Skip to Content

Another Argument For Tau

Just to give most everyone in the Alzheimer’s field a little more reason to doubt themselves, the Mayo Clinic has published a study suggesting that the pathology associated with tau protein (neurofibrillary tangles, NFT) is (as some have always maintained) more important than that associated with beta-amyloid.

Our observations reaffirm the concept that amyloid-β burden itself is not the proximate causal pathology for cognitive decline. . .Abnormal accumulation of amyloid-β and NFT pathology likely occurs in parallel with one another, eventually resulting in incompletely understood crosstalk between the pathologies. . .Although cortical accumulation of amyloid-β was considered sufficient to influence the risk of clinical Alzheimer’s disease, we and others demonstrate that NFT accumulation mediates the contribution of amyloid-β pathology to cognitive decline.

This won’t settle the argument, by any means, but it is more food for thought for the many organizations targeting amyloid itself. It’s always possible that (because of the crosstalk mentioned above) that amyloid-centric approaches could still be of benefit (just as it’s possible that amyloid still is the real driver of the disease state). But neither of these are sure things. Alzheimer’s has no sure things.

27 comments on “Another Argument For Tau”

  1. Lab Mouse says:

    “This comes the same week that a postmortem study of 3,600 brains claims to show that amyloid-ß is not causative for Alzheimer disease, making results like those above are either wrong or irrelevant. Reportedly, the culprit is tau protein. The amyloid hypothesis has driven AD research for many years, and cumulative evidence now suggests that this has all been nonsense. It may be that the tau protein hypothesis is also nonsense, and that both tau and amyloid-ß are results, rather than causes, of AD. There is a substantial chance that what we have to show for years of AD research is conflicting results, bogus hypotheses, and huge piles of dead animals.”

  2. Anonymous says:

    So Lane was right all along! Give him a Nobel prize now!

  3. Bioorganic Chemist says:

    Tau correlates with cognitive decline/neurodegeneration – Alzheimer’s Disease is of course *more than just the final (observable) stages*! This large study is consistent with a significant number of smaller animal and human studies that cognitive decline correlates with tau aggregate accumulation, and that Abeta aggregation occurs on an earlier time scale. HOWEVER, this hardly says that Abeta has no role in the disease. A lot of data indicate that Abeta aggregates can accelerate tau aggregation, and the best genetic data for familial (but not sporadic) AD are all indicative of Abeta processes. (Tau-only mutations generally lead to other neurodegenerative disorders.) And it doesn’t address ApoE4. But it does provide additional data suggesting that Abeta therapeutics might be best pre-symptomatically, partially explaining clinical trial failures.

  4. Bioorganic Chemist says:

    It is worth noting (and was done so at last year’s Alzheimer’s Association meeting): there is only 10 mg of Abeta total in an AD brain. So removal of all aggregated Abeta is not an inherently intractable problem (which of course doesn’t make it easy).

