P14.13.B DECODING GLIOBLASTOMA: SIM4ULATING TUMOUR MICROENVIRONMENT TO DETECT NOVEL BIOMARKERS USING ORGAN-ON-CHIP MODELS

Abstract

Abstract BACKGROUND Glioblastoma (GBM) is the most common malignant primary brain tumour in adults, with a survival rate of 14 months even after current treatment. There are two major limitations in understanding how the tumour works and what markers could facilitate its diagnosis. On the one hand, the great heterogeneity is responsible for the ineffectiveness of treatments. This underlines the importance of developing more complex models that can simulate different elements of the tumour microenvironment of GBM, including hypoxic gradients that lead to the generation of the characteristic necrotic core. On the other hand, the diagnostic techniques currently used have their inherent limitations and entail serious risks for the patients. Therefore, the development of diagnostic markers that allow real-time monitoring of tumour status without posing a very high risk, could improve diagnosis and treatment of patients with GBM. METHODS Our approach consisted on a gas impermeable and pillarless microfluidic device composed of a central chamber, where U-251 MG cells were highly dense seeded in a collagen matrix, and two side channels that provided oxygen and nutrients to the cells. The device was used to generate the necrotic region characteristic of glioblastoma, then, the medium that flowed through the lateral channels in contact with the cells was collected on different days of necrotic core formation and analysed by solid-phase microextraction system coupled with gas chromatography-mass spectrometry. This technique was used to detect the presence of differential volatile organic compounds (VOCs) secreted by tumour cells in different microenvironments. RESULTS The microfluidic device enabled the generation of an in vitro glioblastoma tumour microenvironment characterised by a necrotic zone in the centre of the tumour mass after seven days of cell culture. This allowed the study of VOCs in different microenvironment conditions, and showed that devices with a necrotic core exhibited a much more complex VOC footprint. Among the compounds detected were some types of aldehydes, phenols and nitrogen compounds. Clinical validation of the GBM VOC profile is ongoing. CONCLUSION Our impermeable and pillarless device has allowed us to simulate elements of the tumour microenvironment that are difficult to reproduce with other models, such as central necrotic zones surrounded by cells with nutrient availability. In addition, it offers a promising advance in the characterisation of the metabolic fingerprint of glioblastoma that bring us closer to real-time monitoring of the disease with less invasive methods than those currently used.