Twin and familial studies reveal that autism spectrum disorder (ASD) traits are highly heritable.
The genetic landscape of ASD is made of common and rare variants and can be different from one individual to another.
Most of the ASD-risk genes are involved in chromatin remodelling, regulation of protein synthesis and degradation, or synaptic plasticity.
In cellular and animal models, mutations in the ASD-risk genes lead to a distortion of typical neuronal connectivity by decreasing or increasing synapse strength or number.
Compensatory mechanisms, such as genetic buffering and synaptic homeostasis, could modulate the severity of these mutations.
Genetics studies of autism spectrum disorder (ASD) have identified several risk genes that are key regulators of synaptic plasticity. Indeed, many of the risk genes that have been linked to these disorders encode synaptic scaffolding proteins, receptors, cell adhesion molecules or proteins that are involved in chromatin remodelling, transcription, protein synthesis or degradation, or actin cytoskeleton dynamics. Changes in any of these proteins can increase or decrease synaptic strength or number and, ultimately, neuronal connectivity in the brain. In addition, when deleterious mutations occur, inefficient genetic buffering and impaired synaptic homeostasis may increase an individual's risk for ASD.
Autism spectrum disorder (ASD) is an early onset neuropsychiatric disorder that is characterized by impaired social communication, restricted interests and stereotyped and repetitive behaviours1,2,3. These traits exhibit a wide, continuous distribution both in the general population and in individuals who have been diagnosed with ASD4,5,6, and epidemiological studies estimate that more than 1% of the world's population could receive a diagnosis of ASD7,8. The core symptoms of autism rarely emerge in isolation; they usually coexist with other psychiatric and medical conditions, including intellectual disability (ID), epilepsy, motor control difficulties, attention-deficit hyperactivity disorder (ADHD), tics, anxiety, sleep disorders and gastrointestinal problems9,10. ASD affects more males than females; this is especially true among individuals with a normal intelligence quotient (>5 males/1 female affected) but less so in populations with ID (2 males/1 female affected), and the ratio becomes more balanced if patients present with dysmorphic features11. Individuals with ASD also exhibit alterations in sensory processing, including difficulties in the integration of information across different sensory modalities12,13.
The brain regions that are involved in ASD remain difficult to determine, and the neural mechanisms that underlie the altered social communication and integration of sensory information in this disorder remain largely unknown. Abnormal brain growth and connectivity between regions (for example, too many short-range and too few long-range neuronal connections) have been proposed as a cause of ASD14,15, but these findings remain controversial and could be due to methodological artefacts16,17. To better understand the pathophysiology of ASD, much research has focused on identifying genetic mutations associated with ASD and studying their effects in animal models.
A growing number of the genes that have been associated with ASD seem to have a role in synaptic plasticity18,19,20,21,22,23,24. Indeed, animal studies have shown that many ASD-risk genes are regulated by neuronal activity and that their gene products modulate synaptic strength or number18,25,26,27,28,29. It was therefore hypothesized that mutations could lead to the distortion of typical neuronal connectivity30, increasing the risk of ASD26,31 and the difficulties in adapting to external stimuli, such as those received during social interactions27,28,29.
In this Review, I first depict the complex genetic architecture of ASD and emphasize the crucial role of both rare and common genetic variants in this disorder. I then examine how ASD-linked mutations might affect synaptic plasticity at different levels. Finally, I reassess and expand the synaptic homeostasis hypothesis of ASD27,28, building on recent data. I propose that genetic mutations will either reduce or increase the synapse strength or number through the pathways that are typically induced by neuronal activity and sensory inputs. When compensatory mechanisms such as synaptic homeostasis fail, mutations could induce atypical brain connectivity and symptoms of ASD.
The genetic architecture of ASD
Before summarizing epidemiological and molecular genetic studies of ASD, it is important to consider our current understanding of human genome diversity32,33,34. An individual carries, on average, 3 million genetic variants that differ from the reference human genome sequence. These variants are mainly single-nucleotide variants (SNVs), short insertions and deletions and copy-number variants (CNVs). The vast majority of these variants (>95%) are shared with more than 5% of the population worldwide (the so-called common variants). Approximately 130,000 of the variants (∼4% of the genome) are shared with only 0.5–5% of the population (these are known as rare variants), and less than 1% of the variants are either unique to the individual or shared by a very small number of relatives32,33,34. Finally, any one individual carries, on average, 18–74 de novo SNVs in their genome, including 1–4 located in the exons of their genes — these are mutations that occurred either early during development or in the germinal cells of the individual's parents (in most cases, the father)20,21,35.
Remarkably, the range of mutations that affect protein sequences is considerably different today from the range that existed as recently as 200 to 400 generations ago33. Of the putatively deleterious protein-coding variants, the vast majority (>86%) are rare and arose in the last 5,000 to 10,000 years33. It is estimated that each individual carries, on average, 40–110 variants that are classified by the Human Gene Mutation Database as disease-causing mutations32. Thus, carrying a disease mutation is not the exception but the rule. The central question for epidemiologists and molecular geneticists is therefore: how are low-, medium- and high-risk mutations distributed in the population and how can certain individuals remain unaffected despite carrying deleterious ones? As described below, the genetic architecture of ASD is heterogeneous. In one subset of patients, a single fully penetrant mutation can be enough to cause ASD, but in another an accumulation of many low-risk alleles (>1,000) could be the origin of this disorder36,37.
Epidemiology and heritability of ASD. Social skills and general intelligence, as measured by psychological and cognitive tests, are quantitative traits that show a normal distribution in the general population4,5,6. Individuals diagnosed with ASD or ID are considered to be at the lower end of this distribution (Fig. 1a). However, the demarcation between affected and non-affected individuals based on these traits is difficult to draw4. In addition, although people with clinical ASD and individuals with autistic-like traits differ from each other in the severity and/or degree of functional impairment, twin studies have shown that they could share the same susceptibility genes38,39. The relative influence of genetic, epigenetic and environmental factors on the risk of ASD remains under debate, but all epidemiological studies point to there being a strong genetic contribution to the risk of ASD40,41,42. In all twin studies conducted to date on ASD, monozygotic twins have higher concordance rates for ASD (ranging from 60% to 90%) than dizygotic twins have (from 0% to 30%). The estimated heritability of ASD differs from one study to another, but genetic factors always account for at least 38% and up to 90% of the phenotypic variance40,41,42. In families with a history of ASD, the likelihood of having a child with ASD increases with the proportion of the genome that the child shares with one affected sibling or parent41,43,44 (Fig. 1b). In a population-based sample of 14,516 children diagnosed with ASD41, the relative risk (compared with that for the general population) for ASD was estimated to be 153.0 (95% confidence interval (CI): 56.7–412.8) for monozygotic twins, 8.2 (95% CI: 3.7–18.1) for dizygotic twins, 10.3 (95% CI: 9.4–11.3) for full siblings, 3.3 (95% CI: 2.6–4.2) for maternal half siblings, 2.9 (95% CI: 2.2–3.7) for paternal half siblings and 2.0 (95% CI: 1.8–2.2) for cousins. In this study, ASD heritability was estimated to be 0.50 (95% CI: 0.45–0.56) and the non-shared environmental influence was also 0.50 (95% CI: 0.44–0.55). Surprisingly, both shared environment and dominance components seemed to have only a small role in the heritability of ASD41. Twin studies further revealed that a substantial proportion of the genetic contribution to ASD was shared with other neurodevelopmental disorders, such as ADHD (>50%) and ID (>40%)40,45,46,47.
In summary, epidemiological studies provide crucial information on the genetic contribution to ASD. However, they do not inform us on the genes involved or on the number and frequency of their variants. In the past 15 years, candidate-gene and whole-genome analyses have been performed to address these questions.
The molecular genetics of ASD. Based on studies using quantitative molecular genetics48, the proportion of ASD explained by common genotyped single-nucleotide polymorphisms (SNPs) has been estimated to be between 17% and 60%49,50,51. The contribution of common variants to these disorders is therefore important and should not be neglected. Unfortunately, the causative SNPs remain unknown, as they are numerous (>1,000) and each is associated with only a low risk. To date, the largest genome-wide association studies (GWASs) performed on <5,000 families with ASD were not sufficiently powered to identify a single SNP with genome-wide significance52,53. Hence, the majority of what we know about the genes for ASD has come from studies that identified de novo, rare and highly penetrant variants in clinical cases (Supplementary information S1 (table))36,37.
