ocular itching associated P1732PC00 with seasonal allergic conjunctivitis. R-Ketorolac inhibits Rael and Cdc42, two proteins playing a major role in the control of the dynamical properties of the cytoskeleton.

[0009] Bumetanide is a potent sulfamoyl anthranilic acid derivative belonging to the class of loop diuretics. In the brain, Bumetanide may prevent seizures in neonates by blocking the Bumetanide-sensitive Na / K / Cl cotransporter (NKCC1 and NKCC2) (Ward A, et al. Bumetanide. A review of its pharmacodynamic and pharmacokinetic properties and therapeutic use. Drugs. 1984; 28: 426-64), thereby inhibiting chloride uptake, thus decreasing the internal chloride concentration in neurons. Bumetanide may block the excitatory effect of GABA in neonates.

[0010] WO2023021175A1 discloses R-Ketorolac and Bumetanide as compounds that reduce infiltration and motility of U87 GBM cells.

[0011] Niflumic acid (NF A) is an aromatic carboxylic acid and a member of pyridines. It is widely used as an analgesic and anti-inflammatory agent in the treatment of rheumatoid arthritis. Niflumic acid targets primarily chloride channels (Shane Perrine, Niflumic Acid, xPharm: The Comprehensive Pharmacology Reference, Elsevier, 2007; 1-3).

[0012] These three drugs target distinct components involved in cellular motility and volume regulation.

[0013] However, there remains an unmet need for agents capable of reducing and inhibiting GBM replication not only in IDH1 mutant tumors but also in a broader spectrum of GBM cases, potentially encompassing the majority of GBM patients.

[0014] DESCRIPTION

[0015] It forms the first object of the present invention a composition comprising R- Ketorolac, Bumetanide and NFA for use in the treatment of low and high-grade glioblastoma.

[0016] In an embodiment, glioblastoma cells have low levels of the subunit LRRC8C of the VRAC channels, wherein VRAC channels are pentameric chloride channel composed by 5 different subunits: LRRC8A, LRRC8B, LRRC8C, LRRC8D and LRRC8E.

[0017] It forms a further object of the invention a method to select a population responding to the composition according to the present invention, wherein said method comprises evaluating the ratio of LRRC8C / LRRC8A in Glioma Stem Cells (GSCs) from a subject wherein, when said ratio is low, said subjects are the best responders to the proposed therapeutic approach. Low value of LRRC8C / LRRC8A means a ratio equal or lower than P1732PC00

[0018] In a further embodiment, it is here claimed a method to select a subset of patients, wherein said method comprises:

[0019] Evaluate the expression levels of at least five, or at least ten, or at least 19 of the genes selected in the group comprising: LHFPL, STXBP5L, KCNH4, RAMP2, FLG, IE1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SUEF1, ORS2, CADM1, LTBR, SPINT1, LRRK1, TGM2, MST4 in a subject; wherein, when the expression level of EHFPE, STXBP5E, KCNH4, RAMP2, is higher than the expression level of said gene in the averaged population value, said subject is classified as a Responders; wherein, when the expression level of FEG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SUEF1, ORS2, CADM1, ETBR, SPINT1, LRRK1, TGM2, MST4 is lower than the expression level of said gene in the average population, said subject is classified as a NonResponders.

[0020] In a further embodiment, it is here claimed a composition comprising R-Ketorolac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing LHFPL, STXBP5L, KCNH4, RAMP2 at higher levels with respect to the levels measured in the average population.

[0021] In a further embodiment, it is here claimed a composition comprising R-Ketorolac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing FLG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SULF1, ORS2, CADM1, LTBR, SPINT1, LRRK1, TGM2, MST4 at lower levels with respect to the levels measured in the average population.

[0022] In a further embodiment, it is here claimed a composition comprising R-Ketorolac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing LHFPL, STXBP5L, KCNH4, RAMP2 at higher levels with respect to the levels measured in the average population ad expressing FLG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SULF1, ORS2, CADM1, LTBR, SPINT1, LRRK1, TGM2, MST4 at lower levels with respect to the levels measured in the average population.

[0023] Surprisingly, the inventor demonstrated that the composition according to the present invention does not harm neurons and healthy glial cells, selectively targeting glioblastoma cells. P1732PC00

[0024] DRAWINGS DESCRIPTION

[0025] Figure 1: (A) Number of mitosis per cell / hour in U87 cell line in control conditions and in the presence of 100 pM Cocktail (COC). (B) Effect of the tested active agent alone and in combination. (C) Effect of 100 pM COC on the area covered by U87 cells in a dish. The total area decreased to about 30% of what was initially observed.

[0026] Figure 2: (A) Examples of fluorescent recordings at the indicated conditions. (B) Fraction of alive neurons measured in the above indicated conditions.

[0027] Figure 3: COC blocked replication in 5 out of 9 patients. The mean number of mitoses for different patients was normalized to 1. In GSCs from 4 patients replication continued to occur for 3-5 days after the addition of the COC. COC 100 pM, white bars. Control (no treatment), grey bars.

[0028] Figure 4: Intra and inter patients' variability of currents in GSCs. (A) electrical recordings from patient 1 showing large VRAC-like currents and from patient 5 where no VRAC-like currents are observed. (B) Scatter plot of the amplitude of the VRAC-like currents in 9 patients. (C) Current inhibition of the VRAC-like current by 100 pM NFA.

[0029] Figure 5: Genomic screening. (A) RNA-seq data for the different Na channels; (B) RNA-seq for the 5 subunits of the VRAC channel. Data from 5 patients and U87 cell line (black bars) and Human Astrocytes (HA, white bars).

[0030] Figure 6: Impact of the ratio LRRC8C / LRRC8A. (A) Dependence of the amplitude of the VRAC-like current on the ratio LRRC8C / LRRC8A. (B) Replication inhibition as induced by the composition described in the invention.

[0031] Figure 7: (A) Number of mitosis per cell / h after 3 days with the indicated treatment: 300 pM NFA, R-Ketorolac and Bumetanide and in the presence of different combination of 150 pM NFA, R-Ketorolac and Bumetanide. In these experiments we kept the total concentration of all drugs equal to 300 pM and therefore when we tested the combination of NFA / Bumetanide, NFA / R-Ketorolac and Bumetanide / R-Ketorolac the concentration of each drug was 150 pM. When we tested the effect of a single drug its concentration was 300 pM. (B) Representative images before and after treatment. After treatment, glioma cells become spherical and exhibit reduced motility compared to before treatment. This loss of motility induced by the treatment effectively blocks infiltration. Glioma cells in bright grey, inhibitory neurons in white and in bright field (here in grey) excitatory neurons and glial cells. (C) Quantification of glioma cell replication over time under control conditions and after treatment with 100 pM of the composition disclosed herein. P1732PC00

[0032] Figure 8: Variability of ion currents in GSCs across and within patients. (A, B, D, E) Currentvoltage (I-V) curves, histograms, and scatter plots of sodium currents in GSCs from patient 567, showing maximum current activation at -10 mV in some cells (B) and complete absence in others (C). (F) Representative traces of currents recorded in GSCs from patient 610 treated with 0 and 0.5 pM of internal calcium. (H-K) Evidence that chloride currents in GSCs are VRAC currents, activated by hypotonic solutions. (G, L) Inhibition of chloride currents by NFA.

[0033] Figure 9: Effects of the COC on neuronal electrical activity in vitro and in vivo. (A, B) Number of calcium events per minute and fraction of responding cells under control and treatment conditions. (C) Acute effect of COC on electrical activity in the occipital cortex (VI) of anesthetized mice. Recordings were performed using a Neuronexus probe, with local microinjection of the COC (0.3 mm anterior and 1 mm lateral to the electrode). (D) Representative local field potential (LFP) recording, including baseline (600 s) and 45 minutes post-cocktail injection. Dotted lines indicate the injection period. The magnified upper portion highlights US and DS events before and after treatment, while the spectrogram (bottom) depicts normalized data (Y-axis: frequency in Hz, X-axis: time in seconds). Dark grey denotes baseline activity, and the shades of grey (see Look Up Table nearby) represent the spectral power in dB. (E-G) Box plots showing frequency, duration, and power of US events. Each box includes the first and third quartiles, 5th and 95th percentiles, mean, and median. Dots represent median values for five animals. No significant differences were observed in US parameters before and after treatment (n=5, Mann-Whitney test, p>0.05). (H) Average power spectra of US events from five animals. Shaded regions indicate standard deviation. Two-way ANOVA revealed no significant differences between groups (p>0.05).

[0034] Figure 10: Mouse model and in vitro validation of composition. (A) Schematic representation of the plasmid carrying the red fluorescent protein mDsRed2. (B) Cell electroporation method used to transfect GL261 cells with the plasmid. (C) Mosaic two-photon microscopy image showing the tumor area (dark grey) with the injection site marked as a dark central region. Tumor growth formed an ellipsoid shape around the injection area (scale bar: 100 pm). Collagen fibers (pale grey) are visualized by second harmonic generation (SHG). (D) In vitro proliferation assay: cell proliferation measured over six consecutive days using the Burker chamber method. Error bars represent standard deviation (n=3 per condition). (E) In vitro migration assay: Quantified wound closure rate over time. Error bars represent standard deviation. COC treatment significantly slowed gap closure across all three cell lines (n=3 per condition), with statistical significance determined by the Mann-Whitney test.

