Search

One prominent feature of brain computation is the excitation inhibition balance (E/I balance) that represents one of the main homeostatic functions of the brain. Its aim is to maintain the neural circuits in a narrow and safe range of action. Within this range, the brain network can receive and analyze sensory inputs and produce a modulated output, proportional to the stimuli intensity. Any imbalance in this equilibrium leads to abnormal responses to external stimuli and results in pathological behavior. Indeed, neurological pathologies known for featuring a deep alteration of the E‐I balance are epilepsy and autism, which often occur together in the same patient. Several human genetic syndromes caused by alterations of genes involved in neural development feature signs like autism and epilepsy. Thus, they represent important cases for studying and understanding the role of these single altered genes in the development and regulation of the brain balance. In return, we hope that this knowledge of these genes and more generally of human brain network can be useful in treating the patients affected by these conditions and can help us improve their quality of life. In my work, I studied the regulation of the E/I balance in mouse models of neurological diseases from three different points of view. In the first set of experiments, I studied the E/I balance in a focal model of epileptiform activity. This model is produced by the local application of bicuculline to the mouse cortex. Bicuculline is a competitive GABAA receptor antagonist that, when applied, leads to the development of persistent and periodic interictal spikes at the injection site, while activity appears to be normal in nearby areas that are not reached by bicuculline. In our experiments, we showed that, even in the apparently normal area, there is a disruption of cortical computation. Specifically, the disruption occurs whenever an interictal spike is generated in the epileptic focus. This can have important impact on our understanding of epilepsy and of its treatment since interictal spikes are a common feature not only of epileptic patients, but can also appear in non‐epileptic subject, apparently without any consequence. From our results, we concluded that interictal activity can actually interfere with brain operation not only in the center of the epileptiform activity, but also in the connected areas, where the E/I balance is not directly disrupted. These results provide an example of the fact that apparently non‐symptomatic interictal spikes can affect brain computation. The second experimental model that I studied is a mouse model for a specific human genetic disease: the Phelan‐McDermid syndrome. This is a developmental disease, caused by a genomic deletion at site 22q13. The main suspect for causing the disease is one gene, Shank3, which encodes for a scaffold protein localized in the post‐synaptic density of glutamatergic synapses. In this model, we studied the computation of visual stimuli and we found an alteration of the contrast‐response curve. This is a defining relationship of visual processing: it is the transfer function that converts the visual input into a neural output. This means that to each intensity of visual stimulation corresponds a certain intensity of the neural response, of the visual cortex. We determined that, in Shank3 mutant mice, this curve was altered and showed an increased response to less intense stimuli and showed also a poor modulation of responses to high‐contrast stimulations. An interpretation of this can be that these mice are more sensitive to low‐contrast stimuli, but completely lose the ability of telling apart different high‐contrast stimuli from each other. Therefore, the Phelan McDermid mouse becomes “blinded” by weak stimulations as if they were seeing strong stimulus. Finally, we studied the behavior of the chloride ion in a drug‐induced epileptic seizure model. Chloride ion is of pivotal importance in neurons were the activation of ionotropic GABA and glycine receptors, which increase chloride membrane conductance in response to GABA or glycine release respectively. The intracellular concentration of chloride ions decides what is the effect of GABA release. Traditionally, ionotropic GABA receptors activation was thought to be inhibitory only, but the excitatory or inhibitory nature of these receptors is determined by the intracellular concentration of chloride ions. This concentration in normal adult neurons is thought to be around 5 mM: at this concentration, the effect of the activation of GABA receptors is an inhibition of the postsynaptic element. We investigated if the chloride concentration can be varied under extreme pathologic conditions as during epileptic seizures in a drug induced mouse model. In these animals, the epileptic seizures were produced by local administration of 4‐aminopyridine (4‐AP), a potassium channel antagonist. The effect of 4‐AP is to cause accumulation of chloride ions in neurons and this suggests that, in epileptic crisis, the role of inhibitory neurons can actually favor excitation.