Indeed, in 10–25% of individuals with ASD, a single genetic alteration (for example, a chromosomal rearrangement, CNV, short insertion, short deletion or SNV) seems to be penetrant enough to cause the main symptoms of these disorders. Overall, 5–15% of individuals with ASD carry a de novo CNV versus 1–2% in the general population54,55,56. This enrichment in CNVs applies to both deletions and duplications. Interestingly, in some cases, both deletions and duplications of the same locus could increase the risk of ASD, illustrating the importance of gene dosage in ASD27. The majority of causative CNVs are observed in only one individual, but some loci (for example, loci at chromosomal regions 15q11-q13, 16p11 and 22q11) are recurrently found to be deleted or duplicated. These recurrent deletions or duplications are often caused by recombination events occurring in the genomic regions made of low-copy repeats that flank the CNV54,55,56. In some individuals with ASD, more than one deleterious CNV has been identified, suggesting a 'multiple-hit model' for the disorders in these people56,57,58,59. For example, in 2,312 children with neurodevelopmental disorders known to carry a CNV associated with ID and congenital abnormalities, 10% carried a second large CNV58. No parental transmission bias was observed for the primary de novo or inherited CNV, but 72% of the second CNVs were inherited from the mother58.
Large-scale exome-sequencing studies have found only a slightly elevated rate of de novo SNVs in ASD probands compared with their unaffected siblings (1.02 versus 0.79 mutations per offspring), but the probands had two- to threefold more disruptive de novo mutations than their siblings or than the expected probability to observe such de novo events19,20,21,22,23,24. Interestingly, de novo mutations are not restricted to sporadic cases of ASD, they also seem to contribute to the risk of ASD in families with two affected children60. The de novo SNVs arise three times as often in the paternal background, and mutation rates rise with age of either parent20,21,35, whereas CNVs originated on the paternal and maternal chromosomes with equal likelihood54,55,56. Interestingly, some regions of the genome are more likely to harbour mutations than other regions (by 100-fold), and genes associated with ASD seem to fall into such regions of hypermutability61. This hypermutability of ASD-risk genes was not observed in controls from ASD exome-sequencing studies, and the mutations identified in ASD are markedly enriched in brain-expressed exons that are, from an evolutionary perspective, under negative selective pressure62.
Importantly, the de novo loss-of-function mutations identified in individuals with ASD were not distributed equally in males and females and across the whole spectrum of ASD. They were more enriched in females than males and they were frequently associated with comorbidities, such as ID or epilepsy22,23,24. In individuals with normal IQ or Asperger syndrome, causative de novo mutations are rarely found63,64,65. Hence, as previously suggested, these rare and penetrant variants might not directly cause ASD, but they might 'reveal' autism by decreasing an individual's IQ and hence their ability to compensate for the effects of a high-risk genetic background for ASD66. Indeed, as pointed by Skuse, low levels of autistic vulnerability are more likely to lead to a recognizable ASD syndrome among individuals with ID66. Consequently, the apparent association between ID and ASD would not be because they usually have common causes but rather because the presence of both features greatly increases the probability of clinical ascertainment66.
In summary, common, rare and de novo mutations all contribute to the genetic risk of ASD. The ASD-risk genes (including the level of confidence for the associations with ASD) can be found at the Simons Foundation Autism Research Initiative or at the AutismKb websites. Overall, none of the genes identified so far accounts for more of 1% of all cases of ASD36,67. The number of genes responsible for 'monogenic' forms of ASD (for example, those caused by a de novo deleterious mutation) is estimated to be more than 400 (Refs 19,20,21,68), and these forms of the disorder account for 10–20% of all ASD cases. Monogenic autosomal recessive forms of ASD could account for an additional 3–6% of the cases69. Altogether, karyotypes, CNV analysis and exome sequencing can detect a genetic cause for ASD in almost 25% of the cases. We can expect that whole-genome sequencing will increase the number of cases with a genetic cause of ASD by identifying deleterious mutations that are not screened for yet, such as those affecting regulatory elements of the genes located outside the coding regions35,61,70.
The interplay between common and rare variants in ASD. The epidemiological and molecular data summarized above suggest that the genetic contribution to ASD may be shaped by a combination of rare deleterious variants and a myriad of low-risk alleles (also defined as the genetic background made of common SNPs). The interplay between rare deleterious variants and common low-risk alleles will therefore influence the phenotypic diversity observed in the population. Depending on the ability of the common low-risk alleles to buffer the impact of rare deleterious variants, a very different phenotype can be produced. A genetic background with a strong genetic buffer will cope with the effect of rare genetic variations by canalizing the phenotypic outcome71,72 (Box 1). By contrast, a genetic background with low level of genetic buffering in the child could reveal the effect of mutations that were kept silent in the parents. In this rare deleterious variants versus common low-risk alleles model, we can illustrate three extreme situations (Fig. 2).
In the first situation, an individual has a high genetic buffer and therefore is at low risk for ASD, even if they have a relatively high burden of rare deleterious ASD-risk genes (Fig. 2a). In this situation, only a highly penetrant mutation appearing de novo in the proband would be able to cause ASD (Fig. 2b). This model might account for 10–25% of ASD cases, especially those with ID. Indeed, as previously indicated, de novo truncating mutations are observed in the majority of individuals with ID19,20,21,22,23,24. In the second situation, an individual has a genetic background that acts as a moderate genetic buffer for ASD. As a consequence, they will develop ASD when receiving only a relatively moderate burden of rare mutations (Fig. 2a). This model could account for the observation of multiple hits in some patients with ASD56,57,58,59 (Fig. 2b).
Finally, in the third situation, the individual inherits a genetic background with a very low genetic buffer and is therefore at high risk for ASD. In this case, there is no need of a burden of rare deleterious mutations (Fig. 2a); the accumulation of a high load of low-risk alleles will be enough to cause ASD (Fig. 2b). This model might account for families in which parents have subclinical autistic traits4.
In summary, the genetic landscape of ASD varies greatly between individuals. However, as discussed below, a majority of the mutated genes converge on specific biological pathways that influence synaptic plasticity and connectivity at different levels.
Synaptic plasticity in ASD
Synaptic plasticity is the property of synapses to strengthen or weaken in response to changes in both the amplitude and the temporal dynamics of neuronal activity. Sensory inputs and intrinsic brain activity can effect long-term changes in synaptic efficacy and eventually increase or decrease neuronal connectivity by modulating the number of synapses. Many mutations associated with ASD are predicted to influence the structure and the turnover of synapses at different levels (Box 2) because they encode proteins involved in chromatin remodelling and transcription, protein synthesis and degradation, actin cytoskeleton dynamics or synaptic transmission (Supplementary information S2 (figure)). Below, I discuss the involvement of ASD-risk genes in these biological processes in vitro and in animal models of ASD.
Chromatin remodelling and transcription. Mutations in genes encoding key regulators of chromatin remodelling and gene transcription (for example, methyl-CpG-binding protein 2 (MECP2), myocyte-specific enhancer factor 2C (MEF2C), histone deacetylase 4 (HDAC4), chromodomain-helicase-DNA-binding protein 8 (CHD8) and catenin β1 (CTNNB1)) have been reported in individuals with ASDs (Supplementary information S1 (table)). Interestingly, a subset of these genes is regulated by neuronal activity and influences neuronal connectivity and synaptic plasticity73,74,75.
An example of a relevant protein involved in chromatin remodelling is MECP2. Females with a deletion of, or point mutation in, the X-linked MECP2 gene display Rett syndrome76, whereas boys or girls with duplication of MECP2 develop an ASD and ID77 (Supplementary information S1 (table)). MECP2 is a transcription repressor that coats almost the entire genome78. A lack of MECP2 causes a modification of the global chromatin structure, with enhanced histone acetylation levels79 and elevated transcriptional noise80 (Fig. 3a). Interestingly, MECP2 was shown to act as a transcriptional repressor of long genes81. In response to activity, MECP2 is phosphorylated, which allows or facilitates long-gene transcription. In the absence of MEPC2, these genes are derepressed81. MECP2 should therefore be seen not as a single-gene regulation complex, but more as a genome-wide constituent of the neuronal chromatin that facilitates neuronal activity-dependent gene, especially long gene, regulation73,81. Post-mortem brain samples from patients with Rett syndrome and MECP2-deficient mice show postsynaptic morphological defects such as reduced dendritic branching, reduced dendritic spine density and defects in spine morphology78. Conversely, increasing MECP2 level in neuronal culture was associated with greater spine density82.
An example of a relevant protein involved in transcriptional regulation is MEF2C. De novo MEF2C mutations cause a severe form of ASD83 (Supplementary information S1 (table)). MEF2C is a member of the MEF2 transcription factor family and a negative regulator of synapse number84 (Fig. 3b). In mice, haploinsufficiency of MEF2C is associated with increased synaptic density and impaired hippocampus-dependent learning and memory, whereas expression of a superactivating form of MEF2C results in a reduction of the number of excitatory postsynaptic terminals without changes in learning and memory performance84. Interestingly, in a mouse model of fragile X syndrome, the process of MEF2C-mediated synapse elimination is repressed85. Individuals with mutations in MEF2C and people with fragile X syndrome could therefore present with a similar increase in the level of synaptic density that is caused by a common mechanism85.
