Relaxin, a peptide hormone and the endogenous agonist of the relaxin family peptide receptor 1 (RXFP1), has shown substantial promise for the therapeutic treatment of cardiovascular disorders and fibrosis. RXFP1 has received considerable therapeutic interest as a drug target over the years, and a lot of effort has been put into the development of novel ligands targeting this receptor. However, we do not understand how novel RXFP1 agonists work because we do not understand RXFP1 signalling in enough detail, and do not fully understand which pathways are important for the therapeutic actions, where they are activated in the signal transduction cascade, and in what cell types. Furthermore, development of novel ligands raises the question of biased signalling, which is the ability of different ligands for the same receptor to preferentially activate certain signal transduction pathways relative to one another. It may be possible to utilise bias to create effective drugs that have fewer side effects, but this requires us to first understand which effectors and signal transduction pathways are important for the therapeutic versus harmful actions. Furthermore, recent research has highlighted the importance of measuring the temporal aspects of signalling in order to understand bias, as signalling and bias can change over time, and using end point assays can produce a misleading picture of efficacy and bias depending on which time point was chosen. Additionally, validating findings from recombinant cells in physiologically-relevant cells is also important to understand how ligands signal across cell types, and to distinguish biased signalling from cell-specific signalling. The development of sensitive, real-time assays that can be used across cell types will aid our understanding of the molecular mechanisms underlying the actions of relaxin and other ligands targeting this promising receptor. The general aim of this thesis was to apply and develop bioluminescence resonance energy transfer (BRET)-based methods in conjunction with human native and primary cells in order to examine real-time signalling and bias at RXFP1. First, the functionality and versatility of the CAMYEL (cAMP sensor using YFP-Epac-Rluc) real-time BRET-based cAMP biosensor was demonstrated for RXFP1 and the related GPCRs RXFP2, RXFP3, and RXFP4. CAMYEL was a sensitive alternative to end point assays, as it detected concentration-dependent changes in cAMP activity at all receptors in recombinant cell lines, was dynamic and reversible, detected kinetic differences between different ligands for the same receptor, and showed potencies comparable to those seen in end point assays. CAMYEL was cloned into a lentiviral vector, and lentivirus was used to transduce THP-1 cells, which endogenously express low levels of RXFP1. THP-1 CAMYEL cells showed robust cAMP activation after relaxin stimulation and will therefore streamline the process of screening novel RXFP1 ligands. The lentiviral vector will also allow for the transduction of many mammalian cell types for real-time analysis of cAMP activity at various GPCRs, including in primary cells. However, it appeared that the CAMYEL assay was unable to detect a delayed, Gi3-mediated phase of RXFP1 cAMP activity that has been demonstrated using other assays, suggesting that CAMYEL might not detect cAMP generated in specific compartments of the cell. Second, we developed, validated, and characterised a BRET-based biosensor for cGMP activity, known as CYGYEL (cyclic GMP sensor using YFP-PDE5-Rluc8), based on the Forster/fluorescence resonance energy transfer (FRET) biosensor cGES-DE5. CYGYEL was cloned into a lentiviral vector, enabling its use across different mammalian cell types. CYGYEL was initially characterised in HEK293T cells, where it was shown to be sensitive, dynamic, reversible, and also very selective for the detection of cGMP over cAMP. CYGYEL was then used to detect cGMP after transduction of human primary vascular cells, namely endothelial and smooth muscle cells. CYGYEL detected differences in cGMP signalling kinetics both between cell types, and also between ligands that increased cGMP production via soluble versus membrane guanylate cyclases. So far we have been unsuccessful at detecting GPCR-mediated cGMP using CYGYEL, but further work is required in this area. Regardless, CYGYEL still has many uses for drug discovery. Finally, we used a variety of BRET-based assays for G protein association, second messenger activity, and ERK1/2 activity, as well as physiologically-relevant primary cells, in order to understand the mechanisms of action underlying the beneficial actions of the relaxin peptide analogue B7-33. According to previous work, B7-33 appeared to show cell-specific signalling and biological responses, whereby it had weak activity in recombinant and cancer cells, but potent activity in fibroblasts and vascular cells, as well as in vivo. Our results across several cell types indicated that B7-33 is a biased agonist that favours signalling via Gi3/cGMP over Gs/cAMP, relative to relaxin which signals potently via both pathways. These findings are consistent with B7-33’s actions as a potent vasodilator and anti-fibrotic, which depend on cGMP rather than cAMP. Relatedly, we demonstrated that B7-33 shows transient cAMP activity relative to relaxin in all cell types tested, and that in a real-time cAMP assay involving ligand washout, the cAMP response from B7-33 dropped drastically relative to relaxin, suggesting that B7-33 dissociated from RXFP1 far more readily than relaxin. We thus hypothesised that the bias shown by B7-33 is related to kinetics, whereby the relaxin/RXFP1 complex catalyses more cycles of Gs activation due to the sustained duration of the active receptor conformation, relative to B7-33 which has a faster off-rate associated with its weaker activation of Gs. However, both agonists equally activate Gi3 suggesting that the relative rates of activation and deactivation of the different G proteins may also be important. Finally, it was also observed that ERK1/2 is activated by its upstream effectors in a cell type-specific manner. Specifically, whereas previous findings have shown that ERK1/2 is primarily downstream of Gi/o in native cells, our findings show that ERK1/2 is activated downstream of Gs in HEK-RXFP1 cells, which explains B7-33’s weak ERK1/2 activation in HEK-RXFP1 cells but potent activation in native cells. These findings have implications for the development of novel biased drugs targeting RXFP1, as it is believed that the negative actions of exogenously-administered relaxin, including for example its ability to promote tumour growth in mouse models in vivo, are related to its potent cAMP activity. Conversely, equi-molar doses of B7-33 do not promote tumour growth but do retain the beneficial actions of relaxin which occur via Gi. Thus, we could potentially aim to retain the kinetic bias to maintain potent cGMP signalling, while minimising cAMP activity, and at the same time aim to develop compounds that are more drug-like with longer half-lives.

