More Than mTOR? Novel Roles for MEK-ERK1/2 and FLNA in Tuberous Sclerosis Complex

Commentary

MEK-ERK1/2-Dependent FLNA overexpression promotes abnormal dendritic patterning in tuberous sclerosis independent of mTOR.

Zhang L, Bartley CM, Gong X, Hsieh LS, Lin TV, Feliciano DM, Bordey A. Neuron 2014;84:78–91.

Abnormal dendritic complexity is a shared feature of many neurodevelopmental disorders associated with neurological defects. Here, we found that the actin-crosslinking protein filamin A (FLNA) is overexpressed in tuberous sclerosis complex (TSC) mice, a PI3K-mTOR model of neurodevelopmental disease that is associated with abnormal dendritic complexity. Both under- and overexpression of FLNA in wild-type neurons led to more complex dendritic arbors in vivo, suggesting that an optimal level of FLNA expression is required for normal dendritogenesis. In Tsc1^null neurons, knocking down FLNA in vivo prevented dendritic abnormalities. Surprisingly, FLNA overexpression in Tsc1^null neurons was dependent on MEK1/2 but not mTOR activity, despite both pathways being hyperactive. In addition, increasing MEK-ERK1/2 activity led to dendritic abnormalities via FLNA, and decreasing MEK-ERK1/2 signaling in Tsc1^null neurons rescued dendritic defects. These data demonstrate that altered FLNA expression increases dendritic complexity and contributes to pathologic dendritic patterning in TSC in an mTOR-independent, ERK1/2-dependent manner.

Neurodevelopmental disorders are highly associated with epilepsy and result from diverse genetic lesions. Recent work has revealed an enormous diversity in genes that can lead to developmental epilepsy, including those involved in cytoskeletal organization, neuronal migration, and synapse formation (1). Even with wide the array of mutations associated with these diseases, there is a surprising degree of neuropathological similarity across syndromes. For example, cortical malformations (2), disrupted neuronal morphology, and synaptic reorganization (3) are seen in etiologically diverse neurodevelopmental diseases. Tuberous sclerosis complex (TSC) is a prototypical neurodevelopmental disease highly associated with epilepsy (4). TSC is caused by mutations in the Tsc1 or Tsc2 genes. Tsc1 and Tsc2 are negative regulators of Rheb, an activator of mTOR signaling. Therefore, the loss of Tsc1 or Tsc2 leads to hyperactivation of mTOR pathways. The most striking feature of TSC is the presence of cortical tubers, areas of disorganized cortex containing a mix of phenotypically immature neuronal and glial cells (5). Upon closer histopathological examination of resected tissue from TSC patients, a wide variety of disrupted cell types are seen, often with bizarre cellular morphology (6). Soma size, dendritic branching, spine density, and structure are all altered in TSC, so much so that identification of cell types themselves can be challenging.

mTOR is a well known regulator of cell size, so it is not surprising that its dysregulation leads to altered neuronal morphology. Mechanistic studies of TSC have begun to offer a more specific understanding of how loss of Tsc1 or Tsc2 disrupts cellular morphology. In 2005, an important study demonstrated that loss of Tsc1 or Tsc2 caused mTOR-dependent enlargement of neuronal soma and dendritic spines via cofilin, a molecule involved in actin depolymerization (7). More recently, it was shown that sporadic loss of Tsc1 also resulted in cells with large soma and increased dendritic complexity (8). Interestingly, this model suggests that cellular mosaicism, thought to be involved in many neurological diseases (9), may be an important component of the morphological disruptions in TSC. Building on these studies, Zhang and colleagues contributed another important piece to the puzzle by identifying filamin A (FLNA), an actin crosslinking protein, as a player in the disrupted cellular morphology associated with TSC.

