. Mechanistic basis of phenothiazine-driven inhibition of tau aggregation. Angew Chem Int Ed Engl. 2013 Mar 18;52(12):3511-5. PubMed.


Please login to recommend the paper.


Make a Comment

To make a comment you must login or register.

Comments on this content

  1. When we first isolated a fragment of tau protein from the repeat domain for the proteolytically stable core (1), this represented the first rigorous demonstration that tau protein contributes to the structural core of the paired helical filaments (PHFs) which comprise the neurofibrillary tangle originally discovered by Alzheimer. Prior claims linking tau protein with tangles based on immunohistochemical staining of tangles included neurofilament protein, vimentin, and MAP2 as well as tau (2), and did not help answer the questions, How is the PHF put together? How much of the PHF is tau? How might one design a therapy targeting this pathology?

    A widely quoted paper by Lee et al. claimed the complete dissolution of a subclass of PHFs isolated on the basis of insolubility in sarkosyl, and that the PHF is composed entirely of full-length hyperphosphorylated tau (3). It turned out this was based on a circular argument in which only antibodies recognizing hyperphosphorylated tau were used. When an assay was developed which permitted measurement of total PHF-tau as well as phosphorylated PHF-tau, it was possible to show that, even in the sarkosyl preparation, the quantity of hyperphosphorylated tau accounts for less than 10 percent of the tau content of the PHFs (4). This was also supported by immune-electron microscopy and scanning mass-spectroscopy evidence (5,6) showing that the N-terminal domain which is characteristically hyperphosphorylated contributes only to the proteolytically susceptible fuzzy outer coat of the PHF, and not to the structural core. In all, the N-terminal domain of tau contributes less than 5 percent of the mass of the PHF (4), the rest made up only of the truncated repeat domain fragment, which contributes at least 90 percent of the total protein composition of the structural core of the PHF (7). This has direct implications for a range of therapeutic strategies based on targeting tau phosphorylation, two of which have recently failed in Phase 2 (8,9).

    We later showed that this “fragment” actually comprises a characteristic mixture of 3- and 4-repeat isoforms, made up of repeats 1-3 of the 3-repeat, and either 1-3 or 2-4 of the 4-repeat isoform, but always restricted to three repeats. The repeats as they appear in the PHF core are phase shifted by half a repeat with respect to the organization of the tubulin-binding domains, such that they extend C-terminally to Glu-391, which represents a characteristic C-terminal truncation marker of the tau-tau binding footprint (10,11). This arrangement has significance for understanding how many of the cysteine residues implicated by Akoury et al. (12) could conceivably contribute to the core of the PHF. In fact, only the fragment comprising residues 1-3 of the 4-repeat isoform would contain both C291 and C322. The others would contain only one. This implies that the IC50 potentially relevant to the anti-aggregation effect of methylthioninium (MT) would be closer to 30 μM than to the 2 μM figure.

    The repeat domain-fragment isolated from the PHF core has the ability to propagate aggregation through cycles of binding and proteolysis (1), making the process prion-like (7). The prion-like spread of tau aggregation is now seen as the underlying mechanism responsible for characteristic Braak staging, which maps the spread of neurofibrillary degeneration through a highly stereotyped sequence of brain regions (13,14).

    In this context, the question of exactly how tau aggregates form becomes very important for understanding the process as a potential therapeutic target (7). Since our first demonstration that diaminophenothiazines are able to block the tau-tau binding interaction which reproduces in vitro the conditions required for the characteristic C-terminal truncation at Glu-391, there have been several studies confirming that MT is indeed a tau-aggregation inhibitor (1,15,16). It is generally the chloride salt of the oxidized form (MT+, i.e., MTC) that is used, and this is commonly known as methylene blue. There are limitations in the pharmaceutics of methylene blue which make it unsuitable as a candidate for further pharmaceutical development, including poor tolerability and limitation in absorption in the presence of food (17). It is for this reason that we have chosen to develop a stabilized reduced form of MT, namely, LMTX, as the product we are taking into Phase 3. LMTX has better activity in reducing tau aggregation in two tau transgenic mouse models, is better tolerated, and its absorption is not limited in the presence of food.

    A redox mechanism of action of MT is potentially attractive, as this appears to be the basis for several of its biological effects, including reducing methemoglobin to normal hemoglobin for the intravenous treatment of methemoglobinemia; redox effects on electron shuttling enhancing β-oxidation by mitochondria (18); and inhibition of guanylate cyclase and nitric oxide synthase (19).