Other transcription factors encoded by ASD-risk genes, such as CHD8 and CTNNB1, contribute to early brain development19,20,86,87,88. CHD8 is one of the major risk genes for ASD and is identified in 0.2% of patients (mostly those with ID). This transcription factor acts as a negative regulator of the WNT signalling pathway by repressing CTNNB1 activity (Fig. 3c). Mutations in CHD8 and CTNNB1 have opposite effects on brain development: individuals with CHD8 mutations have larger heads than members of the general population, whereas people with mutations in CTNNB1 have microcephaly87,88. These developmental abnormalities are consistent with the observation that zebrafish with reduced CHD8 expression have a larger brain88 and that transgenic mice overexpressing CTNNB1 in neural precursors have increased cerebral cortical surface area and folding89. In this situation, the change of the head size is probably due to a change in the proliferation of neuronal progenitors in the brain88,89. However, it is important to note that, in contrast to the severe macrocephaly observed in patients with phosphatase and tensin homologue (PTEN) mutations, probands carrying de novo CHD8 mutations have larger heads than members of the general population but they are not always very different from their siblings when adjusting for weight, height and parental head circumference16.
Protein synthesis and degradation. Levels of synaptic proteins can be influenced by neuronal activity through global and local synaptic mRNA translation90. Indeed, glutamate and brain-derived neurotrophic factor (BDNF) induce a cascade of phosphorylation of the mammalian target of rapamycin (mTOR) and fragile X mental retardation protein 1 (FMRP; also known as FMR1 protein) pathways, leading to an increase in mRNA translation (Fig. 4). The newly translated proteins then help to locally increase synaptic strength or number. Many genes involved in such activity-driven regulation of synaptic proteins have been found to be mutated in individuals with ASD (Supplementary information S1 (table))91.
The mTOR pathway controls global mRNA translation (Fig. 4), and its deregulation not only causes diseases associated with increased cell proliferation and loss of autophagy, including cancer90, but also increases the risk for ASD (Supplementary information S1 (table)). Interestingly, mutations in genes encoding repressors of the mTOR pathway, such as neurofibromin (NF1), PTEN and synaptic RAS GTPase-activating protein 1 (SYNGAP1), cause an increase in translation in neurons and at the synapse26 (with the exception of tuberous sclerosis 2 (TSC2) mutations, which lead to reduced protein synthesis in the hippocampus). This hyperactivation of mTOR leads to impaired autophagy and to an increase in dendritic spine density in layer V of the temporal lobes, as seen in post-mortem brains from patients with ASDs92. These reductions in autophagy and developmental spine pruning were suggested to be the origin of the macrocephaly that is present in the majority of the individuals with such mutations93.
Mutations affecting the FMRP–eukaryotic translation initiation 4E (EIF4E)–cytoplasmic FMRP-interacting protein 1 (CYFIP1) complex cause fragile X syndrome and increase the risk of ASD94 (Supplementary information S1 (table)). This protein complex controls local translation of mRNA at the synapse and acts downstream of the RAS–extracellular signal-regulated kinase (ERK) signalling pathway (Fig. 4). As for mutations affecting the mTOR pathway, mutations affecting the FMRP–EIF4E–CYFIP1 complex were shown to increase mRNA translation instead of decreasing it. Interestingly, inhibition of RAS or ERK activity corrects the excessive protein synthesis observed in the fragile X mouse model, whereas rapamycin, an mTOR inhibitor, does not. The FMRP–EIF4E–CYFIP1 complex regulates the translation of more than 1,000 genes, many of which are ASD-risk genes95,96,97,98. An alteration in the level of expression of the FMRP–EIF4E–CYFIP1 complex should therefore create an imbalance in the level of many synaptic proteins that are associated with ASD.
The ubiquitin–proteasome system (UPS) is central to the degradation of proteins and, consequently, for the regulation of synapse composition, assembly and elimination99 (Fig. 4). The gene encoding ubiquitin–protein ligase E3A (UBE3A) is mutated in patients with Angelman syndrome and is duplicated on maternal chromosome 15q11 in some individuals with ASD (Supplementary information S1 (table)). Neuronal activity increases UBE3A transcription through the MEF2 complex and regulates excitatory synapse development by controlling the degradation of activity-regulated cytoskeleton-associated protein (ARC), a synaptic protein that decreases long-term potentiation (LTP) by promoting the internalization of AMPA receptors (AMPARs)100. Disruption of UBE3A function in neurons leads to an increase in ARC expression and a concomitant decrease in the number of AMPARs at excitatory synapses. However, it is not clear whether a duplication of UBE3A (as observed in patients with ASD and duplication of chromosome 15q11) could lead to the inverse synaptic phenotype (a decrease in ARC expression and an increase in AMPARs at the membrane).
In summary, although there are exceptions, most of the mutations that have been identified in the mTOR, the FMRP–EIF4E–CYFIP and the UPS pathways result in an abnormally high level of synaptic proteins and an increase in synaptic density, as reported in several animal models92,97,98 and in the brains of patients with ASD92.
Synaptic receptors and cell adhesion molecules. Many proteins encoded by ASD-risk genes participate in different aspects of neuronal connectivity, such as glutamatergic (for example, glutamate receptor ionotropic, NMDA 2B (GluN2B)), GABAergic (for example, GABAA receptor subunit α3 (GABRA3) and GABAR β3 (GABRB3)) and glycinergic (for example, glycine receptor subunit α2) neurotransmission, neuritogenesis (for example, contactin 6 (CNTN6)), the establishment of synaptic identity (for example, cadherins and protocadherins), neuronal conduction (for example, contactin-associated protein-like 2 (CNTNAP2)) and permeability to ions (voltage-dependent calcium channel subunit α1 (CACNA1) proteins, CACNA2 δ3 (CACNA2D3) and sodium channel protein type 1 subunit-α (SCN1A)). Some of these proteins, such as the neurexins (NRXNs) and neuroligins (NLGNs), are directly involved in activity-driven synapse formation (Fig. 4).
NRXNs and NLGNs are binding partners that are located at the presynapse and postsynapse, respectively25. The three NRXN genes undergo extensive alternative splicing that can produce up to 3,000 putative distinct isoforms with specific biochemical interactions and synapse assembly functions25,101,102,103,104. Interestingly, neuronal activity can modulate the alternative splicing of NRXN genes via calcium/calmodulin-dependent kinase IV (CaMKIV) signalling103. Specific NLGN subtypes are located at the postsynaptic membrane of glutamatergic (NLGN1 and NLGN3), GABAergic (NLGN2 and NLGN4) and glycinergic (NLGN4) synapses25. The intracellular, carboxy-terminal part of the NLGNs binds to several scaffolding proteins of the postsynaptic density that are encoded by certain ASD-risk genes (Supplementary information S2 (figure))105,106. NLGNs can trigger the formation of glutamatergic and GABAergic synapses, but only in an activity-dependent manner107,108,109,110. Indeed, inhibition of NMDA receptors (NMDARs) or CaMKII (which signals downstream of NMDARs) suppresses the formation of glutamatergic synapses through the activity of NLGN1, whereas chronic inhibition of synaptic activity suppresses the formation of GABAergic synapses through the activity of NLGN2 (Ref. 110). As such, NLGNs and NRXNs could regulate synaptic density to maintain local synaptic homeostasis111. Consistent with this idea, mice lacking NLGN3 exhibit extensive misplaced synapse formation and perturbed metabotropic glutamate receptor-dependent synaptic plasticity, similar to that observed in an animal model of fragile X syndrome112.
Scaffolding proteins and the actin cytoskeleton. The positioning of cell-adhesion molecules and neurotransmitter receptors at the synapse involves a complex architecture of postsynaptic proteins, including scaffolding proteins, signalling molecules and the actin cytoskeleton113,114 (Fig. 4). Many of these proteins are encoded by ASD-risk genes. For example, deletions, duplications and coding mutations in the three SH3 and multiple ankyrin repeat domains protein (SHANK) genes (SHANK1, SHANK2 and SHANK3) have been recurrently reported in individuals with ASD115 (Supplementary information S1 (table)). SHANKs assemble into large molecular platforms that interact with glutamate receptors and actin-associated proteins105. In vitro, SHANK3 mutations identified in individuals with ASD reduce actin accumulation in spines, affecting the development and morphology of dendrites as well as axonal growth cone motility116. Neurons derived from induced pluripotent stem cells of patients with 22q13 deletions (which include the deletion of SHANK3) display major defects in excitatory, but not inhibitory, synaptic transmission117. In mice, mutations in Shank3 decrease spine density in the hippocampus and increase dendritic arborizations in striatal neurons118. Mice with mutations in SHANK genes also present with behaviour resembling that in humans with ASDs. Shank1-knockout mice display increased anxiety-like behaviour, decreased vocal communication, decreased locomotion and, remarkably, enhanced working memory but decreased long-term memory119,120,121. Shank2-knockout mice present with hyperactivity, increased anxiety-like behaviour, repetitive grooming and abnormalities in vocal and social behaviours122,123. Shank3-knockout mice show self-injurious repetitive grooming and deficits in social interaction and communication118,124,125,126.
