DISSERTATION (PH.D.)--THE UNIVERSITY OF MICHIGAN

NEUROSCIENCE Evaluating Multisensory Processing in the Mouse Model: Clinical and Translational Implications Justin Kyle Siemann Dissertation under the direction of Mark T. Wallace Multisensory integration is the synthesis of information from the different sensory systems. Numerous investigations have evaluated multisensory function in humans and in various animal models with these studies demonstrating significant neural, behavioral and perceptual benefits conferred under multisensory conditions. In addition, emerging studies have described atypical multisensory function in clinical populations such as those with autism spectrum disorder. While a variety of animal models have been used to assess multisensory processing, there were no current behavioral tasks available for the mouse model. Here, we demonstrate the first behavioral paradigm to assess and evaluate multisensory processing in the mouse, and which show similar behavioral benefits as compared to larger animal species. In addition, we then demonstrate the first characterization of atypical multisensory behavioral function in a mouse model associated with autism spectrum disorder. Lastly, we observed that depending on the type of behavioral task implemented, mice can demonstrate vastly different behavioral responses under these multisensory conditions. Overall, we describe a series of novel studies that examine and characterize multisensory behavioral function in the mouse model system with the belief that this work may provide opportunities for further insights into the underlying neural mechanisms of multisensory function in both typical development and in autism spectrum disorder.

NEUROSCIENCE Electrophysiological Signatures of Multisensory Temporal Processing in the Human Brain David Michael Simon Dissertation under the direction of Professor Mark T. Wallace Events in the natural world frequently generate sensory signals in more than one sensory modality. Multisensory integration, the process of combining these signals into a single coherent perceptual representation, conveys numerous behavioral and perceptual benefits. Integration of audiovisual inputs is particularly relevant to everyday function, as many ecologically important signals such as speech have both auditory and visual elements. To accomplish appropriate integration of audiovisual speech inputs, the nervous system utilizes a number of cues, one of which is the temporal relationship between the auditory and visual signals. We investigated the neurophysiological bases of how the brain uses temporal information to appropriately integrate audiovisual speech inputs using a combination of psychophysics and electroencephalography (EEG). This series of investigations elucidated that, for ecologically valid speech signals, the brain uses temporal concordance to reduce the magnitude of auditory cortical responses and increase the efficiency of cortical processing. Furthermore, we demonstrate that when temporal relationships are task relevant, the neural signals associated with temporal processing are distributed to other brain regions through the formation of functional neural networks. Lastly, perceptual plasticity at the single trial level during temporal processing was found to be associated with changes in the physiological signatures of sensory evidence accumulation in decisional circuits. These experiments offer unique insights into how the brain utilizes temporal concordance to control multiple levels of sensory integration which span low-level cortical responses, transfer of temporal information to higher cognitive systems, and the formation of flexible perceptual decisions. Together they offer the first comprehensive physiological characterization of audiovisual temporal processing in the human brain.

The ability to integrate information across the senses is vital for coherent perception of and interaction with the surrounding world. In the mammalian brain, the superior colliculus (SC) is critical for this multisensory processing to occur. Much is known regarding the organization and function of neurons within the SC, including the necessity of normal sensory experience for proper development of these neurons, influence of modulatory neurotransmitter systems, and how these specific neurons are involved in multisensory integrative processing. However, open questions regarding how neurotransmitter systems and developmental parameters are involved in multisensory processing in adulthood remain. Previous work has shown that sensory experience throughout development is essential for proper multisensory integrative capacity of SC neurons. Here, it is established that this normal sensory experience requirement is maintained throughout a lifetime; perturbation of visual experience in adulthood also affects multisensory integrative capacities of SC neurons. In addition, studies detailed here sought to determine the role of the serotonergic (5-HT) system in multisensory processing occurring within the SC. Through electrophysiological and pharmacological methods, a modulatory role of the 5-HT system was demonstrated, as alterations in the serotonergic signaling system within the SC affected responsivity, receptive field characteristics and integrative capacities of multisensory neurons. These studies help to further our knowledge and understanding of the mechanisms at work in the SC in order to produce and maintain proper multisensory processing capabilities.