[0035] Figure 11: Survival, tumor volume, and infiltration analysis. (A) Kaplan-Meier survival curve comparing mice treated with COC (treated) versus vehicle (control). Treatment was P1732PC00 delivered via osmotic minipumps for 7 days. After treatment, minipumps were removed, and mice were monitored until tumor-related death. (B) Tumor volume at day 17 posttreatment for control and treated groups. For treated survivors, tumor volume was measured at day 120. (C) Tumor infiltration analysis. Quantification of tumor border distance from the center of mass, showing greater infiltration in control tumors compared to treated tumors. Tumors in treated survivors were rounded and encapsulated.

[0036] Figure 12: In vivo cellular morphology analysis. (A) Quantification of cellular shape using the mean aspect ratio (cell width-to-height ratio) as a shape descriptor. (B) Quantification of average cellular size (pm2) based on the manually drawn regions of interest (ROIs) in ImageJ. Statistical analysis was performed using the Kolmogorov-Smirnov test.

[0037] Figure 13: Brain images before, during and after the treatment with COC.

[0038] Figure 14: (A) relation between the amplitude of the outward CF current at +100mV and the ratio of the level of expression of LRRC8C / 8A. Data from the RNA-seq of GSCs from that patient. The outward current seen in these electrophysiological experiments was identified as carried by CT ions in appropriate electrophysiological experiments. (B): relation between the amplitude of the outward CT current at +100mV and the ratio of the level of expression of LRRC8C / 8A. Data from the RNA-seq of GSCs from that patient. (C) Volcano plot of 22,524 genes detected by bulk RNA sequencing from seven patients, divided into cocktail- Responders and Non-responders groups. (D) expression level of gene KCNH4 for Responders and Non-responders GSCs. In this case, there is an excellent separation of the expression level. (E) Unsupervised clustering of the full data of RNA-seq from patient GSCs. Responder and Non- responder patients are well separated.

[0039] Figure 15: Time course of mitosis rate of U87 cell line untreated (Control) and in the presence of 50 pM COC and 50 pM COC + 10 pM Tamoxifen. Mitosis blockage induced by 50 pM COC (square) was increased in the presence of 10 pM Tamoxifen (triangle).

[0040] Figure 16: Blockage of replication of U87 cell line when the concentration of one of the drugs is changed by 50%. After 72 hours similar blockage is observed also when one of the drugs is increased by 50 % or decreased by 50 % . In these experiments replication was monitored by a metabolic test measuring cell viability.

[0041] Figure 17: Time course of mitosis rate of U87 cell line untreated (Control) and in the presence of different concentrations of the COC (100, 50 and 25 pM COC).

[0042] Figure 18: Relation between the amplitude of the Chloride current and the level of expression of the two membrane receptors TNC and CD44. When the expression of these two receptors is high, the CF current is low and the COC containing NFA does not block replication. Responder patients have a low expression of TNC and CD44. P1732PC00

[0043] Figure 19: Examples of GSCs showing large Ca2+-activated Ch currents (A) and only an inward current (B). The large inward current observed in the presence of 0.5 pM Ca2+inside the patch pipette is likely to be carried by the electrogenic Na+-Ca2+exchanger, as shown in the next figures. Whole cell recordings from GSCs from two distinct patients (S557 and S568). Voltage commands stepped by 20 mV from -100 to 100 mV. Holding potential at 0 mV.

[0044] Figure 20: The inward current occasionally observed immediately after membrane rupture with patch pipettes containing no Ca2+disappears in 2-4 minutes. Whole cell recordings as in Figure 19. The disappearance of the inward current shows that this current is carried by the Na+-Ca2+exchanger.

[0045] Figure 21: The inward current carried by the Na+-Ca2+exchanger is blocked quickly and completely by the specific NiCL and SN6 inhibitors and by 0.5 pM doxorubicin. Voltage clamp recordings in whole cell configuration. Voltage clamp varied from -100 to +100 mV. Recordings from GSCs obtained from two different patients.

[0046] Figure 22: Blockage of the inward current by Tamoxifen. Voltage clamp recordings in whole cell configuration. Voltage clamp varied from -100 to +100 mV. Electrical Recordings from GSCs as in Figure 21.

[0047] Figure 23: Blockage of replication in distinct GSCs from 9 patients. Data obtained from live cell imaging experiments after 3 days of incubation with 100 pM COC or 0.5 pM Doxorubicin.

[0048] Chloride channels play a major role in U87 cell replication by enabling an increase in cell volume, particularly during the final stage of mitosis — cytokinesis— when volume expansion is required for successful cell division. Li et al. 2024 (Li Y, et al. Mechanisms of glioblastoma replication: Ca2+flares and Cl- currents. Mol Cancer Res. 2024; doi: 10.1158 / 1541-7786.MCR-23-0934) describe the capability of NFA above 300 pM and of Carbenoxolone (CBX) above 200 pM to drastically reduce GBM replications, but these high concentrations of NFA and CBX damage neurons.

[0049] Surprisingly, here we firstly demonstrate that, by adding R-Ketorolac and Bumetanide to NFA, the obtained composition is capable of blocking GBM replication and infiltration, with negligible damage to neurons and healthy glial cells. The observed synergistic effect enhances efficacy without increasing toxicity, which is both unexpected and advantageous.

[0050] It forms the first object of the present invention a pharmaceutical composition for use in the treatment of high-grade and low-grade glioblastoma, wherein said composition comprises as active compounds R-Ketorolac, Bumetanide and Niflumic acid, wherein the P1732PC00 dose of each one of said actives during the combination treatment is equal to or inferior to 1000 pM at the injection site.

[0051] In an embodiment, the total amount of actives comprised in said composition is 300 pM.

[0052] In an embodiment, said composition comprises: R-Ketorolac, Bumetanide and Niflumic acid and at least one pharmaceutically acceptable carrier.

[0053] In an embodiment, said composition consists of R-Ketorolac, Bumetanide and Niflumic acid as actives and at least one pharmaceutically acceptable carrier.

[0054] In an embodiment, said composition comprises additional actives. As an example, said composition further comprise Tamoxifen. In an embodiment, Tamoxifen is comprised in said composition at 5 pM or at 10 pM.

[0055] In an embodiment, said actives are comprised in said composition at the following ratio, by weight: R-Ketorolac 1, Bumetanide between 0.3 and 1, Niflumic acid between 0.3 and 1. In an embodiment, said actives are comprised in said composition at the following ratio, by weight: R-Ketorolac 1, Bumetanide between 0.5 and 1, Niflumic acid between 0.5 and 1. In an embodiment, said ratio is R-Ketorolac 1 : Bumetanide 1: Niflumic acid 1.

[0056] It forms a further object of the present invention a method to select a responder population to the composition according to the present invention, wherein said method comprises evaluating the ratio of LRRC8C / LRRC8A in GSCs from a subject wherein, when said ratio is equal or lower to 2.8, said subject will respond better to the here proposed therapeutic approach.

[0057] In a further embodiment, it is here claimed a method to select a subset of patients, wherein said method comprises:

[0058] - Evaluate the expression levels of at least five, or at least ten, or at least 19 of the genes selected in the group comprising: LHFPL, STXBP5L, KCNH4, RAMP2, FLG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, EINC02984, SUEF1, ORS2, CADM1, ETBR, SPINT1, ERRK1, TGM2, MST4 in a subject; wherein, when the expression level of EHFPE, STXBP5E, KCNH4, RAMP2, is higher than the expression level of said gene in the averaged population value, said subject is classified as a Responder; wherein, when the expression level of FEG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, EINC02984, SUEF1, ORS2, CADM1, ETBR, SPINT1, LRRK1, TGM2, MST4 is lower than the P1732PC00 expression level of said gene in the average population, said subject is classified as a Nonresponder.

[0059] In a further embodiment, it is here claimed a composition comprising R-Ketoralac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing LHFPL, STXBP5L, KCNH4, RAMP2 at higher level with respect to the levels measured in the average population.

[0060] In a further embodiment, it is here claimed a composition comprising R-Ketorolac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing FLG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SULF1, ORS2, CADM1, LTBR, SPINT1, LRRK1, TGM2, MST4 at lower levels with respect to the levels measured in the average population.

[0061] In a further embodiment, it is here claimed a composition comprising R-Ketorolac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing LHFPL, STXBP5L, KCNH4, RAMP2 at higher levels with respect to the levels measured in the average population ad expressing FLG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SULF1, ORS2, CADM1, LTBR, SPINT1, LRRK1, TGM2, MST4 at lower levels with respect to the levels measured in the average population.

[0062] In vitro experiments showed that the composition effectively blocks replication in about 40% of patient-derived GSCs. For the in vitro experiments, U87 cell line was initially used as a suitable model of GBM, results were verified in patient-derived GSCs.

[0063] In vivo experiments using GL261 murine glioma cells showed a 30-50% reduction in tumor volume and decreased infiltration of glioma cells into healthy brain region. In addition, survival studies demonstrated that the average lifespan of treated mice was doubled and, in many cases, tripled, proving the efficacy of the COG. For these experiments, murine glioma GL261 cells were grafted into the brains of wild-type (WT) mice. This approach avoided the complexities associated with the use of immunocompromised mice.