The incidence of eating disorders and obesity has reached epidemic levels/proportions worldwide and this has been further exacerbated by the COVID19 pandemic. Binge eating is a hallmark feature of most eating disorders and some subtypes of obesity. Negative affect, such as stress, can strongly influence eating behaviour and drive bingeing episodes and overconsumption of palatable food independent of homeostatic need. This is often followed by feelings of guilt and shame, resulting in greater distress and further perpetuating the stress binge cycle. Notably, stress related binge eating disproportionately affects females more than males, yet the precise biological mechanisms underscoring behavioural sex differences remain underexplored. This has in part been due to the lack of robust animal models that faithfully recapitulate the human condition. In this regard, during my PhD, I developed a highly reproducible model of stress-induced binge eating in female mice that is independent of caloric restriction. Cyclic caloric restriction is commonly applied in animal models of binge eating but is suboptimal for examining reward based feeding and binge eating driven by hedonic processes independent of the metabolic memory of negative energy balance. Having this model established, I then aimed to investigate 1) possible endocrine and centrally mediated processes involved with stress-induced binge eating in female mice; 2) changes in neuronal activation across multiple brain regions following aberrant eating behaviours precipitated by stress; 3) neuronal circuits involved in driving stress induced binge eating. To address these questions, I used a combination of behavioural testing with anatomical tracing and chemogenetic manipulation. The results presented in this thesis demonstrate the reproducibility and applicability of the animal model here developed to investigate multiple aspects of stress eating. Crucially, the stress induce binge eating seen in this model occurred in ovariectomised female mice, suggesting factors other than ovarian hormones must be involved, and that centrally mediated processes instigate this behaviour. Indeed, additional results here reported show the possible involvement of key regions from the mesocorticolimbic, thalamic, and hypothalamic systems in stress induced binge eating and provide insight into a possible neural circuit driving this behaviour. Finally, I show that chemogenetic inhibition of a discrete projection from the paraventricular nucleus of the thalamus to the insular cortex suggested a causal role for this pathway in the emergence of overeating following stress in female mice. The results presented in this thesis provide important insights into underlying biological factors involved in stress-induced binge eating and overeating in female mice.

Introduction. Traumatic brain injuries, including sports-related concussions, are a major public health concern. Although participation in sport brings with it many benefits, the potential risk of sustaining adverse injuries such as sports-related concussion is an escalating societal and ethical issue. The onset of short-lived impairment of central nervous system dysfunction in concussed athletes following a sport-related concussion can lead to complications if the condition is not recognised and removed from participation in sport in a timely manner. Rationale and Study aims. This series of studies aim to understand the nature and magnitude of head biomechanics, short-lived impairment of central nervous system (CNS) function, and clinical signs and symptoms associated with concussions and repetitive head impacts. More specifically, it is hypothesised that (i) the biomechanics of repetitive head impacts in amateur boxers, (ii) endorsement of symptoms, and performance on cognition and balance component scores immediately post-contest and, (iii) a decline in function on the King-Devick test, will be worse for boxers clinically classified with a concussion, be associated with high-risk for CNS injury circumstances, and be reflective of exposure to repetitive head impacts through exposure to repeat contests in the tournament. Method. Healthy male collegiate students (n = 376) between the ages of 18-34 years (M = 20.97, SD = 2.30) consented to participate in the Notre Dame Bengal Bouts amateur boxing tournament in 2012 and/or 2013. Boxers completed the Sports Concussion Assessment tool, Cogstate Brief Battery, and the King-Devick test prior to participation in any sparring and following each of the four elimination contests. A sample of boxers were also instrumented with X2 Biosystems headbands to monitor exposure to impacts during contests. Results. The results of this study support the hypothesis that in amateur boxers the number of head impacts, and the acceleration of those impacts are associated with the progress through the tournament, outcome of bouts, weight class, and concussion. However, there were no systematic differences in the number or magnitude of impacts between concussed and non-concussed boxers, although average rotational forces did show potential for the classification of concussed boxers. Symptom, cognition, and balance scores did not systematically change with multiple assessments nor with repeat exposure to head blows further into the tournament, however, symptom component scores were sensitive to and worse for boxers classified on performance factors representative of high-risk for CNS injury, such as sustaining a concussion, receiving a standing eight-count, and losing a contest. The cognitive component was successful in identifying concussed boxers from non-concussed boxers. But there were no meaningful differences on the balance component clinically relevant to the assessment of concussion-related changes. The King-Devick test was largely not sensitive to concussion in the context of amateur boxing and unable to appropriately classify most concussed boxers. Conclusion. This study offers guidance on the measures which provide the best predictive power for the classification of central nervous system dysfunction following a sport-related concussion and adds to the literature on repetitive head impacts. Though innovative, further study of the biomechanics of concussive head injury, amateur boxers, and the use of screening tools are warranted.