Synaptic homeostasis and ASDs
The sections above consider several examples of synaptic changes that could be caused by some of the rare or de novo mutations identified in individuals with ASD. Interestingly, mice carrying mutations in these genes often display ASD-like behaviours, such as self-injurious repetitive grooming and deficits in social interaction and communication118,119,120,121,122,123,124,125,126 (see Ref. 127 for a review). However, a gene mutation will not always manifest behaviourally in the same way in a mouse (or a fly, worm or fish) and a human. Thus, any robust phenotype found in these models could be relevant to understanding the biological roles of the ASD-risk genes. In some models, pharmacological intervention has been shown to reduce these behavioural changes, even in adult animals (see Ref. 128 for a review). That a late-onset drug treatment can correct a behavioural phenotype in mice carrying ASD-risk gene mutations is actually very encouraging news for the treatment of humans. However, it is not clear whether these positive results obtained in animal models can be translated to positive results in humans.
Based on the different pathways associated with ASD depicted above, two main observations can be made regarding the shared properties of the ASD-risk genes. First, they participate in synaptic plasticity by modulating synaptic strength and/or number. Second, mutations in these genes seem to affect neuronal networks in animal models, resulting in either activity that is too high and increased synaptic density or activity that is too low and decreased synaptic density (Fig. 5). Surprisingly, the majority of ASD-related mutations — specifically, mutations in MEF2C, FMR1, NF1, PTEN, SYNGAP1, EIF4E and CYFIP1 — result in increased gene transcription and mRNA translation; effects that are also seen with increased neuronal activity. These mutations therefore could lead to an abnormal increase in the strength and/or number of synapses within specific neuronal networks. This could account for the observation of ectopic synapses in mouse models of ASD112 and in the brains of patients with ASD129,130, as well as for the higher risk of epilepsy in patients with ASD (for example, in patients with SCN1A or SCN2A mutations).
To prevent neuronal networks from reaching such extreme states (network activity that is too low or too high), homeostatic regulation mechanisms adjust synapse strength, size and number to stabilizing firing131,132,133,134,135,136 (Fig. 5). Many mechanisms are used to achieve this. One is the process of synaptic scaling131, whereas another is metaplasticity136. Other mechanisms to achieve network activity regulation do not involve synapses and are mediated by changes in neuronal excitability137.
The first report of synaptic scaling at central synapses revealed that activity blockade increased excitatory synaptic transmission through a simple change in the accumulation of postsynaptic glutamate receptors, with no changes in presynaptic function131. Although studies agree that synaptic scaling is induced by changes in AMPAR accumulation, there is less agreement on the subunit composition of the newly accumulated receptors. The signalling pathways involved in this synaptic scaling process include voltage-gated calcium channels and proteins such as the cytokine tumour necrosis factor, BDNF and ARC132.
Metaplasticity is induced by synaptic or cellular activity, but it is not necessarily expressed as a change in the efficacy of normal synaptic transmission. Instead, it manifests as a change in the ability to induce subsequent synaptic plasticity, such as LTP or long-term depression (LTD). For example, if activity in the cortex or hippocampus is reduced for a period of time, LTP is enhanced and induced at lower stimulation frequencies, and LTD is repressed136,138. One molecular basis for this is known to be the upregulation of the NMDAR subunit GluN2B (encoded by the ASD-risk gene GRIN2B)139,140. Another basis for this metaplasticity is the downregulation of GABAergic inhibition (regulated by the ASD-risk genes GABRA3 and GABRB3)141,142.
Such homeostatic mechanisms might be especially crucial for restraining network activity and avoiding potential epileptogenic states. There is emerging evidence that sleep is necessary for synaptic homeostasis143. For example, in Drosophila melanogaster, the number and size of synapses in neuronal circuits increase after a few hours of waking and decrease only after sleep144. Interestingly, the NLGNs play an important part in this sleep-dependent synaptic homeostasis in both flies and rodents by regulating levels of glutamatergic and GABAergic currents after sleep deprivation144,145,146,147. Such problems could lead to the abnormal sleep and circadian rhythms that are present in a majority of individuals with ASD148.
BDNF release from the postsynaptic neuron was shown to be essential for homeostatic retrograde enhancement of presynaptic function149. Several studies150,151,152, which have not all been replicated153, have reported an increase in BDNF levels in the blood of patients with ASD. In addition, high plasma levels of serotonin and N-acetylserotonin, as well as low plasma levels of melatonin, are more frequently observed in individuals with ASD than in controls154,155. Given that N-acetylserotonin is a potent agonist of the BDNF receptor tropomyosin-related kinase B (TRKB; also known as NTRK2)156, an excess of this molecule could increase TRKB-induced phosphoinositide 3-kinase signalling, thus leading to higher protein synthesis, similarly to the effect of mutations affecting regulators of the mTOR pathway.
In summary, it can be hypothesized that, in a subset of patients with ASDs who carry genetic mutations, the process of synaptic homeostasis might not be able to counter the effects of the mutations on synaptic activity, leading to neuronal activity that is too high or too low. Abnormal sprouting (or pruning) of synapses as a result of the mutations could therefore lead to abnormal coordination and competition between neuronal networks; this could explain the problems in integrating information across different sensory modalities that are observed in individuals with ASD12,13. This abnormal interference between sensory stimuli could be revealed especially during critical phases of cognitive development, when a tight balance between sensory signals is essential157.
In this Review, I discussed how an individual's genomic background made of common SNPs may buffer the deleterious effect of rare variants and the ability of the nervous system to scale up or down synapses to prevent neuronal networks from reaching extreme states. It is therefore tempting to unify both proposals and to hypothesize that the de novo and rare mutations associated with ASD strongly affect synaptic plasticity and that an individual's genetic background will determine their ability to maintain synaptic homeostasis. We have very little information on the common SNPs associated with ASD. Nevertheless, the common SNPs shared by different neuropsychiatric disorders, such as bipolar disorder and schizophrenia, have started to be identified from large-scale GWASs of >30,000 cases and controls. Interestingly, genes encoding voltage-gated calcium channel subunits CACNA1C, CACNB2 and CACNA1I are key players in synaptic homeostasis132 and were among the top 108 loci identified in schizophrenia158.
Two main questions that are not addressed in this Review relate to where and when in the brain autism 'emerges' (Refs 159,160). Almost all brain regions seem to be implicated in ASD, from the cortex to the cerebellum. Although glutamatergic synapses seem to have a major role in the susceptibility to ASD, GABAergic and glycinergic synapses might also be affected in a subset of patients161,162. As has been suggested by Waterhouse and Gillberg163, it might be better to abandon the belief that there is a single defining dysfunction in the ASD brain and, instead, try to understand the diversity of ASD. The variety of ASD phenotypes may be linked to the many different mutations linked to these disorders but also to the capacity of each individual's genetic background to buffer the effect of their mutations on synaptic function and each individual's capacity to maintain synaptic homeostasis. The question of whether people with ASD are living in an 'intense world' (Ref. 29) or a 'disconnected world' (Ref. 15) might be valid for a subset of individuals, but not necessarily for all. Consideration of autism not as a single entity (affected versus non-affected) but as a continuum of human diversity, and identification of this heterogeneity through the use of information coming from different fields of research (including direct information from the affected individuals and their family164), should allow better diagnosis, care and integration into society of individuals with ASD165.
Kanner, L. Autistic disturbances of affective contact. Nerv. Child 2, 217–250 (1943).
Asperger, H. Die “autistischen Psychopathen” im Kindesalter. Arch. Psychiatr. Nervenkr. 177, 76–137 (in German) (1944). References 1 and 2 are the first reports of individuals diagnosed with autism.
Coleman, M. & Gillberg, C. The Autisms 4th edn (Oxford University Press, 2012).
Constantino, J. N. The quantitative nature of autistic social impairment. Pediatr. Res. 69, 55R–62R (2011).
Skuse, D. H. et al. Social communication competence and functional adaptation in a general population of children: preliminary evidence for sex-by-verbal IQ differential risk. J. Am. Acad. Child Adolesc. Psychiatry 48, 128–137 (2009).