Spreading depression (SD) is a neurophysiological phenomenon characterized by abrupt changes in intracellular ion gradients and sustained depolarization of neurons. It leads to loss of electrical activity, changes in the synaptic architecture, and an altered vascular response. Although SD is often described as a unique phenomenon with homogeneous characteristics, it may be strongly affected by the particular triggering event and by genetic background. Furthermore, SD may contribute differently to the pathogenesis of widely heterogeneous clinical conditions. Indeed, clinical disorders related to SD vary in their presentation and severity, ranging from benign headache conditions (migraine syndromes) to severely disabling events, such as cerebral ischemia, or even death in people with epilepsy. Although the characteristics and mechanisms of SD have been dissected using a variety of approaches, ranging from cells to human models, this phenomenon remains only partially understood because of its complexity and the difficulty of obtaining direct experimental data. Currently, clinical monitoring of SD is limited to patients who require neurosurgical interventions and the placement of subdural electrode strips. Significantly, SD events recorded in humans display electrophysiological features that are essentially the same as those observed in animal models. Further research using existing and new experimental models of SD may allow a better understanding of its core mechanisms, and of their differences in different clinical conditions, fostering opportunities to identify and develop targeted therapies for SD-related disorders and their worst consequences.

Epileptiform activity is associated with impairment of brain function even in absence of seizures, as demonstrated by failures in various testing paradigm in presence of hypersynchronous interictal spikes (ISs). Clinical evidence suggests that cognitive deficits might be directly caused by the anomalous activity rather than by its underlying etiology. Indeed, we seek to understand whether ISs interfere with neuronal processing in connected areas not directly participating in the hypersynchronous activity in an acute model of epilepsy. Here we cause focal ISs in the visual cortex of anesthetized mice and we determine that, even if ISs do not invade the opposite hemisphere, the local field potential is subtly disrupted with a modulation of firing probability imposed by the contralateral IS activity. Finally, we find that visual processing is altered depending on the temporal relationship between ISs and stimulus presentation. We conclude that focal ISs interact with normal cortical dynamics far from the epileptic focus, disrupting endogenous oscillatory rhythms and affecting information processing.

Background: Glioblastomas are largely unresponsive to all available treatments and there is therefore an urgent need for novel therapeutics. Here we have probed the antineoplastic effects of a bacterial protein toxin, the cytotoxic necrotizing factor 1 (CNF1), in the syngenic GL261 glioma cell model. CNF1 produces a long-lasting activation of Rho GTPases, with consequent blockade of cytodieresis in proliferating cells and promotion of neuron health and plasticity. Methods: We have tested the antiproliferative effects of CNF1 on GL261 cells and human glioma cells obtained from surgical specimens. For the in vivo experiments, we injected GL261 cells into the adult mouse visual cortex, and five days later we administered either a single intracerebral dose of CNF1 or vehicle. To compare CNF1 with a canonical antitumoral drug, we infused temozolomide (TMZ) via minipumps for 1 week in an additional animal group. Results: In culture, CNF1 was very effective in blocking proliferation of GL261 cells, leading them to multinucleation, senescence and death within 15 days. CNF1 had a similar cytotoxic effect in primary human glioma cells. CNF1 also inhibited motility of GL261 cells in a scratch-wound migration assay. Low dose (2 nM) CNF1 and continuous TMZ infusion significantly prolonged animal survival (median survival 35 days vs. 28 days in vehicle controls). Remarkably, increasing CNF1 concentration to 80 nM resulted in a dramatic enhancement of survival with no obvious toxicity. Indeed, 57% of the CNF1-treated animals survived up to 60 days following GL261 glioma cell transplant. Conclusions: The activation of Rho GTPases by CNF1 represents a novel potential therapeutic strategy for the treatment of central nervous system tumors.