|Neurodegeneration: From Hyperexcitability to
Pathologic Tau, TDP-43 and α-Synuclein
Hyperphosphorylated Tau (Grundke-Iqbal et al., 1986), phosphorylated and cleaved TDP-43 (Hasegawa et al., 2008; Yamashita et al., 2012), and oligomeric α-synuclein (Roberts and Brown, 2015) are the pathological hallmarks for Alzheimer's disease, amyotrophic lateral sclerosis (ALS) and Parkinson's disease, respectively. Frontotemporal dementia (FTD) is characterized by Tau and TDP-43 pathologies, with roughly equal prevalence (Seeley, 2008). Paper 5 suggests that they may all arise from neuronal hyperexcitability. This paper will discuss how hyperexcitability can lead to these pathologies.
Hyperexcitability results in excessive calcium ions (Ca2+) inside the neuron. It has been well documented that calcium dysregulation may cause neurodegeneration (Mattson, 2007), but the underlying mechanism remains unclear. Previous works postulated that the aberrant Ca2+ influx might originate from the toxic forms of amyloid β-peptide (Aβ) (Bezprozvanny and Mattson, 2008; Zündorf and Reiser, 2011). This idea could be influenced by the Amyloid Cascade Hypothesis (Hardy and Higgins, 1992) which asserts that Aβ toxicity is the origin of Alzheimer's disease. Since its inception, numerous Aβ-targeting drugs have been developed, but to date none of them showed any reduction in neurofibrillary tangles or improvement in cognitive performance of patients with Alzheimer's disease (Dai et al., 2017).
In contrast to the Amyloid Cascade Hypothesis, the Wireless Communication Model centers on the Tau protein. Within this framework, the Ca2+ overload originates from hyperexcitability, rather than Aβ. The last two decades have revealed great insight into the pathogenic cascade following Ca2+ overload, including α-synuclein aggregation into oligomers. Interestingly, the oligomeric α-synuclein was found to inhibit tubulin polymerization (Chen et al., 2007). More than three decades ago, experiments have demonstrated that tubulin polymerization is linked to excitability (Sakai et al., 1985), consistent with the notion that tubulin and microtubule may play a critical role in regulating excitability (see Paper 2). This underappreciated function can readily explain why oligomeric α-synuclein impairs dopamine release.
From Ca2+ overload to Tau hyperphosphorylation
Glycogen synthase kinase-3 (GSK-3) and cyclin dependent kinase 5 (Cdk5) are the major protein kinases capable of phosphorylating Tau at multiple sites (Lovestone et al., 1994; Kimura et al., 2014). GSK-3 has been demonstrated to be essential in the pathogenesis of Alzheimer's disease (Takashima, 2006). It is a promising therapeutic target for this devastating disease (Maqbool et al., 2016). Both GSK-3 and Cdk5 can be regulated by calpain (Figure 1) which is a Ca2+-dependent protease that cleaves proteins (Ferreira, 2012). Hence, Ca2+ overload may stimulate calpain to activate GSK-3 and/or Cdk5, resulting in Tau hyperphosphorylation.
From Tau hyperphosphorylation to microtubule breakdown
The Tau protein is expressed only in neurons, and localized predominantly to the axon. Normal Tau binds with tubulin and promotes its assembly into a microtubule. Hyperphosphorylated Tau may aggregate with normal Tau to form neurofibrillary tangles. This will reduce the amount of normal Tau and result in microtubule disintegration at the axon (Alonso et al., 1996). Furthermore, the hyperphosphorylated Tau can pass the axon initial segment (AIS) (Li et al., 2011) and translocate to dendrites where they may trigger microtubule severing by TTLL6 and spastin (Zempel et al., 2013). The microtubule breakdown will eventually lead to neuronal death.
From Ca2+ overload to abnormal TDP-43
TDP-43 can be cleaved by calpain (Yamashita et al., 2012). Its pathologic phosphorylation is catalyzed by casein kinase-1 (CK1) (Hasegawa et al., 2008). As described above, calpain is directly regulated by Ca2+. CK1 also depends on Ca2+, but indirectly. Ca2+ may stimulate protein phosphatase 2B (PP2B or calcineurin) to dephosphorylate CK1, resulting in its activation (Liu et al., 2002; Nakano et al., 2010).