In mice containing a floxed Tsc1 gene and a tdTomato reporter allele, Cre-containing vector was electroporated into neural stem cells in the subventricular zone (SVZ) unilaterally at P0. Newborn neurons generated in the SVZ migrated to the olfactory bulb and were collected at P14 for analysis of mRNA levels. Using a PCR array of genes involved in neurogenesis, FLNA was identified as one of five genes that were significantly upregulated after loss of Tsc1. This led to an increase in FLNA protein level, which could be mimicked by experimental induction of Rheb hyperactivation. Importantly, FLNA levels were shown to be significantly increased in human brain resected from TSC patients as compared to postmortem brain samples with no known pathology. The effects of both increasing and decreasing levels of FLNA on dendritic complexity in wild-type (WT) neurons were next examined via electroporation of plasmids containing either the FLNA gene (increased expression) or short hairpin RNAs (shRNA) directed at FLNA (decreased expression). Surprisingly, both manipulations led to increased dendritic complexity as measured by Scholl analysis and total dendrite length. This suggests that there is a “sweet spot” for FLNA expression, and deviations in either direction will lead to morphological abnormalities. To address the specificity of FLNA in mediating the change in dendritic complexity, FLNA levels were decreased using shRNA in Tsc1^null neurons. Reducing FLNA levels restored normal dendritic morphology and reduced the increased synaptic activity seen in Tsc1 ^null neurons, presumably due to a reduction in the total dendrite length.

The work of Zhang and colleagues is especially important as they went on to examine the signal transduction mechanism linking loss of Tsc1 to FLNA. Many studies have investigated the role of mTOR in mediating the cellular changes caused by loss of Tsc1/2. Unexpectedly, the increase in FLNA was insensitive to mTOR inhibition via rapamycin and Torin 1. Inhibition of MEK-ERK1/2, another pathway implicated in cell growth and known to be overactive in TSC, did restore FLNA levels to normal. Both mTOR and MEK1/2 are downstream targets of Rheb, which is overactive in Tsc1^null neurons. The effects of Tsc1 loss on dendritic complexity could be mimicked in WT neurons by increasing MEK-ERK1/2 activity and, most significantly, pharmacological inhibition of MEK-ERK1/2 restored dendritic morphology in Tsc1^null neurons. Together, these studies identify FLNA as a gene relevant to the morphological abnormalities seen in TSC and show that its aberrant levels are Rheb and MEK-ERK1/2 dependent but mTOR independent. This is an important distinction and demonstrates our limited knowledge of Rheb-dependent changes in cell properties outside of mTOR signaling.

Implicating FLNA in neurodevelopmental epilepsy might not surprise some readers, as mutations in FLNA are a known genetic cause of periventricular heterotopia, a disease characterized by groups of neurons which remain in the walls of the lateral ventricle (10). Loss of FLNA leads to a failure to migrate out of the SVZ and into their proper location in the cortical mantle. Interestingly, this tends to be a late-onset epilepsy. Additionally, ERK1/2 has been demonstrated to be increased in human TSC (11) and plays a role in disruptions of synaptic plasticity in TSC (12). That said, this study greatly contributes to our understanding of TSC-related cellular pathology by elegantly connecting the dots between Tsc1/2, Rheb, and FLNA. There are a number of questions that this study highlights: First, suppression of mTOR pathways can reduce dendritic complexity without affecting FLNA levels (13). How then do cytoskeletal components downstream of mTOR interact with FLNA and other MEK-ERK1/2-dependent changes in cytoskeleton? Second, there are extremely diverse morphological changes seen in human TSC. Neurons can have both increased and decreased density of dendritic spines, glial cells are affected, and dendritic complexity varies wildly across neurons. Does disruption of FLNA mediate some or all of these effects? What other cytoskeletal proteins may shape FLNA-dependent changes? Does MEK-ERK1/2 activity vary on a cell-by-cell basis in TSC? Lastly, TSC is a disease characterized by tubers. In these studies, anatomical organization was not investigated, but FLNA is known to play a role in neuronal migration and motility. Do alterations in FLNA mediate tuber formation in TSC? Clearly, many questions remain, but the direct demonstration of the role of FLNA and MEK-ERK1/2 in TSC opens many new and promising research directions and will hopefully lead to novel therapeutic strategies.