Adult brain is plastic: Re-expressing Shank3 in adult mice rescues autistic like phenotype



Autism spectrum diseases (ASD) are associated with hundreds of genes; most of the genes that are discovered came from genome wide association studies. Whole genome wide association studies on families affected with ASD helps us in understanding the co-occurrence of SNPs (Single nucleotide polymorphism) and rare variants associated with the disease and shed light on how structural changes such as CNVs (Copy number variants), SNPs are involved in cognitive disorders. Mapping how functional regulatory variants influencing the expression of synaptic proteins in neurocognitive disorders is a hot bed of intense translational research. In my previous blog, ( I had described how methods such as genome wide association studies, (let us informing the frequency and the number of the genetic variants), comparative genomics, transcriptome analysis, quantitative trait loci mapping studies combining with rigorous statistical and computational methods is helping in gleaning information about ASD. As expected, the usual suspects for the ASD discovered from the above mentioned studies are those genes which are involved in synaptic transmission, scaffold proteins to maintain the synaptic architecture, genes involved in modulating or maintaining neuronal activity. As much as it is important to characterize how the function of these genes are compromised in a disease state such as Autism spectrum disorders, caution has to be taken while using the mouse model to understand the causality of dysfunction of genes, especially when a dysfunction of a single gene is translated to changes in behavioural responses. Anatomical changes such as spine density, synapse number can be caused not only by original mutation but also fuelled by a cascade of secondary changes associated with neuronal activity, their history of activity and other compensatory mechanisms, especially the compensatory mechanisms are hard to dissect. Though molecular research on a single gene responsible for a spectrum of diseases has its own shortcomings, with the advent of tools such as selective manipulation of subset of neurons, high resolution imaging approaches, connectomics to quantify the convergence and divergence of connections to determine the optimal mode of information transfer, activity manipulation, array of behavioural experiments gave us a great opportunity to understand how genes causes cognitive disorders.

Genome wide association and studies in patients had discovered that mutations in Shank (Shank1, 2 and 3) are one of the common causes of autism spectrum disorders and they have also been implicated in other cognitive disorders such as Schizophrenia and bipolar diseases. Mutation in Shank3 is the best-characterized mutations in patients with ASD and this mutation is mapped to 22q1.3 deletion syndrome nonetheless both Shank1 and Shank2 are also involved in ASD. Shank, a scaffolding protein present in the post-synaptic density (PSD) region of excitatory neurons is involved in forming a polymer like structure by the virtue of protein-protein interaction with other molecules keeping the PSD regions at the tip of the dendritic spines stable. The electron dense PSDs are involved in tethering the receptors to the post-synaptic membrane thereby coupling the downstream synaptic signalling pathways. Besides they also act as an anchor for attaching other important PSD molecules as well as involved in the recruitment of AMPA and NMDA receptors in the excitatory synapses. Shank is also involved in triggering signal cascade pathway through its PDZ binding region thereby bidirectionally activating the synapses during potentiation and depression. Moreover Shank and other proteins can be ubiquitinated and proteasomally degraded by changes in synaptic activity and hence they are required for homeostatic synaptic scaling. Due to its important function in keeping the excitatory synapses intact, it is not surprising Shank mouse models displays some of the autistic like behaviour and this phenotype has been reproduced in different labs that made them a valid model for autism studies. In a recent study published in Nature, Guoping Feng and colleagues had shown that by expressing functional Shank3 at its endogenous loci in the adult animal using an inducible Cre-dependent genetic switch restores the function of Shank3 culminating in reversing some of the selective autistic like phenotypes displayed by the Shank3 mutant. Shank3 is highly enriched in striatum (To note: They are also expressed in hippocampus, cerebellum and also in the cortex) , striatum is associated with some of the autistic behaviour and hence they decided to characterize the medium spiny neurons in the striatum. One of the striking phenotype is the re-enrichment of the scaffolding proteins such as Homer, glutamate receptors subunits such as NR2A, NR2B, GluA2 in the PSD of the excitatory synapses of the mouse model that re-express Shank3 conditionally in adult stage, whilst the scaffolding proteins and receptor subunits are very less or completely lost in the knockout mutant. Moreover, Shank3 mutants display reduced spine density, reduction in spine volume and reduced PSD electron dense region, which are also rescued to that of the wild type in the conditional animal by functional restoration of Shank3 in adult animals insinuating a high degree of plasticity exhibited by the spiny neurons in the striatum. It is noteworthy that the ability of Shank3 to promote spine development and maintenance even after critical period of development suggesting an on-going structural plasticity in the adult striatum. Besides, they have also shown that the re-expression of Shank3 could rescue the reduced miniature excitatory postsynaptic current frequency to that of the wild type denoting the functional recruitment or stable tethering of the post-synaptic receptors back to the normal level. Some of the behavioural deficits such as compulsive and social deficit could be rescued, whilst anxiety and motor co-ordination deficits couldn’t be rescued in adult animals by the functional expression of Shank3 in the mutant. The authors speculated that this might be due to the consequence of connectivity defect though it has not been tested. Strikingly, both the anxiety-like behaviour and the motor co-ordination deficit could be partially rescued when they express Shank3 early postnatal (20 -21 days) period of development. The highlight of this work points out the on-going plasticity exhibited by the adult brain and intervention in the early stages of development can revert morphological, functional and behavioural abnormalities in the ASD models to that of the wild type.

