. β-amyloid redirects norepinephrine signaling to activate the pathogenic GSK3β/tau cascade. Sci Transl Med. 2020 Jan 15;12(526) PubMed.


Please login to recommend the paper.


  1. We have known for some time that Aβ and tau interact with each other and that Aβ can trigger and exacerbate tau pathology, but the molecular mechanisms have not been fully defined. The present study shows that Aβ oligomers (Aβo) can act as a positive allosteric modulator of α2A-adrenergic receptors (α2AAR), and that α2AAR signaling increases tau hyperphosphorylation via inappropriate engagement of the protein kinase GSK3β. This study represents an important advance in the field, particularly given the rigor of the experiments, which included cell culture and transgenic mouse models as well as postmortem human brain tissue, controls for the specificity of the Aβo-α2AAR interaction, and effects on cognition across species. Moreover, identification of the α2AAR as a mediator of Aβo-triggered tau pathology provides a new target for Alzheimer’s disease pharmacotherapies.

    These findings have important implications for both the basic neurobiology underlying Alzheimer’s disease as well as potential treatments, and suggest new directions for future research. For example, although Aβ-based transgenic mice show some tau hyperphosphorylation, wild-type mouse tau is relatively resistant to aggregation and toxicity, and it would be informative to assess the consequences of Aβo2AAR signaling on tau pathology in other models that accumulate bona-fide neurofibrillary tangles.

    Noradrenergic neurons of the brainstem locus coeruleus are among the first to display tau pathology during the early progression of Alzheimer’s and degenerate later in the disease, in part due to the effects of norepinephrine metabolism on tau cleavage and aggregation (Braak et al., 2011; Weinshenker, 2018Kang et al., 2020). Given the high expression of α2AAR in the locus coeruleus, the current study suggests that Aβo-enhanced α2AAR signaling may also contribute to the selective vulnerability of this nucleus, which would necessitate experiments specifically examining locus coeruleus neurons. Although the authors show that α2AAR antagonists can prevent Aβo-induced GSK3β activation, tau phosphorylation, and cognitive deficits in mice, these drugs are not ideal for clinical use because they block all endogenous α2AAR transmission throughout the body and have cardiovascular and anxiogenic effects.

    More detailed knowledge of the structural basis mediating the Aβo2AAR interaction via the use of x-ray crystallography, cryo-electron microscopy, and/or in silico modeling could inform the development of therapeutic molecules that specifically interfere with Aβo2AAR binding without impairing endogenous norepinephrine signaling.


    . Stages of the pathologic process in Alzheimer disease: age categories from 1 to 100 years. J Neuropathol Exp Neurol. 2011 Nov;70(11):960-9. PubMed.

    . Long Road to Ruin: Noradrenergic Dysfunction in Neurodegenerative Disease. Trends Neurosci. 2018 Apr;41(4):211-223. Epub 2018 Feb 20 PubMed.

    . Norepinephrine metabolite DOPEGAL activates AEP and pathological Tau aggregation in locus coeruleus. J Clin Invest. 2020 Jan 2;130(1):422-437. PubMed.

    View all comments by David Weinshenker
  2. The paper shows a series of logically linked experiments using human brain tissue, wild-type mice, two lines of mice producing excessive Aβ, and neuronal cultures. It demonstrates that Aβ oligomers can bind allosterically to α2AAR receptors of the norepinephrine signaling system, which redirects norepinephrine-induced α2AAR signaling to activate Glycogen synthase kinase 3 (GSK3β), known to phosphorylate tau and increase tau phosphorylation. Blockage of the proposed pathway via mutation of the Aβ oligomer allosteric binding site to α2AAR, agents to block α2AAR receptors, or GSK3 activation and others, prevent the cascade and also reverse behavioral phenotypes in animal experiments.

    This elegant study provides a possible pathway to explain the interaction between Aβ and hyperphosphorylated tau via α2AAR in the pathogenesis and progression of AD. Modulation of this pathway has the potential to slow down AD progression if the same mechanism proves to be relevant to AD in humans.

    AD-type misfolded tau first aggregates in selected subcortical neurons, followed by the trans-entorhinal and entorhinal cortices. This early tau accumulation often precedes Aβ deposits, thus constituting the first evidence of AD in humans. The distribution of AD-tau aggregates is the best correlate to neuronal loss and clinical symptoms in AD.

    Interestingly, neuropathological and radiological evidence suggest that AD-tau changes are likely to remain restricted to subcortical and limbic areas and only produce mild symptomatology until Aβ starts aggregating as plaques, first in the neocortex and later in limbic and subcortical regions. The mechanisms underlying the interaction between Aβ and AD-tau that, in turn, modulate AD-tau spread to the neocortex, remain mostly elusive.