    However, it is not clear that this is the mechanism underlying the effect on tau aggregation, notwithstanding the evidence presented by Akoury et al. (12). One particularly important difficulty with the proposal that MT+ oxidizes cysteine to produce a species which cannot form a disulphide bridge hinges on the fact that MT actually disaggregates PHFs isolated from AD brain (1). The mechanism proposed by Akoury et al. (12) could only potentially explain a preventative effect on tau aggregation, but not on disaggregation of already assembled native PHFs. Indeed, the potency of the disaggregation effect (IC50 0.15 μM), which is more than two orders of magnitude greater than the effects shown by Akoury et al. (12) on cysteine residues. Furthermore, the brain concentration at which MT is able to reduce tau aggregation in vivo is in the range of 0.2-0.7 μM, which is also the predicted brain concentration at steady state in humans (unpublished data). It is difficult to relate the benefits seen in vivo, and the evidence of therapeutic benefit we reported in our Phase 2 trial (20), to the concentrations required for the effects demonstrated by Akoury et al. (12).

    A further difficulty is that the NMR data provided by Akoury et al. (12) relates solely to an aqueous phase binding interaction between MT and monomeric soluble tau. However, it is increasingly understood that the critical binding interaction occurs in the solid phase (7,21,22) and, indeed, Akoury et al. (12) make use of heparin sulfate to facilitate aggregation. As shown by Lansbury and others, there is an energy barrier to spontaneous tau aggregation in the aqueous phase (23). It is entirely plausible, therefore, that the critical binding interaction between MT and tau depends on a binding site available only in the aggregated state, and perhaps entirely unavailable in monomeric soluble tau. It is not easy to see how this problem could be addressed experimentally, since tau aggregates disaggregate in the presence of MT and are converted to a monomeric form which is highly susceptible to proteases (1).

    The basic problem is that tau is likely to be able to aggregate in vitro in a number of different ways, including the disulphide crosslinking mechanism, but not all of these may have biological or therapeutic relevance. The significance of tau protein aggregation as an important substrate of clinical dementia has come to be better understood in the AD field, particularly in the light of the failure now of 18 Phase 2 and Phase 3 clinical trials targeting different aspects of the amyloid pathway. Consequently, it becomes important to gain a better understanding of the mechanism of action of the MT family as the first chemical class of tau aggregation inhibitors to be tested in Phase 2 and Phase 3 clinical trials. As it stands, the underlying mechanism of action still remains an open question, notwithstanding the valuable contribution from Akoury et al. (12).

    See also:

    Wischik, CM, Lai, RYK, and Harrington, CR. (1997) Modelling prion-like processing of tau protein in Alzheimer’s disease for pharmaceutical development. In "Microtubule-Associated Proteins: Modifications in Disease." (Avila, J., Brandt, R., and Kosik, K. S., Eds.), pp 185-241, Harwood Academic Publishers, Amsterdam.

    Tau-Targeting Drug Davunetide Washes Out in Phase 3 Trials. Alzheimer Research Forum

    Noscira announces results from ARGO Phase IIb trial of tideglusib for the treatment of Alzheimer's disease. 2012. See website.

    Will Tau Drug Show Its True Colors in Phase 3 Trials? Alzheimer Research Forum


    . Selective inhibition of Alzheimer disease-like tau aggregation by phenothiazines. Proc Natl Acad Sci U S A. 1996 Oct 1;93(20):11213-8. PubMed.

    . Cell biology of the Alzheimer tangle. Curr Opin Cell Biol. 1989 Feb;1(1):115-22. PubMed.

    . A68: a major subunit of paired helical filaments and derivatized forms of normal Tau. Science. 1991 Feb 8;251(4994):675-8. PubMed.

    . Quantitative analysis of tau protein in paired helical filament preparations: implications for the role of tau protein phosphorylation in PHF assembly in Alzheimer's disease. Neurobiol Aging. 1995 May-Jun;16(3):409-17; discussion 418-31. PubMed.

    . Isolation of a fragment of tau derived from the core of the paired helical filament of Alzheimer disease. Proc Natl Acad Sci U S A. 1988 Jun;85(12):4506-10. PubMed.

    . Subunit structure of paired helical filaments in Alzheimer's disease. J Cell Biol. 1985 Jun;100(6):1905-12. PubMed.

