La Maladie d’Alzheimer Sporadique (MAS) est à ce jour la cause majoritaire de démence à travers le monde. Cette pathologie est aujourd’hui une question de santé publique avec des conséquences économiques et sociales du fait du vieillissement de la population et du coût de la prise en charge des patients. Les études visant à élucider les mécanismes moléculaires de cette maladie se basent pourtant majoritairement sur des modèles animaux de la forme familiale, qui ne représente que 2% des cas de Maladie d’Alzheimer. La première partie de cette thèse de doctorat a eu pour objectif d’explorer de nouvelles voies de signalisation neuronales pouvant être impliquées dans la MAS. Elle a permis l’identification de nouveaux partenaires synaptiques de la protéine BIN1, dont le gène représente le second locus de susceptibilité pour la MAS après APOE. Dans un second temps, les travaux de cette thèse ont consisté en la caractérisation d’un nouveau modèle murin de cette maladie, après les modèles APOE : une souris mimant la surexpression du gène BIN1 mise en évidence chez les patients. Il a été possible d’inclure la protéine BIN1 dans un interactome synaptique impliqué dans la régulation de deux marqueurs majeurs de la physiopathologie de la MAS : la morphologie des épines dendritiques et la régulation du peptide amyloïde β, et de les relier à une atteinte du cortex entorhinal latéral, la première structure impactée chez les patients. Au total, cette thèse a permis de mettre en évidence des voies de signalisation impliquées dans les phases pré-symptomatiques de la maladie. Ces voies de signalisation sont des cibles thérapeutiques potentielles pour le traitement de la Maladie d’Alzheimer Sporadique. / Late-onset Alzheimer’s Disease (LOAD) is the most common form of dementia and one of the most challenging disease of modern society. Although familial Alzheimer’s Disease (FAD) only account for 2% of the cases, research on the molecular basis of this pathology has mostly been performed on animal models of FAD. The first objective of this doctoral thesis was to identify new pathways involved in LOAD physiopathology. New synaptic partners of BIN1, the second most associated locus with LOAD after APOE, have been characterized. The second objective of this thesis was to study a new mouse model of LOAD, after APOE models: a mouse mimicking the overexpression of BIN1 found in patients. Together, the results place BIN1 in a synaptic interactome involved in the regulation of two major hallmarks of LOAD: dendritic spine morphology and amyloid β peptide regulation. They also link BIN1 with an alteration of the Lateral Entorhinal Cortex, which is the first brain structure affected in patients. Altogether, this doctoral thesis gives new insights on molecular pathways involved in early pathologic features of LOAD, at early presymptomatic stages of the disease, attainable by new therapeutic strategies.

Pas de résumé en français / Posttraumatic stress disorder (PTSD) is a highly heterogeneous condition, ranging from individual traumatic incidents such as car accidents to national tragedies such as natural disasters. Every individual has a different depending on their personality and past experiences, especially regarding their tendency to depression. Hence the condition is better termed psychotrauma spectrum disorder (PSD). Its heterogeneity hinders reliable diagnosis, as detection is entirely dependent upon a clinician’s subjective impression and sensitivity to comorbidities and there is always the possibility of concealment. Yet early diagnosis is essential, as the earlier PSD is detected the more likely treatment will be successful. Furthermore, reliable biomarkers of PSD would allow for much more accurate detection and monitoring of progression. Here we propose a new computational approach building on our prior work on the early detection of Parkinson’s, Alzheimer’s and depression. We will use a new analysis tool, called the Brain Code (BC). This concept was developed to integrate many different kinds of data, for e.g. the often fragmented and incomplete outputs from body sensors that record balance, dexterity, postural, facial and vocal movements combined together with cognitive or clinical outputs such as the intentional or emotive content of speech. The Brain Code allows us to fit all these different data streams together in such a way as to compensate for the deficiencies of each individually. It can put disparate physiological and cognitive data into the same ‘coordinate system’, so that we will be able to develop a reliable quantitative ‘signature’ of PSD. These quantitative biomarkers will be designed so that they are useful for both physicians in a clinical setting and for communities affected by a large-scale traumatic event.

No description available.

