Autism spectrum disorders (ASDs) are a heterogeneous group of neurodevelopmental disorders with shared symptoms in the area of communication and language, restricted interests, and stereotyped and social behaviors. Causes lie in perturbations of brain development, which can be manifold, but genetic factors are prominent among these. Genetic studies have pointed to hundreds of causative or susceptibility genes in ASD, making it difficult to find common underlying pathogenic mechanisms. Careful dissection of molecular and cellular mechanisms are needed to define the molecular targets that can translate into therapeutic strategies. On page 1199 of this issue, Bidinosti et al. (1) uncover defects in a molecular machinery of a genetic ASD mouse model. This allowed the authors to design specific chemical interventions that relieve cellular and behavioral autistic-like features. In addition, Yi et al. (2) report a channelopathy in neurons that may predispose to autism. The discoveries raise hope for developing new drugs that help patients with ASD.
Current thinking about pharmacological therapies for ASDs has been stimulated by two scientific milestones. One major advance has been the large number of genes associated with risk for autism (from human genetic studies). This has extended the clinical notion that ASDs include heterogeneous conditions ranging from severe intellectual disability to high-functioning forms (3). Furthermore, identified gene variants in ASDs all appear to be rare, and recurrence is very low (less than 1%) in sporadic cases. However, a number of syndromes with autistic-like features—in addition to fragile X and Rett syndromes—have been recognized. One of these is the Phelan-McDermid syndrome.
Considering the human genetics of ASDs, the spectrum of properties of proteins encoded by ASD genes can be aggregated in a number of molecular and cellular functions. Thus, protein synthesis and degradation, signal transduction, transcription, and synaptic transmission emerge as major cellular processes from which ASDs may originate (3, 4). These processes are not independent of each other. For example, transcription, translation, and degradation together control the quantity and quality of the total pool of proteins of the cell. Signal transduction couples extracellular chemical signals, such as neurotransmitters and growth factors, to intracellular responses including protein synthesis and degradation, and transcription. These are all essential activity-dependent pathways that remain highly dynamic in adult stages. At an integrated level, these cellular pathways are apparent in biological functions relevant for ASDs, in particular synaptogenesis, axon guidance, dendritic and spine morphology, and synaptic plasticity. This has led to the hypothesis that abnormal synaptic homeostasis could play a key role in the pathogenesis of ASDs (3, 4).
The other milestone in the field is the notion that neurodevelopmental defects are not necessarily permanent, but may be reversible. There has been a long-standing view that neurodevelopmental disorders are congenital inborn errors of brain development that leave the patient with irreversible defects. This traditional view was first challenged by the reactivation of a silenced gene encoding methyl CpG-binding protein 2 (MeCP2) in a mouse model of Rett syndrome (5). Induction of Mecp2 expression dramatically reversed behavioral and electrophysiological abnormalities in developing and adult mice. Selective reversal of abnormalities was also observed in other ASD models. For example, phenotypes in mice lacking the gene encoding the protein tuberous sclerosis 1 (TSC1) could be reversed by the small molecule rapamycin. Rapamycin blocks mammalian/mechanistic target of rapamycin complex 1 (mTORC1), which controls protein synthesis. The TSC1-TSC2 complex controls mTORC1 activity (6). In mouse models of fragile X syndrome [mice that lack the gene encoding fragile X mental retardation protein 1 (FMRP1)], treatment with an antagonist of the metabotropic glutamate receptor 1/5 class (mGluR1/5) also reversed disease characteristics (6). Signaling by mGluR1/5 is coupled to synaptic response involving FMRP1. Moreover, insulin-like growth factor I has been successfully used to ameliorate autistic-like phenotypes in mouse models of Rett syndrome and Phelan-McDermid syndrome. SHANK3 is the prime gene culprit causing the latter disorder. Interestingly, selective rescue of autistic-like phenotypes in a mouse model was established by reexpression of Shank3 (7).
SHANK3 is a synaptic scaffolding protein in the postsynapse that connects receptors and ion channels in the membrane with intracellular signaling proteins and downstream processes (see the figure). Yi et al. propose that through direct interaction, SHANK3 may enrich hyperpolarization-activated cyclic nucleotide–gated channels at postsynaptic sites. SHANK3(haplo)deficiency severely impaired hyperpolarization-activation (Ih) current conductance, explaining increased input resistance, a neuronal phenotype in Phelan-McDermid syndrome. Bidinosti et al. generated cells with a genetic defect reminiscent of SHANK3 variants seen in Phelan-McDermid syndrome and sporadic ASD, and encountered a deregulated pathway that has been implicated in other forms of ASD. The AKT-mTORC1 signaling pathway is a hub for many cellular processes and is down-regulated as a consequence of Shank3 deletion in mice. This is opposite of the effect of several other ASD gene mutations on the AKT-mTORC1 pathway. Apparently, an imbalance in this pathway in either direction can elicit autistic-like features. Bidinosti et al. discovered that the down-regulation involves a cascade of events tracing back to an increase in Cdc2-like kinase (CLK2); this is attributed to reduced CLK2 degradation by the ubiquitin-proteosome pathway. How mutated Shank3 affects ubiquitination remains unclear. This may result from a loss-of-function of SHANK3 protein, or perhaps a gain-of-function of other SHANK3 isoforms as a consequence of genetic interference. An intriguing speculation is that it relates to Ih-channel impairment.
The findings of Bidinosti et al. suggest that small molecules that activate AKT or inhibit CLK2 may be used to adjust the activity of a critical signaling pathway in ASDs. Indeed, Bidinosti et al. reversed abnormalities at the molecular level (AKT phosphorylation) and cellular level (density of dendritic spines; miniature excitatory postsynaptic currents) with such compounds in Shank3-deficient neurons, and also reversed abnormal social behaviors inShank3-deficient mice. These are important proofs of principle for drug targets to be taken further in the direction of drug development.
Previously, mGluR1/5 antagonists that successfully rescued phenotypes in genetic animal models of fragile X syndrome had disappointing results in patients with the disorder (8). Other candidate compounds are in queue to take this translational route, like compounds related to the mTORC1-inhibitor rapamycin. Bidinosti et al. add new targets to intervene with the pathogenesis of ASDs. The decade to come will show whether this finding can reach patients.