Establishing, applying, and enhancing a multiple particle tracking technique for characterization of the brain extracellular microenvironment

Abstract

Neurological diseases place a significant burden on modern society, both financially and socially. The development of effective therapeutics hinges on our ability to fully understand the impacts these diseases have on brain tissue. This includes a better awareness of the barriers that prevent the delivery of therapeutics to and throughout diseased areas and the identification of possible mechanisms that can be actively targeted by drug interventions. While significant resources have gone into characterizing the cellular processes associated with neurological disease progression and the eventual loss of neuronal function, the role of the brain extracellular environment is still not well understood. This is partly due to a lack of techniques capable of probing it. The brain extracellular environment is extremely heterogenous and has harsh geometry. It consists of highly tortuous, interconnected channels known collectively as the brain extracellular space (ECS) and is filled with a dynamic, glycan-based extracellular matrix (ECM) that can degrade, rearrange, and condense to form specific structures that play particular roles. Its heterogeneity at the nanoscale and ability to dynamically change requires a characterization technique that is both live tissue compatible and has submicron resolution. Multiple particle tracking (MPT) provides both. Particle tracking methods like MPT are live tissue compatible, use colloidal-sized (between 10nm- and 1μm-diameter) tracer particles that can access small compartments like those observed in brain ECS, and leverage high-resolution fluorescence microscopy to quantify particle behavior with single-particle resolution. Quantification of particle behavior provides insight on relevant transport, geometric, and microrheological properties, all of which impact extracellular function. An additional advantage of MPT is the sheer amount of data it generates: single experiments can result in hundreds tothousands of nanoparticle trajectories, making them an obvious candidate for the incorporation of machine learning methods. With this dissertation, we first establish MPT as a viable tool for characterizing changes in brain extracellular matrix structure. We then incorporate machine learning methods to enhance the utility of MPT datasets and generate predictive models. Through the use of both neural networks and boosted decision trees, we demonstrate an ability to predict both nanoparticle properties, such as size and surface functionality, and biological properties like neurodevelopmental age, based solely on nanoparticle trajectory feature data. Finally, we apply MPT to an ex vivo model of cerebral ischemia to better elucidate the impact an extended period of oxygen- and glucose-deprivation has on the extracellular environment of different brain regions. Collectively, we hope to demonstrate the utility of MPT in characterizing the brain extracellular microenvironment and inspire its future use in work directed at better understanding the extracellular impacts of neurological disease. We believe a more comprehensive understanding of both the cellular and extracellular processes affected by neurological disease will result in the development of more effective therapeutics and potentially provide additional tools for disease diagnosis and prognosis.