The autonomic and sensory nervous systems are derived from the neural crest and are required for the normal functioning of visceral organ systems, including the bladder and urethra (lower urinary tract, LUT). Surprisingly little is known about the molecular factors involved in the normal development and maturation of LUT innervation. Serotonin receptor 5-HT3A (encoded by gene Htr3a) was found to be significantly enriched in differentiating autonomic neurons innervating the LUT. A variety of pharmacological, physiological, and behavioral approaches were used to determine the roles of 5-HT3A in the development of LUT innervation. Over-stimulating 5-HT3A in sacral neural crest cells in vitro disrupted neuronal differentiation outcomes and inhibited neurite outgrowth in pelvic ganglia explants. Loss of 5-HT3A in vivo resulted in a transient disturbance of PG autonomic neuronal subtypes and an increase in autonomic and sensory neuronal fibers innervating the bladder. Male Htr3a knockout mice exhibited increased urinary voiding frequency and decreased bladder voiding efficiency. Overall, the work presented in this dissertation highlights a previously unknown role for 5-HT3A signaling in peripheral nervous system development and its requirement for normal adult urinary tract function.

Glutamate (Glu) signaling plays a critical role in regulating neural excitability, thus supporting many behaviors. Perturbed Glu homeostasis in the brain is implicated in multiple psychiatric and neurodegenerative disorders including Parkinsonâs disease, where theories implicate excitotoxic Glu signaling in dopamine (DA) neuron degeneration. Microscopy studies demonstrate that mutation to a glial expressed gene in C. elegans, swip-10, induces premature and progressive DA neuron degeneration typified by dystrophic dendritic processes, as well as shrunken and/or missing cell soma. DA neuron degeneration in swip-10 mutants is rescued by glial-specific expression of WT swip-10, and genetic studies implicate Glu signaling, Ca2+-permeable Glu receptors, intracellular Ca2+ signaling, and apoptotic cell death in swip-10 DA neurodegeneration. Like swip-10, the putative mammalian ortholog, Mblac1, encodes a protein containing a metallo beta-lactamase domain. To gain insight into the role of MBLAC1 in vivo, CRISPR/Cas9 methods were employed to generate a MBLAC1 knockout (KO) model. Using serum from MBLAC1 KO and WT mice, untargeted metabolomic analyses were performed to nominate metabolic pathways responsive to MBLAC1 loss. Findings point to taurine metabolism, primary bile acid biosynthesis, and linoleate metabolism as pathways sensitive to loss of MBLAC1. The swip-10/MBLAC1 KO models serve as platforms for the elucidation of mechanisms that enhance risk for neurodegenerative diseases and/or the identification of agents that can limit excitotoxicity.

Epileptic encephalopathies (EEs) are a devastating group of severe childhood onset epilepsies with medication resistant seizures and poor developmental outcomes. Many EEs have a genetic etiology and are often associated with de novo mutations in genes coding for proteins involved in synaptic transmission, including GABAA receptor subunit genes. A better understanding of GABAA receptor subunit mutations associated with EEs in vitro and in vivo will facilitate epilepsy diagnosis as well as treatments in the future. Here we employed a combination of next generation sequencing and in vitro functional assays and established for the first time that missense GABRG2 mutations are genetic risk factors for EEs. In addition, we focused on three nonsense GABRG2 mutations associated with epilepsies of different severities and demonstrated that they resulted in different structural disturbance and different suppression of wild-type partnering subunits. Finally we investigated the performance of heterozygous knock-in (KI) mice which bear the GABRB3(N110D) mutation associated with infantile spasms (Gabrb3+/N110D KI mice) in a battery of behavioral tasks, showing that they had significantly abnormal neurobehavioral profiles persisting into adulthood. To conclude, we have shown meaningful functional and structural changes for EE-associated GABRG2 mutations in vitro, and have determined the behavioral comorbidities of KI mice harboring a human infantile spasms GABRB3 mutation in vivo.