AMPA receptors are implicated in both ALS (Kawahara et al., 2004) and FTD (Gascon et al., 2014). They are composed of four types of subunits, designated as GluR1 - GluR4. The presence of the GluR2 subunit renders the AMPA receptor impermeable to Ca2+. Over-expression and/or over-stimulation of Ca2+ permeable AMPA receptors may cause excessive Ca2+ influx through dendritic spines, thereby resulting in neurodegeneration.
AMPA receptors belong to glutamate receptors which can be activated by the excitatory neurotransmitter glutamate. As discussed in Paper 5, ALS may begin in the motor cortex where a population of neurons oscillate at the beta band. These neurons will be referred to as "beta neurons". Hyperexcitable beta neurons may release glutamate to stimulate motor neurons in the motor cortex, i.e., Betz cells, leading to Ca2+ overload inside the Betz cells. In many types of cells, the toxicity of Ca2+ overload can be alleviated by Ca2+ buffering proteins, which are a special class of Ca2+ binding proteins that bind transiently with Ca2+ ions to prevent further Ca2+ signaling (Schwaller, 2010). Motor neurons express low level of Ca2+ buffering proteins (e.g., calbindin and parvalbumin) (Jaiswal, 2014), which makes them particularly vulnerable to Ca2+ toxicity.
Tau-positive vs.Tau-negative ALS
From the previous section, it is clear that neurons may display Tau pathology if their AIS is enriched with calcium channels, calpain, GSK-3 and Tau proteins. Generally speaking, they are the neurons whose AIS contains microtubule antennas for long range synchronization, which requires Ca2+ to regulate excitability via microtubule dynamics (Paper 2). Motor neurons do not participate in beta synchronization. Hence Tau-negative ALS may occur if Ca2+ overload in Betz cells results mainly from over-expression of Ca2+-permeable AMPA receptors, rather than over-stimulation by beta neurons, consequently leading to severe TDP-43 pathology in motor neurons, while beta neurons are spared.
In the Guam type ALS, neurofibrillary tangles are abundant not only in the primary motor cortex (Hof and Perl, 2002), but also in the entorhinal, frontal and temporal cortex (Hof et al., 1991; Hof et al., 1994). The Guam type ALS was originally discovered on the island of Guam in the Pacific (Arnold et al., 1953). It is caused by a non-protein amino acid, β-N-methylamino-L-alanine (BMAA), which can be misincorporated into normal proteins, resulting in protein misfolding (Dunlop et al., 2013). Misfolded proteins are known to enhance mTOR activity, thereby increasing the risk for Tau pathology (Paper 3). The BMAA-induced ALS also exhibits features of Parkinson's disease and Alzheimer disease (Miklossy et al., 2008). This demonstrates the common origin of these neurodegenerative disorders.
From abnormal TDP-43 to neurofilament dysregulation
Neurofilaments are a major component of the neuronal cytoskeleton, located primarily in the axon. Their dysregulation, which impairs axonal transport, has been shown to play an important role in the pathogenesis of ALS (Collard et al., 1995). In patients with ALS, the level of neurofilaments in the cerebrospinal fluid (CSF) is significantly elevated (Xu et al., 2016). A neurofilament is composed of polypeptide chains (subunits), which are classified into light, medium or heavy chains, on the basis of their molecular weight. The FTD, particularly its TDP-43 subtype, is associated with elevated neurofilament light chains in CSF (Landqvist-Waldö et al., 2013).
TDP-43 is involved in the biogenesis of microRNAs, including miR-132 and miR-9 (Paper 4). miR-132 regulates Tau expression while miR-9 targets neurofilaments. A single miR-9 binding site was observed on the mRNA of the neurofilament light chain, compared to nine miR-9 binding sites in the heavy chain mRNA (Haramati et al., 2010). Experiments have shown that TDP-43 mutations result in miR-9 down-regulation (Zhang et al., 2013), which may account for neurofilament dysregulation in TDP-43 pathology.
α-Synuclein localizes primarily to the presynaptic terminal (Iwai et al., 1995). Their aggregation into oligomers is the major cause of neurodegeneration. A plethora of possible mechanisms were proposed, but no consensus has been reached yet (Roberts and Brown, 2015). Several lines of evidence suggest that tubulin polymerization into microtubules plays a pivotal role.
