Early in the course of Alzheimer disease, blockages in axonal traffic lead to sick axons swollen with the jumbled pile-up of traffic components. The blockages precede overt amyloid pathology in AD mouse models by a year and mark the brains of people who died at early stages of AD. Axonal traffic blockage should be studied as a potential cause for this devastating disease and might open new research avenues toward early diagnosis. These are the provocative claims of a paper by Larry Goldstein’s group at the University of California, San Diego, which will appear tomorrow in Science magazine. In collaboration with Eliezer Masliah and others at UCSD and Peter Davies at Albert Einstein College of Medicine in New York, the Goldstein group extends to mice and humans a line of investigation they originally established in fruit flies (see ARF related news story and ARF news story).

The axonal transport hypothesis is appealing to some scientists because its underlying mechanism engages both major pathologies of Alzheimer disease—neurofibrillary tangles and amyloid plaques. Tau protein, the major component of tangles, is well-established as regulating axonal traffic through its microtubule-binding function (see, for example, ARF related news story and ARF news story). Based on his earlier work, Goldstein had suggested that APP does so, too, possibly by way of serving as a binding partner for the anterograde motor protein kinesin. A slowdown of axonal traffic, so the hypothesis goes, may cause aberrant generation of the Aβ peptide en route to nerve terminals, leading to synaptic damage and, ultimately, plaques. This research has created a buzz but also ruffled feathers in the AD field. Some researchers have said that the findings are hard to reproduce, and many have eagerly awaited follow-up data from mouse and humans to put the hypothesis to rigorous scrutiny.

In the present paper, first author Gorazd Stokin and colleagues looked for axonal defects that represent transport deficits. In doing so, they harked back to early observations about axonal pathology by AD research pioneer Bob Terry (Terry et al., 1964) and others. The UCSD researchers first examined various fiber tracts in two different transgenic APP mouse models. As had earlier investigators before them, they, too, found axonal swellings distended up to 3 micrometers in diameter and filled with axonal transport components. These occurred in axons of the nucleus basalis of Meynert (NBM), an area that provides cholinergic input to the cerebral cortex and atrophies in AD, as well as in cortex and hippocampus of transgenic, but not wild-type mice. At four months of age, long before amyloid deposits have formed, and prior to the established loss of fibers in the NBM of these mouse models, these axonal swellings were as abundant as at 20 months, according to the paper. The sample sizes in all groups were 4 to 5 animals each. The scientists also assessed fiber changes in the NBM of a small human sample of three people each who had died at Braak AD stages 0, I-III, or IV-VI. Similarly, nearly half the cholinergic fibers at stages I-III had axonal swellings, well prior to when amyloid deposition becomes detectable.

What’s in those swellings? Electron microscopy from the young mice indicated that they brim with mitochondria, vacuoles, and vesicles. Dense bodies, dense axoplasm, and debris gave some swellings the appearance of early-stage axonal degeneration. Light microscopy indicated accumulation of kinesin, hinting at a transport deficit.

To explore the role of transport and kinesin further, Stokin and colleagues recapitulated in mice their earlier genetic experiment in flies, in which they bred strains to cut the dosage of kinesin light chain by half. As before in the flies, tightening the kinesin supply in this way almost doubled the formation of axonal swellings in the APP-transgenic mice. In one test of how the amount of kinesin affects APP transport, the scientists cultured hippocampal neurons from kinesin wild-type and kinesin-reduced mice and then transfected the cells with APP linked to yellow fluorescent protein. In the neurons from the kinesin-reduced mice, anterograde transport of APP-containing particles decreased, and retrograde transport increased. Beyond that, the paper does not quantify how axonal transport rates change in the APP-kinesin knockout crosses. Even so, this data indicates that if kinesin is in short supply, APP transport changes and swellings form, the authors state. To date, one small genetic association study has linked kinesin polymorphisms to AD, but it has not yet been independently confirmed (see Alzgene entry).

Furthermore, the paper suggests that slowing down traffic to axon terminals by limiting the kinesin supply leads to a selective increase in the ratio of Aβ42 to Aβ40 peptides, and also to the peptide’s intraneuronal accumulation, in APP-transgenic mice. The topic of intraneuronal Aβ accumulation has garnered growing interest in the research community in recent years, though some scientists question its relevance to the primary disease process (see ARF related conference story). In the present model, kinesin reduction not only boosts Aβ42 generation prior to amyloid deposition, it also accelerates amyloid deposition. The brains of old APP-transgenic mice with reduced kinesin contained more plaques, larger plaques, and more diffuse extracellular deposits than did APP-transgenics with normal kinesin levels. The brains of middle-aged APP-transgenic mice with reduced kinesin had as much amyloid deposition as did old APP-transgenics with both copies of kinesin. This, and further analysis, suggests “that swellings precede, and participate in the formation of, amyloid plaques,” the authors write.