    This paper goes beyond providing a possible explanation underlying Aβ and AD-tau interaction, but supports the importance of norepinephrine dysfunction to AD pathogenesis. Degeneration of the pontine locus coeruleus (LC), the main seat of norepinephrine-producing neurons in the brain, is an early and integral part of AD pathophysiology. The LC is one of the first brain regions to accumulate tau-AD, already at stage 0 of neurofibrillary pathology proposed by Braak and Braak. By late AD stages, only 20 percent of the original neurons survive in the LC.

    Interestingly, an early effect of LC accumulation of AD-tau is an increase in norepinephrine release and an increase in norepinephrine receptors in remote areas. Although this paper suggests a feed-forward mechanism in which Aβ oligomers make α2AAR more responsive to norepinephrine, it does not answer the question of what first caused these receptors to overreact or how early this happens in the relation to increasing levels of Aβ oligomers.

    It would be interesting to learn more about temporal LC changes, if any, in mouse models used in this study. Would it be plausible that early LC dysfunction causes changes in α2AAR that, in turn, will enter a vicious cycle with Aβ oligomers perpetuating disease progression? Also, given the robustness of the experiments, it would be essential to understand how the proposed pathway complements changes in neuroinflammatory pathways mediated by LC, or a recently reported pathway showing that the norepinephrine metabolite DOPEGAL may directly trigger tau phosphorylation.

    Finally, changes in beta-adrenergic signaling also have been reported in AD and Down syndrome, and it is worth looking into how both mechanisms interact. In summary, this paper sheds light into the mechanisms of AD acceleration and spreading and supports that modulating neurotransmitters systems may modify AD progression, rather than only target symptoms.

    View all comments by Lea T. Grinberg
  3. Different regions of the brain are differentially affected by plaque and tangle pathology, with the cerebral cortex and hippocampus being the primary targets. Subcortical regions are also affected. The noradrenergic locus coeruleus (LC) is involved in the control of complex behavior, including arousal, alertness, as well as cognitive and endocrine functions. As such, the LC is a target in many psychiatric and neurodegenerative disorders, including Alzheimer’s disease (AD). In fact, the LC is largely affected by cell death early in AD. However, since this brain region mostly lacks neuritic plaques, the prevailing view was that neuronal loss in the LC is caused by the cortical Aβ, which “poisons” the projections of brainstem cells (German et al., 2005). 

    Subsequent studies provided some evidence that the degenerating LC neurons may deprive brain regions supplied by their axons of norepinephrine, and promote AD pathogenesis in the brain regions that are primarily affected by Aβ and Tau pathology: the cortical areas, the hippocampus, the entorhinal cortex, and the frontal cortex (Heneka et al., 2006; O'Neil et al., 2007). 

    At about the same time, we proposed that the Aβ pathology in AD could actually begin in the LC, with the production of Aβ oligomers (Aβo) inside the LC neurons, and their transport throughout the entire brain via axonal transport (Muresan and Muresan, 2006; Muresan and Muresan, 2008). Our hypothesis envisioned that the pathogenic Aβo would be released from the axon terminals in the AD vulnerable brain regions, but not in the LC itself. This would explain the lack of Aβ accumulation, and lack of plaque formation, in the LC.

    Several years after we published our hypothesis, Braak and Del Tredici also arrived at the conclusion that the LC could be the site where the pathology in AD begins—and from where it spreads throughout the brain—based on tau pathology. Their studies have shown that AT8-immunoreactive abnormal tau aggregates (pretangles) develop within proximal axons of noradrenergic LC projection neurons in the absence of both tau lesions in the trans-entorhinal region as well as cortical Aβ pathology (Braak and Del Tredici, 2011; Braak and Del Tredici, 2012). Thus, LC neurons appear to be particularly vulnerable to tau pathology. Why this would be the case remained unknown. Zhang et al. now provide a possible explanation.

    This paper addresses the mechanism of preferential accumulation of AT8-immunoreactive abnormal tau aggregates in LC neurons, in the AD brain. Using mouse models of AD, AD brain samples, and data mining, the authors find that this tau pathology is initiated by low amounts of Aβo, which alter norepinephrine signaling, and lead to tau phosphorylation by activating GSK3β. The activation of the GSK3β/tau cascade requires direct binding of Aβo to α2AAR. Unexpectedly, this binding does not block binding of norepinephrine to α2AAR; moreover, activation of the pathogenic mechanism absolutely requires binding of norepinephrine to its receptor (simultaneously with Aβo, which binds to an allosteric site).

    Zhang et al.’s paper is important for several reasons, some of which are listed below. We also list some issues that remain unsolved.