[0064] Individually, the selected compounds exhibit limited toxicity but, in combination, they effectively hinder GBM replication and infiltration (FIG. 7, FIG. 8). Importantly, this drug combination induced only negligible changes to neuronal activity, both in vitro and in vivo, without significantly altering the firing properties of neurons (FIG. 9). P1732PC00

[0065] The drug combination comprises actives that do not easily cross the Blood Brain Barrier (BBB). Moreover, all of these compounds also affect major organs in the human body making localized delivery to the brain a desirable objective. Technologies are available to enable such targeted administration. As an example, lus et al (lus T, et al. Combining Clinical and Molecular Data to Predict the Benefits of Carmustine Wafers in Newly Diagnosed High- Grade Gliomas. Curr Treat Options Neurol. 2018; 20:3.) describe a localized delivery using wafers. Actives can be conjugated to nanocarrier (Alshawwa SZ, et al. Nanocarrier Drug Delivery Systems: Characterization, Limitations, Future Perspectives and Implementation of Artificial Intelligence. Pharmaceutics. 2022; 14:883). In the patient brain a cannula or micro-needles can be implanted (Mendez I, et al. Neural transplantarion cannula and microinjector system: experimental and clinical experience. Technical note. J Neurosurg. 2000; 92:493-9; Kawabori M, et al. Evaluation of Novel Stereotactic Cannula for Stem Cell Transplantation against Central Nervous System Disease. Stem Cells Int. 2020: 4085617. Erratum in: Stem Cells Int. 2021: 9796010). In an embodiment, the composition is delivered via Slow-Release Devices, utilizing post-surgical implants, such as wafers. In an embodiment, via nasal sprays.

[0066] Accordingly, the present invention further relates to a method for the treatment of glioblastoma, comprising the local delivery of the composition disclosed herein to target cells in the brain.

[0067] In an embodiment, said local delivery is obtained via nanocarriers, and / or cannulas, and / or Micro-Needles, and / or wafers and / or nasal sprays.

[0068] EXPERIMENTAL DATA

[0069] The drugs selected for testing were: Niflumic acid NF A, Bumetanide BUM, R- Ketorolac KET, Tamoxifen TMX.

[0070] The following combinations were tested, too:

[0071] - NFA 150 pM + BUM 150 pM

[0072] - NFA 150 pM + KET 150 pM

[0073] - BUM 150 pM + KET 150 pM

[0074] - NFA 50 pM + BUM 50 pM + KET 50 pM, named the Cocktail COC, 50 pM.

[0075] - NFA 100 pM + BUM 100 pM + KET 100 pM, named the Cocktail COC, 100 pM

[0076] - COC 50 pM + TMX 10 pM

[0077] The experimental models were: P1732PC00

[0078] - U87 cells

[0079] - GSC cells obtained from patients.

[0080] Example 1: effect on U87 GBM cells mitosis

[0081] The experiments were performed on 24 wells plates. Images were acquired every 8- 12 minutes, and four neighboring images were acquired either in bright field and / or fluorescence.

[0082] By using a combination of visual inspection and automatized analysis, the following parameters were acquired: i) the number of mitotic events occurring within a defined observation period; ii) the number of cells undergoing cell death during the same period; and iii) the number of viable cells at a given time point.

[0083] By using a 20X objective, the shape of the cells was observed and their morphology characterized before and after the treatment.

[0084] In some experiments, U87 GBM were transfected with mCherry to measure their total area and to monitor how the drugs affect the total area.

[0085] The results are shown in FIG. 1. In panel A and B, the number of mitoses per cell per hour is reported, after 2 days and 5 days in culture, respectively. The COG according to the present invention shows a clear synergic effect of the three actives. COG 100 pM blocked replication in U87 cell line and, after approximately 5 days, the rate of mitosis was about or less than 0.002 per cell per hour.

[0086] When tested at lower concentrations, the COG according to the present invention is still effective but the effect develops on a longer time scale, i.e. after a week of treatment.

[0087] Moreover, data collected from experiments using only fluorescent U87 cells — without the addition of neurons or glial cells— showed that after three days of exposure to the three- drug combination, the total area occupied by fluorescent U87 cells was reduced by 50-70%. Glioblastoma cells covered approximately 30% of the initial area (FIG. 1C). This result suggests that a significant reduction of the area / volume of GBM is obtained by perfusion with the COG with no surgery.

[0088] The experiment was repeated by using COG at the concentration of 25, 50 and 100 pM. This in vitro experiment confirmed the rate of mitosis is blocked 2-3 days after the addition of 100 pM COG. In the presence of 50 pM COG, blockage of replication is still P1732PC00 observed. When COC concentration is further reduced to 25 pM, blockage of replication is still observed but at a much slower rate (FIG. 17).

[0089] The COC was then tested on the same cellular model by varying the concentration of one of its components, i.e. keeping two of them at 100 pM and lowering the third at 50 pM, or increasing the third at 150 pM. Data shown in FIG. 16 are indicative of the fact that the synergistic activity is maintained.

[0090] In a further experiment, COC 50 pM is tested on U87 cells in the presence or in the absence of Tamoxifen, an estrogen receptor blocker, at 10 pM. The addition of Tamoxifen further increases replication blockage caused by 50 pM COC (FIG. 15). Similar results were obtained by adding doxorubicin.

[0091] Example 2: effect on patients GSC mitosis

[0092] The effect of COC was analyzed on GSC from patients. Results obtained are shown in FIG. 3. The effect is statistically significative on GSC obtained from 4 out of the 9 subjects analyzed. In 3 population, the COC was not effective, in 2 the COC had a partial effect. The data allow to conclude that COC is efficient in blocking replication in approximately half of patients

[0093] Example 3: Impact on healthy neurons and glial cells

[0094] Having seen that COC blocks replication in U87 cell line (FIG. 1) and in GSCs from a subgroup of patients (FIG. 3) the impact of the same on healthy neurons and glial cells was evaluated. m-Cherry U87 were co-cultured with neurons and glial cells obtained from healthy hippocampal tissue from GAD mice, wherein inhibitory neurons are engineered to express green fluorescence. The co-cultures were treated with the actives BUM, NF A and KET and with the COC 100 pM.

[0095] In this experimental setting, motility and replication of U87 GMB was evaluated, together with neurons and glial cells viability.

[0096] The COC blocked motion and replication of red U87 GBM but left almost unaltered glial cells and excitatory and inhibitory green neurons. This treatment - rather remarkably - also changed the shape of glioma, this change of shape is likely to be associated to a reduced infiltrative capacity (data not shown). P1732PC00

[0097] As an additional experiment, a series of neuronal cultures from rats / mice hippocampus / cortex was prepared and their electrical activity was evaluated by Calcium imaging.

[0098] In a first assay, hippocampal cultures from healthy rats in the absence of U87 cells and of GSCs from patients were tested. Neurons were loaded with the Calcium indicator Calcium Green or Oregon Green and the health of neurons and of their firing was established by looking to emitted fluorescence signals, where a burst of fluorescence indicated a burst of neuronal firing.

[0099] Cells were incubated for two days with concentrations of COC varying from 10 to 300 pM. As shown in FIG. 2A, control neuronal cultures have the usual Calcium transients typical of a healthy preparation: fast transients originating from neuronal spikes, usually a burst of them, and slower and larger transients produced by glial cells, which often could be distinguished by their larger soma. When 50 pM of COC was added, Calcium transients were almost indistinguishable from those seen in control cultures, but less frequent calcium transients were observed when COC concentration was increased to 100 pM. This data demonstrates COC, at concentrations suitable to exert effect on GBM, do not impair neuronal viability.

[0100] Notably, no calcium transients are observed in cultures exposed to 100 pM NF A alone. This is surprising, considering that when used at the same concentration in the COC according to the present invention, NFA does not affect neuronal viability.

[0101] Example 4: electrical properties of GSC from patients

[0102] The electrophysiological properties of GSCs from patients and the presence of Ca- activated Chloride channels have been evaluated.

[0103] Recordings from U87 cell line were rather reproducible. The same is not true when coming to GSCs from patients. Indeed, some electrical recordings from a specific patient showed a clear inward Na+current. However, this inward Na+current was not observed in all GSCs from the same patient, and the amplitude of this inward Na current was rather variable ranging from -270 to 0 pA. This inward Na+current was activated at about -35 / 30 mV, indicating a high and not a low threshold Na+current.

[0104] The Ca2+activated Ck currents / channels have been characterized by means of electrical recordings with patch clamp electrodes in whole cell configuration. In these experiments the ionic medium inside the patch pipette contained either 140 mM of CsCl and concentrations of Ca2+varying from 0 (no added Ca2+with ImM EGTA) to 1.5 pM. The high concentration of Cs+in the patch pipette was used to block K+currents. After sealing the P1732PC00 patch pipette on the cell membrane the seal was ruptured, allowing electrical recording in the presence of controlled amount of [Ca2+]i.