The authors also write that their data might warrant a reconsideration of dystrophic neurites, which are widely assumed to form because of the presence of plaques. By contrast, axonal swellings do not form in response to plaques, the authors contend, suggesting that they may instead be precursors to some dystrophic neurites.

Reports implicating axonal transport to AD are not new (for a recent review on axonal transport deficits in the neurodegeneration literature, see Roy et al. 2005). What is new is that Stokin and colleagues have placed them prior to other aspects of AD pathogenesis in vivo and staked out a claim for a possible cause for AD. The study does not prove a cause-effect relationship. It also does not address how axonal swellings might relate to other factors that are in contention as early-stage insults, including oligomeric forms of Aβ or tau, oxidative stress, or risk factors such as the apoE4 allele. The study also does not go so far as to compare behavioral differences between the APP-transgenics and their kinesin-reduced brethren.

In summary, the paper puts forth this scenario for testing by the research community: Impaired axonal transport could lead to axonal swellings, which promote aberrant Aβ generation locally at the sites of blockage. This local Aβ increase leads to Aβ secretion or lysis of the swellings, which could trigger amyloid deposition in these spots. Additional transport vesicles coming upon a blockage would get diverted back to the cell body and dendrites, where aberrant Aβ generation, accumulation, and deposition might then occur. People with genetic reductions of kinesin would be especially prone to this latter process. The authors connect these processes into a hypothetical autocatalytic spiral, in which blockages and APP processing would keep stimulating each other and lead to synaptic loss. Finally, differences in how fast axonal transport slows down with age might underlie some cases of sporadic AD, the researchers speculate.—Gabrielle Strobel.

Stokin GB, Lillo C, Falzone T, Brusch R, Rockenstein E, Mount S, Raman R, Davies P, Masliah E, Williams DS, Goldstein LSB. Axonopathy and Transport Deficits Early in the Pathogenesis of Alzheiimer's Disease. Science. 2005 Feb 25, (307):1282-1288. Abstract


Make a Comment

To make a comment you must login or register.

Comments on this content

  1. Building on their earlier provocative findings linking APP function to fast axonal transport, Stokin and colleagues, in this latest report, reinforce several important themes that are emerging from recent studies. First, significant neuronal pathobiology, especially evidence of altered vesicular trafficking, can be detected very early in Alzheimer disease (AD), before classical Alzheimer neuropathology appears. Second, these early disturbances at least partly stem from a behavior of APP or one of its processed forms; however, the issue of whether Aβ generation is an effect rather than the cause of this pathophysiology needs to be considered seriously. Finally, beyond its implications for Aβ generation, the defective vesicular transport observed in this study, and early endosomal-lysosomal dysfunction seen in other studies, are in their own right very likely to impair synapse function and axon/dendrite maintenance (Nixon, 2005). The new studies by the Goldstein group will hopefully encourage further exploration of these research themes, which are relatively understudied.

    The report provides evidence for an early failure of anterograde axonal transport in AD and implicates the transport motor, kinesin-1, as one route to this failure. This could nicely explain an initial report suggesting that KLC1 polymorphisms may influence risk for AD. A more generalized defect of vesicular transport in AD could also be envisioned. A dysfunctional microtubule "track," possibly involving tau, or an altered vesicular cargo, perhaps involving post-translationally modified APP, would be expected to impair not only anterograde axonal transport, but also retrograde traffic in dendrites, where dystrophy and accumulation of vesicular cargoes is more profoundly affected than in axons. The accumulating vesicles in dystrophic neurites in the Alzheimer brain include many of lysosomal origin, as initially pointed out by Robert Terry and colleagues. At the same time, many, if not most, correspond to autophagic vacuoles, which are early and late compartments of macroautophagy, a pathway for the turnover of organelles and long-lived proteins (Nixon et al. 2005). Interestingly, autophagic vacuoles are enriched in γ-secretase activity and contain Aβ in addition to the necessary components to generate Aβ (Yu et al., 2004). In the Stokin et al. study, a proportion of the vesicles accumulating in pathologic axons of the mouse model appear to have the distinctive double limiting-membrane morphology of early autophagic vacuoles, suggesting one possible source for the extra Aβ in these mice. Endosomes, another site of amyloidogenic APP processing, are known to be abundant anterograde vesicular cargoes in axons, so it will be interesting in future studies to sort out the relative contributions of these different vesicular compartments to the Aβ effect.