    . Difference between the tau protein of Alzheimer paired helical filament core and normal tau revealed by epitope analysis of monoclonal antibodies 423 and 7.51. Proc Natl Acad Sci U S A. 1991 Jul 1;88(13):5837-41. PubMed.

    . Measurement of distinct immunochemical presentations of tau protein in Alzheimer disease. Proc Natl Acad Sci U S A. 1991 Jul 1;88(13):5842-6. PubMed.

    . Mechanistic basis of phenothiazine-driven inhibition of tau aggregation. Angew Chem Int Ed Engl. 2013 Mar 18;52(12):3511-5. PubMed.

    . Alzheimer's pathogenesis: is there neuron-to-neuron propagation?. Acta Neuropathol. 2011 May;121(5):589-95. PubMed.

    . Transmission and spreading of tauopathy in transgenic mouse brain. Nat Cell Biol. 2009 Jul;11(7):909-13. PubMed.

    . Inhibition of heparin-induced tau filament formation by phenothiazines, polyphenols, and porphyrins. J Biol Chem. 2005 Mar 4;280(9):7614-23. PubMed.

    . Inhibition of tau aggregation in a novel Caenorhabditis elegans model of tauopathy mitigates proteotoxicity. Hum Mol Genet. 2012 Aug 15;21(16):3587-603. PubMed.

    . Stimulation of respiration by methylene blue in rat liver mitochondria. FEBS Lett. 1997 Jul 21;412(1):157-60. PubMed.

    . 7-Nitroindazole and methylene blue, inhibitors of neuronal nitric oxide synthase and NO-stimulated guanylate cyclase, block MK-801-elicited behaviors in mice. Neuropsychopharmacology. 1996 Jul;15(1):37-43. PubMed.

    . Tau aggregation inhibitor (TAI) therapy with RemberTM arrests disease progression in mild and moderate Alzheimer's disease over 50 weeks. Alzheimer's and Dementia 4, T167. (2008)

    . Assembly of microtubule-associated protein tau into Alzheimer-like filaments induced by sulphated glycosaminoglycans. Nature. 1996 Oct 10;383(6600):550-3. PubMed.

    . RNA stimulates aggregation of microtubule-associated protein tau into Alzheimer-like paired helical filaments. FEBS Lett. 1996 Dec 16;399(3):344-9. PubMed.

    . Structural neurology: are seeds at the root of neuronal degeneration?. Neuron. 1997 Dec;19(6):1151-4. PubMed.

  2. The commentary by Claude Wischik contains a review of some key discoveries in the study of the structure of paired helical fragments (PHFs) of tau, many by him and his collaborators, including the inhibitory effects of methylene blue on tau aggregation. With regard to our paper, his main argument is that the oxidative effects of methylene blue on tau cannot explain its effects in cells, contrary to what we suggest. In our view, his critique does not really address the point we are trying to make in our paper. We also question some of his claims about PHFs.

    In 2008, Wischik and colleagues claimed that methylene blue was beneficial as a treatment of AD (Wischik et al., 2008). This approach was based on the assumption that methylene blue inhibits the aggregation of tau (Wischik et al., 2006). Indeed, several authors (e.g., Taniguchi et al., 2005; Schirmer et al., 2011) have confirmed that methylene blue can act directly as an inhibitor of tau aggregation in vitro. Tau aggregation is considered to be a key step in AD pathology and, therefore, it is appropriate to ask how methylene blue could inhibit the process at the molecular level. This is the issue we tried to address in our paper. The answer, in short, is that methylene blue, which is a redox-shuttling reagent, interacts with the repeat domain of tau and modifies cysteines C291 and/or C322, depending on the tau isoform. With normal tau, cysteine disulfide bond formation can lead to tau dimers that nucleate and greatly accelerate aggregation (Schweers et al., 1995). Thus, the blockage of such bonds effectively amounts to inhibition of aggregation. In cells, this would allow extra time for the cellular degradation pathways to clear unwanted aggregates, e.g., oligomers.

