This paper introduces the first results of dielec-7 tric spectroscopy characterization of glioblastoma cells, measur-8 ing their crossover frequencies in the ultra-high-frequency range 9 (above 50 MHz) by dielectrophoresis (DEP) techniques. Exper-10 iments were performed on two glioblastoma lines U87-MG and 11 LN18 that were cultured following different conditions, in order 12 to achieve different phenotypic profiles. We demonstrate here that 13 the presented DEP electrokinetic method can be used to discrim-14 inate the undifferentiated from the differentiated cells. In this 15 study, microfluidic lab-on-chip systems implemented on bipolar-16 complementary oxide semiconductor technology are used allowing 17 single cell handling and analysis. Based on the characterizations 18 of their own intracellular features, both the selected glioblastoma 19 (GBM) cell lines cultured in distinct culture conditions have shown 20 clear differences of DEP crossover frequency signatures compared 21 to the differentiated cells cultured in a normal medium. These re-22 sults support the concept and validate the efficiency for cell char-23 acterization in glioblastoma pathology. 24 Index Terms-BiCMOS chip, biological cell manipulation, 25 glioblastoma cells, high frequency dielectrophoresis. 26 I. INTRODUCTION 27 G LIOBLASTOMA (GBM) is one of the most frequent and 28 the most aggressive tumors of the central nervous system.

Medulloblastoma (MB) is the most common malignant pediatric brain tumor. Despite the progress of new treatments, the risk of recurrence, morbidity, and death remains significant and the long-term adverse effects in survivors are substantial. The fraction of cancer stem-like cells (CSCs) because of their self-renewal ability and multi-lineage differentiation potential is critical for tumor initiation, growth, and resistance to therapies. For the development of new CSC-targeted therapies, further in-depth studies are needed using enriched and stable MB-CSCs populations. This work, aimed at identifying the amount of CSCs in three available human cell lines (DAOY, D341, and D283), describes different approaches based on the expression of stemness markers. First, we explored potential differences in gene and protein expression patterns of specific stem cell markers. Then, in order to identify and discriminate undifferentiated from differentiated cells, MB cells were characterized using a physical characterization method based on a high-frequency dielectrophoresis approach. Finally, we compared their tumorigenic potential in vivo, through engrafting in nude mice. Concordantly, our findings identified the D283 human cell line as an ideal model of CSCs, providing important evidence on the use of a commercial human MB cell line for the development of new strategic CSC-targeting therapies.

Glioblastoma (GBM) is one of the most aggressive solid tumors, particularly due to the presence of cancer stem cells (CSCs). Nowadays, the characterization of this cell type with an efficient, fast and low-cost method remains an issue. Hence, we have developed a microfluidic lab-on-a-chip based on dielectrophoresis (DEP) single cell electro-manipulation to measure the two crossover frequencies: fx01 in the low-frequency range (below 500 kHz) and fx02 in the ultra-high-frequency range (UHF, above 50 MHz). First, in vitro conditions were investigated. An U87-MG cell line was cultured in different conditions in order to induce an undifferentiated phenotype. Then, ex vivo GBM cells from patients’ primary cell culture were passed through the developed microfluidic system and characterized in order to reflect clinical conditions. This article demonstrates that the usual exploitation of low-frequency range DEP does not allow the discrimination of the undifferentiated GBM cells from the differentiated one. However, the presented study highlights the use of UHF-DEP as a relevant discriminant parameter. The proposed microfluidic lab-on-a-chip is able to follow the kinetics of U87-MG phenotype transformation in a CSC enrichment medium and the cancer stem cells phenotype acquirement.

No abstract