    See also:

    ARF related conference report


    . Endosome function and dysfunction in Alzheimer's disease and other neurodegenerative diseases. Neurobiol Aging. 2005 Mar;26(3):373-82. PubMed.

    . Extensive involvement of autophagy in Alzheimer disease: an immuno-electron microscopy study. J Neuropathol Exp Neurol. 2005 Feb;64(2):113-22. PubMed.

    . Autophagic vacuoles are enriched in amyloid precursor protein-secretase activities: implications for beta-amyloid peptide over-production and localization in Alzheimer's disease. Int J Biochem Cell Biol. 2004 Dec;36(12):2531-40. PubMed.

  2. The paper by Stokin et al is most remarkable and very convincing. Reducing axonal transport enhanced axonopathy, increased intracellular Aβ levels and extracellular deposition. Stimulation of APP cleavage may be the consequence of enhanced presence of APP-containing vesicles in axonal and/or somatodendritic compartments due to mistrafficking. Increased intraneuronal Aβ accumulation as a consequence has been earlier shown to trigger neuronal death in APP/PS1 mouse models. Impaired axonal transport may be the result of age-dependent processes leading to axonal deafferentiation and loss of synaptic contacts.

    In my opinion, this is a milestone paper, because it shows that intraneuronal deficits, like axonopathy, are observed prior to plaque induction. It provides further evidence for a central role of intraneuronal Aβ accumulation in the pathological processes of Alzheimer disease.

  3. This paper by Stokin et al. from the lab of Larry Goldstein has some interesting and important findings. I think the finding that APPsw transgenics having half the dose of kinesin-1 have increased Aβ deposition and pathology strongly argues that normal axonal transport is involved in the development of Aβ-related pathologies in AD. This is important, as it suggests that augmentation of this function or factors that prevent axonopathy may be protective against AD.

    The finding that there are neuritic swellings in very young APP transgenic mice is interesting, but whether this is relevant to AD is unclear. First, these swellings are smaller and different in appearance than the neuritic dystrophy around amyloid deposits. Second, and more importantly, the APP transgenic mice being studied overexpress mutant APP many-fold. Humans with AD of any type do not overexpress mutant APP (except in Down syndrome, in which there is APP overexpression but at a much lower level than in these mice). The overexpression of human APP increases human Aβ (required for Aβ aggregation in mice), but also may be resulting in other biological effects of mutant APP overexpression.

    It is possible that the neuritic changes described in the young APPsw mice are secondary to increased soluble Aβ. It is also possible that they are due to APPsw overexpression. Appropriate controls to sort this out might be overexpression of APPsw with the Aβ region changed in sequence or determining whether pharmacological or other inhibition of Aβ blocks the early neuritic changes. While the neuritic swellings seen in young APPsw mice are interesting and may have relevance to AD, I think this remains unclear at this point.

  4. Kinesin molecular motor protein is involved axonal transport along microtubules. Tau protein is a major constituent of mircrotubules and thus disruption of tau (hyperphophorylation as an example) or any other part of microtubules have been shown to interfere with anterograde transport and retrograde transport. In the case of AD the research seems to point more towards APP buildup as a result of neuronal structure degradation. A drastic reduction of kinesin is merely a symptom and not directly causal of APP and amyloid beta. Presenilin mutations that affect the enzyme's activity in cutting APP are shown in a wide variety of axonal dysfucntion in AD patients.

  5. Aluminum could be a co-factor in the findings of Stokin and collegues. Aluminum was found to inhibit neurofilament assembly, cytoskeletal incorporation, and axonal transport by Shea et al, 1997. Deloncle et al, 2001 found that aluminum L-glutamate causes massive mitochondrial swelling in the hippocampus of younger laboratory rats that mimics similar effects of the aging process in older animals. Stokin et al. found mitochondria in the axons. Aluminum is known to interfere with ATP and is linked with neurofibrillary degeneration. Bioaccumulation of aluminum in the human brain over the lifespan exposes the aging brain to potentially significant dosages.