Mast cells are immune cells that are found in the brain. Behavioral and endocrine states increase the number and activation of brain mast cells, independent of the animal's immune status. Activation causes the release of many neuro-active mediators into the brain parenchyma. However, the function or impact of mast cells in the brain has not been studied. The recruitment of mast cells to the brain, and their subsequent activation following a stressor suggests that they may have a role in regulating the stress response through interactions with neural systems. The goal of this thesis is to examine the functional role of brain mast cells using a mouse model. Mast cells are present in the mouse brain parenchyma, meninges and choroid plexus from birth throughout adulthood. A mast cell deficient (KitW-sh /W-sh) mouse is a strong model to study the effects of mast cells on brain physiology and behavior. The homozygote mutant lacks all brain mast cells resulting in reductions of mast cell-derived mediators. Interestingly, mast cell deficient mice have increased levels of anxiety-like behavior and stress-induced defecation compared to heterozygote (mast cell competent) littermate controls. Since mast cells are activated by stressors via corticotrophin releasing factor, it is surprising that no differences in the hypothalamic-pituitary-adrenal axis reactivity are seen between mast cell deficient mice and littermate controls. Instead, the effects of mast cells on anxiety behavior and physiology may be mediated through mast cell contribution of serotonin to the hippocampus, a brain region where many mast cells reside. In vitro, application of a mast cell activator to hippocampal slices causes a rise in serotonin levels in the hippocampus of control, but not mast cell deficient mice. Given the known effects of hippocampal serotonin as a trophic factor and transmitter, hippocampal function is likely affected by the absence of mast cells. There are deficits in hippocampal neurogenesis, but not subventricular zone neurogenesis (a brain region with no mast cells), in mast cell deficient mice. This deficit can be reversed by increasing serotonin signaling with SSRI treatment or by enriched housing conditions. Mast cell deficient mice also have deficits in hippocampal dependent spatial learning and memory which can be reversed by enriched housing. Overall these results show that mast cells affect neural systems and behavior in the absence of an immune stimulus. These studies link an immune cell to the brain and behavior, and suggest a beneficial role for the recruitment of mast cells and subsequent neuroimmune interactions.

Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder characterized by a progressive loss of cognition. Histopathologically, AD is defined by the presence of two lesions, senile plaques (SP) and neurofibrillary tangles (NFT), which result from the accumulation and deposition of the amyloid-β peptide (Aβ) and the aggregation of hyperphosphorylated tau protein, respectively. Aβ is formed upon sequential cleavage of the amyloid precursor protein (APP) by β- and γ-secretases and is secreted extracellularly. The accumulation of extracellular Aβ is thought to initiate a pathogenic cascade resulting in synaptic dysfunction in neurons, followed by the their eventual demise through apoptosis. However, while Aβ has been shown to be increased in AD patients' brains, little is known about how the cleavage of APP and the subsequent generation of Aβ is influenced or if the cleavage process changes over time. Moreover, while the effects of Aβ on neurons are known, the exact mechanism remains unclear. Many have postulated that Aβ exerts its effects by binding a putative receptor, but the search for an Aβ receptor has so far remained inconclusive. Interestingly, one of the proposed potential receptor for Aβ is APP itself. In this model, soluble oligomeric Aβ binds cell-surface APP, inducing its dimerization leading to all the downstream effects of Aβ in cells -- e.g. cell death and/or synaptic dysfunction. Moreover, it has been proposed that Aβ can promote its own production in neurons, thereby initiating a pathogenic loop. However, isolating Aβ-induced APP signaling has remained challenging due to the promiscuous nature of Aβ binding. To work around this problem, we used an antibody-mediated approach to artificially trigger the dimerization of cell-surface APP in cells. We found that dimerization of APP could recapitulate all of the effects of oligomeric Aβ in hippocampal neurons, triggering neuronal death at high concentrations and interfering with normal synaptic functions low concentrations. We also found that dimerization of APP is sufficient to promote the amyloidogenic pathway, by increasing levels of the β-secretase BACE1, resulting in increased Aβ production. Finally, we found that dimerization of APP triggered caspase-dependent cleavage of APP and the formation of a second neurotoxic fragment, termed C31, which also mimics the effects of Aβ in hippocampal neurons. Taken together, our data provides support for the occurrence of a positive pathogenic feedback loop involving Aβ, APP and C31 in neurons.

The ASE neurons of C. elegans are an excellent model to study neuronal asymmetry. Lateralization with respect to their genetic fate and function has been well studied, but their more subtle asymmetries have not. This work describes three such asymmetries: that of amino acid gustation, associative learning, and morphological size. In the first two of these, I found a previously uncharacterized asymmetric neuronal response with respect to amino acid gustation, and expand on the known asymmetry with respect to associative salt learning. Most of this thesis focuses on a discovered size asymmetry in the ASE pair of neurons: characterizing it, providing a functional significance, and describing some of its genetic underpinnings. Size asymmetry and the mechanisms of overall neuron growth are not well-studied, but do have functional consequences in higher organisms. This work hopefully furthers our understandings of these processes and of neuronal development in general.

