(69c) Investigation of Drug Efficacy Under in Vitro Hypoxic Gradients in Glioblastoma Multiforme

Introduction: Oxygen remains the central source of life in multicellular organisms. The varying vasculature network in different respective organs and tissues helps maintain optimal and unique tissue oxygen tension (pO₂) necessary for physiological functions. The high metabolic activity in the central nervous system (CNS) results in drastic variation of pO₂ along different regions of the brain, which is strictly regulated, ranging from 0.55-8%. Maintenance of such oxygen niches is mandatory for the regulation, proliferation and appropriate differentiation of different types of neural precursor cells. Presence of chemical gradients, such as localized densities of neurotransmitters and ions, also affect neural cell differentiation. This combinatorial effect of hypoxic and chemical gradients or niches is exaggerated in the case of cancers of the CNS, particularly glioblastoma multiforme (GBM). The hypoxic gradients also diminish chemotherapeutic efficacy, resulting in high mortalities and lack of proper clinical investigations thereof. Recent studies have concluded the nullifying effect of hypoxia on the activity temozolomide, the gold standard drug for treating GBM. Therefore, the necessity of proper in vitro investigations on the efficacy of chemotherapeutic regimens on GBM mortality under hypoxic gradients is clearly evident. Specialized O₂ incubators capable of recreating hypoxia are limited to singular concentrations, therefore do not reflect the tumor microenvironment. Microfluidic technologies enable in vitro oxygen gradient formation but are restricted to low-to-moderate resolution gradients. Utilizing a novel microfluidic technology previously characterized our group, we present a novel method of testing chemotherapeutic drug efficacy under in vitro hypoxic gradients replicating the microenvironment of GBM tumor masses. The split and recombine strategy utilized efficiently generates dissolved oxygen gradients within the channels with high temporal and spatial resolution, maintained by a glass coating deposited on the channel walls. As preliminary investigation to validate device functionality, viability analysis of grade IV GBM cells (SF-539) was conducted using the hypoxia-activated drug tirapazamine. Tirapazamine is converted into its active form only in hypoxic conditions, which is reminiscent of the center of tumor masses.

Materials and Methods: PDMS devices were fabricated using standard lithography techniques. A 3-sided glass coating (cured sol-gel) was deposited on the inner walls of the microfluidic channels to prevent diffusion of ambient oxygen into the media. Temperature and reaction time of the curing process were characterized to attain an optimal and reproducible glass-coating. Calibration for both gaseous and DO were conducted prior to oxygen gradient generation experiments. SF-539 cells were then seeded in a multi cell-outlet device and subjected to hypoxic gradient by flowing regular (21% O₂) and deoxygenated (0% O₂) media at 1.5µLmin^-1 for 24 hours incubation period. Prior to seeding, the cells were stained with 10µM Cell Tracker Green CMFDA dye for live cell staining.

Results and Discussion: The average glass coating thickness progressively increases with both temperature and time. Heating at 80⁰C for 60s or 100⁰C for 20s produced optimal glass thickness of approximately 7µm with lower frequency of cracks and no unpolymerized spaces. Stable oxygen concentration gradients with high spatial resolution were generated over various flow rates, ranging from 1µLmin^-1 to 100µLmin^-1. Linear Stern-Volmer relationship (I₀:I₁₀₀) was maintained throughout. Viability analysis in the multiple-outlet device showed that SF-539 cells perished under oxygen levels below 2%, confirming the activation of the prodrug tirapazamine under the hypoxic concentrations generated by the device. Viability reduced to 40% under these conditions whereas tirapazamine had no effect on the SF-539 cells seeded in the chambers containing higher pO₂ levels. The cellular proliferation profile was comparable to controls conducted in separate hypoxia chambers using the WST-1 assay.

Conclusion: Our proposed technique enables convenient generation of in vitro DO gradients and allows us to properly investigate effect of localized hypoxia on chemotherapeutic efficacy on GBM cells. This is one of the few such studies capable of doing so. Modifications to design of the device can be done to investigate potential cellular phenomenon, such as epithelial-to-mesenchymal transition in GBM cell lines, a feat that has yet to be achieved.

Acknowledgments: Funding for this project was provided by the National Science Foundation Award # 1642794 and 1645195.