Multimodal Imaging of Physiologic Changes Induced by Anti-Angiogenic Therapy in Glioblastoma

Abstract

Glioblastoma (GBM) is the most frequent and malignant form of primary brain tumors. The standard of care for this disease consists in surgical resection, followed by radiotherapy and chemotherapy. Yet, the infiltrative nature of the disease and the resistance to current therapies cause GBMs to inevitably recur, limiting the prognosis to a little more than a year. GBMs are highly vascularized, and it has long been proposed that interfering with the supply of oxygen and nutrients to the tumor could be used as a therapeutic strategy. Research in this domain has recently lead to the approval in the US of the anti-angiogenic agent bevacizumab for second line treatment of GBMs. An accelerated approval by the Food and Drug Administration was granted on the basis of clinical trials that showed strong radiological response and improved progression free survival in comparison to historical controls. Since then, questions have however been raised about the true antitumor effect of the drug since, despite early improvement in general patient condition, resistance to therapy seems to occur. Benefit in overall survival has not been demonstrated, whether bevacizumab is given as single agent or in combination with chemotherapy. In the present thesis, multimodal imaging techniques were used to assess the changes induced by antiangiogenic therapy in a clinically relevant model of GBM. Radiological findings obtained by Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET) were compared to histological and molecular analyses to provide insight into the radiological, physiological, metabolic and molecular responses to the bevacizumab therapy. Using perfusion and diffusion MRI, we observed that anti-angiogenic therapy normalizes the tumor vasculature, and strongly reduces the permeability of blood vessels. While this probably contributes to the improvement in patients condition by reducing peritumoral edema and associated side effects, it also leads to drastic radiological changes which may misleadingly suggest an antitumor effect while the tumor actually continues to grow, possibly adopting a more infiltrative progression pattern. We also found that blood supply to the tumor was strongly reduced after the treatment, an observation of clinical importance that suggests that anti-angiogenic treatment could impair the delivery of systemic chemotherapeutic drugs. The reduced blood supply is also consistent with an increased hypoxia observed in our models when we assessed it by PET. This again has therapeutic significance since increased hypoxia is also associated with reduced efficacy of radiotherapy and chemotherapy. Metabolic changes induced by anti-angiogenic therapy were evaluated in vivo by Magnetic Resonance Spectroscopy and PET analysis of glucose uptake. These studies highlighted an increased glucose consumption and increased glycolytic metabolism in the treated animals, that was later confirmed by metabolomic analysis of tissue extracts. The putative increased acidification of the tumor microenvironment that may result from this glycolytic activity is a factor that favors the infiltration of tumor cells in the brain parenchyma. This suggests that therapies that combine anti-angiogenic compounds with drugs designed to interfere with the impaired metabolic activity of tumors could be interesting as a new treatment. Preliminary results from preclinical studies that use this strategy seem to support this hypothesis. We finally examined the hypothesis that anti-angiogenic therapy could impair neuro-cognitive function, a finding that has recently been suggested in a phase III clinical study. In a small preclinical study using measurements of Long Term Potentiation, a technique classically used in neuroscience to determine memory function, we observed that, in comparison with controls, animals treated with anti-angiogenic therapy showed reduced neuronal plasticity in the hippocampus, a region of the brain associated with spatial learning and short-term memory. These results therefore support the findings of the clinical study but deserve further clinical investigation. In conclusion, the findings in the present thesis show that MRI and PET have complementary roles in the imaging of brain tumors and can be combined to obtain insight into the mechanisms through which tumor cells adapt to anti-angiogenic therapies. Imaging protocols commonly used in the clinic today provide a partial and sometimes misguiding view of the physiological changes induced by the treatment. The addition of new physiologic and cellular imaging techniques could in the future improve our ability to detect, characterize and treat malignant brain tumors, especially in the context of evolving cellular and molecular therapies.