[0105] Chloride currents were present in approximately half of the tested GSCs (FIG. 4). This Chloride current was carried through VRAC channels. This VRAC like Chloride current was blocked by the addition of 300 pM NFA. Indeed, GSC from patients which have large Chloride currents blocked by NFA almost coincide with the GSCs which are blocked by the composition according to the invention (FIG. 6A).

[0106] When no Ca2+is present in the patch pipette, after going whole cell (in the presence of CsCl inside the pipette), the current / voltage relations are almost flat, see left column of FIG. 19A. However, when 0.5 mM Ca2+is added inside the patch pipette, profound differences are observed. In GSCs from some patients (such as patient S577) large currents - both inward and outward - are observed, but in other patients (such as patient S568) only a large inward current is seen (FIG. 19B).

[0107] Li et al. in Mechanisms of Glioblastoma Replication: Ca2+Flares and CF Currents. Mol Cancer Res. 2024, 22(9):852-863 describe that the outward current seen in the presence of high intracellular Ca - i.e. 0.5 pM - is carried by CF ions and is blocked by NFA. In addition, COC blocks replication in these GSCs.

[0108] Further support to the identification of the inward current seen in the GSCs of some patients - such as S568 - is obtained by the data shown in FIG. 20. Here the GSC was patched with a pipette containing no Ca2+ions, and immediately after obtaining a whole cell recording an inward current of some hundreds of pA was observed. However, the amplitude of this current declined irreversibly within 2-4 minutes, i.e. the time during which intracellular Ca was washed out. These experiments demonstrate that the inward current - observed in the presence of 0.5 pM Ca2+inside the path pipette - is indeed carried by the Na- Ca exchanger. This inward current is strongly identified as carried by the Na-Ca exchanger by the use of specific inhibitors, such as NiCF (Tepel M, et al. Activation of Na+, H+exchanger produces vasoconstriction of renal resistance vessels. Am J Hypertens. 1998; 11:1214-21) and SN6 (Ottolia M, et al. Na / Ca exchange and contraction of the heart. J Mol Cell Cardiol. 2013; 61:28-33) as shown in FIG. 21. Indeed, these compounds at micromolar amounts block very quickly - i.e., within seconds - this inward current. When these inhibitors are removed, the inward current comes back in 1-3 minutes.

[0109] Similar results were observed with the FDA approved drug Tamoxifen (V. Craig Jordan, Tamoxifen (1046,474) as a targeted therapy to treat and prevent breast cancer. British J. Pharmacol. 2006; 147: S269-S276), as shown in FIG. 22. P1732PC00

[0110] These results clearly show that Ca flares occurring in GBM during mitosis activate:

[0111] - an outward current through Ca-activated Cl' currents;

[0112] - an inward current carried by the Na-Ca exchanger.

[0113] The magnitude of these currents and their relative ratio depend on the biological characteristics of GSCs, and may vary between patients and potentially with disease progression. These currents contribute to the GBM cell swelling, which is required for the completion of the final stage of cytokinesis, thereby enabling the conclusion of mitosis. As a consequence:

[0114] - in the presence of large outward currents carried by CT, the COC according to the present invention blocks replication;

[0115] - in the absence of large outward CT currents and in the presence of inward currents associated to the Na-Ca exchanger, Doxorubicin and or Tamoxifen blocks GSC replication.

[0116] Example 5: Genomic screening of Chloride Channels in patient derived GSCs

[0117] Having observed that Ca-activated Chloride Channels are not present in all GSCs from patients, RNA-seq experiments were performed to determine the genomic pattern of expression of Chloride channels in patient GSCs.

[0118] Patients were divided into two groups: a first group, Responders, consisting of patients whose replication was blocked by the composition, and a second group, Nonresponders, consisting of patients whose replication was not blocked by the composition.

[0119] The expression level of voltage gated Chloride channel measured in the two groups was similar, as reported in FIG. 5AZ5B. The same applies to TMEM16A and TMEM16B Ca activated Chloride channels.

[0120] However, a significative difference was observed in the expression level of the subunits of the VRAC channels. The expression level of the major subunit LRRC8A was very similar in all tested GSCs from patients and in U87 and HA cell lines. Conversely, the LRRC8C subunit was present at lower levels in Responders compared to Non-responders.

[0121] It has been recently reported that a high level of expression of LRRC8C inhibits the activation of the VRAC channel by oxidation (Bertelli S, et al. J. Physiol. 2022;600: 3965-3982). The here presented findings demonstrate that the composition is active in Responders because the VRAC channel is operative.

[0122] Moreover, GSCs from patients with significant VRAC-like Chloride currents had a very low expression of the ratio LRRC8C / LRRC8A. GSCs patients with a high ratio of P1732PC00

[0123] LRRC8C / LRRC8A VRAC-like currents could not be detected or were very small. Data reported in FIG. 6B demonstrate that the efficacy of the composition inversely correlates with the ratio LRRC8C / LRRC8A.

[0124] Example 6: Synergistic effect

[0125] We utilized 24-well plates to monitor 24 preparations simultaneously within the same experiment. Some preparations served as biological controls, while others were biological samples treated differently.

[0126] Cell proliferation was analyzed by capturing images of cultured cells every 8-12 minutes. At each time point, four neighboring images were taken, either in bright field or in fluorescence. These cultures were observed for at least 3-5 days, often extending up to a week. Using a combination of visual inspection and automated analysis, we could estimate with good precision the following metrics within the imaging time window:

[0127] - The number of mitotic events.

[0128] - The number of cells undergoing apoptosis.

[0129] - The number of viable cells at defined time points.

[0130] By employing a 20x objective lens, we were able to evaluate cell morphology and characterize their shape before and after treatment, as well as throughout the treatment process.

[0131] In this experimental setting, a com position comprising R-Ketorolac, Bumetanide and Niflumic acid effectively blocks the final step of replication— when GBM cell volume must approximately double— without causing substantial damage to neurons. Notably, neurons neither replicate nor undergo significant volume changes, thereby ensuring that their functionality remains largely unaffected.

[0132] Under control conditions, the rate of mitosis was approximately 0.02 per cell per hour. However, this rate was drastically reduced to 0.002 per cell per hour by 300 mM NFA. Unfortunately, this high concentration also caused significant damage to neurons and healthy glial cells (FIG. 9). In contrast, a composition consisting of 100 pM NFA, 100 pM Bumetanide, and 100 pM R-Ketorolac reduced the mitosis rate to the same extent as 300 pM NFA but without the adverse effects.

[0133] This observed synergy required the simultaneous presence of all three drugs and was not replicated with any pairwise combinations of the drugs (FIG. 7). P1732PC00

[0134] In some experiments, we co-cultured m-Cherry U87 cells with neurons and glial cells from healthy hippocampal tissue derived from GAD mice, in which inhibitory neurons expressed green fluorescence (GFP). In these co-cultures, it was confirmed that the COG effectively inhibited both the motility and proliferation of red fluorescent U87 GBM cells with minimal impact on glial cells and on both excitatory and inhibitory green fluorescent neurons. Remarkably, this treatment also altered the shape of glioma cells. As will be discussed later, this change in shape is likely associated with reduced infiltrative capacity.

[0135] In some experiments, U87 GBM cells were transfected with m-Cherry, enabling precise measurement of their total area and allowing to monitor how the drugs reduced the total area of GBM cells in the dish. Collected data showed that, after three days of exposure to the composition according to the invention, the total area of GBM cells was reduced by 50-70%, leaving them covering only about 30% of the area occupied at the beginning of the treatment. It was confirmed that simultaneous exposure to all three drugs or sequential addition of the drugs produced the same final effect on proliferation and motility. A concentration of 100 pM of the COG effectively blocked proliferation in both U87 and GL261 cell lines. After approximately five days, the rate of mitosis in both cell lines was reduced to about 0.002 per cell per hour or lower. Lower concentrations of the composition were also effective, but their effects developed over a longer time frame, typically about a week.

[0136] The synergistic effect was then tested by reducing the concentration of the three drugs from 100 pM to 25 pM (FIG. 16). Under these conditions, the replication blockage effect of the low-dose COG (i.e., 25 pM) was delayed compared to the usual 100 pM concentration, but it was still present, and synergy occurred at later time points. We also investigated the presence of synergy when the concentration of one of the three drugs was either reduced or increased by 50% (FIG. 17). In all these conditions, replication blockage of U87 cell line was observed.

[0137] These experiments show that synergy is observed when:

[0138] • all three drugs are present simultaneously, but not when only two are used (FIG. 7);

[0139] • and their relative concentrations are varied up to ±50% (FIG. 16 and FIG. 17).

[0140] Synergy was also observed by adding a low dose of Tamoxifen, i.e., 5 or 10 pM, to the COG (FIG. 15). In this case, the blockage effect of 50 pM COG combined with 10 pM Tamoxifen is equivalent to that of observed with 100 pM COG alone.