  5. Lane Simonian says:

    #1 This is it. Both amyloid oligomers and plaques and hyperphosporylated, nitrated, and tangled tau are all primarily a response to oxidative stress which is the key to the initiation and progression of Alzheimer’s disease. Amyloid oligomers increase hydrogen peroxide production and hyperphosphorylated, nitrated, and tangled tau inhibit neurotransmissions and limit the transport of nutrients in the brain, so both contribute something to Alzheimer’s disease. Amyloid plaques apparently contribute nothing.
    Recently I searched for Alzheimer’s cases without neurofibrillary tangles (tau tangles) and found this article:
    Neurology. 2004 Apr 13;62(7):1141-7.
    Alzheimer disease without neocortical neurofibrillary tangles: “a second look”.
    Tiraboschi P1, Sabbagh MN, Hansen LA, Salmon DP, Merdes A, Gamst A, Masliah E, Alford M, Thal LJ, Corey-Bloom J.
    CONCLUSIONS: Dementing syndromes virtually indistinguishable from each other can, and do, develop in the presence or absence of neocortical NFT. Patients without neocortical NFT are, on average, older at disease onset and death, and show a trend toward a shorter disease duration with somewhat slower deterioration. Although neocortical NFT per se are not obligatory for the development of clinical dementia, more subtle neocortical cytoskeletal tau pathology may contribute to cognitive decline in these subjects.
    All you need for Alzheimer’s disease is high levels of peroxynitrites which lead to the c terminal fragment of the amyloid precursor protein and hyperphosphorylated and nitrated tau. Peroxynitrites usually but do not necessarily lead to amyloid oligomers, plaques, and neurofibrillary tangles. None of the latter are prerequisites for the disease.
    http://www.ncbi.nlm.nih.gov/pubmed/16816118
    J Neurochem. 2001 Jul;78(1):109-20.
    C-terminal fragment of amyloid precursor protein induces astrocytosis.
    Bach JH1, Chae HS, Rah JC, Lee MW, Park CH, Choi SH, Choi JK, Lee SH, Kim YS, Kim KY, Lee WB, Suh YH, Kim SS.
    Shift the focus from removing amyloid plaques and oligomers and neurofibrillary tangles to preventing, scavenging, and reversing the damage done by peroxynitrites and you can effectively treat Alzheimer’s disease (as multiple small-scale clinical trials have indicated). It is not a matter of starting earlier or targeting oligomers rather than plaques; it is a matter of targeting what actually causes the disease.
    As far as a Nobel Prize, I have been wondering do you have to go to Stockholm to pick it up or can they FedEx it to your door? Of course, I would have to share it with a number of people, but I don’t mind that.

