UC Berkeley

Glioblastoma remains a deadly disease due to the diffuse infiltration of single tumor

cells into the surrounding tissue. Even with the current treatment regimens of surgery,

radiation, and chemotherapy, the median survival time is approximately one year. Like

many solid tumors, glioblastoma is extremely heterogeneous with respect to multiple

phenotypes such as invasive capacity, therapeutic resistance, and tumorigenicity. This

heterogeneity complicates our understanding of glioblastoma and consequently, our

ability to treat this disease. Unfortunately, standard population-based assays can mask

the properties of rare subpopulations within a tumor and therefore, obscure our

understanding of these subpopulations. As a result, without tools that allow for single

cell analysis, we are unable to interrogate how different subpopulation phenotypes may

serve as indicators of glioblastoma tumor growth and progression.

In this dissertation, we sought to develop and optimize new microfluidic tools to analyze

single glioblastoma cells for a range of phenotypes: viscoelasticity, motility, and

invasion. First, we developed a cross-slot based platform and a corresponding

analytical model that enables the determination of cellular viscoelastic properties

(stiffness and fluidity) in a high-throughput manner. Using this platform, we quantified

the viscoelastic properties of 3T3 fibroblasts and glioblastoma tumor initiating cells

(TICs) and observed the expected changes in the cellular elastic modulus in response

to agents that soften or stiffen the cytoskeleton. Second, we developed a microfluidic

device and workflows that integrates measurements of invasive motility and targeted

protein expression with single cell resolution, which we named SCAMPR (Single Cell

Analysis of Motility and Proteotype). Using this platform, we identified two proteins,

Nestin and EphA2, which positively correlates with TIC invasive motility.

In summary, this dissertation focuses on the development of microfluidic platforms for

single cell analysis. Our developed platforms provide a method in which to interrogate

single cells in a high throughput manner and to identify novel relationships between

varied cellular phenotypes. These insights are crucial for the identification of both rare

subpopulations within a given tumor and patient-specific protein markers that describe

these subpopulations.