Ronald, A., Happe, F., Price, T. S., Baron-Cohen, S. & Plomin, R. Phenotypic and genetic overlap between autistic traits at the extremes of the general population. J. Am. Acad. Child Adolesc. Psychiatry 45, 1206–1214 (2006).
Developmental Disabilities Monitoring Network Surveillance Year Principal Ivestigators, Centers for Disease Control and Prevention (CDC). Prevalence of autism spectrum disorder among children aged 8 years — autism and developmental disabilities monitoring network, 11 sites, United States, 2010. MMWR Surveill. Summ. 63, 1–21 (2014).
Elsabbagh, M. et al. Global prevalence of autism and other pervasive developmental disorders. Autism Res. 5, 160–179 (2012).
Gillberg, C. The ESSENCE in child psychiatry: Early Symptomatic Syndromes Eliciting Neurodevelopmental Clinical Examinations. Res. Dev. Disabil. 31, 1543–1551 (2010).
Moreno-De-Luca, A. et al. Developmental brain dysfunction: revival and expansion of old concepts based on new genetic evidence. Lancet Neurol. 12, 406–414 (2013). References 9 and 10 highlight the need to take into account the different symptoms and developmental brain dysfunctions observed in individuals with ASD.
Miles, J. H. et al. Essential versus complex autism: definition of fundamental prognostic subtypes. Am. J. Med. Genet. A 135, 171–180 (2005).
Foxe, J. J. et al. Severe multisensory speech integration deficits in high-functioning school-aged children with autism spectrum disorder (ASD) and their resolution during early adolescence. Cereb. Cortex 25, 298–312 (2013).
Marco, E. J., Hinkley, L. B., Hill, S. S. & Nagarajan, S. S. Sensory processing in autism: a review of neurophysiologic findings. Pediatr. Res. 69, 48R–54R (2011).
Courchesne, E. & Pierce, K. Why the frontal cortex in autism might be talking only to itself: local over-connectivity but long-distance disconnection. Curr. Opin. Neurobiol. 15, 225–230 (2005).
Just, M. A., Cherkassky, V. L., Keller, T. A., Kana, R. K. & Minshew, N. J. Functional and anatomical cortical underconnectivity in autism: evidence from an fMRI study of an executive function task and corpus callosum morphometry. Cereb. Cortex 17, 951–961 (2007).
Chaste, P. et al. Adjusting head circumference for covariates in autism: clinical correlates of a highly heritable continuous trait. Biol. Psychiatry 74, 576–584 (2013). An important paper highlighting the importance of considering the covariates of genetic ancestry, height and age in interpreting the results on head circumference in patients with ASD.
Kaiser, M. D. et al. Neural signatures of autism. Proc. Natl Acad. Sci. USA 107, 21223–21228 (2010).
Ebert, D. H. & Greenberg, M. E. Activity-dependent neuronal signalling and autism spectrum disorder. Nature 493, 327–337 (2013).
Neale, B. M. et al. Patterns and rates of exonic de novo mutations in autism spectrum disorders. Nature 485, 242–245 (2012).
O'Roak, B. J. et al. Sporadic autism exomes reveal a highly interconnected protein network of de novo mutations. Nature 485, 246–250 (2012).
Krumm, N., O'Roak, B. J., Shendure, J. & Eichler, E. E. A de novo convergence of autism genetics and molecular neuroscience. Trends Neurosci. 37, 95–105 (2014).
Samocha, K. E. et al. A framework for the interpretation of de novo mutation in human disease. Nat. Genet. 46, 944–950 (2014).
Iossifov, I. et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature 515, 216–221 (2014).
De Rubeis, S. et al. Synaptic, transcriptional and chromatin genes disrupted in autism. Nature 515, 209–215 (2014). References 22–24 report the results of whole-exome sequencing in large samples of patients with ASD.
Sudhof, T. C. Neuroligins and neurexins link synaptic function to cognitive disease. Nature 455, 903–911 (2008).
Auerbach, B. D., Osterweil, E. K. & Bear, M. F. Mutations causing syndromic autism define an axis of synaptic pathophysiology. Nature 480, 63–68 (2011).
Toro, R. et al. Key role for gene dosage and synaptic homeostasis in autism spectrum disorders. Trends Genet. 26, 363–372 (2010).
Ramocki, M. B. & Zoghbi, H. Y. Failure of neuronal homeostasis results in common neuropsychiatric phenotypes. Nature 455, 912–918 (2008).
Markram, H., Rinaldi, T. & Markram, K. The intense world syndrome — an alternative hypothesis for autism. Front. Neurosci. 1, 77–96 (2007).
Hahamy, A., Behrmann, M. & Malach, R. The idiosyncratic brain: distortion of spontaneous connectivity patterns in autism spectrum disorder. Nat. Neurosci. 18, 302–309 (2015).
Belmonte, M. K. & Bourgeron, T. Fragile X syndrome and autism at the intersection of genetic and neural networks. Nat. Neurosci. 9, 1221–1225 (2006).
Xue, Y. et al. Deleterious- and disease-allele prevalence in healthy individuals: insights from current predictions, mutation databases, and population-scale resequencing. Am. J. Hum. Genet. 91, 1022–1032 (2012).
Fu, W. et al. Analysis of 6,515 exomes reveals the recent origin of most human protein-coding variants. Nature 493, 216–220 (2013).
Genome of the Netherlands Consortium. Whole-genome sequence variation, population structure and demographic history of the Dutch population. Nat. Genet. 46, 818–825 (2014).
Kong, A. et al. Rate of de novo mutations and the importance of father's age to disease risk. Nature 488, 471–475 (2012).
Huguet, G., Ey, E. & Bourgeron, T. The genetic landscapes of autism spectrum disorders. Annu. Rev. Genom. Hum. Genet. 14, 191–213 (2013).
Betancur, C. Etiological heterogeneity in autism spectrum disorders: more than 100 genetic and genomic disorders and still counting. Brain Res. 1380, 42–77 (2011).
Robinson, E. B. et al. Evidence that autistic traits show the same etiology in the general population and at the quantitative extremes (5%, 2.5%, and 1%). Arch. Gen. Psychiatry 68, 1113–1121 (2011).
Lundstrom, S. et al. Autism spectrum disorders and autistic like traits: similar etiology in the extreme end and the normal variation. Arch. Gen. Psychiatry 69, 46–52 (2012). References 38 and 39 are epidemiological papers that report independent results indicating that autistic traits in individuals from the general population and in clinical cases of ASD have similar aetiology.
Ronald, A. & Hoekstra, R. A. Autism spectrum disorders and autistic traits: a decade of new twin studies. Am. J. Med. Genet. B Neuropsychiatr. Genet. 156, 255–274 (2011).
Sandin, S. et al. The familial risk of autism. JAMA 311, 1770–1777 (2014).
Hallmayer, J. et al. Genetic heritability and shared environmental factors among twin pairs with autism. Arch. Gen. Psychiatry 68, 1095–1102 (2011).
Constantino, J. N., Zhang, Y., Frazier, T., Abbacchi, A. M. & Law, P. Sibling recurrence and the genetic epidemiology of autism. Am. J. Psychiatry 167, 1349–1356 (2010).
Risch, N. et al. Familial recurrence of autism spectrum disorder: evaluating genetic and environmental contributions. Am. J. Psychiatry 171, 1206–1213 (2014).
Lichtenstein, P., Carlstrom, E., Rastam, M., Gillberg, C. & Anckarsater, H. The genetics of autism spectrum disorders and related neuropsychiatric disorders in childhood. Am. J. Psychiatry 167, 1357–1363 (2010).
Ronald, A., Larsson, H., Anckarsater, H. & Lichtenstein, P. A twin study of autism symptoms in Sweden. Mol. Psychiatry 16, 1039–1047 (2010).
Lundstrom, S. et al. Autistic-like traits and their association with mental health problems in two nationwide twin cohorts of children and adults. Psychol. Med. 41, 2423–2433 (2011).
Gratten, J., Wray, N. R., Keller, M. C. & Visscher, P. M. Large-scale genomics unveils the genetic architecture of psychiatric disorders. Nat. Neurosci. 17, 782–790 (2014).
Klei, L. et al. Common genetic variants, acting additively, are a major source of risk for autism. Mol. Autism 3, 9 (2012).
Cross-Disorder Group of the Psychiatric Genomics Consortium et al. Genetic relationship between five psychiatric disorders estimated from genome-wide SNPs. Nat. Genet. 45, 984–994 (2013).
Gaugler, T. et al. Most genetic risk for autism resides with common variation. Nat. Genet. 46, 881–885 (2014). References 49–51 use quantitative genetics methods to show that an important proportion of the heritability of ASD is captured by common SNPs.