  6. Anonymous says:

    Read up a review on the concept of the O-GlcNAc modification:
    O-GlcNAc and neurodegeneration: biochemical mechanisms and potential roles in Alzheimer’s disease and beyond
    Chem Soc Rev. 2014 Oct 7;43(19):6839-58
    The Emerging Link Between O-GlcNAc and Alzheimer’s Disease
    J Biol Chem. 2014 Dec 12;289(50):34472-81
    Protein physiology was long thought to be controlled by phosphorylation, where phospho groups added to proteins can behave like on/off switches to regulate physiology. In fact, Krebs and Fischer won the Nobel Prize in 1992 for this concept. Years later after protein phosphorylation was discovered, however, it was found that a sugar called N-Acetylglucosamine (GlcNAc) gets added to proteins at SER/THR sites and often times these sites are the same exact sites where protein phosphorylation occurs. This concept is now known as the “O-glcnac modification” and the discovery of this sugar is quite profound. If you can visualize this concept in your head, O-glcnac acts as almost like a “cap” on top of SER/THR residues in proteins that must removed before a protein can be phosphorylated. In otherwords protein physiology is much, much more complex than Krebs and Fischer originally recognized since proteins do not act in a simple on/off switch-like manner due to phosphorylation. There must be some extraordinarily complex cycling mechanism between many multiple combinations of O-GlcNAc modifications and phosphorylated sites within proteins. This cycling between phosphorylation and modification by O-GlcNAc has been dubbed the “Yin and Yang Hypothesis”. What’s really, really cool about O-GlcNAc is the fact that only 2 enzymes regulate the entire patterns of O-GlcNAc on the proteome–OGT which adds O-GlcNAc and OGA which removes it–and that’s it. This is in very stark contrast to protein phosphorylation which requires hundreds if not thousands of phosphatases and kinases to regulate protein phosphorylation. We now know that it is likely probable that >90% of all proteins within the proteome are modified by O-GlcNAc.
    What does this have to do with AZ? Major disease such as cancer, diabetes, and AZ are characterized by abnormal glucose metabolism. Most scientists outside of biochemistry or glycobiology probably don’t know this either, but the substrate to perform the O-GlcNAc modification (GlcNAc) directly descends from glucose metabolism. In otherwords, there’s some very good scientific reasons as to why a disease like AZ could in fact be described as type 3 diabetes. Very important proteins involved in AZ are modified by O-GlcNAc, and this includes a protein like tau which is often hyperphosphorylated in AZ. In fact, O-glcnac has already been proven to regulate the phosphorylation status of tau:
    “O-GlcNAcylation regulates phosphorylation of tau: A mechanism involved in Alzheimer’s disease”
    Proc Natl Acad Sci U S A. 2004 Jul 20; 101(29): 10804–10809
    There are some quite interesting and novel strategies coming out now to treat AZ by attacking abberrant glucose metabolism. This may go a long way to explain why patients who take drugs like Metformin are less likely to develop dimentia. If tau is often hyperphosphorylated in AZ, and tau hyperphosphorylation exists in a ‘ying and yang’ like nature with O-glcnac, what if you could inhibit the enzyme OGA to treat tauopathies? In fact, this has been tried, and thus far the results have been quite intriguing in animal models (some recent presentations at conferences too on OGA inhibitors to treat AZ have quite amazing data):
    “Pharmacological inhibition of O-GlcNAcase (OGA) prevents cognitive decline and amyloid plaque formation in bigenic tau/APP mutant mice”
    Mol Neurodegener. 2014; 9: 42.
    “Differential Effects of an O-GlcNAcase Inhibitor on Tau Phosphorylation”
    PLoS One. 2012;7(4):e35277
    There are a lot of people at recent conferences on AZ and glycobiology that have are very interested in coming up with new OGA inhibitors or using existing OGA inhibitors as a new class of treatments for AZ. In any case, when you start messing up carbohydrate/glucose metabolism, you mess up a loooooot of physiology because of the fact that the O-glcnac modification and glucose metabolism go hand in hand. In fact, not only does O-GlcNAc regulate the phosphorylation status of tau proteins, it also is known to attenuate the build up of amyloid plauqes, so maybe Aβ and tau are actually connected (rather than one being more important than the other), and that connection may very well be O-GlcNAc. AZ may very well be more a metabolic disease than anything. Decreased glucose metabolism, which is almost always observed in an AZ brain, may be one of the early keys that drives the development of AZ in an aging brain.

  7. Bioorganic Chemist says:

    I love the O-GlcNAc story for the protective effects on preventing phosphorylation (and potential structural effects opposing those of phosphorylation) – the animal data indeed look encouraging, and Gong has presented some nice unpublished data at conferences too. The major question is stoichiometry: O-GlcNAc levels seem quite low even in normal (human/mouse model) brains (often

  8. Harrison says:

    My reading of this data is that you essentially have two types of Alzheimer’s disease: ApoE4 positive disease, which results in a higher amyloid burden, and ApoE4 negative disease, which results in a high tau burden. It’s quite possible Tau is a more important driver for both, but it’s time to stop pretending that there is a unitary construct that is “Alzheimer’s Disease.” If you throw in insulin-resistance / diabetic induced AD, you probably have 3 types of Alzheimer’s disease, as diabetics tend to show more white matter infarcts and fewer plaques.

  9. Anonymous says:

    ApoE4 and insulin signaling/glucose metabolism appear to go hand in hand from some recent reports:
    “Reduced phosphorylation of brain insulin receptor substrate and Akt proteins in apolipoprotein-E4 targeted replacement mice”
    Scientific Reports 4, Article number: 3754
    E4 may manifest itself as perturbed sugar metabolism, hence patterns of O-GlcNacylation will change in response to altered insulin signaling; something like hyperphosphorylation of Tau may or poor clearance of AB may result (it has been shown before that OGA inhibitors also attenuate AB formation).

