Mechanical compliance has been demonstrated to be a key determinant of cell behavior, directing processes such as spreading, migration, and differentiation. Durotaxis, directional migration from softer to more stiff regions of a substrate, has been observed for a variety of cell types on mechanical gradients with absolute stiffnesses and gradient rates spanning multiple orders of magnitude. Recent stiffness mapping experiments have shown that local changes in tissue stiffness in disease are often accompanied by an altered extracellular matrix composition in vivo. However, the importance of extracellular matrix composition in cellular responses to mechanical gradients has not yet been thoroughly explored. To address this problem, we have developed a method to produce polyacrylamide hydrogels featuring highly tunable gradients in mechanical stiffness that allow for independent control of the absolute substrate stiffness and gradient rate. Maskless lithography is used to micropattern glass slides with hydrophobic and hydrophilic silanes to constrain the geometry of pre-gel solutions and establish a predictable cross-linker diffusion gradient, resulting in consistent linear mechanical gradients upon polymerization. This feature, together with the ability to control ECM composition independent of substrate stiffness, allows us to isolate the effects of mechanical and biological signals on cell migratory behavior. Using this system, we have tracked the migration of vascular smooth muscle cells and NIH 3T3 fibroblasts in vitro on mechanical gradient and uniform stiffness hydrogels and quantitatively analyzed differences in cell migration as a function of extracellular matrix composition. Our results show that both vascular smooth muscle cells and NIH 3T3 fibroblasts will exhibit durotaxis on mechanical gradients coated with fibronectin but not on those coated with laminin, demonstrating that extracellular matrix type can act as a regulator of a cell’s response to mechanical gradients. Interestingly, NIH 3T3 fibroblasts were also observed to migrate randomly on gradients coated with a mixture of both fibronectin and laminin, suggesting that there may be a complex interplay in the cellular response to mechanical gradients in the presence of multiple extracellular matrix signals. These findings indicate that the composition of the adhesion ligand is a critical determinant of a cell’s migratory response to mechanical gradients. / 2018-03-09T00:00:00Z

Cells within the aortic wall are exposed to variable cyclic mechanical stimuli due to variability in blood pressure. Vascular smooth muscle cells, which have been shown to play an important role in maintaining aortic wall tension and stiffness, respond to mechanical stretch by reorganizing their cytoskeletal and contractile elements, which requires energy produced by the mitochondria. Mitochondria have also been implicated as a sensor in mechanotransduction pathways. We have shown recently that cells transduce fluctuations in their mechanical environment by reorganizing their cytoskeletal and mitochondrial networks, which can alter energy. Here we hypothesize that fluctuations in the mechanical environment regulates cell contractility through mitochondrial function. In order to test this hypothesis we first developed a cell stretching device that can deliver equi-biaxial stretch during simultaneous live imaging of subcellular structures. Cells were seeded on deformable silicone membranes which were stretched both monotonously and variably. To test the effect of substrate stiffness on mitochondrial function and to measure traction forces a gel with a tunable stiffness was added to the silicone membrane. Cells undergoing variable stretch had larger mitochondrial clusters and higher membrane potential than those undergoing monotonous stretch. To confirm the effects of increased mitochondrial membrane potential on contractility of vascular smooth muscle, aortic rings from Wistar Kyoto rats were stretched with variable and monotonous stretch patterns at physiological levels of variability and strain. Variable stretch maintained the contractile force of the vessel, while monotonous stretch decreased the force. Inhibition of ATP synthase function abated the difference between stretch patterns, implicating the role of the mitochondria during fluctuation-driven mechanotransduction. To measure the effects of variable stretch on aorta wall mechanics, the storage and loss moduli were computed for each cycle. While stretch pattern did not have an effect on vessel stiffness it did have an effect on the ratio of force generation to stiffness. Changes in force to stiffening ratio during contraction have implications on maintaining aortic wall mechanical homeostasis, particularly in diseased states characterized by increases in blood pressure variability such as hypertension.

