Among the explanations offered for tau’s detrimental effects on neurons in Alzheimer disease is that elevations in tau levels, and/or changes in its cellular distribution, disrupt microtubule function and axonal transport. This hypothesis rests on evidence garnered from cultured neurons and from studies with purified microtubules and motor proteins in vitro. That work suggests that simply raising tau levels is enough to inhibit axonal transport and cause degeneration of synapses (see ARF related news story and ARF news story).

However, the result does not readily translate in vivo, according to new data from Aidong Yuan, Ralph Nixon, and colleagues at the Nathan Kline Institute, Orangeburg, New York, and New York University. Measuring axonal transport in the optic nerve of mice that either overexpress the long form of human tau, or have no tau at all, these investigators detect no alteration whatsoever in the overall speed of axonal transport or in the movement of select cargo proteins.

”These findings show that axonal transport is not necessarily dependent on the presence of tau and is not significantly inhibited by moderately elevated levels of tau,” the investigators write in a paper published in the February 13 issue of the Journal of Neuroscience.

To assess slow axonal transport, the researchers injected radiolabeled 35S-methionine into the vitreous of the eye, after which they dissected the optic nerve and looked at how far the radioactivity had traveled up the nerve tract. In tau knockout mice, global transport was unchanged, as measured by bulk radioactivity levels along the optic nerve tract 1 to 2 weeks after injection. Neither was slow transport affected, as measured by the progress of labeled neurofilament proteins and other cargos over the same time frame. Fast transport was measured by tracking 35S-labeling 5 hours after injection, and likewise revealed no alteration in the absence of tau.

Mice overexpressing tau yielded similar results. The radioactivity measurements were supported by additional immunohistochemistry and electron microscopy imaging, which revealed a normal steady state distribution of proteins and organelles in all the mice.

The reasons for the discrepancies between the in vivo and in vitro results of tau overexpression are unclear. The animals used in this study had twofold elevated tau in their retina and fourfold elevated protein in their optic nerve, but no evidence of changes in tau phosphorylation or solubility. The results, the authors write, “do not exclude the possibility that tau may interfere with axonal transport when overexpressed at extremely high levels or when its isoform composition or phosphorylation state is altered significantly.”

In support of the in vivo results, previous in vitro studies of transport in squid axon preparations found no effect of monomeric tau, unless the protein was added at very high concentrations, where it began to inhibit transport non-specifically (Morfini et al., 2007).

Both elevation and depression of tau levels have been reported in AD brain (Ksiezak-Reding et al., 1988; Khatoon et al., 1992). Changes in tau’s subcellular distribution from its normal place in axons to the somatodendritic compartment in the form of neurofibrillary tangles could result in a lack of properly functioning protein in AD brain. However, the current study raises questions about whether these alterations have pathological consequences on axonal transport, and pose anew the question of exactly how tau relates to neurodegeneration. Evidence from AD and also from the tauopathies, a separate family of dementias where mutations in tau itself cause neurodegeneration, suggests that hyperphosphorylation, truncation, or aggregation of tau are important in neurodegeneration. More in vivo studies will surely help settle the question of whether those changes influence axonal transport.—Pat McCaffrey


Make a Comment

To make a comment you must login or register.

Comments on News and Primary Papers

  1. Reply by Aidong Yuan and Ralph Nixon to comments
    Our study was an in vivo test of the hypothesis that moderate overexpression of tau directly impairs axonal transport. We expected our in vivo results to support in vitro data, but the surprising absence of a significant effect was clear and is compatible with existing data. The thoughtful comments in this forum highlight important general issues for consideration in future investigations in this area. As discussed in our report, we agree with the view expressed by Fred Van Leuven that our findings do not exclude the possibility that pathological states of tau might directly or indirectly alter axonal transport. Going forward, however, the task of defining any meaningful direct connection between pathological tau and transport disruption will require that primary effects on transport be distinguished from indirect effects that are secondary to neurodegeneration, which inevitably disrupts transport. In the studies of mice overexpressing tau 4R cited by Van Leuven (1-4), the inference that axonal transport may be disrupted is based on neuropathological detection of focal organelle accumulations in the degenerating axons of tg mice. While the conclusion may well be correct, evidence from the only one of these studies that investigated axonal transport directly (3) suggests that effects on transport in this case are more likely secondary to neurodegeneration than related to a primary effect of tau 4R on transport mechanisms.

