The Stress-Induced Atf3-Gelsolin Cascade Underlies Dendritic Spine Deficits in Neuronal Models of Tuberous Sclerosis Complex

Abstract

Hyperactivation of the mechanistic target of rapamycin (mTOR) kinase, as a result of loss-of-function mutations in tuberous sclerosis complex 1 (TSC1) or TSC2 genes, causes protein synthesis dysregulation, increased cell size, and aberrant neuronal connectivity. Dysregulated synthesis of synaptic proteins has been implicated in the pathophysiology of autism spectrum disorder (ASD) associated with TSC and fragile X syndrome. However, cell type-specific translational profiles in these disease models remain to be investigated. Here, we used high-fidelity and unbiased Translating Ribosome Affinity Purification (TRAP) methodology to purify ribosome-associated mRNAs and identified translational alterations in a rat neuronal culture model of TSC. We find that expression of many stress and/or activity-dependent proteins is highly induced while some synaptic proteins are repressed. Importantly, transcripts for the activating transcription factor-3 (Atf3) and mitochondrial uncoupling protein-2 (Ucp2) are highly induced in Tsc2-deficient neurons, as well as in a neuron-specific Tsc1 conditional knock-out mouse model, and show differential responses to the mTOR inhibitor rapamycin. Gelsolin, a known target of Atf3 transcriptional activity, is also upregulated. shRNA-mediated block of Atf3 induction suppresses expression of gelsolin, an actin-severing protein, and rescues spine deficits found in Tsc2-deficient neurons. Together, our data demonstrate that a cell-autonomous program consisting of a stress-induced Atf3-gelsolin cascade affects the change in dendritic spine morphology following mTOR hyperactivation. This previously unidentified molecular cascade could be a therapeutic target for treating mTORopathies.

Dysfunctional mTORC1 Signaling: A Convergent Mechanism between Syndromic and Nonsyndromic Forms of Autism Spectrum Disorder?

“…Atypical number and/or length and/or morphology of dendritic spines were mechanistically linked to disinhibited mTORC1 signaling in different mouse models of ASD-related syndromes, including TSC [75,76,86], PTEN-ASD [100], FXS [160,166] and AS [44,45], as well as in NS-ASD models, such as Cyfip1 transgenic mice [52], eIF4E-NS-ASD mice [49,50] and idiopathic ASD patients [51]. Importantly, a recent study reported that disruption of mTORC1-dependent macroautophagy reduces spine pruning and consequently increases spine density in neurons of individuals with TSC or idiopathic ASD [51].…”

“…Consistent brain functional and morphological abnormalities observed in these animals include seizures [68,69,70,71], larger brains [69,70,72], deficits in neuronal migration and cortical lamination [68,69,70,72,73,74], enlarged and dysplastic neurons [73,75,76,77] astrogliosis [68,69,70,73], reduced myelination [72,76], multiple and ectopic axons [78,79], enhanced excitatory network [65,80,81], and disrupted synaptic plasticity in the form of impaired hippocampal long-term potentiation (LTP) [82,83] and metabotropic glutamate receptor-mediated long-term depression (mGluR-LTD) [80,84] (Table 1). Notwithstanding, some conflicting results exist regarding normal/increased neurite length [78,79], reduced/normal/increased dendritic spine density and length [51,75,76,80,85,86] and increased spine head width/immature shape [75,85] (Table 1), possibly due to the use of different animal models or experimental conditions. Importantly, while it is still unknown whether increased neurite length and reduced LTP are dependent exclusively on mTORC1 overactivation, the majority of the other brain alterations in rodent models were partially or completely rescued or prevented by the mTORC1 inhibitor rapamycin [51,69,70,72,75,76,78,81,84,86] (Table 1), including astrogliosis, red...…”

“…Notwithstanding, some conflicting results exist regarding normal/increased neurite length [78,79], reduced/normal/increased dendritic spine density and length [51,75,76,80,85,86] and increased spine head width/immature shape [75,85] (Table 1), possibly due to the use of different animal models or experimental conditions. Importantly, while it is still unknown whether increased neurite length and reduced LTP are dependent exclusively on mTORC1 overactivation, the majority of the other brain alterations in rodent models were partially or completely rescued or prevented by the mTORC1 inhibitor rapamycin [51,69,70,72,75,76,78,81,84,86] (Table 1), including astrogliosis, reduced myelination and ASD-relevant behaviors (not described in Table 1), even when treatment is begun in adulthood [77,87,88], demonstrating that they are dependent on mTORC1 overactivation. Notably, it was shown that cell-autonomous and non-cell autonomous mechanisms still not understood drive astrogliosis in Tsc1 -deficient mice [68,69,70,73], and that neuron-specific deletion of Tsc1 impairs oligodentrocyte maturation and myelination through a non-cell autonomously manner [89].…”

“…On the other hand, an in vitro study has shown that Tsc2 -deficient neurons have increased autophagy through 5' adenosine monophosphate (AMP)-activated protein kinase (AMPK) stimulation [92], which could potentially explain the controversial results regarding dendritic spine density observed in different studies. In addition, recent reports revealed mTORC1-dependent alterations in hippocampal protein synthesis in TSC mouse models, including decreased expression of synaptic proteins, such as the plasticity-related Arc protein required for mGluR-LTD [84], which could contribute to the impaired synaptic plasticity, but increased expression of stress-responsive proteins and anti-inflammatory cytokines [86], which may suggest an attempt to deal with the oxidative stress and inflammation that are frequently observed in the brain of individuals with ASD [93]. …”

Dendritic spine actin cytoskeleton in autism spectrum disorder

“…shown to underlie dendritic spine deficits in mice with loss-of-function mutations in tuberous sclerosis complex 1 (Tsc1) or Tsc2 genes 96 .…”