Our auditory systems are evolved to process a myriad of acoustic environments. In complex listening scenarios, we can tune our attention to one sound source (e.g., a conversation partner), while monitoring the entire acoustic space for cues we might be interested in (e.g., our names being called, or the fire alarm going off). While normal hearing listeners handle complex listening scenarios remarkably well, hearing-impaired listeners experience difficulty even when wearing hearing-assist devices. This thesis presents both theoretical work in understanding the neural mechanisms behind this process, as well as the application of neural models to segregate mixed sources and potentially help the hearing impaired population. On the theoretical side, auditory spatial processing has been studied primarily up to the midbrain region, and studies have shown how individual neurons can localize sounds using spatial cues. Yet, how higher brain regions such as the cortex use this information to process multiple sounds in competition is not clear. This thesis demonstrates a physiology-based spiking neural network model, which provides a mechanism illustrating how the auditory cortex may organize up-stream spatial information when there are multiple competing sound sources in space. Based on this model, an engineering solution to help hearing-impaired listeners segregate mixed auditory inputs is proposed. Using the neural model to perform sound-segregation in the neural domain, the neural outputs (representing the source of interest) are reconstructed back to the acoustic domain using a novel stimulus reconstruction method. / 2017-09-22T00:00:00Z

Self-amplifying RNA replicons are an attractive alternative to traditional nucleic acid therapeutics, providing high, sustained expression from a low dose, without the risk of genomic integration. Replicons are derived from a viral genome, but utilize viral self-amplification to produce heterologous genes instead of the structural proteins necessary to form virions. Despite a variety of therapeutic applications, ranging from vaccination to genetic reprogramming, regulating expression from replicons remains relatively unexplored, as the field continues to rely solely on constitutive expression. However, we will demonstrate that dose cannot be used to control constitutive expression levels from a replicon. Without any means of regulation, the inability to adjust expression is a major safety concern that must be addressed before this platform can be used for more clinically-driven specifications. In this dissertation, we have employed synthetic biology to expand the potential of replicon-based platforms to include sequence-level and small molecule mediated control of expression levels. Synthetic biology aims to create and characterize libraries of highly predictable and modular genetic parts that can be combined to produce genetic circuits. To this end, we generated a collection of parts that can modulate replicon subgenomic transcription, explored existing and novel replicon-based expression platforms, and designed small molecule responsive replicon circuits. We established sequence elements that can be used to predictably control constitutive expression levels of up to three genes driven from a single self-amplifying RNA strand. We verified that this regulatory framework was functional for multiple replicon-based platforms, including multi-SGP replicons, DNA-launched replicons, and a novel self-cleaving, amplifying RNA platform. Finally, we coupled these genetic parts with small molecule responsive elements to form RNA-only circuits delivered on a single replicon that could control expression of multiple proteins based on external inputs. By introducing regulatory genetic circuits to self-amplifying RNA, we demonstrate control over the strength, timing, and location of expression, enhancing the utility of RNA for gene delivery and establishing a framework for the next generation of RNA-based therapeutics. / 2018-07-09T00:00:00Z

Genome-scale flux balance models of metabolism provide testable predictions of all metabolic rates in an organism, by assuming that the cell is optimizing a metabolic goal known as the objective function. In the first chapter of this dissertation, we introduce an efficient inverse flux balance analysis (invFBA) approach, based on linear programming duality, to characterize the space of possible objective functions compatible with measured fluxes. After testing our algorithm on simulated E. coli data and time-dependent S. oneidensis fluxes inferred from gene expression, we apply our inverse approach to flux measurements in long-term evolved E. coli strains, revealing objective functions that provide insight into metabolic adaptation trajectories. For over a hundred years, enzymes, or the proteins that catalyze metabolic reactions, have been characterized in vitro, even though the aqueous solution of a test tube little resembles the crowded intracellular milieu. Since few metabolites show unique fluorescent signatures, metabolism is all but invisible, greatly complicating efforts to describe fluxes in vivo. In the second chapter of this dissertation, we introduce a new technique called EIFFL (Estimation of Intracellular Flux through Fluorescence Loss) for visualizing the flux through a reaction inside single E. coli cells, using a substrate that undergoes an enzyme-catalyzed loss of fluorescence. EIFFL would not only further our quantitative understanding of metabolism, but enable us to promptly detect enzymes that confer clinically meaningful states, such as antibiotic resistance. We present a particular instance of EIFFL that couples nfsA, the major nitroreductase of E. coli responsible for its antibiotic sensitivity to nitrofurantoin, to 2-NBDG, a glucose derivative that loses fluorescence upon being reduced by nfsA with NADPH. We correlate the flux through the reaction with the concentration of a fluorescently tagged nfsA and measure the “flux noise” across a population of E. coli cells. Given that nfsA abolishes 2-NBDG fluorescence by the same molecular mechanism that it activates nitrofurantoin, EIFFL could serve as a means to rapidly infer the antibiotic resistance of single pathogenic E. coli cells directly from clinical samples. / 2020-02-20T00:00:00Z