This thesis develops and applies statistical methods for the analysis of neural data. In the second chapter we incorporate a latent process to the Generalized Linear Model framework. We develop and apply our framework to estimate the linear filters of an entire population of retinal ganglion cells while taking into account the effects of common-noise the cells might share. We are able to capture the encoding and decoding of visual stimulus to neural code. Our formalism gives us insight into the underlying architecture of the neural system. And we are able to estimate the common-noise that the cells receive. In the third chapter we discuss methods for optimally inferring the synaptic inputs to an electrotonically compact neuron, given intracellular voltage-clamp or current-clamp recordings from the postsynaptic cell. These methods are based on sequential Monte Carlo techniques ("particle filtering"). We demonstrate, on model data, that these methods can recover the time course of excitatory and inhibitory synaptic inputs accurately on a single trial. In the fourth chapter we develop a more general approach to the state-space filtering problem. Our method solves the same recursive set of Markovian filter equations as the particle filter, but we replace all importance sampling steps with a more general Markov chain Monte Carlo (MCMC) step. Our algorithm is especially well suited for problems where the model parameters might be misspecified.

Individuals with schizophrenia exhibit substantial deficits in their ability to perform working memory (WM) tasks, and these deficits have a critical impact on health and life outcomes for these patients, and may be fundamental to the neurophysiological basis of the disorder itself. However, neuroimaging investigations into the nature of these deficits in these patients over the last decade and a half have been stymied by inconsistent findings that leave no clear answer as to their cognitive or neural basis. One hypothesis that has been proposed to account for these inconsistent findings is that the response of some brain regions subserving WM task performance to parametrically increasing WM load, most critically dorsolateral prefrontal cortex, may in fact be non-monotonic in nature; that is, at sufficiently high loads activation in these regions may begin to decrease. If true, this could account for the inconsistent findings in comparisons of patients with schizophrenia and matched controls, as the two groups may be at different points along this putative activation-load 'inverted-U' curve, resulting in different findings depending on the degree of load utilized in any given study. To date, this hypothesis has not been directly tested; however, I report here the results of a series of studies using the self-ordered working memory task that clearly demonstrate such an 'inverted-U' in healthy participants that is absent in patients with schizophrenia. The pattern of findings in the studies reported here are consistent with healthy individuals switching from WM-mediated strategies to long-term memory-mediated strategies as WM load is increased, while patients with schizophrenia fail to make this switch, instead attempting to utilize WM to subserve task performance even when their WM capacity is exceeded.

The role of adult hippocampal neurogenesis in behavior, especially contextual fear conditioning (CFC), is debated. Several studies demonstrated that blocking adult hippocampal neurogenesis in rodents impairs CFC, while several other studies failed to observe impairment. We sought to determine whether different CFC methods vary in their sensitivity to the arrest of adult neurogenesis. Adult neurogenesis was arrested in mice using low-dose, targeted x-irradiation, and the effects of x-irradiation were assayed in conditioning procedures that varied in the use of a discrete conditioned stimulus, the number of trials administered, and the final level of conditioning produced. We demonstrate that x-irradiation impairs CFC in mice when a single-trial CFC procedure is used but not when multiple-trial procedures are used, regardless of the final level of contextual fear produced. In addition, we show that the x-irradiation-induced deficit in single-trial CFC can be rescued by providing pre-exposure to the conditioning context. These results indicate that adult hippocampal neurogenesis is required for CFC in mice only when brief training is provided. We next sought to determine the age at which adult-born hippocampal neurons contribute to behaviors such as CFC and novel object recognition (NOR). NOR was assessed in mice after neurogenesis was arrested using focal x-irradiation of the hippocampus, or a reversible, genetic method in which glial fibrillary acidic protein-positive neural progenitor cells are ablated with ganciclovir. Arresting neurogenesis did not alter general activity or object investigation during four exposures with two constant objects. However, when a novel object replaced a constant object, mice with neurogenesis arrested by either ablation method showed increased exploration of the novel object when compared with control mice. The increased novel object exploration did not manifest until 4-6 weeks after x-irradiation or 6 weeks following a genetic ablation, indicating that exploration of the novel object is increased specifically by the elimination of 4- to 6-week-old adult born neurons. The increased novel object exploration was also observed in older mice, which exhibited a marked reduction in neurogenesis relative to young mice. Mice with neurogenesis arrested by either ablation method were also impaired in one-trial contextual fear conditioning (CFC) at 6 weeks but not at 4 weeks following ablation, further supporting the idea that 4- to 6-week-old adult born neurons are necessary for specific forms of hippocampus-dependent learning, and suggesting that the NOR and CFC effects have a common underlying mechanism. These data suggest that the transient enhancement of plasticity observed in young adult-born neurons contributes to cognitive functions. Finally, we sought to understand how a memory trace is formed and retrieved in the DG and in CA3, and whether adult hippocampal neurogenesis modulates these events. Our hypothesis is that the cells reactivated during recall of a memory are a component of the memory trace. We have designed a novel tool to test this hypothesis not just in a short time period but also on a longer timescale. We, therefore, created an ArcCreERT2 BAC transgenic mouse that allows for the permanent labeling of cells expressing the IEG Arc/Arg3.1 and allows for a comparison between the cells that are activated during the encoding of an experience and those that are activated during the retrieval of the corresponding memory. To test our hypothesis of the memory trace, we have performed various manipulations that affect hippocampus-dependent memory: 1) contextual differences (fearful versus non-fearful context), 2) time (recent versus remote memory), and 3) arrest of adult hippocampal neurogenesis (x-irradiation and social defeat). We find that levels of reactivation in CA3, both in the presence and in the absence of neurogenesis, are correlated with the strength of the memory, which suggests that the cells reactivated in CA3, are a component of the memory trace.