Studies of genetics and neural structure and function have implicated reward circuitry in the pathogenesis of autism spectrum disorders (ASD). Consequences of atypical reward have been studied most thoroughly in the context of the social motivation theory of autism to describe the lack of development of socio-communicative behaviors in ASD. Recent research indicates alterations in reward processing may also play a role in both the development restricted, repetitive behaviors and interests and the presence of certain co-morbid diagnoses in ASD. My dissertation studies have sought to address how alterations in reward processing may contribute to the development of two such clinical features in ASD, namely, restricted interests, which may be pathognomonic of ASD, and highly prevalent co-morbid diagnoses of depression. To this end, I have worked to modify / develop novel tasks to quantify attentional and motivational biases to specific types of nonsocial and affective information, using behavioral phenotyping, eye-tracking, and EEG. Overall results indicate that reward processing in ASD may be characterized by a nonsocial bias that is present early in development and across a range of functional ability and may interfere with processing of social reward. Further, altered reward processing may guide motivational biases that confer risk for co-morbid depression in ASD.

G protein-coupled receptors (GPCRs) are a large and diverse group of transmembrane receptors which convert extracellular signals into intracellular responses via coupling to heterotrimeric G proteins. In order to integrate diverse extracellular signals into a message the cell can recognize and respond to, conformational changes occur that rewire the interactions between the receptor and heterotrimer in a specific and coordinated manner. By interrogating the structural and sequence-based constraints of these proteins across each of the signaling states, we can infer which residues are necessary for function and selectivity. Two opportunities emerged to construct predictive models for G protein interactions that invite the application of informatics: 1) With advances in genome sequencing, we can reconstruct and reconcile fully resolved phylogenetic histories of G alpha subunit subfamilies; 2) With experimental G protein structures in complex with protein partners, we can model interactions affiliated with signal mechanics. Here 1 and 2 were combined to create quantitative, predictive models of G protein signaling to identify conserved patterns and characteristics necessary for subfamily-specific protein-protein interactions that will ultimately aid in drug discovery. We were able to successfully model and predict a number of residues across the G protein structure acting as the underlying communication network necessary for function. We then turned to evaluate the sequence-based constraints which imping on subfamily-specific function and selectivity. By integrating sequence information, we were able to predict residue motifs necessary for G protein activation and signaling. Key positions from these predictions have been biochemically validated through mutational studies to verify requirements for G protein subfamily-specific interaction with activated GPCRs and to improve the in silico methodologies in an iterative fashion. Overall, our studies have resulted in new understanding of G protein activation, evolution, and function. As GPCRs represent the targets of roughly half of all therapeutics, increasing our understanding of the intracellular transducing element and the system around these proteins is critical for continued improvement and development of therapeutics. As many diseases are the direct cause of erroneous G protein signaling, study of the mechanism of G protein evolution, activation and signaling remains paramount for the improvement of human health.

Neurons are specialized cells that communicate through electrochemical signals: a neuron receives input through dendrites and sends information through a single axon. The receptive field for each neuron is defined by sister dendrites that occupy discrete domains. Two neurons in C. elegans, PVDL and PVDR, are model nociceptors for studying dendrite development because they exhibit an elaborate but well-characterized dendritic arbor that is readily visible beneath the skin. Previous studies of the PVD neuron showed that the LIM-homeodomain transcription factor MEC-3 is required for higher order dendritic branching. Microarray profiles of wild-type and mec-3 mutant animals identified targets of MEC-3 that may be involved in this developmental process. One of those targets, HPO-30/Claudin, was shown to be required for pioneer branch stabilization. Another target of MEC-3, the TFIIA-like zinc finger transcription factor EGL-46, was also found to be required for 2° branches, but the extent of the defect in egl-46 mutants was not as severe as those of mec-3. The work in this thesis explores the genetic pathways required for proper development of dendritic branches using the PVD nociceptive neuron as a model. Specifically, I found that EGL-46 works cell-autonomously in PVD to promote commissural 2° branches and that EGL-44 works with EGL-46 in this context. This EGL-44/EGL-46 pathway works in parallel to the previously reported HPO-30 pathway. In addition to being a target of MEC-3, EGL-46 is regulated by other factors as well. MEC-3 is also required for 1° branch length and axon length. Finally, I generated a strain that was optimal for isolating PVD neurons from worms by fluorescence-activated cell sorting (FACS). Additional targets of MEC-3 were identified from a differential expression analysis of mec-3 mutants versus wild-type worms using these FACS-isolated PVD cells. This dataset provides a foundation for future work on specific components downstream of MEC-3 that are required for dendrite development.