Anney, R. et al. Individual common variants exert weak effects on the risk for autism spectrum disorderspi. Hum. Mol. Genet. 21, 4781–4792 (2012).
Cross-Disorder Group of the Psychiatric Genomics Consortium. Identification of risk loci with shared effects on five major psychiatric disorders: a genome-wide analysis. Lancet 381, 1371–1379 (2013).
Glessner, J. T. et al. Autism genome-wide copy number variation reveals ubiquitin and neuronal genes. Nature 459, 569–573 (2009).
Sanders, S. J. et al. Multiple recurrent de novo CNVs, including duplications of the 7q11.23 Williams syndrome region, are strongly associated with autism. Neuron 70, 863–885 (2011).
Pinto, D. et al. Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am. J. Hum. Genet. 94, 677–694 (2014).
Girirajan, S. et al. A recurrent 16p12.1 microdeletion supports a two-hit model for severe developmental delay. Nat. Genet. 42, 203–209 (2010).
Girirajan, S. et al. Phenotypic heterogeneity of genomic disorders and rare copy-number variants. N. Engl. J. Med. 367, 1321–1331 (2012). References 55–58 report the detection of CNVs in large, independent samples of patients with ASD or developmental delay and highlight the presence of multiple hits.
Leblond, C. S. et al. Genetic and functional analyses of SHANK2 mutations suggest a multiple hit model of autism spectrum disorders. PLoS Genet. 8, e1002521 (2012).
Yuen, R. K. et al. Whole-genome sequencing of quartet families with autism spectrum disorder. Nat. Med. 21, 185–191 (2015).
Michaelson, J. J. et al. Whole-genome sequencing in autism identifies hot spots for de novo germline mutation. Cell 151, 1431–1442 (2012).
Uddin, M. et al. Brain-expressed exons under purifying selection are enriched for de novo mutations in autism spectrum disorder. Nat. Genet. 46, 742–747 (2014).
Jamain, S. et al. Mutations of the X-linked genes encoding neuroligins NLGN3 and NLGN4 are associated with autism. Nat. Genet. 34, 27–29 (2003).
Durand, C. M. et al. Mutations in the gene encoding the synaptic scaffolding protein SHANK3 are associated with autism spectrum disorders. Nat. Genet. 39, 25–27 (2007). References 63 and 64 are the first reports of mutations affecting synaptic proteins in people with ASD.
Sato, D. et al. SHANK1 deletions in males with autism spectrum disorder. Am. J. Hum. Genet. 90, 879–887 (2012).
Skuse, D. H. Rethinking the nature of genetic vulnerability to autistic spectrum disorders. Trends Genet. 23, 387–395 (2007).
Jeste, S. S. & Geschwind, D. H. Disentangling the heterogeneity of autism spectrum disorder through genetic findings. Nat. Rev. Neurol. 10, 74–81 (2014).
Ronemus, M., Iossifov, I., Levy, D. & Wigler, M. The role of de novo mutations in the genetics of autism spectrum disorders. Nat. Rev. Genet. 15, 133–141 (2014).
Lim, E. T. et al. Rare complete knockouts in humans: population distribution and significant role in autism spectrum disorders. Neuron 77, 235–242 (2013).
Jiang, Y. H. et al. Detection of clinically relevant genetic variants in autism spectrum disorder by whole-genome sequencing. Am. J. Hum. Genet. 93, 249–263 (2013).
Rutherford, S. L. From genotype to phenotype: buffering mechanisms and the storage of genetic information. Bioessays 22, 1095–1105 (2000).
Hartman, J. L. 4th, Garvik, B. & Hartwell, L. Principles for the buffering of genetic variation. Science 291, 1001–1004 (2001).
Cohen, S. et al. Genome-wide activity-dependent MeCP2 phosphorylation regulates nervous system development and function. Neuron 72, 72–85 (2011).
Ebert, D. H. et al. Activity-dependent phosphorylation of MeCP2 threonine 308 regulates interaction with NCoR. Nature 499, 341–345 (2013).
Sando, R. 3rd et al. HDAC4 governs a transcriptional program essential for synaptic plasticity and memory. Cell 151, 821–834 (2012).
Amir, R. E. et al. Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-binding protein 2. Nat. Genet. 23, 185–188 (1999).
Ramocki, M. B. et al. Autism and other neuropsychiatric symptoms are prevalent in individuals with MeCP2 duplication syndrome. Ann. Neurol. 66, 771–782 (2009).
Guy, J., Cheval, H., Selfridge, J. & Bird, A. The role of MeCP2 in the brain. Annu. Rev. Cell Dev. Biol. 27, 631–652 (2011).
Yazdani, M. et al. Disease modeling using embryonic stem cells: MeCP2 regulates nuclear size and RNA synthesis in neurons. Stem Cells 30, 2128–2139 (2012).
Muotri, A. R. et al. L1 retrotransposition in neurons is modulated by MeCP2. Nature 468, 443–446 (2010).
Gabel, H. W. et al. Disruption of DNA-methylation-dependent long gene repression in Rett syndrome. Nature 522, 89–93 (2015).
Chao, H. T., Zoghbi, H. Y. & Rosenmund, C. MeCP2 controls excitatory synaptic strength by regulating glutamatergic synapse number. Neuron 56, 58–65 (2007).
Le Meur, N. et al. MEF2C haploinsufficiency caused either by microdeletion of the 5q14.3 region or mutation is responsible for severe mental retardation with stereotypic movements, epilepsy and/or cerebral malformations. J. Med. Genet. 47, 22–29 (2009).
Barbosa, A. C. et al. MEF2C, a transcription factor that facilitates learning and memory by negative regulation of synapse numbers and function. Proc. Natl Acad. Sci. USA 105, 9391–9396 (2008).
Tsai, N. P. et al. Multiple autism-linked genes mediate synapse elimination via proteasomal degradation of a synaptic scaffold PSD-95. Cell 151, 1581–1594 (2012).
Zahir, F. et al. Novel deletions of 14q11.2 associated with developmental delay, cognitive impairment and similar minor anomalies in three children. J. Med. Genet. 44, 556–561 (2007).
O'Roak, B. J. et al. Multiplex targeted sequencing identifies recurrently mutated genes in autism spectrum disorders. Science 338, 1619–1622 (2012).
Bernier, R. et al. Disruptive CHD8 mutations define a subtype of autism early in development. Cell 158, 263–276 (2014). This paper reports a comprehensive genetic and clinical exploration of patients with CHD8 mutations, one of the most frequent genetic causes of ASD.
Chenn, A. & Walsh, C. A. Regulation of cerebral cortical size by control of cell cycle exit in neural precursors. Science 297, 365–369 (2002).
Ma, X. M. & Blenis, J. Molecular mechanisms of mTOR-mediated translational control. Nat. Rev. Mol. Cell Biol. 10, 307–318 (2009).
Kelleher, R. J. 3rd & Bear, M. F. The autistic neuron: troubled translation? Cell 135, 401–406 (2008).
Tang, G. et al. Loss of mTOR-dependent macroautophagy causes autistic-like synaptic pruning deficits. Neuron 83, 1131–1143 (2014).
Williams, C. A., Dagli, A. & Battaglia, A. Genetic disorders associated with macrocephaly. Am. J. Med. Genet. A 146, 2023–2037 (2008).
Budimirovic, D. B. & Kaufmann, W. E. What can we learn about autism from studying fragile X syndrome? Dev. Neurosci. 33, 379–394 (2011).
Fernandez, E., Rajan, N. & Bagni, C. The FMRP regulon: from targets to disease convergence. Front. Neurosci. 7, 191 (2013).
De Rubeis, S. et al. CYFIP1 coordinates mRNA translation and cytoskeleton remodeling to ensure proper dendritic spine formation. Neuron 79, 1169–1182 (2013).
Santini, E. et al. Exaggerated translation causes synaptic and behavioural aberrations associated with autism. Nature 493, 411–415 (2013).
Gkogkas, C. G. et al. Autism-related deficits via dysregulated eIF4E-dependent translational control. Nature 493, 371–377 (2013).
Mabb, A. M. & Ehlers, M. D. Ubiquitination in postsynaptic function and plasticity. Annu. Rev. Cell Dev. Biol. 26, 179–210 (2010).
Greer, P. L. et al. The Angelman syndrome protein Ube3A regulates synapse development by ubiquitinating Arc. Cell 140, 704–716 (2010).
Chih, B., Gollan, L. & Scheiffele, P. Alternative splicing controls selective trans-synaptic interactions of the neuroligin–neurexin complex. Neuron 51, 171–178 (2006).
Comoletti, D. et al. Gene selection, alternative splicing, and post-translational processing regulate neuroligin selectivity for β-neurexins. Biochemistry 45, 12816–12827 (2006).