  10. Robert says:

    Lane–
    Interesting thoughts about peroxynitrite. There is a company (Genkyotex) that is developing inhibitors of NADPH oxidase. These are actual inhibitors of the enzyme, not free radical scavengers, and would suppress formation of peroxynitrite (via suppression of superoxide formation) as well as other derived ROS. I have long been dubious about free radical scavengers, but inhibition of NADPH oxidase has intrigued me for some time (might be some risk to it, however, as ROS have a role other than killing things). Didn’t have the opportunity to work on it, however (retired now). Their initial target seems to be diabetic nephropathy, but they mention neurodegenerative diseases as of interest in this context.
    I have no connection with this company, but have been attempting to follow the story for some years. They seem to have a compound in Phase II now, with others coming behind.

  11. Lane Simonian says:

    Thank you, Robert. I will try to read up on the work of Genkyotex.
    I think the better target might be inducible nitric oxide synthase (although tricky as well) rather than NADPH oxidase as more superoxide appears to be produced in Alzheimer’s disease than inducible nitric oxide. Enough such that as when amyloid oligomers attract copper and zinc, some of the superoxides are converted into hydrogen peroxide well those remaining continue to combine with inducible nitric oxide to produce peroxynitrites. Having said that here is an argument for the inhibition of NADPH oxidase for the treatment of Alzheimer’s disease:
    http://www.biomedcentral.com/1471-2202/9/S2/S8
    The following researchers identify the dual activation of NADPH oxidase and Inducible Nitric Oxide synthase in cell death in neurodegeneration.
    Activation of microglial NADPH oxidase is synergistic with glial iNOS expression in inducing neuronal death: a dual-key mechanism of inflammatory neurodegeneration
    Palwinder Mander and Guy C Brown*
    I always appreciate it when scientists provide valuable information and leads on this site and elsewhere.

  12. bank says:

    (@Bioorganic Chemist — do you have a citation for there only being 10 mg of Abeta in the brain?)
    I think it it worth remembering that the amyloid hypothesis does not depend only on amyloid per se, but also on the fact that mutations in the amyloid precursor protein cause early onset Alzheimer’s disease, as do mutations in presenilin, which is required for production of Abeta. The early onset disease is not readily distinguishable pathologically from the late onset disease, and both show excess deposition of Abeta.
    That tau is involved is shown only by its appearance as tangles of hyperphosphorylated tau; mutations in the tau gene cause neurodegenerative diseases that do not resemble Alzheimer’s, either pathologically or clinically.

  13. anon says:

    AD breakthrough CLR01 (molecular tweezers).

  14. Lane Simonian says:

    The problem with the mutation argument in favor of the amyloid hypothesis is that it assumes that amyloid is the first result of the mutation where it is the last result.
    The first thing that most if not all of the mutations do is to activate (or prevent the deactivation) of g protein-coupled receptors. Some presenilin-1 gene mutations may suppress the activation of the neuroprotective phosphatiylinositol 3-kinase.
    Wild-Type But Not FAD Mutant Presenilin-1 Prevents
    Neuronal Degeneration by Promoting Phosphatidylinositol
    3-Kinase Neuroprotective Signaling
    Both aberrant g protein-coupled receptor activity and suppression of the phosphatidylinositol 3-kinase lead to the production of peroxynitrites. The subsequent formation of amyloid requires intracellular calcium release which does not always happen. Some researchers have labelled Alzheimer’s disease without
    amyloid Part Age Related Taupathy perhaps in part to try to rescue the notion that Alzheimer’s disease is caused by amyloid. But the critical point is this you can have the death of neurons and Alzheimer’s disease with widespread peroxynitrite-mediated damage itself, it does not require amyloid.
    Widespread Peroxynitrite-Mediated Damage in Alzheimer’s Disease
    Mark A. Smith1, Peggy L. Richey Harris1, Lawrence M. Sayre2, Joseph S. Beckman3, and George Perry1
    Alzheimer’s Disease Without Amyloid Plaques
    Amyloid plaques have long been thought to be the cause of neuron loss in Alzheimer’s disease. Now researchers report that excess of mutated amyloid precursor protein (APP) inside the neurons is sufficient to induce neuron death. The report challenges the notion that amyloid deposits outside of the cells are necessary for neuron death in Alzheimer’s disease.

