It cuts two ways: microtubule loss during Alzheimer disease

EMBO J 32: 2920–2937 10.1038/emboj.2013.207; published online September242013Microtubule loss from axons and dendrites is a key contributor to nervous system degeneration during Alzheimer disease. Previous evidence suggested a simple pathway by which tau dissociation from microtubules in the axon allows excess severing of microtubules by katanin. Now, new evidence has emerged for a more complex pathway by which abnormal tau invasion into dendrites, triggered by Aβ oligomers, results in excess severing of microtubules by spastin.Alzheimer disease (AD) is a member of a category of neurodegenerative disorders called tauopathies (Wang and Liu, 2008). These are diseases of the nervous system in which tau becomes abnormally phosphorylated, and thereby detaches from microtubules. As the microtubules lose tau, they diminish in number and density, and this loss of microtubule mass negatively impacts the capacity of the neuron to maintain axonal transport and synaptic connections. Terms such as disintegrate or ‘fall apart'' are often used to describe the effect on the microtubules as they lose tau, but to date there has been very little information on how this happens. There is no mechanistic evidence to support the view that the microtubules become less stable and simply disassemble by their normal dynamic properties.One possibility is that tau normally protects microtubules from being destroyed by various proteins in the axon that would otherwise cut them into pieces or in some other way break them down. This presumably reflects a physiological mechanism wherein the regulation of tau dissociation from the microtubule via signalling pathways controls when and where microtubule breakage normally occurs. When a pathological condition causes tau to detach from microtubules, they become extremely sensitive to such factors. In addition, there is strong evidence that the abnormal tau, whether soluble or filamentous, can elicit toxic gain-of-function effects on the axon (Wang and Liu, 2008).To make matters even more complex, AD is not a pure tauopathy. Beta amyloid (Aβ) accumulates abnormally in the brain during AD, and this prompts tau to become hyperphosphorylated and lose association with microtubules. However, the Aβ can also elicit microtubule loss, independent of tau dissociation from the microtubules. In AD, there is also a loss of microtubules from dendrites, and this introduces an additional degree of complexity. Tau is normally less enriched in dendrites than axons. In AD, tau invades dendrites abnormally through deregulation of its normal sorting mechanism, and this somehow leads to microtubule loss from dendrites (Zempel et al, 2010).Microtubule loss is a common end point of multiple pathways, some involving loss of tau function, others involving gain-of-function effects of abnormal tau, and still others working through tau-dependent Aβ toxicity. All of this is not to say that the effects on microtubules are the only reason or even the principal explanation for axonal degeneration in AD, but the loss of microtubules is an important contributor to nervous system degeneration. Preventing or reversing the effects on microtubules could help stave off degeneration and hence provide patients with additional years of cognitive health and better quality of life.Microtubule assembly and disassembly occur from the ends of a microtubule, mainly (and often exclusively) at the plus end of the microtubule in living cells. Proteins that regulate microtubule stability affect the rate of these dynamics at microtubule ends. In recent years, a great deal of attention has focused on a category of proteins, termed microtubule-severing proteins. These proteins are enzymes that yank at the microtubule anywhere along its length to pull out a tubulin subunit, and thereby ‘cut'' the microtubule by causing it to break into pieces (Roll-Mecak and Vale, 2008). If the microtubule is sufficiently stable in the region of the break, the parent microtubule is cut into two shorter microtubules that persist, with minimal disassembly of either of the two pieces. If a microtubule is severed in its more labile region, the breakage could cause a great deal of disassembly. If the tubulin being yanked is situated at one of the ends of a microtubule, the result would be a shortening of the microtubule from that end; that is, disassembly. Thus, microtubule severing in the axon can certainly lead to microtubule loss, either by cutting the polymer all the way to subunits, inducing disassembly directly from microtubule ends or promoting disassembly as a secondary effect to the cutting.To date, the AAA enzymes katanin and spastin are the best studied of the microtubule-severing proteins (Yu et al, 2008). Spastin was originally identified as the product of the gene whose mutations are the chief cause of hereditary spastic paraplegia. Curiously, neurons express levels of spastin and katanin that are theoretically high enough to completely sever all of the microtubules in the neuron to subunits (Solowska et al, 2008), and yet this does not happen. Various regulatory mechanisms presumably keep the activities of the severing proteins in