Relaxin is a peptide hormone that is involved in several physiological processes such as pregnancy, collagen breakdown, fibrosis inhibition and vasodilation. It has been investigated for the use of several disease states such as scleroderma, fibrosis, cancer and most recently acute heart failure. Relaxin’s cognate receptor is the relaxin family peptide receptor 1 (RXFP1), an integral membrane protein belonging to the G protein-coupled receptor (GPCR) family with a complex, multistep activation mechanism which is still not well understood. Given the physiological roles of relaxin, RXFP1 is a promising target for the treatment of abovementioned conditions. However, there is currently a lack of a detailed mechanism in which relaxin mediated activation of RXFP1 occurs and this makes the design of relaxin-like compounds such as long active peptide mimetics, small molecules or biologics targeting RXFP1, or understanding and optimizing existing compounds that act at RXFP1 difficult. The lack of a detailed mechanism of RXFP1 activation can be attributed to the lack of full-length RXFP1 structures. While there are proposed models of this activation mechanism, these models were derived from studies on isolated domains of RXFP1 and thus it cannot be assumed that the findings are similar to that of a full-length RXFP1. Thus, the aim of this thesis was to work toward active and inactive state structures of full-length RXFP1 using cryo-electron microscopy (EM). By solving active and inactive state structures, we can overlay these structures to determine key conformational changes and key residues that interact with relaxin to determine a complete mode of relaxin mediated activation of RXFP1. However, these studies are hampered by the limitations of cryo-EM to study inactive state GPCRs and the low recombinant expression of WT RXFP1 which makes producing sufficient amounts of purified RXFP1 for these studies very difficult. In this thesis we optimised the expression and purification of RXFP1 for the purposes of cryo-EM studies. We also developed and optimised a novel tool, monomeric ultra-stable GFP (muGFP) as an intracellular loop 3 (ICL3) fusion partner to overcome the limitations of inactive state cryo-EM studies. We applied this to a thermostabilised variant of the alpha1A-adrenoceptor and demonstrated its utility for cryo-EM studies before applying it to RXFP1. Next, we applied an established workflow for the production of active state GPCR-G protein complexes in insect cells for cryo-EM studies to WT RXFP1 for the active state studies of the receptor. We also experimented with the expression and formation of an RXFP1-G protein complex in a mammalian expression systems. However, we were unable to proceed to cryo-EM studies of either inactive or active state RXFP1 due to inability to produce sufficient quantities of protein To overcome the limitation of poor protein yield, we developed a novel mammalian cell-based method of directed evolution. Existing methods of GPCR directed evolution are primarily E. coli based, and as RXFP1 is unable to be expressed in E. coli due to requiring post-translational modification, a mammalian system was required. We applied this novel method to RXFP1 and were able to evolve mutant #35, which demonstrated an ~9x increase in recombinant RXFP1 expression. Additionally, we also identified 2 mutants that demonstrated interesting pharmacological changes from WT. This includes a mutant that demonstrated an increase in basal signalling, and another mutant that demonstrates a decreased pEC50 for relaxin, that is a higher concentration of relaxin is required to produce an equivalent response in WT. By evolving high expressing mutant #35, we could potentially overcome the bottleneck of insufficient purified protein yield for cryo-EM studies. By applying mutant #35 to the workflows developed in this thesis, we can potentially enable downstream cryo-EM studies of RXFP1 through the ability to produce ~9x more protein than WT. Through enabling these studies, we may be able to elucidate the mechanism in which relaxin triggers RXFP1 activation in a full-length receptor. Understanding this mechanism in atomic resolution detail through cryo-EM studies could then facilitate rational drug design of novel relaxin-like mimetics for the treatment of acute heart failure or fibrosis or antagonists for the treatment of certain cancers.