This work functionally rescues the autistic related behavioural phenotypes in adult animal nonetheless it implores a few interesting questions to address.

The authors have shown that they can rescue the spine density in conditional animals compared to that of the knock out, nonetheless the number of spines in the conditional is significantly higher than that of the wild type. The authors speculated that this increase in spine density is due to the absence of pruning in the developmental stages. This posits a legitimate question whether

i) All this newly formed spines are functionally active? Restoration of the miniature excitatory post-synaptic current frequency partially explains that the newly formed spines will have functional glutamate receptors anchored to the PSD. However, it warrants a question to ask, what happens to the pre-synaptic machinery in these mutants?

ii) Are there de novo additions of new pre-synaptic boutons juxtaposed to the newly formed spines?

iii) Are there a degeneration of boutons in the mutant and then an appearance of functional boutons during the re-expression of shank3, which might retrogradely transduce signals for the pre-synaptic maturation? Or there exist ghost boutons, which are silent in the mutant and become functionally active later in the conditional animals?

iv) What is the turn over of these spines in the conditional animals compared to that of the wild type?

v) Are they more stable or transient? Given Shank3 is expressed in cerebellum, cortex and in hippocampus, does it hold true that there is an increase in spine density in the above-mentioned regions?

vi) Recently work from Mark Schnitzer’s lab has shown that the neocortex possess more permanent spines while that of the CA1 in hippocampus have unstable population of spines, and they also possess different turn –over ratios. I had written a blog about this work on the kinetics and dynamics of neocortical spines to that of the hippocampal spines, ( ). It will be interesting to see how Shank3 regulates the structural plasticity in cortex and in hippocampus in comparison to that of the adult striatum. In other words does functional expression of Shank3 forces new rules of structural plasticity in cortex and in hippocampus?

vii) It will be interesting too see how the connectivity matrix such as axon arborisation is different in mutant, conditional and in wild type animals. This is important in light of the finding that both the motor cortex deficit as well as anxiety like behaviours couldn’t be rescued in adult stage in Shank-3 mutant while the above mentioned defects was partially rescued during early stages of development.

There are 3 shank proteins (1, 2, & 3) and tens of isoforms of the shank proteins. Moreover all the shank mutants cause autistic like behaviour in the mouse model. Given some of the protein domain structure are same in all Shank proteins, it will be interesting to see whether each of them have an overlapping phenotype or do they exhibit distinct behavioural deficits? Mutants of different isoforms are created for different shank proteins to model the role of isoforms in disease state, which yield different and overlapping behavioural deficits. It is also speculated that phenotypic heterogeneity in patients are due to the mutations harboured in different isoforms. It will be important to validate the broad spectrum of different isoforms of Shanks and translate how mutations in isoforms lead to ASD phenotype or how they influence each other? In other words, will it be possible by comparing different shank genes as well as substituting one with another can lead to better therapeutic interventions. An important question, which needs to be addressed from this study, would be that the rescue of structural, functional and behavioural deficits by expressing Shank3 in the later stages of development can also produce identical phenotypes in other Shank mouse models or do they diverge? This might enable in finding common mode of operations of Shank proteins and their isoforms in different levels ranging from structural to functional to behavioural in different regions of the brain.

P.S Due to the scarcity of space, I am not enlisting the detailed references except the scientific paper of discussion. The header image was downloaded from the Google images, originally from, history of autism.


  1. Mei Y et al; Adult restoration of Shank3 expression rescues selective autistic like phenotypes. Nature, 2016.
  2. Attardo A et al; Impermanence of dendritic spines in live adult CA1 hippocampus. Nature, 2015.


Leave a Comment