Increasing evidence suggests that anxiety and depression disorders have neurodevelopmental roots, with a strong case for early-life serotonergic dysfunction in their ontogeny. For instance, disrupting serotonergic tone during critical developmental periods results in dysregulated adult affective behaviors, in mice. However, insight into the mechanism by which early-life serotonin levels impact emotional behaviors in later-life, remains scarce. Based on its potent neurotrophic properties, we hypothesized that serotonin acts during development to guide the maturation of brain regions implicated in the modulation of both the emotional and cognitive features of anxiety and depression disorders, namely the medial prefrontal cortex (mPFC) and hippocampus (HC). To address this hypothesis, we employed a pharmacological mouse model in which serotonin levels were increased using a selective serotonin reuptake inhibitor, fluoxetine (FLX), during a previously established critical post-natal time window (post-natal day 2-11; "PN-FLX" mouse model). The effect of this post-natal serotonergic disruption on the structural, physiological and functional properties of the adult mPFC and HC was assessed. In the mPFC, we found morphological and electrophysiological changes specific to cortical layer 2/3. Pyramidal neurons in layer 2/3 of the infralimbic (IL) sub-region of the mPFC, displayed decreased dendritic arborization, concomitant with reduced intrinsic excitability. Conversely, prelimbic (PL) layer 2/3 neurons displayed normal dendritic arborization, but increased intrinsic excitability. Together, these changes produce a reduced IL/PL output ratio. In order to probe the functional consequences of these changes, we investigated the ability of PN-FLX mice to extinguish learned fear ("fear extinction"), as performance in this behavioral paradigm is modulated by IL/PL output. Congruent with their altered IL/PL balance, PN-FLX mice display deficits in fear extinction learning and recall. Mimicking the diminished IL/PL ratio by selectively lesioning the IL, in control mice, phenocopied some features of the PN-FLX anxiety and depression-behavioral profile, demonstrating a causal relationship between mPFC changes and the emotional phenotype of PN-FLX mice. In the HC, PN-FLX treatment produced retraction of dendritic arbors but increased branching in the cornu ammonis 3 (CA3) sub-field. PN-FLX treatment also reduced total synapse number and synaptic density of CA3 neurons. In order to probe the functional consequences of these morphological changes, we investigated spatial learning in the Morris water maze and contextual fear conditioning in PN-FLX mice, as these behaviors are sensitive to CA3 manipulations. In line with reduced CA3 function, PN-FLX mice displayed deficits in both these learning paradigms. Taken together, our findings demonstrate that serotonin acts during a key developmental epoch to set the structural and physiological properties, and ultimately the functional output, of two key brain regions that underlie affective and cognitive behaviors, the mPFC and HC. Our findings provide a mechanism by which reported genetic (polymorphisms, mutations) and environmental factors (exposure to pharmacological agents or stress) that alter serotonergic tone during development, have long-term, often indelible, consequences on adult affective and cognitive function.