    . Aluminum inhibits neurofilament assembly, cytoskeletal incorporation, and axonal transport. Dynamic nature of aluminum-induced perikaryal neurofilament accumulations as revealed by subunit turnover. Mol Chem Neuropathol. 1997 Sep-Dec;32(1-3):17-39. PubMed.

    . Ultrastructural study of rat hippocampus after chronic administration of aluminum L-glutamate: an acceleration of the aging process. Exp Gerontol. 2001 Feb;36(2) PubMed.

  6. This excellent study clearly demonstrates that axonal damage occurs long
    before amyloid deposition in both early stage AD and an APP mouse model.
    Furthermore, the authors demonstrate that reduced expression of the motor
    protein KCL-1 increases both the production and deposition of Aβ. However,
    it is unclear which comes first, the generation of soluble toxic Aβ
    species and then disruption of axonal transport, or disruption of
    transport leading to increased Aβ production and subsequent generation of
    toxic assemblies. A clear understanding of the pathogenic sequence is
    essential for the rational development of therapies and thus the temporal
    relationship between axonopathy and soluble Aβ species demands further
    investigation. Specifically, in light of the finding that anti-Aβ
    antibodies can lead to the clearance of early hyperphosphorylated forms of
    tau, it would be worthwhile determining if either passive or active
    immunization can rescue the pre-amyloid axonopathy.

Comments on Primary Papers for this Article

  1. More support for what might be called the axonal "traffic jam" hypothesis for the pathogenesis of AD - from Larry Goldstein's lab. This paper should be read in conjunction with Orly Lazarov et al., J Neurosci March 2, 2005, which integrates work from Sam Sisodia's lab and five other labs and which provides evidence against that hypothesis. It would be nice if experiments could sharply differentiate between axonal transport peripherally and centrally. One would expect fierce traffic jams in peripheral axons, but AD patients do not appear to be particularly susceptible to peripheral neuropathy. (Peripheral neuropathy is very common in older people and is a sadly neglected research topic.)

    View all comments by George M. Martin
  2. This interesting paper shows that our perception of AD physiopathology is getting more complex, but more realistic. We are far away from the simple explanation of the amyloid cascade hypothesis (ACH). To summarize, neurodegeneration is associated with a defect of the axonal transport: key players involved are the microtubules stabilized by tau proteins, the motor proteins that transport the cargo- vesicle along microtubules, and especially kinesin-I, and APP as well as PS1 in the cargo-vesicles.

    One big surprise is that the axonal transport defect generated by reducing the genetic dosage of kinesin increases Ab42 secretion and deposition. This sequence of events is the opposite of the ACH.

    To conclude, kinesin-I is likely to be an additional risk factor of AD. But behind the paper, even if bypassed, is the role of tau to stabilize and control axonal transport. Cause and effects have still to be untangled in AD.

    Among the references related to this approach, I recommend also the papers of Beyreuther on the fast axonal transport of APP and those of the Mandelkow's related to kinesin, tau, APP and the axonal transport (J Cell Biol. 2002 Mar 18;156(6):1051-63 and other related papers)

    View all comments by Andre Delacourte


News Citations

  1. Axonal Transport Suggested as Function for APP
  2. Suspects for Aβ Generation Spotted Together, En Route to Nerve Terminal
  3. Tau Accused of Blocking Transport, Causing APP to Linger and Nerve Processes to Wither
  4. Tau Kinase Clears Microtubules—Keeps Axonal Transport on Track
  5. Philadelphia: The Enemy Within—Neurodegeneration From Intraneuronal Aβ

Paper Citations

  1. . Ultrastructural studies in Alzheimer's presenile dementia. Am J Pathol. 1964 Feb;44:269-87.
  2. . Axonal transport defects: a common theme in neurodegenerative diseases. Acta Neuropathol. 2005 Jan;109(1):5-13. PubMed.
  3. . Axonopathy and transport deficits early in the pathogenesis of Alzheimer's disease. Science. 2005 Feb 25;307(5713):1282-8. PubMed.

External Citations

  1. Alzgene entry

Further Reading


  1. San Diego: Too Much APP Blocks Transport, Starves Down's Neurons
  2. New Orleans: Symposium Probes Why Synapses Are Suffering
  3. Huntington’s Protein Snarls Axonal Traffic
  4. Role of the Motor in Motor Neuron Diseases
  5. Can Travel, Will Deposit: Aβ via the Perforant Pathway?

Primary Papers

  1. . Axonopathy and transport deficits early in the pathogenesis of Alzheimer's disease. Science. 2005 Feb 25;307(5713):1282-8. PubMed.