From Ca2+ overload to α-synuclein aggregation
At the nerve terminal, Ca2+ entry is required to trigger the release of neurotransmitters, including dopamine (Okita et al., 2000). In most cases, Ca2+ enters the nerve terminal via L-type channels, whose antagonists are protective for Parkinson's disease (Ilijic et al., 2011; Swart and Hurley, 2016). This supports the view that Ca2+ overload at the nerve terminal leads to α-synuclein pathology. Mechanistically, the terminal Ca2+ overload, together with induced oxidative stress, promote α-synuclein aggregation (Nath et al., 2011; Goodwin et al., 2013; Rcom-H'cheo-Gauthier et al., 2014; Rcom-H'cheo-Gauthier et al., 2016).
From α-synuclein aggregation to microtubule deficiency
The toxicity of α-synuclein aggregation is a subject of intensive research. As described above, tubulin and microtubule may play a central role. A tubulin dimer contains about 50 more negatively than positively charged residues (Minoura and Muto, 2006). In a solution, tubulin dimers are surrounded by counterions and polar water molecules which may reduce the electrostatic interaction with external fields. The effective charge on a tubulin dimer was estimated to be 12 - 20 e– (van den Heuvel et al., 2006; Minoura and Muto, 2006 ). Such negative charges can exert a hyperpolarizing field on voltage-activated ion channels, thereby regulating neuronal excitability (see Paper 1, Figure 1). Therefore, at the axon terminal, depolymerized free tubulin dimers may move close to the membrane to inhibit neuronal firing via their hyperpolarizing fields. Tubulin polymerization into microtubules will reduce the amount of free tubulin dimers, thus enhancing excitability. Oligomeric α-synuclein has been shown to inhibit tubulin polymerization (Chen et al., 2007). This explains why α-synuclein oligomers reduce neuronal excitability (Kaufmann et al., 2016) and why dopamine release is impaired in Parkinson's disease.
In Tau pathology, hyperexcitability due to excessive Tau proteins eventually leads to hypoexcitability caused by Tau hyperphosphorylation (Paper 2). Similarly, in α-synuclein pathology, hyperexcitability due to excessive glutamate stimulation also leads to hypoexcitability, but in this case it is caused by high level of free tubulin dimers. Reduction in tubulin polymerization affects not only excitability, but also the amount of functional microtubules. Microtubule deficiency may impair axonal transport of essential cellular components (Pellegrini et al., 2017), eventually leading to neuronal death.
When a neuron fires, Ca2+ may enter the neuron through voltage-gated calcium channels, such as T-type calcium channels at AIS or L-type calcium channels at axon terminals. Moreover, the firing of glutamatergic neurons may release glutamate to excite postsynaptc neurons. Over-stimulation of Ca2+-permeable glutamate receptors may lead to Ca2+ overload in the postsynaptc neurons. The Wireless Communication Model suggests that hyperexcitability of the neurons (theta or beta) whose AIS contains microtubule antennas for long range synchronization may cause Ca2+ overload via AIS, leading to Tau pathology. The TDP-43 pathology may result from over-stimulation of glutamate receptors (the AMPA type) at the dendritic spines that synapse with hyperexcitable theta or beta neurons. The α-synuclein pathology may arise from Tau-mediated hyperexcitability in the subthalamic nucleus (STN), resulting in glutamate-mediated hyperactive dopaminergic (DA) neurons in substantia nigra pars compacta (SNc). Subsequently, the hyperactive DA neurons may cause excessive Ca2+ influx through L-type calcium channels at axon terminals. The pathogenic cascade following Ca2+ overload is depicted in Figure 2.
The Guam type ALS, also known as ALS/parkinsonism-dementia complex of Guam (ALS/PDC), exhibits all three pathologies discussed above (Miklossy et al., 2008). This demonstrates their common origin. ALS/PDC is caused by the non-protein amino acid, BMAA, which can be misincorporated into normal proteins, resulting in protein misfolding (Dunlop et al., 2013). Experiments have shown that protein misfolding may promote mTOR activation, and consequently hyperexcitability (Paper 3). Figure 3 illustrates the BMAA-induced pathogenic cascade. This figure does not imply that activation of mTOR will definitely lead to all of these pathologies. Rather, mTOR activation simply increase the probability of developing pathologies. Many other factors are also involved in neurodegeneration, such as the expression level of miR-132, TDP-43, α-synuclein, GluR2, calpain, GSK-3, etc. The reason why BMAA induces all three pathologies is probably because mTOR activation by BMAA-containing proteins is overwhelming.
Author: Frank Lee