[0141] Example 7: Electrical properties of GSCs from patients P1732PC00

[0142] We performed electrophysiological recordings on U87 cell line and several GSC lines obtained from patients. Recordings from U87 cell line were relatively consistent; however, as previously noted, GSCs from patients yielded heterogeneous results. For example, electrical recordings from a specific patient (S567) showed a clear inward sodium (Na+) current in some cells (not shown). However, this inward Na+current was not consistently observed across all GSCs from the same patient (FIG. 9A), and its amplitude varied considerably, ranging from -270 to 0 pA (FIG. 9B). The inward Na+current was activated at approximately -35 to -30 mV, indicating a high-threshold Na+current. These results highlight heterogeneity even among cells derived from the same patient. Notably, such a Na+current was never observed in over 100 U87 GBM recordings.

[0143] RNA-seq analysis of patient-derived GSCs, compared to U87 and HA cell lines, revealed higher transcription levels of the SNC3 gene. SNC3 encodes the Navl.3 channel, a high-voltage Na+channel activated at membrane potential around -30 mV. This activation threshold is significantly more depolarized than low-threshold Na+channels, such as Navi.6, which activate at membrane potentials near -55 mV. Consistent with this finding, electrophysiological recordings from GSCs confirmed the presence of high-threshold Na+currents. These findings suggest that at least some patient-derived GSCs express Navl.3 channels, potentially contributing to slight depolarization in GBM cells.

[0144] We also characterized calcium-activated chloride (Ca2+-activated Cl”) currents using patch-clamp recordings in the whole-cell configuration. The ionic medium inside the patch pipette contained 140 mM CsCl, with calcium concentrations ranging from 0 (no added Ca2+, 1 mM EGTA) to 1.5 pM. The high Cs+concentration in the patch pipette was used to block potassium (K+) currents. After achieving a seal between the patch pipette and the cell membrane, the seal was ruptured to enable electrical recordings under controlled intracellular calcium concentrations ([Ca2+]i).

[0145] Within 10 seconds of establishing whole-cell recording, stepping the voltage from 0 mV to between -100 and +100 mV revealed no Cl” or K+currents in the absence of Ca2+. However, when the same experiment was conducted with KC1 in the patch pipette, a clear outward K+current was observed. In contrast, with 1.5 pM Ca2+present in the CsCl-based patch pipette, we observed highly rectifying currents with a slow rise time. These currents resemble those observed when TMEM16A / B Cl” channels are expressed heterologously or constitutively in various cell types (REFs 16, 41-44). We termed this the "early current."

[0146] The amplitude of the early current declined within one minute, while a second current emerged. This "late current" exhibited a much faster rise time and closely resembled the activity of volume-regulated anion channels (VRACs) (right panel of FIG. 9D). Given the P1732PC00 highly transient nature of the early current, we focused on characterizing the late current in whole-cell recordings obtained at least 90 seconds after seal rupture. This late current was nearly undetectable at 0 intracellular Ca2+but was strongly activated when intracellular Ca2+levels were increased to 1.5 gM (FIG. 9E). At 0.5 gM intracellular Ca2+, the current exhibited outward rectification, but this rectification diminished as intracellular Ca2+levels increased. The late current was carried by anions, as confirmed by replacing most extracellular Cl” ions with the non-permeant anion gluconate (reducing Cl” from 150 mM to 15 mM). This substitution caused a reversal potential shift of approximately 55 mV, consistent with the theoretical value for a perfectly anion-selective channel. Additionally, replacing extracellular Na+had no effect on the current's amplitude or reversal potential. However, this chloride current was not observed in GSCs from all tested patients and was detected in approximately half of the tested GSCs (FIG. 9F). Further experiments confirmed that this chloride current was mediated through VRAC channels, as it was also activated by hypotonic solutions (FIG. 9G). The VRAC-like chloride current was blocked by the addition of 300 gM NFA (FIG. 9J-K). Notably, GSCs from patients with large chloride currents sensitive to NFA and CBX also coincide with those blocked by the composition according to the present invention (compare FIG. 9G and L).

[0147] Example 8: The effect of the cocktail on the electrical properties of neurons

[0148] Having observed that the composition according to the invention inhibits replication in U87 cell line and in 40% of GSCs derived from patients, we next investigated its effects on healthy neurons and glial cells. To this end, we prepared neuronal cultures from rat / mouse hippocampus and cortex and assessed their electrical activity using calcium imaging. Approximately 20 dishes of neuronal cultures were prepared, with three serving as controls. The remaining dishes were treated for two days with varying COC concentrations, ranging from 10 to 300 mM. As shown in FIG. 10, control neuronal cultures exhibit typical calcium transients characteristic of a healthy preparation. These included fast transients originating from neuronal spikes, often appearing as bursts, and slower, larger transients produced by glial cells, which could often be identified by their larger soma. When 50 gM of the composition of the invention were applied, calcium transients were nearly indistinguishable from those in control cultures. However, at 100 gM, calcium transients became less frequent. At 300 gM of the COC or NFA, no calcium transients were observed. Conversely, 100 gM CBX had only a minor effect on calcium transients. We also compared VRAC currents in glial cells and GSCs, finding that glial cells exhibited a lower density of Cl” currents, likely mediated by VRAC channels. In conclusion, concentrations of the COC below 50 gM did not significantly impact the electrical activity of neurons and had only a minor effect on glial cells. P1732PC00

[0149] To assess the effects of COC on cortical activity, we recorded local field potential (LFP) signals from the occipital cortex of anesthetized mice before and after treatment administration, using a multielectrode array (FIG. 10D). FIG. 10E shows representative LFP recordings, with magnified traces illustrating cortical activity before and after treatment.

[0150] Baseline activity was recorded for 10 minutes (highlighted by the rectangle), followed by the injection of the composition of the invention near the recording site using a Hamilton syringe. Cortical activity was subsequently recorded for 45 minutes. The LFP recordings were performed under urethane anesthesia, which induces slow -wave activity (SWA; 0.5-4 Hz), characteristic of non-REM deep sleep. This state enables the neocortex, due to its extensive recurrent connectivity, to sustain patterned network activity even in the absence of sensory stimulation. Under these conditions, the neocortex alternates between Up States (US), characterized by high neuronal firing, and Down States (DS), marked by network silence, reflecting the activity of pyramidal cells and inhibitory interneurons (Neske, G.T., et al., 2016 Distinct Roles of SOM and VIP Interneurons during Cortical Up States. Front Neural Circuits. 10:52).

[0151] The spectrogram of the full trace within the 0-60 Hz range is shown in FIG. 10E (bottom), with the window representing the baseline used for data normalization. FIGS. 4F- H show box plots of the duration, power, and frequency of USs measured prior to and following treatment. Although there was a slight reduction in the frequency, duration, and power of USs following drug injection, these changes were not statistically significant. SWA patterns persisted after treatment, indicating that cortical activity remained largely unaffected.

[0152] Additionally, analysis of the power spectral density of USs in the 4-400 Hz range revealed no significant differences before and after treatment (FIG. 101). These findings indicate that acute administration of the composition according to the invention does not significantly alter slow-wave cortical activity, indicating that the proposed treatment does not disrupt normal brain function during deep sleep states.

[0153] Example 9: In Vitro Effect of the composition according to the invention on the GL261 Murine Glioma Cell Line

[0154] To assess cellular proliferation, we seeded cells on a P60 Petri dish at an initial density of lxlO4cells. Cells were detached and counted daily over six consecutive days. FIG. 4D shows the proliferation rates of GL261-WT and GL261-B4 cells. The COC, administered in a 1:1:1 ratio with a final concentration of 100 pM, significantly reduced the proliferative potential of both cell lines. P1732PC00

[0155] Next, we evaluated the migration capacity of the cells using a wound-healing assay. A silicon culture insert with two chambers was placed on a P60 Petri dish, and cells were seeded at a concentration of 4xl05cells / mL in each chamber. After 24 hours, the insert was removed, and the distance between the two cell layers was measured daily over eight consecutive days until the gap closure was complete (FIG. 4E). The composition of the invention exhibited a similar inhibitory effect on the migratory potential of both cell lines. Under control conditions, both wild-type and transfected cells achieved gap closure within 48 hours. However, in the presence of the COC, migration was significantly delayed. Wildtype cells required 7 days (168 hours) to close the gap, while transfected cells took 8 days (192 hours).

[0156] These results demonstrate that the treatment effectively inhibits both cellular proliferation and migration in vitro.

[0157] Example 10: In Vivo Experimental Pipeline

[0158] FIG. 11 illustrates the in vivo experimental timeline designed to evaluate the efficacy of the composition of the invention on GBM progression and survival. The study began with the intracerebral injection of GL261-B4 cells into the occipital cortex of wild-type C57BL / 6 mice, followed by sequential phases of intervention and analysis.

[0159] On day 10, osmotic minipumps were implanted to deliver the COC continuously for a period of 7 days in animals designated for survival studies and tumor characterization. At the end of the treatment period (day 17), a subset of animals was sacrificed to assess tumor volume and cellular morphology.

[0160] In a parallel follow-up experiment, instead of sacrificing the animals on day 17, the minipumps were removed, and the animals were monitored for survival over an extended period. By day 120, animals that exhibited prolonged survival were sacrificed to further analyze tumor progression, volume, and cellular morphology.

[0161] This comprehensive experimental pipeline provides insights into both the short-term and long-term effects of the treatment, enabling a thorough evaluation of tumor progression and therapeutic efficacy.