Protein interactions with other proteins or small molecules are critical to most physiological processes. These interactions may be characterized experimentally, but this can be time consuming and expensive; computational methods for predicting how two proteins interact, or which regions of a protein are most favorable for binding, are thus valuable tools for understanding how proteins of interest function, and have applications in drug discovery and identifying proteins of therapeutic interest. The ClusPro and FTMap algorithms for docking or solvent mapping, respectively, model protein-protein and protein-small molecule interactions, and can be used to identify the most likely orientations of a protein complex or the regions on a protein surface with the greatest propensity for binding. Here we describe three applications of ClusPro and FTMap. ClusPro was used to develop a method for determining whether a protein-protein interface is biologically relevant, by docking the proteins and comparing the results to the given interface; a larger number of near-native structures--which have interfaces similar to that of the given complex--was found to correspond to a greater probability that an interface is biological. In another project, ClusPro was used to predict whether a mutation in a multimeric complex would trigger the formation of a supramolecular assembly, based on how often that mutated residue appeared in the interfaces of the docking results; if a mutation caused such a residue to be present in the docked interfaces more often, in comparison to those of the wild-type structure, then it was likely to induce self-assembly. FTMap was used to detect and analyze the druggability of potential allosteric sites in kinases, with mapping performed on all available kinase structures to identify and determine the potential binding affinity of binding hot spots located outside of the active site. Discrimination of proteins as dimers or monomers was implemented as an addition to the ClusPro server, ClusPro-DC, and the results of the druggability analysis of kinases were organized into an online resource, the Kinase Atlas. / 2019-02-20T00:00:00Z

Human listeners can direct top-down spatial auditory attention to listen selectively to one sound source amidst competing sounds. However, many listeners with hearing loss (HL) have trouble on tasks requiring selective auditory attention; even listeners with normal hearing thresholds (NHTs) differ in this ability. Selective attention depends on both top-down executive control and coding fidelity of the peripheral auditory system. Here we explore how low-level sensory perception and high-level attentional modulation interact to contribute to auditory selective attention for listeners with NHTs and HL. In the first study, we designed a paradigm to allow simultaneous measurement of envelope following responses (EFRs), onset event-related potentials (ERPs), and behavioral performance. We varied conditions to alter the degree to which the bottleneck limiting behavior was due to the coding of fine stimulus details vs. top-down control of attentional focus. We found attention modulated ERPs, from cortex, but not EFRs from the brainstem. Importantly, when coding fidelity limited the task, EFRs but not ERPs correlated with behavior; conversely, when sensory cues for segregation were robust, individual behavior correlated with both EFR strength and strength of attentional modulation of cortical responses. In the second study, we explored how HL affects control of auditory selective attention. Listeners with NHTs or with HL identified a simple melody presented simultaneously with two competing melodies, each from different spatial locations. Compared to NHT listeners, HL listeners both performed more poorly and showed less robust attentional modulation of cortical ERPs. While both groups showed some cortical suppression of distracting streams, this modulation was weaker in HL listeners, especially when spatial separation between attended and distracting streams was small. In the final study, we compared temporal coding precision in listeners with NHT and HL using both behavioral and physiological measures. We found that listeners with HL are more sensitive than listeners with NHT to amplitude modulation in both measures. Within the NHT listener group, we found a strong correlation between behavioral and electrophysiological measurements, consistent with cochlear synaptopathy. Overall, these studies demonstrate that everyday communication abilities depend jointly on both low-level differences in sensory coding and high-level ability to control attention.