  15. Lane Simonian says:

    Even better:
    Familial amyloid precursor protein mutants cause caspase-6-dependent but amyloid β-peptide-independent neuronal degeneration in primary human neuron cultures.
    Because Casp6 is activated early in AD and is involved in axonal degeneration, these results suggest that the inhibition of Casp6 may represent an efficient early intervention against familial forms of AD. Furthermore, these results indicate that removing Aβ without inhibiting Casp6 may have little effect in preventing the progressive dementia associated with sporadic or familial AD.
    The appropriate caspase to target early in Alzheimer’s disease (either early or late onset) is probably caspase3 not caspase6, but regardless this is the clearest indication yet that removing amyloid (either oligomers or plaques) does very little to slow the progression of Alzheimer’s disease. The key to treating the disease is the use of specific antioxidants.

  16. bank says:

    @Lane
    The citation you provide by Smith et al, 1997, was headed by George Perry. I’m sure you will find that Perry has not persisted with the peroxynitirite hypothesis today (Mark Smith has unfortunately passed away).
    That mutations in APP cause both Alzheimer’s and (excess) amyloid plaques is unambiguous. Whether the plaques or Abeta cause neurodegeneration is indeed a vexing question.
    The issue will not be truly settled unless someone finds a new reason for those mutations causing Alzheimer’s disease that doesn’t rely on Abeta.

  17. Bioorganic Chemist says:

    @bank This was from a talk by Colin L. Masters last year at AAIC. (The age-matched control had 2.7 mg Abeta average – so, yes, there is more TOTAL Abeta in AD brains by a factor of 3-4.) I don’t have a reference on hand (these are from my notes), but it should be in his papers (sorry). They also used Abeta imaging over time (AIBL study) to predict likelihood of onset of future AD, including as a function of ApoE genotype.

  18. bank says:

    @Bioorganic Chemist
    Thanks for the information I will check it out. I wonder how it compares to the argument presented by Karran and de Strooper in their review (its a few years old: “The amyloid cascade hypothesis for Alzheimer’s disease: an appraisal for the development of therapeutics”; Pubmed id 21852788) where they state “the amount of deposited amyloid-[beta] in the AD brain is equivalent to approximately 7 years of total amyloid-[beta] production”
    The amount of Abeta deposited in the brain is an important question, since it is now well established that Abeta in the CSF *declines* in late onset Alzheimer’s disease, both for Ab40 *and* Ab42, although Ab40 declines more than Ab42, giving the now well-established signature of an increased Ab42/40 ratio in the CSF; this is OK for the amyloid hypothesis as long as one can account for that by stating that this is due to increased amyloid deposition in the brain.
    However… it is also known (although longitudinal studies done to date are somewhat underpowered) that amyloid plaques (as measured by PiB/PET) do not increase as the disease progresses, instead they plateau.
    Vexing questions….

  19. Anonymous says:

    @blank
    the imaging of beta amyloid plaques with PiB or other tracers like this is not quantitative in terms of total beta amyloid protein deposited in plaque, this target is a degraded, agglomerating mess that crashes out at some point in time, from the typical radiotracer point of view it has a ‘not too hydrophilic surface’ to lean on – literally.
    there is no established correlation between total beta amyloid mass and tracer uptake.

  20. Lane Simonian says:

    Bank, I talked with George Perry last year and he gave no indication to me that he has given up on the peroxynitrite hypothesis for Alzheimer’s disease. Rather, he reminded me with pride that he was one of the founders of the hypothesis. Here is a relatively recent interview with George Perry (2011).
    http://archive.sciencewatch.com/ana/st/alz2/11monSTAlz2Perr/
    Unfortunately, I never had the opportunity to meet or talk to Mark A. Smith.
    The first part of your statement is correct: “That mutations in APP cause both Alzheimer’s and (excess) amyloid plaques is unambiguous.” This however is a correlation; it does not prove causation. Indeed it is possible to have the death of neurons with the amyloid precursor protein mutations alone. Neither plaques nor oligomers cause Alzheimer’s disease, although the latter are not blameless as they increase hydrogen peroxide production.