    Wischik states that the interaction of methylene blue with the cysteines of tau cannot be the modus operandi in cells because methylene blue can dissolve "PHFs prepared from AD brain." We question this claim because in the cited reference (Wischik et al., 1996), the authors use PHFs heavily processed by proteolytic cleavage. In our view, the claim still needs substantiation. In principle, an aggregation inhibitor can not only inhibit aggregation, but also disassemble preformed aggregates, as long as the aggregates are in dynamic equilibrium with their free subunits. The reason is that subunits released from polymers can be poisoned by the aggregation inhibitors, and fail to reassemble. The net effect is disassembly. The dynamic nature of tau filaments has been demonstrated in a number of experimental settings, for example, by the disappearance of certain tau species from aggregates, once their supply has been turned off (e.g., after switching off transgene expression in mice; see Sydow et al., 2011). Indeed, the various aggregation inhibitors we have tested in the past often show similar half-maximal values for inhibition and disassembly (IC50 and DC50 values; e.g., Pickhardt et al., 2007; Bulic et al., 2010). The fact that methylene blue rescues the tau-induced paralysis of C. elegans models shows that this approach can work, at least in simple systems (Fatouros et al., 2012).

    A caveat that we probably agree on is that methylene blue's inhibition of tau aggregation does not necessarily mean that this is the only way it has beneficial effects in cells and organisms. Being a charged and planar molecule, methylene blue has the potential to interact with hundreds of enzymes, especially nucleotide-binding proteins. Which of these might be the crucial effect for treating AD is unknown at present, but other therapeutic applications point to considerable diversity (see examples quoted by Wischik, and review by Schirmer et al., 2011). Time will tell.

    We would also draw attention to some statements that we consider to be misleading. The statement that "the N-terminal domain of tau contributes less than 5 percent of the mass of the PHF" is incorrect. As shown by various authors, PHFs contain all six full-length isoforms (e.g., Delacourte et al., 1990; Goedert et al., 1992). The confusion arises because Wischik's preparation of "AD-derived PHFs" is, in fact, an extensively processed preparation involving digestion by pronase. This preparation has been useful in identifying the repeat domain as the core of PHFs, but one cannot claim that PHFs lack the N-terminal domain when one cleaves it off in the first place. Actually, scanning transmission electron microscopy experiments are consistent with mostly full-length tau in PHFs (von Bergen et al., 2006).

    The antibody MN423, recognizing tau truncated at E391 (Novak et al., 1991), has played an important role in identifying tau in PHFs, but this truncation is not a necessary requirement for PHF assembly.

    Although Wischik uses cycles of binding and proteolysis as a method to obtain PHFs, this is not a necessary requirement, since even full-length tau can assemble into PHFs once they are nucleated, both in vitro and in vivo (Goedert et al., 1996; Sydow et al., 2011). Apart from that, PHF assembly can be simply described by the well-known laws of self-assembling systems (e.g., Flory, Principles of Polymer Chemistry, 1953; Oosawa and Kasai, 1962). There is no need to invoke the label "prion-like," especially since tau does not form infectious particles.

    Finally, the argument about aggregation in "solid phase" is inappropriate. PHF assembly from tau subunits works well in solution, including critical binding interactions, and the use of cofactors such as heparin does not make the system solid phase.

    In conclusion: If one assumes that methylene blue's therapeutic effect is related to tau, then the interaction with tau's cysteines and inhibition of aggregation is a likely scenario. Whether or not this is the main pathway of methylene blue's action in neurons remains to be seen.


    . Tau protein and tau aggregation inhibitors. Neuropharmacology. 2010 Sep-Oct;59(4-5):276-89. PubMed.

    . Pathological proteins Tau 64 and 69 are specifically expressed in the somatodendritic domain of the degenerating cortical neurons during Alzheimer's disease. Demonstration with a panel of antibodies against Tau proteins. Acta Neuropathol. 1990;80(2):111-7. PubMed.

    . Inhibition of tau aggregation in a novel Caenorhabditis elegans model of tauopathy mitigates proteotoxicity. Hum Mol Genet. 2012 Aug 15;21(16):3587-603. PubMed.

    . Tau proteins of Alzheimer paired helical filaments: abnormal phosphorylation of all six brain isoforms. Neuron. 1992 Jan;8(1):159-68. PubMed.

    . Assembly of microtubule-associated protein tau into Alzheimer-like filaments induced by sulphated glycosaminoglycans. Nature. 1996 Oct 10;383(6600):550-3. PubMed.

    . Difference between the tau protein of Alzheimer paired helical filament core and normal tau revealed by epitope analysis of monoclonal antibodies 423 and 7.51. Proc Natl Acad Sci U S A. 1991 Jul 1;88(13):5837-41. PubMed.