In order to advance understanding of the brain, more specific and widely applicable methods are needed for genetically defining and labelling neuron subtypes in vivo. This dissertation describes the development and application of a novel strategy for designing viral gene delivery vectors that target and classify specific neurons in the adult rodent brain based primarily on endogenous microRNA (miRNA) expression profiles. We call these vectors microRNA-guided neuron tags (mAGNETs). We first demonstrate the feasibility of using neuron-type-specific miRNA profiles to guide viral-mediated transgene expression. Using lentiviral (LV) vectors bearing “signature” miRNA recognition sites, we explore several microRNA cassette design principles from a synthetic biology perspective and demonstrate the feasibility of preferentially targeting inhibitory (GABA+) neurons in the mouse cortex. Next, we maximize interneuron targeting specificity by altering the strength of the constitutive promoter driving transgene expression and switching the viral packaging platform from LV to adeno-associated virus (AAV), producing a GABA mAGNET that can label cortical interneurons with 98% targeting specificity in the mouse brain. The utility of this GABA mAGNET is highlighted by demonstrations of inhibitory neuron targeting in a Ube3a 2X Tg mouse model of autism, cross-species functionality in the rat cortex and hippocampus, and viral-mediated dual-color optogenetic manipulation of two separate neural subsets in the mouse cortex. Finally, we design and implement a mAGNET vector to characterize a novel neuron subtype: calcium/calmodulin dependent protein kinase II alpha (CamKIIα) neurons with low expression of miR-128. Calcium imaging in behaving mice, slice physiology and immunofluorescence characterization reveal that low-miR-128-CamKIIα+ (Lm128C) cells exhibit several unique biophysical properties, and resemble fast-spiking interneurons. This work highlights the potential of mAGNETs to help define neuron subtypes, and characterize neural activity and physiology. Overall, this dissertation presents a novel platform for viral-mediated, neuron-type specific labelling based upon endogenous miRNA expression, provides a useful tool for targeting cortical interneurons in the rodent brain with high specificity, and demonstrates the classification of a previously uncharacterized neuron subtype. / 2020-10-22T00:00:00Z

Breast cancer remains one of the leading causes of cancer death in women with one in eight women expected to develop breast cancer. Breast cancer progression causes several adverse changes in the extracellular matrix (ECM) composition and organization including an increase in stromal collagen and stiffening of the ECM. Clinical studies have recently discovered that stiff and dense breast tissue, a result of the abnormal architecture of the tumor microenvironment, correlates with breast tumor growth and increases the likelihood of tumor metastasis. However, the tumor microenvironment influence on cancer progression and intracellular behavior is not well understood due to the lack of physiologically relevant three dimensional (3D) in vitro models that are able to capture the mechanical and structural in vivo complexity and are also able to provide rigorous and quantitative understanding. The goal of this dissertation is to investigate how the mechanical components of the microenvironment influence intracellular and molecular activity to drive cancer progression in a robust and scalable 3D system. In order to address these gaps, our work studied the the impact of collagen concentration, cell type, and drug incubation time on drug response in 2D and 3D environments. To understand the role of local cellular mechanics in mediating drug response, we optimized and utilized particle-tracking microrheology to quantify the intracellular activity of single cells and spheroids embedded in 3D collagen gels. Finally, our study connected both structure and mechanics with cell signaling by investigating the relationship between the mechanical components of the ECM and the YAP/TAZ pathway. Furthermore, we integrated our 3D embedded spheroid model with tissue clearing methods to allow for complete visualization of YAP/YAZ activity throughout the dense spheroid structure. Collectively, the results showed that matrix properties interact with matrix dimensionality to influence drug response. This interaction also was found to affect intracellular activity, even in the presence of chemotherapeutic and anti-MMP drugs. We then showed how this interaction in mechanics and ECM properties affects the spatial and temporal heterogeneity of YAP/YAZ activity within a 3D spheroid. Overall, the work in this dissertation provides new insights into how the physical properties of the tumor microenvironment influence cellular form and function, as well as response to therapy of cancer cells, which may have implications on development of novel treatment strategies and patient outcome.

Endowing cells with enhanced decision-making capacities is essential for creating smarter therapeutics and for dissecting phenotypes. Implementation of synthetic gene circuits affords a means for enhanced cellular control and genetic processing; however, genetic circuits for mammalian cells often require extensive fine-tuning to perform as intended. Here, a robust, general, and scalable system, called 'Boolean logic and arithmetic through DNA excision' (BLADE) is presented that is used to engineer genetic circuits with multiple inputs and outputs in mammalian cells with minimal optimization. The reliability of BLADE arises from its reliance on site-specific recombinases that regulate genes under the control of a single promoter that integrates circuit signals on a single transcriptional layer. Using BLADE, >100 circuits were tested in human embryonic kidney and Jurkat T cells and a quantitative metric was used to evaluate their performance. The circuits include a 3-input, two-output full adder; a 6-input, one-output Boolean logic look-up table; and circuits that incorporate CRISPR–Cas9 to regulate endogenous genes. Moreover, a large library of over 15 small-molecule, light and temperature-inducible recombinases has been established for fine-tuned control. BLADE enables execution of sophisticated cellular computation in mammalian cells, with applications in cell and tissue engineering.