Iijima, T. et al. SAM68 regulates neuronal activity-dependent alternative splicing of neurexin-1. Cell 147, 1601–1614 (2011).
Schreiner, D. et al. Targeted combinatorial alternative splicing generates brain region-specific repertoires of neurexins. Neuron 84, 386–398 (2014).
Grabrucker, A. M., Schmeisser, M. J., Schoen, M. & Boeckers, T. M. Postsynaptic ProSAP/Shank scaffolds in the cross-hair of synaptopathies. Trends Cell Biol. 21, 594–603 (2011).
Ting, J. T., Peca, J. & Feng, G. Functional consequences of mutations in postsynaptic scaffolding proteins and relevance to psychiatric disorders. Annu. Rev. Neurosci. 35, 49–71 (2012).
Graf, E. R., Zhang, X., Jin, S. X., Linhoff, M. W. & Craig, A. M. Neurexins induce differentiation of GABA and glutamate postsynaptic specializations via neuroligins. Cell 119, 1013–1026 (2004).
Chih, B., Engelman, H. & Scheiffele, P. Control of excitatory and inhibitory synapse formation by neuroligins. Science 307, 1324–1328 (2005).
Varoqueaux, F. et al. Neuroligins determine synapse maturation and function. Neuron 51, 741–754 (2006).
Chubykin, A. A. et al. Activity-dependent validation of excitatory versus inhibitory synapses by neuroligin-1 versus neuroligin-2. Neuron 54, 919–931 (2007). The first paper describing that NLGNs specify and validate synapses via an activity-dependent mechanism.
Kwon, H. B. et al. Neuroligin-1-dependent competition regulates cortical synaptogenesis and synapse number. Nat. Neurosci. 15, 1667–1674 (2012).
Baudouin, S. J. et al. Shared synaptic pathophysiology in syndromic and nonsyndromic rodent models of autism. Science 338, 128–132 (2012). This paper reports the convergence of synaptic pathophysiology in fragile X syndrome mouse models and NLGN3-knockout mice, and the rescue of the NLGN3 alterations by re-expression of NLGN3 in juvenile mice.
Sheng, M. & Kim, E. The postsynaptic organization of synapses. Cold Spring Harb. Perspect. Biol. 3, a005678 (2011).
Choquet, D. & Triller, A. The dynamic synapse. Neuron 80, 691–703 (2013).
Leblond, C. S. et al. Meta-analysis of SHANK mutations in autism spectrum disorders: a gradient of severity in cognitive impairments. PLoS Genet. 10, e1004580 (2014). This paper provides a comprehensive genetic and clinical exploration of patients with SHANK mutations.
Durand, C. M. et al. SHANK3 mutations identified in autism lead to modification of dendritic spine morphology via an actin-dependent mechanism. Mol. Psychiatry 17, 71–84 (2012).
Shcheglovitov, A. et al. SHANK3 and IGF1 restore synaptic deficits in neurons from 22q13 deletion syndrome patients. Nature 503, 267–271 (2013).
Peca, J. et al. Shank3 mutant mice display autistic-like behaviours and striatal dysfunction. Nature 472, 437–442 (2011).
Hung, A. Y. et al. Smaller dendritic spines, weaker synaptic transmission, but enhanced spatial learning in mice lacking Shank1. J. Neurosci. 28, 1697–1708 (2008).
Silverman, J. L. et al. Sociability and motor functions in Shank1 mutant mice. Brain Res. 1380, 120–137 (2011).
Wohr, M., Roullet, F. I., Hung, A. Y., Sheng, M. & Crawley, J. N. Communication impairments in mice lacking Shank1: reduced levels of ultrasonic vocalizations and scent marking behavior. PLoS ONE 6, e20631 (2011).
Schmeisser, M. J. et al. Autistic-like behaviours and hyperactivity in mice lacking ProSAP1/Shank2. Nature 486, 256–260 (2012).
Won, H. et al. Autistic-like social behaviour in Shank2-mutant mice improved by restoring NMDA receptor function. Nature 486, 261–265 (2012).
Bozdagi, O. et al. Haploinsufficiency of the autism-associated Shank3 gene leads to deficits in synaptic function, social interaction, and social communication. Mol. Autism 1, 15 (2010).
Wang, X. et al. Synaptic dysfunction and abnormal behaviors in mice lacking major isoforms of Shank3. Hum. Mol. Genet. 20, 3093–3108 (2011).
Yang, M. et al. Reduced excitatory neurotransmission and mild autism-relevant phenotypes in adolescent Shank3 null mutant mice. J. Neurosci. 32, 6525–6541 (2012).
Ey, E., Leblond, C. S. & Bourgeron, T. Behavioral profiles of mouse models for autism spectrum disorders. Autism Res. 4, 5–16 (2010).
Delorme, R. et al. Progress toward treatments for synaptic defects in autism. Nat. Med. 19, 685–694 (2013).
Chung, K. et al. Structural and molecular interrogation of intact biological systems. Nature 497, 332–337 (2013).
Stoner, R. et al. Patches of disorganization in the neocortex of children with autism. N. Engl. J. Med. 370, 1209–1219 (2014).
Turrigiano, G. G., Leslie, K. R., Desai, N. S., Rutherford, L. C. & Nelson, S. B. Activity-dependent scaling of quantal amplitude in neocortical neurons. Nature 391, 892–896 (1998). This paper reports the first evidence of synaptic scaling in neocortical neurons.
Turrigiano, G. Too many cooks? Intrinsic and synaptic homeostatic mechanisms in cortical circuit refinement. Annu. Rev. Neurosci. 34, 89–103 (2011).
Yu, L. M. & Goda, Y. Dendritic signalling and homeostatic adaptation. Curr. Opin. Neurobiol. 19, 327–335 (2009).
Vitureira, N., Letellier, M. & Goda, Y. Homeostatic synaptic plasticity: from single synapses to neural circuits. Curr. Opin. Neurobiol. 22, 516–521 (2012).
Pozo, K. & Goda, Y. Unraveling mechanisms of homeostatic synaptic plasticity. Neuron 66, 337–351 (2010).
Abraham, W. C. & Bear, M. F. Metaplasticity: the plasticity of synaptic plasticity. Trends Neurosci. 19, 126–130 (1996).
Mozzachiodi, R. & Byrne, J. H. More than synaptic plasticity: role of nonsynaptic plasticity in learning and memory. Trends Neurosci. 33, 17–26 (2010).
Abraham, W. C. Metaplasticity: tuning synapses and networks for plasticity. Nat. Rev. Neurosci. 9, 387 (2008).
Chen, W. S. & Bear, M. F. Activity-dependent regulation of NR2B translation contributes to metaplasticity in mouse visual cortex. Neuropharmacology 52, 200–214 (2007).
Yang, Q. et al. Hippocampal synaptic metaplasticity requires the activation of NR2B-containing NMDA receptors. Brain Res. Bull. 84, 137–143 (2011).
Carlson, G., Wang, Y. & Alger, B. E. Endocannabinoids facilitate the induction of LTP in the hippocampus. Nat. Neurosci. 5, 723–724 (2002).
Chevaleyre, V. & Castillo, P. E. Endocannabinoid-mediated metaplasticity in the hippocampus. Neuron 43, 871–881 (2004).
Tononi, G. & Cirelli, C. Sleep and synaptic homeostasis: a hypothesis. Brain Res. Bull. 62, 143–150 (2003).
Gilestro, G. F., Tononi, G. & Cirelli, C. Widespread changes in synaptic markers as a function of sleep and wakefulness in Drosophila. Science 324, 109–112 (2009).
Huber, R., Ghilardi, M. F., Massimini, M. & Tononi, G. Local sleep and learning. Nature 430, 78–81 (2004).
El Helou, J. et al. Neuroligin-1 links neuronal activity to sleep–wake regulation. Proc. Natl Acad. Sci. USA 110, 9974–9979 (2013).
Li, Y. et al. Drosophila neuroligin 4 regulates sleep through modulating GABA transmission. J. Neurosci. 33, 15545–15554 (2013).
Bourgeron, T. The possible interplay of synaptic and clock genes in autism spectrum disorders. Cold Spring Harb. Symp. Quant. Biol. 72, 645–654 (2007).
Jakawich, S. K. et al. Local presynaptic activity gates homeostatic changes in presynaptic function driven by dendritic BDNF synthesis. Neuron 68, 1143–1158 (2010).
Kasarpalkar, N. J., Kothari, S. T. & Dave, U. P. Brain-derived neurotrophic factor in children with autism spectrum disorder. Ann. Neurosci. 21, 129–133 (2014).