[0162] Example 11: Drug Treatment Significantly Affects Survival, GBM Tumor Volume, and Tumor Cell Phenotype

[0163] FIG. 11 provides a comprehensive analysis of survival outcomes, tumor volume, and infiltration patterns in mice treated with the COC compared to control animals. The survival curve (FIG. 11A) demonstrates a significant extension in lifespan for treated mice, with P1732PC00

[0164] 63.6% surviving up to 120 days post-tumor implantation. These results strongly indicate the efficacy of the treatment.

[0165] To further investigate the impact of the drug combination, tumor volume and infiltration were examined. Tumor volume was quantified at two key time points: day 17 for both control and treated groups, and day 120 for long-term surviving treated mice (FIG. 11B). For in vivo tumor volume assessment, mice were perfused at the end of the treatment period, and their brains were fixed in a 4% paraformaldehyde solution. The brains were sectioned into 100 pm slices, imaged using confocal microscopy, and reconstructed using MosaicJ, an image analysis tool in ImageJ. Tumor regions of interest (ROIs) were manually delineated, and the total tumor volume was calculated by summing the volumes obtained from each slice.

[0166] Obtained data revealed a significant reduction in tumor volume in treated animals compared to controls. Specifically:

[0167] • Treated animals exhibited a 66.2% reduction in tumor volume.

[0168] • Long-term surviving treated mice showed an even more dramatic reduction of 99.4 % .

[0169] The mean tumor volumes were:

[0170] • 0.787 mm3in the control group,

[0171] • 0.266 mm3in the treated group, and

[0172] • 0.005 mm3in long-term survivors.

[0173] The observed decrease in tumor volume is likely due to a combination of reduced proliferation rates and inhibited glioma motility. This inhibition is primarily attributed to the drugs' effects on the glioma cells' ability to swell and shrink, a process critical for percolating through narrow pores and pathways to infiltrate healthy brain tissues.

[0174] To verify the role of reduced infiltration in decreasing tumor volume, we examined the distribution of glioma cells at the tumor edges (FIG. 11C). Long tails in the glioma density profile indicate infiltration, whereas sharper boundaries suggest a reduced ability to infiltrate surrounding tissue.

[0175] To compute the mean glioma profile at the tumor boundaries, we followed these steps:

[0176] 1. For each section, we identified the tumor center, marked by the cross in FIG. 11C P1732PC00

[0177] (right).

[0178] 2. From this center, we traced outward straight lines and measured the glioma density along these lines.

[0179] 3. This process was repeated for all traced lines and across all tumor sections.

[0180] 4. Finally, the density profiles were averaged across all directions, tumor sections, and groups of animals.

[0181] The resulting density profile for control mice displayed a long tail, as shown by the histogram in FIG. 11C (left), indicating significant infiltration into healthy tissue. In contrast, the density profile for treated mice lacked this long tail, as shown by the red histogram, suggesting that the treatment reduced glioma infiltration and effectively encapsulated the tumor.

[0182] These findings are further supported by representative images of tumor growth shown in FIG. 13, which provide a visual comparison of tumor volume and infiltration patterns under different treatment conditions.

[0183] Tumors of treated survivor mice had a remarkable feature which could explain their long survival: as shown in FIG. 13A tumor cells in survivor animals were clearly present indicating that xenotransplant was successful. However, in survival animals' tumor cells remained packed in a restricted region as if they were encapsulated so that they could not exit or escape.

[0184] Finally, we examined the impact of the treatment on glioma cellular morphology, as changes in cell.

[0185] To quantify these morphological changes, we measured two parameters:

[0186] 1. Aspect Ratio (AR): A metric describing cell shape (higher AR indicates elongated morphology).

[0187] 2. Cell Area: A measure of cell size (in pm2).

[0188] Cell outlines were manually drawn using ImageJ by two independent observers to reduce bias. The AR and cell area were then calculated for cells in control animals, treated animals, and long-term treated survivors.

[0189] The results observed are:

[0190] • Aspect Ratio (FIG. 12A): Control animals had a significantly higher AR compared to P1732PC00 treated and treated survivor groups, reflecting the elongated shape of untreated glioma cells. This indicates that the treatment effectively induced a more compact, circular morphology in glioma cells.

[0191] • Cell Area (FIG. 12B): Treated animals showed a significant reduction in tumor cell area compared to controls. However, in long-term treated survivors, the cell area was not significantly different from that of controls, although these cells retained their rounded morphology.

[0192] These findings suggest that the COC disrupts glioma cell structure, inducing a shift from an elongated to a compact shape. While this treatment does not necessarily reduce cell size in long-term survivors, the altered morphology likely contributes to the observed reduction in infiltration and malignancy by impairing the cells' invasive capacity.

[0193] Example 12: Genomic Analysis of sensitivity to the composition of GSCs from patients

[0194] As shown in FIG. 14A the amplitude of the VRAC or Chloride current is inversely proportional to the ratio of the level of expression of LRRC8C / LRRC8A obtained from RNA- seq data of GSCs from the patient. The presence of large chloride currents indicates sensitivity to the composition according to the invention, whereas the absence of such Cl' current suggests non-responders patients characterized by a high LRRC8C / LRRC8A ratio. Very similar results are obtained considering the ratio LRRC8B / LRRC8A (FIG. 14B), although with a lower correlation value of 0.44, against 0.62 when using the ratio LRRC8C / LRRC8A.

[0195] The determination of sensitivity to the composition of the invention based on the LRRC8C / LRRC8A and LRRC8B / LRRC8A ratio provides a useful indication, although with limited confidence. Therefore, we looked for a limited set of genes - in the order of 10-20 - which provide a stronger confidence of the sensitivity to our COC.

[0196] Therefore, we constructed the volcano plot of genes from responders and non- responders patients. A volcano plot is a graph that shows the relationship between the magnitude of change and statistical significance in data. It's a key tool in statistical and genomic analyses. In this way we identified 4 strongly up regulated genes in responders patients (shown in FIG. 14C).

[0197] As shown in FIG. 14D, the different expression level of these genes is very high. By using this set of identified genes the prediction of the sensitivity of the GSCs to the composition becomes very high, with a statistical significance larger than 99% .

[0198] The 19 identified genes are listed in Table 1: P1732PC00

[0199] Table 1

[0200] Then, the authors looked to the full RNA-seq data containing the level of expression of more than 20.000 genes. In this case, unsupervised clustering has been used, and two clusters were obtained, corresponding to Responders and Non-responders GSCs (Fig. 14E).

[0201] This analysis provides a better characterization of the genomic profile of patients GSC. The genomic signature of GSCs from patients exhibits a high heterogeneity. High heterogeneity means a high propensity to mutate which has a high clinical relevance.

[0202] The analysis of the genes encoding for the five subunits of the VRAC channel LRRC8A, 8B, 8C, 8D and 8E identifies Responders and Non-responders GSCs, but not with a high statistical significance, because the p-value is not very low.

[0203] When considering the Volcano plot, the determination of Responders and Non- responders GSCs has a statistical significance higher than 99% .

[0204] Example 13: additional Genomic Analysis of sensitivity to the composition of GSCs from patients

[0205] Responders / Non-responders GSCs can be identified also by looking to the level of expression of other membrane proteins, such as CD44 (Senbanjo L & Chellaiah M. 2017 P1732PC00

[0206] CD44: A Multifunctional Cell Surface Adhesion Receptor Is a Regulator of Progression and Metastasis of Cancer Cells. Front. Cell Dev. Biol., Sec. Cell Adhesion and Migration Volume 5 and TNC (Abedsaeidi M, et al. Biology of

[0207] Tenascin C and its Role in Physiology and Pathology. Curr Med Chem. 2024; 31:2706-2731). These two membrane proteins are not in any way related to Chloride channels and to the subunits of the VRAC channels. The protein encoded by the gene CD44 is a cell-surface glycoprotein involved in cell-cell interactions, cell adhesion and migration. It is a receptor for hyaluronic acid (HA) and can also interact with other ligands, such as osteopontin. The protein TNC is an extracellular matrix protein with a spatially and temporally restricted tissue distribution. This protein is homohexameric with disulfide-linked subunits and contains multiple EGF-like and fibronectin type-III domains, collagens, and matrix metalloproteinases (MMPs).

[0208] Data are provided in FIG. 18.

[0209] Example 14: Blockage of replication of GSCs Non-responders to the Cocktail

[0210] The here described COC blocks replication of GSCs in about 40% of patients, as shown in FIG. 23. The COC does not show activity in about 60 % of patients, referred as non- responders patients. Data shown in previous experiments show that non-responders GSCs have significant inward currents carried by the Na+-Ca2+exchanger. The authors of the present invention surprisingly demonstrated that GSCs non-responders to the COC are blocked by Doxorubicin.

[0211] From the here reported experiment, it is therefore possible to conclude that:

[0212] - COC blocks replication in 40% of patients, when swelling is primarily caused by Ca2+-sensitive CF channels;

[0213] Doxorubicin blocks replication in 30% of patients who do not respond to the COC, when swelling is primarily caused by the action of the Na+-Ca2+exchanger.

[0214] Accordingly, when both currents are significant, treatment with the COC together with Doxorubicin and / or Tamoxifen is recommended (FIG. 15).