Epilepsy is a common and chronic neurological condition characterised by the emergence of excessive or hypersynchronous electrical activity within the brain. A significant proportion of epilepsy is caused by gene mutations, many of which disrupt the function of subcellular protein structures known as ion channels that regulate the excitability of nerve cells (neurons). Despite the prevalence of epilepsy and its societal and economic impact, the mechanisms relating ion channel dysfunction to abnormal electrical activity within neuronal networks remain unclear. This is a matter of importance as approximately one-third of patients with epilepsy suffer intractable seizures despite treatment with modern anti-seizure pharmacotherapy. A more comprehensive understanding of epilepsy pathophysiology that that links ion channel pathology to network dysfunction may reveal new avenues for treatment. In this thesis, the biophysical consequences of two ion channel mutations associated with genetic forms of human epilepsy are explored using computational modelling and experimental electrophysiology. The first is a mutation of the NaV1.1 channel: a voltage-gated sodium channel that serves as an important regulator of neuronal excitability. In this work, we find that a NaV1.1 mutation associated with a severe form of epilepsy leads to impaired cortical inhibition through depolarisation block of inhibitory interneurons. Our results also suggest that NaV1.1 plays a central physiological role for sustaining high firing rates within cortical inhibitory interneurons. The second is a mutation of the GABAA (gamma-aminobutyric acid) receptor: a ligand-gated ion channel that mediates a powerful inhibitory influence within the brain known as tonic inhibition. Using computational modelling we predict that tonic inhibition can selectively modulate the excitability of subtypes of cortical interneurons according to their intrinsic electrophysiological properties. Our models suggest that differential modulation of neuronal excitability occurs via a novel electrophysiological mechanism that is mediated through the dendritic tree. These predictions are supported by in-vitro experiments, and further analysis suggests that modulation of neuronal excitability is dependent upon the expression of certain subtypes of voltage-gated potassium channels, such as the KV3.1 channel. A theme arising from this work is the relevance of distinct subtypes of inhibitory interneurons for regulating excitability in the brain. Therefore, this idea is explored in further detail using a cortical network model that incorporates different interneuron subtypes. Our model suggests that interneurons with properties typical of Parvalbumin-positive subtypes – a prevalent interneuron class within the cortex – are crucial for regulating the extent of internally-driven excitatory activity within a neuronal network. Reductions of excitability in Parvalbumin-positive interneurons promote a network state characterised by strong coupling between excitatory neurons. Known as an inhibition-stabilised network, this network regime is associated with certain cortical computational abilities and the potential to generate epileptic seizures.

Amyotrophic lateral sclerosis (ALS) is a fatal adult-onset neurodegenerative disorder, resulting in the death of motor neurons. Sexual dimorphism is clear in ALS patients and mouse models. Capable of modifying disease onset and progression, sex differences can also manifest in the responsiveness and outcomes to drug therapies and interventions. Understanding the biological components contributing to the clinical heterogeneity in ALS is of great interest. The androgen receptor (AR), a nuclear steroid hormone receptor which mediates the biological actions of androgens, has strong links to motor neuron function and survival. Androgens are neuroprotective to cultured motor neurons and in response to nerve injury in vivo. An expansion mutation in the AR gene gives rise to lower motor neuron degeneration in the disorder known as spinal and bulbar muscular atrophy (SBMA). This thesis explores a potential role for AR in mediating the sex-differences in ALS using the transgenic SOD1G93A mouse model of ALS. Firstly, changes to the level and cellular localisation of AR, estrogen receptor alpha (ERa) and beta (ERb), and progesterone receptor (PR), were analysed within the lumbar spinal cord of male and female SOD1G93A mice relative to wildtype control mice, over the disease course. A robust decrease in spinal cord AR level occurred in symptomatic male SOD1G93A mice. This occurred in parallel with decreases in transcript levels of androgen metabolising enzyme, 5a-reductase type II. AR level in females was half that of males, whereas levels and localisation of ER and PR were comparable between sexes. ERa showed a robust localisation to astrocytic processes in symptomatic SOD1G93A mice; ERb was upregulated, possibly contributing to multiple glial cell responses. Secondly, a closer inspection of AR expression within the various motor neurons (MN) throughout the central nervous system (CNS) neuraxis was conducted. It was strikingly apparent that by endstage disease, the majority of MNs remaining in SOD1G93A male mice had reduced nuclear AR. A clear correlation between AR expression and vulnerability in ALS was not immediately evident, although unique MN clusters with high AR content were all preserved in endstage disease. In the vulnerable lumbar spinal cord MNs, AR levels were reduced in presymptomatic stages of disease. This was paralleled by a downregulation in AR within target skeletal muscle. The consequences of these perturbations to AR level remains unclear, although their early disruption could suggest their contribution to disease pathogenesis. Finally, this thesis explored the impact of disrupted AR signalling and expression levels, to SOD1G93A disease course. Firstly, chronic administration of AR antagonist, flutamide, worsened muscle pathology and exacerbated onset without affecting neurodegeneration and survival. Secondly, the Cre/LoxP system was used to conditionally delete AR from neuronal and glial cells of SOD1G93A male mice using the Nestin-Cre transgenic. Central AR deletion was not observed to have an impact on disease course, however, the metabolic phenotype of the Nestin-Cre mouse had a noticeable, and confounding, effect the on SOD1G93A lifespan. Thirdly, global overexpression of AR in SOD1G93A mice, showed subtle effects on disease onset, although did not have an appreciable impact on ALS disease course. In summary, the work of this thesis provides a robust characterisation of sex steroid hormone receptors in the SOD1G93A mouse model of ALS. Evidence of reduced AR level in MNs through disrupted local androgen metabolism may present a target for future intervention studies. Peripheral AR was identified as a modulator of disease onset, through chronic AR antagonism. Genetic manipulation studies did not identify AR as a modifier of disease outcome in the SOD1G93A.