    In these time-lapse transport analyses of fluorescent dextran-loaded vesicles in DRG axons of tau 4R tg mice, “the calculated mean of absolute velocities of fluorescent vesicles moving >0.1 μm/min reveals that both retrograde and anterograde vesicle transport did not differ significantly between transgenic and control mice” (3). Also, in vivo tracking of fluorescent vesicle numbers after intranerve injection of fluorescent dextran in live mice showed that tau 4R and wt mice differed at 1 hour, but not at 2 hours or 3 hours (3). Therefore, in the one context where axonal transport was measured directly, overexpression of tau 4R had no or equivocal effects on transport velocities. This observation is consistent with our data and with the Alzforum comment of Akihiko Takashima that “if tau overexpression induces impairment of axonal transport, tau tg mice must show neuronal dysfunction in the entire brain from a young age on” and not only after 20 months of age, when aging-related factors might combine with tau cytotoxicity to cause axonal degeneration and secondary transport failure. Rescue of this pathology by GSK-3β overexpression in tau 4R mice (4) may well be directly related to tau hyperphosphorylation, although neuroprotection could also involve reversal of transport-independent cytotoxic effects of tau.

    We agree with Virginia Lee and John Trojanowski that our results are compatible with findings from their labs. The overexpression of the shortest isoform of tau at levels that were higher than those in our study was associated with neurofibrillary degeneration and reduced fast transport rates measured directly in vivo in side-by-side comparisons with appropriate control groups. Our discussion of these studies was not a misread of the two papers we cited (5-6). It instead drew attention to a comparison specifically of the untreated (or sham-treated) tau tg mice across the two studies showing that the range of fast transport rates for these tg mice overlapped with the rate measured for wt mice. Regardless of the precise extent of transport impairment, this model features significant neurofibrillary degeneration unlike the model we studied, as they pointed out. If high tau levels have direct effects on transport in this model, it would need to be shown in the absence of indirect effects on transport arising from tau-mediated neurodegeneration.

    The Alzforum comments and data from in vivo and in vitro studies emphasize that the degree of tau overexpression is critical to any observed effects on axonal transport. Transport impairments have been associated so far only with very high tau expression, although Fred Van Leuven points out that the emergence of pathological effects at any level of tau overexpression may be influenced by additional vulnerabilities of the animal contributed by genotype and phenotype. Finally, any observed effect of tau at altered expression levels will need to be evaluated in the light of disease relevance, namely, clear evidence that tau levels are comparably altered in Alzheimer disease or other tauopathies. Existing evidence on this issue is not so clear.

    View all comments by Ralph Nixon
  2. In this paper, Yuan et al. report elegant studies of axonal transport in vivo using tau transgenic and tau knockout mice that overexpress human tau isoforms or completely lack tau expression, respectively. These studies sought to elucidate the consequences of too much tau or a complete lack of tau on axonal transport in living mice. This is a most welcome study by the Nixon lab, which has made important contributions to the understanding of axonal transport dynamics for over 2 decades. This study makes increasingly clear that there is a critical need for more studies of this kind to understand how perturbations in tau expression levels or tau pathologies are linked to axonal transport failure and tau-mediated neurodegeneration in Alzheimer disease (AD) and related tauopathies. Indeed, there is growing evidence that failed axonal transport might be the underlying basis for several neurodegenerative diseases in addition to tauopathies (8). It is especially important and timely to undertake in vivo axonal transport studies using the classic Lasek paradigm for measuring rates of axonal transport in living animals (8). This was done in the present study by Yuan et al., as well as in previous studies we conducted on tau transgenic mice (4,13).

    Cell culture studies to examine this issue are being reported, but it is not always clear how well these model systems recapitulate what occurs in living animals. This point is driven home by the data reported here because the results on tau overexpression do not support the cell culture reports by Stamer et al. (10) suggesting that excess tau in the absence of fibrillary tau inclusions “clogs” microtubules (MTs) and impedes axonal transport. The in vivo data reported by Yuan et al. do not confirm or support these prior in vitro studies. It will be important to understand the basis of these discrepant findings in cell culture versus animal model systems. However, it also is noteworthy that Yuan et al. point out that there is no unequivocal evidence that tau is overexpressed in AD or any other known human tauopathy.

    At first blush, the data reported by Yuan et al. also appear to differ from studies we have reported on tau transgenic mice that overexpress the smallest human tau isoform to perturb the 3R-to-4R tau ratio in these mice. However, there are significant differences between the transgenic mice in our studies (T44 line) and those studies in the report by Yuan (8c line). Most significantly, the 8c mice overexpress human tau isoforms but do not develop neurofibrillary tau pathology, as do our T44 transgenic mice (4,5,14). Thus, human tau overexpression that results in the development of neurofibrillary tau pathology can model authentic human neurodegenerative tauopathies, whereas there is no clear human counterpart, disease or otherwise, of tau overexpression alone. To understand the significance of this, some background on tau pathology in AD and related tauopathies is important.