[0215] Methods

[0216] Cell culture

[0217] U87 GBM cells (#89081402, Sigma-Aldrich) were cultured in DMEM supplemented with 10% fetal bovine serum (FBS; Invitrogen, Life Technologies, Gaithersburg, MD), 1% PenStrep (100 U / ml penicillin and 100 pg / ml streptomycin; Invitrogen). P1732PC00

[0218] Human GBM samples were collected by the Neurosurgery Department of the Azienda Ospedaliera Universitaria di Udine, after informed consent was obtained, in accordance with the Declaration of Helsinki, and with approval by the Independent Ethics Committee of the University Hospital of Udine (Approval 196 / 2014Em) and human GSCs were expanded in adherent culture following the protocol optimized by Dirk's group to maintain undifferentiated GSCs in adherent condition [17, 33, 34], Briefly, tissue samples were mechanically and enzymatically dissociated and single-cell suspensions were cultured with adhesion on laminin-coated dishes in a growing medium composed of the following: Neurobasal™-A Medium (Gibco by Invitrogen) supplemented with 2 mM L-glutamine (Sigma- Aldrich), lx N2 supplement [7.5% bovine serum albumin (BSA), 0.63 pg / ml progesterone, 1.6 mg / ml putrescine dihydrochloride, 0.52 pg / ml sodium selenite], 25 pg / ml insulin (Sigma- Aldrich), lx penicillin-streptomycin (Gibco by Invitrogen), 100 pg / ml h-apo-transferrin (Sigma- Aldrich), lx B-27 supplement (Gibco by Invitrogen), 20 ng / ml h- FGF-basic (Peprotech), 20 ng / ml h-EGF (Peprotech).

[0219] U87 cells were cultured in an incubator at 37°C, 5% CO2, 95% relative humidity, while GSCs in an atmosphere of 1.5% O2, 5% CO2, balanced nitrogen and 95% relative humidity.

[0220] Transfection

[0221] Addgene mQY-Nl and pGP-CMV-NES-jRCaMPlb plasmids were used. Twenty -four hours after plating, U87 GBM cells were transfected with the indicated plasmids with Eipofectamine 3000® transfection reagent (Invitrogen), following the manufacturer's protocol. The empty vector was used as a control.

[0222] Live-cell imaging

[0223] Cells were plated at a density of 8.0xl04cells into 35-mm dishes with a glass bottom and cultured for 2 days. Five-cell imaging experiments were performed on an epi-fluorescence microscope (Olympus IX-83, Olympus) equipped with a chamber incubator (Okolab, Pozzuoli, Italy) and light-emitting diode (LED) illumination (X=365 nm for MQAE; X=480 nm for OGB-1; X=500 nm for mQY; X=560 nm for jRCaMPlb,). During all imaging experiments, cells were kept at 37°C, 5 % CO2 and 95% humidity. Time-lapse images were taken with 300 - 500 ms of exposure time. Images were acquired with a charge-coupled device (CCD) sensor at 12-bit depth (ORCA-D2, Hamamatsu) using a 20X air objective (Olympus, NA=0.75) or a 40x oil objective (Olympus, NAM.3) with a spatial resolution of 1280x960 pixels.

[0224] Ca2+imaging

[0225] In Ca2+-imaging experiments, 2.0xl04cells were plated on a flat coverslip and cultured for 1-6 days, and subsequently were loaded with the membrane-permeable Ca2+dye Fluo-4 AM P1732PC00 or Oregon Green 488 BAPTA, AM (Life Technologies) by incubation with 4 pM Fluo-4 or OGB-1 (dissolved in anhydrous DM SO, 4 mM stock solution) in DMSO (Life Technologies) at a ratio of 1:1 in Ringer's solution (145 mM NaCl, 3 mM KC1, 1.5 mM CaCL, 1 mM MgCE, 10 mM glucose and 10 mM Hepes, pH 7.4) at 37°C for 30 min. After incubation, the cultures were washed with Ringer's solution for 20 min and then transferred to the stage of a Nikon Eclipse Ti-U inverted microscope equipped with a piezoelectric table (Nano-ZI Series 500 pm range, Mad City Labs), an HBO 103 W / 2 mercury short arc lamp (Osram, Munich, Germany), a mirror unit (465-495 nm excitation bandpass filter, 505 nm dichroic, 515-555 nm emission bandpass filter) and an Electron Multiplier CCD Camera C9100-13 (Hamamatsu Photonics, Japan). Ca2+-imaging recordings were performed at room temperature and images were acquired using the NIS Element software (Nikon, Japan) with an S-Fluor 20x / 0.75 NA objective, at a sampling rate of 3-10 Hz, with a spatial resolution of 256x256 pixels for 15 min. To avoid saturation of the signals, excitation light intensity was attenuated by ND4 and ND8 neutral density filters (Nikon, Tokyo, Japan).

[0226] For ratiometric Ca2+-imaging experiments, cells were loaded with 1.5 pM Fluo-4 or OGB-1 AM [dissolved in anhydrous DMSO (Sigma- Aldrich), 1.5 mM stock solution], 2.5 pM Fura Red AM [dissolved in anhydrous DMSO (Sigma- Aldrich), 2.5 mM stock solution] in DMSO (Life Technologies) in 1 ml Ringer's solution for 30 min. After incubation, the cultures were washed with Ringer's solution for 20 min and then transferred to the stage of an epifluorescence microscope (Olympus IX-83, Olympus) equipped with a chamber incubator (Okolab, Pozzuoli, Italy), LED illumination (X=490 nm for both Fluo-4 / OGB-l and Fura Red AM) and a CCD camera (ORCA-D2, Hamamatsu) with a dual sensor to record the fluorescence images from OGB-1 (520±17.5 nm) and Fura Red AM (640±37.5 nm) simultaneously. During all imaging experiments, cells were kept at 37°C, 5 %CO2 and 95 % humidity. Time-lapse images were taken with 300 ms of exposure time. All acquisitions were operated with a 20X air objective (Olympus, NA=0.75) or a 40x oil objective (Olympus, NAM.3).

[0227] Immunofluorescence

[0228] Cells were grown on coverslips for 2 days and subsequently washed with ice-cold PBS, then fixed with 4 % paraformaldehyde for 10 min at room temperature, followed by permeabilization with PBS plus 0.1 % Triton X-100, blocked with 3 % BSA and incubated overnight with primary antibody anti-TMEM 16 A, TMEM 16B, LRRC and Piezol, which are all rabbit monoclonal (1:500). The cells were then washed with PBS three times for 5 min each and incubated with Alexa Fluor 488-labelled goat anti-mouse secondary antibody (1:1000; Life Technologies) at room temperature for 1 h. The cells were examined with a confocal microscope (Nikon AIR).

[0229] Optical manipulation P1732PC00

[0230] To mechanically stimulate the cell, a custom-built optical tweezer combined with an inverted microscope was used. Briefly, a polystyrene bead with a 3.5-pm diameter (G. Kisker Biotech Germany) optically trapped in a controllable oscillatory optical trap (OOT) was used, with an infrared (IR) laser at 1064 nm for trapping (IPG Laser Germany). The axial position of the trap could be adjusted within a range of 0 - 20 pm using a custom designed two-lens system composed of a focus tunable lens (EL-10-30-NIR-LD, Optotune AG, Switzerland), of which the focal length can be varied by computer control (Lab VIEW code).

[0231] Data and statistical analysis

[0232] For calcium and chloride experiments, the DF / F values were obtained through custom MATLAB (The MathWorks, Natick, MA, USA) code and the ImageJ software vl.6 (National Institutes of Health, Bethesda, MD, USA). All results are presented as mean ± SD, and significant differences were determined using a t test with p < 0.05 (GraphPad Prism 7, GraphPad soft-ware, San Diego, GA, USA). In vivo: Z-series were analyzed as Z-stacks using ImageJ (NIH) program. Tumor mosaics were reconstructed either using the tool MosaicJ or automatic tiles construction using Zeiss software. ROIs (region of interest) were manually drawn using ImageJ to calculate tumor area and volume in reconstructed mosaics as well as to evaluate cell morphology in high-resolution images. Tumor volume was calculated by adding up the tumor area present in each brain slice. Data was entered into the Origin Pro 9.0 application to generate graphs. When appropriate, box charts have been used to represent data. Boxes represent the 25th and 75th percentiles, whereas whiskers represent the 5th and 95th percentiles. The non-parametric Mann- Whitney and Kolmogorov-Smirnov tests, were used to analyze statistical significance.

[0233] GL261 cell line

[0234] Wild-type GL261 (GL261-WT) cells were stably transfected with a plasmid carrying a red fluorescent protein (RFP; mDsRed2) under the constitutive promoter GAG, flanked by piggyBac ITRs, which allows its insertion in a target cell genome in the presence of the helper plasmid, piggyBac transposase (PBase). Transfections were performed through electroporation on cell suspension and later sorted using Bio-Rad S3e Cell Sorter to achieve 100 %. Transfected cells (GL261-B4) were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco) containing 10 % fetal bovine serum, 1% penicillin-streptomycin (100 units / mF of penicillin and 100 pg / mL of streptomycin, Gibco), 2 mM L-glutamine (Gibco), 1 mM sodium pyruvate (Gibco) and lOmM HEPES (Gibco) and maintained, in a humidified atmosphere of 5 % CO2 at 37 °C. For injection, GL261-B4 were grown up to 90% confluency in a P60 Petri dish and harvested using 0.05% Trypsin-EDTA. Cells were then pelleted by centrifugation at 1200 rpm for 5 minutes and re-suspended to a cell concentration of 20.000 cells / pL with Dulbecco's Phosphate-Buffered Saline (PBS). P1732PC00

[0235] Animals

[0236] The mice strain used was a C57BL6 / J line, obtained from The Jackson Laboratory. All animal care and experimental procedures were performed in strict accordance with the recommendations of the Italian Ministry of Health.