Patients with neurodegenerative diseases, such as Parkinson’s, Alzheimer’s, and frontotemporal dementia may present laryngeal motor dysfunction even during the early stages of the disease progression. The larynx, also called the glottis, is the main valve that regulates respiratory airflow. Moreover, the muscles of the glottis are critically involved in the mediation of orofacial behaviours, such as vocalisation and swallowing. For example, glottis prevents the aspiration of foods into the lungs during swallowing. Furthermore, during speech, the vocal folds located within the glottis are vibrated by the expiratory airflow to produce sound. Post-mortem studies in patients with neurodegenerative diseases indicate that neuronal degeneration within the brainstem respiratory network may be the cause of laryngeal motor dysfunction. Not surprisingly, patients often present respiratory, vocalisation, and swallowing malfunction. These clinical reports motivated my research group to study a transgenic mouse model of neurodegeneration (i.e., tauopathy) in the brainstem respiratory network. Data from the respiratory neurobiology team at the Florey found that similarly to clinical observations, this mouse model presents laryngeal motor dysfunction, leading to respiratory and swallowing impairments. Thus, the next step was to investigate whether tauopathy in the brainstem can also be linked to impaired ultrasonic vocalisation (USV). This working hypothesis was part of my first experimental study of my PhD (chapter 2). I found that the transgenic mouse model presented USV patterns similar to the wild-type mice (control group), even during the late stage of diseases progression, when many brainstem circuits have degenerated. However, it is crucial to note that orofacial behaviours, including vocalisation, can be volitionally controlled by supra-pontine brains regions, such as the cortex and the limbic system in the forebrain, and I speculated that these control circuits for vocalisation might be able to compensate for brainstem neurodegeneration. Surprisingly, the neuroanatomical pathways that might be involved with volitional control of respiration during orofacial behaviours were poorly defined. Thus, the following experimental study (Chapter 3) examined the anatomical projections from the forebrain to the brainstem respiratory network using a retrograde tracing approach. I reported for the first time widespread descending cortical and subcortical descending projections to a subset of primary respiratory control nuclei. Functionally, this neuroanatomical framework suggests that volitional information is sent to the brainstem respiratory network to adapt breathing with orofacial behaviours, such as vocalisation. Moreover, I found a highly redundant network connectome within the brainstem respiratory network that could be implicated in the compensation of local brainstem tauopathy (Chapter 4). Collectively, data from my thesis suggest the presence of numerous pathways in the brain that may compensate for localised neurodegeneration in the respiratory network of patients with neurodegenerative diseases. In summary, my PhD thesis proposes a new neuroanatomical framework to study behavioural adaptation of breathing in healthy and neurological diseases.