    Briefly, as reviewed recently (2), most early insights into AD and other tauopathies came from studies of AD. At the same time, a substantial amount of data has come more recently from studies of non-AD tauopathies, and it has shown that tau pathology is the critical underlying abnormality that links AD and these disorders to a shared mechanism of neurodegeneration. (View Slide reprinted from Journal of Alzheimer Disease)

    For example, when tau becomes hyperphosphorylated or, as a result of other mechanisms, disengages from MTs, higher concentrations of cytosolic tau lead to tau fibrillization and the formation amyloid-like paired helical filaments that aggregate to form neurofibrillary tangles (NFTs). Thus, as a consequence of this tau amyloidosis in the CNS, normal tau proteins will be sequestered. This depletes the levels of normal tau in affected CNS neurons. As a result, this leaves less normal tau available to stabilize MTs, and, when MTs are destabilized, this compromises intraneuronal transport leading to neurodegeneration. Most elements of this tau-mediated neurodegeneration hypothesis in AD and related tauopathies were demonstrated to occur in experimental animals, when Ishihara et al. showed for the first time that the development of fibrillary tau pathology was linked to MT loss, impaired fast axonal transport (FAT) using the Lasek et al. paradigm, and neurodegeneration (4).

    This study and subsequent studies by our group indicated that the T44 line recapitulates features of AD tau pathology. However, the overall phenotype of the T44 mice is most similar to Guam amyotrophic lateral sclerosis/Parkinson’s dementia complex or ALS/PDC (5,11). Thus, the T44 mice do produce a phenotype that recapitulates features of authentic human neurodegenerative tauopathies.

    Yuan et al. misread the Zhang et al. paper (13) when they infer on page 1686-1687 that the Zhang et al. study of the T44 mice contradicts the data in Ishihara et al. on these same mice. The faster rate of FAT reported in Zhang et al. was the result of a treatment intervention with an MT-stabilizing drug, i.e., taxol. The use of taxol to treat the T44 mice was designed to functionally replace tau and stabilize MTs to offset the sequestration of tau into inclusions in the T44 mice. Specifically, Zhang et al. used the same Lasek et al. paradigm described in Ishihara et al. (and also by Yuan et al.). Zhang et al. confirmed that sham-treated T44 tau-transgenic mice had impaired FAT and the other phenotypic features described by Ishihara et al. (4). By contrast, taxol treatment corrected the FAT deficit as well as the motor impairment in these mice, and this was associated with increased numbers of MTs relative to sham-treated T44 mice (14). Thus, taxol made up for the loss of tau function that resulted from tau sequestration in fibrillary tau inclusions in the T44 line. While there are several problems with taxol for use as therapy for AD and related tauopathies, work is in progress to develop other MT-stabilizing compounds that could go on to become potential disease-modifying therapies for tauopathy patients (1,12).

    These and other studies of transgenic mouse models of tauopathies have increased efforts to develop tau-focused interventions for AD and related tauopathies. Some interventions are directed at abrogating/reversing tau fibrillization or tau hyperphosphorylation; others are designed to stabilize MTs by compensating for the sequestration of tau in NFTs (9). However, while it may be desirable to suppress mutant forms of tau in hereditary tauopathies or in tauopathies with an abnormal ratio of 3R versus 4R tau isoforms (6), reducing tau levels in AD (7), especially to a degree that compromises MT stability, is likely to have long-term deleterious effects as exemplified by the studies cited above (4,13). For example, Hirokowa’s laboratory reported that tau knockout mice develop cognitive and motor abnormalities with age, thereby signifying that reducing tau levels may have negative consequences (3). However, there are few studies examining the behavioral and functional consequences of reducing CNS tau levels over the life span. The present study makes it clear that far more research is needed on this topic, as well as on the role of axonal transport in animal models of tauopathies in order to identify the optimal targets for tau-focused drug discovery for AD and related tauopathies.


    . Paclitaxel C-10 carbamates: potential candidates for the treatment of neurodegenerative tauopathies. Bioorg Med Chem Lett. 2007 Jul 1;17(13):3642-6. PubMed.

    . Tau-mediated neurodegeneration in Alzheimer's disease and related disorders. Nat Rev Neurosci. 2007 Sep;8(9):663-72. PubMed.

    . Muscle weakness, hyperactivity, and impairment in fear conditioning in tau-deficient mice. Neurosci Lett. 2000 Feb 4;279(3):129-32. PubMed.