check. One of these mechanisms, in the case of katanin, is microtubule-bound tau, which protects the microtubule from being accessed by katanin (Qiang et al, 2006).Could microtubule loss in AD be due, at least in part, to excess microtubule severing due to deregulation of microtubule-severing proteins? We have posited that heightened severing of the microtubules by katanin, as the microtubules lose association with tau, is a contributing factor to the degradation of microtubules in the axons of AD patients (Sudo and Baas, 2011). A role for spastin in this pathway is questionable, because tau does not appear to protect microtubules against spastin as effectively as it does against katanin (Qiang et al, 2006). However, we now know that spastin is far from irrelevant to AD, as an exciting new article from the Mandelkow and Dawson laboratories implicates spastin in an entirely different pathway for microtubule loss in AD (Zempel et al, 2013). Whereas the katanin pathway is more relevant to axons, this new spastin pathway is more relevant to dendrites.In this new work, Zempel et al (2013) exposed mature primary neurons to oligomers of Aβ and observed microtubule breakdown in dendrites that had been invaded by tau. They found that the missorting of tau leads to an elevation of TTLL6 (Tubulin-Tyrosine-Ligase-Like-6) in dendrites, and this results in a marked increase in the polyglutamylation status of the microtubules. Because spastin has a strong preference for polyglutamylated microtubules, the microtubules become more sensitive to spastin-induced severing. Exactly why katanin is not a factor remains unclear, as polyglutamylation renders microtubules more sensitive to both of the severing proteins, not just spastin (Lacroix et al, 2010). Perhaps some of the tau that invades the dendrite is able to bind to microtubules and protect them from katanin, or perhaps katanin is less potent in dendrites because their microtubules are poorly acetylated, as katanin prefers acetylated microtubules to unacetylated ones (Sudo and Baas, 2011). Whatever the case, these new studies suggest that spastin, a protein whose mutations cause an entirely different neurodegenerative disease, is also a major factor in AD.What are the implications of these findings for AD treatment? In recent years, there have been encouraging results on rodent models for AD, in which behavioural improvement and enhanced neuronal vitality were observed when the animals were treated with drugs that stabilize microtubules against disassembly (Zhang et al, 2012). Such drugs are currently in clinical trials for AD (Barten et al, 2012). This strategy is based on the presumption that the microtubule loss that occurs in AD is due to destabilization of the microtubules. However, the results discussed here suggest that the primary cause of the microtubule loss could be something quite different, namely excess severing of microtubules. In this regard, it is relevant that both katanin and spastin seem to have a preference for severing stable microtubules (Lacroix et al, 2010; Sudo and Baas, 2011). Therefore, while treatment with a microtubule-stabilizing drug would mitigate disassembly that occurs as an aftereffect of microtubule severing, the severing events themselves would likely be increased (Figure 1). Heightened microtubule severing in axons and dendrites, even if the total levels of microtubule mass are preserved, could result in a gradual shift from a normal distribution of long and short microtubules to a predominance of microtubules too short to support sustained excursions of organelle transport. Long microtubules are also necessary as compression-bearing struts that prevent axons and dendrites from collapsing on themselves. We suspect that appropriate treatment regimes can be devised to prevent such dire consequences from happening, but we would advocate for the development of new drugs that inhibit microtubule-severing proteins. Such drugs may prove to be a better approach (on their own or in combination with a stabilizing drug) for preserving the fidelity of axonal and dendritic microtubules in AD patients. Given that the structure of the severing proteins is known, it may be straightforward to develop inhibitors, especially to their ATPase domains.Open in a separate windowFigure 1Microtubules in axons and dendrites consist of a stable region towards the minus end of the microtubule and a labile region towards the plus end, as well as a pool of free tubulin subunits. Microtubule severing is a normal event in the neuron, when tightly regulated. Abnormal (deregulated) microtubule severing is posited to account for microtubule loss in AD. Severing in the stable region of the microtubule would create two new microtubules, with fairly minimal disassembly of either one. Severing in the labile region of the microtubule would result in notably more disassembly. Severing at the end of the microtubule would result in disassembly. Because known microtubule-severing proteins favour the stable region of the microtubule, treatment of AD with a microtubule-stabilizing drug may mitigate disassembly that occurs as an aftereffect of the severing, but the severing events themselves would likely increase.     