Among the questions addressed by modern neuroscience is how distinct cellular structures within the nervous system give rise to particular behaviours and cognitive functions. Our ability to address such questions has been greatly enhanced with the emergence of ‘transgenic’ technologies, which allow the introduction of foreign or engineered genes into biological systems of interest, such as neural circuits. This capacity to introduce and express transgenes in neurons has led to the development of a diverse catalogue of genetically encoded markers, sensors, and actuators, which provide powerful insights into the structure and function of the nervous system. In addition to their intrinsic utility, mechanisms exist by which expression of these transgenic tools can be selectively directed to cell-types and subcellular structures of interest. Studies described in this thesis sought to establish new molecular tools and approaches for localised transgene expression in neural structures of scientific and therapeutic importance. Transgene delivery to neural circuits of interest was mediated here in rats using viral vectors, particularly adeno-associated viral (AAV) vectors, from which targeted transduction of neurons expressing the highly conserved neuropeptide, relaxin-3, was achieved using a novel cell-type specific promoter. Relaxin-3 is most abundantly expressed by a subpopulation of neurons in the pontine nucleus incertus (NI) and has been identified as an important modulator of arousal and other interrelated cognitive functions and behaviours, making the relaxin-3 system a promising therapeutic target. To better understand this circuitry using transgenic approaches, I sought to identify a promoter sequence capable of regulating cell-type specific transgene expression in relaxin-3 neurons. In parallel to a relaxin-3 promoter sequence (1,736 bp), I also characterised transgene expression under an 880 bp tropomyosin receptor kinase A (TrkA) promoter, as TrkA is exclusively co-expressed with relaxin-3 in rat NI neurons. Eight weeks following stereotaxic injection of an AAV vector, expressing mCherry under the TrkA promoter, to rat NI, widespread non-specific transduction was observed. However, rats receiving injections of an equivalent AAV vector, where transgene expression was regulated by the relaxin-3 promoter, displayed almost exclusive mCherry production in relaxin-3-expressing neurons, demonstrating targeted, cell-type specific transgene delivery. Localisation of transgene expression at the subcellular level is also an important consideration. The need for rapid and efficient electrochemical communication between neurons has led these cells to evolve highly complex morphologies, with subcellular compartments each having distinct structural, biochemical, and functional properties. These differences can impact transgene function, as seen with the chloride-conducting opsin, GtACR2, where differing subcellular ion gradients cause GtACR2 to have hyper- and depolarising effects in either the somatodendritic or axonal membrane, respectively. To resolve these paradoxical effects, whilst simultaneously providing a mechanism for the functional dissection of closely interconnected circuitry, I developed fusion constructs between GtACR2 and trafficking motifs derived from either neurexin 1a (GtACR2nrxn) or the potassium channel, Kv2.1 (GtACR2Kv2.1), which endogenously localise to the axonal or somatic membrane, respectively. The preBotzinger complex (preBotC), a key driver of inspiration and respiratory rhythm, with close but poorly understood functional connections to neighbouring cardiovascular nuclei in the medulla, was selected as an ideal system for characterisation of these GtACR2 constructs. AAV vectors expressing GtACR2Kv2.1, GtACR2nrxn, or a control GtACR2-EYFP construct, under constitutively active promoters, were injected into the rat preBotC. Fusion of GtACR2 with the Kv2.1 motif was associated with enriched neuronal expression of this construct, but did not eliminate axonal expression, while subcellular localisation of GtACR2nrxn could not be determined histologically. However, a selective dissociation of respiratory and cardiac responses to optical manipulation, both of which were seen in rats expressing GtACR2-EYFP, was observed in rats expressing GtACR2Kv2.1 or GtACR2nrxn, respectively, supporting their differential trafficking in vivo. To further clarify these results and optimise the utility of GtACR2 fusion constructs in future experiments, a capacity for cell-type specific expression was required. I therefore developed an additional AAV vector for recombinase-dependent GtACR2Kv2.1 expression in neurons. In characterising expression from this construct, I first validated a novel approach for cell-type specific transduction of an amygdaloid efferent population (CeA) projecting to the nucleus of the solitary tract (NST) in rats (referred to as CeA>NST neurons), using locally injected AAV vectors in the amygdala and a retrogradely transported, Cre-recombinase expressing canine adenovirus (CAV) vector delivered into the rostral NST. The CeA>NST pathway was chosen as a model system, given its empirically supported but poorly understood role in modulating autonomic function in response to emotional stimuli, which would benefit from cell-type specific transgenic investigations, and its long-range axonal projections, which provide a further opportunity to assess GtACR2Kv2.1 trafficking in vivo. While cell-type specific GtACR2Kv2.1 expression in CeA>NST circuitry was achieved, some axonal expression of GtACR2Kv2.1 was again present in these neurons. Overall, studies described in this thesis provide valuable contributions to the available toolset for targeted neuronal transgene expression. Changes in GtACR2 trafficking observed, following fusion with the Kv2.1 motif, are supported by independent publications which appeared in the literature concurrently with my studies, where a somatic enrichment, and significant, but not complete, reduction in axonal expression, was demonstrated for analogous GtACR2-Kv2.1 fusion constructs. As such, although effects arising from residual axonal expression must be considered and minimised, GtACR2Kv2.1 provides an improved and potent optogenetic construct for neuronal inhibition. Further studies may also establish GtACR2nrxn as a robust tool for manipulation of the axonal membrane. These, as well as a variety of other established and emerging transgenic actuators, sensors, and markers, can be selectively expressed in relaxin-3 NI or CeA>NST circuitry, among other precisely defined networks, using the strategies and resources developed here. In doing so, powerful insights into a range of neural systems of both scientific and therapeutic importance can be gained through future research.