    • The presented mechanism would explain the vulnerability of LC neurons in AD. The noradrenergic LC neurons express high levels of α2AAR, the target site of both the Aβo and norepinephrine.
    • The paper fully supports the amyloid cascade hypothesis, where the cascade of events leading to AD involves an initial production of Aβo. It is these oligomers that then cause the hyperphosphorylation and aggregation of tau, as well as more production of Aβ. Together, Aβ and phosphorylated tau are toxic to neurons, causing neuronal death.
    • The amyloid cascade hypothesis is still the one on which most therapeutic strategies, and clinical trials, are based today. Yet, the AD field is constantly reminded that this hypothesis may not explain the initiation and progress of AD. Clinical trials aimed at reducing Aβ levels largely fail. The authors provide a possible explanation for why these clinical trials show no effect on the progress of AD, in spite of decreasing the brain Aβ loads. According to this paper, the amount of Aβ that triggers the pathological cascade leading to tau pathology is minuscule—not much higher than the normal, physiological levels of Aβ. The authors suggest that the mechanism they identify would not be blocked by strategies aimed at reducing Aβ levels; those, the authors reason, do not reduce enough, and cannot reduce enough, the Aβ levels.
    • The paper also implies that studies that refute the amyloid cascade hypothesis on the basis that memory deficits are seen prior to the detection of Aβ pathology may not use assays sensitive enough to detect the low levels of pathogenic Aβo. In such studies, it is easy to miss the presence of the pathogenic Aβo during initial stages of the disease, to dismiss a role of Aβ in the pathology, and to discard altogether the amyloid cascade hypothesis.
    • The paper shows that mouse models of AD are not completely obsolete, and could still be useful to some extent.
    • The paper suggests that the tau pathology, not the Aβ (even at high levels), causes the memory deficits. The paper shows that the cognitive deficits in a mouse model of AD are completely restored by blocking the signaling via α2AAR (which eliminates the tau pathology), despite the presence of substantial Aβ burden.
    • The paper also suggests that blocking norepinephrine signaling through α2AAR could be an effective strategy to ameliorate pathological and cognitive deficits associated with both Aβ and tau.
    • The paper does not address what is the primary event that initiates AD, i.e., where the initial pool of Aβo—even infinitesimal in amount—comes from. Is this Aβ a deviant form of the normal Aβ released in low quantities throughout life? Is there some other event that initiates the production of pathogenic Aβo? Which is the “primordial” event that initiates the disease? The authors avoid addressing the origin of the Aβo which initiate the LC pathology. Could the Aβo be generated by the LC neurons themselves? Our published data (Muresan and Muresan, 2006; Muresan and Muresan, 2008) show that neuronal cells originating in the LC copiously produce and release Aβo.
    • Certainly, there are other mechanisms that lead to tau pathology, in addition to the one described in this paper. Yet, the authors report that the memory deficits in two mouse models of AD are significantly corrected just by blocking this aberrant, norepinephrine signaling pathway.
    • The proposed treatment strategy, aimed at blocking the redirected norepinephrine signaling, will need to be eventually tested in clinical trials. In the end, it is only the clinical trials that validate—or invalidate—hypotheses and therapeutic strategies. As we all know, it is easy to cure the disease in mouse models of AD; it has been done in uncountable studies. The success of a therapeutic strategy in mice is encouraging, but in no way assures that it will work in humans. This is probably because none of the current mouse models of AD reproduce the human disease. Not even remotely.


    . The PDAPP mouse model of Alzheimer's disease: locus coeruleus neuronal shrinkage. J Comp Neurol. 2005 Nov 28;492(4):469-76. PubMed.

    . Locus ceruleus degeneration promotes Alzheimer pathogenesis in amyloid precursor protein 23 transgenic mice. J Neurosci. 2006 Feb 1;26(5):1343-54. PubMed.

    . Catecholaminergic neuronal loss in locus coeruleus of aged female dtg APP/PS1 mice. J Chem Neuroanat. 2007 Nov;34(3-4):102-7. PubMed.

    . Neuritic deposits of amyloid-beta peptide in a subpopulation of central nervous system-derived neuronal cells. Mol Cell Biol. 2006 Jul;26(13):4982-97. PubMed.

    . Seeding neuritic plaques from the distance: a possible role for brainstem neurons in the development of Alzheimer's disease pathology. Neurodegener Dis. 2008;5(3-4):250-3. Epub 2008 Mar 6 PubMed.

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

    . Where, when, and in what form does sporadic Alzheimer's disease begin?. Curr Opin Neurol. 2012 Dec;25(6):708-14. PubMed.

    View all comments by Virgil Muresan

Make a Comment

To make a comment you must login or register.

This paper appears in the following:


  1. With a Shot of Adrenaline, Amyloid-β Sparks Tau Cascade