Halepoto, D. M., Bashir, S. & AL-Ayadhi, L. Possible role of brain-derived neurotrophic factor (BDNF) in autism spectrum disorder: current status. J. Coll. Physicians Surg. Pak. 24, 274–278 (2014).
Masi, A. et al. Cytokine aberrations in autism spectrum disorder: a systematic review and meta-analysis. Mol. Psychiatry 20, 440–446 (2014).
Rodrigues, D. H. et al. Circulating levels of neurotrophic factors in autism spectrum disorders. Neuro Endocrinol. Lett. 35, 380–384 (2014).
Melke, J. et al. Abnormal melatonin synthesis in autism spectrum disorders. Mol. Psychiatry 13, 90–98 (2008).
Pagan, C. et al. The serotonin–N-acetylserotonin–melatonin pathway as a biomarker for autism spectrum disorders. Transl Psychiatry 4, e479 (2014).
Jang, S. W. et al. N-acetylserotonin activates TrkB receptor in a circadian rhythm. Proc. Natl Acad. Sci. USA 107, 3876–3881 (2010).
Hensch, T. K. Critical period regulation. Annu. Rev. Neurosci. 27, 549–579 (2004).
Schizophrenia Working Group of the Psychiatric Genomics Consortium. Biological insights from 108 schizophrenia-associated genetic loci. Nature 511, 421–427 (2014). This article describes the largest GWAS in psychiatric disorders, highlighting the role of the common variants in schizophrenia and revealing that large numbers of patients and controls are required to identify risk alleles.
Parikshak, N. N. et al. Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell 155, 1008–1021 (2013).
Willsey, A. J. et al. Coexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell 155, 997–1007 (2013).
Karayannis, T. et al. Cntnap4 differentially contributes to GABAergic and dopaminergic synaptic transmission. Nature 511, 236–240 (2014).
Lionel, A. C. et al. Rare exonic deletions implicate the synaptic organizer Gephyrin (GPHN) in risk for autism, schizophrenia and seizures. Hum. Mol. Genet. 22, 2055–2066 (2013).
Waterhouse, L. & Gillberg, C. Why autism must be taken apart. J. Autism Dev. Disord. 44, 1788–1792 (2014).
Kohane, I. S. & Eran, A. Can we measure autism? Sci. Transl Med. 5, 209ed18 (2013).
Mottron, L. Changing perceptions: the power of autism. Nature 479, 33–35 (2011). References 163–165 highlight the importance of studying autism from a new perspective, with the integration of clinical, molecular and biochemical characteristics in a patient-information commons and by taking into account not only the deficits but also the abilities and strengths of people with autism.
Jarosz, D. F., Taipale, M. & Lindquist, S. Protein homeostasis and the phenotypic manifestation of genetic diversity: principles and mechanisms. Annu. Rev. Genet. 44, 189–216 (2010). A comprehensive review on the concept of genetic buffering.
Toma, C. et al. Exome sequencing in multiplex autism families suggests a major role for heterozygous truncating mutations. Mol. Psychiatry 19, 784–790 (2014).
Ardehali, M. B. & Lis, J. T. Tracking rates of transcription and splicing in vivo. Nat. Struct. Mol. Biol. 16, 1123–1124 (2009).
King, I. F. et al. Topoisomerases facilitate transcription of long genes linked to autism. Nature 501, 58–62 (2013).
McMahon, A. C. et al. SynGAP isoforms exert opposing effects on synaptic strength. Nat. Commun. 3, 900 (2012).
Jamain, S. et al. Linkage and association of the glutamate receptor 6 gene with autism. Mol. Psychiatry 7, 302–310 (2002).
Li, Q., Lee, J. A. & Black, D. L. Neuronal regulation of alternative pre-mRNA splicing. Nat. Rev. Neurosci. 8, 819–831 (2007).
Voineagu, I. et al. Transcriptomic analysis of autistic brain reveals convergent molecular pathology. Nature 474, 380–384 (2011).
Khare, T. et al. 5-hmC in the brain is abundant in synaptic genes and shows differences at the exon–intron boundary. Nat. Struct. Mol. Biol. 19, 1037–1043 (2012).
Peixoto, R. T. et al. Transsynaptic signaling by activity-dependent cleavage of neuroligin-1. Neuron 76, 396–409 (2012).
Suzuki, K. et al. Activity-dependent proteolytic cleavage of neuroligin-1. Neuron 76, 410–422 (2012).
Napoli, I. et al. The fragile X syndrome protein represses activity-dependent translation through CYFIP1, a new 4E-BP. Cell 134, 1042–1054 (2008).
Splawski, I. et al. CaV1.2 calcium channel dysfunction causes a multisystem disorder including arrhythmia and autism. Cell 119, 19–31 (2004).
The author thanks Sophie Calderari, Guillaume Dumas, Elodie Ey, Eva Loth, Thomas Rolland and Roberto Toro for their discussions and reading of the manuscript. This work was funded by the Institut Pasteur, Fondation Bettencourt Schueller, Centre National de la Recherche Scientifique, University Paris Diderot, Agence Nationale de la Recherche (SynDivAutism), the Conny-Maeva Charitable Foundation, Fondation Cognacq-Jay, the Orange Foundation and Fondation FondaMental. The research leading to these results has also received support from the Innovative Medicines Initiative Joint Undertaking under grant agreement no. 115300, resources of which are composed of financial contribution from the European Union's Seventh Framework Programme (FP7/2007–2013) and the European Federation of Pharmaceutical Industries and Associations (EFPIA) companies' in-kind contribution.
The author declares no competing financial interests.
- Dysmorphic features
Differences in body structure compared with that in the general population. Dysmorphic features can be isolated or multiple and vary from mild anomalies, such as minor malformations of the fingers, to more severe differences, such as microcephaly.
- Synaptic homeostasis
Crosstalk between the presynaptic and the postsynaptic sides of the synapse that allows the tight regulation of synaptic strength and thus maintains excitability within a narrow range. It stabilizes neuronal circuits and ensures the fidelity of communication within the neuronal network despite sensory and/or growth-dependent changes.
- Single-nucleotide variants
(SNVs). DNA-sequence variations occurring within a population. SNV is the general term for all such variations. Single-nucleotide polymorphism is usually used for SNVs occurring in > 1% of the population.
- Copy-number variants
(CNVs). Variations in the number of copies of one segment of DNA. CNVs include deletions and duplications.
- Psychological and cognitive tests
Various self- and parent-reports designed to reliably quantify autistic traits in the general population and in clinical cases. Examples of theses scales are the social responsiveness scale and the autism spectrum quotient.
The proportion of the phenotypic variance that is due to genetic factors. In the narrow sense, heritability includes only the additive genetic component. However, in the broad sense, heritability includes both the additive and the dominance genetic components.
- Dominance components
Part of the genetic contribution to a phenotype, the other part of which is the additive component. Some genes have an additive effect on the quantitative trait, whereas other genes may exhibit a dominant gene action, which will mask the contribution of the recessive alleles at the locus.
- Quantitative molecular genetics
A branch of population genetics that assesses the heritability of continuously distributed phenotypes based on their molecular genetic signatures. For example, the Genome-Wide Complex Trait Analysis (GCTA) software estimates genomic relationships between pairs of conventionally unrelated individuals using single-nucleotide polymorphism (SNP) data.
- Gene dosage
The number of copies of a gene that are present in a cell. An abnormal gene dosage (by gene deletion or duplication) can result in abnormal levels of gene product formation. Gene dosage compensation to adjust the normal level of gene product can occur at different levels (transcription, translation and degradation).
- Genetic buffer
The process by which an individual's genetic background can moderate or counteract the phenotypic effect of deleterious mutations.
- Chromatin remodelling
The dynamic modification of chromatin architecture to allow or deny access of condensed genomic DNA to the regulatory transcription machinery proteins, thereby controlling gene expression.
- X inactivation
A process by which one of the two copies of the X chromosome present in female mammals is inactivated.
- Synaptic scaling
A form of synaptic plasticity that adjusts the strength of a neuron's excitatory synapses up or down to stabilize firing.
The plasticity of synaptic plasticity (that is, the prior history of activity of a synapse determines its current plasticity).
About this article
Cite this article
Bourgeron, T. From the genetic architecture to synaptic plasticity in autism spectrum disorder. Nat Rev Neurosci 16, 551–563 (2015). https://doi.org/10.1038/nrn3992
Neuroligin dependence of social behaviour in Caenorhabditis elegans provides a model to investigate an autism-associated gene
Human Molecular Genetics (2021)
Journal of Neuroscience Research (2021)
Biological Psychiatry: Cognitive Neuroscience and Neuroimaging (2021)
Bixafen, a succinate dehydrogenase inhibitor fungicide, causes microcephaly and motor neuron axon defects during development
Psychiatry Research (2021)