Short tandem repeats (STRs) are repetitive DNA sequences composed of repeat units with 2–6 base pairs in length. STRs make up around 3% of the human genome and have higher mutation rate than single nucleotide variants. Mutations in STRs loci have been linked to around 60 genetic disorders known today, the majority of them being neurological. Genotyping STRs has been difficult due to technical limitations but advances in DNA sequencing methods and the emergence of bioinformatic analysis tools have led to the discoveries of many new disease-causing loci. In this thesis, we focused on exploring different methods of STR analysis, their performance on high-throughput sequencing data and improving these existing methods. At first, we compared four genome-wide STR genotyping tools and determined their performance on whole exome sequencing data. We found that all tools have their merits and there is no clear winner. Difference between tools were observed when genotyping some repeat types with some tools having significantly higher error rates on two base-pair repeats. Next, we characterised STRs in whole genome sequencing data to determine genotyping accuracy of different types of repeats in various genomic settings and library preparation methods. We saw differences in accuracy between repeat unit sizes, repeat lengths and repeat composition. This demonstrates the need to use customised filtering strategies for different repeats that would improve quality of calls while retaining as much data as possible, which could be helpful detecting de novo mutations. Finally, we focused on known disease-causing STRs to improve and extend the current method of analysis. A comprehensive list of disease-causing STR loci was collated based on the literature. This was used to create a variant catalogue for a robust genotyping tool ExpansionHunter allowing genotyping of more than twice as many loci than it currently allows. In addition, it is now possible to genotype alleles longer than the fragment length in over forty loci. This catalogue was validated with an extensive series of simulations. We have used this updated catalogue on our sequencing dataset of individuals with rare disease to detect expansions in any of these loci. Moreover, we developed our own computational tool called STRipy, which has been made publicly available, to greatly simplify STR genotyping in sequencing data and enable researchers to genotype the highest number of known pathogenic STR loci compared to other existing methods.

Methamphetamine use is a major health concern globally, with ever-expanding market and an increasing number of users worldwide. In Australia, the number of people with methamphetamine use disorder has increased over the past 10 years, specifically amongst adolescents and young adults. This is a particular concern, as the age of onset of substance use predicts the severity of substance use disorder later in life. In addition, young people are more resistant to treatment. While it is still poorly understood why methamphetamine dependence is increasing in adolescents, literature suggests that methamphetamine use is associated with other psychiatric disorders, deficits in cognition, and certain genes involved in the neurocircuitry of addiction. Therefore, the aim of this thesis was to explore psychiatric, cognitive, and genetic factors associated with early onset of methamphetamine use to gain a holistic understanding of methamphetamine use disorder. To investigate these factors, I conducted a cross-sectional two-group study, recruiting people with a current diagnosis of stimulant use disorder, methamphetamine-type, and controls with no history of substance use disorder. All participants were administered a clinical interview to collect demographic and drug use characteristics. Psychiatric comorbidities and psychotic symptoms were also assessed. Following the interview, participants completed a cognitive task battery assessing attention, speed of processing, cognitive flexibility, working memory, and inhibitory control. Inhibitory control was also assessed using a cue reactivity task that I specifically developed for this project. It consisted of the pseudorandomized presentation of methamphetamine-related cues counterbalanced with control, food-related cues. Upon completion of the study session, whole blood was collected for single nucleotide polymorphism analysis in genes of interest. Genes were selected based on robust preclinical data and results from the meta-analysis presented in this thesis. Poor inhibitory control was identified as an age-dependent risk factor in this thesis, with an earlier age of onset associated with greater deficits. In addition, poor inhibition was associated with an increase in craving upon exposure to methamphetamine-related cues. This is critical as cue-induced craving may lead to relapse after abstinence, and therefore poor treatment outcome. Results from this thesis therefore suggest that pre-existing reduced inhibitory control in adolescence is a risk factor for developing methamphetamine use disorder when methamphetamine is first taken early in life, potentially by perpetuating methamphetamine use and inducing repeated relapses. This thesis also identified comorbid antisocial personality disorder as a strong predictor for age of onset of methamphetamine use. This highlights the need to treat cooccurring mental disorders in young people who use drugs to prevent them from transitioning into problematic use. Lastly, a mutation in the neuregulin-1 gene was associated with early onset methamphetamine use, suggesting that people carrying the gene variant are more likely to develop methamphetamine use disorder when exposed to methamphetamine earlier in life. Taken together, this thesis identified a range of psychiatric, cognitive, and genetic factors associated with early onset methamphetamine use. Findings from this work will contribute to the development of larger studies and clinical trials investigating new early interventions to prevent young people who casually use methamphetamine transitioning into a formal methamphetamine use disorder, thus alleviating the rising burden of disease