    . A theory of linear and helical aggregations of macromolecules. J Mol Biol. 1962 Jan;4:10-21. PubMed.

    . Phenylthiazolyl-hydrazide and its derivatives are potent inhibitors of tau aggregation and toxicity in vitro and in cells. Biochemistry. 2007 Sep 4;46(35):10016-23. PubMed.

    . "Lest we forget you--methylene blue...". Neurobiol Aging. 2011 Dec;32(12):2325.e7-16. PubMed.

    . Oxidation of cysteine-322 in the repeat domain of microtubule-associated protein tau controls the in vitro assembly of paired helical filaments. Proc Natl Acad Sci U S A. 1995 Aug 29;92(18):8463-7. PubMed.

    . Tau-induced defects in synaptic plasticity, learning, and memory are reversible in transgenic mice after switching off the toxic Tau mutant. J Neurosci. 2011 Feb 16;31(7):2511-25. PubMed.

    . Inhibition of heparin-induced tau filament formation by phenothiazines, polyphenols, and porphyrins. J Biol Chem. 2005 Mar 4;280(9):7614-23. PubMed.

    . The core of tau-paired helical filaments studied by scanning transmission electron microscopy and limited proteolysis. Biochemistry. 2006 May 23;45(20):6446-57. PubMed.

    . Selective inhibition of Alzheimer disease-like tau aggregation by phenothiazines. Proc Natl Acad Sci U S A. 1996 Oct 1;93(20):11213-8. PubMed.

  3. Our study investigated the mechanism of inhibition of tau aggregation by methylene blue and its metabolites, azure A and B, when starting from the tau monomer—that is, the conformation that is able to bind to microtubules. NMR spectroscopy together with mass spectrometry unambiguously proved that the cysteine residues of tau are modified to sulfenic, sulfinic, and sulfonic acids. Importantly, the methylene blue-induced oxidation of cysteines is key to the inhibition of tau aggregation, as methylene blue does not influence aggregation of a cysteine-free tau variant.

    Claude Wischik now argues that this cannot be the mechanism of aggregation inhibition, as it could not explain methylene blue-induced dissociation of tau aggregates. Apparently, there might be two different things going on—inhibition of conversion from monomer to tau aggregates might work by a different mechanism than dissociation of preformed aggregates. Indeed, using NMR spectroscopy, we showed that upon oxidation of the cysteine residues of tau, conformational changes also occur in the second hexapeptide. This hexapeptide is located in the core of tau amyloid fibrils (Daebel et al., 2012). Thus, this secondary effect (secondary in terms of inhibition of conversion from monomer to aggregate) could be important for dissociation of aggregated tau. In addition, new binding sites might be formed in aggregated tau, further contributing to dissociation of tau aggregates.

    Thus, our study does not address the mechanism of dissociation of tau aggregates by methylene blue, but it unambiguously proves that the methylene blue-driven modification of the tau cysteine residues to sulfenic, sulfinic, and sulfonic acids is essential for inhibition of aggregation from the tau monomer to amyloid fibrils and avoids the formation of potentially toxic tau intermediates.


    . β-Sheet core of tau paired helical filaments revealed by solid-state NMR. J Am Chem Soc. 2012 Aug 29;134(34):13982-9. PubMed.

  4. The assumption behind all the work with methylene blue is that inhibition of tau aggregation is going to be good for people. Some of the same arguments have been made about Aβ, but that field seems to have drifted toward the idea that it is not the aggregates that are the problem, at least not the large aggregates. In the tau arena, the work of Mel Feany's group with flies seems to say that tau toxicity can occur in the absence of visible aggregates, and Karen Ashe's group working with the Tg4510 mouse seems to be heading in the same direction. Reading the literature leaves me with the notion that tau aggregation could be good, bad, or irrelevant, and it is hard to be convinced which of these is correct. The Mandelkows, and here Zweckstetter, do their usual elegant job of working out precisely what methylene blue and its metabolites do to tau, and I suppose that further testing in the various fly and mouse models might tell us whether or not that was a good thing or a bad thing. The final answer really has to come from human clinical trials, which are underway. I have not been very impressed with the data from humans that I have seen so far, but the Phase 3 studies, if they were clearly positive, would be a very strong argument that aggregation was an important event in human disease.

This paper appears in the following:


  1. Does TauRx Drug Work by Oxidizing Tau?