Combination therapies in a patient-derived glioblastoma model

Abstract

Glioblastoma is the most malignant primary brain tumor. Glioblastoma is originating from the supportive tissue of the brain, the glial cells. Despite increased research, mortality rates have not decreased significantly. Standard therapy consists of surgical resection, temozolomide and radiation therapy. The survival rate is despite maximum treatment 14.6 months. We are currently learning more about markers that correlate to response to treatment, the most important being the MGMT promoter methylation status. One of the issues that prevents treatments from being effective is the multiple escape routes glioblastoma has, to circumvent cell death. We aim to tackle this issue using combination therapies in a representative model, namely the patient-derived glioblastoma stem-like cell (GSC) model. This model reflects the molecular characteristics of the original tumor. This model allows studying novel treatment strategies in the context of heterogeneity in response to therapy, and in the context of analyzing molecular differences underlying response. In this thesis, we apply these concepts in a broad range of treatment modalities including combination therapeutics. We aim at optimizing the treatment with enhancers of radiation and of experimental oncolytic virotherapy. We have also used the approach of clinical drug screening in order to repurpose clinical drugs. We aimed at finding clinically applied drugs that may be candidates for glioblastoma treatment.

Part I of this manuscript describes studies on new therapeutic regimens using HDACi in combination with radiation, and using a large clinical drug screen. In Chapter 3 we show that the histone deacetylase inhibitors (HDACi) SAHA, VPA, MS275, LBH589 and Scriptaid, are radiosensitizers in a significant proportion of GSCs. The observed variations in sensitivity show a relationship with molecular characteristics of the specific cultures. Regarding the clinically most relevant HDACi (SAHA and LBH589), differences in the DNA damage and apoptotic response were found between responsive and resistant cultures. Various identified associated molecules that warrant further exploration as candidate response markers are pChek2 for both SAHA/radiation (RTx) and LBH589/RTx, and in addition Bcl-XL for LBH589/RTx with positive predictive values of 90% and 100%. In Chapter 4 we emphasize the efficacy of the Bcl-2 family pathway inhibition by Obatoclax to reach sensitization of GSCs to HDACi and HDACi/RTx, circumventing a tumor-related resistance mechanism to treatment. The Bcl-2 family proteins are heterogeneously expressed in glioblastoma and 30% to 60% of tumors shows overexpression. We demonstrate that Obatoclax synergized with HDACi and showed efficacy in a large set of GSCs. This pathway is an adequate target to inhibit, in order to achieve better therapeutic efficacy, also in HDACi/RTx therapy. We identified predictive gene expression profiles for the combination treatments with Obatoclax that are associated with cellular regulatory functions. In Chapter 5 we screened a large drug collection in order to repurpose clinical drugs for the treatment of glioblastoma. This has led to the identification of three clinical compounds that have potential for treating glioblastoma: amiodarone, clofazimine and triptolide. The drugs were selected based on efficacy in GSCs and lack of toxicity in normal human astrocytes. Twenty GSCs were tested for this purpose. Clofazimine and triptolide have previously been identified as potent inhibitors of immortalized glioma cell lines. Amiodarone is a novel candidate for glioblastoma treatment as a single agent.

Part II of this manuscript describes studies on combination therapeutics for oncolytic virus therapy with Delta24-RGD. Delta24-RGD is a conditionally replicating oncolytic virus that has ended phase I/II testing in patients with glioblastoma. To enhance efficacy of oncolytic virus therapy, combination therapies are needed. First of all we tested whether anti-epileptic drugs interacted with Delta24-RGD efficacy. The three agents valproic acid (a weak HDACi), phenytoin and levetiracetam were studied. HDACi have been reported to alter oncolytic viral activity. In Chapter 6 we illustrate that therapeutic levels of the most frequently prescribed anticonvulsants valproic acid, phenytoin and levetiracetam do not negatively influence the oncolytic activity of Delta24-RGD. In some cells, additive effects between drugs and the virus were observed. In Chapter 7 we show that the novel pan-HDACi Scriptaid and LBH589, which have stronger HDACi activity than valproic acid, exert enhanced anti-tumor activity in combination with Delta24-RGD in GSCs. These HDACi induced slight up-regulation of cell surface integrins, facilitating adenoviral entry and leading to increased levels of viral gene expression. The HDACi induced cell death pathways in the GSCs, thereby accelerating the virus-induced killing of the infected cells but slightly hampering the viral progeny production. The concerted action of these two treatment modalities leads to improved anti-tumor efficacy and shows limited toxicity in normal human astrocytes. Taken together, Scriptaid and LBH589 offer opportunities as potential candidates for future Delta24-RGD combination studies. Triggered by the long duration of the process of implementation of novel drugs in clinical practice, we aimed at finding effective viral sensitizers by performing a clinical drug screen. Chapter 8 describes the identification of four effective viral sensitizers for oncolytic virotherapy, including fluphenazine, indirubin, lofepramine and ranolazine. We reveal interaction of the drugs with important viral oncolytic cell death mechanism as shown in silico and in vitro. These drugs, that are known to pass the blood brain barrier, are not only applicable in glioblastoma but show synergy with Delta24-RGD in multiple cancer types. Moreover, all drugs showed synergistic activity with other oncolytic viruses as well.