Multiple ASD susceptibility genes converge on cellular
Multiple ASD susceptibility genes converge on cellular pathways that intersect at the postsynaptic site of glutamatergic synapses (Bourgeron, 2015; Peca & Feng, 2012), the development and maturation of synaptic contacts (Gilman et al., 2011) or synaptic transmission (Li et al., 2014). The majority of excitatory glutamatergic synapses are located in small dendritic protrusions known as spines. The formation, maturation and elimination of dendritic spines lie at the core of synaptic transmission and memory formation (Roberts et al., 2010; Yang et al., 2009). Many of the ASD risk genes encode synaptic scaffolding proteins, receptors, cell adhesion molecules or proteins that control mma f receptor cytoskeleton dynamics, all of which directly affect synaptic strength and number and, ultimately, neuronal connectivity in the brain. In addition, ASD risk genes encode proteins that are involved in chromatin remodeling, transcription, and protein synthesis or degradation, all of which can act as upstream controllers of the expression levels of synaptic proteins. Changes in any of these proteins can affect dendritic spine density in the brain. When deleterious mutations occur, an individual's risk for ASD is further affected by how well the defects of a single mutation can be compensated in the brain (Bourgeron, 2015). Increased spine density has been observed in the frontal, temporal, and parietal lobes of human ASD brains (Hutsler & Zhang, 2010) and recent studies indicate a defect in dendritic spine pruning from 13–18years of age (Tang et al., 2014). Tang et al. proposed that this pruning deficit may contribute to abnormalities in the cognitive functions that humans acquire in their late childhood, teenage, or early adult years, including the acquisition of executive skills such as reasoning, motivation, judgment, language, and abstract thinking (Goda & Davis, 2003). Many children diagnosed with ASD reach adolescence and adulthood with functional disability in these skills, in addition to social and communication deficits (Seltzer et al., 2004). Although human ASD studies have reported increased dendritic spine density (Hutsler & Zhang, 2010; Tang et al., 2014), only a few genetic ASD animal models display this effect, making it unclear whether the increased spine density causes ASD-like behavior or whether any change in the number or function of synapse leads to atypical brain connectivity and symptoms of ASD (Bourgeron, 2015). One of the mouse models that does mimic the spine-pruning defect seen in humans has a constitutively overactive mammalian target of rapamycin (mTOR) (Tang et al., 2014). In this model, the mechanism behind the defective spine pruning is an autophagy deficiency in synapses. Autophagy is an evolutionarily conserved cellular process that provides nutrients during starvation and eliminates defective proteins and organelles via lysosomal degradation. Hyperactive mTOR inhibits autophagy at an early step in autophagosome formation (Kim et al., 2011). Inhibition of neuronal autophagy produces ASD-like inhibition of normal developmental spine depletion, without affecting the rate of spine formation, leading to increased spine density and ASD-like behaviors (Tang et al., 2014). This model offers one mechanism underlying increased spine density and ASD-type behavior, but considering the heterogeneity of ASD genetics, it is likely that other mechanisms affecting spine pruning are defective in different ASD patients. Interestingly, it has been reported that mutations of five autism-risk genes with diversified molecular functions all lead to a similar ASD phenotype of behavioral inflexibility, indicated by impaired reversal-learning in Drosophila (Dong et al., 2016). These reversal-learning defects result from an inability to activate Ras-related C3 botulinum toxin substrate 1 (Rac1)-dependent forgetting (Dong et al., 2016). In mammalian neurons, Rac1 affects spine structure, regulates synaptic plasticity in the hippocampus and is required for proper hippocampus-dependent spatial learning (Haditsch et al., 2009; Bongmba et al., 2011). Active Rac1, when targeted to potentiated spines, is able to deplete targeted synapses (Hayashi-Takagi et al., 2015). Rac1 is one of the main regulators of the actin cytoskeleton and, therefore, one of the possible mechanisms underlying the defective spine pruning is deficient regulation of the synaptic actin cytoskeleton. It is also possible that spine pruning is normal, but spine formation is overactive, leading to increased spine density.