    . Age-dependent emergence and progression of a tauopathy in transgenic mice overexpressing the shortest human tau isoform. Neuron. 1999 Nov;24(3):751-62. PubMed.

    . Age-dependent induction of congophilic neurofibrillary tau inclusions in tau transgenic mice. Am J Pathol. 2001 Feb;158(2):555-62. PubMed.

    . Allele-specific silencing of dominant disease genes. Proc Natl Acad Sci U S A. 2003 Jun 10;100(12):7195-200. PubMed.

    . Reducing endogenous tau ameliorates amyloid beta-induced deficits in an Alzheimer's disease mouse model. Science. 2007 May 4;316(5825):750-4. PubMed.

    . Axonal transport defects: a common theme in neurodegenerative diseases. Acta Neuropathol. 2005 Jan;109(1):5-13. PubMed.

    . Neurodegenerative diseases: new concepts of pathogenesis and their therapeutic implications. Annu Rev Pathol. 2006;1:151-70. PubMed.

    . Tau blocks traffic of organelles, neurofilaments, and APP vesicles in neurons and enhances oxidative stress. J Cell Biol. 2002 Mar 18;156(6):1051-63. PubMed.

    . Amyotrophic lateral sclerosis/parkinsonism dementia complex: transgenic mice provide insights into mechanisms underlying a common tauopathy in an ethnic minority on Guam. Exp Neurol. 2002 Jul;176(1):1-11. PubMed.

    . Microtubule-stabilising drugs for therapy of Alzheimer's disease and other neurodegenerative disorders with axonal transport impairments. Expert Opin Pharmacother. 2005 May;6(5):683-6. PubMed.

    . Microtubule-binding drugs offset tau sequestration by stabilizing microtubules and reversing fast axonal transport deficits in a tauopathy model. Proc Natl Acad Sci U S A. 2005 Jan 4;102(1):227-31. PubMed.

  3. This interesting paper uses classical assays to measure slow transport and some markers for fast axonal transport. It sees no differences in gross rates of axonal transport in the absence of tau, or upon tau overexpression. Thus, this work differs significantly from observations made by the Mandelkow lab looking at the effects of tau expression on axonal transport and organelle localization.

    The reasons for the apparent discrepancies between these observations remain to be determined. One possibility is the nature of the cargos under investigation, as the paper by Yuan et al. is focusing primarily on cargos undergoing slow transport along the axon. An alternate possibility is the relatively insensitive nature of the transport assay used by Yuan et al. For example, previous work using this approach in the SOD1 model for familial ALS did not reveal significant defects in anterograde transport until relatively late in disease (Zhang et al., 1997; Williamson and Cleveland, 1999), whereas live cell assays for vesicular transport that we have done on neurons from the SOD1 mouse (Perlson et al., submitted) do reveal significant changes in transport velocities. Alternatively, isoform expression may affect the observations, as we and others have seen pronounced differences in the effects of tau isoforms on microtubule motors at the single molecule level.

    In our recent work on the differential effects of tau on kinesin and dynein (Dixit et al., 2008), the motility of individual kinesin motors was strongly affected when the motors encountered patches of tau bound along the microtubule. These encounters most frequently resulted in either pausing of the motor or detachment of the kinesin from the microtubule. It is possible that for larger cargos undergoing longer-distance transport, these molecular-level changes are damped out. This is an interesting question that needs to be investigated further with high-resolution assays. Alternatively, as Yuan et al. suggest, there may be compensatory mechanisms operating in vivo that mitigate the effects seen at the molecular level.

    Thus, their argument is a good one: we need to keep examining in vivo biology in concert with higher-resolution studies that offer more mechanistic insights.

    View all comments by Erika Holzbaur
  4. In this paper, Randy Nixon’s group first demonstrated in vivo that axonal transport rates are not significantly affected by tau deletion or overexpression in mouse brain. The results are highly convincing.

    In in vitro studies, the Mandelkows’ group and Hirokawa’s group have suggested that tau overexpression inhibits anterograde transport in cultured cells and neurons. Recently, Holzbaur’s group indicated that when kinesin motor protein encountered tau patches on microtubules, composed of 10 tau molecules, it detached from microtubules (Dixit et al., 2008). However, monomeric tau levels 20-fold above physiological concentration did not affect axonal transport in squid axon (Morfini et al., 2007). Taken together, aggregated tau on microtubules, but not monomeric tau, may induce inhibition of axonal transport.

