Abstract Autism spectrum disorders (ASD) area group of neurodevelopmental disorders that affect up to 1.5% of population in the world. Recent largescale genomic studies show that genetic causes of ASD are very heterogeneous. Gene ontology, pathwayanalysis and animal model studies have revealed several potential converging mechanisms including postsynaptic dysfunction of excitatory synapses. In this review, we focus on the structural and functional specializations of dendritic spines, and describe their defects in ASD.We use Fragile X syndrome, Rett syndrome and Phelan-McDermid syndrome, three of the most studied neurodevelopmental disorders with autism features, as examples to demonstrate the significant contribution made by mouse models towards the understanding of monogenic ASD. We envision that the development and application of new technologies to study the function of dendritic spines in valid animal models will eventually lead to innovative treatments for ASD.

Abstract The quest to find novel therapeutics for mental and neurological disorders has been hindered by the lack of access to live human brain samples and relevant experimental models. Conventional 2D human pluripotent stem cell-derived neuronal cultures and animal models do not fully recapitulate many endogenous human biochemical processes and diseasephenotypes. Currently, the majority of candidate drugs obtained from preclinical testing in conventional systems does not usually translate into success and have a high failure rate in clinical trials. Recent advancements in bioengineering and stem cell technologies have resulted in three-dimensional brain-like tissues, such as oragnoids,which better resemble endogenous tissue and are more physiologically relevant than monolayer cultures. These brain-like tissues can bridge the gap between existing models and the patient, and may revolutionize the field of translational neuroscience. Here, we discuss utilities and challenges of using stem cell-derived human brain tissues in basic research and pharmacotherapy

Abstract Multiple sclerosis (MS) is characterized by neurological symptoms that are separated in time and space, which correlate with demyelination and white matter lesions. The conventional pathophysiological model is that an autoimmune reaction against the myelinated nerve sheath results in demyelination, accompanied by axon damage and the death of oligodendrocytes that produce myelin. There is no cure for MS, but current treatments are primarily aimed at suppressing the autoimmune reaction, with the goal of reducing demyelination. These treatments have limited efficacy and developing better treatments for MS remains an important goal. Here we argue that the autoimmune reaction may be secondary to neurodegeneration or neurotoxicity, and that protecting neurons from glutamate-mediated toxicity may be a better therapeutic strategy than targeting the immune system. We have recently demonstrated that a protein-protein interaction between the GluR2 subunit of the AMPA (α-Amino-3-hydroxy-5-methyl-4- isoxazolepropionic acid) glutamate receptor and GAPDH (Glyceraldehyde 3-phosphate dehydrogenase) is elevated in human MS plaques and in an animal model of MS. Disrupting this interaction in a rodent model restores neurological function, preserves myelin, and protects neurons, oligodendrocytes and axons. The peptide we created to block the GluR2-GAPDH interaction also reduces immune system activation, suggesting that autoimmunity is not necessarily the primary etiology in MS. The GluR2-GAPDH interaction may promote cell death via increased calcium influx through non-GluR2-containing AMPA receptors, or through the p53 and Siah1 cell death pathways.

Neurological and psychiatric disorders collectively constitute the greatest burden of disease. However, the human brain is the most complex of biological systems and therefore accurately modeling brain disorders presents enormous challenges. A large range of therapeutic approaches across a diverse collection of brain disorders have been found to show great promise in preclinical testing and then failed during clinical trials. There are a variety of potential reasons for such failures, on both the preclinical and clinical sides of the equation. In this article, I will focus on the key issues of validity in animal models. I will discuss two forms of construct validity, ‘genetic construct validity’ and ‘environmental construct validity’, which model specific aspects of the genome and ‘envirome’ relevant to the disorder in question. The generation of new gene-edited animal models has been facilitated by new technologies, the most notable of which are CRISPR-Cas systems. These and other technologies can be used to enhance construct validity. Finally, I will discuss how face validity can be optimized, via more sophisticated cognitive, affective and motor behavioural tests, translational tools and the integration of molecular, cellular and systems data. Predictive validity cannot yet be assessed for the many preclinical models where we currently lack effective clinical interventions, however this will change as the translational pipeline is honed to deliver therapies for a range of devastating disorders.