Peptide hormone arginine vasopressin (AVP) and its cognate G protein-coupled receptor (GPCR), V1A, belong to the vasopressin-oxytocin signalling system – a highly complex and widespread endocrine system in the human body. AVP activation of V1A causes a stimulatory contractile effect on vascular and uterine smooth muscle, while, in the brain, AVP and V1A play an important role in anxiety, aggression and social behaviours. Due to these physiological mechanisms, AVP and V1A have been of therapeutic interest for decades; initially for the treatment of peripheral conditions such as heart failure and menstrual cramps, while recently, V1A antagonists have shown promise for treating aspects of autism spectrum disorder (ASD). Despite this longstanding interest, there are currently no approved drugs selectively targeting V1A. Subtype-selectivity is important among the vasopressin-oxytocin receptor family, as these receptors share a high degree of sequence similarity and structural homology, yet have differing, and sometimes opposing, effects. Furthermore, as there are currently no experimental 3D structures of V1A, there is a lack of understanding on exactly how AVP binds to V1A at the molecular level, which limits the design and optimisation of new, V1A-selective drugs. GPCR structural biology is a rapidly developing field; however, the low-expression level and instability of many GPCRs, including V1A, can be problematic for purification and subsequent structural study. This thesis aimed to apply various protein engineering techniques to V1A, in order to facilitate the purification and subsequent biochemical study of this important, but long unfulfilled, potential therapeutic target. Chapter 2 covers the first GPCR-engineering venture, which was to introduce expression and stability augmenting missense mutations to V1A via a well-characterised method – directed evolution. This method, which uses E. coli display of receptor mutants, has successfully produced high-expressing, stabilised receptor mutants for the structural study of other neuropeptide receptors, including NTS1 and NK1. However, upon application of this method to V1A, stabilised V1A mutants were not produced. Instead, a V1A truncate was selected, which contained a small section of the V1A N terminus. The reason this protein was selected was unclear – but can primarily be attributed to the issues that arise when using E. coli for GPCR expression. It was evident that further engineering of V1A would require the use a different cell type, such as insect or mammalian cells. Chapter 3 explores the development of a mammalian cell-based directed evolution method. This new method was developed to overcome the expression issues encountered using E. coli, and required optimisation and innovation to accommodate the new cell type, including the use of lentivirus as the gene-delivery method. The method was developed, optimised and applied to V1A – producing V1A mutants that exhibited improved expression levels compared to wild-type V1A. Furthermore, several mutations were identified that appeared in multiple V1A mutants, indicating that they were contributing to this expression-boosting effect. Chapter 4 covers the purification of V1A using a combination of protein engineering approaches – including the introduction of the expression-boosting point mutations discovered in Chapter 3. Ultimately, functional V1A was able to be purified with the aid of these mutations, which provided a substantial increase in functional yield of purified receptor. The two main outcomes of this thesis are: the development of a mammalian cell-based directed evolution method that improves the functional yield of purified receptor – and the successful purification of functional V1A. Firstly, the new, mammalian cell-based selection method presents as a generic, rapid and viable means of improving the heterologous expression of low-expressing GPCRs – and potentially other complex membrane proteins with expression issues. Secondly, the substantial increase the yield of purified, functional V1A is a significant breakthrough in the biochemical study of V1A, as it enables the application of structural biology techniques, such as cryo-electron microscopy (cryo-EM), as well as ligand screening methods for new, V1A-selective ligands.

Early life stress is a known antecedent to anxiety- and fear-related disorders. It may disrupt the ability to regulate fear and act to trigger the development of an anxiety- or fear-related disorder. Hence, in this thesis, I examined the link between early life stress and fear regulation across development. To do this, I used Pavlovian fear extinction as a model of fear regulation to assess the effects of infancy and peri-adolescent stress in rats. In Chapter 2, I examined whether a chronic infancy stressor altered fear extinction in juvenile rats. This series of experiments utilized the limited bedding and nesting model of chronic infancy stress. I found that chronic infancy stress resulted in the precocious emergence of relapse-prone fear in male juvenile rats. However, chronic infancy stress had little effect on extinction behavior in female juvenile rats, as they exhibited relapse-prone fear regardless of stress condition. In Chapter 3, I investigated whether chronic peri-adolescent stress altered fear extinction in adolescence and adulthood. This series of experiments utilized social isolation as a model of chronic peri-adolescent stress. Peri-adolescent stress had a sex-specific effect on fear extinction. In males, peri-adolescent stress resulted in impairments in extinction recall in adolescent and adult male rats. However, in females, peri-adolescent stress had no effect on fear extinction behavior. In Chapter 4, I explored the interplay between peri-adolescent stress and physical activity on fear extinction in adolescence. Peri-adolescent stress, as modelled by social isolation, impaired extinction recall in male adolescents, but this effect was prevented by increased physical activity. Extinction recall in female adolescents was again unaffected by peri-adolescent stress. Surprisingly, increased physical activity was disruptive to extinction recall in peri-adolescent stressed females. Pharmacological suppression of cellular proliferation in peri-adolescent stressed adolescents blocked the effect of physical activity on extinction recall in both sexes. This suggests that peri-adolescent born cells mediate the interplay between peri-adolescent stress and physical activity effects on extinction behavior. Together, these findings highlight sex-specific outcomes of peri-adolescent stress and physical activity on adolescent brain and behavior. In the final Chapter, I propose the Variable Speed Stress model to interpret the effects of stress in infancy and peri-adolescence. Further, I outline potential reasons for the sex effects and possible avenues for future research. Overall, the findings in this thesis contribute to our understanding of how early life stress affects fear extinction throughout development. It suggests that after early life adversity, stress-induced changes to extinction learning development in males may contribute to a reduction in treatment efficacy for exposure-based anxiety disorder therapies.