    Ishihara and colleagues showed that expressing the shortest human tau fivefold to 10-fold over endogenous tau inhibited axonal transport (Ishihara et al., 1999), although Nixon’s group reports no inhibition of axonal transport in fourfold overexpression of human tau. Therefore, it is possible that the effect of tau on axonal transport in vivo may be dependent on the level of tau overexpression.

    Even if we accept that tau overexpression induces axonal transport, and that it may be a cause of neuronal dysfunction or synapse loss in tauopathy, the question still remains how tau impairs neuronal function in a specific brain region in tauopathy. If tau overexpression induces impairment of axonal transport, tau Tg mice must show neuronal dysfunction in the entire brain from a young age on. We recently showed that human wild-type tau (the longest form), expressed about threefold over endogenous levels, induces neural dysfunction in entorhinal cortex at old age (more than 20 months), accompanied by synapse loss and accumulation of hyperphosphorylated tau resulting in a memory deficit, while adult mice (>12 months old) are no different from non-Tg (Kimura et al., 2007). The level of tau on microtubules shows no significant difference between adult and aged mice. Therefore, our results suggest that hyperphosphorylation of tau at old age itself, rather than tau-induced impairment of axonal transport, may be a cause of neuronal dysfunction in the entorhinal cortex, which shows the earliest pathological change in AD.


    . Differential regulation of dynein and kinesin motor proteins by tau. Science. 2008 Feb 22;319(5866):1086-9. PubMed.

    . Tau binding to microtubules does not directly affect microtubule-based vesicle motility. J Neurosci Res. 2007 Sep;85(12):2620-30. PubMed.

    . Age-dependent emergence and progression of a tauopathy in transgenic mice overexpressing the shortest human tau isoform. Neuron. 1999 Nov;24(3):751-62. PubMed.

    . Hyperphosphorylated tau in parahippocampal cortex impairs place learning in aged mice expressing wild-type human tau. EMBO J. 2007 Dec 12;26(24):5143-52. PubMed.

  5. The picture is more complicated than the title of Yuan et al. would lead us to believe. Our group has generated many tau transgenic mice strains, and at least Tau-4R mice have impaired axonal transport (Spittaels et al., 1999; Künzi et al., 2002), which, moreover, can be rescued by GSK-3β (Spittaels et al., 2000).

    Whether or not axonal transport is impaired depends not only on expression levels, as our Tau-4R mice expressed only about twofold over endogenous mouse tau, and we did not observe aggregates of tau.

    Other factors must play a role, from the actual tau isoform and promoter used, up to integration site effects. The latter is illustrated by the "selection" of tau mutant mice (Schindowsky et al., 2006). Other, as yet unknown factors play a role, based on heterogeneity of phenotype, gender differences, variability in response to treatments, etc.

    There is clearly more to tau and transport than currently meets the eye (just as is the case with APP).


    . Prominent axonopathy in the brain and spinal cord of transgenic mice overexpressing four-repeat human tau protein. Am J Pathol. 1999 Dec;155(6):2153-65. PubMed.

    . Glycogen synthase kinase-3beta phosphorylates protein tau and rescues the axonopathy in the central nervous system of human four-repeat tau transgenic mice. J Biol Chem. 2000 Dec 29;275(52):41340-9. PubMed.

    . Unhampered prion neuroinvasion despite impaired fast axonal transport in transgenic mice overexpressing four-repeat tau. J Neurosci. 2002 Sep 1;22(17):7471-7. PubMed.

    . Alzheimer's disease-like tau neuropathology leads to memory deficits and loss of functional synapses in a novel mutated tau transgenic mouse without any motor deficits. Am J Pathol. 2006 Aug;169(2):599-616. PubMed.


News Citations

  1. Tau Accused of Blocking Transport, Causing APP to Linger and Nerve Processes to Wither
  2. Tau Roundup: Inducible Mice Accentuate Aggregation and More

Paper Citations

  1. . Tau binding to microtubules does not directly affect microtubule-based vesicle motility. J Neurosci Res. 2007 Sep;85(12):2620-30. PubMed.
  2. . Immunochemical and biochemical characterization of tau proteins in normal and Alzheimer's disease brains with Alz 50 and Tau-1. J Biol Chem. 1988 Jun 15;263(17):7948-53. PubMed.
  3. . Brain levels of microtubule-associated protein tau are elevated in Alzheimer's disease: a radioimmuno-slot-blot assay for nanograms of the protein. J Neurochem. 1992 Aug;59(2):750-3. PubMed.

Further Reading

Primary Papers

  1. . Axonal transport rates in vivo are unaffected by tau deletion or overexpression in mice. J Neurosci. 2008 Feb 13;28(7):1682-7. PubMed.