[0237] Mice were kept on a 12h light-cycle and fed ad libitum. For all experiments, mice of either sex were studied unless stated otherwise.

[0238] Tumor implantation

[0239] To obtain a GBM-grafted mouse, the animal was anesthetized with Avertin for single anesthesia doses used in brief surgeries (2,2,2-tribromoethanol; 250 mg / kg IP) or with zoletil / xylazine for chronic experiments (50 mg / kg zoletil+10 mg / kg xilaxine; dose: 0.1 ml / 10 g for each mouse plus eventual adding). The mouse was fixed on a stereotaxic apparatus where two metal ear bars and a nose clamp kept the head in position. A skin incision was made between the ears using fine scissors and the exposed subcutaneous tissue was delicately removed. Using a hand drill, a small hole was made in the skull around visual area VI (3 mm posterior to bregma; 1 mm lateral to the midline) and the injection area was washed with a small amount of PBS. Next, 40,000 GL261-B4 cells suspended in 2 pL of PBS were slowly infused at a velocity of 500 nL / min for 5 minutes with a Hamilton syringe guided by an automatized pump (Legato 130; Phymep) at a depth of 400 pm from the pial surface. After injection, the incision between the ears was sutured with stitches and antiseptic cream was applied to avoid infection. A single dose of antipyretic, i.e., 5 drops of paracetamol (100 mg / mL) per 100 mL of drinking water was provided for the animals after surgery. Mice were closely monitored for behavior, reactivity, and appearance.

[0240] Intracerebral drug application via osmotic minipumps

[0241] Since the COG do not pass the blood-brain barrier effectively, subcutaneous osmotic minipumps (ALZETOOsmotic Pumps, Model 1007D) were used to administer the treatment. These minipumps are linked to a brain infusion kit (ALZET Brain Infusion Kit 3, l-3mm) (FIG. 9). A day before implantation, osmotic minipumps were filled with COG (ratio 1:1:1; 100 pM) in ACSF for the treated group and ACSF+ 1% DMSO for the control group. The prepared minipumps allow for sustained delivery over 7 days. They were immersed in sterile PBS and left to prime overnight at 37°C in a sterile incubator to be activated. Minipumps were then implanted 10 days post tumor implantation. Mice were anesthetized with Avertine and fixed on a stereotaxic apparatus as previously described. A subcutaneous sack in the back of the animal was created to place the minipump, the brain infusion kit was then placed in the hole where injection was made by the experimenter. Dental cement (Paladur dental) was used to secure the rest of the cannula in order to avoid its detachment during the treatment (infusion rate of 0.5 pL / h). P1732PC00

[0242] Brain fixation and slices preparation

[0243] To collect brains, mice were anesthetized deeply with 20% urethane (ethyl carbamate, Sigma Aldrich) and perfused transcardially with PBS lx followed by fixative paraformaldehyde (PFA; 4). Brains were gently collected and post-fixed with 4% PFA for 1 day and then placed in 30% sucrose solution (in PBS) for cry opreservation. Serial sections of 60 pm were prepared with a freezing microtome (Leica) and mounted on glass slides. To preserve fluorescence, Vecta shield (VectorLab) was added on brain slices before securing them with a cover glass.

[0244] Two photon microscopy

[0245] Slices were imaged at high resolution with a Prairie Ultima Multiphoton microscope (Bruker) equipped with a mode-locked Ti:Sapphire laser (Chameleon Ultra II, Coherent) through a 20X Olympus XLUMPLFLN water immersion objective (numerical aperture 1.0). The power at the sample surface during acquisitions was kept at 25 mW, while the laser was set on an excitation wavelength of 1040 nm unless otherwise mentioned. Z-series for Tumor volume reconstruction was done by acquiring Z-stacks at a resolution of 512x512 pixels, zoom 1, leading to a field of about 608.3 pm and a linear resolution of 0.922pm per pixel.

[0246] Confocal microscopy

[0247] Images on brain slices were acquired using Zeiss LSM-900 Airyscan 2 fluorescence microscope equipped with diode laser (488nm) through a 10 or 20x Plan- Apochromat (Thorlabs) air objective (numerical aperture 0.3). The detection system of this microscope is the GaAsP (Gallium Arsenide Phosphide) photomultiplier, and the microscope is interfaced with an Axiocam CCD camera. Z-series for tumor volume reconstruction was done at a resolution of 512x512 pixels, zoom 1, leading to a field of about 319.45 pm and a linear resolution of 0.624 pm per pixel, unless otherwise stated.

Claims

P1732PC00CLAIMS1. A pharmaceutical composition for use in the treatment of low and high-grade glioblastoma in a human subject, comprising R-Ketorolac, Bumetanide and Niflumic acid.

2. The pharmaceutical composition for use according to claim 1, wherein the dose of each one of said R-Ketorolac, Bumetanide and Niflumic acid is equal to or inferior to 1000 pM at the injection site.

3. The pharmaceutical composition for use according to claim 1, wherein the total amounts of the active compounds in said composition is equal or lower than 300 pM.

4. The pharmaceutical composition for use according to claim 1, comprising at least one pharmaceutical acceptable carrier.

5. The pharmaceutical composition for use according to claim 1, wherein said active compounds are comprised in said composition at the following weight ratio: R-Ketorolac between 0,3 and 1, Bumetanide between 0.3 and 1, and Niflumic acid 1.

6. The pharmaceutical composition for use according to claim 1, wherein said active compounds are comprised in said composition at the following weight ratio: R-Ketorolac between 0,5 and 1, Bumetanide between 0.5 and 1, and Niflumic acid 1.

7. The pharmaceutical composition for use according to one of the claims 1-6, comprising an additional active compound, preferably comprising Tamoxifen and / or Doxorubicin.

8. The pharmaceutical composition for use according to one of the claims 1-6, further comprising Tamoxifen in an amount in the range 5 - 10 pM.

9. The pharmaceutical composition for use according to one of the claims 1-8, wherein said composition is delivered locally to target cells within the brain.

10. The pharmaceutical composition for use according to claim 9, wherein said local delivery is obtained via nanocarriers, and / or cannulas, and / or Micro-Needles.

11. A method to select, among a population of glioblastoma affected subjects, a subpopulation of therapy responders, said method comprising:- Measuring the expression level of the LRRC8A VRAC channel subunity in GSCs from a biopsy of the tumor of a given subject;P1732PC00- Measuring, in the said GSCs, the expression level of the LRRC8C VRAC channel subunity;- Calculating the ratio of LRRC8A / LRRC8C expression level, wherein, when said ratio is equal or higher to 2.5, said subject is a therapy responder subject, said therapy consisting of the pharmaceutical composition for use according to claim 1.

12. A method for selecting a subset of patients responding to the composition according to claim 1, wherein said method comprises:Evaluate the expression levels of at least five, or of at least ten, or of at least 19 of the genes selected in the group comprising: LHFPL, STXBP5L, KCNH4, RAMP2, FLG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SUEF1, ORS2, CADM1, ETBR, SPINT1, ERRK1, TGM2, MST4 in a subject; wherein, when the expression level of at least one of the genes selected in the group comprising EHFPE, STXBP5E, KCNH4, RAMP2, is higher than the expression level of said gene in the average population, said subject is classified as a Responder; wherein, when the expression level of at least one of the genes selected in the group comprising FEG, IE1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SUEF1, ORS2, CADM1, LTBR, SPINT1, LRRK1, TGM2, MST4 is lower than the expression level of said gene in the average population, said subject is classified as a Nonresponder.

13. A pharmaceutical composition comprising R-Ketorolac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing LHFPL, STXBP5L, KCNH4, RAMP2 at higher level with respect to the levels measured in the average population.

14. A pharmaceutical composition comprising R-Ketorolac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing FLG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SULF1, ORS2, CADM1, LTBR, SPINT1, LRRK1, TGM2, MST4 at lower levels with respect to the levels measured in the average population.

15. A pharmaceutical composition comprising R-Ketorolac, Bumetanide and Niflumic acid for use in the treatment of glioblastoma in a subpopulation of subjects in need thereof, wherein said subpopulation consists of subjects expressing LHFPL, STXBP5L, KCNH4, RAMP2 at higher levels with respect to the levels measured in the average population adP1732PC00 expressing FLG, IL1R1, TPBG, FEZF1, ARHGEF5, DOCK5, LINC02984, SULF1, ORS2, CAD M l , LTBR, SPINTl, LRRK1, TGM2, MST4 at lower levels with respect to the levels measured in the average population.

Images