Such recycling of proteins is natural because actin network dynamics are essential for such processes as growth of axonal filopodia, which are used in searching learn more for growth cone guidance cues (Tessier-Lavigne and Goodman, 1996). The presence of DCC protein in the identified network ( Figure 2 and Figure 3), also suggests an important role of perturbed axonal guidance in autism. Although DCC is also involved in dendrite development ( Suli et al., 2006), this receptor and its signaling protein, netrin, are primarily essential for guiding
axons to their final destinations ( Tessier-Lavigne and Goodman, 1996). Several signaling pathways highlighted in Figure 3, such as the WNT and reelin pathways, also play prominent roles in neuron motility ( Reiner and Sapir, 2005 and Salinas and Zou, 2008). In addition,
several specific proteins, such as PAKs and LIMK, which regulate the dynamics of actin network, are reused in axonal morphogenesis. Consequently, malfunction of many proteins shown in Figure 3 may influence autistic phenotypes through their role in either dendrite or axon signaling, or possibly a combination of these processes. Considering the genes hit by rare de novo variants from the perspective of the functional molecular network (Figure 3) allowed us to investigate the likely morphological consequences of some CNVs. There is growing evidence that changes in dendritic spine morphology contribute to a number of neurological disorders (Halpain et al., 2005). A decrease high throughput screening compounds in the density of dendritic spines in regions of the cerebral cortex has been linked
to schizophrenia (Blanpied and Ehlers, 2004, Garey et al., 1998 and Glantz and Lewis, 2000). On the other hand, an increase in spine size or density has been connected to Fragile X syndrome, a disorder frequently associated with autism (Fiala et al., 2002 and Kaufmann and Moser, 2000). Following the logic that CNV deletions should decrease while duplications increase the dosage of the affected genes, we can infer—based on the structure and regulatory logic of the functional network in Figure 3—the morphological effects of 13 gene perturbations on dendritic spines. Specifically, we found that in 11 out of 13 cases (∼85%) the gene perturbations caused by the observed CNV events should increase either dendritic 3-mercaptopyruvate sulfurtransferase spine growth or their density (see Table S4). This result is consistent with recent findings that autistic individuals have increased spine density in portions of their cerebral cortex (Hutsler and Zhang, 2010 and Woolfrey et al., 2009) and possibly a local brain overconnectivity (Scott-Van Zeeland et al., 2010). Overall, the results of this study, the first to our knowledge, demonstrate that autism-associated rare de novo CNVs, observed in an unbiased genome-wide study, form a large and statistically significant functional network responsible for synaptogenesis, axon guidance, and related molecular processes.