  21. bank says:

    Lane,
    That Abeta causes Alzheimer’s is indeed correlation, and not causation. However, that (some) APP mutations cause Alzheimer’s is causation, but not correlation.

  22. SAR Screener says:

    AD is well outside my field of expertise but whenever I’ve seen discussions around Tau vs Abeta I do wonder if the situation with Alzheimer’s is not that dissimilar to cancer where it’s not actually one disease caused by one single underlying mechanism.
    This may also explain the poor outcome of clinical trials in AD as the situation could be similar to something like Herceptin treatment outcomes for breast cancer in general vs targeting patients that are HER2 positive.

  23. Lane Simonian says:

    Bank, exactly!

  24. Harrison says:

    @ SAR Screener. This is precisely the point I was trying to make up back in post #8. This most recent study suggests at least two types of Late Onset Alzheimer’s disease, an ApoE4 positive disease with higher amyloid burden and an ApoE4 negative disease with higher tau burden. Attempts at better diagnosis, such as those through the ADNI consortium suggest that 25% of patients have “pure” Alzheimer’s disease, 50% have mixed dementia (indications of some AD pathology plus something else, like fronto-temporal or Vascular Dementia), and a full 25% are just mis-diagnosed. And this doesn’t include the 5% of patients with early-onset AD caused by APP/PS1/PS2 mutations.
    @Lane: You might want to check out an new paper in Neuron 85, 967-981 that suggests the PS1 mutations may be a loss of function, at least in mice. This also suggests the inhibiting gamma-secretase function (perhaps independent of notch signaling) might be a bad thing, which is exactly what Lilly found in their semagacestat trial.
    @Bank: To me, data like this shows the danger of generalizing from the familial mutations to the sporadic, late-onset disease. In some ways, AD plaque deposition is like a skin rash. There are dozens of different ways to get a similar looking skin rash that can vary widely in etiology.

  25. SAR Screener says:

    @24 sorry I did see your post and dithered about addressing it directly, but your original post seemed focused on their being 2 types of AD and I was more wondering if there could be dozens (or more) of diseases that cause similar symptoms.

  26. Bioorganic Chemist says:

    @SAR Screener tau *versus* Abeta was the old battle that died a decade ago when it was recognized that they clearly interact and that the disease is clearly multifactorial. It is certainly why trials and understanding the underlying biology is so difficult: there is not a single smoking-gun cause, but it requires multiple components acting in tandem to different extents that modulate age of onset and severity. Which is why the field is looking at the equivalent of combination therapies as what may be necessary to truly change disease course. In that way, I think it is less cancer-like (that is, multiple components are central for all forms of the disease; and we are saved from the more complex genotypes of cancer, and the absence of resistance mutations!), though the “hallmarks of cancer” multifactorial aspect is a useful paradigm.

  27. Bioorganic Chemist says:

    @SAR Screener tau *versus* Abeta was the old battle that died a decade ago when it was recognized that they clearly interact and that the disease is clearly multifactorial. It is certainly why trials and understanding the underlying biology is so difficult: there is not a single smoking-gun cause, but it requires multiple components acting in tandem to different extents that modulate age of onset and severity. Which is why the field is looking at the equivalent of combination therapies as what may be necessary to truly change disease course. In that way, I think it is less cancer-like (that is, multiple components are central for all forms of the disease; and we are saved from the more complex genotypes of cancer, and the absence of resistance mutations!), though the “hallmarks of cancer” multifactorial aspect is a useful paradigm.

Comments are closed.