The cytoskeleton of the neuron--an essay in celebration of Paul Letourneau's career

The neuronal cytoskeleton consists of microtubules, actin filaments, neurofilaments, and an array of accessory proteins that regulate and modify these three main filament systems. This essay celebrates the career of Paul Letourneau, a pioneer of the neuronal cytoskeleton, to whom the community owes a debt of gratitude.     

Tau Does Not Stabilize Axonal Microtubules but Rather Enables Them to Have Long Labile Domains

1. Download high-res image (265KB)
2. Download full-size image

     

Cytomegalovirus Viral Load Kinetics in Patients with HIV/AIDS Admitted to a Medical Intensive Care Unit: A Case for Pre-Emptive Therapy

Background

Cytomegalovirus (CMV) infection is associated with severe diseases in immunosuppressed patients; however, there is a lack of data for pre-emptive therapy in patients with HIV/AIDS.

Method

This was a retrospective study, which enrolled patients diagnosed with HIV/AIDS (CD4<200 cells/μl), who had detectable CMV viral load (VL) during their stay in an adult medical intensive care unit between 2009–2012.

Results

After screening 82 patients’ records, 41 patients met the enrolment criteria. Their median age was 37 (interquartile range [IQR]: 31–46), and median CD4 count was 29 cells/μl (IQR: 5–55). Sixteen patients (39%) had serial measurements of CMV VL before treatment with ganciclovir. Patients whose baseline CMV VL values were between 1,000–3,000 copies/ml had significantly higher values (median of 14,650 copies/ml) on follow-up testing done 4–12 days later. Those with undetectable VLs at baseline testing had detectable VLs (median of 1,590 copies/ml) mostly within 20 days of follow-up testing. Patients who had VLs >1,000 copies/ml at baseline testing had significantly higher mortality compared to those who had <1,000 copies/ml {hazard ratio of 3.46, p = 0.003 [95% confidence interval (CI): 1.55–7.71]}. Analysis of the highest CMV VL per patient showed that patients who had VLs of >5,100 copies/ml and did not receive ganciclovir had 100% mortality compared to 58% mortality in those who received ganciclovir at VLs of >5,100 copies/ml, 50% mortality in those who were not treated and had low VLs of <5,100 copies/ml, and 44% mortality in those who had ganciclovir treatment at VLs of <5,100 copies/ml (p = 0.084, 0.046, 0.037, respectively).

Conclusion

This study showed a significantly increased mortality in patients with HIV/AIDS who had high CMV VLs, and suggests that a threshold value of 1,000 copies/ml may be appropriate for pre-emptive treatment in this group.     

Fine specificity of TCR complementarity-determining region residues and lipid antigen hydrophilic moieties in the recognition of a CD1-lipid complex

alphabeta TCR can recognize peptides presented by MHC molecules or lipids and glycolipids presented by CD1 proteins. Whereas the structural basis for peptide/MHC recognition is now clearly understood, it is not known how the TCR can interact with such disparate molecules as lipids. Recently, we demonstrated that the alphabeta TCR confers specificity for both the lipid Ag and CD1 isoform restriction, indicating that the TCR is likely to recognize a lipid/CD1 complex. We hypothesized that lipids may bind to CD1 via their hydrophobic alkyl and acyl chains, exposing the hydrophilic sugar, phosphate, and other polar functions for interaction with the TCR complementarity-determining regions (CDRs). To test this model, we mutated the residues in the CDR3 region of the DN1 TCR beta-chain that were predicted to project between the CD1b alpha helices in a model of the TCR/CD1 complex. In addition, we tested the requirement for the negatively charged and polar functions of mycolic acid for Ag recognition. Our findings indicate that the CDR loops of the TCR form the Ag recognition domain of CD1-restricted TCRs and suggest that the hydrophilic domains of a lipid Ag can form a combinatorial epitope recognized by the TCR.